diff options
Diffstat (limited to 'kernel')
184 files changed, 38784 insertions, 13664 deletions
diff --git a/kernel/Makefile b/kernel/Makefile index 170a9213c1b..b833bd5cc12 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -11,6 +11,7 @@ obj-y = sched.o fork.o exec_domain.o panic.o printk.o \ hrtimer.o rwsem.o nsproxy.o srcu.o semaphore.o \ notifier.o ksysfs.o pm_qos_params.o sched_clock.o cred.o \ async.o +obj-y += groups.o ifdef CONFIG_FUNCTION_TRACER # Do not trace debug files and internal ftrace files @@ -51,6 +52,7 @@ obj-$(CONFIG_UID16) += uid16.o obj-$(CONFIG_MODULES) += module.o obj-$(CONFIG_KALLSYMS) += kallsyms.o obj-$(CONFIG_PM) += power/ +obj-$(CONFIG_FREEZER) += power/ obj-$(CONFIG_BSD_PROCESS_ACCT) += acct.o obj-$(CONFIG_KEXEC) += kexec.o obj-$(CONFIG_BACKTRACE_SELF_TEST) += backtracetest.o @@ -67,20 +69,20 @@ obj-$(CONFIG_IKCONFIG) += configs.o obj-$(CONFIG_RESOURCE_COUNTERS) += res_counter.o obj-$(CONFIG_STOP_MACHINE) += stop_machine.o obj-$(CONFIG_KPROBES_SANITY_TEST) += test_kprobes.o -obj-$(CONFIG_AUDIT) += audit.o auditfilter.o +obj-$(CONFIG_AUDIT) += audit.o auditfilter.o audit_watch.o obj-$(CONFIG_AUDITSYSCALL) += auditsc.o +obj-$(CONFIG_GCOV_KERNEL) += gcov/ obj-$(CONFIG_AUDIT_TREE) += audit_tree.o obj-$(CONFIG_KPROBES) += kprobes.o obj-$(CONFIG_KGDB) += kgdb.o obj-$(CONFIG_DETECT_SOFTLOCKUP) += softlockup.o +obj-$(CONFIG_DETECT_HUNG_TASK) += hung_task.o obj-$(CONFIG_GENERIC_HARDIRQS) += irq/ obj-$(CONFIG_SECCOMP) += seccomp.o obj-$(CONFIG_RCU_TORTURE_TEST) += rcutorture.o -obj-$(CONFIG_CLASSIC_RCU) += rcuclassic.o obj-$(CONFIG_TREE_RCU) += rcutree.o -obj-$(CONFIG_PREEMPT_RCU) += rcupreempt.o +obj-$(CONFIG_TREE_PREEMPT_RCU) += rcutree.o obj-$(CONFIG_TREE_RCU_TRACE) += rcutree_trace.o -obj-$(CONFIG_PREEMPT_RCU_TRACE) += rcupreempt_trace.o obj-$(CONFIG_RELAY) += relay.o obj-$(CONFIG_SYSCTL) += utsname_sysctl.o obj-$(CONFIG_TASK_DELAY_ACCT) += delayacct.o @@ -91,7 +93,11 @@ obj-$(CONFIG_LATENCYTOP) += latencytop.o obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o obj-$(CONFIG_FUNCTION_TRACER) += trace/ obj-$(CONFIG_TRACING) += trace/ +obj-$(CONFIG_X86_DS) += trace/ +obj-$(CONFIG_RING_BUFFER) += trace/ obj-$(CONFIG_SMP) += sched_cpupri.o +obj-$(CONFIG_SLOW_WORK) += slow-work.o +obj-$(CONFIG_PERF_COUNTERS) += perf_counter.o ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) # According to Alan Modra <alan@linuxcare.com.au>, the -fno-omit-frame-pointer is diff --git a/kernel/acct.c b/kernel/acct.c index 7afa3156416..9a4715a2f6b 100644 --- a/kernel/acct.c +++ b/kernel/acct.c @@ -215,6 +215,7 @@ static void acct_file_reopen(struct bsd_acct_struct *acct, struct file *file, static int acct_on(char *name) { struct file *file; + struct vfsmount *mnt; int error; struct pid_namespace *ns; struct bsd_acct_struct *acct = NULL; @@ -256,11 +257,12 @@ static int acct_on(char *name) acct = NULL; } - mnt_pin(file->f_path.mnt); + mnt = file->f_path.mnt; + mnt_pin(mnt); acct_file_reopen(ns->bacct, file, ns); spin_unlock(&acct_lock); - mntput(file->f_path.mnt); /* it's pinned, now give up active reference */ + mntput(mnt); /* it's pinned, now give up active reference */ kfree(acct); return 0; @@ -489,13 +491,17 @@ static void do_acct_process(struct bsd_acct_struct *acct, u64 run_time; struct timespec uptime; struct tty_struct *tty; + const struct cred *orig_cred; + + /* Perform file operations on behalf of whoever enabled accounting */ + orig_cred = override_creds(file->f_cred); /* * First check to see if there is enough free_space to continue * the process accounting system. */ if (!check_free_space(acct, file)) - return; + goto out; /* * Fill the accounting struct with the needed info as recorded @@ -576,6 +582,8 @@ static void do_acct_process(struct bsd_acct_struct *acct, sizeof(acct_t), &file->f_pos); current->signal->rlim[RLIMIT_FSIZE].rlim_cur = flim; set_fs(fs); +out: + revert_creds(orig_cred); } /** diff --git a/kernel/async.c b/kernel/async.c index 608b32b4281..27235f5de19 100644 --- a/kernel/async.c +++ b/kernel/async.c @@ -49,11 +49,13 @@ asynchronous and synchronous parts of the kernel. */ #include <linux/async.h> +#include <linux/bug.h> #include <linux/module.h> #include <linux/wait.h> #include <linux/sched.h> #include <linux/init.h> #include <linux/kthread.h> +#include <linux/delay.h> #include <asm/atomic.h> static async_cookie_t next_cookie = 1; @@ -90,19 +92,18 @@ extern int initcall_debug; static async_cookie_t __lowest_in_progress(struct list_head *running) { struct async_entry *entry; + if (!list_empty(running)) { entry = list_first_entry(running, struct async_entry, list); return entry->cookie; - } else if (!list_empty(&async_pending)) { - entry = list_first_entry(&async_pending, - struct async_entry, list); - return entry->cookie; - } else { - /* nothing in progress... next_cookie is "infinity" */ - return next_cookie; } + list_for_each_entry(entry, &async_pending, list) + if (entry->running == running) + return entry->cookie; + + return next_cookie; /* "infinity" value */ } static async_cookie_t lowest_in_progress(struct list_head *running) @@ -132,21 +133,23 @@ static void run_one_entry(void) entry = list_first_entry(&async_pending, struct async_entry, list); /* 2) move it to the running queue */ - list_del(&entry->list); - list_add_tail(&entry->list, &async_running); + list_move_tail(&entry->list, entry->running); spin_unlock_irqrestore(&async_lock, flags); /* 3) run it (and print duration)*/ if (initcall_debug && system_state == SYSTEM_BOOTING) { - printk("calling %lli_%pF @ %i\n", entry->cookie, entry->func, task_pid_nr(current)); + printk("calling %lli_%pF @ %i\n", (long long)entry->cookie, + entry->func, task_pid_nr(current)); calltime = ktime_get(); } entry->func(entry->data, entry->cookie); if (initcall_debug && system_state == SYSTEM_BOOTING) { rettime = ktime_get(); delta = ktime_sub(rettime, calltime); - printk("initcall %lli_%pF returned 0 after %lld usecs\n", entry->cookie, - entry->func, ktime_to_ns(delta) >> 10); + printk("initcall %lli_%pF returned 0 after %lld usecs\n", + (long long)entry->cookie, + entry->func, + (long long)ktime_to_ns(delta) >> 10); } /* 4) remove it from the running queue */ @@ -205,18 +208,44 @@ static async_cookie_t __async_schedule(async_func_ptr *ptr, void *data, struct l return newcookie; } +/** + * async_schedule - schedule a function for asynchronous execution + * @ptr: function to execute asynchronously + * @data: data pointer to pass to the function + * + * Returns an async_cookie_t that may be used for checkpointing later. + * Note: This function may be called from atomic or non-atomic contexts. + */ async_cookie_t async_schedule(async_func_ptr *ptr, void *data) { - return __async_schedule(ptr, data, &async_pending); + return __async_schedule(ptr, data, &async_running); } EXPORT_SYMBOL_GPL(async_schedule); -async_cookie_t async_schedule_special(async_func_ptr *ptr, void *data, struct list_head *running) +/** + * async_schedule_domain - schedule a function for asynchronous execution within a certain domain + * @ptr: function to execute asynchronously + * @data: data pointer to pass to the function + * @running: running list for the domain + * + * Returns an async_cookie_t that may be used for checkpointing later. + * @running may be used in the async_synchronize_*_domain() functions + * to wait within a certain synchronization domain rather than globally. + * A synchronization domain is specified via the running queue @running to use. + * Note: This function may be called from atomic or non-atomic contexts. + */ +async_cookie_t async_schedule_domain(async_func_ptr *ptr, void *data, + struct list_head *running) { return __async_schedule(ptr, data, running); } -EXPORT_SYMBOL_GPL(async_schedule_special); +EXPORT_SYMBOL_GPL(async_schedule_domain); +/** + * async_synchronize_full - synchronize all asynchronous function calls + * + * This function waits until all asynchronous function calls have been done. + */ void async_synchronize_full(void) { do { @@ -225,13 +254,30 @@ void async_synchronize_full(void) } EXPORT_SYMBOL_GPL(async_synchronize_full); -void async_synchronize_full_special(struct list_head *list) +/** + * async_synchronize_full_domain - synchronize all asynchronous function within a certain domain + * @list: running list to synchronize on + * + * This function waits until all asynchronous function calls for the + * synchronization domain specified by the running list @list have been done. + */ +void async_synchronize_full_domain(struct list_head *list) { - async_synchronize_cookie_special(next_cookie, list); + async_synchronize_cookie_domain(next_cookie, list); } -EXPORT_SYMBOL_GPL(async_synchronize_full_special); +EXPORT_SYMBOL_GPL(async_synchronize_full_domain); -void async_synchronize_cookie_special(async_cookie_t cookie, struct list_head *running) +/** + * async_synchronize_cookie_domain - synchronize asynchronous function calls within a certain domain with cookie checkpointing + * @cookie: async_cookie_t to use as checkpoint + * @running: running list to synchronize on + * + * This function waits until all asynchronous function calls for the + * synchronization domain specified by the running list @list submitted + * prior to @cookie have been done. + */ +void async_synchronize_cookie_domain(async_cookie_t cookie, + struct list_head *running) { ktime_t starttime, delta, endtime; @@ -247,14 +293,22 @@ void async_synchronize_cookie_special(async_cookie_t cookie, struct list_head *r delta = ktime_sub(endtime, starttime); printk("async_continuing @ %i after %lli usec\n", - task_pid_nr(current), ktime_to_ns(delta) >> 10); + task_pid_nr(current), + (long long)ktime_to_ns(delta) >> 10); } } -EXPORT_SYMBOL_GPL(async_synchronize_cookie_special); +EXPORT_SYMBOL_GPL(async_synchronize_cookie_domain); +/** + * async_synchronize_cookie - synchronize asynchronous function calls with cookie checkpointing + * @cookie: async_cookie_t to use as checkpoint + * + * This function waits until all asynchronous function calls prior to @cookie + * have been done. + */ void async_synchronize_cookie(async_cookie_t cookie) { - async_synchronize_cookie_special(cookie, &async_running); + async_synchronize_cookie_domain(cookie, &async_running); } EXPORT_SYMBOL_GPL(async_synchronize_cookie); @@ -315,7 +369,11 @@ static int async_manager_thread(void *unused) ec = atomic_read(&entry_count); while (tc < ec && tc < MAX_THREADS) { - kthread_run(async_thread, NULL, "async/%i", tc); + if (IS_ERR(kthread_run(async_thread, NULL, "async/%i", + tc))) { + msleep(100); + continue; + } atomic_inc(&thread_count); tc++; } @@ -329,18 +387,11 @@ static int async_manager_thread(void *unused) static int __init async_init(void) { - if (async_enabled) - kthread_run(async_manager_thread, NULL, "async/mgr"); - return 0; -} + async_enabled = + !IS_ERR(kthread_run(async_manager_thread, NULL, "async/mgr")); -static int __init setup_async(char *str) -{ - async_enabled = 1; - return 1; + WARN_ON(!async_enabled); + return 0; } -__setup("fastboot", setup_async); - - core_initcall(async_init); diff --git a/kernel/audit.c b/kernel/audit.c index ce6d8ea3131..defc2e6f1e3 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -115,9 +115,6 @@ static atomic_t audit_lost = ATOMIC_INIT(0); /* The netlink socket. */ static struct sock *audit_sock; -/* Inotify handle. */ -struct inotify_handle *audit_ih; - /* Hash for inode-based rules */ struct list_head audit_inode_hash[AUDIT_INODE_BUCKETS]; @@ -136,7 +133,7 @@ static DECLARE_WAIT_QUEUE_HEAD(kauditd_wait); static DECLARE_WAIT_QUEUE_HEAD(audit_backlog_wait); /* Serialize requests from userspace. */ -static DEFINE_MUTEX(audit_cmd_mutex); +DEFINE_MUTEX(audit_cmd_mutex); /* AUDIT_BUFSIZ is the size of the temporary buffer used for formatting * audit records. Since printk uses a 1024 byte buffer, this buffer @@ -375,6 +372,25 @@ static void audit_hold_skb(struct sk_buff *skb) kfree_skb(skb); } +/* + * For one reason or another this nlh isn't getting delivered to the userspace + * audit daemon, just send it to printk. + */ +static void audit_printk_skb(struct sk_buff *skb) +{ + struct nlmsghdr *nlh = nlmsg_hdr(skb); + char *data = NLMSG_DATA(nlh); + + if (nlh->nlmsg_type != AUDIT_EOE) { + if (printk_ratelimit()) + printk(KERN_NOTICE "type=%d %s\n", nlh->nlmsg_type, data); + else + audit_log_lost("printk limit exceeded\n"); + } + + audit_hold_skb(skb); +} + static void kauditd_send_skb(struct sk_buff *skb) { int err; @@ -427,14 +443,8 @@ static int kauditd_thread(void *dummy) if (skb) { if (audit_pid) kauditd_send_skb(skb); - else { - if (printk_ratelimit()) - printk(KERN_NOTICE "%s\n", skb->data + NLMSG_SPACE(0)); - else - audit_log_lost("printk limit exceeded\n"); - - audit_hold_skb(skb); - } + else + audit_printk_skb(skb); } else { DECLARE_WAITQUEUE(wait, current); set_current_state(TASK_INTERRUPTIBLE); @@ -495,42 +505,25 @@ int audit_send_list(void *_dest) return 0; } -#ifdef CONFIG_AUDIT_TREE -static int prune_tree_thread(void *unused) -{ - mutex_lock(&audit_cmd_mutex); - audit_prune_trees(); - mutex_unlock(&audit_cmd_mutex); - return 0; -} - -void audit_schedule_prune(void) -{ - kthread_run(prune_tree_thread, NULL, "audit_prune_tree"); -} -#endif - struct sk_buff *audit_make_reply(int pid, int seq, int type, int done, int multi, void *payload, int size) { struct sk_buff *skb; struct nlmsghdr *nlh; - int len = NLMSG_SPACE(size); void *data; int flags = multi ? NLM_F_MULTI : 0; int t = done ? NLMSG_DONE : type; - skb = alloc_skb(len, GFP_KERNEL); + skb = nlmsg_new(size, GFP_KERNEL); if (!skb) return NULL; - nlh = NLMSG_PUT(skb, pid, seq, t, size); - nlh->nlmsg_flags = flags; - data = NLMSG_DATA(nlh); + nlh = NLMSG_NEW(skb, pid, seq, t, size, flags); + data = NLMSG_DATA(nlh); memcpy(data, payload, size); return skb; -nlmsg_failure: /* Used by NLMSG_PUT */ +nlmsg_failure: /* Used by NLMSG_NEW */ if (skb) kfree_skb(skb); return NULL; @@ -766,6 +759,9 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) audit_log_format(ab, " msg="); size = nlmsg_len(nlh); + if (size > 0 && + ((unsigned char *)data)[size - 1] == '\0') + size--; audit_log_n_untrustedstring(ab, data, size); } audit_set_pid(ab, pid); @@ -923,28 +919,29 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh) } /* - * Get message from skb (based on rtnetlink_rcv_skb). Each message is - * processed by audit_receive_msg. Malformed skbs with wrong length are - * discarded silently. + * Get message from skb. Each message is processed by audit_receive_msg. + * Malformed skbs with wrong length are discarded silently. */ static void audit_receive_skb(struct sk_buff *skb) { - int err; - struct nlmsghdr *nlh; - u32 rlen; + struct nlmsghdr *nlh; + /* + * len MUST be signed for NLMSG_NEXT to be able to dec it below 0 + * if the nlmsg_len was not aligned + */ + int len; + int err; - while (skb->len >= NLMSG_SPACE(0)) { - nlh = nlmsg_hdr(skb); - if (nlh->nlmsg_len < sizeof(*nlh) || skb->len < nlh->nlmsg_len) - return; - rlen = NLMSG_ALIGN(nlh->nlmsg_len); - if (rlen > skb->len) - rlen = skb->len; - if ((err = audit_receive_msg(skb, nlh))) { + nlh = nlmsg_hdr(skb); + len = skb->len; + + while (NLMSG_OK(nlh, len)) { + err = audit_receive_msg(skb, nlh); + /* if err or if this message says it wants a response */ + if (err || (nlh->nlmsg_flags & NLM_F_ACK)) netlink_ack(skb, nlh, err); - } else if (nlh->nlmsg_flags & NLM_F_ACK) - netlink_ack(skb, nlh, 0); - skb_pull(skb, rlen); + + nlh = NLMSG_NEXT(nlh, len); } } @@ -956,13 +953,6 @@ static void audit_receive(struct sk_buff *skb) mutex_unlock(&audit_cmd_mutex); } -#ifdef CONFIG_AUDITSYSCALL -static const struct inotify_operations audit_inotify_ops = { - .handle_event = audit_handle_ievent, - .destroy_watch = audit_free_parent, -}; -#endif - /* Initialize audit support at boot time. */ static int __init audit_init(void) { @@ -988,12 +978,6 @@ static int __init audit_init(void) audit_log(NULL, GFP_KERNEL, AUDIT_KERNEL, "initialized"); -#ifdef CONFIG_AUDITSYSCALL - audit_ih = inotify_init(&audit_inotify_ops); - if (IS_ERR(audit_ih)) - audit_panic("cannot initialize inotify handle"); -#endif - for (i = 0; i < AUDIT_INODE_BUCKETS; i++) INIT_LIST_HEAD(&audit_inode_hash[i]); @@ -1067,18 +1051,20 @@ static struct audit_buffer * audit_buffer_alloc(struct audit_context *ctx, goto err; } - ab->skb = alloc_skb(AUDIT_BUFSIZ, gfp_mask); - if (!ab->skb) - goto err; - ab->ctx = ctx; ab->gfp_mask = gfp_mask; - nlh = (struct nlmsghdr *)skb_put(ab->skb, NLMSG_SPACE(0)); - nlh->nlmsg_type = type; - nlh->nlmsg_flags = 0; - nlh->nlmsg_pid = 0; - nlh->nlmsg_seq = 0; + + ab->skb = nlmsg_new(AUDIT_BUFSIZ, gfp_mask); + if (!ab->skb) + goto nlmsg_failure; + + nlh = NLMSG_NEW(ab->skb, 0, 0, type, 0, 0); + return ab; + +nlmsg_failure: /* Used by NLMSG_NEW */ + kfree_skb(ab->skb); + ab->skb = NULL; err: audit_buffer_free(ab); return NULL; @@ -1382,7 +1368,7 @@ void audit_log_n_string(struct audit_buffer *ab, const char *string, int audit_string_contains_control(const char *string, size_t len) { const unsigned char *p; - for (p = string; p < (const unsigned char *)string + len && *p; p++) { + for (p = string; p < (const unsigned char *)string + len; p++) { if (*p == '"' || *p < 0x21 || *p > 0x7e) return 1; } @@ -1437,18 +1423,27 @@ void audit_log_d_path(struct audit_buffer *ab, const char *prefix, /* We will allow 11 spaces for ' (deleted)' to be appended */ pathname = kmalloc(PATH_MAX+11, ab->gfp_mask); if (!pathname) { - audit_log_format(ab, "<no memory>"); + audit_log_string(ab, "<no_memory>"); return; } p = d_path(path, pathname, PATH_MAX+11); if (IS_ERR(p)) { /* Should never happen since we send PATH_MAX */ /* FIXME: can we save some information here? */ - audit_log_format(ab, "<too long>"); + audit_log_string(ab, "<too_long>"); } else audit_log_untrustedstring(ab, p); kfree(pathname); } +void audit_log_key(struct audit_buffer *ab, char *key) +{ + audit_log_format(ab, " key="); + if (key) + audit_log_untrustedstring(ab, key); + else + audit_log_format(ab, "(null)"); +} + /** * audit_log_end - end one audit record * @ab: the audit_buffer @@ -1472,15 +1467,7 @@ void audit_log_end(struct audit_buffer *ab) skb_queue_tail(&audit_skb_queue, ab->skb); wake_up_interruptible(&kauditd_wait); } else { - if (nlh->nlmsg_type != AUDIT_EOE) { - if (printk_ratelimit()) { - printk(KERN_NOTICE "type=%d %s\n", - nlh->nlmsg_type, - ab->skb->data + NLMSG_SPACE(0)); - } else - audit_log_lost("printk limit exceeded\n"); - } - audit_hold_skb(ab->skb); + audit_printk_skb(ab->skb); } ab->skb = NULL; } diff --git a/kernel/audit.h b/kernel/audit.h index 16f18cac661..208687be4f3 100644 --- a/kernel/audit.h +++ b/kernel/audit.h @@ -53,18 +53,7 @@ enum audit_state { }; /* Rule lists */ -struct audit_parent; - -struct audit_watch { - atomic_t count; /* reference count */ - char *path; /* insertion path */ - dev_t dev; /* associated superblock device */ - unsigned long ino; /* associated inode number */ - struct audit_parent *parent; /* associated parent */ - struct list_head wlist; /* entry in parent->watches list */ - struct list_head rules; /* associated rules */ -}; - +struct audit_watch; struct audit_tree; struct audit_chunk; @@ -108,19 +97,28 @@ struct audit_netlink_list { int audit_send_list(void *); -struct inotify_watch; -/* Inotify handle */ -extern struct inotify_handle *audit_ih; - -extern void audit_free_parent(struct inotify_watch *); -extern void audit_handle_ievent(struct inotify_watch *, u32, u32, u32, - const char *, struct inode *); extern int selinux_audit_rule_update(void); extern struct mutex audit_filter_mutex; extern void audit_free_rule_rcu(struct rcu_head *); extern struct list_head audit_filter_list[]; +/* audit watch functions */ +extern unsigned long audit_watch_inode(struct audit_watch *watch); +extern dev_t audit_watch_dev(struct audit_watch *watch); +extern void audit_put_watch(struct audit_watch *watch); +extern void audit_get_watch(struct audit_watch *watch); +extern int audit_to_watch(struct audit_krule *krule, char *path, int len, u32 op); +extern int audit_add_watch(struct audit_krule *krule); +extern void audit_remove_watch(struct audit_watch *watch); +extern void audit_remove_watch_rule(struct audit_krule *krule, struct list_head *list); +extern void audit_inotify_unregister(struct list_head *in_list); +extern char *audit_watch_path(struct audit_watch *watch); +extern struct list_head *audit_watch_rules(struct audit_watch *watch); + +extern struct audit_entry *audit_dupe_rule(struct audit_krule *old, + struct audit_watch *watch); + #ifdef CONFIG_AUDIT_TREE extern struct audit_chunk *audit_tree_lookup(const struct inode *); extern void audit_put_chunk(struct audit_chunk *); @@ -130,10 +128,9 @@ extern int audit_add_tree_rule(struct audit_krule *); extern int audit_remove_tree_rule(struct audit_krule *); extern void audit_trim_trees(void); extern int audit_tag_tree(char *old, char *new); -extern void audit_schedule_prune(void); -extern void audit_prune_trees(void); extern const char *audit_tree_path(struct audit_tree *); extern void audit_put_tree(struct audit_tree *); +extern void audit_kill_trees(struct list_head *); #else #define audit_remove_tree_rule(rule) BUG() #define audit_add_tree_rule(rule) -EINVAL @@ -142,6 +139,7 @@ extern void audit_put_tree(struct audit_tree *); #define audit_put_tree(tree) (void)0 #define audit_tag_tree(old, new) -EINVAL #define audit_tree_path(rule) "" /* never called */ +#define audit_kill_trees(list) BUG() #endif extern char *audit_unpack_string(void **, size_t *, size_t); @@ -160,7 +158,10 @@ static inline int audit_signal_info(int sig, struct task_struct *t) return 0; } extern void audit_filter_inodes(struct task_struct *, struct audit_context *); +extern struct list_head *audit_killed_trees(void); #else #define audit_signal_info(s,t) AUDIT_DISABLED #define audit_filter_inodes(t,c) AUDIT_DISABLED #endif + +extern struct mutex audit_cmd_mutex; diff --git a/kernel/audit_tree.c b/kernel/audit_tree.c index 8ad9545b8db..2451dc6f328 100644 --- a/kernel/audit_tree.c +++ b/kernel/audit_tree.c @@ -2,6 +2,7 @@ #include <linux/inotify.h> #include <linux/namei.h> #include <linux/mount.h> +#include <linux/kthread.h> struct audit_tree; struct audit_chunk; @@ -385,6 +386,7 @@ static int tag_chunk(struct inode *inode, struct audit_tree *tree) mutex_lock(&inode->inotify_mutex); if (inotify_clone_watch(&old->watch, &chunk->watch) < 0) { mutex_unlock(&inode->inotify_mutex); + put_inotify_watch(&old->watch); free_chunk(chunk); return -ENOSPC; } @@ -394,6 +396,7 @@ static int tag_chunk(struct inode *inode, struct audit_tree *tree) chunk->dead = 1; inotify_evict_watch(&chunk->watch); mutex_unlock(&inode->inotify_mutex); + put_inotify_watch(&old->watch); put_inotify_watch(&chunk->watch); return 0; } @@ -439,13 +442,11 @@ static void kill_rules(struct audit_tree *tree) if (rule->tree) { /* not a half-baked one */ ab = audit_log_start(NULL, GFP_KERNEL, AUDIT_CONFIG_CHANGE); - audit_log_format(ab, "op=remove rule dir="); + audit_log_format(ab, "op="); + audit_log_string(ab, "remove rule"); + audit_log_format(ab, " dir="); audit_log_untrustedstring(ab, rule->tree->pathname); - if (rule->filterkey) { - audit_log_format(ab, " key="); - audit_log_untrustedstring(ab, rule->filterkey); - } else - audit_log_format(ab, " key=(null)"); + audit_log_key(ab, rule->filterkey); audit_log_format(ab, " list=%d res=1", rule->listnr); audit_log_end(ab); rule->tree = NULL; @@ -517,6 +518,8 @@ static void trim_marked(struct audit_tree *tree) } } +static void audit_schedule_prune(void); + /* called with audit_filter_mutex */ int audit_remove_tree_rule(struct audit_krule *rule) { @@ -566,7 +569,7 @@ void audit_trim_trees(void) if (err) goto skip_it; - root_mnt = collect_mounts(path.mnt, path.dentry); + root_mnt = collect_mounts(&path); path_put(&path); if (!root_mnt) goto skip_it; @@ -658,7 +661,7 @@ int audit_add_tree_rule(struct audit_krule *rule) err = kern_path(tree->pathname, 0, &path); if (err) goto Err; - mnt = collect_mounts(path.mnt, path.dentry); + mnt = collect_mounts(&path); path_put(&path); if (!mnt) { err = -ENOMEM; @@ -718,7 +721,7 @@ int audit_tag_tree(char *old, char *new) err = kern_path(new, 0, &path); if (err) return err; - tagged = collect_mounts(path.mnt, path.dentry); + tagged = collect_mounts(&path); path_put(&path); if (!tagged) return -ENOMEM; @@ -732,9 +735,6 @@ int audit_tag_tree(char *old, char *new) dentry = dget(path.dentry); path_put(&path); - if (dentry == tagged->mnt_root && dentry == mnt->mnt_root) - follow_up(&mnt, &dentry); - list_add_tail(&list, &tagged->mnt_list); mutex_lock(&audit_filter_mutex); @@ -825,10 +825,11 @@ int audit_tag_tree(char *old, char *new) /* * That gets run when evict_chunk() ends up needing to kill audit_tree. - * Runs from a separate thread, with audit_cmd_mutex held. + * Runs from a separate thread. */ -void audit_prune_trees(void) +static int prune_tree_thread(void *unused) { + mutex_lock(&audit_cmd_mutex); mutex_lock(&audit_filter_mutex); while (!list_empty(&prune_list)) { @@ -845,6 +846,40 @@ void audit_prune_trees(void) } mutex_unlock(&audit_filter_mutex); + mutex_unlock(&audit_cmd_mutex); + return 0; +} + +static void audit_schedule_prune(void) +{ + kthread_run(prune_tree_thread, NULL, "audit_prune_tree"); +} + +/* + * ... and that one is done if evict_chunk() decides to delay until the end + * of syscall. Runs synchronously. + */ +void audit_kill_trees(struct list_head *list) +{ + mutex_lock(&audit_cmd_mutex); + mutex_lock(&audit_filter_mutex); + + while (!list_empty(list)) { + struct audit_tree *victim; + + victim = list_entry(list->next, struct audit_tree, list); + kill_rules(victim); + list_del_init(&victim->list); + + mutex_unlock(&audit_filter_mutex); + + prune_one(victim); + + mutex_lock(&audit_filter_mutex); + } + + mutex_unlock(&audit_filter_mutex); + mutex_unlock(&audit_cmd_mutex); } /* @@ -855,6 +890,8 @@ void audit_prune_trees(void) static void evict_chunk(struct audit_chunk *chunk) { struct audit_tree *owner; + struct list_head *postponed = audit_killed_trees(); + int need_prune = 0; int n; if (chunk->dead) @@ -870,15 +907,21 @@ static void evict_chunk(struct audit_chunk *chunk) owner->root = NULL; list_del_init(&owner->same_root); spin_unlock(&hash_lock); - kill_rules(owner); - list_move(&owner->list, &prune_list); - audit_schedule_prune(); + if (!postponed) { + kill_rules(owner); + list_move(&owner->list, &prune_list); + need_prune = 1; + } else { + list_move(&owner->list, postponed); + } spin_lock(&hash_lock); } list_del_rcu(&chunk->hash); for (n = 0; n < chunk->count; n++) list_del_init(&chunk->owners[n].list); spin_unlock(&hash_lock); + if (need_prune) + audit_schedule_prune(); mutex_unlock(&audit_filter_mutex); } diff --git a/kernel/audit_watch.c b/kernel/audit_watch.c new file mode 100644 index 00000000000..0e96dbc60ea --- /dev/null +++ b/kernel/audit_watch.c @@ -0,0 +1,543 @@ +/* audit_watch.c -- watching inodes + * + * Copyright 2003-2009 Red Hat, Inc. + * Copyright 2005 Hewlett-Packard Development Company, L.P. + * Copyright 2005 IBM Corporation + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#include <linux/kernel.h> +#include <linux/audit.h> +#include <linux/kthread.h> +#include <linux/mutex.h> +#include <linux/fs.h> +#include <linux/namei.h> +#include <linux/netlink.h> +#include <linux/sched.h> +#include <linux/inotify.h> +#include <linux/security.h> +#include "audit.h" + +/* + * Reference counting: + * + * audit_parent: lifetime is from audit_init_parent() to receipt of an IN_IGNORED + * event. Each audit_watch holds a reference to its associated parent. + * + * audit_watch: if added to lists, lifetime is from audit_init_watch() to + * audit_remove_watch(). Additionally, an audit_watch may exist + * temporarily to assist in searching existing filter data. Each + * audit_krule holds a reference to its associated watch. + */ + +struct audit_watch { + atomic_t count; /* reference count */ + char *path; /* insertion path */ + dev_t dev; /* associated superblock device */ + unsigned long ino; /* associated inode number */ + struct audit_parent *parent; /* associated parent */ + struct list_head wlist; /* entry in parent->watches list */ + struct list_head rules; /* associated rules */ +}; + +struct audit_parent { + struct list_head ilist; /* entry in inotify registration list */ + struct list_head watches; /* associated watches */ + struct inotify_watch wdata; /* inotify watch data */ + unsigned flags; /* status flags */ +}; + +/* Inotify handle. */ +struct inotify_handle *audit_ih; + +/* + * audit_parent status flags: + * + * AUDIT_PARENT_INVALID - set anytime rules/watches are auto-removed due to + * a filesystem event to ensure we're adding audit watches to a valid parent. + * Technically not needed for IN_DELETE_SELF or IN_UNMOUNT events, as we cannot + * receive them while we have nameidata, but must be used for IN_MOVE_SELF which + * we can receive while holding nameidata. + */ +#define AUDIT_PARENT_INVALID 0x001 + +/* Inotify events we care about. */ +#define AUDIT_IN_WATCH IN_MOVE|IN_CREATE|IN_DELETE|IN_DELETE_SELF|IN_MOVE_SELF + +static void audit_free_parent(struct inotify_watch *i_watch) +{ + struct audit_parent *parent; + + parent = container_of(i_watch, struct audit_parent, wdata); + WARN_ON(!list_empty(&parent->watches)); + kfree(parent); +} + +void audit_get_watch(struct audit_watch *watch) +{ + atomic_inc(&watch->count); +} + +void audit_put_watch(struct audit_watch *watch) +{ + if (atomic_dec_and_test(&watch->count)) { + WARN_ON(watch->parent); + WARN_ON(!list_empty(&watch->rules)); + kfree(watch->path); + kfree(watch); + } +} + +void audit_remove_watch(struct audit_watch *watch) +{ + list_del(&watch->wlist); + put_inotify_watch(&watch->parent->wdata); + watch->parent = NULL; + audit_put_watch(watch); /* match initial get */ +} + +char *audit_watch_path(struct audit_watch *watch) +{ + return watch->path; +} + +struct list_head *audit_watch_rules(struct audit_watch *watch) +{ + return &watch->rules; +} + +unsigned long audit_watch_inode(struct audit_watch *watch) +{ + return watch->ino; +} + +dev_t audit_watch_dev(struct audit_watch *watch) +{ + return watch->dev; +} + +/* Initialize a parent watch entry. */ +static struct audit_parent *audit_init_parent(struct nameidata *ndp) +{ + struct audit_parent *parent; + s32 wd; + + parent = kzalloc(sizeof(*parent), GFP_KERNEL); + if (unlikely(!parent)) + return ERR_PTR(-ENOMEM); + + INIT_LIST_HEAD(&parent->watches); + parent->flags = 0; + + inotify_init_watch(&parent->wdata); + /* grab a ref so inotify watch hangs around until we take audit_filter_mutex */ + get_inotify_watch(&parent->wdata); + wd = inotify_add_watch(audit_ih, &parent->wdata, + ndp->path.dentry->d_inode, AUDIT_IN_WATCH); + if (wd < 0) { + audit_free_parent(&parent->wdata); + return ERR_PTR(wd); + } + + return parent; +} + +/* Initialize a watch entry. */ +static struct audit_watch *audit_init_watch(char *path) +{ + struct audit_watch *watch; + + watch = kzalloc(sizeof(*watch), GFP_KERNEL); + if (unlikely(!watch)) + return ERR_PTR(-ENOMEM); + + INIT_LIST_HEAD(&watch->rules); + atomic_set(&watch->count, 1); + watch->path = path; + watch->dev = (dev_t)-1; + watch->ino = (unsigned long)-1; + + return watch; +} + +/* Translate a watch string to kernel respresentation. */ +int audit_to_watch(struct audit_krule *krule, char *path, int len, u32 op) +{ + struct audit_watch *watch; + + if (!audit_ih) + return -EOPNOTSUPP; + + if (path[0] != '/' || path[len-1] == '/' || + krule->listnr != AUDIT_FILTER_EXIT || + op != Audit_equal || + krule->inode_f || krule->watch || krule->tree) + return -EINVAL; + + watch = audit_init_watch(path); + if (IS_ERR(watch)) + return PTR_ERR(watch); + + audit_get_watch(watch); + krule->watch = watch; + + return 0; +} + +/* Duplicate the given audit watch. The new watch's rules list is initialized + * to an empty list and wlist is undefined. */ +static struct audit_watch *audit_dupe_watch(struct audit_watch *old) +{ + char *path; + struct audit_watch *new; + + path = kstrdup(old->path, GFP_KERNEL); + if (unlikely(!path)) + return ERR_PTR(-ENOMEM); + + new = audit_init_watch(path); + if (IS_ERR(new)) { + kfree(path); + goto out; + } + + new->dev = old->dev; + new->ino = old->ino; + get_inotify_watch(&old->parent->wdata); + new->parent = old->parent; + +out: + return new; +} + +static void audit_watch_log_rule_change(struct audit_krule *r, struct audit_watch *w, char *op) +{ + if (audit_enabled) { + struct audit_buffer *ab; + ab = audit_log_start(NULL, GFP_NOFS, AUDIT_CONFIG_CHANGE); + audit_log_format(ab, "auid=%u ses=%u op=", + audit_get_loginuid(current), + audit_get_sessionid(current)); + audit_log_string(ab, op); + audit_log_format(ab, " path="); + audit_log_untrustedstring(ab, w->path); + audit_log_key(ab, r->filterkey); + audit_log_format(ab, " list=%d res=1", r->listnr); + audit_log_end(ab); + } +} + +/* Update inode info in audit rules based on filesystem event. */ +static void audit_update_watch(struct audit_parent *parent, + const char *dname, dev_t dev, + unsigned long ino, unsigned invalidating) +{ + struct audit_watch *owatch, *nwatch, *nextw; + struct audit_krule *r, *nextr; + struct audit_entry *oentry, *nentry; + + mutex_lock(&audit_filter_mutex); + list_for_each_entry_safe(owatch, nextw, &parent->watches, wlist) { + if (audit_compare_dname_path(dname, owatch->path, NULL)) + continue; + + /* If the update involves invalidating rules, do the inode-based + * filtering now, so we don't omit records. */ + if (invalidating && current->audit_context) + audit_filter_inodes(current, current->audit_context); + + nwatch = audit_dupe_watch(owatch); + if (IS_ERR(nwatch)) { + mutex_unlock(&audit_filter_mutex); + audit_panic("error updating watch, skipping"); + return; + } + nwatch->dev = dev; + nwatch->ino = ino; + + list_for_each_entry_safe(r, nextr, &owatch->rules, rlist) { + + oentry = container_of(r, struct audit_entry, rule); + list_del(&oentry->rule.rlist); + list_del_rcu(&oentry->list); + + nentry = audit_dupe_rule(&oentry->rule, nwatch); + if (IS_ERR(nentry)) { + list_del(&oentry->rule.list); + audit_panic("error updating watch, removing"); + } else { + int h = audit_hash_ino((u32)ino); + list_add(&nentry->rule.rlist, &nwatch->rules); + list_add_rcu(&nentry->list, &audit_inode_hash[h]); + list_replace(&oentry->rule.list, + &nentry->rule.list); + } + + audit_watch_log_rule_change(r, owatch, "updated rules"); + + call_rcu(&oentry->rcu, audit_free_rule_rcu); + } + + audit_remove_watch(owatch); + goto add_watch_to_parent; /* event applies to a single watch */ + } + mutex_unlock(&audit_filter_mutex); + return; + +add_watch_to_parent: + list_add(&nwatch->wlist, &parent->watches); + mutex_unlock(&audit_filter_mutex); + return; +} + +/* Remove all watches & rules associated with a parent that is going away. */ +static void audit_remove_parent_watches(struct audit_parent *parent) +{ + struct audit_watch *w, *nextw; + struct audit_krule *r, *nextr; + struct audit_entry *e; + + mutex_lock(&audit_filter_mutex); + parent->flags |= AUDIT_PARENT_INVALID; + list_for_each_entry_safe(w, nextw, &parent->watches, wlist) { + list_for_each_entry_safe(r, nextr, &w->rules, rlist) { + e = container_of(r, struct audit_entry, rule); + audit_watch_log_rule_change(r, w, "remove rule"); + list_del(&r->rlist); + list_del(&r->list); + list_del_rcu(&e->list); + call_rcu(&e->rcu, audit_free_rule_rcu); + } + audit_remove_watch(w); + } + mutex_unlock(&audit_filter_mutex); +} + +/* Unregister inotify watches for parents on in_list. + * Generates an IN_IGNORED event. */ +void audit_inotify_unregister(struct list_head *in_list) +{ + struct audit_parent *p, *n; + + list_for_each_entry_safe(p, n, in_list, ilist) { + list_del(&p->ilist); + inotify_rm_watch(audit_ih, &p->wdata); + /* the unpin matching the pin in audit_do_del_rule() */ + unpin_inotify_watch(&p->wdata); + } +} + +/* Get path information necessary for adding watches. */ +static int audit_get_nd(char *path, struct nameidata **ndp, struct nameidata **ndw) +{ + struct nameidata *ndparent, *ndwatch; + int err; + + ndparent = kmalloc(sizeof(*ndparent), GFP_KERNEL); + if (unlikely(!ndparent)) + return -ENOMEM; + + ndwatch = kmalloc(sizeof(*ndwatch), GFP_KERNEL); + if (unlikely(!ndwatch)) { + kfree(ndparent); + return -ENOMEM; + } + + err = path_lookup(path, LOOKUP_PARENT, ndparent); + if (err) { + kfree(ndparent); + kfree(ndwatch); + return err; + } + + err = path_lookup(path, 0, ndwatch); + if (err) { + kfree(ndwatch); + ndwatch = NULL; + } + + *ndp = ndparent; + *ndw = ndwatch; + + return 0; +} + +/* Release resources used for watch path information. */ +static void audit_put_nd(struct nameidata *ndp, struct nameidata *ndw) +{ + if (ndp) { + path_put(&ndp->path); + kfree(ndp); + } + if (ndw) { + path_put(&ndw->path); + kfree(ndw); + } +} + +/* Associate the given rule with an existing parent inotify_watch. + * Caller must hold audit_filter_mutex. */ +static void audit_add_to_parent(struct audit_krule *krule, + struct audit_parent *parent) +{ + struct audit_watch *w, *watch = krule->watch; + int watch_found = 0; + + list_for_each_entry(w, &parent->watches, wlist) { + if (strcmp(watch->path, w->path)) + continue; + + watch_found = 1; + + /* put krule's and initial refs to temporary watch */ + audit_put_watch(watch); + audit_put_watch(watch); + + audit_get_watch(w); + krule->watch = watch = w; + break; + } + + if (!watch_found) { + get_inotify_watch(&parent->wdata); + watch->parent = parent; + + list_add(&watch->wlist, &parent->watches); + } + list_add(&krule->rlist, &watch->rules); +} + +/* Find a matching watch entry, or add this one. + * Caller must hold audit_filter_mutex. */ +int audit_add_watch(struct audit_krule *krule) +{ + struct audit_watch *watch = krule->watch; + struct inotify_watch *i_watch; + struct audit_parent *parent; + struct nameidata *ndp = NULL, *ndw = NULL; + int ret = 0; + + mutex_unlock(&audit_filter_mutex); + + /* Avoid calling path_lookup under audit_filter_mutex. */ + ret = audit_get_nd(watch->path, &ndp, &ndw); + if (ret) { + /* caller expects mutex locked */ + mutex_lock(&audit_filter_mutex); + goto error; + } + + /* update watch filter fields */ + if (ndw) { + watch->dev = ndw->path.dentry->d_inode->i_sb->s_dev; + watch->ino = ndw->path.dentry->d_inode->i_ino; + } + + /* The audit_filter_mutex must not be held during inotify calls because + * we hold it during inotify event callback processing. If an existing + * inotify watch is found, inotify_find_watch() grabs a reference before + * returning. + */ + if (inotify_find_watch(audit_ih, ndp->path.dentry->d_inode, + &i_watch) < 0) { + parent = audit_init_parent(ndp); + if (IS_ERR(parent)) { + /* caller expects mutex locked */ + mutex_lock(&audit_filter_mutex); + ret = PTR_ERR(parent); + goto error; + } + } else + parent = container_of(i_watch, struct audit_parent, wdata); + + mutex_lock(&audit_filter_mutex); + + /* parent was moved before we took audit_filter_mutex */ + if (parent->flags & AUDIT_PARENT_INVALID) + ret = -ENOENT; + else + audit_add_to_parent(krule, parent); + + /* match get in audit_init_parent or inotify_find_watch */ + put_inotify_watch(&parent->wdata); + +error: + audit_put_nd(ndp, ndw); /* NULL args OK */ + return ret; + +} + +void audit_remove_watch_rule(struct audit_krule *krule, struct list_head *list) +{ + struct audit_watch *watch = krule->watch; + struct audit_parent *parent = watch->parent; + + list_del(&krule->rlist); + + if (list_empty(&watch->rules)) { + audit_remove_watch(watch); + + if (list_empty(&parent->watches)) { + /* Put parent on the inotify un-registration + * list. Grab a reference before releasing + * audit_filter_mutex, to be released in + * audit_inotify_unregister(). + * If filesystem is going away, just leave + * the sucker alone, eviction will take + * care of it. */ + if (pin_inotify_watch(&parent->wdata)) + list_add(&parent->ilist, list); + } + } +} + +/* Update watch data in audit rules based on inotify events. */ +static void audit_handle_ievent(struct inotify_watch *i_watch, u32 wd, u32 mask, + u32 cookie, const char *dname, struct inode *inode) +{ + struct audit_parent *parent; + + parent = container_of(i_watch, struct audit_parent, wdata); + + if (mask & (IN_CREATE|IN_MOVED_TO) && inode) + audit_update_watch(parent, dname, inode->i_sb->s_dev, + inode->i_ino, 0); + else if (mask & (IN_DELETE|IN_MOVED_FROM)) + audit_update_watch(parent, dname, (dev_t)-1, (unsigned long)-1, 1); + /* inotify automatically removes the watch and sends IN_IGNORED */ + else if (mask & (IN_DELETE_SELF|IN_UNMOUNT)) + audit_remove_parent_watches(parent); + /* inotify does not remove the watch, so remove it manually */ + else if(mask & IN_MOVE_SELF) { + audit_remove_parent_watches(parent); + inotify_remove_watch_locked(audit_ih, i_watch); + } else if (mask & IN_IGNORED) + put_inotify_watch(i_watch); +} + +static const struct inotify_operations audit_inotify_ops = { + .handle_event = audit_handle_ievent, + .destroy_watch = audit_free_parent, +}; + +static int __init audit_watch_init(void) +{ + audit_ih = inotify_init(&audit_inotify_ops); + if (IS_ERR(audit_ih)) + audit_panic("cannot initialize inotify handle"); + return 0; +} +subsys_initcall(audit_watch_init); diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c index fbf24d121d9..a70604047f3 100644 --- a/kernel/auditfilter.c +++ b/kernel/auditfilter.c @@ -27,7 +27,6 @@ #include <linux/namei.h> #include <linux/netlink.h> #include <linux/sched.h> -#include <linux/inotify.h> #include <linux/security.h> #include "audit.h" @@ -44,36 +43,6 @@ * be written directly provided audit_filter_mutex is held. */ -/* - * Reference counting: - * - * audit_parent: lifetime is from audit_init_parent() to receipt of an IN_IGNORED - * event. Each audit_watch holds a reference to its associated parent. - * - * audit_watch: if added to lists, lifetime is from audit_init_watch() to - * audit_remove_watch(). Additionally, an audit_watch may exist - * temporarily to assist in searching existing filter data. Each - * audit_krule holds a reference to its associated watch. - */ - -struct audit_parent { - struct list_head ilist; /* entry in inotify registration list */ - struct list_head watches; /* associated watches */ - struct inotify_watch wdata; /* inotify watch data */ - unsigned flags; /* status flags */ -}; - -/* - * audit_parent status flags: - * - * AUDIT_PARENT_INVALID - set anytime rules/watches are auto-removed due to - * a filesystem event to ensure we're adding audit watches to a valid parent. - * Technically not needed for IN_DELETE_SELF or IN_UNMOUNT events, as we cannot - * receive them while we have nameidata, but must be used for IN_MOVE_SELF which - * we can receive while holding nameidata. - */ -#define AUDIT_PARENT_INVALID 0x001 - /* Audit filter lists, defined in <linux/audit.h> */ struct list_head audit_filter_list[AUDIT_NR_FILTERS] = { LIST_HEAD_INIT(audit_filter_list[0]), @@ -97,56 +66,21 @@ static struct list_head audit_rules_list[AUDIT_NR_FILTERS] = { DEFINE_MUTEX(audit_filter_mutex); -/* Inotify events we care about. */ -#define AUDIT_IN_WATCH IN_MOVE|IN_CREATE|IN_DELETE|IN_DELETE_SELF|IN_MOVE_SELF - -void audit_free_parent(struct inotify_watch *i_watch) -{ - struct audit_parent *parent; - - parent = container_of(i_watch, struct audit_parent, wdata); - WARN_ON(!list_empty(&parent->watches)); - kfree(parent); -} - -static inline void audit_get_watch(struct audit_watch *watch) -{ - atomic_inc(&watch->count); -} - -static void audit_put_watch(struct audit_watch *watch) -{ - if (atomic_dec_and_test(&watch->count)) { - WARN_ON(watch->parent); - WARN_ON(!list_empty(&watch->rules)); - kfree(watch->path); - kfree(watch); - } -} - -static void audit_remove_watch(struct audit_watch *watch) -{ - list_del(&watch->wlist); - put_inotify_watch(&watch->parent->wdata); - watch->parent = NULL; - audit_put_watch(watch); /* match initial get */ -} - static inline void audit_free_rule(struct audit_entry *e) { int i; - + struct audit_krule *erule = &e->rule; /* some rules don't have associated watches */ - if (e->rule.watch) - audit_put_watch(e->rule.watch); - if (e->rule.fields) - for (i = 0; i < e->rule.field_count; i++) { - struct audit_field *f = &e->rule.fields[i]; + if (erule->watch) + audit_put_watch(erule->watch); + if (erule->fields) + for (i = 0; i < erule->field_count; i++) { + struct audit_field *f = &erule->fields[i]; kfree(f->lsm_str); security_audit_rule_free(f->lsm_rule); } - kfree(e->rule.fields); - kfree(e->rule.filterkey); + kfree(erule->fields); + kfree(erule->filterkey); kfree(e); } @@ -156,50 +90,6 @@ void audit_free_rule_rcu(struct rcu_head *head) audit_free_rule(e); } -/* Initialize a parent watch entry. */ -static struct audit_parent *audit_init_parent(struct nameidata *ndp) -{ - struct audit_parent *parent; - s32 wd; - - parent = kzalloc(sizeof(*parent), GFP_KERNEL); - if (unlikely(!parent)) - return ERR_PTR(-ENOMEM); - - INIT_LIST_HEAD(&parent->watches); - parent->flags = 0; - - inotify_init_watch(&parent->wdata); - /* grab a ref so inotify watch hangs around until we take audit_filter_mutex */ - get_inotify_watch(&parent->wdata); - wd = inotify_add_watch(audit_ih, &parent->wdata, - ndp->path.dentry->d_inode, AUDIT_IN_WATCH); - if (wd < 0) { - audit_free_parent(&parent->wdata); - return ERR_PTR(wd); - } - - return parent; -} - -/* Initialize a watch entry. */ -static struct audit_watch *audit_init_watch(char *path) -{ - struct audit_watch *watch; - - watch = kzalloc(sizeof(*watch), GFP_KERNEL); - if (unlikely(!watch)) - return ERR_PTR(-ENOMEM); - - INIT_LIST_HEAD(&watch->rules); - atomic_set(&watch->count, 1); - watch->path = path; - watch->dev = (dev_t)-1; - watch->ino = (unsigned long)-1; - - return watch; -} - /* Initialize an audit filterlist entry. */ static inline struct audit_entry *audit_init_entry(u32 field_count) { @@ -260,31 +150,6 @@ static inline int audit_to_inode(struct audit_krule *krule, return 0; } -/* Translate a watch string to kernel respresentation. */ -static int audit_to_watch(struct audit_krule *krule, char *path, int len, - u32 op) -{ - struct audit_watch *watch; - - if (!audit_ih) - return -EOPNOTSUPP; - - if (path[0] != '/' || path[len-1] == '/' || - krule->listnr != AUDIT_FILTER_EXIT || - op != Audit_equal || - krule->inode_f || krule->watch || krule->tree) - return -EINVAL; - - watch = audit_init_watch(path); - if (IS_ERR(watch)) - return PTR_ERR(watch); - - audit_get_watch(watch); - krule->watch = watch; - - return 0; -} - static __u32 *classes[AUDIT_SYSCALL_CLASSES]; int __init audit_register_class(int class, unsigned *list) @@ -766,7 +631,8 @@ static struct audit_rule_data *audit_krule_to_data(struct audit_krule *krule) break; case AUDIT_WATCH: data->buflen += data->values[i] = - audit_pack_string(&bufp, krule->watch->path); + audit_pack_string(&bufp, + audit_watch_path(krule->watch)); break; case AUDIT_DIR: data->buflen += data->values[i] = @@ -818,7 +684,8 @@ static int audit_compare_rule(struct audit_krule *a, struct audit_krule *b) return 1; break; case AUDIT_WATCH: - if (strcmp(a->watch->path, b->watch->path)) + if (strcmp(audit_watch_path(a->watch), + audit_watch_path(b->watch))) return 1; break; case AUDIT_DIR: @@ -844,32 +711,6 @@ static int audit_compare_rule(struct audit_krule *a, struct audit_krule *b) return 0; } -/* Duplicate the given audit watch. The new watch's rules list is initialized - * to an empty list and wlist is undefined. */ -static struct audit_watch *audit_dupe_watch(struct audit_watch *old) -{ - char *path; - struct audit_watch *new; - - path = kstrdup(old->path, GFP_KERNEL); - if (unlikely(!path)) - return ERR_PTR(-ENOMEM); - - new = audit_init_watch(path); - if (IS_ERR(new)) { - kfree(path); - goto out; - } - - new->dev = old->dev; - new->ino = old->ino; - get_inotify_watch(&old->parent->wdata); - new->parent = old->parent; - -out: - return new; -} - /* Duplicate LSM field information. The lsm_rule is opaque, so must be * re-initialized. */ static inline int audit_dupe_lsm_field(struct audit_field *df, @@ -904,8 +745,8 @@ static inline int audit_dupe_lsm_field(struct audit_field *df, * rule with the new rule in the filterlist, then free the old rule. * The rlist element is undefined; list manipulations are handled apart from * the initial copy. */ -static struct audit_entry *audit_dupe_rule(struct audit_krule *old, - struct audit_watch *watch) +struct audit_entry *audit_dupe_rule(struct audit_krule *old, + struct audit_watch *watch) { u32 fcount = old->field_count; struct audit_entry *entry; @@ -977,137 +818,6 @@ static struct audit_entry *audit_dupe_rule(struct audit_krule *old, return entry; } -/* Update inode info in audit rules based on filesystem event. */ -static void audit_update_watch(struct audit_parent *parent, - const char *dname, dev_t dev, - unsigned long ino, unsigned invalidating) -{ - struct audit_watch *owatch, *nwatch, *nextw; - struct audit_krule *r, *nextr; - struct audit_entry *oentry, *nentry; - - mutex_lock(&audit_filter_mutex); - list_for_each_entry_safe(owatch, nextw, &parent->watches, wlist) { - if (audit_compare_dname_path(dname, owatch->path, NULL)) - continue; - - /* If the update involves invalidating rules, do the inode-based - * filtering now, so we don't omit records. */ - if (invalidating && current->audit_context) - audit_filter_inodes(current, current->audit_context); - - nwatch = audit_dupe_watch(owatch); - if (IS_ERR(nwatch)) { - mutex_unlock(&audit_filter_mutex); - audit_panic("error updating watch, skipping"); - return; - } - nwatch->dev = dev; - nwatch->ino = ino; - - list_for_each_entry_safe(r, nextr, &owatch->rules, rlist) { - - oentry = container_of(r, struct audit_entry, rule); - list_del(&oentry->rule.rlist); - list_del_rcu(&oentry->list); - - nentry = audit_dupe_rule(&oentry->rule, nwatch); - if (IS_ERR(nentry)) { - list_del(&oentry->rule.list); - audit_panic("error updating watch, removing"); - } else { - int h = audit_hash_ino((u32)ino); - list_add(&nentry->rule.rlist, &nwatch->rules); - list_add_rcu(&nentry->list, &audit_inode_hash[h]); - list_replace(&oentry->rule.list, - &nentry->rule.list); - } - - call_rcu(&oentry->rcu, audit_free_rule_rcu); - } - - if (audit_enabled) { - struct audit_buffer *ab; - ab = audit_log_start(NULL, GFP_KERNEL, - AUDIT_CONFIG_CHANGE); - audit_log_format(ab, "auid=%u ses=%u", - audit_get_loginuid(current), - audit_get_sessionid(current)); - audit_log_format(ab, - " op=updated rules specifying path="); - audit_log_untrustedstring(ab, owatch->path); - audit_log_format(ab, " with dev=%u ino=%lu\n", - dev, ino); - audit_log_format(ab, " list=%d res=1", r->listnr); - audit_log_end(ab); - } - audit_remove_watch(owatch); - goto add_watch_to_parent; /* event applies to a single watch */ - } - mutex_unlock(&audit_filter_mutex); - return; - -add_watch_to_parent: - list_add(&nwatch->wlist, &parent->watches); - mutex_unlock(&audit_filter_mutex); - return; -} - -/* Remove all watches & rules associated with a parent that is going away. */ -static void audit_remove_parent_watches(struct audit_parent *parent) -{ - struct audit_watch *w, *nextw; - struct audit_krule *r, *nextr; - struct audit_entry *e; - - mutex_lock(&audit_filter_mutex); - parent->flags |= AUDIT_PARENT_INVALID; - list_for_each_entry_safe(w, nextw, &parent->watches, wlist) { - list_for_each_entry_safe(r, nextr, &w->rules, rlist) { - e = container_of(r, struct audit_entry, rule); - if (audit_enabled) { - struct audit_buffer *ab; - ab = audit_log_start(NULL, GFP_KERNEL, - AUDIT_CONFIG_CHANGE); - audit_log_format(ab, "auid=%u ses=%u", - audit_get_loginuid(current), - audit_get_sessionid(current)); - audit_log_format(ab, " op=remove rule path="); - audit_log_untrustedstring(ab, w->path); - if (r->filterkey) { - audit_log_format(ab, " key="); - audit_log_untrustedstring(ab, - r->filterkey); - } else - audit_log_format(ab, " key=(null)"); - audit_log_format(ab, " list=%d res=1", - r->listnr); - audit_log_end(ab); - } - list_del(&r->rlist); - list_del(&r->list); - list_del_rcu(&e->list); - call_rcu(&e->rcu, audit_free_rule_rcu); - } - audit_remove_watch(w); - } - mutex_unlock(&audit_filter_mutex); -} - -/* Unregister inotify watches for parents on in_list. - * Generates an IN_IGNORED event. */ -static void audit_inotify_unregister(struct list_head *in_list) -{ - struct audit_parent *p, *n; - - list_for_each_entry_safe(p, n, in_list, ilist) { - list_del(&p->ilist); - inotify_rm_watch(audit_ih, &p->wdata); - /* the unpin matching the pin in audit_do_del_rule() */ - unpin_inotify_watch(&p->wdata); - } -} - /* Find an existing audit rule. * Caller must hold audit_filter_mutex to prevent stale rule data. */ static struct audit_entry *audit_find_rule(struct audit_entry *entry, @@ -1145,134 +855,6 @@ out: return found; } -/* Get path information necessary for adding watches. */ -static int audit_get_nd(char *path, struct nameidata **ndp, - struct nameidata **ndw) -{ - struct nameidata *ndparent, *ndwatch; - int err; - - ndparent = kmalloc(sizeof(*ndparent), GFP_KERNEL); - if (unlikely(!ndparent)) - return -ENOMEM; - - ndwatch = kmalloc(sizeof(*ndwatch), GFP_KERNEL); - if (unlikely(!ndwatch)) { - kfree(ndparent); - return -ENOMEM; - } - - err = path_lookup(path, LOOKUP_PARENT, ndparent); - if (err) { - kfree(ndparent); - kfree(ndwatch); - return err; - } - - err = path_lookup(path, 0, ndwatch); - if (err) { - kfree(ndwatch); - ndwatch = NULL; - } - - *ndp = ndparent; - *ndw = ndwatch; - - return 0; -} - -/* Release resources used for watch path information. */ -static void audit_put_nd(struct nameidata *ndp, struct nameidata *ndw) -{ - if (ndp) { - path_put(&ndp->path); - kfree(ndp); - } - if (ndw) { - path_put(&ndw->path); - kfree(ndw); - } -} - -/* Associate the given rule with an existing parent inotify_watch. - * Caller must hold audit_filter_mutex. */ -static void audit_add_to_parent(struct audit_krule *krule, - struct audit_parent *parent) -{ - struct audit_watch *w, *watch = krule->watch; - int watch_found = 0; - - list_for_each_entry(w, &parent->watches, wlist) { - if (strcmp(watch->path, w->path)) - continue; - - watch_found = 1; - - /* put krule's and initial refs to temporary watch */ - audit_put_watch(watch); - audit_put_watch(watch); - - audit_get_watch(w); - krule->watch = watch = w; - break; - } - - if (!watch_found) { - get_inotify_watch(&parent->wdata); - watch->parent = parent; - - list_add(&watch->wlist, &parent->watches); - } - list_add(&krule->rlist, &watch->rules); -} - -/* Find a matching watch entry, or add this one. - * Caller must hold audit_filter_mutex. */ -static int audit_add_watch(struct audit_krule *krule, struct nameidata *ndp, - struct nameidata *ndw) -{ - struct audit_watch *watch = krule->watch; - struct inotify_watch *i_watch; - struct audit_parent *parent; - int ret = 0; - - /* update watch filter fields */ - if (ndw) { - watch->dev = ndw->path.dentry->d_inode->i_sb->s_dev; - watch->ino = ndw->path.dentry->d_inode->i_ino; - } - - /* The audit_filter_mutex must not be held during inotify calls because - * we hold it during inotify event callback processing. If an existing - * inotify watch is found, inotify_find_watch() grabs a reference before - * returning. - */ - mutex_unlock(&audit_filter_mutex); - - if (inotify_find_watch(audit_ih, ndp->path.dentry->d_inode, - &i_watch) < 0) { - parent = audit_init_parent(ndp); - if (IS_ERR(parent)) { - /* caller expects mutex locked */ - mutex_lock(&audit_filter_mutex); - return PTR_ERR(parent); - } - } else - parent = container_of(i_watch, struct audit_parent, wdata); - - mutex_lock(&audit_filter_mutex); - - /* parent was moved before we took audit_filter_mutex */ - if (parent->flags & AUDIT_PARENT_INVALID) - ret = -ENOENT; - else - audit_add_to_parent(krule, parent); - - /* match get in audit_init_parent or inotify_find_watch */ - put_inotify_watch(&parent->wdata); - return ret; -} - static u64 prio_low = ~0ULL/2; static u64 prio_high = ~0ULL/2 - 1; @@ -1282,7 +864,6 @@ static inline int audit_add_rule(struct audit_entry *entry) struct audit_entry *e; struct audit_watch *watch = entry->rule.watch; struct audit_tree *tree = entry->rule.tree; - struct nameidata *ndp = NULL, *ndw = NULL; struct list_head *list; int h, err; #ifdef CONFIG_AUDITSYSCALL @@ -1296,8 +877,8 @@ static inline int audit_add_rule(struct audit_entry *entry) mutex_lock(&audit_filter_mutex); e = audit_find_rule(entry, &list); - mutex_unlock(&audit_filter_mutex); if (e) { + mutex_unlock(&audit_filter_mutex); err = -EEXIST; /* normally audit_add_tree_rule() will free it on failure */ if (tree) @@ -1305,22 +886,16 @@ static inline int audit_add_rule(struct audit_entry *entry) goto error; } - /* Avoid calling path_lookup under audit_filter_mutex. */ - if (watch) { - err = audit_get_nd(watch->path, &ndp, &ndw); - if (err) - goto error; - } - - mutex_lock(&audit_filter_mutex); if (watch) { /* audit_filter_mutex is dropped and re-taken during this call */ - err = audit_add_watch(&entry->rule, ndp, ndw); + err = audit_add_watch(&entry->rule); if (err) { mutex_unlock(&audit_filter_mutex); goto error; } - h = audit_hash_ino((u32)watch->ino); + /* entry->rule.watch may have changed during audit_add_watch() */ + watch = entry->rule.watch; + h = audit_hash_ino((u32)audit_watch_inode(watch)); list = &audit_inode_hash[h]; } if (tree) { @@ -1358,11 +933,9 @@ static inline int audit_add_rule(struct audit_entry *entry) #endif mutex_unlock(&audit_filter_mutex); - audit_put_nd(ndp, ndw); /* NULL args OK */ return 0; error: - audit_put_nd(ndp, ndw); /* NULL args OK */ if (watch) audit_put_watch(watch); /* tmp watch, matches initial get */ return err; @@ -1372,7 +945,7 @@ error: static inline int audit_del_rule(struct audit_entry *entry) { struct audit_entry *e; - struct audit_watch *watch, *tmp_watch = entry->rule.watch; + struct audit_watch *watch = entry->rule.watch; struct audit_tree *tree = entry->rule.tree; struct list_head *list; LIST_HEAD(inotify_list); @@ -1394,29 +967,8 @@ static inline int audit_del_rule(struct audit_entry *entry) goto out; } - watch = e->rule.watch; - if (watch) { - struct audit_parent *parent = watch->parent; - - list_del(&e->rule.rlist); - - if (list_empty(&watch->rules)) { - audit_remove_watch(watch); - - if (list_empty(&parent->watches)) { - /* Put parent on the inotify un-registration - * list. Grab a reference before releasing - * audit_filter_mutex, to be released in - * audit_inotify_unregister(). - * If filesystem is going away, just leave - * the sucker alone, eviction will take - * care of it. - */ - if (pin_inotify_watch(&parent->wdata)) - list_add(&parent->ilist, &inotify_list); - } - } - } + if (e->rule.watch) + audit_remove_watch_rule(&e->rule, &inotify_list); if (e->rule.tree) audit_remove_tree_rule(&e->rule); @@ -1438,8 +990,8 @@ static inline int audit_del_rule(struct audit_entry *entry) audit_inotify_unregister(&inotify_list); out: - if (tmp_watch) - audit_put_watch(tmp_watch); /* match initial get */ + if (watch) + audit_put_watch(watch); /* match initial get */ if (tree) audit_put_tree(tree); /* that's the temporary one */ @@ -1527,11 +1079,9 @@ static void audit_log_rule_change(uid_t loginuid, u32 sessionid, u32 sid, security_release_secctx(ctx, len); } } - audit_log_format(ab, " op=%s rule key=", action); - if (rule->filterkey) - audit_log_untrustedstring(ab, rule->filterkey); - else - audit_log_format(ab, "(null)"); + audit_log_format(ab, " op="); + audit_log_string(ab, action); + audit_log_key(ab, rule->filterkey); audit_log_format(ab, " list=%d res=%d", rule->listnr, res); audit_log_end(ab); } @@ -1595,7 +1145,7 @@ int audit_receive_filter(int type, int pid, int uid, int seq, void *data, return PTR_ERR(entry); err = audit_add_rule(entry); - audit_log_rule_change(loginuid, sessionid, sid, "add", + audit_log_rule_change(loginuid, sessionid, sid, "add rule", &entry->rule, !err); if (err) @@ -1611,7 +1161,7 @@ int audit_receive_filter(int type, int pid, int uid, int seq, void *data, return PTR_ERR(entry); err = audit_del_rule(entry); - audit_log_rule_change(loginuid, sessionid, sid, "remove", + audit_log_rule_change(loginuid, sessionid, sid, "remove rule", &entry->rule, !err); audit_free_rule(entry); @@ -1793,7 +1343,7 @@ static int update_lsm_rule(struct audit_krule *r) list_del(&r->list); } else { if (watch) { - list_add(&nentry->rule.rlist, &watch->rules); + list_add(&nentry->rule.rlist, audit_watch_rules(watch)); list_del(&r->rlist); } else if (tree) list_replace_init(&r->rlist, &nentry->rule.rlist); @@ -1829,27 +1379,3 @@ int audit_update_lsm_rules(void) return err; } - -/* Update watch data in audit rules based on inotify events. */ -void audit_handle_ievent(struct inotify_watch *i_watch, u32 wd, u32 mask, - u32 cookie, const char *dname, struct inode *inode) -{ - struct audit_parent *parent; - - parent = container_of(i_watch, struct audit_parent, wdata); - - if (mask & (IN_CREATE|IN_MOVED_TO) && inode) - audit_update_watch(parent, dname, inode->i_sb->s_dev, - inode->i_ino, 0); - else if (mask & (IN_DELETE|IN_MOVED_FROM)) - audit_update_watch(parent, dname, (dev_t)-1, (unsigned long)-1, 1); - /* inotify automatically removes the watch and sends IN_IGNORED */ - else if (mask & (IN_DELETE_SELF|IN_UNMOUNT)) - audit_remove_parent_watches(parent); - /* inotify does not remove the watch, so remove it manually */ - else if(mask & IN_MOVE_SELF) { - audit_remove_parent_watches(parent); - inotify_remove_watch_locked(audit_ih, i_watch); - } else if (mask & IN_IGNORED) - put_inotify_watch(i_watch); -} diff --git a/kernel/auditsc.c b/kernel/auditsc.c index 8cbddff6c28..68d3c6a0ecd 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -66,6 +66,7 @@ #include <linux/syscalls.h> #include <linux/inotify.h> #include <linux/capability.h> +#include <linux/fs_struct.h> #include "audit.h" @@ -198,6 +199,7 @@ struct audit_context { struct audit_tree_refs *trees, *first_trees; int tree_count; + struct list_head killed_trees; int type; union { @@ -328,6 +330,14 @@ static int audit_match_filetype(struct audit_context *ctx, int which) */ #ifdef CONFIG_AUDIT_TREE +static void audit_set_auditable(struct audit_context *ctx) +{ + if (!ctx->prio) { + ctx->prio = 1; + ctx->current_state = AUDIT_RECORD_CONTEXT; + } +} + static int put_tree_ref(struct audit_context *ctx, struct audit_chunk *chunk) { struct audit_tree_refs *p = ctx->trees; @@ -539,9 +549,9 @@ static int audit_filter_rules(struct task_struct *tsk, } break; case AUDIT_WATCH: - if (name && rule->watch->ino != (unsigned long)-1) - result = (name->dev == rule->watch->dev && - name->ino == rule->watch->ino); + if (name && audit_watch_inode(rule->watch) != (unsigned long)-1) + result = (name->dev == audit_watch_dev(rule->watch) && + name->ino == audit_watch_inode(rule->watch)); break; case AUDIT_DIR: if (ctx) @@ -741,17 +751,9 @@ void audit_filter_inodes(struct task_struct *tsk, struct audit_context *ctx) rcu_read_unlock(); } -static void audit_set_auditable(struct audit_context *ctx) -{ - if (!ctx->prio) { - ctx->prio = 1; - ctx->current_state = AUDIT_RECORD_CONTEXT; - } -} - static inline struct audit_context *audit_get_context(struct task_struct *tsk, int return_valid, - int return_code) + long return_code) { struct audit_context *context = tsk->audit_context; @@ -852,6 +854,7 @@ static inline struct audit_context *audit_alloc_context(enum audit_state state) if (!(context = kmalloc(sizeof(*context), GFP_KERNEL))) return NULL; audit_zero_context(context, state); + INIT_LIST_HEAD(&context->killed_trees); return context; } @@ -1023,8 +1026,8 @@ static int audit_log_single_execve_arg(struct audit_context *context, { char arg_num_len_buf[12]; const char __user *tmp_p = p; - /* how many digits are in arg_num? 3 is the length of a=\n */ - size_t arg_num_len = snprintf(arg_num_len_buf, 12, "%d", arg_num) + 3; + /* how many digits are in arg_num? 5 is the length of ' a=""' */ + size_t arg_num_len = snprintf(arg_num_len_buf, 12, "%d", arg_num) + 5; size_t len, len_left, to_send; size_t max_execve_audit_len = MAX_EXECVE_AUDIT_LEN; unsigned int i, has_cntl = 0, too_long = 0; @@ -1109,7 +1112,7 @@ static int audit_log_single_execve_arg(struct audit_context *context, * so we can be sure nothing was lost. */ if ((i == 0) && (too_long)) - audit_log_format(*ab, "a%d_len=%zu ", arg_num, + audit_log_format(*ab, " a%d_len=%zu", arg_num, has_cntl ? 2*len : len); /* @@ -1129,15 +1132,14 @@ static int audit_log_single_execve_arg(struct audit_context *context, buf[to_send] = '\0'; /* actually log it */ - audit_log_format(*ab, "a%d", arg_num); + audit_log_format(*ab, " a%d", arg_num); if (too_long) audit_log_format(*ab, "[%d]", i); audit_log_format(*ab, "="); if (has_cntl) audit_log_n_hex(*ab, buf, to_send); else - audit_log_format(*ab, "\"%s\"", buf); - audit_log_format(*ab, "\n"); + audit_log_string(*ab, buf); p += to_send; len_left -= to_send; @@ -1165,7 +1167,7 @@ static void audit_log_execve_info(struct audit_context *context, p = (const char __user *)axi->mm->arg_start; - audit_log_format(*ab, "argc=%d ", axi->argc); + audit_log_format(*ab, "argc=%d", axi->argc); /* * we need some kernel buffer to hold the userspace args. Just @@ -1372,11 +1374,7 @@ static void audit_log_exit(struct audit_context *context, struct task_struct *ts audit_log_task_info(ab, tsk); - if (context->filterkey) { - audit_log_format(ab, " key="); - audit_log_untrustedstring(ab, context->filterkey); - } else - audit_log_format(ab, " key=(null)"); + audit_log_key(ab, context->filterkey); audit_log_end(ab); for (aux = context->aux; aux; aux = aux->next) { @@ -1478,7 +1476,7 @@ static void audit_log_exit(struct audit_context *context, struct task_struct *ts case 0: /* name was specified as a relative path and the * directory component is the cwd */ - audit_log_d_path(ab, " name=", &context->pwd); + audit_log_d_path(ab, "name=", &context->pwd); break; default: /* log the name's directory component */ @@ -1549,6 +1547,8 @@ void audit_free(struct task_struct *tsk) /* that can happen only if we are called from do_exit() */ if (context->in_syscall && context->current_state == AUDIT_RECORD_CONTEXT) audit_log_exit(context, tsk); + if (!list_empty(&context->killed_trees)) + audit_kill_trees(&context->killed_trees); audit_free_context(context); } @@ -1692,6 +1692,9 @@ void audit_syscall_exit(int valid, long return_code) context->in_syscall = 0; context->prio = context->state == AUDIT_RECORD_CONTEXT ? ~0ULL : 0; + if (!list_empty(&context->killed_trees)) + audit_kill_trees(&context->killed_trees); + if (context->previous) { struct audit_context *new_context = context->previous; context->previous = NULL; @@ -2149,7 +2152,7 @@ int audit_set_loginuid(struct task_struct *task, uid_t loginuid) * __audit_mq_open - record audit data for a POSIX MQ open * @oflag: open flag * @mode: mode bits - * @u_attr: queue attributes + * @attr: queue attributes * */ void __audit_mq_open(int oflag, mode_t mode, struct mq_attr *attr) @@ -2196,7 +2199,7 @@ void __audit_mq_sendrecv(mqd_t mqdes, size_t msg_len, unsigned int msg_prio, /** * __audit_mq_notify - record audit data for a POSIX MQ notify * @mqdes: MQ descriptor - * @u_notification: Notification event + * @notification: Notification event * */ @@ -2525,3 +2528,11 @@ void audit_core_dumps(long signr) audit_log_format(ab, " sig=%ld", signr); audit_log_end(ab); } + +struct list_head *audit_killed_trees(void) +{ + struct audit_context *ctx = current->audit_context; + if (likely(!ctx || !ctx->in_syscall)) + return NULL; + return &ctx->killed_trees; +} diff --git a/kernel/cgroup.c b/kernel/cgroup.c index c29831076e7..c7ece8f027f 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -46,6 +46,8 @@ #include <linux/cgroupstats.h> #include <linux/hash.h> #include <linux/namei.h> +#include <linux/smp_lock.h> +#include <linux/pid_namespace.h> #include <asm/atomic.h> @@ -94,7 +96,6 @@ struct cgroupfs_root { char release_agent_path[PATH_MAX]; }; - /* * The "rootnode" hierarchy is the "dummy hierarchy", reserved for the * subsystems that are otherwise unattached - it never has more than a @@ -102,6 +103,39 @@ struct cgroupfs_root { */ static struct cgroupfs_root rootnode; +/* + * CSS ID -- ID per subsys's Cgroup Subsys State(CSS). used only when + * cgroup_subsys->use_id != 0. + */ +#define CSS_ID_MAX (65535) +struct css_id { + /* + * The css to which this ID points. This pointer is set to valid value + * after cgroup is populated. If cgroup is removed, this will be NULL. + * This pointer is expected to be RCU-safe because destroy() + * is called after synchronize_rcu(). But for safe use, css_is_removed() + * css_tryget() should be used for avoiding race. + */ + struct cgroup_subsys_state *css; + /* + * ID of this css. + */ + unsigned short id; + /* + * Depth in hierarchy which this ID belongs to. + */ + unsigned short depth; + /* + * ID is freed by RCU. (and lookup routine is RCU safe.) + */ + struct rcu_head rcu_head; + /* + * Hierarchy of CSS ID belongs to. + */ + unsigned short stack[0]; /* Array of Length (depth+1) */ +}; + + /* The list of hierarchy roots */ static LIST_HEAD(roots); @@ -185,6 +219,8 @@ struct cg_cgroup_link { static struct css_set init_css_set; static struct cg_cgroup_link init_css_set_link; +static int cgroup_subsys_init_idr(struct cgroup_subsys *ss); + /* css_set_lock protects the list of css_set objects, and the * chain of tasks off each css_set. Nests outside task->alloc_lock * due to cgroup_iter_start() */ @@ -564,9 +600,13 @@ static struct inode_operations cgroup_dir_inode_operations; static struct file_operations proc_cgroupstats_operations; static struct backing_dev_info cgroup_backing_dev_info = { + .name = "cgroup", .capabilities = BDI_CAP_NO_ACCT_AND_WRITEBACK, }; +static int alloc_css_id(struct cgroup_subsys *ss, + struct cgroup *parent, struct cgroup *child); + static struct inode *cgroup_new_inode(mode_t mode, struct super_block *sb) { struct inode *inode = new_inode(sb); @@ -585,13 +625,18 @@ static struct inode *cgroup_new_inode(mode_t mode, struct super_block *sb) * Call subsys's pre_destroy handler. * This is called before css refcnt check. */ -static void cgroup_call_pre_destroy(struct cgroup *cgrp) +static int cgroup_call_pre_destroy(struct cgroup *cgrp) { struct cgroup_subsys *ss; + int ret = 0; + for_each_subsys(cgrp->root, ss) - if (ss->pre_destroy) - ss->pre_destroy(ss, cgrp); - return; + if (ss->pre_destroy) { + ret = ss->pre_destroy(ss, cgrp); + if (ret) + break; + } + return ret; } static void free_cgroup_rcu(struct rcu_head *obj) @@ -685,6 +730,34 @@ static void cgroup_d_remove_dir(struct dentry *dentry) remove_dir(dentry); } +/* + * A queue for waiters to do rmdir() cgroup. A tasks will sleep when + * cgroup->count == 0 && list_empty(&cgroup->children) && subsys has some + * reference to css->refcnt. In general, this refcnt is expected to goes down + * to zero, soon. + * + * CGRP_WAIT_ON_RMDIR flag is set under cgroup's inode->i_mutex; + */ +DECLARE_WAIT_QUEUE_HEAD(cgroup_rmdir_waitq); + +static void cgroup_wakeup_rmdir_waiter(struct cgroup *cgrp) +{ + if (unlikely(test_and_clear_bit(CGRP_WAIT_ON_RMDIR, &cgrp->flags))) + wake_up_all(&cgroup_rmdir_waitq); +} + +void cgroup_exclude_rmdir(struct cgroup_subsys_state *css) +{ + css_get(css); +} + +void cgroup_release_and_wakeup_rmdir(struct cgroup_subsys_state *css) +{ + cgroup_wakeup_rmdir_waiter(css->cgroup); + css_put(css); +} + + static int rebind_subsystems(struct cgroupfs_root *root, unsigned long final_bits) { @@ -784,6 +857,11 @@ static int parse_cgroupfs_options(char *data, struct cgroup_sb_opts *opts) { char *token, *o = data ?: "all"; + unsigned long mask = (unsigned long)-1; + +#ifdef CONFIG_CPUSETS + mask = ~(1UL << cpuset_subsys_id); +#endif opts->subsys_bits = 0; opts->flags = 0; @@ -828,6 +906,15 @@ static int parse_cgroupfs_options(char *data, } } + /* + * Option noprefix was introduced just for backward compatibility + * with the old cpuset, so we allow noprefix only if mounting just + * the cpuset subsystem. + */ + if (test_bit(ROOT_NOPREFIX, &opts->flags) && + (opts->subsys_bits & mask)) + return -EINVAL; + /* We can't have an empty hierarchy */ if (!opts->subsys_bits) return -EINVAL; @@ -842,6 +929,7 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) struct cgroup *cgrp = &root->top_cgroup; struct cgroup_sb_opts opts; + lock_kernel(); mutex_lock(&cgrp->dentry->d_inode->i_mutex); mutex_lock(&cgroup_mutex); @@ -857,18 +945,19 @@ static int cgroup_remount(struct super_block *sb, int *flags, char *data) } ret = rebind_subsystems(root, opts.subsys_bits); + if (ret) + goto out_unlock; /* (re)populate subsystem files */ - if (!ret) - cgroup_populate_dir(cgrp); + cgroup_populate_dir(cgrp); if (opts.release_agent) strcpy(root->release_agent_path, opts.release_agent); out_unlock: - if (opts.release_agent) - kfree(opts.release_agent); + kfree(opts.release_agent); mutex_unlock(&cgroup_mutex); mutex_unlock(&cgrp->dentry->d_inode->i_mutex); + unlock_kernel(); return ret; } @@ -885,6 +974,7 @@ static void init_cgroup_housekeeping(struct cgroup *cgrp) INIT_LIST_HEAD(&cgrp->children); INIT_LIST_HEAD(&cgrp->css_sets); INIT_LIST_HEAD(&cgrp->release_list); + INIT_LIST_HEAD(&cgrp->pids_list); init_rwsem(&cgrp->pids_mutex); } static void init_cgroup_root(struct cgroupfs_root *root) @@ -969,15 +1059,13 @@ static int cgroup_get_sb(struct file_system_type *fs_type, /* First find the desired set of subsystems */ ret = parse_cgroupfs_options(data, &opts); if (ret) { - if (opts.release_agent) - kfree(opts.release_agent); + kfree(opts.release_agent); return ret; } root = kzalloc(sizeof(*root), GFP_KERNEL); if (!root) { - if (opts.release_agent) - kfree(opts.release_agent); + kfree(opts.release_agent); return -ENOMEM; } @@ -1071,13 +1159,13 @@ static int cgroup_get_sb(struct file_system_type *fs_type, mutex_unlock(&cgroup_mutex); } - return simple_set_mnt(mnt, sb); + simple_set_mnt(mnt, sb); + return 0; free_cg_links: free_cg_links(&tmp_cg_links); drop_new_super: - up_write(&sb->s_umount); - deactivate_super(sb); + deactivate_locked_super(sb); return ret; } @@ -1115,13 +1203,15 @@ static void cgroup_kill_sb(struct super_block *sb) { } write_unlock(&css_set_lock); - list_del(&root->root_list); - root_count--; + if (!list_empty(&root->root_list)) { + list_del(&root->root_list); + root_count--; + } mutex_unlock(&cgroup_mutex); - kfree(root); kill_litter_super(sb); + kfree(root); } static struct file_system_type cgroup_fs_type = { @@ -1277,6 +1367,12 @@ int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) set_bit(CGRP_RELEASABLE, &oldcgrp->flags); synchronize_rcu(); put_css_set(cg); + + /* + * wake up rmdir() waiter. the rmdir should fail since the cgroup + * is no longer empty. + */ + cgroup_wakeup_rmdir_waiter(cgrp); return 0; } @@ -1622,10 +1718,10 @@ static struct inode_operations cgroup_dir_inode_operations = { .rename = cgroup_rename, }; -static int cgroup_create_file(struct dentry *dentry, int mode, +static int cgroup_create_file(struct dentry *dentry, mode_t mode, struct super_block *sb) { - static struct dentry_operations cgroup_dops = { + static const struct dentry_operations cgroup_dops = { .d_iput = cgroup_diput, }; @@ -1668,7 +1764,7 @@ static int cgroup_create_file(struct dentry *dentry, int mode, * @mode: mode to set on new directory. */ static int cgroup_create_dir(struct cgroup *cgrp, struct dentry *dentry, - int mode) + mode_t mode) { struct dentry *parent; int error = 0; @@ -1686,6 +1782,33 @@ static int cgroup_create_dir(struct cgroup *cgrp, struct dentry *dentry, return error; } +/** + * cgroup_file_mode - deduce file mode of a control file + * @cft: the control file in question + * + * returns cft->mode if ->mode is not 0 + * returns S_IRUGO|S_IWUSR if it has both a read and a write handler + * returns S_IRUGO if it has only a read handler + * returns S_IWUSR if it has only a write hander + */ +static mode_t cgroup_file_mode(const struct cftype *cft) +{ + mode_t mode = 0; + + if (cft->mode) + return cft->mode; + + if (cft->read || cft->read_u64 || cft->read_s64 || + cft->read_map || cft->read_seq_string) + mode |= S_IRUGO; + + if (cft->write || cft->write_u64 || cft->write_s64 || + cft->write_string || cft->trigger) + mode |= S_IWUSR; + + return mode; +} + int cgroup_add_file(struct cgroup *cgrp, struct cgroup_subsys *subsys, const struct cftype *cft) @@ -1693,6 +1816,7 @@ int cgroup_add_file(struct cgroup *cgrp, struct dentry *dir = cgrp->dentry; struct dentry *dentry; int error; + mode_t mode; char name[MAX_CGROUP_TYPE_NAMELEN + MAX_CFTYPE_NAME + 2] = { 0 }; if (subsys && !test_bit(ROOT_NOPREFIX, &cgrp->root->flags)) { @@ -1703,7 +1827,8 @@ int cgroup_add_file(struct cgroup *cgrp, BUG_ON(!mutex_is_locked(&dir->d_inode->i_mutex)); dentry = lookup_one_len(name, dir, strlen(name)); if (!IS_ERR(dentry)) { - error = cgroup_create_file(dentry, 0644 | S_IFREG, + mode = cgroup_file_mode(cft); + error = cgroup_create_file(dentry, mode | S_IFREG, cgrp->root->sb); if (!error) dentry->d_fsdata = (void *)cft; @@ -2091,12 +2216,30 @@ err: return ret; } +/* + * Cache pids for all threads in the same pid namespace that are + * opening the same "tasks" file. + */ +struct cgroup_pids { + /* The node in cgrp->pids_list */ + struct list_head list; + /* The cgroup those pids belong to */ + struct cgroup *cgrp; + /* The namepsace those pids belong to */ + struct pid_namespace *ns; + /* Array of process ids in the cgroup */ + pid_t *tasks_pids; + /* How many files are using the this tasks_pids array */ + int use_count; + /* Length of the current tasks_pids array */ + int length; +}; + static int cmppid(const void *a, const void *b) { return *(pid_t *)a - *(pid_t *)b; } - /* * seq_file methods for the "tasks" file. The seq_file position is the * next pid to display; the seq_file iterator is a pointer to the pid @@ -2111,45 +2254,47 @@ static void *cgroup_tasks_start(struct seq_file *s, loff_t *pos) * after a seek to the start). Use a binary-search to find the * next pid to display, if any */ - struct cgroup *cgrp = s->private; + struct cgroup_pids *cp = s->private; + struct cgroup *cgrp = cp->cgrp; int index = 0, pid = *pos; int *iter; down_read(&cgrp->pids_mutex); if (pid) { - int end = cgrp->pids_length; + int end = cp->length; while (index < end) { int mid = (index + end) / 2; - if (cgrp->tasks_pids[mid] == pid) { + if (cp->tasks_pids[mid] == pid) { index = mid; break; - } else if (cgrp->tasks_pids[mid] <= pid) + } else if (cp->tasks_pids[mid] <= pid) index = mid + 1; else end = mid; } } /* If we're off the end of the array, we're done */ - if (index >= cgrp->pids_length) + if (index >= cp->length) return NULL; /* Update the abstract position to be the actual pid that we found */ - iter = cgrp->tasks_pids + index; + iter = cp->tasks_pids + index; *pos = *iter; return iter; } static void cgroup_tasks_stop(struct seq_file *s, void *v) { - struct cgroup *cgrp = s->private; + struct cgroup_pids *cp = s->private; + struct cgroup *cgrp = cp->cgrp; up_read(&cgrp->pids_mutex); } static void *cgroup_tasks_next(struct seq_file *s, void *v, loff_t *pos) { - struct cgroup *cgrp = s->private; + struct cgroup_pids *cp = s->private; int *p = v; - int *end = cgrp->tasks_pids + cgrp->pids_length; + int *end = cp->tasks_pids + cp->length; /* * Advance to the next pid in the array. If this goes off the @@ -2176,26 +2321,33 @@ static struct seq_operations cgroup_tasks_seq_operations = { .show = cgroup_tasks_show, }; -static void release_cgroup_pid_array(struct cgroup *cgrp) +static void release_cgroup_pid_array(struct cgroup_pids *cp) { + struct cgroup *cgrp = cp->cgrp; + down_write(&cgrp->pids_mutex); - BUG_ON(!cgrp->pids_use_count); - if (!--cgrp->pids_use_count) { - kfree(cgrp->tasks_pids); - cgrp->tasks_pids = NULL; - cgrp->pids_length = 0; + BUG_ON(!cp->use_count); + if (!--cp->use_count) { + list_del(&cp->list); + put_pid_ns(cp->ns); + kfree(cp->tasks_pids); + kfree(cp); } up_write(&cgrp->pids_mutex); } static int cgroup_tasks_release(struct inode *inode, struct file *file) { - struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); + struct seq_file *seq; + struct cgroup_pids *cp; if (!(file->f_mode & FMODE_READ)) return 0; - release_cgroup_pid_array(cgrp); + seq = file->private_data; + cp = seq->private; + + release_cgroup_pid_array(cp); return seq_release(inode, file); } @@ -2214,6 +2366,8 @@ static struct file_operations cgroup_tasks_operations = { static int cgroup_tasks_open(struct inode *unused, struct file *file) { struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); + struct pid_namespace *ns = current->nsproxy->pid_ns; + struct cgroup_pids *cp; pid_t *pidarray; int npids; int retval; @@ -2240,20 +2394,37 @@ static int cgroup_tasks_open(struct inode *unused, struct file *file) * array if necessary */ down_write(&cgrp->pids_mutex); - kfree(cgrp->tasks_pids); - cgrp->tasks_pids = pidarray; - cgrp->pids_length = npids; - cgrp->pids_use_count++; + + list_for_each_entry(cp, &cgrp->pids_list, list) { + if (ns == cp->ns) + goto found; + } + + cp = kzalloc(sizeof(*cp), GFP_KERNEL); + if (!cp) { + up_write(&cgrp->pids_mutex); + kfree(pidarray); + return -ENOMEM; + } + cp->cgrp = cgrp; + cp->ns = ns; + get_pid_ns(ns); + list_add(&cp->list, &cgrp->pids_list); +found: + kfree(cp->tasks_pids); + cp->tasks_pids = pidarray; + cp->length = npids; + cp->use_count++; up_write(&cgrp->pids_mutex); file->f_op = &cgroup_tasks_operations; retval = seq_open(file, &cgroup_tasks_seq_operations); if (retval) { - release_cgroup_pid_array(cgrp); + release_cgroup_pid_array(cp); return retval; } - ((struct seq_file *)file->private_data)->private = cgrp; + ((struct seq_file *)file->private_data)->private = cp; return 0; } @@ -2285,6 +2456,7 @@ static struct cftype files[] = { .write_u64 = cgroup_tasks_write, .release = cgroup_tasks_release, .private = FILE_TASKLIST, + .mode = S_IRUGO | S_IWUSR, }, { @@ -2324,6 +2496,17 @@ static int cgroup_populate_dir(struct cgroup *cgrp) if (ss->populate && (err = ss->populate(ss, cgrp)) < 0) return err; } + /* This cgroup is ready now */ + for_each_subsys(cgrp->root, ss) { + struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; + /* + * Update id->css pointer and make this css visible from + * CSS ID functions. This pointer will be dereferened + * from RCU-read-side without locks. + */ + if (css->id) + rcu_assign_pointer(css->id->css, css); + } return 0; } @@ -2335,6 +2518,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, css->cgroup = cgrp; atomic_set(&css->refcnt, 1); css->flags = 0; + css->id = NULL; if (cgrp == dummytop) set_bit(CSS_ROOT, &css->flags); BUG_ON(cgrp->subsys[ss->subsys_id]); @@ -2349,7 +2533,7 @@ static void cgroup_lock_hierarchy(struct cgroupfs_root *root) for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { struct cgroup_subsys *ss = subsys[i]; if (ss->root == root) - mutex_lock_nested(&ss->hierarchy_mutex, i); + mutex_lock(&ss->hierarchy_mutex); } } @@ -2373,7 +2557,7 @@ static void cgroup_unlock_hierarchy(struct cgroupfs_root *root) * Must be called with the mutex on the parent inode held */ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, - int mode) + mode_t mode) { struct cgroup *cgrp; struct cgroupfs_root *root = parent->root; @@ -2410,6 +2594,10 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, goto err_destroy; } init_cgroup_css(css, ss, cgrp); + if (ss->use_id) + if (alloc_css_id(ss, parent, cgrp)) + goto err_destroy; + /* At error, ->destroy() callback has to free assigned ID. */ } cgroup_lock_hierarchy(root); @@ -2434,7 +2622,9 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, err_remove: + cgroup_lock_hierarchy(root); list_del(&cgrp->sibling); + cgroup_unlock_hierarchy(root); root->number_of_cgroups--; err_destroy: @@ -2507,7 +2697,7 @@ static int cgroup_clear_css_refs(struct cgroup *cgrp) for_each_subsys(cgrp->root, ss) { struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; int refcnt; - do { + while (1) { /* We can only remove a CSS with a refcnt==1 */ refcnt = atomic_read(&css->refcnt); if (refcnt > 1) { @@ -2521,7 +2711,10 @@ static int cgroup_clear_css_refs(struct cgroup *cgrp) * css_tryget() to spin until we set the * CSS_REMOVED bits or abort */ - } while (atomic_cmpxchg(&css->refcnt, refcnt, 0) != refcnt); + if (atomic_cmpxchg(&css->refcnt, refcnt, 0) == refcnt) + break; + cpu_relax(); + } } done: for_each_subsys(cgrp->root, ss) { @@ -2547,9 +2740,11 @@ static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) struct cgroup *cgrp = dentry->d_fsdata; struct dentry *d; struct cgroup *parent; + DEFINE_WAIT(wait); + int ret; /* the vfs holds both inode->i_mutex already */ - +again: mutex_lock(&cgroup_mutex); if (atomic_read(&cgrp->count) != 0) { mutex_unlock(&cgroup_mutex); @@ -2562,20 +2757,51 @@ static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) mutex_unlock(&cgroup_mutex); /* + * In general, subsystem has no css->refcnt after pre_destroy(). But + * in racy cases, subsystem may have to get css->refcnt after + * pre_destroy() and it makes rmdir return with -EBUSY. This sometimes + * make rmdir return -EBUSY too often. To avoid that, we use waitqueue + * for cgroup's rmdir. CGRP_WAIT_ON_RMDIR is for synchronizing rmdir + * and subsystem's reference count handling. Please see css_get/put + * and css_tryget() and cgroup_wakeup_rmdir_waiter() implementation. + */ + set_bit(CGRP_WAIT_ON_RMDIR, &cgrp->flags); + + /* * Call pre_destroy handlers of subsys. Notify subsystems * that rmdir() request comes. */ - cgroup_call_pre_destroy(cgrp); + ret = cgroup_call_pre_destroy(cgrp); + if (ret) { + clear_bit(CGRP_WAIT_ON_RMDIR, &cgrp->flags); + return ret; + } mutex_lock(&cgroup_mutex); parent = cgrp->parent; - - if (atomic_read(&cgrp->count) - || !list_empty(&cgrp->children) - || !cgroup_clear_css_refs(cgrp)) { + if (atomic_read(&cgrp->count) || !list_empty(&cgrp->children)) { + clear_bit(CGRP_WAIT_ON_RMDIR, &cgrp->flags); mutex_unlock(&cgroup_mutex); return -EBUSY; } + prepare_to_wait(&cgroup_rmdir_waitq, &wait, TASK_INTERRUPTIBLE); + if (!cgroup_clear_css_refs(cgrp)) { + mutex_unlock(&cgroup_mutex); + /* + * Because someone may call cgroup_wakeup_rmdir_waiter() before + * prepare_to_wait(), we need to check this flag. + */ + if (test_bit(CGRP_WAIT_ON_RMDIR, &cgrp->flags)) + schedule(); + finish_wait(&cgroup_rmdir_waitq, &wait); + clear_bit(CGRP_WAIT_ON_RMDIR, &cgrp->flags); + if (signal_pending(current)) + return -EINTR; + goto again; + } + /* NO css_tryget() can success after here. */ + finish_wait(&cgroup_rmdir_waitq, &wait); + clear_bit(CGRP_WAIT_ON_RMDIR, &cgrp->flags); spin_lock(&release_list_lock); set_bit(CGRP_REMOVED, &cgrp->flags); @@ -2630,6 +2856,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) BUG_ON(!list_empty(&init_task.tasks)); mutex_init(&ss->hierarchy_mutex); + lockdep_set_class(&ss->hierarchy_mutex, &ss->subsys_key); ss->active = 1; } @@ -2699,6 +2926,8 @@ int __init cgroup_init(void) struct cgroup_subsys *ss = subsys[i]; if (!ss->early_init) cgroup_init_subsys(ss); + if (ss->use_id) + cgroup_subsys_init_idr(ss); } /* Add init_css_set to the hash table */ @@ -2991,20 +3220,21 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, mutex_unlock(&cgroup_mutex); return 0; } - task_lock(tsk); - cg = tsk->cgroups; - parent = task_cgroup(tsk, subsys->subsys_id); /* Pin the hierarchy */ - if (!atomic_inc_not_zero(&parent->root->sb->s_active)) { + if (!atomic_inc_not_zero(&root->sb->s_active)) { /* We race with the final deactivate_super() */ mutex_unlock(&cgroup_mutex); return 0; } /* Keep the cgroup alive */ + task_lock(tsk); + parent = task_cgroup(tsk, subsys->subsys_id); + cg = tsk->cgroups; get_css_set(cg); task_unlock(tsk); + mutex_unlock(&cgroup_mutex); /* Now do the VFS work to create a cgroup */ @@ -3043,7 +3273,7 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, mutex_unlock(&inode->i_mutex); put_css_set(cg); - deactivate_super(parent->root->sb); + deactivate_super(root->sb); /* The cgroup is still accessible in the VFS, but * we're not going to try to rmdir() it at this * point. */ @@ -3069,23 +3299,24 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, mutex_lock(&cgroup_mutex); put_css_set(cg); mutex_unlock(&cgroup_mutex); - deactivate_super(parent->root->sb); + deactivate_super(root->sb); return ret; } /** - * cgroup_is_descendant - see if @cgrp is a descendant of current task's cgrp + * cgroup_is_descendant - see if @cgrp is a descendant of @task's cgrp * @cgrp: the cgroup in question + * @task: the task in question * - * See if @cgrp is a descendant of the current task's cgroup in - * the appropriate hierarchy. + * See if @cgrp is a descendant of @task's cgroup in the appropriate + * hierarchy. * * If we are sending in dummytop, then presumably we are creating * the top cgroup in the subsystem. * * Called only by the ns (nsproxy) cgroup. */ -int cgroup_is_descendant(const struct cgroup *cgrp) +int cgroup_is_descendant(const struct cgroup *cgrp, struct task_struct *task) { int ret; struct cgroup *target; @@ -3095,7 +3326,7 @@ int cgroup_is_descendant(const struct cgroup *cgrp) return 1; get_first_subsys(cgrp, NULL, &subsys_id); - target = task_cgroup(current, subsys_id); + target = task_cgroup(task, subsys_id); while (cgrp != target && cgrp!= cgrp->top_cgroup) cgrp = cgrp->parent; ret = (cgrp == target); @@ -3128,10 +3359,12 @@ void __css_put(struct cgroup_subsys_state *css) { struct cgroup *cgrp = css->cgroup; rcu_read_lock(); - if ((atomic_dec_return(&css->refcnt) == 1) && - notify_on_release(cgrp)) { - set_bit(CGRP_RELEASABLE, &cgrp->flags); - check_for_release(cgrp); + if (atomic_dec_return(&css->refcnt) == 1) { + if (notify_on_release(cgrp)) { + set_bit(CGRP_RELEASABLE, &cgrp->flags); + check_for_release(cgrp); + } + cgroup_wakeup_rmdir_waiter(cgrp); } rcu_read_unlock(); } @@ -3231,3 +3464,232 @@ static int __init cgroup_disable(char *str) return 1; } __setup("cgroup_disable=", cgroup_disable); + +/* + * Functons for CSS ID. + */ + +/* + *To get ID other than 0, this should be called when !cgroup_is_removed(). + */ +unsigned short css_id(struct cgroup_subsys_state *css) +{ + struct css_id *cssid = rcu_dereference(css->id); + + if (cssid) + return cssid->id; + return 0; +} + +unsigned short css_depth(struct cgroup_subsys_state *css) +{ + struct css_id *cssid = rcu_dereference(css->id); + + if (cssid) + return cssid->depth; + return 0; +} + +bool css_is_ancestor(struct cgroup_subsys_state *child, + const struct cgroup_subsys_state *root) +{ + struct css_id *child_id = rcu_dereference(child->id); + struct css_id *root_id = rcu_dereference(root->id); + + if (!child_id || !root_id || (child_id->depth < root_id->depth)) + return false; + return child_id->stack[root_id->depth] == root_id->id; +} + +static void __free_css_id_cb(struct rcu_head *head) +{ + struct css_id *id; + + id = container_of(head, struct css_id, rcu_head); + kfree(id); +} + +void free_css_id(struct cgroup_subsys *ss, struct cgroup_subsys_state *css) +{ + struct css_id *id = css->id; + /* When this is called before css_id initialization, id can be NULL */ + if (!id) + return; + + BUG_ON(!ss->use_id); + + rcu_assign_pointer(id->css, NULL); + rcu_assign_pointer(css->id, NULL); + spin_lock(&ss->id_lock); + idr_remove(&ss->idr, id->id); + spin_unlock(&ss->id_lock); + call_rcu(&id->rcu_head, __free_css_id_cb); +} + +/* + * This is called by init or create(). Then, calls to this function are + * always serialized (By cgroup_mutex() at create()). + */ + +static struct css_id *get_new_cssid(struct cgroup_subsys *ss, int depth) +{ + struct css_id *newid; + int myid, error, size; + + BUG_ON(!ss->use_id); + + size = sizeof(*newid) + sizeof(unsigned short) * (depth + 1); + newid = kzalloc(size, GFP_KERNEL); + if (!newid) + return ERR_PTR(-ENOMEM); + /* get id */ + if (unlikely(!idr_pre_get(&ss->idr, GFP_KERNEL))) { + error = -ENOMEM; + goto err_out; + } + spin_lock(&ss->id_lock); + /* Don't use 0. allocates an ID of 1-65535 */ + error = idr_get_new_above(&ss->idr, newid, 1, &myid); + spin_unlock(&ss->id_lock); + + /* Returns error when there are no free spaces for new ID.*/ + if (error) { + error = -ENOSPC; + goto err_out; + } + if (myid > CSS_ID_MAX) + goto remove_idr; + + newid->id = myid; + newid->depth = depth; + return newid; +remove_idr: + error = -ENOSPC; + spin_lock(&ss->id_lock); + idr_remove(&ss->idr, myid); + spin_unlock(&ss->id_lock); +err_out: + kfree(newid); + return ERR_PTR(error); + +} + +static int __init cgroup_subsys_init_idr(struct cgroup_subsys *ss) +{ + struct css_id *newid; + struct cgroup_subsys_state *rootcss; + + spin_lock_init(&ss->id_lock); + idr_init(&ss->idr); + + rootcss = init_css_set.subsys[ss->subsys_id]; + newid = get_new_cssid(ss, 0); + if (IS_ERR(newid)) + return PTR_ERR(newid); + + newid->stack[0] = newid->id; + newid->css = rootcss; + rootcss->id = newid; + return 0; +} + +static int alloc_css_id(struct cgroup_subsys *ss, struct cgroup *parent, + struct cgroup *child) +{ + int subsys_id, i, depth = 0; + struct cgroup_subsys_state *parent_css, *child_css; + struct css_id *child_id, *parent_id = NULL; + + subsys_id = ss->subsys_id; + parent_css = parent->subsys[subsys_id]; + child_css = child->subsys[subsys_id]; + depth = css_depth(parent_css) + 1; + parent_id = parent_css->id; + + child_id = get_new_cssid(ss, depth); + if (IS_ERR(child_id)) + return PTR_ERR(child_id); + + for (i = 0; i < depth; i++) + child_id->stack[i] = parent_id->stack[i]; + child_id->stack[depth] = child_id->id; + /* + * child_id->css pointer will be set after this cgroup is available + * see cgroup_populate_dir() + */ + rcu_assign_pointer(child_css->id, child_id); + + return 0; +} + +/** + * css_lookup - lookup css by id + * @ss: cgroup subsys to be looked into. + * @id: the id + * + * Returns pointer to cgroup_subsys_state if there is valid one with id. + * NULL if not. Should be called under rcu_read_lock() + */ +struct cgroup_subsys_state *css_lookup(struct cgroup_subsys *ss, int id) +{ + struct css_id *cssid = NULL; + + BUG_ON(!ss->use_id); + cssid = idr_find(&ss->idr, id); + + if (unlikely(!cssid)) + return NULL; + + return rcu_dereference(cssid->css); +} + +/** + * css_get_next - lookup next cgroup under specified hierarchy. + * @ss: pointer to subsystem + * @id: current position of iteration. + * @root: pointer to css. search tree under this. + * @foundid: position of found object. + * + * Search next css under the specified hierarchy of rootid. Calling under + * rcu_read_lock() is necessary. Returns NULL if it reaches the end. + */ +struct cgroup_subsys_state * +css_get_next(struct cgroup_subsys *ss, int id, + struct cgroup_subsys_state *root, int *foundid) +{ + struct cgroup_subsys_state *ret = NULL; + struct css_id *tmp; + int tmpid; + int rootid = css_id(root); + int depth = css_depth(root); + + if (!rootid) + return NULL; + + BUG_ON(!ss->use_id); + /* fill start point for scan */ + tmpid = id; + while (1) { + /* + * scan next entry from bitmap(tree), tmpid is updated after + * idr_get_next(). + */ + spin_lock(&ss->id_lock); + tmp = idr_get_next(&ss->idr, &tmpid); + spin_unlock(&ss->id_lock); + + if (!tmp) + break; + if (tmp->depth >= depth && tmp->stack[depth] == rootid) { + ret = rcu_dereference(tmp->css); + if (ret) { + *foundid = tmpid; + break; + } + } + /* continue to scan from next id */ + tmpid = tmpid + 1; + } + return ret; +} + diff --git a/kernel/cgroup_debug.c b/kernel/cgroup_debug.c index daca6209202..0c92d797baa 100644 --- a/kernel/cgroup_debug.c +++ b/kernel/cgroup_debug.c @@ -40,9 +40,7 @@ static u64 taskcount_read(struct cgroup *cont, struct cftype *cft) { u64 count; - cgroup_lock(); count = cgroup_task_count(cont); - cgroup_unlock(); return count; } diff --git a/kernel/compat.c b/kernel/compat.c index 42d56544460..f6c204f07ea 100644 --- a/kernel/compat.c +++ b/kernel/compat.c @@ -882,6 +882,17 @@ compat_sys_rt_sigtimedwait (compat_sigset_t __user *uthese, } +asmlinkage long +compat_sys_rt_tgsigqueueinfo(compat_pid_t tgid, compat_pid_t pid, int sig, + struct compat_siginfo __user *uinfo) +{ + siginfo_t info; + + if (copy_siginfo_from_user32(&info, uinfo)) + return -EFAULT; + return do_rt_tgsigqueueinfo(tgid, pid, sig, &info); +} + #ifdef __ARCH_WANT_COMPAT_SYS_TIME /* compat_time_t is a 32 bit "long" and needs to get converted. */ diff --git a/kernel/cpu.c b/kernel/cpu.c index 79e40f00dcb..8ce10043e4a 100644 --- a/kernel/cpu.c +++ b/kernel/cpu.c @@ -34,14 +34,11 @@ static struct { * an ongoing cpu hotplug operation. */ int refcount; -} cpu_hotplug; - -void __init cpu_hotplug_init(void) -{ - cpu_hotplug.active_writer = NULL; - mutex_init(&cpu_hotplug.lock); - cpu_hotplug.refcount = 0; -} +} cpu_hotplug = { + .active_writer = NULL, + .lock = __MUTEX_INITIALIZER(cpu_hotplug.lock), + .refcount = 0, +}; #ifdef CONFIG_HOTPLUG_CPU @@ -281,7 +278,7 @@ int __ref cpu_down(unsigned int cpu) goto out; } - cpu_clear(cpu, cpu_active_map); + set_cpu_active(cpu, false); /* * Make sure the all cpus did the reschedule and are not @@ -296,7 +293,7 @@ int __ref cpu_down(unsigned int cpu) err = _cpu_down(cpu, 0); if (cpu_online(cpu)) - cpu_set(cpu, cpu_active_map); + set_cpu_active(cpu, true); out: cpu_maps_update_done(); @@ -333,7 +330,7 @@ static int __cpuinit _cpu_up(unsigned int cpu, int tasks_frozen) goto out_notify; BUG_ON(!cpu_online(cpu)); - cpu_set(cpu, cpu_active_map); + set_cpu_active(cpu, true); /* Now call notifier in preparation. */ raw_notifier_call_chain(&cpu_chain, CPU_ONLINE | mod, hcpu); diff --git a/kernel/cpuset.c b/kernel/cpuset.c index a85678865c5..7e75a41bd50 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -61,6 +61,14 @@ #include <linux/cgroup.h> /* + * Workqueue for cpuset related tasks. + * + * Using kevent workqueue may cause deadlock when memory_migrate + * is set. So we create a separate workqueue thread for cpuset. + */ +static struct workqueue_struct *cpuset_wq; + +/* * Tracks how many cpusets are currently defined in system. * When there is only one cpuset (the root cpuset) we can * short circuit some hooks. @@ -89,12 +97,6 @@ struct cpuset { struct cpuset *parent; /* my parent */ - /* - * Copy of global cpuset_mems_generation as of the most - * recent time this cpuset changed its mems_allowed. - */ - int mems_generation; - struct fmeter fmeter; /* memory_pressure filter */ /* partition number for rebuild_sched_domains() */ @@ -120,10 +122,6 @@ static inline struct cpuset *task_cs(struct task_struct *task) return container_of(task_subsys_state(task, cpuset_subsys_id), struct cpuset, css); } -struct cpuset_hotplug_scanner { - struct cgroup_scanner scan; - struct cgroup *to; -}; /* bits in struct cpuset flags field */ typedef enum { @@ -172,27 +170,6 @@ static inline int is_spread_slab(const struct cpuset *cs) return test_bit(CS_SPREAD_SLAB, &cs->flags); } -/* - * Increment this integer everytime any cpuset changes its - * mems_allowed value. Users of cpusets can track this generation - * number, and avoid having to lock and reload mems_allowed unless - * the cpuset they're using changes generation. - * - * A single, global generation is needed because cpuset_attach_task() could - * reattach a task to a different cpuset, which must not have its - * generation numbers aliased with those of that tasks previous cpuset. - * - * Generations are needed for mems_allowed because one task cannot - * modify another's memory placement. So we must enable every task, - * on every visit to __alloc_pages(), to efficiently check whether - * its current->cpuset->mems_allowed has changed, requiring an update - * of its current->mems_allowed. - * - * Since writes to cpuset_mems_generation are guarded by the cgroup lock - * there is no need to mark it atomic. - */ -static int cpuset_mems_generation; - static struct cpuset top_cpuset = { .flags = ((1 << CS_CPU_EXCLUSIVE) | (1 << CS_MEM_EXCLUSIVE)), }; @@ -224,8 +201,9 @@ static struct cpuset top_cpuset = { * If a task is only holding callback_mutex, then it has read-only * access to cpusets. * - * The task_struct fields mems_allowed and mems_generation may only - * be accessed in the context of that task, so require no locks. + * Now, the task_struct fields mems_allowed and mempolicy may be changed + * by other task, we use alloc_lock in the task_struct fields to protect + * them. * * The cpuset_common_file_read() handlers only hold callback_mutex across * small pieces of code, such as when reading out possibly multi-word @@ -327,75 +305,22 @@ static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask) BUG_ON(!nodes_intersects(*pmask, node_states[N_HIGH_MEMORY])); } -/** - * cpuset_update_task_memory_state - update task memory placement - * - * If the current tasks cpusets mems_allowed changed behind our - * backs, update current->mems_allowed, mems_generation and task NUMA - * mempolicy to the new value. - * - * Task mempolicy is updated by rebinding it relative to the - * current->cpuset if a task has its memory placement changed. - * Do not call this routine if in_interrupt(). - * - * Call without callback_mutex or task_lock() held. May be - * called with or without cgroup_mutex held. Thanks in part to - * 'the_top_cpuset_hack', the task's cpuset pointer will never - * be NULL. This routine also might acquire callback_mutex during - * call. - * - * Reading current->cpuset->mems_generation doesn't need task_lock - * to guard the current->cpuset derefence, because it is guarded - * from concurrent freeing of current->cpuset using RCU. - * - * The rcu_dereference() is technically probably not needed, - * as I don't actually mind if I see a new cpuset pointer but - * an old value of mems_generation. However this really only - * matters on alpha systems using cpusets heavily. If I dropped - * that rcu_dereference(), it would save them a memory barrier. - * For all other arch's, rcu_dereference is a no-op anyway, and for - * alpha systems not using cpusets, another planned optimization, - * avoiding the rcu critical section for tasks in the root cpuset - * which is statically allocated, so can't vanish, will make this - * irrelevant. Better to use RCU as intended, than to engage in - * some cute trick to save a memory barrier that is impossible to - * test, for alpha systems using cpusets heavily, which might not - * even exist. - * - * This routine is needed to update the per-task mems_allowed data, - * within the tasks context, when it is trying to allocate memory - * (in various mm/mempolicy.c routines) and notices that some other - * task has been modifying its cpuset. +/* + * update task's spread flag if cpuset's page/slab spread flag is set + * + * Called with callback_mutex/cgroup_mutex held */ - -void cpuset_update_task_memory_state(void) +static void cpuset_update_task_spread_flag(struct cpuset *cs, + struct task_struct *tsk) { - int my_cpusets_mem_gen; - struct task_struct *tsk = current; - struct cpuset *cs; - - rcu_read_lock(); - my_cpusets_mem_gen = task_cs(tsk)->mems_generation; - rcu_read_unlock(); - - if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) { - mutex_lock(&callback_mutex); - task_lock(tsk); - cs = task_cs(tsk); /* Maybe changed when task not locked */ - guarantee_online_mems(cs, &tsk->mems_allowed); - tsk->cpuset_mems_generation = cs->mems_generation; - if (is_spread_page(cs)) - tsk->flags |= PF_SPREAD_PAGE; - else - tsk->flags &= ~PF_SPREAD_PAGE; - if (is_spread_slab(cs)) - tsk->flags |= PF_SPREAD_SLAB; - else - tsk->flags &= ~PF_SPREAD_SLAB; - task_unlock(tsk); - mutex_unlock(&callback_mutex); - mpol_rebind_task(tsk, &tsk->mems_allowed); - } + if (is_spread_page(cs)) + tsk->flags |= PF_SPREAD_PAGE; + else + tsk->flags &= ~PF_SPREAD_PAGE; + if (is_spread_slab(cs)) + tsk->flags |= PF_SPREAD_SLAB; + else + tsk->flags &= ~PF_SPREAD_SLAB; } /* @@ -513,6 +438,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) return 0; } +#ifdef CONFIG_SMP /* * Helper routine for generate_sched_domains(). * Do cpusets a, b have overlapping cpus_allowed masks? @@ -807,6 +733,18 @@ static void do_rebuild_sched_domains(struct work_struct *unused) put_online_cpus(); } +#else /* !CONFIG_SMP */ +static void do_rebuild_sched_domains(struct work_struct *unused) +{ +} + +static int generate_sched_domains(struct cpumask **domains, + struct sched_domain_attr **attributes) +{ + *domains = NULL; + return 1; +} +#endif /* CONFIG_SMP */ static DECLARE_WORK(rebuild_sched_domains_work, do_rebuild_sched_domains); @@ -831,7 +769,7 @@ static DECLARE_WORK(rebuild_sched_domains_work, do_rebuild_sched_domains); */ static void async_rebuild_sched_domains(void) { - schedule_work(&rebuild_sched_domains_work); + queue_work(cpuset_wq, &rebuild_sched_domains_work); } /* @@ -990,14 +928,6 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, * other task, the task_struct mems_allowed that we are hacking * is for our current task, which must allocate new pages for that * migrating memory region. - * - * We call cpuset_update_task_memory_state() before hacking - * our tasks mems_allowed, so that we are assured of being in - * sync with our tasks cpuset, and in particular, callbacks to - * cpuset_update_task_memory_state() from nested page allocations - * won't see any mismatch of our cpuset and task mems_generation - * values, so won't overwrite our hacked tasks mems_allowed - * nodemask. */ static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from, @@ -1005,17 +935,64 @@ static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from, { struct task_struct *tsk = current; - cpuset_update_task_memory_state(); - - mutex_lock(&callback_mutex); tsk->mems_allowed = *to; - mutex_unlock(&callback_mutex); do_migrate_pages(mm, from, to, MPOL_MF_MOVE_ALL); - mutex_lock(&callback_mutex); guarantee_online_mems(task_cs(tsk),&tsk->mems_allowed); - mutex_unlock(&callback_mutex); +} + +/* + * cpuset_change_task_nodemask - change task's mems_allowed and mempolicy + * @tsk: the task to change + * @newmems: new nodes that the task will be set + * + * In order to avoid seeing no nodes if the old and new nodes are disjoint, + * we structure updates as setting all new allowed nodes, then clearing newly + * disallowed ones. + * + * Called with task's alloc_lock held + */ +static void cpuset_change_task_nodemask(struct task_struct *tsk, + nodemask_t *newmems) +{ + nodes_or(tsk->mems_allowed, tsk->mems_allowed, *newmems); + mpol_rebind_task(tsk, &tsk->mems_allowed); + mpol_rebind_task(tsk, newmems); + tsk->mems_allowed = *newmems; +} + +/* + * Update task's mems_allowed and rebind its mempolicy and vmas' mempolicy + * of it to cpuset's new mems_allowed, and migrate pages to new nodes if + * memory_migrate flag is set. Called with cgroup_mutex held. + */ +static void cpuset_change_nodemask(struct task_struct *p, + struct cgroup_scanner *scan) +{ + struct mm_struct *mm; + struct cpuset *cs; + int migrate; + const nodemask_t *oldmem = scan->data; + nodemask_t newmems; + + cs = cgroup_cs(scan->cg); + guarantee_online_mems(cs, &newmems); + + task_lock(p); + cpuset_change_task_nodemask(p, &newmems); + task_unlock(p); + + mm = get_task_mm(p); + if (!mm) + return; + + migrate = is_memory_migrate(cs); + + mpol_rebind_mm(mm, &cs->mems_allowed); + if (migrate) + cpuset_migrate_mm(mm, oldmem, &cs->mems_allowed); + mmput(mm); } static void *cpuset_being_rebound; @@ -1024,104 +1001,48 @@ static void *cpuset_being_rebound; * update_tasks_nodemask - Update the nodemasks of tasks in the cpuset. * @cs: the cpuset in which each task's mems_allowed mask needs to be changed * @oldmem: old mems_allowed of cpuset cs + * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks() * * Called with cgroup_mutex held - * Return 0 if successful, -errno if not. + * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0 + * if @heap != NULL. */ -static int update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem) +static void update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem, + struct ptr_heap *heap) { - struct task_struct *p; - struct mm_struct **mmarray; - int i, n, ntasks; - int migrate; - int fudge; - struct cgroup_iter it; - int retval; + struct cgroup_scanner scan; cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ - fudge = 10; /* spare mmarray[] slots */ - fudge += cpumask_weight(cs->cpus_allowed);/* imagine 1 fork-bomb/cpu */ - retval = -ENOMEM; - - /* - * Allocate mmarray[] to hold mm reference for each task - * in cpuset cs. Can't kmalloc GFP_KERNEL while holding - * tasklist_lock. We could use GFP_ATOMIC, but with a - * few more lines of code, we can retry until we get a big - * enough mmarray[] w/o using GFP_ATOMIC. - */ - while (1) { - ntasks = cgroup_task_count(cs->css.cgroup); /* guess */ - ntasks += fudge; - mmarray = kmalloc(ntasks * sizeof(*mmarray), GFP_KERNEL); - if (!mmarray) - goto done; - read_lock(&tasklist_lock); /* block fork */ - if (cgroup_task_count(cs->css.cgroup) <= ntasks) - break; /* got enough */ - read_unlock(&tasklist_lock); /* try again */ - kfree(mmarray); - } - - n = 0; - - /* Load up mmarray[] with mm reference for each task in cpuset. */ - cgroup_iter_start(cs->css.cgroup, &it); - while ((p = cgroup_iter_next(cs->css.cgroup, &it))) { - struct mm_struct *mm; - - if (n >= ntasks) { - printk(KERN_WARNING - "Cpuset mempolicy rebind incomplete.\n"); - break; - } - mm = get_task_mm(p); - if (!mm) - continue; - mmarray[n++] = mm; - } - cgroup_iter_end(cs->css.cgroup, &it); - read_unlock(&tasklist_lock); + scan.cg = cs->css.cgroup; + scan.test_task = NULL; + scan.process_task = cpuset_change_nodemask; + scan.heap = heap; + scan.data = (nodemask_t *)oldmem; /* - * Now that we've dropped the tasklist spinlock, we can - * rebind the vma mempolicies of each mm in mmarray[] to their - * new cpuset, and release that mm. The mpol_rebind_mm() - * call takes mmap_sem, which we couldn't take while holding - * tasklist_lock. Forks can happen again now - the mpol_dup() - * cpuset_being_rebound check will catch such forks, and rebind - * their vma mempolicies too. Because we still hold the global - * cgroup_mutex, we know that no other rebind effort will - * be contending for the global variable cpuset_being_rebound. + * The mpol_rebind_mm() call takes mmap_sem, which we couldn't + * take while holding tasklist_lock. Forks can happen - the + * mpol_dup() cpuset_being_rebound check will catch such forks, + * and rebind their vma mempolicies too. Because we still hold + * the global cgroup_mutex, we know that no other rebind effort + * will be contending for the global variable cpuset_being_rebound. * It's ok if we rebind the same mm twice; mpol_rebind_mm() * is idempotent. Also migrate pages in each mm to new nodes. */ - migrate = is_memory_migrate(cs); - for (i = 0; i < n; i++) { - struct mm_struct *mm = mmarray[i]; - - mpol_rebind_mm(mm, &cs->mems_allowed); - if (migrate) - cpuset_migrate_mm(mm, oldmem, &cs->mems_allowed); - mmput(mm); - } + cgroup_scan_tasks(&scan); /* We're done rebinding vmas to this cpuset's new mems_allowed. */ - kfree(mmarray); cpuset_being_rebound = NULL; - retval = 0; -done: - return retval; } /* * Handle user request to change the 'mems' memory placement * of a cpuset. Needs to validate the request, update the - * cpusets mems_allowed and mems_generation, and for each - * task in the cpuset, rebind any vma mempolicies and if - * the cpuset is marked 'memory_migrate', migrate the tasks - * pages to the new memory. + * cpusets mems_allowed, and for each task in the cpuset, + * update mems_allowed and rebind task's mempolicy and any vma + * mempolicies and if the cpuset is marked 'memory_migrate', + * migrate the tasks pages to the new memory. * * Call with cgroup_mutex held. May take callback_mutex during call. * Will take tasklist_lock, scan tasklist for tasks in cpuset cs, @@ -1133,6 +1054,7 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, { nodemask_t oldmem; int retval; + struct ptr_heap heap; /* * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY]; @@ -1167,12 +1089,17 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, if (retval < 0) goto done; + retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL); + if (retval < 0) + goto done; + mutex_lock(&callback_mutex); cs->mems_allowed = trialcs->mems_allowed; - cs->mems_generation = cpuset_mems_generation++; mutex_unlock(&callback_mutex); - retval = update_tasks_nodemask(cs, &oldmem); + update_tasks_nodemask(cs, &oldmem, &heap); + + heap_free(&heap); done: return retval; } @@ -1184,8 +1111,10 @@ int current_cpuset_is_being_rebound(void) static int update_relax_domain_level(struct cpuset *cs, s64 val) { +#ifdef CONFIG_SMP if (val < -1 || val >= SD_LV_MAX) return -EINVAL; +#endif if (val != cs->relax_domain_level) { cs->relax_domain_level = val; @@ -1198,6 +1127,46 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val) } /* + * cpuset_change_flag - make a task's spread flags the same as its cpuset's + * @tsk: task to be updated + * @scan: struct cgroup_scanner containing the cgroup of the task + * + * Called by cgroup_scan_tasks() for each task in a cgroup. + * + * We don't need to re-check for the cgroup/cpuset membership, since we're + * holding cgroup_lock() at this point. + */ +static void cpuset_change_flag(struct task_struct *tsk, + struct cgroup_scanner *scan) +{ + cpuset_update_task_spread_flag(cgroup_cs(scan->cg), tsk); +} + +/* + * update_tasks_flags - update the spread flags of tasks in the cpuset. + * @cs: the cpuset in which each task's spread flags needs to be changed + * @heap: if NULL, defer allocating heap memory to cgroup_scan_tasks() + * + * Called with cgroup_mutex held + * + * The cgroup_scan_tasks() function will scan all the tasks in a cgroup, + * calling callback functions for each. + * + * No return value. It's guaranteed that cgroup_scan_tasks() always returns 0 + * if @heap != NULL. + */ +static void update_tasks_flags(struct cpuset *cs, struct ptr_heap *heap) +{ + struct cgroup_scanner scan; + + scan.cg = cs->css.cgroup; + scan.test_task = NULL; + scan.process_task = cpuset_change_flag; + scan.heap = heap; + cgroup_scan_tasks(&scan); +} + +/* * update_flag - read a 0 or a 1 in a file and update associated flag * bit: the bit to update (see cpuset_flagbits_t) * cs: the cpuset to update @@ -1210,8 +1179,10 @@ static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, int turning_on) { struct cpuset *trialcs; - int err; int balance_flag_changed; + int spread_flag_changed; + struct ptr_heap heap; + int err; trialcs = alloc_trial_cpuset(cs); if (!trialcs) @@ -1226,9 +1197,16 @@ static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, if (err < 0) goto out; + err = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL); + if (err < 0) + goto out; + balance_flag_changed = (is_sched_load_balance(cs) != is_sched_load_balance(trialcs)); + spread_flag_changed = ((is_spread_slab(cs) != is_spread_slab(trialcs)) + || (is_spread_page(cs) != is_spread_page(trialcs))); + mutex_lock(&callback_mutex); cs->flags = trialcs->flags; mutex_unlock(&callback_mutex); @@ -1236,6 +1214,9 @@ static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, if (!cpumask_empty(trialcs->cpus_allowed) && balance_flag_changed) async_rebuild_sched_domains(); + if (spread_flag_changed) + update_tasks_flags(cs, &heap); + heap_free(&heap); out: free_trial_cpuset(trialcs); return err; @@ -1347,19 +1328,22 @@ static int cpuset_can_attach(struct cgroup_subsys *ss, struct cgroup *cont, struct task_struct *tsk) { struct cpuset *cs = cgroup_cs(cont); - int ret = 0; if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed)) return -ENOSPC; - if (tsk->flags & PF_THREAD_BOUND) { - mutex_lock(&callback_mutex); - if (!cpumask_equal(&tsk->cpus_allowed, cs->cpus_allowed)) - ret = -EINVAL; - mutex_unlock(&callback_mutex); - } + /* + * Kthreads bound to specific cpus cannot be moved to a new cpuset; we + * cannot change their cpu affinity and isolating such threads by their + * set of allowed nodes is unnecessary. Thus, cpusets are not + * applicable for such threads. This prevents checking for success of + * set_cpus_allowed_ptr() on all attached tasks before cpus_allowed may + * be changed. + */ + if (tsk->flags & PF_THREAD_BOUND) + return -EINVAL; - return ret < 0 ? ret : security_task_setscheduler(tsk, 0, NULL); + return security_task_setscheduler(tsk, 0, NULL); } static void cpuset_attach(struct cgroup_subsys *ss, @@ -1374,15 +1358,20 @@ static void cpuset_attach(struct cgroup_subsys *ss, if (cs == &top_cpuset) { cpumask_copy(cpus_attach, cpu_possible_mask); + to = node_possible_map; } else { - mutex_lock(&callback_mutex); guarantee_online_cpus(cs, cpus_attach); - mutex_unlock(&callback_mutex); + guarantee_online_mems(cs, &to); } err = set_cpus_allowed_ptr(tsk, cpus_attach); if (err) return; + task_lock(tsk); + cpuset_change_task_nodemask(tsk, &to); + task_unlock(tsk); + cpuset_update_task_spread_flag(cs, tsk); + from = oldcs->mems_allowed; to = cs->mems_allowed; mm = get_task_mm(tsk); @@ -1444,11 +1433,9 @@ static int cpuset_write_u64(struct cgroup *cgrp, struct cftype *cft, u64 val) break; case FILE_SPREAD_PAGE: retval = update_flag(CS_SPREAD_PAGE, cs, val); - cs->mems_generation = cpuset_mems_generation++; break; case FILE_SPREAD_SLAB: retval = update_flag(CS_SPREAD_SLAB, cs, val); - cs->mems_generation = cpuset_mems_generation++; break; default: retval = -EINVAL; @@ -1698,6 +1685,7 @@ static struct cftype files[] = { .read_u64 = cpuset_read_u64, .write_u64 = cpuset_write_u64, .private = FILE_MEMORY_PRESSURE, + .mode = S_IRUGO, }, { @@ -1787,8 +1775,6 @@ static struct cgroup_subsys_state *cpuset_create( struct cpuset *parent; if (!cont->parent) { - /* This is early initialization for the top cgroup */ - top_cpuset.mems_generation = cpuset_mems_generation++; return &top_cpuset.css; } parent = cgroup_cs(cont->parent); @@ -1800,7 +1786,6 @@ static struct cgroup_subsys_state *cpuset_create( return ERR_PTR(-ENOMEM); } - cpuset_update_task_memory_state(); cs->flags = 0; if (is_spread_page(parent)) set_bit(CS_SPREAD_PAGE, &cs->flags); @@ -1809,7 +1794,6 @@ static struct cgroup_subsys_state *cpuset_create( set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); cpumask_clear(cs->cpus_allowed); nodes_clear(cs->mems_allowed); - cs->mems_generation = cpuset_mems_generation++; fmeter_init(&cs->fmeter); cs->relax_domain_level = -1; @@ -1828,8 +1812,6 @@ static void cpuset_destroy(struct cgroup_subsys *ss, struct cgroup *cont) { struct cpuset *cs = cgroup_cs(cont); - cpuset_update_task_memory_state(); - if (is_sched_load_balance(cs)) update_flag(CS_SCHED_LOAD_BALANCE, cs, 0); @@ -1850,21 +1832,6 @@ struct cgroup_subsys cpuset_subsys = { .early_init = 1, }; -/* - * cpuset_init_early - just enough so that the calls to - * cpuset_update_task_memory_state() in early init code - * are harmless. - */ - -int __init cpuset_init_early(void) -{ - alloc_bootmem_cpumask_var(&top_cpuset.cpus_allowed); - - top_cpuset.mems_generation = cpuset_mems_generation++; - return 0; -} - - /** * cpuset_init - initialize cpusets at system boot * @@ -1875,11 +1842,13 @@ int __init cpuset_init(void) { int err = 0; + if (!alloc_cpumask_var(&top_cpuset.cpus_allowed, GFP_KERNEL)) + BUG(); + cpumask_setall(top_cpuset.cpus_allowed); nodes_setall(top_cpuset.mems_allowed); fmeter_init(&top_cpuset.fmeter); - top_cpuset.mems_generation = cpuset_mems_generation++; set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags); top_cpuset.relax_domain_level = -1; @@ -1905,10 +1874,9 @@ int __init cpuset_init(void) static void cpuset_do_move_task(struct task_struct *tsk, struct cgroup_scanner *scan) { - struct cpuset_hotplug_scanner *chsp; + struct cgroup *new_cgroup = scan->data; - chsp = container_of(scan, struct cpuset_hotplug_scanner, scan); - cgroup_attach_task(chsp->to, tsk); + cgroup_attach_task(new_cgroup, tsk); } /** @@ -1924,15 +1892,15 @@ static void cpuset_do_move_task(struct task_struct *tsk, */ static void move_member_tasks_to_cpuset(struct cpuset *from, struct cpuset *to) { - struct cpuset_hotplug_scanner scan; + struct cgroup_scanner scan; - scan.scan.cg = from->css.cgroup; - scan.scan.test_task = NULL; /* select all tasks in cgroup */ - scan.scan.process_task = cpuset_do_move_task; - scan.scan.heap = NULL; - scan.to = to->css.cgroup; + scan.cg = from->css.cgroup; + scan.test_task = NULL; /* select all tasks in cgroup */ + scan.process_task = cpuset_do_move_task; + scan.heap = NULL; + scan.data = to->css.cgroup; - if (cgroup_scan_tasks(&scan.scan)) + if (cgroup_scan_tasks(&scan)) printk(KERN_ERR "move_member_tasks_to_cpuset: " "cgroup_scan_tasks failed\n"); } @@ -2025,7 +1993,7 @@ static void scan_for_empty_cpusets(struct cpuset *root) remove_tasks_in_empty_cpuset(cp); else { update_tasks_cpumask(cp, NULL); - update_tasks_nodemask(cp, &oldmems); + update_tasks_nodemask(cp, &oldmems, NULL); } } } @@ -2061,7 +2029,9 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb, } cgroup_lock(); + mutex_lock(&callback_mutex); cpumask_copy(top_cpuset.cpus_allowed, cpu_online_mask); + mutex_unlock(&callback_mutex); scan_for_empty_cpusets(&top_cpuset); ndoms = generate_sched_domains(&doms, &attr); cgroup_unlock(); @@ -2084,11 +2054,12 @@ static int cpuset_track_online_nodes(struct notifier_block *self, cgroup_lock(); switch (action) { case MEM_ONLINE: - top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; - break; case MEM_OFFLINE: + mutex_lock(&callback_mutex); top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; - scan_for_empty_cpusets(&top_cpuset); + mutex_unlock(&callback_mutex); + if (action == MEM_OFFLINE) + scan_for_empty_cpusets(&top_cpuset); break; default: break; @@ -2111,6 +2082,9 @@ void __init cpuset_init_smp(void) hotcpu_notifier(cpuset_track_online_cpus, 0); hotplug_memory_notifier(cpuset_track_online_nodes, 10); + + cpuset_wq = create_singlethread_workqueue("cpuset"); + BUG_ON(!cpuset_wq); } /** @@ -2195,26 +2169,24 @@ static const struct cpuset *nearest_hardwall_ancestor(const struct cpuset *cs) } /** - * cpuset_zone_allowed_softwall - Can we allocate on zone z's memory node? - * @z: is this zone on an allowed node? + * cpuset_node_allowed_softwall - Can we allocate on a memory node? + * @node: is this an allowed node? * @gfp_mask: memory allocation flags * - * If we're in interrupt, yes, we can always allocate. If - * __GFP_THISNODE is set, yes, we can always allocate. If zone - * z's node is in our tasks mems_allowed, yes. If it's not a - * __GFP_HARDWALL request and this zone's nodes is in the nearest - * hardwalled cpuset ancestor to this tasks cpuset, yes. - * If the task has been OOM killed and has access to memory reserves - * as specified by the TIF_MEMDIE flag, yes. + * If we're in interrupt, yes, we can always allocate. If __GFP_THISNODE is + * set, yes, we can always allocate. If node is in our task's mems_allowed, + * yes. If it's not a __GFP_HARDWALL request and this node is in the nearest + * hardwalled cpuset ancestor to this task's cpuset, yes. If the task has been + * OOM killed and has access to memory reserves as specified by the TIF_MEMDIE + * flag, yes. * Otherwise, no. * - * If __GFP_HARDWALL is set, cpuset_zone_allowed_softwall() - * reduces to cpuset_zone_allowed_hardwall(). Otherwise, - * cpuset_zone_allowed_softwall() might sleep, and might allow a zone - * from an enclosing cpuset. + * If __GFP_HARDWALL is set, cpuset_node_allowed_softwall() reduces to + * cpuset_node_allowed_hardwall(). Otherwise, cpuset_node_allowed_softwall() + * might sleep, and might allow a node from an enclosing cpuset. * - * cpuset_zone_allowed_hardwall() only handles the simpler case of - * hardwall cpusets, and never sleeps. + * cpuset_node_allowed_hardwall() only handles the simpler case of hardwall + * cpusets, and never sleeps. * * The __GFP_THISNODE placement logic is really handled elsewhere, * by forcibly using a zonelist starting at a specified node, and by @@ -2253,20 +2225,17 @@ static const struct cpuset *nearest_hardwall_ancestor(const struct cpuset *cs) * GFP_USER - only nodes in current tasks mems allowed ok. * * Rule: - * Don't call cpuset_zone_allowed_softwall if you can't sleep, unless you + * Don't call cpuset_node_allowed_softwall if you can't sleep, unless you * pass in the __GFP_HARDWALL flag set in gfp_flag, which disables * the code that might scan up ancestor cpusets and sleep. */ - -int __cpuset_zone_allowed_softwall(struct zone *z, gfp_t gfp_mask) +int __cpuset_node_allowed_softwall(int node, gfp_t gfp_mask) { - int node; /* node that zone z is on */ const struct cpuset *cs; /* current cpuset ancestors */ int allowed; /* is allocation in zone z allowed? */ if (in_interrupt() || (gfp_mask & __GFP_THISNODE)) return 1; - node = zone_to_nid(z); might_sleep_if(!(gfp_mask & __GFP_HARDWALL)); if (node_isset(node, current->mems_allowed)) return 1; @@ -2295,15 +2264,15 @@ int __cpuset_zone_allowed_softwall(struct zone *z, gfp_t gfp_mask) } /* - * cpuset_zone_allowed_hardwall - Can we allocate on zone z's memory node? - * @z: is this zone on an allowed node? + * cpuset_node_allowed_hardwall - Can we allocate on a memory node? + * @node: is this an allowed node? * @gfp_mask: memory allocation flags * - * If we're in interrupt, yes, we can always allocate. - * If __GFP_THISNODE is set, yes, we can always allocate. If zone - * z's node is in our tasks mems_allowed, yes. If the task has been - * OOM killed and has access to memory reserves as specified by the - * TIF_MEMDIE flag, yes. Otherwise, no. + * If we're in interrupt, yes, we can always allocate. If __GFP_THISNODE is + * set, yes, we can always allocate. If node is in our task's mems_allowed, + * yes. If the task has been OOM killed and has access to memory reserves as + * specified by the TIF_MEMDIE flag, yes. + * Otherwise, no. * * The __GFP_THISNODE placement logic is really handled elsewhere, * by forcibly using a zonelist starting at a specified node, and by @@ -2311,20 +2280,16 @@ int __cpuset_zone_allowed_softwall(struct zone *z, gfp_t gfp_mask) * any node on the zonelist except the first. By the time any such * calls get to this routine, we should just shut up and say 'yes'. * - * Unlike the cpuset_zone_allowed_softwall() variant, above, - * this variant requires that the zone be in the current tasks + * Unlike the cpuset_node_allowed_softwall() variant, above, + * this variant requires that the node be in the current task's * mems_allowed or that we're in interrupt. It does not scan up the * cpuset hierarchy for the nearest enclosing mem_exclusive cpuset. * It never sleeps. */ - -int __cpuset_zone_allowed_hardwall(struct zone *z, gfp_t gfp_mask) +int __cpuset_node_allowed_hardwall(int node, gfp_t gfp_mask) { - int node; /* node that zone z is on */ - if (in_interrupt() || (gfp_mask & __GFP_THISNODE)) return 1; - node = zone_to_nid(z); if (node_isset(node, current->mems_allowed)) return 1; /* diff --git a/kernel/cred.c b/kernel/cred.c index 3a039189d70..006fcab009d 100644 --- a/kernel/cred.c +++ b/kernel/cred.c @@ -18,6 +18,18 @@ #include <linux/cn_proc.h> #include "cred-internals.h" +#if 0 +#define kdebug(FMT, ...) \ + printk("[%-5.5s%5u] "FMT"\n", current->comm, current->pid ,##__VA_ARGS__) +#else +static inline __attribute__((format(printf, 1, 2))) +void no_printk(const char *fmt, ...) +{ +} +#define kdebug(FMT, ...) \ + no_printk("[%-5.5s%5u] "FMT"\n", current->comm, current->pid ,##__VA_ARGS__) +#endif + static struct kmem_cache *cred_jar; /* @@ -36,6 +48,10 @@ static struct thread_group_cred init_tgcred = { */ struct cred init_cred = { .usage = ATOMIC_INIT(4), +#ifdef CONFIG_DEBUG_CREDENTIALS + .subscribers = ATOMIC_INIT(2), + .magic = CRED_MAGIC, +#endif .securebits = SECUREBITS_DEFAULT, .cap_inheritable = CAP_INIT_INH_SET, .cap_permitted = CAP_FULL_SET, @@ -48,6 +64,31 @@ struct cred init_cred = { #endif }; +static inline void set_cred_subscribers(struct cred *cred, int n) +{ +#ifdef CONFIG_DEBUG_CREDENTIALS + atomic_set(&cred->subscribers, n); +#endif +} + +static inline int read_cred_subscribers(const struct cred *cred) +{ +#ifdef CONFIG_DEBUG_CREDENTIALS + return atomic_read(&cred->subscribers); +#else + return 0; +#endif +} + +static inline void alter_cred_subscribers(const struct cred *_cred, int n) +{ +#ifdef CONFIG_DEBUG_CREDENTIALS + struct cred *cred = (struct cred *) _cred; + + atomic_add(n, &cred->subscribers); +#endif +} + /* * Dispose of the shared task group credentials */ @@ -85,9 +126,22 @@ static void put_cred_rcu(struct rcu_head *rcu) { struct cred *cred = container_of(rcu, struct cred, rcu); + kdebug("put_cred_rcu(%p)", cred); + +#ifdef CONFIG_DEBUG_CREDENTIALS + if (cred->magic != CRED_MAGIC_DEAD || + atomic_read(&cred->usage) != 0 || + read_cred_subscribers(cred) != 0) + panic("CRED: put_cred_rcu() sees %p with" + " mag %x, put %p, usage %d, subscr %d\n", + cred, cred->magic, cred->put_addr, + atomic_read(&cred->usage), + read_cred_subscribers(cred)); +#else if (atomic_read(&cred->usage) != 0) panic("CRED: put_cred_rcu() sees %p with usage %d\n", cred, atomic_read(&cred->usage)); +#endif security_cred_free(cred); key_put(cred->thread_keyring); @@ -106,12 +160,90 @@ static void put_cred_rcu(struct rcu_head *rcu) */ void __put_cred(struct cred *cred) { + kdebug("__put_cred(%p{%d,%d})", cred, + atomic_read(&cred->usage), + read_cred_subscribers(cred)); + BUG_ON(atomic_read(&cred->usage) != 0); +#ifdef CONFIG_DEBUG_CREDENTIALS + BUG_ON(read_cred_subscribers(cred) != 0); + cred->magic = CRED_MAGIC_DEAD; + cred->put_addr = __builtin_return_address(0); +#endif + BUG_ON(cred == current->cred); + BUG_ON(cred == current->real_cred); call_rcu(&cred->rcu, put_cred_rcu); } EXPORT_SYMBOL(__put_cred); +/* + * Clean up a task's credentials when it exits + */ +void exit_creds(struct task_struct *tsk) +{ + struct cred *cred; + + kdebug("exit_creds(%u,%p,%p,{%d,%d})", tsk->pid, tsk->real_cred, tsk->cred, + atomic_read(&tsk->cred->usage), + read_cred_subscribers(tsk->cred)); + + cred = (struct cred *) tsk->real_cred; + tsk->real_cred = NULL; + validate_creds(cred); + alter_cred_subscribers(cred, -1); + put_cred(cred); + + cred = (struct cred *) tsk->cred; + tsk->cred = NULL; + validate_creds(cred); + alter_cred_subscribers(cred, -1); + put_cred(cred); + + cred = (struct cred *) tsk->replacement_session_keyring; + if (cred) { + tsk->replacement_session_keyring = NULL; + validate_creds(cred); + put_cred(cred); + } +} + +/* + * Allocate blank credentials, such that the credentials can be filled in at a + * later date without risk of ENOMEM. + */ +struct cred *cred_alloc_blank(void) +{ + struct cred *new; + + new = kmem_cache_zalloc(cred_jar, GFP_KERNEL); + if (!new) + return NULL; + +#ifdef CONFIG_KEYS + new->tgcred = kzalloc(sizeof(*new->tgcred), GFP_KERNEL); + if (!new->tgcred) { + kfree(new); + return NULL; + } + atomic_set(&new->tgcred->usage, 1); +#endif + + atomic_set(&new->usage, 1); + + if (security_cred_alloc_blank(new, GFP_KERNEL) < 0) + goto error; + +#ifdef CONFIG_DEBUG_CREDENTIALS + new->magic = CRED_MAGIC; +#endif + return new; + +error: + abort_creds(new); + return NULL; +} + /** * prepare_creds - Prepare a new set of credentials for modification * @@ -132,16 +264,19 @@ struct cred *prepare_creds(void) const struct cred *old; struct cred *new; - BUG_ON(atomic_read(&task->real_cred->usage) < 1); + validate_process_creds(); new = kmem_cache_alloc(cred_jar, GFP_KERNEL); if (!new) return NULL; + kdebug("prepare_creds() alloc %p", new); + old = task->cred; memcpy(new, old, sizeof(struct cred)); atomic_set(&new->usage, 1); + set_cred_subscribers(new, 0); get_group_info(new->group_info); get_uid(new->user); @@ -157,6 +292,7 @@ struct cred *prepare_creds(void) if (security_prepare_creds(new, old, GFP_KERNEL) < 0) goto error; + validate_creds(new); return new; error: @@ -167,7 +303,7 @@ EXPORT_SYMBOL(prepare_creds); /* * Prepare credentials for current to perform an execve() - * - The caller must hold current->cred_exec_mutex + * - The caller must hold current->cred_guard_mutex */ struct cred *prepare_exec_creds(void) { @@ -229,9 +365,12 @@ struct cred *prepare_usermodehelper_creds(void) if (!new) return NULL; + kdebug("prepare_usermodehelper_creds() alloc %p", new); + memcpy(new, &init_cred, sizeof(struct cred)); atomic_set(&new->usage, 1); + set_cred_subscribers(new, 0); get_group_info(new->group_info); get_uid(new->user); @@ -250,6 +389,7 @@ struct cred *prepare_usermodehelper_creds(void) #endif if (security_prepare_creds(new, &init_cred, GFP_ATOMIC) < 0) goto error; + validate_creds(new); BUG_ON(atomic_read(&new->usage) != 1); return new; @@ -276,7 +416,7 @@ int copy_creds(struct task_struct *p, unsigned long clone_flags) struct cred *new; int ret; - mutex_init(&p->cred_exec_mutex); + mutex_init(&p->cred_guard_mutex); if ( #ifdef CONFIG_KEYS @@ -286,6 +426,10 @@ int copy_creds(struct task_struct *p, unsigned long clone_flags) ) { p->real_cred = get_cred(p->cred); get_cred(p->cred); + alter_cred_subscribers(p->cred, 2); + kdebug("share_creds(%p{%d,%d})", + p->cred, atomic_read(&p->cred->usage), + read_cred_subscribers(p->cred)); atomic_inc(&p->cred->user->processes); return 0; } @@ -331,6 +475,8 @@ int copy_creds(struct task_struct *p, unsigned long clone_flags) atomic_inc(&new->user->processes); p->cred = p->real_cred = get_cred(new); + alter_cred_subscribers(new, 2); + validate_creds(new); return 0; error_put: @@ -355,13 +501,20 @@ error_put: int commit_creds(struct cred *new) { struct task_struct *task = current; - const struct cred *old; + const struct cred *old = task->real_cred; - BUG_ON(task->cred != task->real_cred); - BUG_ON(atomic_read(&task->real_cred->usage) < 2); + kdebug("commit_creds(%p{%d,%d})", new, + atomic_read(&new->usage), + read_cred_subscribers(new)); + + BUG_ON(task->cred != old); +#ifdef CONFIG_DEBUG_CREDENTIALS + BUG_ON(read_cred_subscribers(old) < 2); + validate_creds(old); + validate_creds(new); +#endif BUG_ON(atomic_read(&new->usage) < 1); - old = task->real_cred; security_commit_creds(new, old); get_cred(new); /* we will require a ref for the subj creds too */ @@ -390,12 +543,14 @@ int commit_creds(struct cred *new) * cheaply with the new uid cache, so if it matters * we should be checking for it. -DaveM */ + alter_cred_subscribers(new, 2); if (new->user != old->user) atomic_inc(&new->user->processes); rcu_assign_pointer(task->real_cred, new); rcu_assign_pointer(task->cred, new); if (new->user != old->user) atomic_dec(&old->user->processes); + alter_cred_subscribers(old, -2); sched_switch_user(task); @@ -428,6 +583,13 @@ EXPORT_SYMBOL(commit_creds); */ void abort_creds(struct cred *new) { + kdebug("abort_creds(%p{%d,%d})", new, + atomic_read(&new->usage), + read_cred_subscribers(new)); + +#ifdef CONFIG_DEBUG_CREDENTIALS + BUG_ON(read_cred_subscribers(new) != 0); +#endif BUG_ON(atomic_read(&new->usage) < 1); put_cred(new); } @@ -444,7 +606,20 @@ const struct cred *override_creds(const struct cred *new) { const struct cred *old = current->cred; - rcu_assign_pointer(current->cred, get_cred(new)); + kdebug("override_creds(%p{%d,%d})", new, + atomic_read(&new->usage), + read_cred_subscribers(new)); + + validate_creds(old); + validate_creds(new); + get_cred(new); + alter_cred_subscribers(new, 1); + rcu_assign_pointer(current->cred, new); + alter_cred_subscribers(old, -1); + + kdebug("override_creds() = %p{%d,%d}", old, + atomic_read(&old->usage), + read_cred_subscribers(old)); return old; } EXPORT_SYMBOL(override_creds); @@ -460,7 +635,15 @@ void revert_creds(const struct cred *old) { const struct cred *override = current->cred; + kdebug("revert_creds(%p{%d,%d})", old, + atomic_read(&old->usage), + read_cred_subscribers(old)); + + validate_creds(old); + validate_creds(override); + alter_cred_subscribers(old, 1); rcu_assign_pointer(current->cred, old); + alter_cred_subscribers(override, -1); put_cred(override); } EXPORT_SYMBOL(revert_creds); @@ -502,11 +685,15 @@ struct cred *prepare_kernel_cred(struct task_struct *daemon) if (!new) return NULL; + kdebug("prepare_kernel_cred() alloc %p", new); + if (daemon) old = get_task_cred(daemon); else old = get_cred(&init_cred); + validate_creds(old); + *new = *old; get_uid(new->user); get_group_info(new->group_info); @@ -526,7 +713,9 @@ struct cred *prepare_kernel_cred(struct task_struct *daemon) goto error; atomic_set(&new->usage, 1); + set_cred_subscribers(new, 0); put_cred(old); + validate_creds(new); return new; error: @@ -589,3 +778,95 @@ int set_create_files_as(struct cred *new, struct inode *inode) return security_kernel_create_files_as(new, inode); } EXPORT_SYMBOL(set_create_files_as); + +#ifdef CONFIG_DEBUG_CREDENTIALS + +/* + * dump invalid credentials + */ +static void dump_invalid_creds(const struct cred *cred, const char *label, + const struct task_struct *tsk) +{ + printk(KERN_ERR "CRED: %s credentials: %p %s%s%s\n", + label, cred, + cred == &init_cred ? "[init]" : "", + cred == tsk->real_cred ? "[real]" : "", + cred == tsk->cred ? "[eff]" : ""); + printk(KERN_ERR "CRED: ->magic=%x, put_addr=%p\n", + cred->magic, cred->put_addr); + printk(KERN_ERR "CRED: ->usage=%d, subscr=%d\n", + atomic_read(&cred->usage), + read_cred_subscribers(cred)); + printk(KERN_ERR "CRED: ->*uid = { %d,%d,%d,%d }\n", + cred->uid, cred->euid, cred->suid, cred->fsuid); + printk(KERN_ERR "CRED: ->*gid = { %d,%d,%d,%d }\n", + cred->gid, cred->egid, cred->sgid, cred->fsgid); +#ifdef CONFIG_SECURITY + printk(KERN_ERR "CRED: ->security is %p\n", cred->security); + if ((unsigned long) cred->security >= PAGE_SIZE && + (((unsigned long) cred->security & 0xffffff00) != + (POISON_FREE << 24 | POISON_FREE << 16 | POISON_FREE << 8))) + printk(KERN_ERR "CRED: ->security {%x, %x}\n", + ((u32*)cred->security)[0], + ((u32*)cred->security)[1]); +#endif +} + +/* + * report use of invalid credentials + */ +void __invalid_creds(const struct cred *cred, const char *file, unsigned line) +{ + printk(KERN_ERR "CRED: Invalid credentials\n"); + printk(KERN_ERR "CRED: At %s:%u\n", file, line); + dump_invalid_creds(cred, "Specified", current); + BUG(); +} +EXPORT_SYMBOL(__invalid_creds); + +/* + * check the credentials on a process + */ +void __validate_process_creds(struct task_struct *tsk, + const char *file, unsigned line) +{ + if (tsk->cred == tsk->real_cred) { + if (unlikely(read_cred_subscribers(tsk->cred) < 2 || + creds_are_invalid(tsk->cred))) + goto invalid_creds; + } else { + if (unlikely(read_cred_subscribers(tsk->real_cred) < 1 || + read_cred_subscribers(tsk->cred) < 1 || + creds_are_invalid(tsk->real_cred) || + creds_are_invalid(tsk->cred))) + goto invalid_creds; + } + return; + +invalid_creds: + printk(KERN_ERR "CRED: Invalid process credentials\n"); + printk(KERN_ERR "CRED: At %s:%u\n", file, line); + + dump_invalid_creds(tsk->real_cred, "Real", tsk); + if (tsk->cred != tsk->real_cred) + dump_invalid_creds(tsk->cred, "Effective", tsk); + else + printk(KERN_ERR "CRED: Effective creds == Real creds\n"); + BUG(); +} +EXPORT_SYMBOL(__validate_process_creds); + +/* + * check creds for do_exit() + */ +void validate_creds_for_do_exit(struct task_struct *tsk) +{ + kdebug("validate_creds_for_do_exit(%p,%p{%d,%d})", + tsk->real_cred, tsk->cred, + atomic_read(&tsk->cred->usage), + read_cred_subscribers(tsk->cred)); + + __validate_process_creds(tsk, __FILE__, __LINE__); +} + +#endif /* CONFIG_DEBUG_CREDENTIALS */ diff --git a/kernel/dma-coherent.c b/kernel/dma-coherent.c index 038707404b7..962a3b574f2 100644 --- a/kernel/dma-coherent.c +++ b/kernel/dma-coherent.c @@ -98,7 +98,7 @@ EXPORT_SYMBOL(dma_mark_declared_memory_occupied); * @size: size of requested memory area * @dma_handle: This will be filled with the correct dma handle * @ret: This pointer will be filled with the virtual address - * to allocated area. + * to allocated area. * * This function should be only called from per-arch dma_alloc_coherent() * to support allocation from per-device coherent memory pools. @@ -118,31 +118,32 @@ int dma_alloc_from_coherent(struct device *dev, ssize_t size, mem = dev->dma_mem; if (!mem) return 0; - if (unlikely(size > mem->size)) - return 0; + + *ret = NULL; + + if (unlikely(size > (mem->size << PAGE_SHIFT))) + goto err; pageno = bitmap_find_free_region(mem->bitmap, mem->size, order); - if (pageno >= 0) { - /* - * Memory was found in the per-device arena. - */ - *dma_handle = mem->device_base + (pageno << PAGE_SHIFT); - *ret = mem->virt_base + (pageno << PAGE_SHIFT); - memset(*ret, 0, size); - } else if (mem->flags & DMA_MEMORY_EXCLUSIVE) { - /* - * The per-device arena is exhausted and we are not - * permitted to fall back to generic memory. - */ - *ret = NULL; - } else { - /* - * The per-device arena is exhausted and we are - * permitted to fall back to generic memory. - */ - return 0; - } + if (unlikely(pageno < 0)) + goto err; + + /* + * Memory was found in the per-device area. + */ + *dma_handle = mem->device_base + (pageno << PAGE_SHIFT); + *ret = mem->virt_base + (pageno << PAGE_SHIFT); + memset(*ret, 0, size); + return 1; + +err: + /* + * In the case where the allocation can not be satisfied from the + * per-device area, try to fall back to generic memory if the + * constraints allow it. + */ + return mem->flags & DMA_MEMORY_EXCLUSIVE; } EXPORT_SYMBOL(dma_alloc_from_coherent); diff --git a/kernel/exec_domain.c b/kernel/exec_domain.c index 667c841c295..c35452cadde 100644 --- a/kernel/exec_domain.c +++ b/kernel/exec_domain.c @@ -18,6 +18,7 @@ #include <linux/syscalls.h> #include <linux/sysctl.h> #include <linux/types.h> +#include <linux/fs_struct.h> static void default_handler(int, struct pt_regs *); @@ -145,28 +146,6 @@ __set_personality(u_long personality) return 0; } - if (atomic_read(¤t->fs->count) != 1) { - struct fs_struct *fsp, *ofsp; - - fsp = copy_fs_struct(current->fs); - if (fsp == NULL) { - module_put(ep->module); - return -ENOMEM; - } - - task_lock(current); - ofsp = current->fs; - current->fs = fsp; - task_unlock(current); - - put_fs_struct(ofsp); - } - - /* - * At that point we are guaranteed to be the sole owner of - * current->fs. - */ - current->personality = personality; oep = current_thread_info()->exec_domain; current_thread_info()->exec_domain = ep; diff --git a/kernel/exit.c b/kernel/exit.c index f80dec3f187..ae5d8660ddf 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -12,7 +12,6 @@ #include <linux/completion.h> #include <linux/personality.h> #include <linux/tty.h> -#include <linux/mnt_namespace.h> #include <linux/iocontext.h> #include <linux/key.h> #include <linux/security.h> @@ -46,8 +45,10 @@ #include <linux/blkdev.h> #include <linux/task_io_accounting_ops.h> #include <linux/tracehook.h> +#include <linux/fs_struct.h> #include <linux/init_task.h> -#include <trace/sched.h> +#include <linux/perf_counter.h> +#include <trace/events/sched.h> #include <asm/uaccess.h> #include <asm/unistd.h> @@ -55,17 +56,8 @@ #include <asm/mmu_context.h> #include "cred-internals.h" -DEFINE_TRACE(sched_process_free); -DEFINE_TRACE(sched_process_exit); -DEFINE_TRACE(sched_process_wait); - static void exit_mm(struct task_struct * tsk); -static inline int task_detached(struct task_struct *p) -{ - return p->exit_signal == -1; -} - static void __unhash_process(struct task_struct *p) { nr_threads--; @@ -118,6 +110,8 @@ static void __exit_signal(struct task_struct *tsk) * We won't ever get here for the group leader, since it * will have been the last reference on the signal_struct. */ + sig->utime = cputime_add(sig->utime, task_utime(tsk)); + sig->stime = cputime_add(sig->stime, task_stime(tsk)); sig->gtime = cputime_add(sig->gtime, task_gtime(tsk)); sig->min_flt += tsk->min_flt; sig->maj_flt += tsk->maj_flt; @@ -126,6 +120,7 @@ static void __exit_signal(struct task_struct *tsk) sig->inblock += task_io_get_inblock(tsk); sig->oublock += task_io_get_oublock(tsk); task_io_accounting_add(&sig->ioac, &tsk->ioac); + sig->sum_sched_runtime += tsk->se.sum_exec_runtime; sig = NULL; /* Marker for below. */ } @@ -159,6 +154,9 @@ static void delayed_put_task_struct(struct rcu_head *rhp) { struct task_struct *tsk = container_of(rhp, struct task_struct, rcu); +#ifdef CONFIG_PERF_COUNTERS + WARN_ON_ONCE(tsk->perf_counter_ctxp); +#endif trace_sched_process_free(tsk); put_task_struct(tsk); } @@ -175,6 +173,7 @@ repeat: atomic_dec(&__task_cred(p)->user->processes); proc_flush_task(p); + write_lock_irq(&tasklist_lock); tracehook_finish_release_task(p); __exit_signal(p); @@ -359,16 +358,12 @@ static void reparent_to_kthreadd(void) void __set_special_pids(struct pid *pid) { struct task_struct *curr = current->group_leader; - pid_t nr = pid_nr(pid); - if (task_session(curr) != pid) { + if (task_session(curr) != pid) change_pid(curr, PIDTYPE_SID, pid); - set_task_session(curr, nr); - } - if (task_pgrp(curr) != pid) { + + if (task_pgrp(curr) != pid) change_pid(curr, PIDTYPE_PGID, pid); - set_task_pgrp(curr, nr); - } } static void set_special_pids(struct pid *pid) @@ -379,9 +374,8 @@ static void set_special_pids(struct pid *pid) } /* - * Let kernel threads use this to say that they - * allow a certain signal (since daemonize() will - * have disabled all of them by default). + * Let kernel threads use this to say that they allow a certain signal. + * Must not be used if kthread was cloned with CLONE_SIGHAND. */ int allow_signal(int sig) { @@ -389,14 +383,14 @@ int allow_signal(int sig) return -EINVAL; spin_lock_irq(¤t->sighand->siglock); + /* This is only needed for daemonize()'ed kthreads */ sigdelset(¤t->blocked, sig); - if (!current->mm) { - /* Kernel threads handle their own signals. - Let the signal code know it'll be handled, so - that they don't get converted to SIGKILL or - just silently dropped */ - current->sighand->action[(sig)-1].sa.sa_handler = (void __user *)2; - } + /* + * Kernel threads handle their own signals. Let the signal code + * know it'll be handled, so that they don't get converted to + * SIGKILL or just silently dropped. + */ + current->sighand->action[(sig)-1].sa.sa_handler = (void __user *)2; recalc_sigpending(); spin_unlock_irq(¤t->sighand->siglock); return 0; @@ -426,7 +420,6 @@ EXPORT_SYMBOL(disallow_signal); void daemonize(const char *name, ...) { va_list args; - struct fs_struct *fs; sigset_t blocked; va_start(args, name); @@ -459,11 +452,7 @@ void daemonize(const char *name, ...) /* Become as one with the init task */ - exit_fs(current); /* current->fs->count--; */ - fs = init_task.fs; - current->fs = fs; - atomic_inc(&fs->count); - + daemonize_fs_struct(); exit_files(current); current->files = init_task.files; atomic_inc(¤t->files->count); @@ -562,30 +551,6 @@ void exit_files(struct task_struct *tsk) } } -void put_fs_struct(struct fs_struct *fs) -{ - /* No need to hold fs->lock if we are killing it */ - if (atomic_dec_and_test(&fs->count)) { - path_put(&fs->root); - path_put(&fs->pwd); - kmem_cache_free(fs_cachep, fs); - } -} - -void exit_fs(struct task_struct *tsk) -{ - struct fs_struct * fs = tsk->fs; - - if (fs) { - task_lock(tsk); - tsk->fs = NULL; - task_unlock(tsk); - put_fs_struct(fs); - } -} - -EXPORT_SYMBOL_GPL(exit_fs); - #ifdef CONFIG_MM_OWNER /* * Task p is exiting and it owned mm, lets find a new owner for it @@ -624,7 +589,7 @@ retry: /* * Search in the siblings */ - list_for_each_entry(c, &p->parent->children, sibling) { + list_for_each_entry(c, &p->real_parent->children, sibling) { if (c->mm == mm) goto assign_new_owner; } @@ -729,119 +694,6 @@ static void exit_mm(struct task_struct * tsk) } /* - * Return nonzero if @parent's children should reap themselves. - * - * Called with write_lock_irq(&tasklist_lock) held. - */ -static int ignoring_children(struct task_struct *parent) -{ - int ret; - struct sighand_struct *psig = parent->sighand; - unsigned long flags; - spin_lock_irqsave(&psig->siglock, flags); - ret = (psig->action[SIGCHLD-1].sa.sa_handler == SIG_IGN || - (psig->action[SIGCHLD-1].sa.sa_flags & SA_NOCLDWAIT)); - spin_unlock_irqrestore(&psig->siglock, flags); - return ret; -} - -/* - * Detach all tasks we were using ptrace on. - * Any that need to be release_task'd are put on the @dead list. - * - * Called with write_lock(&tasklist_lock) held. - */ -static void ptrace_exit(struct task_struct *parent, struct list_head *dead) -{ - struct task_struct *p, *n; - int ign = -1; - - list_for_each_entry_safe(p, n, &parent->ptraced, ptrace_entry) { - __ptrace_unlink(p); - - if (p->exit_state != EXIT_ZOMBIE) - continue; - - /* - * If it's a zombie, our attachedness prevented normal - * parent notification or self-reaping. Do notification - * now if it would have happened earlier. If it should - * reap itself, add it to the @dead list. We can't call - * release_task() here because we already hold tasklist_lock. - * - * If it's our own child, there is no notification to do. - * But if our normal children self-reap, then this child - * was prevented by ptrace and we must reap it now. - */ - if (!task_detached(p) && thread_group_empty(p)) { - if (!same_thread_group(p->real_parent, parent)) - do_notify_parent(p, p->exit_signal); - else { - if (ign < 0) - ign = ignoring_children(parent); - if (ign) - p->exit_signal = -1; - } - } - - if (task_detached(p)) { - /* - * Mark it as in the process of being reaped. - */ - p->exit_state = EXIT_DEAD; - list_add(&p->ptrace_entry, dead); - } - } -} - -/* - * Finish up exit-time ptrace cleanup. - * - * Called without locks. - */ -static void ptrace_exit_finish(struct task_struct *parent, - struct list_head *dead) -{ - struct task_struct *p, *n; - - BUG_ON(!list_empty(&parent->ptraced)); - - list_for_each_entry_safe(p, n, dead, ptrace_entry) { - list_del_init(&p->ptrace_entry); - release_task(p); - } -} - -static void reparent_thread(struct task_struct *p, struct task_struct *father) -{ - if (p->pdeath_signal) - /* We already hold the tasklist_lock here. */ - group_send_sig_info(p->pdeath_signal, SEND_SIG_NOINFO, p); - - list_move_tail(&p->sibling, &p->real_parent->children); - - /* If this is a threaded reparent there is no need to - * notify anyone anything has happened. - */ - if (same_thread_group(p->real_parent, father)) - return; - - /* We don't want people slaying init. */ - if (!task_detached(p)) - p->exit_signal = SIGCHLD; - - /* If we'd notified the old parent about this child's death, - * also notify the new parent. - */ - if (!ptrace_reparented(p) && - p->exit_state == EXIT_ZOMBIE && - !task_detached(p) && thread_group_empty(p)) - do_notify_parent(p, p->exit_signal); - - kill_orphaned_pgrp(p, father); -} - -/* * When we die, we re-parent all our children. * Try to give them to another thread in our thread * group, and if no such member exists, give it to @@ -880,31 +732,68 @@ static struct task_struct *find_new_reaper(struct task_struct *father) return pid_ns->child_reaper; } +/* +* Any that need to be release_task'd are put on the @dead list. + */ +static void reparent_thread(struct task_struct *father, struct task_struct *p, + struct list_head *dead) +{ + if (p->pdeath_signal) + group_send_sig_info(p->pdeath_signal, SEND_SIG_NOINFO, p); + + list_move_tail(&p->sibling, &p->real_parent->children); + + if (task_detached(p)) + return; + /* + * If this is a threaded reparent there is no need to + * notify anyone anything has happened. + */ + if (same_thread_group(p->real_parent, father)) + return; + + /* We don't want people slaying init. */ + p->exit_signal = SIGCHLD; + + /* If it has exited notify the new parent about this child's death. */ + if (!task_ptrace(p) && + p->exit_state == EXIT_ZOMBIE && thread_group_empty(p)) { + do_notify_parent(p, p->exit_signal); + if (task_detached(p)) { + p->exit_state = EXIT_DEAD; + list_move_tail(&p->sibling, dead); + } + } + + kill_orphaned_pgrp(p, father); +} + static void forget_original_parent(struct task_struct *father) { struct task_struct *p, *n, *reaper; - LIST_HEAD(ptrace_dead); + LIST_HEAD(dead_children); + + exit_ptrace(father); write_lock_irq(&tasklist_lock); reaper = find_new_reaper(father); - /* - * First clean up ptrace if we were using it. - */ - ptrace_exit(father, &ptrace_dead); list_for_each_entry_safe(p, n, &father->children, sibling) { p->real_parent = reaper; if (p->parent == father) { - BUG_ON(p->ptrace); + BUG_ON(task_ptrace(p)); p->parent = p->real_parent; } - reparent_thread(p, father); + reparent_thread(father, p, &dead_children); } - write_unlock_irq(&tasklist_lock); + BUG_ON(!list_empty(&father->children)); - ptrace_exit_finish(father, &ptrace_dead); + list_for_each_entry_safe(p, n, &dead_children, sibling) { + list_del_init(&p->sibling); + release_task(p); + } } /* @@ -947,8 +836,7 @@ static void exit_notify(struct task_struct *tsk, int group_dead) */ if (tsk->exit_signal != SIGCHLD && !task_detached(tsk) && (tsk->parent_exec_id != tsk->real_parent->self_exec_id || - tsk->self_exec_id != tsk->parent_exec_id) && - !capable(CAP_KILL)) + tsk->self_exec_id != tsk->parent_exec_id)) tsk->exit_signal = SIGCHLD; signal = tracehook_notify_death(tsk, &cookie, group_dead); @@ -977,12 +865,9 @@ static void check_stack_usage(void) { static DEFINE_SPINLOCK(low_water_lock); static int lowest_to_date = THREAD_SIZE; - unsigned long *n = end_of_stack(current); unsigned long free; - while (*n == 0) - n++; - free = (unsigned long)n - (unsigned long)end_of_stack(current); + free = stack_not_used(current); if (free >= lowest_to_date) return; @@ -1016,6 +901,8 @@ NORET_TYPE void do_exit(long code) tracehook_report_exit(&code); + validate_creds_for_do_exit(tsk); + /* * We're taking recursive faults here in do_exit. Safest is to just * leave this task alone and wait for reboot. @@ -1037,6 +924,8 @@ NORET_TYPE void do_exit(long code) schedule(); } + exit_irq_thread(); + exit_signals(tsk); /* sets PF_EXITING */ /* * tsk->flags are checked in the futex code to protect against @@ -1087,16 +976,19 @@ NORET_TYPE void do_exit(long code) module_put(tsk->binfmt->module); proc_exit_connector(tsk); + + /* + * Flush inherited counters to the parent - before the parent + * gets woken up by child-exit notifications. + */ + perf_counter_exit_task(tsk); + exit_notify(tsk, group_dead); #ifdef CONFIG_NUMA mpol_put(tsk->mempolicy); tsk->mempolicy = NULL; #endif #ifdef CONFIG_FUTEX - /* - * This must happen late, after the PID is not - * hashed anymore: - */ if (unlikely(!list_empty(&tsk->pi_state_list))) exit_pi_state_list(tsk); if (unlikely(current->pi_state_cache)) @@ -1119,7 +1011,10 @@ NORET_TYPE void do_exit(long code) if (tsk->splice_pipe) __free_pipe_info(tsk->splice_pipe); + validate_creds_for_do_exit(tsk); + preempt_disable(); + exit_rcu(); /* causes final put_task_struct in finish_task_switch(). */ tsk->state = TASK_DEAD; schedule(); @@ -1189,6 +1084,18 @@ SYSCALL_DEFINE1(exit_group, int, error_code) return 0; } +struct wait_opts { + enum pid_type wo_type; + int wo_flags; + struct pid *wo_pid; + + struct siginfo __user *wo_info; + int __user *wo_stat; + struct rusage __user *wo_rusage; + + int notask_error; +}; + static struct pid *task_pid_type(struct task_struct *task, enum pid_type type) { struct pid *pid = NULL; @@ -1199,13 +1106,12 @@ static struct pid *task_pid_type(struct task_struct *task, enum pid_type type) return pid; } -static int eligible_child(enum pid_type type, struct pid *pid, int options, - struct task_struct *p) +static int eligible_child(struct wait_opts *wo, struct task_struct *p) { int err; - if (type < PIDTYPE_MAX) { - if (task_pid_type(p, type) != pid) + if (wo->wo_type < PIDTYPE_MAX) { + if (task_pid_type(p, wo->wo_type) != wo->wo_pid) return 0; } @@ -1214,8 +1120,8 @@ static int eligible_child(enum pid_type type, struct pid *pid, int options, * set; otherwise, wait for non-clone children *only*. (Note: * A "clone" child here is one that reports to its parent * using a signal other than SIGCHLD.) */ - if (((p->exit_signal != SIGCHLD) ^ ((options & __WCLONE) != 0)) - && !(options & __WALL)) + if (((p->exit_signal != SIGCHLD) ^ !!(wo->wo_flags & __WCLONE)) + && !(wo->wo_flags & __WALL)) return 0; err = security_task_wait(p); @@ -1225,14 +1131,15 @@ static int eligible_child(enum pid_type type, struct pid *pid, int options, return 1; } -static int wait_noreap_copyout(struct task_struct *p, pid_t pid, uid_t uid, - int why, int status, - struct siginfo __user *infop, - struct rusage __user *rusagep) +static int wait_noreap_copyout(struct wait_opts *wo, struct task_struct *p, + pid_t pid, uid_t uid, int why, int status) { - int retval = rusagep ? getrusage(p, RUSAGE_BOTH, rusagep) : 0; + struct siginfo __user *infop; + int retval = wo->wo_rusage + ? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0; put_task_struct(p); + infop = wo->wo_info; if (!retval) retval = put_user(SIGCHLD, &infop->si_signo); if (!retval) @@ -1256,19 +1163,18 @@ static int wait_noreap_copyout(struct task_struct *p, pid_t pid, uid_t uid, * the lock and this task is uninteresting. If we return nonzero, we have * released the lock and the system call should return. */ -static int wait_task_zombie(struct task_struct *p, int options, - struct siginfo __user *infop, - int __user *stat_addr, struct rusage __user *ru) +static int wait_task_zombie(struct wait_opts *wo, struct task_struct *p) { unsigned long state; int retval, status, traced; pid_t pid = task_pid_vnr(p); uid_t uid = __task_cred(p)->uid; + struct siginfo __user *infop; - if (!likely(options & WEXITED)) + if (!likely(wo->wo_flags & WEXITED)) return 0; - if (unlikely(options & WNOWAIT)) { + if (unlikely(wo->wo_flags & WNOWAIT)) { int exit_code = p->exit_code; int why, status; @@ -1281,8 +1187,7 @@ static int wait_task_zombie(struct task_struct *p, int options, why = (exit_code & 0x80) ? CLD_DUMPED : CLD_KILLED; status = exit_code & 0x7f; } - return wait_noreap_copyout(p, pid, uid, why, - status, infop, ru); + return wait_noreap_copyout(wo, p, pid, uid, why, status); } /* @@ -1296,11 +1201,13 @@ static int wait_task_zombie(struct task_struct *p, int options, } traced = ptrace_reparented(p); - - if (likely(!traced)) { + /* + * It can be ptraced but not reparented, check + * !task_detached() to filter out sub-threads. + */ + if (likely(!traced) && likely(!task_detached(p))) { struct signal_struct *psig; struct signal_struct *sig; - struct task_cputime cputime; /* * The resource counters for the group leader are in its @@ -1313,26 +1220,23 @@ static int wait_task_zombie(struct task_struct *p, int options, * p->signal fields, because they are only touched by * __exit_signal, which runs with tasklist_lock * write-locked anyway, and so is excluded here. We do - * need to protect the access to p->parent->signal fields, + * need to protect the access to parent->signal fields, * as other threads in the parent group can be right * here reaping other children at the same time. - * - * We use thread_group_cputime() to get times for the thread - * group, which consolidates times for all threads in the - * group including the group leader. */ - thread_group_cputime(p, &cputime); - spin_lock_irq(&p->parent->sighand->siglock); - psig = p->parent->signal; + spin_lock_irq(&p->real_parent->sighand->siglock); + psig = p->real_parent->signal; sig = p->signal; psig->cutime = cputime_add(psig->cutime, - cputime_add(cputime.utime, - sig->cutime)); + cputime_add(p->utime, + cputime_add(sig->utime, + sig->cutime))); psig->cstime = cputime_add(psig->cstime, - cputime_add(cputime.stime, - sig->cstime)); + cputime_add(p->stime, + cputime_add(sig->stime, + sig->cstime))); psig->cgtime = cputime_add(psig->cgtime, cputime_add(p->gtime, @@ -1354,7 +1258,7 @@ static int wait_task_zombie(struct task_struct *p, int options, sig->oublock + sig->coublock; task_io_accounting_add(&psig->ioac, &p->ioac); task_io_accounting_add(&psig->ioac, &sig->ioac); - spin_unlock_irq(&p->parent->sighand->siglock); + spin_unlock_irq(&p->real_parent->sighand->siglock); } /* @@ -1363,11 +1267,14 @@ static int wait_task_zombie(struct task_struct *p, int options, */ read_unlock(&tasklist_lock); - retval = ru ? getrusage(p, RUSAGE_BOTH, ru) : 0; + retval = wo->wo_rusage + ? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0; status = (p->signal->flags & SIGNAL_GROUP_EXIT) ? p->signal->group_exit_code : p->exit_code; - if (!retval && stat_addr) - retval = put_user(status, stat_addr); + if (!retval && wo->wo_stat) + retval = put_user(status, wo->wo_stat); + + infop = wo->wo_info; if (!retval && infop) retval = put_user(SIGCHLD, &infop->si_signo); if (!retval && infop) @@ -1417,42 +1324,51 @@ static int wait_task_zombie(struct task_struct *p, int options, return retval; } +static int *task_stopped_code(struct task_struct *p, bool ptrace) +{ + if (ptrace) { + if (task_is_stopped_or_traced(p)) + return &p->exit_code; + } else { + if (p->signal->flags & SIGNAL_STOP_STOPPED) + return &p->signal->group_exit_code; + } + return NULL; +} + /* * Handle sys_wait4 work for one task in state TASK_STOPPED. We hold * read_lock(&tasklist_lock) on entry. If we return zero, we still hold * the lock and this task is uninteresting. If we return nonzero, we have * released the lock and the system call should return. */ -static int wait_task_stopped(int ptrace, struct task_struct *p, - int options, struct siginfo __user *infop, - int __user *stat_addr, struct rusage __user *ru) +static int wait_task_stopped(struct wait_opts *wo, + int ptrace, struct task_struct *p) { - int retval, exit_code, why; + struct siginfo __user *infop; + int retval, exit_code, *p_code, why; uid_t uid = 0; /* unneeded, required by compiler */ pid_t pid; - if (!(options & WUNTRACED)) + /* + * Traditionally we see ptrace'd stopped tasks regardless of options. + */ + if (!ptrace && !(wo->wo_flags & WUNTRACED)) return 0; exit_code = 0; spin_lock_irq(&p->sighand->siglock); - if (unlikely(!task_is_stopped_or_traced(p))) - goto unlock_sig; - - if (!ptrace && p->signal->group_stop_count > 0) - /* - * A group stop is in progress and this is the group leader. - * We won't report until all threads have stopped. - */ + p_code = task_stopped_code(p, ptrace); + if (unlikely(!p_code)) goto unlock_sig; - exit_code = p->exit_code; + exit_code = *p_code; if (!exit_code) goto unlock_sig; - if (!unlikely(options & WNOWAIT)) - p->exit_code = 0; + if (!unlikely(wo->wo_flags & WNOWAIT)) + *p_code = 0; /* don't need the RCU readlock here as we're holding a spinlock */ uid = __task_cred(p)->uid; @@ -1473,14 +1389,15 @@ unlock_sig: why = ptrace ? CLD_TRAPPED : CLD_STOPPED; read_unlock(&tasklist_lock); - if (unlikely(options & WNOWAIT)) - return wait_noreap_copyout(p, pid, uid, - why, exit_code, - infop, ru); + if (unlikely(wo->wo_flags & WNOWAIT)) + return wait_noreap_copyout(wo, p, pid, uid, why, exit_code); - retval = ru ? getrusage(p, RUSAGE_BOTH, ru) : 0; - if (!retval && stat_addr) - retval = put_user((exit_code << 8) | 0x7f, stat_addr); + retval = wo->wo_rusage + ? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0; + if (!retval && wo->wo_stat) + retval = put_user((exit_code << 8) | 0x7f, wo->wo_stat); + + infop = wo->wo_info; if (!retval && infop) retval = put_user(SIGCHLD, &infop->si_signo); if (!retval && infop) @@ -1507,15 +1424,13 @@ unlock_sig: * the lock and this task is uninteresting. If we return nonzero, we have * released the lock and the system call should return. */ -static int wait_task_continued(struct task_struct *p, int options, - struct siginfo __user *infop, - int __user *stat_addr, struct rusage __user *ru) +static int wait_task_continued(struct wait_opts *wo, struct task_struct *p) { int retval; pid_t pid; uid_t uid; - if (!unlikely(options & WCONTINUED)) + if (!unlikely(wo->wo_flags & WCONTINUED)) return 0; if (!(p->signal->flags & SIGNAL_STOP_CONTINUED)) @@ -1527,7 +1442,7 @@ static int wait_task_continued(struct task_struct *p, int options, spin_unlock_irq(&p->sighand->siglock); return 0; } - if (!unlikely(options & WNOWAIT)) + if (!unlikely(wo->wo_flags & WNOWAIT)) p->signal->flags &= ~SIGNAL_STOP_CONTINUED; uid = __task_cred(p)->uid; spin_unlock_irq(&p->sighand->siglock); @@ -1536,17 +1451,17 @@ static int wait_task_continued(struct task_struct *p, int options, get_task_struct(p); read_unlock(&tasklist_lock); - if (!infop) { - retval = ru ? getrusage(p, RUSAGE_BOTH, ru) : 0; + if (!wo->wo_info) { + retval = wo->wo_rusage + ? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0; put_task_struct(p); - if (!retval && stat_addr) - retval = put_user(0xffff, stat_addr); + if (!retval && wo->wo_stat) + retval = put_user(0xffff, wo->wo_stat); if (!retval) retval = pid; } else { - retval = wait_noreap_copyout(p, pid, uid, - CLD_CONTINUED, SIGCONT, - infop, ru); + retval = wait_noreap_copyout(wo, p, pid, uid, + CLD_CONTINUED, SIGCONT); BUG_ON(retval == 0); } @@ -1556,19 +1471,16 @@ static int wait_task_continued(struct task_struct *p, int options, /* * Consider @p for a wait by @parent. * - * -ECHILD should be in *@notask_error before the first call. + * -ECHILD should be in ->notask_error before the first call. * Returns nonzero for a final return, when we have unlocked tasklist_lock. * Returns zero if the search for a child should continue; - * then *@notask_error is 0 if @p is an eligible child, + * then ->notask_error is 0 if @p is an eligible child, * or another error from security_task_wait(), or still -ECHILD. */ -static int wait_consider_task(struct task_struct *parent, int ptrace, - struct task_struct *p, int *notask_error, - enum pid_type type, struct pid *pid, int options, - struct siginfo __user *infop, - int __user *stat_addr, struct rusage __user *ru) +static int wait_consider_task(struct wait_opts *wo, struct task_struct *parent, + int ptrace, struct task_struct *p) { - int ret = eligible_child(type, pid, options, p); + int ret = eligible_child(wo, p); if (!ret) return ret; @@ -1580,16 +1492,17 @@ static int wait_consider_task(struct task_struct *parent, int ptrace, * to look for security policy problems, rather * than for mysterious wait bugs. */ - if (*notask_error) - *notask_error = ret; + if (wo->notask_error) + wo->notask_error = ret; + return 0; } - if (likely(!ptrace) && unlikely(p->ptrace)) { + if (likely(!ptrace) && unlikely(task_ptrace(p))) { /* * This child is hidden by ptrace. * We aren't allowed to see it now, but eventually we will. */ - *notask_error = 0; + wo->notask_error = 0; return 0; } @@ -1600,34 +1513,30 @@ static int wait_consider_task(struct task_struct *parent, int ptrace, * We don't reap group leaders with subthreads. */ if (p->exit_state == EXIT_ZOMBIE && !delay_group_leader(p)) - return wait_task_zombie(p, options, infop, stat_addr, ru); + return wait_task_zombie(wo, p); /* * It's stopped or running now, so it might * later continue, exit, or stop again. */ - *notask_error = 0; + wo->notask_error = 0; - if (task_is_stopped_or_traced(p)) - return wait_task_stopped(ptrace, p, options, - infop, stat_addr, ru); + if (task_stopped_code(p, ptrace)) + return wait_task_stopped(wo, ptrace, p); - return wait_task_continued(p, options, infop, stat_addr, ru); + return wait_task_continued(wo, p); } /* * Do the work of do_wait() for one thread in the group, @tsk. * - * -ECHILD should be in *@notask_error before the first call. + * -ECHILD should be in ->notask_error before the first call. * Returns nonzero for a final return, when we have unlocked tasklist_lock. * Returns zero if the search for a child should continue; then - * *@notask_error is 0 if there were any eligible children, + * ->notask_error is 0 if there were any eligible children, * or another error from security_task_wait(), or still -ECHILD. */ -static int do_wait_thread(struct task_struct *tsk, int *notask_error, - enum pid_type type, struct pid *pid, int options, - struct siginfo __user *infop, int __user *stat_addr, - struct rusage __user *ru) +static int do_wait_thread(struct wait_opts *wo, struct task_struct *tsk) { struct task_struct *p; @@ -1636,9 +1545,7 @@ static int do_wait_thread(struct task_struct *tsk, int *notask_error, * Do not consider detached threads. */ if (!task_detached(p)) { - int ret = wait_consider_task(tsk, 0, p, notask_error, - type, pid, options, - infop, stat_addr, ru); + int ret = wait_consider_task(wo, tsk, 0, p); if (ret) return ret; } @@ -1647,22 +1554,12 @@ static int do_wait_thread(struct task_struct *tsk, int *notask_error, return 0; } -static int ptrace_do_wait(struct task_struct *tsk, int *notask_error, - enum pid_type type, struct pid *pid, int options, - struct siginfo __user *infop, int __user *stat_addr, - struct rusage __user *ru) +static int ptrace_do_wait(struct wait_opts *wo, struct task_struct *tsk) { struct task_struct *p; - /* - * Traditionally we see ptrace'd stopped tasks regardless of options. - */ - options |= WUNTRACED; - list_for_each_entry(p, &tsk->ptraced, ptrace_entry) { - int ret = wait_consider_task(tsk, 1, p, notask_error, - type, pid, options, - infop, stat_addr, ru); + int ret = wait_consider_task(wo, tsk, 1, p); if (ret) return ret; } @@ -1670,65 +1567,59 @@ static int ptrace_do_wait(struct task_struct *tsk, int *notask_error, return 0; } -static long do_wait(enum pid_type type, struct pid *pid, int options, - struct siginfo __user *infop, int __user *stat_addr, - struct rusage __user *ru) +static long do_wait(struct wait_opts *wo) { DECLARE_WAITQUEUE(wait, current); struct task_struct *tsk; int retval; - trace_sched_process_wait(pid); + trace_sched_process_wait(wo->wo_pid); add_wait_queue(¤t->signal->wait_chldexit,&wait); repeat: /* * If there is nothing that can match our critiera just get out. - * We will clear @retval to zero if we see any child that might later - * match our criteria, even if we are not able to reap it yet. + * We will clear ->notask_error to zero if we see any child that + * might later match our criteria, even if we are not able to reap + * it yet. */ - retval = -ECHILD; - if ((type < PIDTYPE_MAX) && (!pid || hlist_empty(&pid->tasks[type]))) - goto end; + wo->notask_error = -ECHILD; + if ((wo->wo_type < PIDTYPE_MAX) && + (!wo->wo_pid || hlist_empty(&wo->wo_pid->tasks[wo->wo_type]))) + goto notask; - current->state = TASK_INTERRUPTIBLE; + set_current_state(TASK_INTERRUPTIBLE); read_lock(&tasklist_lock); tsk = current; do { - int tsk_result = do_wait_thread(tsk, &retval, - type, pid, options, - infop, stat_addr, ru); - if (!tsk_result) - tsk_result = ptrace_do_wait(tsk, &retval, - type, pid, options, - infop, stat_addr, ru); - if (tsk_result) { - /* - * tasklist_lock is unlocked and we have a final result. - */ - retval = tsk_result; + retval = do_wait_thread(wo, tsk); + if (retval) goto end; - } - if (options & __WNOTHREAD) + retval = ptrace_do_wait(wo, tsk); + if (retval) + goto end; + + if (wo->wo_flags & __WNOTHREAD) break; - tsk = next_thread(tsk); - BUG_ON(tsk->signal != current->signal); - } while (tsk != current); + } while_each_thread(current, tsk); read_unlock(&tasklist_lock); - if (!retval && !(options & WNOHANG)) { +notask: + retval = wo->notask_error; + if (!retval && !(wo->wo_flags & WNOHANG)) { retval = -ERESTARTSYS; if (!signal_pending(current)) { schedule(); goto repeat; } } - end: - current->state = TASK_RUNNING; + __set_current_state(TASK_RUNNING); remove_wait_queue(¤t->signal->wait_chldexit,&wait); - if (infop) { + if (wo->wo_info) { + struct siginfo __user *infop = wo->wo_info; + if (retval > 0) retval = 0; else { @@ -1757,6 +1648,7 @@ end: SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *, infop, int, options, struct rusage __user *, ru) { + struct wait_opts wo; struct pid *pid = NULL; enum pid_type type; long ret; @@ -1786,7 +1678,14 @@ SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *, if (type < PIDTYPE_MAX) pid = find_get_pid(upid); - ret = do_wait(type, pid, options, infop, NULL, ru); + + wo.wo_type = type; + wo.wo_pid = pid; + wo.wo_flags = options; + wo.wo_info = infop; + wo.wo_stat = NULL; + wo.wo_rusage = ru; + ret = do_wait(&wo); put_pid(pid); /* avoid REGPARM breakage on x86: */ @@ -1797,6 +1696,7 @@ SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *, SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr, int, options, struct rusage __user *, ru) { + struct wait_opts wo; struct pid *pid = NULL; enum pid_type type; long ret; @@ -1812,13 +1712,19 @@ SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr, pid = find_get_pid(-upid); } else if (upid == 0) { type = PIDTYPE_PGID; - pid = get_pid(task_pgrp(current)); + pid = get_task_pid(current, PIDTYPE_PGID); } else /* upid > 0 */ { type = PIDTYPE_PID; pid = find_get_pid(upid); } - ret = do_wait(type, pid, options | WEXITED, NULL, stat_addr, ru); + wo.wo_type = type; + wo.wo_pid = pid; + wo.wo_flags = options | WEXITED; + wo.wo_info = NULL; + wo.wo_stat = stat_addr; + wo.wo_rusage = ru; + ret = do_wait(&wo); put_pid(pid); /* avoid REGPARM breakage on x86: */ diff --git a/kernel/extable.c b/kernel/extable.c index e136ed8d82b..7f8f263f852 100644 --- a/kernel/extable.c +++ b/kernel/extable.c @@ -15,11 +15,22 @@ along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ +#include <linux/ftrace.h> +#include <linux/memory.h> #include <linux/module.h> +#include <linux/mutex.h> #include <linux/init.h> -#include <linux/ftrace.h> -#include <asm/uaccess.h> + #include <asm/sections.h> +#include <asm/uaccess.h> + +/* + * mutex protecting text section modification (dynamic code patching). + * some users need to sleep (allocating memory...) while they hold this lock. + * + * NOT exported to modules - patching kernel text is a really delicate matter. + */ +DEFINE_MUTEX(text_mutex); extern struct exception_table_entry __start___ex_table[]; extern struct exception_table_entry __stop___ex_table[]; @@ -41,31 +52,50 @@ const struct exception_table_entry *search_exception_tables(unsigned long addr) return e; } -__notrace_funcgraph int core_kernel_text(unsigned long addr) +static inline int init_kernel_text(unsigned long addr) +{ + if (addr >= (unsigned long)_sinittext && + addr <= (unsigned long)_einittext) + return 1; + return 0; +} + +int core_kernel_text(unsigned long addr) { if (addr >= (unsigned long)_stext && addr <= (unsigned long)_etext) return 1; if (system_state == SYSTEM_BOOTING && - addr >= (unsigned long)_sinittext && - addr <= (unsigned long)_einittext) + init_kernel_text(addr)) return 1; return 0; } -__notrace_funcgraph int __kernel_text_address(unsigned long addr) +int __kernel_text_address(unsigned long addr) { if (core_kernel_text(addr)) return 1; - return __module_text_address(addr) != NULL; + if (is_module_text_address(addr)) + return 1; + /* + * There might be init symbols in saved stacktraces. + * Give those symbols a chance to be printed in + * backtraces (such as lockdep traces). + * + * Since we are after the module-symbols check, there's + * no danger of address overlap: + */ + if (init_kernel_text(addr)) + return 1; + return 0; } int kernel_text_address(unsigned long addr) { if (core_kernel_text(addr)) return 1; - return module_text_address(addr) != NULL; + return is_module_text_address(addr); } /* @@ -81,5 +111,5 @@ int func_ptr_is_kernel_text(void *ptr) addr = (unsigned long) dereference_function_descriptor(ptr); if (core_kernel_text(addr)) return 1; - return module_text_address(addr) != NULL; + return is_module_text_address(addr); } diff --git a/kernel/fork.c b/kernel/fork.c index bf0cef8bbdf..bfee931ee3f 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -17,7 +17,6 @@ #include <linux/module.h> #include <linux/vmalloc.h> #include <linux/completion.h> -#include <linux/mnt_namespace.h> #include <linux/personality.h> #include <linux/mempolicy.h> #include <linux/sem.h> @@ -60,7 +59,9 @@ #include <linux/tty.h> #include <linux/proc_fs.h> #include <linux/blkdev.h> -#include <trace/sched.h> +#include <linux/fs_struct.h> +#include <linux/magic.h> +#include <linux/perf_counter.h> #include <asm/pgtable.h> #include <asm/pgalloc.h> @@ -69,6 +70,8 @@ #include <asm/cacheflush.h> #include <asm/tlbflush.h> +#include <trace/events/sched.h> + /* * Protected counters by write_lock_irq(&tasklist_lock) */ @@ -81,8 +84,6 @@ DEFINE_PER_CPU(unsigned long, process_counts) = 0; __cacheline_aligned DEFINE_RWLOCK(tasklist_lock); /* outer */ -DEFINE_TRACE(sched_process_fork); - int nr_processes(void) { int cpu; @@ -151,8 +152,7 @@ void __put_task_struct(struct task_struct *tsk) WARN_ON(atomic_read(&tsk->usage)); WARN_ON(tsk == current); - put_cred(tsk->real_cred); - put_cred(tsk->cred); + exit_creds(tsk); delayacct_tsk_free(tsk); if (!profile_handoff_task(tsk)) @@ -176,7 +176,7 @@ void __init fork_init(unsigned long mempages) /* create a slab on which task_structs can be allocated */ task_struct_cachep = kmem_cache_create("task_struct", sizeof(struct task_struct), - ARCH_MIN_TASKALIGN, SLAB_PANIC, NULL); + ARCH_MIN_TASKALIGN, SLAB_PANIC | SLAB_NOTRACK, NULL); #endif /* do the arch specific task caches init */ @@ -212,6 +212,8 @@ static struct task_struct *dup_task_struct(struct task_struct *orig) { struct task_struct *tsk; struct thread_info *ti; + unsigned long *stackend; + int err; prepare_to_copy(orig); @@ -237,6 +239,8 @@ static struct task_struct *dup_task_struct(struct task_struct *orig) goto out; setup_thread_stack(tsk, orig); + stackend = end_of_stack(tsk); + *stackend = STACK_END_MAGIC; /* for overflow detection */ #ifdef CONFIG_CC_STACKPROTECTOR tsk->stack_canary = get_random_int(); @@ -279,7 +283,7 @@ static int dup_mmap(struct mm_struct *mm, struct mm_struct *oldmm) mm->free_area_cache = oldmm->mmap_base; mm->cached_hole_size = ~0UL; mm->map_count = 0; - cpus_clear(mm->cpu_vm_mask); + cpumask_clear(mm_cpumask(mm)); mm->mm_rb = RB_ROOT; rb_link = &mm->mm_rb.rb_node; rb_parent = NULL; @@ -562,18 +566,18 @@ void mm_release(struct task_struct *tsk, struct mm_struct *mm) * the value intact in a core dump, and to save the unnecessary * trouble otherwise. Userland only wants this done for a sys_exit. */ - if (tsk->clear_child_tid - && !(tsk->flags & PF_SIGNALED) - && atomic_read(&mm->mm_users) > 1) { - u32 __user * tidptr = tsk->clear_child_tid; + if (tsk->clear_child_tid) { + if (!(tsk->flags & PF_SIGNALED) && + atomic_read(&mm->mm_users) > 1) { + /* + * We don't check the error code - if userspace has + * not set up a proper pointer then tough luck. + */ + put_user(0, tsk->clear_child_tid); + sys_futex(tsk->clear_child_tid, FUTEX_WAKE, + 1, NULL, NULL, 0); + } tsk->clear_child_tid = NULL; - - /* - * We don't check the error code - if userspace has - * not set up a proper pointer then tough luck. - */ - put_user(0, tidptr); - sys_futex(tidptr, FUTEX_WAKE, 1, NULL, NULL, 0); } } @@ -639,6 +643,9 @@ static int copy_mm(unsigned long clone_flags, struct task_struct * tsk) tsk->min_flt = tsk->maj_flt = 0; tsk->nvcsw = tsk->nivcsw = 0; +#ifdef CONFIG_DETECT_HUNG_TASK + tsk->last_switch_count = tsk->nvcsw + tsk->nivcsw; +#endif tsk->mm = NULL; tsk->active_mm = NULL; @@ -676,38 +683,21 @@ fail_nomem: return retval; } -static struct fs_struct *__copy_fs_struct(struct fs_struct *old) -{ - struct fs_struct *fs = kmem_cache_alloc(fs_cachep, GFP_KERNEL); - /* We don't need to lock fs - think why ;-) */ - if (fs) { - atomic_set(&fs->count, 1); - rwlock_init(&fs->lock); - fs->umask = old->umask; - read_lock(&old->lock); - fs->root = old->root; - path_get(&old->root); - fs->pwd = old->pwd; - path_get(&old->pwd); - read_unlock(&old->lock); - } - return fs; -} - -struct fs_struct *copy_fs_struct(struct fs_struct *old) -{ - return __copy_fs_struct(old); -} - -EXPORT_SYMBOL_GPL(copy_fs_struct); - static int copy_fs(unsigned long clone_flags, struct task_struct *tsk) { + struct fs_struct *fs = current->fs; if (clone_flags & CLONE_FS) { - atomic_inc(¤t->fs->count); + /* tsk->fs is already what we want */ + write_lock(&fs->lock); + if (fs->in_exec) { + write_unlock(&fs->lock); + return -EAGAIN; + } + fs->users++; + write_unlock(&fs->lock); return 0; } - tsk->fs = __copy_fs_struct(current->fs); + tsk->fs = copy_fs_struct(fs); if (!tsk->fs) return -ENOMEM; return 0; @@ -808,6 +798,12 @@ static void posix_cpu_timers_init_group(struct signal_struct *sig) sig->cputime_expires.virt_exp = cputime_zero; sig->cputime_expires.sched_exp = 0; + if (sig->rlim[RLIMIT_CPU].rlim_cur != RLIM_INFINITY) { + sig->cputime_expires.prof_exp = + secs_to_cputime(sig->rlim[RLIMIT_CPU].rlim_cur); + sig->cputimer.running = 1; + } + /* The timer lists. */ INIT_LIST_HEAD(&sig->cpu_timers[0]); INIT_LIST_HEAD(&sig->cpu_timers[1]); @@ -817,16 +813,10 @@ static void posix_cpu_timers_init_group(struct signal_struct *sig) static int copy_signal(unsigned long clone_flags, struct task_struct *tsk) { struct signal_struct *sig; - int ret; - if (clone_flags & CLONE_THREAD) { - ret = thread_group_cputime_clone_thread(current); - if (likely(!ret)) { - atomic_inc(¤t->signal->count); - atomic_inc(¤t->signal->live); - } - return ret; - } + if (clone_flags & CLONE_THREAD) + return 0; + sig = kmem_cache_alloc(signal_cachep, GFP_KERNEL); tsk->signal = sig; if (!sig) @@ -836,6 +826,8 @@ static int copy_signal(unsigned long clone_flags, struct task_struct *tsk) atomic_set(&sig->live, 1); init_waitqueue_head(&sig->wait_chldexit); sig->flags = 0; + if (clone_flags & CLONE_NEWPID) + sig->flags |= SIGNAL_UNKILLABLE; sig->group_exit_code = 0; sig->group_exit_task = NULL; sig->group_stop_count = 0; @@ -851,13 +843,14 @@ static int copy_signal(unsigned long clone_flags, struct task_struct *tsk) sig->tty_old_pgrp = NULL; sig->tty = NULL; - sig->cutime = sig->cstime = cputime_zero; + sig->utime = sig->stime = sig->cutime = sig->cstime = cputime_zero; sig->gtime = cputime_zero; sig->cgtime = cputime_zero; sig->nvcsw = sig->nivcsw = sig->cnvcsw = sig->cnivcsw = 0; sig->min_flt = sig->maj_flt = sig->cmin_flt = sig->cmaj_flt = 0; sig->inblock = sig->oublock = sig->cinblock = sig->coublock = 0; task_io_accounting_init(&sig->ioac); + sig->sum_sched_runtime = 0; taskstats_tgid_init(sig); task_lock(current->group_leader); @@ -880,16 +873,6 @@ void __cleanup_signal(struct signal_struct *sig) kmem_cache_free(signal_cachep, sig); } -static void cleanup_signal(struct task_struct *tsk) -{ - struct signal_struct *sig = tsk->signal; - - atomic_dec(&sig->live); - - if (atomic_dec_and_test(&sig->count)) - __cleanup_signal(sig); -} - static void copy_flags(unsigned long clone_flags, struct task_struct *p) { unsigned long new_flags = p->flags; @@ -984,6 +967,8 @@ static struct task_struct *copy_process(unsigned long clone_flags, if (!p) goto fork_out; + ftrace_graph_init_task(p); + rt_mutex_init_task(p); #ifdef CONFIG_PROVE_LOCKING @@ -1007,6 +992,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, * triggers too late. This doesn't hurt, the check is only there * to stop root fork bombs. */ + retval = -EAGAIN; if (nr_threads >= max_threads) goto bad_fork_cleanup_count; @@ -1021,14 +1007,10 @@ static struct task_struct *copy_process(unsigned long clone_flags, copy_flags(clone_flags, p); INIT_LIST_HEAD(&p->children); INIT_LIST_HEAD(&p->sibling); -#ifdef CONFIG_PREEMPT_RCU - p->rcu_read_lock_nesting = 0; - p->rcu_flipctr_idx = 0; -#endif /* #ifdef CONFIG_PREEMPT_RCU */ + rcu_copy_process(p); p->vfork_done = NULL; spin_lock_init(&p->alloc_lock); - clear_tsk_thread_flag(p, TIF_SIGPENDING); init_sigpending(&p->pending); p->utime = cputime_zero; @@ -1041,11 +1023,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->default_timer_slack_ns = current->timer_slack_ns; -#ifdef CONFIG_DETECT_SOFTLOCKUP - p->last_switch_count = 0; - p->last_switch_timestamp = 0; -#endif - task_io_accounting_init(&p->ioac); acct_clear_integrals(p); @@ -1095,12 +1072,16 @@ static struct task_struct *copy_process(unsigned long clone_flags, #ifdef CONFIG_DEBUG_MUTEXES p->blocked_on = NULL; /* not blocked yet */ #endif - if (unlikely(ptrace_reparented(current))) - ptrace_fork(p, clone_flags); + + p->bts = NULL; /* Perform scheduler related setup. Assign this task to a CPU. */ sched_fork(p, clone_flags); + retval = perf_counter_init_task(p); + if (retval) + goto bad_fork_cleanup_policy; + if ((retval = audit_alloc(p))) goto bad_fork_cleanup_policy; /* copy all the process information */ @@ -1120,7 +1101,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, goto bad_fork_cleanup_mm; if ((retval = copy_io(clone_flags, p))) goto bad_fork_cleanup_namespaces; - retval = copy_thread(0, clone_flags, stack_start, stack_size, p, regs); + retval = copy_thread(clone_flags, stack_start, stack_size, p, regs); if (retval) goto bad_fork_cleanup_io; @@ -1137,8 +1118,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, } } - ftrace_graph_init_task(p); - p->pid = pid_nr(pid); p->tgid = p->pid; if (clone_flags & CLONE_THREAD) @@ -1147,7 +1126,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, if (current->nsproxy != p->nsproxy) { retval = ns_cgroup_clone(p, pid); if (retval) - goto bad_fork_free_graph; + goto bad_fork_free_pid; } p->set_child_tid = (clone_flags & CLONE_CHILD_SETTID) ? child_tidptr : NULL; @@ -1179,10 +1158,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, #endif clear_all_latency_tracing(p); - /* Our parent execution domain becomes current domain - These must match for thread signalling to apply */ - p->parent_exec_id = p->self_exec_id; - /* ok, now we should be set up.. */ p->exit_signal = (clone_flags & CLONE_THREAD) ? -1 : (clone_flags & CSIGNAL); p->pdeath_signal = 0; @@ -1220,10 +1195,13 @@ static struct task_struct *copy_process(unsigned long clone_flags, set_task_cpu(p, smp_processor_id()); /* CLONE_PARENT re-uses the old parent */ - if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) + if (clone_flags & (CLONE_PARENT|CLONE_THREAD)) { p->real_parent = current->real_parent; - else + p->parent_exec_id = current->parent_exec_id; + } else { p->real_parent = current; + p->parent_exec_id = current->self_exec_id; + } spin_lock(¤t->sighand->siglock); @@ -1240,10 +1218,12 @@ static struct task_struct *copy_process(unsigned long clone_flags, spin_unlock(¤t->sighand->siglock); write_unlock_irq(&tasklist_lock); retval = -ERESTARTNOINTR; - goto bad_fork_free_graph; + goto bad_fork_free_pid; } if (clone_flags & CLONE_THREAD) { + atomic_inc(¤t->signal->count); + atomic_inc(¤t->signal->live); p->group_leader = current->group_leader; list_add_tail_rcu(&p->thread_group, &p->group_leader->thread_group); } @@ -1259,8 +1239,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->signal->leader_pid = pid; tty_kref_put(p->signal->tty); p->signal->tty = tty_kref_get(current->signal->tty); - set_task_pgrp(p, task_pgrp_nr(current)); - set_task_session(p, task_session_nr(current)); attach_pid(p, PIDTYPE_PGID, task_pgrp(current)); attach_pid(p, PIDTYPE_SID, task_session(current)); list_add_tail_rcu(&p->tasks, &init_task.tasks); @@ -1275,10 +1253,9 @@ static struct task_struct *copy_process(unsigned long clone_flags, write_unlock_irq(&tasklist_lock); proc_fork_connector(p); cgroup_post_fork(p); + perf_counter_fork(p); return p; -bad_fork_free_graph: - ftrace_graph_exit_task(p); bad_fork_free_pid: if (pid != &init_struct_pid) free_pid(pid); @@ -1290,7 +1267,8 @@ bad_fork_cleanup_mm: if (p->mm) mmput(p->mm); bad_fork_cleanup_signal: - cleanup_signal(p); + if (!(clone_flags & CLONE_THREAD)) + __cleanup_signal(p->signal); bad_fork_cleanup_sighand: __cleanup_sighand(p->sighand); bad_fork_cleanup_fs: @@ -1302,6 +1280,7 @@ bad_fork_cleanup_semundo: bad_fork_cleanup_audit: audit_free(p); bad_fork_cleanup_policy: + perf_counter_free_task(p); #ifdef CONFIG_NUMA mpol_put(p->mempolicy); bad_fork_cleanup_cgroup: @@ -1314,8 +1293,7 @@ bad_fork_cleanup_put_domain: module_put(task_thread_info(p)->exec_domain->module); bad_fork_cleanup_count: atomic_dec(&p->cred->user->processes); - put_cred(p->real_cred); - put_cred(p->cred); + exit_creds(p); bad_fork_free: free_task(p); fork_out: @@ -1418,7 +1396,7 @@ long do_fork(unsigned long clone_flags, } audit_finish_fork(p); - tracehook_report_clone(trace, regs, clone_flags, nr, p); + tracehook_report_clone(regs, clone_flags, nr, p); /* * We set PF_STARTING at creation in case tracing wants to @@ -1470,20 +1448,21 @@ void __init proc_caches_init(void) { sighand_cachep = kmem_cache_create("sighand_cache", sizeof(struct sighand_struct), 0, - SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_DESTROY_BY_RCU, - sighand_ctor); + SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_DESTROY_BY_RCU| + SLAB_NOTRACK, sighand_ctor); signal_cachep = kmem_cache_create("signal_cache", sizeof(struct signal_struct), 0, - SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL); + SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL); files_cachep = kmem_cache_create("files_cache", sizeof(struct files_struct), 0, - SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL); + SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL); fs_cachep = kmem_cache_create("fs_cache", sizeof(struct fs_struct), 0, - SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL); + SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL); mm_cachep = kmem_cache_create("mm_struct", sizeof(struct mm_struct), ARCH_MIN_MMSTRUCT_ALIGN, - SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL); + SLAB_HWCACHE_ALIGN|SLAB_PANIC|SLAB_NOTRACK, NULL); + vm_area_cachep = KMEM_CACHE(vm_area_struct, SLAB_PANIC); mmap_init(); } @@ -1539,12 +1518,16 @@ static int unshare_fs(unsigned long unshare_flags, struct fs_struct **new_fsp) { struct fs_struct *fs = current->fs; - if ((unshare_flags & CLONE_FS) && - (fs && atomic_read(&fs->count) > 1)) { - *new_fsp = __copy_fs_struct(current->fs); - if (!*new_fsp) - return -ENOMEM; - } + if (!(unshare_flags & CLONE_FS) || !fs) + return 0; + + /* don't need lock here; in the worst case we'll do useless copy */ + if (fs->users == 1) + return 0; + + *new_fsp = copy_fs_struct(fs); + if (!*new_fsp) + return -ENOMEM; return 0; } @@ -1660,8 +1643,13 @@ SYSCALL_DEFINE1(unshare, unsigned long, unshare_flags) if (new_fs) { fs = current->fs; + write_lock(&fs->lock); current->fs = new_fs; - new_fs = fs; + if (--fs->users) + new_fs = NULL; + else + new_fs = fs; + write_unlock(&fs->lock); } if (new_mm) { @@ -1700,7 +1688,7 @@ bad_unshare_cleanup_sigh: bad_unshare_cleanup_fs: if (new_fs) - put_fs_struct(new_fs); + free_fs_struct(new_fs); bad_unshare_cleanup_thread: bad_unshare_out: diff --git a/kernel/freezer.c b/kernel/freezer.c index 2f4936cf708..bd1d42b17cb 100644 --- a/kernel/freezer.c +++ b/kernel/freezer.c @@ -44,12 +44,19 @@ void refrigerator(void) recalc_sigpending(); /* We sent fake signal, clean it up */ spin_unlock_irq(¤t->sighand->siglock); + /* prevent accounting of that task to load */ + current->flags |= PF_FREEZING; + for (;;) { set_current_state(TASK_UNINTERRUPTIBLE); if (!frozen(current)) break; schedule(); } + + /* Remove the accounting blocker */ + current->flags &= ~PF_FREEZING; + pr_debug("%s left refrigerator\n", current->comm); __set_current_state(save); } diff --git a/kernel/futex.c b/kernel/futex.c index f89d373a9c6..248dd119a86 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -19,6 +19,10 @@ * PRIVATE futexes by Eric Dumazet * Copyright (C) 2007 Eric Dumazet <dada1@cosmosbay.com> * + * Requeue-PI support by Darren Hart <dvhltc@us.ibm.com> + * Copyright (C) IBM Corporation, 2009 + * Thanks to Thomas Gleixner for conceptual design and careful reviews. + * * Thanks to Ben LaHaise for yelling "hashed waitqueues" loudly * enough at me, Linus for the original (flawed) idea, Matthew * Kirkwood for proof-of-concept implementation. @@ -96,8 +100,8 @@ struct futex_pi_state { */ struct futex_q { struct plist_node list; - /* There can only be a single waiter */ - wait_queue_head_t waiter; + /* Waiter reference */ + struct task_struct *task; /* Which hash list lock to use: */ spinlock_t *lock_ptr; @@ -107,14 +111,21 @@ struct futex_q { /* Optional priority inheritance state: */ struct futex_pi_state *pi_state; - struct task_struct *task; + + /* rt_waiter storage for requeue_pi: */ + struct rt_mutex_waiter *rt_waiter; + + /* The expected requeue pi target futex key: */ + union futex_key *requeue_pi_key; /* Bitset for the optional bitmasked wakeup */ u32 bitset; }; /* - * Split the global futex_lock into every hash list lock. + * Hash buckets are shared by all the futex_keys that hash to the same + * location. Each key may have multiple futex_q structures, one for each task + * waiting on a futex. */ struct futex_hash_bucket { spinlock_t lock; @@ -189,9 +200,9 @@ static void drop_futex_key_refs(union futex_key *key) /** * get_futex_key - Get parameters which are the keys for a futex. * @uaddr: virtual address of the futex - * @shared: NULL for a PROCESS_PRIVATE futex, - * ¤t->mm->mmap_sem for a PROCESS_SHARED futex + * @fshared: 0 for a PROCESS_PRIVATE futex, 1 for PROCESS_SHARED * @key: address where result is stored. + * @rw: mapping needs to be read/write (values: VERIFY_READ, VERIFY_WRITE) * * Returns a negative error code or 0 * The key words are stored in *key on success. @@ -200,11 +211,10 @@ static void drop_futex_key_refs(union futex_key *key) * offset_within_page). For private mappings, it's (uaddr, current->mm). * We can usually work out the index without swapping in the page. * - * fshared is NULL for PROCESS_PRIVATE futexes - * For other futexes, it points to ¤t->mm->mmap_sem and - * caller must have taken the reader lock. but NOT any spinlocks. + * lock_page() might sleep, the caller should not hold a spinlock. */ -static int get_futex_key(u32 __user *uaddr, int fshared, union futex_key *key) +static int +get_futex_key(u32 __user *uaddr, int fshared, union futex_key *key, int rw) { unsigned long address = (unsigned long)uaddr; struct mm_struct *mm = current->mm; @@ -227,7 +237,7 @@ static int get_futex_key(u32 __user *uaddr, int fshared, union futex_key *key) * but access_ok() should be faster than find_vma() */ if (!fshared) { - if (unlikely(!access_ok(VERIFY_WRITE, uaddr, sizeof(u32)))) + if (unlikely(!access_ok(rw, uaddr, sizeof(u32)))) return -EFAULT; key->private.mm = mm; key->private.address = address; @@ -236,10 +246,11 @@ static int get_futex_key(u32 __user *uaddr, int fshared, union futex_key *key) } again: - err = get_user_pages_fast(address, 1, 0, &page); + err = get_user_pages_fast(address, 1, rw == VERIFY_WRITE, &page); if (err < 0) return err; + page = compound_head(page); lock_page(page); if (!page->mapping) { unlock_page(page); @@ -277,6 +288,44 @@ void put_futex_key(int fshared, union futex_key *key) drop_futex_key_refs(key); } +/* + * fault_in_user_writeable - fault in user address and verify RW access + * @uaddr: pointer to faulting user space address + * + * Slow path to fixup the fault we just took in the atomic write + * access to @uaddr. + * + * We have no generic implementation of a non destructive write to the + * user address. We know that we faulted in the atomic pagefault + * disabled section so we can as well avoid the #PF overhead by + * calling get_user_pages() right away. + */ +static int fault_in_user_writeable(u32 __user *uaddr) +{ + int ret = get_user_pages(current, current->mm, (unsigned long)uaddr, + 1, 1, 0, NULL, NULL); + return ret < 0 ? ret : 0; +} + +/** + * futex_top_waiter() - Return the highest priority waiter on a futex + * @hb: the hash bucket the futex_q's reside in + * @key: the futex key (to distinguish it from other futex futex_q's) + * + * Must be called with the hb lock held. + */ +static struct futex_q *futex_top_waiter(struct futex_hash_bucket *hb, + union futex_key *key) +{ + struct futex_q *this; + + plist_for_each_entry(this, &hb->chain, list) { + if (match_futex(&this->key, key)) + return this; + } + return NULL; +} + static u32 cmpxchg_futex_value_locked(u32 __user *uaddr, u32 uval, u32 newval) { u32 curval; @@ -299,41 +348,6 @@ static int get_futex_value_locked(u32 *dest, u32 __user *from) return ret ? -EFAULT : 0; } -/* - * Fault handling. - */ -static int futex_handle_fault(unsigned long address, int attempt) -{ - struct vm_area_struct * vma; - struct mm_struct *mm = current->mm; - int ret = -EFAULT; - - if (attempt > 2) - return ret; - - down_read(&mm->mmap_sem); - vma = find_vma(mm, address); - if (vma && address >= vma->vm_start && - (vma->vm_flags & VM_WRITE)) { - int fault; - fault = handle_mm_fault(mm, vma, address, 1); - if (unlikely((fault & VM_FAULT_ERROR))) { -#if 0 - /* XXX: let's do this when we verify it is OK */ - if (ret & VM_FAULT_OOM) - ret = -ENOMEM; -#endif - } else { - ret = 0; - if (fault & VM_FAULT_MAJOR) - current->maj_flt++; - else - current->min_flt++; - } - } - up_read(&mm->mmap_sem); - return ret; -} /* * PI code: @@ -573,29 +587,160 @@ lookup_pi_state(u32 uval, struct futex_hash_bucket *hb, return 0; } +/** + * futex_lock_pi_atomic() - atomic work required to acquire a pi aware futex + * @uaddr: the pi futex user address + * @hb: the pi futex hash bucket + * @key: the futex key associated with uaddr and hb + * @ps: the pi_state pointer where we store the result of the + * lookup + * @task: the task to perform the atomic lock work for. This will + * be "current" except in the case of requeue pi. + * @set_waiters: force setting the FUTEX_WAITERS bit (1) or not (0) + * + * Returns: + * 0 - ready to wait + * 1 - acquired the lock + * <0 - error + * + * The hb->lock and futex_key refs shall be held by the caller. + */ +static int futex_lock_pi_atomic(u32 __user *uaddr, struct futex_hash_bucket *hb, + union futex_key *key, + struct futex_pi_state **ps, + struct task_struct *task, int set_waiters) +{ + int lock_taken, ret, ownerdied = 0; + u32 uval, newval, curval; + +retry: + ret = lock_taken = 0; + + /* + * To avoid races, we attempt to take the lock here again + * (by doing a 0 -> TID atomic cmpxchg), while holding all + * the locks. It will most likely not succeed. + */ + newval = task_pid_vnr(task); + if (set_waiters) + newval |= FUTEX_WAITERS; + + curval = cmpxchg_futex_value_locked(uaddr, 0, newval); + + if (unlikely(curval == -EFAULT)) + return -EFAULT; + + /* + * Detect deadlocks. + */ + if ((unlikely((curval & FUTEX_TID_MASK) == task_pid_vnr(task)))) + return -EDEADLK; + + /* + * Surprise - we got the lock. Just return to userspace: + */ + if (unlikely(!curval)) + return 1; + + uval = curval; + + /* + * Set the FUTEX_WAITERS flag, so the owner will know it has someone + * to wake at the next unlock. + */ + newval = curval | FUTEX_WAITERS; + + /* + * There are two cases, where a futex might have no owner (the + * owner TID is 0): OWNER_DIED. We take over the futex in this + * case. We also do an unconditional take over, when the owner + * of the futex died. + * + * This is safe as we are protected by the hash bucket lock ! + */ + if (unlikely(ownerdied || !(curval & FUTEX_TID_MASK))) { + /* Keep the OWNER_DIED bit */ + newval = (curval & ~FUTEX_TID_MASK) | task_pid_vnr(task); + ownerdied = 0; + lock_taken = 1; + } + + curval = cmpxchg_futex_value_locked(uaddr, uval, newval); + + if (unlikely(curval == -EFAULT)) + return -EFAULT; + if (unlikely(curval != uval)) + goto retry; + + /* + * We took the lock due to owner died take over. + */ + if (unlikely(lock_taken)) + return 1; + + /* + * We dont have the lock. Look up the PI state (or create it if + * we are the first waiter): + */ + ret = lookup_pi_state(uval, hb, key, ps); + + if (unlikely(ret)) { + switch (ret) { + case -ESRCH: + /* + * No owner found for this futex. Check if the + * OWNER_DIED bit is set to figure out whether + * this is a robust futex or not. + */ + if (get_futex_value_locked(&curval, uaddr)) + return -EFAULT; + + /* + * We simply start over in case of a robust + * futex. The code above will take the futex + * and return happy. + */ + if (curval & FUTEX_OWNER_DIED) { + ownerdied = 1; + goto retry; + } + default: + break; + } + } + + return ret; +} + /* * The hash bucket lock must be held when this is called. * Afterwards, the futex_q must not be accessed. */ static void wake_futex(struct futex_q *q) { - plist_del(&q->list, &q->list.plist); + struct task_struct *p = q->task; + /* - * The lock in wake_up_all() is a crucial memory barrier after the - * plist_del() and also before assigning to q->lock_ptr. + * We set q->lock_ptr = NULL _before_ we wake up the task. If + * a non futex wake up happens on another CPU then the task + * might exit and p would dereference a non existing task + * struct. Prevent this by holding a reference on p across the + * wake up. */ - wake_up(&q->waiter); + get_task_struct(p); + + plist_del(&q->list, &q->list.plist); /* - * The waiting task can free the futex_q as soon as this is written, - * without taking any locks. This must come last. - * - * A memory barrier is required here to prevent the following store - * to lock_ptr from getting ahead of the wakeup. Clearing the lock - * at the end of wake_up_all() does not prevent this store from - * moving. + * The waiting task can free the futex_q as soon as + * q->lock_ptr = NULL is written, without taking any locks. A + * memory barrier is required here to prevent the following + * store to lock_ptr from getting ahead of the plist_del. */ smp_wmb(); q->lock_ptr = NULL; + + wake_up_state(p, TASK_NORMAL); + put_task_struct(p); } static int wake_futex_pi(u32 __user *uaddr, u32 uval, struct futex_q *this) @@ -692,9 +837,16 @@ double_lock_hb(struct futex_hash_bucket *hb1, struct futex_hash_bucket *hb2) } } +static inline void +double_unlock_hb(struct futex_hash_bucket *hb1, struct futex_hash_bucket *hb2) +{ + spin_unlock(&hb1->lock); + if (hb1 != hb2) + spin_unlock(&hb2->lock); +} + /* - * Wake up all waiters hashed on the physical page that is mapped - * to this virtual address: + * Wake up waiters matching bitset queued on this futex (uaddr). */ static int futex_wake(u32 __user *uaddr, int fshared, int nr_wake, u32 bitset) { @@ -707,7 +859,7 @@ static int futex_wake(u32 __user *uaddr, int fshared, int nr_wake, u32 bitset) if (!bitset) return -EINVAL; - ret = get_futex_key(uaddr, fshared, &key); + ret = get_futex_key(uaddr, fshared, &key, VERIFY_READ); if (unlikely(ret != 0)) goto out; @@ -717,7 +869,7 @@ static int futex_wake(u32 __user *uaddr, int fshared, int nr_wake, u32 bitset) plist_for_each_entry_safe(this, next, head, list) { if (match_futex (&this->key, &key)) { - if (this->pi_state) { + if (this->pi_state || this->rt_waiter) { ret = -EINVAL; break; } @@ -750,29 +902,25 @@ futex_wake_op(u32 __user *uaddr1, int fshared, u32 __user *uaddr2, struct futex_hash_bucket *hb1, *hb2; struct plist_head *head; struct futex_q *this, *next; - int ret, op_ret, attempt = 0; + int ret, op_ret; -retryfull: - ret = get_futex_key(uaddr1, fshared, &key1); +retry: + ret = get_futex_key(uaddr1, fshared, &key1, VERIFY_READ); if (unlikely(ret != 0)) goto out; - ret = get_futex_key(uaddr2, fshared, &key2); + ret = get_futex_key(uaddr2, fshared, &key2, VERIFY_WRITE); if (unlikely(ret != 0)) goto out_put_key1; hb1 = hash_futex(&key1); hb2 = hash_futex(&key2); -retry: double_lock_hb(hb1, hb2); - +retry_private: op_ret = futex_atomic_op_inuser(op, uaddr2); if (unlikely(op_ret < 0)) { - u32 dummy; - spin_unlock(&hb1->lock); - if (hb1 != hb2) - spin_unlock(&hb2->lock); + double_unlock_hb(hb1, hb2); #ifndef CONFIG_MMU /* @@ -788,26 +936,16 @@ retry: goto out_put_keys; } - /* - * futex_atomic_op_inuser needs to both read and write - * *(int __user *)uaddr2, but we can't modify it - * non-atomically. Therefore, if get_user below is not - * enough, we need to handle the fault ourselves, while - * still holding the mmap_sem. - */ - if (attempt++) { - ret = futex_handle_fault((unsigned long)uaddr2, - attempt); - if (ret) - goto out_put_keys; - goto retry; - } - - ret = get_user(dummy, uaddr2); + ret = fault_in_user_writeable(uaddr2); if (ret) - return ret; + goto out_put_keys; + + if (!fshared) + goto retry_private; - goto retryfull; + put_futex_key(fshared, &key2); + put_futex_key(fshared, &key1); + goto retry; } head = &hb1->chain; @@ -834,9 +972,7 @@ retry: ret += op_ret; } - spin_unlock(&hb1->lock); - if (hb1 != hb2) - spin_unlock(&hb2->lock); + double_unlock_hb(hb1, hb2); out_put_keys: put_futex_key(fshared, &key2); out_put_key1: @@ -845,30 +981,205 @@ out: return ret; } -/* - * Requeue all waiters hashed on one physical page to another - * physical page. +/** + * requeue_futex() - Requeue a futex_q from one hb to another + * @q: the futex_q to requeue + * @hb1: the source hash_bucket + * @hb2: the target hash_bucket + * @key2: the new key for the requeued futex_q + */ +static inline +void requeue_futex(struct futex_q *q, struct futex_hash_bucket *hb1, + struct futex_hash_bucket *hb2, union futex_key *key2) +{ + + /* + * If key1 and key2 hash to the same bucket, no need to + * requeue. + */ + if (likely(&hb1->chain != &hb2->chain)) { + plist_del(&q->list, &hb1->chain); + plist_add(&q->list, &hb2->chain); + q->lock_ptr = &hb2->lock; +#ifdef CONFIG_DEBUG_PI_LIST + q->list.plist.lock = &hb2->lock; +#endif + } + get_futex_key_refs(key2); + q->key = *key2; +} + +/** + * requeue_pi_wake_futex() - Wake a task that acquired the lock during requeue + * q: the futex_q + * key: the key of the requeue target futex + * hb: the hash_bucket of the requeue target futex + * + * During futex_requeue, with requeue_pi=1, it is possible to acquire the + * target futex if it is uncontended or via a lock steal. Set the futex_q key + * to the requeue target futex so the waiter can detect the wakeup on the right + * futex, but remove it from the hb and NULL the rt_waiter so it can detect + * atomic lock acquisition. Set the q->lock_ptr to the requeue target hb->lock + * to protect access to the pi_state to fixup the owner later. Must be called + * with both q->lock_ptr and hb->lock held. + */ +static inline +void requeue_pi_wake_futex(struct futex_q *q, union futex_key *key, + struct futex_hash_bucket *hb) +{ + drop_futex_key_refs(&q->key); + get_futex_key_refs(key); + q->key = *key; + + WARN_ON(plist_node_empty(&q->list)); + plist_del(&q->list, &q->list.plist); + + WARN_ON(!q->rt_waiter); + q->rt_waiter = NULL; + + q->lock_ptr = &hb->lock; +#ifdef CONFIG_DEBUG_PI_LIST + q->list.plist.lock = &hb->lock; +#endif + + wake_up_state(q->task, TASK_NORMAL); +} + +/** + * futex_proxy_trylock_atomic() - Attempt an atomic lock for the top waiter + * @pifutex: the user address of the to futex + * @hb1: the from futex hash bucket, must be locked by the caller + * @hb2: the to futex hash bucket, must be locked by the caller + * @key1: the from futex key + * @key2: the to futex key + * @ps: address to store the pi_state pointer + * @set_waiters: force setting the FUTEX_WAITERS bit (1) or not (0) + * + * Try and get the lock on behalf of the top waiter if we can do it atomically. + * Wake the top waiter if we succeed. If the caller specified set_waiters, + * then direct futex_lock_pi_atomic() to force setting the FUTEX_WAITERS bit. + * hb1 and hb2 must be held by the caller. + * + * Returns: + * 0 - failed to acquire the lock atomicly + * 1 - acquired the lock + * <0 - error + */ +static int futex_proxy_trylock_atomic(u32 __user *pifutex, + struct futex_hash_bucket *hb1, + struct futex_hash_bucket *hb2, + union futex_key *key1, union futex_key *key2, + struct futex_pi_state **ps, int set_waiters) +{ + struct futex_q *top_waiter = NULL; + u32 curval; + int ret; + + if (get_futex_value_locked(&curval, pifutex)) + return -EFAULT; + + /* + * Find the top_waiter and determine if there are additional waiters. + * If the caller intends to requeue more than 1 waiter to pifutex, + * force futex_lock_pi_atomic() to set the FUTEX_WAITERS bit now, + * as we have means to handle the possible fault. If not, don't set + * the bit unecessarily as it will force the subsequent unlock to enter + * the kernel. + */ + top_waiter = futex_top_waiter(hb1, key1); + + /* There are no waiters, nothing for us to do. */ + if (!top_waiter) + return 0; + + /* Ensure we requeue to the expected futex. */ + if (!match_futex(top_waiter->requeue_pi_key, key2)) + return -EINVAL; + + /* + * Try to take the lock for top_waiter. Set the FUTEX_WAITERS bit in + * the contended case or if set_waiters is 1. The pi_state is returned + * in ps in contended cases. + */ + ret = futex_lock_pi_atomic(pifutex, hb2, key2, ps, top_waiter->task, + set_waiters); + if (ret == 1) + requeue_pi_wake_futex(top_waiter, key2, hb2); + + return ret; +} + +/** + * futex_requeue() - Requeue waiters from uaddr1 to uaddr2 + * uaddr1: source futex user address + * uaddr2: target futex user address + * nr_wake: number of waiters to wake (must be 1 for requeue_pi) + * nr_requeue: number of waiters to requeue (0-INT_MAX) + * requeue_pi: if we are attempting to requeue from a non-pi futex to a + * pi futex (pi to pi requeue is not supported) + * + * Requeue waiters on uaddr1 to uaddr2. In the requeue_pi case, try to acquire + * uaddr2 atomically on behalf of the top waiter. + * + * Returns: + * >=0 - on success, the number of tasks requeued or woken + * <0 - on error */ static int futex_requeue(u32 __user *uaddr1, int fshared, u32 __user *uaddr2, - int nr_wake, int nr_requeue, u32 *cmpval) + int nr_wake, int nr_requeue, u32 *cmpval, + int requeue_pi) { union futex_key key1 = FUTEX_KEY_INIT, key2 = FUTEX_KEY_INIT; + int drop_count = 0, task_count = 0, ret; + struct futex_pi_state *pi_state = NULL; struct futex_hash_bucket *hb1, *hb2; struct plist_head *head1; struct futex_q *this, *next; - int ret, drop_count = 0; + u32 curval2; + + if (requeue_pi) { + /* + * requeue_pi requires a pi_state, try to allocate it now + * without any locks in case it fails. + */ + if (refill_pi_state_cache()) + return -ENOMEM; + /* + * requeue_pi must wake as many tasks as it can, up to nr_wake + * + nr_requeue, since it acquires the rt_mutex prior to + * returning to userspace, so as to not leave the rt_mutex with + * waiters and no owner. However, second and third wake-ups + * cannot be predicted as they involve race conditions with the + * first wake and a fault while looking up the pi_state. Both + * pthread_cond_signal() and pthread_cond_broadcast() should + * use nr_wake=1. + */ + if (nr_wake != 1) + return -EINVAL; + } retry: - ret = get_futex_key(uaddr1, fshared, &key1); + if (pi_state != NULL) { + /* + * We will have to lookup the pi_state again, so free this one + * to keep the accounting correct. + */ + free_pi_state(pi_state); + pi_state = NULL; + } + + ret = get_futex_key(uaddr1, fshared, &key1, VERIFY_READ); if (unlikely(ret != 0)) goto out; - ret = get_futex_key(uaddr2, fshared, &key2); + ret = get_futex_key(uaddr2, fshared, &key2, + requeue_pi ? VERIFY_WRITE : VERIFY_READ); if (unlikely(ret != 0)) goto out_put_key1; hb1 = hash_futex(&key1); hb2 = hash_futex(&key2); +retry_private: double_lock_hb(hb1, hb2); if (likely(cmpval != NULL)) { @@ -877,16 +1188,18 @@ retry: ret = get_futex_value_locked(&curval, uaddr1); if (unlikely(ret)) { - spin_unlock(&hb1->lock); - if (hb1 != hb2) - spin_unlock(&hb2->lock); + double_unlock_hb(hb1, hb2); ret = get_user(curval, uaddr1); + if (ret) + goto out_put_keys; - if (!ret) - goto retry; + if (!fshared) + goto retry_private; - goto out_put_keys; + put_futex_key(fshared, &key2); + put_futex_key(fshared, &key1); + goto retry; } if (curval != *cmpval) { ret = -EAGAIN; @@ -894,40 +1207,123 @@ retry: } } + if (requeue_pi && (task_count - nr_wake < nr_requeue)) { + /* + * Attempt to acquire uaddr2 and wake the top waiter. If we + * intend to requeue waiters, force setting the FUTEX_WAITERS + * bit. We force this here where we are able to easily handle + * faults rather in the requeue loop below. + */ + ret = futex_proxy_trylock_atomic(uaddr2, hb1, hb2, &key1, + &key2, &pi_state, nr_requeue); + + /* + * At this point the top_waiter has either taken uaddr2 or is + * waiting on it. If the former, then the pi_state will not + * exist yet, look it up one more time to ensure we have a + * reference to it. + */ + if (ret == 1) { + WARN_ON(pi_state); + task_count++; + ret = get_futex_value_locked(&curval2, uaddr2); + if (!ret) + ret = lookup_pi_state(curval2, hb2, &key2, + &pi_state); + } + + switch (ret) { + case 0: + break; + case -EFAULT: + double_unlock_hb(hb1, hb2); + put_futex_key(fshared, &key2); + put_futex_key(fshared, &key1); + ret = fault_in_user_writeable(uaddr2); + if (!ret) + goto retry; + goto out; + case -EAGAIN: + /* The owner was exiting, try again. */ + double_unlock_hb(hb1, hb2); + put_futex_key(fshared, &key2); + put_futex_key(fshared, &key1); + cond_resched(); + goto retry; + default: + goto out_unlock; + } + } + head1 = &hb1->chain; plist_for_each_entry_safe(this, next, head1, list) { - if (!match_futex (&this->key, &key1)) + if (task_count - nr_wake >= nr_requeue) + break; + + if (!match_futex(&this->key, &key1)) continue; - if (++ret <= nr_wake) { + + /* + * FUTEX_WAIT_REQEUE_PI and FUTEX_CMP_REQUEUE_PI should always + * be paired with each other and no other futex ops. + */ + if ((requeue_pi && !this->rt_waiter) || + (!requeue_pi && this->rt_waiter)) { + ret = -EINVAL; + break; + } + + /* + * Wake nr_wake waiters. For requeue_pi, if we acquired the + * lock, we already woke the top_waiter. If not, it will be + * woken by futex_unlock_pi(). + */ + if (++task_count <= nr_wake && !requeue_pi) { wake_futex(this); - } else { - /* - * If key1 and key2 hash to the same bucket, no need to - * requeue. - */ - if (likely(head1 != &hb2->chain)) { - plist_del(&this->list, &hb1->chain); - plist_add(&this->list, &hb2->chain); - this->lock_ptr = &hb2->lock; -#ifdef CONFIG_DEBUG_PI_LIST - this->list.plist.lock = &hb2->lock; -#endif - } - this->key = key2; - get_futex_key_refs(&key2); - drop_count++; + continue; + } - if (ret - nr_wake >= nr_requeue) - break; + /* Ensure we requeue to the expected futex for requeue_pi. */ + if (requeue_pi && !match_futex(this->requeue_pi_key, &key2)) { + ret = -EINVAL; + break; + } + + /* + * Requeue nr_requeue waiters and possibly one more in the case + * of requeue_pi if we couldn't acquire the lock atomically. + */ + if (requeue_pi) { + /* Prepare the waiter to take the rt_mutex. */ + atomic_inc(&pi_state->refcount); + this->pi_state = pi_state; + ret = rt_mutex_start_proxy_lock(&pi_state->pi_mutex, + this->rt_waiter, + this->task, 1); + if (ret == 1) { + /* We got the lock. */ + requeue_pi_wake_futex(this, &key2, hb2); + continue; + } else if (ret) { + /* -EDEADLK */ + this->pi_state = NULL; + free_pi_state(pi_state); + goto out_unlock; + } } + requeue_futex(this, hb1, hb2, &key2); + drop_count++; } out_unlock: - spin_unlock(&hb1->lock); - if (hb1 != hb2) - spin_unlock(&hb2->lock); + double_unlock_hb(hb1, hb2); - /* drop_futex_key_refs() must be called outside the spinlocks. */ + /* + * drop_futex_key_refs() must be called outside the spinlocks. During + * the requeue we moved futex_q's from the hash bucket at key1 to the + * one at key2 and updated their key pointer. We no longer need to + * hold the references to key1. + */ while (--drop_count >= 0) drop_futex_key_refs(&key1); @@ -936,7 +1332,9 @@ out_put_keys: out_put_key1: put_futex_key(fshared, &key1); out: - return ret; + if (pi_state != NULL) + free_pi_state(pi_state); + return ret ? ret : task_count; } /* The key must be already stored in q->key. */ @@ -944,8 +1342,6 @@ static inline struct futex_hash_bucket *queue_lock(struct futex_q *q) { struct futex_hash_bucket *hb; - init_waitqueue_head(&q->waiter); - get_futex_key_refs(&q->key); hb = hash_futex(&q->key); q->lock_ptr = &hb->lock; @@ -1063,7 +1459,7 @@ static int fixup_pi_state_owner(u32 __user *uaddr, struct futex_q *q, struct futex_pi_state *pi_state = q->pi_state; struct task_struct *oldowner = pi_state->owner; u32 uval, curval, newval; - int ret, attempt = 0; + int ret; /* Owner died? */ if (!pi_state->owner) @@ -1076,11 +1472,9 @@ static int fixup_pi_state_owner(u32 __user *uaddr, struct futex_q *q, * in the user space variable. This must be atomic as we have * to preserve the owner died bit here. * - * Note: We write the user space value _before_ changing the - * pi_state because we can fault here. Imagine swapped out - * pages or a fork, which was running right before we acquired - * mmap_sem, that marked all the anonymous memory readonly for - * cow. + * Note: We write the user space value _before_ changing the pi_state + * because we can fault here. Imagine swapped out pages or a fork + * that marked all the anonymous memory readonly for cow. * * Modifying pi_state _before_ the user space value would * leave the pi_state in an inconsistent state when we fault @@ -1136,7 +1530,7 @@ retry: handle_fault: spin_unlock(q->lock_ptr); - ret = futex_handle_fault((unsigned long)uaddr, attempt++); + ret = fault_in_user_writeable(uaddr); spin_lock(q->lock_ptr); @@ -1158,36 +1552,152 @@ handle_fault: */ #define FLAGS_SHARED 0x01 #define FLAGS_CLOCKRT 0x02 +#define FLAGS_HAS_TIMEOUT 0x04 static long futex_wait_restart(struct restart_block *restart); -static int futex_wait(u32 __user *uaddr, int fshared, - u32 val, ktime_t *abs_time, u32 bitset, int clockrt) +/** + * fixup_owner() - Post lock pi_state and corner case management + * @uaddr: user address of the futex + * @fshared: whether the futex is shared (1) or not (0) + * @q: futex_q (contains pi_state and access to the rt_mutex) + * @locked: if the attempt to take the rt_mutex succeeded (1) or not (0) + * + * After attempting to lock an rt_mutex, this function is called to cleanup + * the pi_state owner as well as handle race conditions that may allow us to + * acquire the lock. Must be called with the hb lock held. + * + * Returns: + * 1 - success, lock taken + * 0 - success, lock not taken + * <0 - on error (-EFAULT) + */ +static int fixup_owner(u32 __user *uaddr, int fshared, struct futex_q *q, + int locked) { - struct task_struct *curr = current; - DECLARE_WAITQUEUE(wait, curr); - struct futex_hash_bucket *hb; - struct futex_q q; - u32 uval; - int ret; - struct hrtimer_sleeper t; - int rem = 0; + struct task_struct *owner; + int ret = 0; - if (!bitset) - return -EINVAL; + if (locked) { + /* + * Got the lock. We might not be the anticipated owner if we + * did a lock-steal - fix up the PI-state in that case: + */ + if (q->pi_state->owner != current) + ret = fixup_pi_state_owner(uaddr, q, current, fshared); + goto out; + } - q.pi_state = NULL; - q.bitset = bitset; -retry: - q.key = FUTEX_KEY_INIT; - ret = get_futex_key(uaddr, fshared, &q.key); - if (unlikely(ret != 0)) + /* + * Catch the rare case, where the lock was released when we were on the + * way back before we locked the hash bucket. + */ + if (q->pi_state->owner == current) { + /* + * Try to get the rt_mutex now. This might fail as some other + * task acquired the rt_mutex after we removed ourself from the + * rt_mutex waiters list. + */ + if (rt_mutex_trylock(&q->pi_state->pi_mutex)) { + locked = 1; + goto out; + } + + /* + * pi_state is incorrect, some other task did a lock steal and + * we returned due to timeout or signal without taking the + * rt_mutex. Too late. We can access the rt_mutex_owner without + * locking, as the other task is now blocked on the hash bucket + * lock. Fix the state up. + */ + owner = rt_mutex_owner(&q->pi_state->pi_mutex); + ret = fixup_pi_state_owner(uaddr, q, owner, fshared); goto out; + } - hb = queue_lock(&q); + /* + * Paranoia check. If we did not take the lock, then we should not be + * the owner, nor the pending owner, of the rt_mutex. + */ + if (rt_mutex_owner(&q->pi_state->pi_mutex) == current) + printk(KERN_ERR "fixup_owner: ret = %d pi-mutex: %p " + "pi-state %p\n", ret, + q->pi_state->pi_mutex.owner, + q->pi_state->owner); + +out: + return ret ? ret : locked; +} + +/** + * futex_wait_queue_me() - queue_me() and wait for wakeup, timeout, or signal + * @hb: the futex hash bucket, must be locked by the caller + * @q: the futex_q to queue up on + * @timeout: the prepared hrtimer_sleeper, or null for no timeout + */ +static void futex_wait_queue_me(struct futex_hash_bucket *hb, struct futex_q *q, + struct hrtimer_sleeper *timeout) +{ + queue_me(q, hb); /* - * Access the page AFTER the futex is queued. + * There might have been scheduling since the queue_me(), as we + * cannot hold a spinlock across the get_user() in case it + * faults, and we cannot just set TASK_INTERRUPTIBLE state when + * queueing ourselves into the futex hash. This code thus has to + * rely on the futex_wake() code removing us from hash when it + * wakes us up. + */ + set_current_state(TASK_INTERRUPTIBLE); + + /* Arm the timer */ + if (timeout) { + hrtimer_start_expires(&timeout->timer, HRTIMER_MODE_ABS); + if (!hrtimer_active(&timeout->timer)) + timeout->task = NULL; + } + + /* + * !plist_node_empty() is safe here without any lock. + * q.lock_ptr != 0 is not safe, because of ordering against wakeup. + */ + if (likely(!plist_node_empty(&q->list))) { + /* + * If the timer has already expired, current will already be + * flagged for rescheduling. Only call schedule if there + * is no timeout, or if it has yet to expire. + */ + if (!timeout || timeout->task) + schedule(); + } + __set_current_state(TASK_RUNNING); +} + +/** + * futex_wait_setup() - Prepare to wait on a futex + * @uaddr: the futex userspace address + * @val: the expected value + * @fshared: whether the futex is shared (1) or not (0) + * @q: the associated futex_q + * @hb: storage for hash_bucket pointer to be returned to caller + * + * Setup the futex_q and locate the hash_bucket. Get the futex value and + * compare it with the expected value. Handle atomic faults internally. + * Return with the hb lock held and a q.key reference on success, and unlocked + * with no q.key reference on failure. + * + * Returns: + * 0 - uaddr contains val and hb has been locked + * <1 - -EFAULT or -EWOULDBLOCK (uaddr does not contain val) and hb is unlcoked + */ +static int futex_wait_setup(u32 __user *uaddr, u32 val, int fshared, + struct futex_q *q, struct futex_hash_bucket **hb) +{ + u32 uval; + int ret; + + /* + * Access the page AFTER the hash-bucket is locked. * Order is important: * * Userspace waiter: val = var; if (cond(val)) futex_wait(&var, val); @@ -1202,121 +1712,116 @@ retry: * A consequence is that futex_wait() can return zero and absorb * a wakeup when *uaddr != val on entry to the syscall. This is * rare, but normal. - * - * for shared futexes, we hold the mmap semaphore, so the mapping - * cannot have changed since we looked it up in get_futex_key. */ +retry: + q->key = FUTEX_KEY_INIT; + ret = get_futex_key(uaddr, fshared, &q->key, VERIFY_READ); + if (unlikely(ret != 0)) + return ret; + +retry_private: + *hb = queue_lock(q); + ret = get_futex_value_locked(&uval, uaddr); - if (unlikely(ret)) { - queue_unlock(&q, hb); - put_futex_key(fshared, &q.key); + if (ret) { + queue_unlock(q, *hb); ret = get_user(uval, uaddr); + if (ret) + goto out; - if (!ret) - goto retry; - return ret; + if (!fshared) + goto retry_private; + + put_futex_key(fshared, &q->key); + goto retry; } - ret = -EWOULDBLOCK; - if (uval != val) - goto out_unlock_put_key; - /* Only actually queue if *uaddr contained val. */ - queue_me(&q, hb); + if (uval != val) { + queue_unlock(q, *hb); + ret = -EWOULDBLOCK; + } - /* - * There might have been scheduling since the queue_me(), as we - * cannot hold a spinlock across the get_user() in case it - * faults, and we cannot just set TASK_INTERRUPTIBLE state when - * queueing ourselves into the futex hash. This code thus has to - * rely on the futex_wake() code removing us from hash when it - * wakes us up. - */ +out: + if (ret) + put_futex_key(fshared, &q->key); + return ret; +} - /* add_wait_queue is the barrier after __set_current_state. */ - __set_current_state(TASK_INTERRUPTIBLE); - add_wait_queue(&q.waiter, &wait); - /* - * !plist_node_empty() is safe here without any lock. - * q.lock_ptr != 0 is not safe, because of ordering against wakeup. - */ - if (likely(!plist_node_empty(&q.list))) { - if (!abs_time) - schedule(); - else { - unsigned long slack; - slack = current->timer_slack_ns; - if (rt_task(current)) - slack = 0; - hrtimer_init_on_stack(&t.timer, - clockrt ? CLOCK_REALTIME : - CLOCK_MONOTONIC, - HRTIMER_MODE_ABS); - hrtimer_init_sleeper(&t, current); - hrtimer_set_expires_range_ns(&t.timer, *abs_time, slack); - - hrtimer_start_expires(&t.timer, HRTIMER_MODE_ABS); - if (!hrtimer_active(&t.timer)) - t.task = NULL; +static int futex_wait(u32 __user *uaddr, int fshared, + u32 val, ktime_t *abs_time, u32 bitset, int clockrt) +{ + struct hrtimer_sleeper timeout, *to = NULL; + struct restart_block *restart; + struct futex_hash_bucket *hb; + struct futex_q q; + int ret; - /* - * the timer could have already expired, in which - * case current would be flagged for rescheduling. - * Don't bother calling schedule. - */ - if (likely(t.task)) - schedule(); + if (!bitset) + return -EINVAL; - hrtimer_cancel(&t.timer); + q.pi_state = NULL; + q.bitset = bitset; + q.rt_waiter = NULL; + q.requeue_pi_key = NULL; - /* Flag if a timeout occured */ - rem = (t.task == NULL); + if (abs_time) { + to = &timeout; - destroy_hrtimer_on_stack(&t.timer); - } + hrtimer_init_on_stack(&to->timer, clockrt ? CLOCK_REALTIME : + CLOCK_MONOTONIC, HRTIMER_MODE_ABS); + hrtimer_init_sleeper(to, current); + hrtimer_set_expires_range_ns(&to->timer, *abs_time, + current->timer_slack_ns); } - __set_current_state(TASK_RUNNING); - /* - * NOTE: we don't remove ourselves from the waitqueue because - * we are the only user of it. - */ + /* Prepare to wait on uaddr. */ + ret = futex_wait_setup(uaddr, val, fshared, &q, &hb); + if (ret) + goto out; + + /* queue_me and wait for wakeup, timeout, or a signal. */ + futex_wait_queue_me(hb, &q, to); /* If we were woken (and unqueued), we succeeded, whatever. */ + ret = 0; if (!unqueue_me(&q)) - return 0; - if (rem) - return -ETIMEDOUT; + goto out_put_key; + ret = -ETIMEDOUT; + if (to && !to->task) + goto out_put_key; /* * We expect signal_pending(current), but another thread may * have handled it for us already. */ + ret = -ERESTARTSYS; if (!abs_time) - return -ERESTARTSYS; - else { - struct restart_block *restart; - restart = ¤t_thread_info()->restart_block; - restart->fn = futex_wait_restart; - restart->futex.uaddr = (u32 *)uaddr; - restart->futex.val = val; - restart->futex.time = abs_time->tv64; - restart->futex.bitset = bitset; - restart->futex.flags = 0; - - if (fshared) - restart->futex.flags |= FLAGS_SHARED; - if (clockrt) - restart->futex.flags |= FLAGS_CLOCKRT; - return -ERESTART_RESTARTBLOCK; - } + goto out_put_key; -out_unlock_put_key: - queue_unlock(&q, hb); - put_futex_key(fshared, &q.key); + restart = ¤t_thread_info()->restart_block; + restart->fn = futex_wait_restart; + restart->futex.uaddr = (u32 *)uaddr; + restart->futex.val = val; + restart->futex.time = abs_time->tv64; + restart->futex.bitset = bitset; + restart->futex.flags = FLAGS_HAS_TIMEOUT; + + if (fshared) + restart->futex.flags |= FLAGS_SHARED; + if (clockrt) + restart->futex.flags |= FLAGS_CLOCKRT; + ret = -ERESTART_RESTARTBLOCK; + +out_put_key: + put_futex_key(fshared, &q.key); out: + if (to) { + hrtimer_cancel(&to->timer); + destroy_hrtimer_on_stack(&to->timer); + } return ret; } @@ -1325,13 +1830,16 @@ static long futex_wait_restart(struct restart_block *restart) { u32 __user *uaddr = (u32 __user *)restart->futex.uaddr; int fshared = 0; - ktime_t t; + ktime_t t, *tp = NULL; - t.tv64 = restart->futex.time; + if (restart->futex.flags & FLAGS_HAS_TIMEOUT) { + t.tv64 = restart->futex.time; + tp = &t; + } restart->fn = do_no_restart_syscall; if (restart->futex.flags & FLAGS_SHARED) fshared = 1; - return (long)futex_wait(uaddr, fshared, restart->futex.val, &t, + return (long)futex_wait(uaddr, fshared, restart->futex.val, tp, restart->futex.bitset, restart->futex.flags & FLAGS_CLOCKRT); } @@ -1347,11 +1855,9 @@ static int futex_lock_pi(u32 __user *uaddr, int fshared, int detect, ktime_t *time, int trylock) { struct hrtimer_sleeper timeout, *to = NULL; - struct task_struct *curr = current; struct futex_hash_bucket *hb; - u32 uval, newval, curval; struct futex_q q; - int ret, lock_taken, ownerdied = 0, attempt = 0; + int res, ret; if (refill_pi_state_cache()) return -ENOMEM; @@ -1365,117 +1871,35 @@ static int futex_lock_pi(u32 __user *uaddr, int fshared, } q.pi_state = NULL; + q.rt_waiter = NULL; + q.requeue_pi_key = NULL; retry: q.key = FUTEX_KEY_INIT; - ret = get_futex_key(uaddr, fshared, &q.key); + ret = get_futex_key(uaddr, fshared, &q.key, VERIFY_WRITE); if (unlikely(ret != 0)) goto out; -retry_unlocked: +retry_private: hb = queue_lock(&q); -retry_locked: - ret = lock_taken = 0; - - /* - * To avoid races, we attempt to take the lock here again - * (by doing a 0 -> TID atomic cmpxchg), while holding all - * the locks. It will most likely not succeed. - */ - newval = task_pid_vnr(current); - - curval = cmpxchg_futex_value_locked(uaddr, 0, newval); - - if (unlikely(curval == -EFAULT)) - goto uaddr_faulted; - - /* - * Detect deadlocks. In case of REQUEUE_PI this is a valid - * situation and we return success to user space. - */ - if (unlikely((curval & FUTEX_TID_MASK) == task_pid_vnr(current))) { - ret = -EDEADLK; - goto out_unlock_put_key; - } - - /* - * Surprise - we got the lock. Just return to userspace: - */ - if (unlikely(!curval)) - goto out_unlock_put_key; - - uval = curval; - - /* - * Set the WAITERS flag, so the owner will know it has someone - * to wake at next unlock - */ - newval = curval | FUTEX_WAITERS; - - /* - * There are two cases, where a futex might have no owner (the - * owner TID is 0): OWNER_DIED. We take over the futex in this - * case. We also do an unconditional take over, when the owner - * of the futex died. - * - * This is safe as we are protected by the hash bucket lock ! - */ - if (unlikely(ownerdied || !(curval & FUTEX_TID_MASK))) { - /* Keep the OWNER_DIED bit */ - newval = (curval & ~FUTEX_TID_MASK) | task_pid_vnr(current); - ownerdied = 0; - lock_taken = 1; - } - - curval = cmpxchg_futex_value_locked(uaddr, uval, newval); - - if (unlikely(curval == -EFAULT)) - goto uaddr_faulted; - if (unlikely(curval != uval)) - goto retry_locked; - - /* - * We took the lock due to owner died take over. - */ - if (unlikely(lock_taken)) - goto out_unlock_put_key; - - /* - * We dont have the lock. Look up the PI state (or create it if - * we are the first waiter): - */ - ret = lookup_pi_state(uval, hb, &q.key, &q.pi_state); - + ret = futex_lock_pi_atomic(uaddr, hb, &q.key, &q.pi_state, current, 0); if (unlikely(ret)) { switch (ret) { - + case 1: + /* We got the lock. */ + ret = 0; + goto out_unlock_put_key; + case -EFAULT: + goto uaddr_faulted; case -EAGAIN: /* * Task is exiting and we just wait for the * exit to complete. */ queue_unlock(&q, hb); + put_futex_key(fshared, &q.key); cond_resched(); goto retry; - - case -ESRCH: - /* - * No owner found for this futex. Check if the - * OWNER_DIED bit is set to figure out whether - * this is a robust futex or not. - */ - if (get_futex_value_locked(&curval, uaddr)) - goto uaddr_faulted; - - /* - * We simply start over in case of a robust - * futex. The code above will take the futex - * and return happy. - */ - if (curval & FUTEX_OWNER_DIED) { - ownerdied = 1; - goto retry_locked; - } default: goto out_unlock_put_key; } @@ -1499,74 +1923,29 @@ retry_locked: } spin_lock(q.lock_ptr); + /* + * Fixup the pi_state owner and possibly acquire the lock if we + * haven't already. + */ + res = fixup_owner(uaddr, fshared, &q, !ret); + /* + * If fixup_owner() returned an error, proprogate that. If it acquired + * the lock, clear our -ETIMEDOUT or -EINTR. + */ + if (res) + ret = (res < 0) ? res : 0; - if (!ret) { - /* - * Got the lock. We might not be the anticipated owner - * if we did a lock-steal - fix up the PI-state in - * that case: - */ - if (q.pi_state->owner != curr) - ret = fixup_pi_state_owner(uaddr, &q, curr, fshared); - } else { - /* - * Catch the rare case, where the lock was released - * when we were on the way back before we locked the - * hash bucket. - */ - if (q.pi_state->owner == curr) { - /* - * Try to get the rt_mutex now. This might - * fail as some other task acquired the - * rt_mutex after we removed ourself from the - * rt_mutex waiters list. - */ - if (rt_mutex_trylock(&q.pi_state->pi_mutex)) - ret = 0; - else { - /* - * pi_state is incorrect, some other - * task did a lock steal and we - * returned due to timeout or signal - * without taking the rt_mutex. Too - * late. We can access the - * rt_mutex_owner without locking, as - * the other task is now blocked on - * the hash bucket lock. Fix the state - * up. - */ - struct task_struct *owner; - int res; - - owner = rt_mutex_owner(&q.pi_state->pi_mutex); - res = fixup_pi_state_owner(uaddr, &q, owner, - fshared); - - /* propagate -EFAULT, if the fixup failed */ - if (res) - ret = res; - } - } else { - /* - * Paranoia check. If we did not take the lock - * in the trylock above, then we should not be - * the owner of the rtmutex, neither the real - * nor the pending one: - */ - if (rt_mutex_owner(&q.pi_state->pi_mutex) == curr) - printk(KERN_ERR "futex_lock_pi: ret = %d " - "pi-mutex: %p pi-state %p\n", ret, - q.pi_state->pi_mutex.owner, - q.pi_state->owner); - } - } + /* + * If fixup_owner() faulted and was unable to handle the fault, unlock + * it and return the fault to userspace. + */ + if (ret && (rt_mutex_owner(&q.pi_state->pi_mutex) == current)) + rt_mutex_unlock(&q.pi_state->pi_mutex); /* Unqueue and drop the lock */ unqueue_me_pi(&q); - if (to) - destroy_hrtimer_on_stack(&to->timer); - return ret != -EINTR ? ret : -ERESTARTNOINTR; + goto out; out_unlock_put_key: queue_unlock(&q, hb); @@ -1576,32 +1955,20 @@ out_put_key: out: if (to) destroy_hrtimer_on_stack(&to->timer); - return ret; + return ret != -EINTR ? ret : -ERESTARTNOINTR; uaddr_faulted: - /* - * We have to r/w *(int __user *)uaddr, and we have to modify it - * atomically. Therefore, if we continue to fault after get_user() - * below, we need to handle the fault ourselves, while still holding - * the mmap_sem. This can occur if the uaddr is under contention as - * we have to drop the mmap_sem in order to call get_user(). - */ queue_unlock(&q, hb); - if (attempt++) { - ret = futex_handle_fault((unsigned long)uaddr, attempt); - if (ret) - goto out_put_key; - goto retry_unlocked; - } + ret = fault_in_user_writeable(uaddr); + if (ret) + goto out_put_key; - ret = get_user(uval, uaddr); - if (!ret) - goto retry; + if (!fshared) + goto retry_private; - if (to) - destroy_hrtimer_on_stack(&to->timer); - return ret; + put_futex_key(fshared, &q.key); + goto retry; } /* @@ -1616,7 +1983,7 @@ static int futex_unlock_pi(u32 __user *uaddr, int fshared) u32 uval; struct plist_head *head; union futex_key key = FUTEX_KEY_INIT; - int ret, attempt = 0; + int ret; retry: if (get_user(uval, uaddr)) @@ -1627,12 +1994,11 @@ retry: if ((uval & FUTEX_TID_MASK) != task_pid_vnr(current)) return -EPERM; - ret = get_futex_key(uaddr, fshared, &key); + ret = get_futex_key(uaddr, fshared, &key, VERIFY_WRITE); if (unlikely(ret != 0)) goto out; hb = hash_futex(&key); -retry_unlocked: spin_lock(&hb->lock); /* @@ -1689,27 +2055,234 @@ out: return ret; pi_faulted: + spin_unlock(&hb->lock); + put_futex_key(fshared, &key); + + ret = fault_in_user_writeable(uaddr); + if (!ret) + goto retry; + + return ret; +} + +/** + * handle_early_requeue_pi_wakeup() - Detect early wakeup on the initial futex + * @hb: the hash_bucket futex_q was original enqueued on + * @q: the futex_q woken while waiting to be requeued + * @key2: the futex_key of the requeue target futex + * @timeout: the timeout associated with the wait (NULL if none) + * + * Detect if the task was woken on the initial futex as opposed to the requeue + * target futex. If so, determine if it was a timeout or a signal that caused + * the wakeup and return the appropriate error code to the caller. Must be + * called with the hb lock held. + * + * Returns + * 0 - no early wakeup detected + * <0 - -ETIMEDOUT or -ERESTARTNOINTR + */ +static inline +int handle_early_requeue_pi_wakeup(struct futex_hash_bucket *hb, + struct futex_q *q, union futex_key *key2, + struct hrtimer_sleeper *timeout) +{ + int ret = 0; + + /* + * With the hb lock held, we avoid races while we process the wakeup. + * We only need to hold hb (and not hb2) to ensure atomicity as the + * wakeup code can't change q.key from uaddr to uaddr2 if we hold hb. + * It can't be requeued from uaddr2 to something else since we don't + * support a PI aware source futex for requeue. + */ + if (!match_futex(&q->key, key2)) { + WARN_ON(q->lock_ptr && (&hb->lock != q->lock_ptr)); + /* + * We were woken prior to requeue by a timeout or a signal. + * Unqueue the futex_q and determine which it was. + */ + plist_del(&q->list, &q->list.plist); + drop_futex_key_refs(&q->key); + + if (timeout && !timeout->task) + ret = -ETIMEDOUT; + else + ret = -ERESTARTNOINTR; + } + return ret; +} + +/** + * futex_wait_requeue_pi() - Wait on uaddr and take uaddr2 + * @uaddr: the futex we initialyl wait on (non-pi) + * @fshared: whether the futexes are shared (1) or not (0). They must be + * the same type, no requeueing from private to shared, etc. + * @val: the expected value of uaddr + * @abs_time: absolute timeout + * @bitset: 32 bit wakeup bitset set by userspace, defaults to all. + * @clockrt: whether to use CLOCK_REALTIME (1) or CLOCK_MONOTONIC (0) + * @uaddr2: the pi futex we will take prior to returning to user-space + * + * The caller will wait on uaddr and will be requeued by futex_requeue() to + * uaddr2 which must be PI aware. Normal wakeup will wake on uaddr2 and + * complete the acquisition of the rt_mutex prior to returning to userspace. + * This ensures the rt_mutex maintains an owner when it has waiters; without + * one, the pi logic wouldn't know which task to boost/deboost, if there was a + * need to. + * + * We call schedule in futex_wait_queue_me() when we enqueue and return there + * via the following: + * 1) wakeup on uaddr2 after an atomic lock acquisition by futex_requeue() + * 2) wakeup on uaddr2 after a requeue + * 3) signal + * 4) timeout + * + * If 3, cleanup and return -ERESTARTNOINTR. + * + * If 2, we may then block on trying to take the rt_mutex and return via: + * 5) successful lock + * 6) signal + * 7) timeout + * 8) other lock acquisition failure + * + * If 6, return -EWOULDBLOCK (restarting the syscall would do the same). + * + * If 4 or 7, we cleanup and return with -ETIMEDOUT. + * + * Returns: + * 0 - On success + * <0 - On error + */ +static int futex_wait_requeue_pi(u32 __user *uaddr, int fshared, + u32 val, ktime_t *abs_time, u32 bitset, + int clockrt, u32 __user *uaddr2) +{ + struct hrtimer_sleeper timeout, *to = NULL; + struct rt_mutex_waiter rt_waiter; + struct rt_mutex *pi_mutex = NULL; + struct futex_hash_bucket *hb; + union futex_key key2; + struct futex_q q; + int res, ret; + + if (!bitset) + return -EINVAL; + + if (abs_time) { + to = &timeout; + hrtimer_init_on_stack(&to->timer, clockrt ? CLOCK_REALTIME : + CLOCK_MONOTONIC, HRTIMER_MODE_ABS); + hrtimer_init_sleeper(to, current); + hrtimer_set_expires_range_ns(&to->timer, *abs_time, + current->timer_slack_ns); + } + /* - * We have to r/w *(int __user *)uaddr, and we have to modify it - * atomically. Therefore, if we continue to fault after get_user() - * below, we need to handle the fault ourselves, while still holding - * the mmap_sem. This can occur if the uaddr is under contention as - * we have to drop the mmap_sem in order to call get_user(). + * The waiter is allocated on our stack, manipulated by the requeue + * code while we sleep on uaddr. */ + debug_rt_mutex_init_waiter(&rt_waiter); + rt_waiter.task = NULL; + + key2 = FUTEX_KEY_INIT; + ret = get_futex_key(uaddr2, fshared, &key2, VERIFY_WRITE); + if (unlikely(ret != 0)) + goto out; + + q.pi_state = NULL; + q.bitset = bitset; + q.rt_waiter = &rt_waiter; + q.requeue_pi_key = &key2; + + /* Prepare to wait on uaddr. */ + ret = futex_wait_setup(uaddr, val, fshared, &q, &hb); + if (ret) + goto out_key2; + + /* Queue the futex_q, drop the hb lock, wait for wakeup. */ + futex_wait_queue_me(hb, &q, to); + + spin_lock(&hb->lock); + ret = handle_early_requeue_pi_wakeup(hb, &q, &key2, to); spin_unlock(&hb->lock); + if (ret) + goto out_put_keys; - if (attempt++) { - ret = futex_handle_fault((unsigned long)uaddr, attempt); - if (ret) - goto out; - uval = 0; - goto retry_unlocked; + /* + * In order for us to be here, we know our q.key == key2, and since + * we took the hb->lock above, we also know that futex_requeue() has + * completed and we no longer have to concern ourselves with a wakeup + * race with the atomic proxy lock acquition by the requeue code. + */ + + /* Check if the requeue code acquired the second futex for us. */ + if (!q.rt_waiter) { + /* + * Got the lock. We might not be the anticipated owner if we + * did a lock-steal - fix up the PI-state in that case. + */ + if (q.pi_state && (q.pi_state->owner != current)) { + spin_lock(q.lock_ptr); + ret = fixup_pi_state_owner(uaddr2, &q, current, + fshared); + spin_unlock(q.lock_ptr); + } + } else { + /* + * We have been woken up by futex_unlock_pi(), a timeout, or a + * signal. futex_unlock_pi() will not destroy the lock_ptr nor + * the pi_state. + */ + WARN_ON(!&q.pi_state); + pi_mutex = &q.pi_state->pi_mutex; + ret = rt_mutex_finish_proxy_lock(pi_mutex, to, &rt_waiter, 1); + debug_rt_mutex_free_waiter(&rt_waiter); + + spin_lock(q.lock_ptr); + /* + * Fixup the pi_state owner and possibly acquire the lock if we + * haven't already. + */ + res = fixup_owner(uaddr2, fshared, &q, !ret); + /* + * If fixup_owner() returned an error, proprogate that. If it + * acquired the lock, clear our -ETIMEDOUT or -EINTR. + */ + if (res) + ret = (res < 0) ? res : 0; + + /* Unqueue and drop the lock. */ + unqueue_me_pi(&q); } - ret = get_user(uval, uaddr); - if (!ret) - goto retry; + /* + * If fixup_pi_state_owner() faulted and was unable to handle the + * fault, unlock the rt_mutex and return the fault to userspace. + */ + if (ret == -EFAULT) { + if (rt_mutex_owner(pi_mutex) == current) + rt_mutex_unlock(pi_mutex); + } else if (ret == -EINTR) { + /* + * We've already been requeued, but cannot restart by calling + * futex_lock_pi() directly. We could restart this syscall, but + * it would detect that the user space "val" changed and return + * -EWOULDBLOCK. Save the overhead of the restart and return + * -EWOULDBLOCK directly. + */ + ret = -EWOULDBLOCK; + } +out_put_keys: + put_futex_key(fshared, &q.key); +out_key2: + put_futex_key(fshared, &key2); + +out: + if (to) { + hrtimer_cancel(&to->timer); + destroy_hrtimer_on_stack(&to->timer); + } return ret; } @@ -1935,7 +2508,7 @@ long do_futex(u32 __user *uaddr, int op, u32 val, ktime_t *timeout, fshared = 1; clockrt = op & FUTEX_CLOCK_REALTIME; - if (clockrt && cmd != FUTEX_WAIT_BITSET) + if (clockrt && cmd != FUTEX_WAIT_BITSET && cmd != FUTEX_WAIT_REQUEUE_PI) return -ENOSYS; switch (cmd) { @@ -1950,10 +2523,11 @@ long do_futex(u32 __user *uaddr, int op, u32 val, ktime_t *timeout, ret = futex_wake(uaddr, fshared, val, val3); break; case FUTEX_REQUEUE: - ret = futex_requeue(uaddr, fshared, uaddr2, val, val2, NULL); + ret = futex_requeue(uaddr, fshared, uaddr2, val, val2, NULL, 0); break; case FUTEX_CMP_REQUEUE: - ret = futex_requeue(uaddr, fshared, uaddr2, val, val2, &val3); + ret = futex_requeue(uaddr, fshared, uaddr2, val, val2, &val3, + 0); break; case FUTEX_WAKE_OP: ret = futex_wake_op(uaddr, fshared, uaddr2, val, val2, val3); @@ -1970,6 +2544,15 @@ long do_futex(u32 __user *uaddr, int op, u32 val, ktime_t *timeout, if (futex_cmpxchg_enabled) ret = futex_lock_pi(uaddr, fshared, 0, timeout, 1); break; + case FUTEX_WAIT_REQUEUE_PI: + val3 = FUTEX_BITSET_MATCH_ANY; + ret = futex_wait_requeue_pi(uaddr, fshared, val, timeout, val3, + clockrt, uaddr2); + break; + case FUTEX_CMP_REQUEUE_PI: + ret = futex_requeue(uaddr, fshared, uaddr2, val, val2, &val3, + 1); + break; default: ret = -ENOSYS; } @@ -1987,7 +2570,8 @@ SYSCALL_DEFINE6(futex, u32 __user *, uaddr, int, op, u32, val, int cmd = op & FUTEX_CMD_MASK; if (utime && (cmd == FUTEX_WAIT || cmd == FUTEX_LOCK_PI || - cmd == FUTEX_WAIT_BITSET)) { + cmd == FUTEX_WAIT_BITSET || + cmd == FUTEX_WAIT_REQUEUE_PI)) { if (copy_from_user(&ts, utime, sizeof(ts)) != 0) return -EFAULT; if (!timespec_valid(&ts)) @@ -1999,11 +2583,11 @@ SYSCALL_DEFINE6(futex, u32 __user *, uaddr, int, op, u32, val, tp = &t; } /* - * requeue parameter in 'utime' if cmd == FUTEX_REQUEUE. + * requeue parameter in 'utime' if cmd == FUTEX_*_REQUEUE_*. * number of waiters to wake in 'utime' if cmd == FUTEX_WAKE_OP. */ if (cmd == FUTEX_REQUEUE || cmd == FUTEX_CMP_REQUEUE || - cmd == FUTEX_WAKE_OP) + cmd == FUTEX_CMP_REQUEUE_PI || cmd == FUTEX_WAKE_OP) val2 = (u32) (unsigned long) utime; return do_futex(uaddr, op, val, tp, uaddr2, val2, val3); diff --git a/kernel/futex_compat.c b/kernel/futex_compat.c index d607a5b9ee2..235716556bf 100644 --- a/kernel/futex_compat.c +++ b/kernel/futex_compat.c @@ -180,7 +180,8 @@ asmlinkage long compat_sys_futex(u32 __user *uaddr, int op, u32 val, int cmd = op & FUTEX_CMD_MASK; if (utime && (cmd == FUTEX_WAIT || cmd == FUTEX_LOCK_PI || - cmd == FUTEX_WAIT_BITSET)) { + cmd == FUTEX_WAIT_BITSET || + cmd == FUTEX_WAIT_REQUEUE_PI)) { if (get_compat_timespec(&ts, utime)) return -EFAULT; if (!timespec_valid(&ts)) @@ -191,7 +192,8 @@ asmlinkage long compat_sys_futex(u32 __user *uaddr, int op, u32 val, t = ktime_add_safe(ktime_get(), t); tp = &t; } - if (cmd == FUTEX_REQUEUE || cmd == FUTEX_CMP_REQUEUE) + if (cmd == FUTEX_REQUEUE || cmd == FUTEX_CMP_REQUEUE || + cmd == FUTEX_CMP_REQUEUE_PI || cmd == FUTEX_WAKE_OP) val2 = (int) (unsigned long) utime; return do_futex(uaddr, op, val, tp, uaddr2, val2, val3); diff --git a/kernel/gcov/Kconfig b/kernel/gcov/Kconfig new file mode 100644 index 00000000000..22e9dcfaa3d --- /dev/null +++ b/kernel/gcov/Kconfig @@ -0,0 +1,48 @@ +menu "GCOV-based kernel profiling" + +config GCOV_KERNEL + bool "Enable gcov-based kernel profiling" + depends on DEBUG_FS && CONSTRUCTORS + default n + ---help--- + This option enables gcov-based code profiling (e.g. for code coverage + measurements). + + If unsure, say N. + + Additionally specify CONFIG_GCOV_PROFILE_ALL=y to get profiling data + for the entire kernel. To enable profiling for specific files or + directories, add a line similar to the following to the respective + Makefile: + + For a single file (e.g. main.o): + GCOV_PROFILE_main.o := y + + For all files in one directory: + GCOV_PROFILE := y + + To exclude files from being profiled even when CONFIG_GCOV_PROFILE_ALL + is specified, use: + + GCOV_PROFILE_main.o := n + and: + GCOV_PROFILE := n + + Note that the debugfs filesystem has to be mounted to access + profiling data. + +config GCOV_PROFILE_ALL + bool "Profile entire Kernel" + depends on GCOV_KERNEL + depends on S390 || X86 + default n + ---help--- + This options activates profiling for the entire kernel. + + If unsure, say N. + + Note that a kernel compiled with profiling flags will be significantly + larger and run slower. Also be sure to exclude files from profiling + which are not linked to the kernel image to prevent linker errors. + +endmenu diff --git a/kernel/gcov/Makefile b/kernel/gcov/Makefile new file mode 100644 index 00000000000..3f761001d51 --- /dev/null +++ b/kernel/gcov/Makefile @@ -0,0 +1,3 @@ +EXTRA_CFLAGS := -DSRCTREE='"$(srctree)"' -DOBJTREE='"$(objtree)"' + +obj-$(CONFIG_GCOV_KERNEL) := base.o fs.o gcc_3_4.o diff --git a/kernel/gcov/base.c b/kernel/gcov/base.c new file mode 100644 index 00000000000..9b22d03cc58 --- /dev/null +++ b/kernel/gcov/base.c @@ -0,0 +1,148 @@ +/* + * This code maintains a list of active profiling data structures. + * + * Copyright IBM Corp. 2009 + * Author(s): Peter Oberparleiter <oberpar@linux.vnet.ibm.com> + * + * Uses gcc-internal data definitions. + * Based on the gcov-kernel patch by: + * Hubertus Franke <frankeh@us.ibm.com> + * Nigel Hinds <nhinds@us.ibm.com> + * Rajan Ravindran <rajancr@us.ibm.com> + * Peter Oberparleiter <oberpar@linux.vnet.ibm.com> + * Paul Larson + */ + +#define pr_fmt(fmt) "gcov: " fmt + +#include <linux/init.h> +#include <linux/module.h> +#include <linux/mutex.h> +#include "gcov.h" + +static struct gcov_info *gcov_info_head; +static int gcov_events_enabled; +static DEFINE_MUTEX(gcov_lock); + +/* + * __gcov_init is called by gcc-generated constructor code for each object + * file compiled with -fprofile-arcs. + */ +void __gcov_init(struct gcov_info *info) +{ + static unsigned int gcov_version; + + mutex_lock(&gcov_lock); + if (gcov_version == 0) { + gcov_version = info->version; + /* + * Printing gcc's version magic may prove useful for debugging + * incompatibility reports. + */ + pr_info("version magic: 0x%x\n", gcov_version); + } + /* + * Add new profiling data structure to list and inform event + * listener. + */ + info->next = gcov_info_head; + gcov_info_head = info; + if (gcov_events_enabled) + gcov_event(GCOV_ADD, info); + mutex_unlock(&gcov_lock); +} +EXPORT_SYMBOL(__gcov_init); + +/* + * These functions may be referenced by gcc-generated profiling code but serve + * no function for kernel profiling. + */ +void __gcov_flush(void) +{ + /* Unused. */ +} +EXPORT_SYMBOL(__gcov_flush); + +void __gcov_merge_add(gcov_type *counters, unsigned int n_counters) +{ + /* Unused. */ +} +EXPORT_SYMBOL(__gcov_merge_add); + +void __gcov_merge_single(gcov_type *counters, unsigned int n_counters) +{ + /* Unused. */ +} +EXPORT_SYMBOL(__gcov_merge_single); + +void __gcov_merge_delta(gcov_type *counters, unsigned int n_counters) +{ + /* Unused. */ +} +EXPORT_SYMBOL(__gcov_merge_delta); + +/** + * gcov_enable_events - enable event reporting through gcov_event() + * + * Turn on reporting of profiling data load/unload-events through the + * gcov_event() callback. Also replay all previous events once. This function + * is needed because some events are potentially generated too early for the + * callback implementation to handle them initially. + */ +void gcov_enable_events(void) +{ + struct gcov_info *info; + + mutex_lock(&gcov_lock); + gcov_events_enabled = 1; + /* Perform event callback for previously registered entries. */ + for (info = gcov_info_head; info; info = info->next) + gcov_event(GCOV_ADD, info); + mutex_unlock(&gcov_lock); +} + +#ifdef CONFIG_MODULES +static inline int within(void *addr, void *start, unsigned long size) +{ + return ((addr >= start) && (addr < start + size)); +} + +/* Update list and generate events when modules are unloaded. */ +static int gcov_module_notifier(struct notifier_block *nb, unsigned long event, + void *data) +{ + struct module *mod = data; + struct gcov_info *info; + struct gcov_info *prev; + + if (event != MODULE_STATE_GOING) + return NOTIFY_OK; + mutex_lock(&gcov_lock); + prev = NULL; + /* Remove entries located in module from linked list. */ + for (info = gcov_info_head; info; info = info->next) { + if (within(info, mod->module_core, mod->core_size)) { + if (prev) + prev->next = info->next; + else + gcov_info_head = info->next; + if (gcov_events_enabled) + gcov_event(GCOV_REMOVE, info); + } else + prev = info; + } + mutex_unlock(&gcov_lock); + + return NOTIFY_OK; +} + +static struct notifier_block gcov_nb = { + .notifier_call = gcov_module_notifier, +}; + +static int __init gcov_init(void) +{ + return register_module_notifier(&gcov_nb); +} +device_initcall(gcov_init); +#endif /* CONFIG_MODULES */ diff --git a/kernel/gcov/fs.c b/kernel/gcov/fs.c new file mode 100644 index 00000000000..ef3c3f88a7a --- /dev/null +++ b/kernel/gcov/fs.c @@ -0,0 +1,673 @@ +/* + * This code exports profiling data as debugfs files to userspace. + * + * Copyright IBM Corp. 2009 + * Author(s): Peter Oberparleiter <oberpar@linux.vnet.ibm.com> + * + * Uses gcc-internal data definitions. + * Based on the gcov-kernel patch by: + * Hubertus Franke <frankeh@us.ibm.com> + * Nigel Hinds <nhinds@us.ibm.com> + * Rajan Ravindran <rajancr@us.ibm.com> + * Peter Oberparleiter <oberpar@linux.vnet.ibm.com> + * Paul Larson + * Yi CDL Yang + */ + +#define pr_fmt(fmt) "gcov: " fmt + +#include <linux/init.h> +#include <linux/module.h> +#include <linux/debugfs.h> +#include <linux/fs.h> +#include <linux/list.h> +#include <linux/string.h> +#include <linux/slab.h> +#include <linux/mutex.h> +#include <linux/seq_file.h> +#include "gcov.h" + +/** + * struct gcov_node - represents a debugfs entry + * @list: list head for child node list + * @children: child nodes + * @all: list head for list of all nodes + * @parent: parent node + * @info: associated profiling data structure if not a directory + * @ghost: when an object file containing profiling data is unloaded we keep a + * copy of the profiling data here to allow collecting coverage data + * for cleanup code. Such a node is called a "ghost". + * @dentry: main debugfs entry, either a directory or data file + * @links: associated symbolic links + * @name: data file basename + * + * struct gcov_node represents an entity within the gcov/ subdirectory + * of debugfs. There are directory and data file nodes. The latter represent + * the actual synthesized data file plus any associated symbolic links which + * are needed by the gcov tool to work correctly. + */ +struct gcov_node { + struct list_head list; + struct list_head children; + struct list_head all; + struct gcov_node *parent; + struct gcov_info *info; + struct gcov_info *ghost; + struct dentry *dentry; + struct dentry **links; + char name[0]; +}; + +static const char objtree[] = OBJTREE; +static const char srctree[] = SRCTREE; +static struct gcov_node root_node; +static struct dentry *reset_dentry; +static LIST_HEAD(all_head); +static DEFINE_MUTEX(node_lock); + +/* If non-zero, keep copies of profiling data for unloaded modules. */ +static int gcov_persist = 1; + +static int __init gcov_persist_setup(char *str) +{ + unsigned long val; + + if (strict_strtoul(str, 0, &val)) { + pr_warning("invalid gcov_persist parameter '%s'\n", str); + return 0; + } + gcov_persist = val; + pr_info("setting gcov_persist to %d\n", gcov_persist); + + return 1; +} +__setup("gcov_persist=", gcov_persist_setup); + +/* + * seq_file.start() implementation for gcov data files. Note that the + * gcov_iterator interface is designed to be more restrictive than seq_file + * (no start from arbitrary position, etc.), to simplify the iterator + * implementation. + */ +static void *gcov_seq_start(struct seq_file *seq, loff_t *pos) +{ + loff_t i; + + gcov_iter_start(seq->private); + for (i = 0; i < *pos; i++) { + if (gcov_iter_next(seq->private)) + return NULL; + } + return seq->private; +} + +/* seq_file.next() implementation for gcov data files. */ +static void *gcov_seq_next(struct seq_file *seq, void *data, loff_t *pos) +{ + struct gcov_iterator *iter = data; + + if (gcov_iter_next(iter)) + return NULL; + (*pos)++; + + return iter; +} + +/* seq_file.show() implementation for gcov data files. */ +static int gcov_seq_show(struct seq_file *seq, void *data) +{ + struct gcov_iterator *iter = data; + + if (gcov_iter_write(iter, seq)) + return -EINVAL; + return 0; +} + +static void gcov_seq_stop(struct seq_file *seq, void *data) +{ + /* Unused. */ +} + +static const struct seq_operations gcov_seq_ops = { + .start = gcov_seq_start, + .next = gcov_seq_next, + .show = gcov_seq_show, + .stop = gcov_seq_stop, +}; + +/* + * Return the profiling data set for a given node. This can either be the + * original profiling data structure or a duplicate (also called "ghost") + * in case the associated object file has been unloaded. + */ +static struct gcov_info *get_node_info(struct gcov_node *node) +{ + if (node->info) + return node->info; + + return node->ghost; +} + +/* + * open() implementation for gcov data files. Create a copy of the profiling + * data set and initialize the iterator and seq_file interface. + */ +static int gcov_seq_open(struct inode *inode, struct file *file) +{ + struct gcov_node *node = inode->i_private; + struct gcov_iterator *iter; + struct seq_file *seq; + struct gcov_info *info; + int rc = -ENOMEM; + + mutex_lock(&node_lock); + /* + * Read from a profiling data copy to minimize reference tracking + * complexity and concurrent access. + */ + info = gcov_info_dup(get_node_info(node)); + if (!info) + goto out_unlock; + iter = gcov_iter_new(info); + if (!iter) + goto err_free_info; + rc = seq_open(file, &gcov_seq_ops); + if (rc) + goto err_free_iter_info; + seq = file->private_data; + seq->private = iter; +out_unlock: + mutex_unlock(&node_lock); + return rc; + +err_free_iter_info: + gcov_iter_free(iter); +err_free_info: + gcov_info_free(info); + goto out_unlock; +} + +/* + * release() implementation for gcov data files. Release resources allocated + * by open(). + */ +static int gcov_seq_release(struct inode *inode, struct file *file) +{ + struct gcov_iterator *iter; + struct gcov_info *info; + struct seq_file *seq; + + seq = file->private_data; + iter = seq->private; + info = gcov_iter_get_info(iter); + gcov_iter_free(iter); + gcov_info_free(info); + seq_release(inode, file); + + return 0; +} + +/* + * Find a node by the associated data file name. Needs to be called with + * node_lock held. + */ +static struct gcov_node *get_node_by_name(const char *name) +{ + struct gcov_node *node; + struct gcov_info *info; + + list_for_each_entry(node, &all_head, all) { + info = get_node_info(node); + if (info && (strcmp(info->filename, name) == 0)) + return node; + } + + return NULL; +} + +static void remove_node(struct gcov_node *node); + +/* + * write() implementation for gcov data files. Reset profiling data for the + * associated file. If the object file has been unloaded (i.e. this is + * a "ghost" node), remove the debug fs node as well. + */ +static ssize_t gcov_seq_write(struct file *file, const char __user *addr, + size_t len, loff_t *pos) +{ + struct seq_file *seq; + struct gcov_info *info; + struct gcov_node *node; + + seq = file->private_data; + info = gcov_iter_get_info(seq->private); + mutex_lock(&node_lock); + node = get_node_by_name(info->filename); + if (node) { + /* Reset counts or remove node for unloaded modules. */ + if (node->ghost) + remove_node(node); + else + gcov_info_reset(node->info); + } + /* Reset counts for open file. */ + gcov_info_reset(info); + mutex_unlock(&node_lock); + + return len; +} + +/* + * Given a string <path> representing a file path of format: + * path/to/file.gcda + * construct and return a new string: + * <dir/>path/to/file.<ext> + */ +static char *link_target(const char *dir, const char *path, const char *ext) +{ + char *target; + char *old_ext; + char *copy; + + copy = kstrdup(path, GFP_KERNEL); + if (!copy) + return NULL; + old_ext = strrchr(copy, '.'); + if (old_ext) + *old_ext = '\0'; + if (dir) + target = kasprintf(GFP_KERNEL, "%s/%s.%s", dir, copy, ext); + else + target = kasprintf(GFP_KERNEL, "%s.%s", copy, ext); + kfree(copy); + + return target; +} + +/* + * Construct a string representing the symbolic link target for the given + * gcov data file name and link type. Depending on the link type and the + * location of the data file, the link target can either point to a + * subdirectory of srctree, objtree or in an external location. + */ +static char *get_link_target(const char *filename, const struct gcov_link *ext) +{ + const char *rel; + char *result; + + if (strncmp(filename, objtree, strlen(objtree)) == 0) { + rel = filename + strlen(objtree) + 1; + if (ext->dir == SRC_TREE) + result = link_target(srctree, rel, ext->ext); + else + result = link_target(objtree, rel, ext->ext); + } else { + /* External compilation. */ + result = link_target(NULL, filename, ext->ext); + } + + return result; +} + +#define SKEW_PREFIX ".tmp_" + +/* + * For a filename .tmp_filename.ext return filename.ext. Needed to compensate + * for filename skewing caused by the mod-versioning mechanism. + */ +static const char *deskew(const char *basename) +{ + if (strncmp(basename, SKEW_PREFIX, sizeof(SKEW_PREFIX) - 1) == 0) + return basename + sizeof(SKEW_PREFIX) - 1; + return basename; +} + +/* + * Create links to additional files (usually .c and .gcno files) which the + * gcov tool expects to find in the same directory as the gcov data file. + */ +static void add_links(struct gcov_node *node, struct dentry *parent) +{ + char *basename; + char *target; + int num; + int i; + + for (num = 0; gcov_link[num].ext; num++) + /* Nothing. */; + node->links = kcalloc(num, sizeof(struct dentry *), GFP_KERNEL); + if (!node->links) + return; + for (i = 0; i < num; i++) { + target = get_link_target(get_node_info(node)->filename, + &gcov_link[i]); + if (!target) + goto out_err; + basename = strrchr(target, '/'); + if (!basename) + goto out_err; + basename++; + node->links[i] = debugfs_create_symlink(deskew(basename), + parent, target); + if (!node->links[i]) + goto out_err; + kfree(target); + } + + return; +out_err: + kfree(target); + while (i-- > 0) + debugfs_remove(node->links[i]); + kfree(node->links); + node->links = NULL; +} + +static const struct file_operations gcov_data_fops = { + .open = gcov_seq_open, + .release = gcov_seq_release, + .read = seq_read, + .llseek = seq_lseek, + .write = gcov_seq_write, +}; + +/* Basic initialization of a new node. */ +static void init_node(struct gcov_node *node, struct gcov_info *info, + const char *name, struct gcov_node *parent) +{ + INIT_LIST_HEAD(&node->list); + INIT_LIST_HEAD(&node->children); + INIT_LIST_HEAD(&node->all); + node->info = info; + node->parent = parent; + if (name) + strcpy(node->name, name); +} + +/* + * Create a new node and associated debugfs entry. Needs to be called with + * node_lock held. + */ +static struct gcov_node *new_node(struct gcov_node *parent, + struct gcov_info *info, const char *name) +{ + struct gcov_node *node; + + node = kzalloc(sizeof(struct gcov_node) + strlen(name) + 1, GFP_KERNEL); + if (!node) { + pr_warning("out of memory\n"); + return NULL; + } + init_node(node, info, name, parent); + /* Differentiate between gcov data file nodes and directory nodes. */ + if (info) { + node->dentry = debugfs_create_file(deskew(node->name), 0600, + parent->dentry, node, &gcov_data_fops); + } else + node->dentry = debugfs_create_dir(node->name, parent->dentry); + if (!node->dentry) { + pr_warning("could not create file\n"); + kfree(node); + return NULL; + } + if (info) + add_links(node, parent->dentry); + list_add(&node->list, &parent->children); + list_add(&node->all, &all_head); + + return node; +} + +/* Remove symbolic links associated with node. */ +static void remove_links(struct gcov_node *node) +{ + int i; + + if (!node->links) + return; + for (i = 0; gcov_link[i].ext; i++) + debugfs_remove(node->links[i]); + kfree(node->links); + node->links = NULL; +} + +/* + * Remove node from all lists and debugfs and release associated resources. + * Needs to be called with node_lock held. + */ +static void release_node(struct gcov_node *node) +{ + list_del(&node->list); + list_del(&node->all); + debugfs_remove(node->dentry); + remove_links(node); + if (node->ghost) + gcov_info_free(node->ghost); + kfree(node); +} + +/* Release node and empty parents. Needs to be called with node_lock held. */ +static void remove_node(struct gcov_node *node) +{ + struct gcov_node *parent; + + while ((node != &root_node) && list_empty(&node->children)) { + parent = node->parent; + release_node(node); + node = parent; + } +} + +/* + * Find child node with given basename. Needs to be called with node_lock + * held. + */ +static struct gcov_node *get_child_by_name(struct gcov_node *parent, + const char *name) +{ + struct gcov_node *node; + + list_for_each_entry(node, &parent->children, list) { + if (strcmp(node->name, name) == 0) + return node; + } + + return NULL; +} + +/* + * write() implementation for reset file. Reset all profiling data to zero + * and remove ghost nodes. + */ +static ssize_t reset_write(struct file *file, const char __user *addr, + size_t len, loff_t *pos) +{ + struct gcov_node *node; + + mutex_lock(&node_lock); +restart: + list_for_each_entry(node, &all_head, all) { + if (node->info) + gcov_info_reset(node->info); + else if (list_empty(&node->children)) { + remove_node(node); + /* Several nodes may have gone - restart loop. */ + goto restart; + } + } + mutex_unlock(&node_lock); + + return len; +} + +/* read() implementation for reset file. Unused. */ +static ssize_t reset_read(struct file *file, char __user *addr, size_t len, + loff_t *pos) +{ + /* Allow read operation so that a recursive copy won't fail. */ + return 0; +} + +static const struct file_operations gcov_reset_fops = { + .write = reset_write, + .read = reset_read, +}; + +/* + * Create a node for a given profiling data set and add it to all lists and + * debugfs. Needs to be called with node_lock held. + */ +static void add_node(struct gcov_info *info) +{ + char *filename; + char *curr; + char *next; + struct gcov_node *parent; + struct gcov_node *node; + + filename = kstrdup(info->filename, GFP_KERNEL); + if (!filename) + return; + parent = &root_node; + /* Create directory nodes along the path. */ + for (curr = filename; (next = strchr(curr, '/')); curr = next + 1) { + if (curr == next) + continue; + *next = 0; + if (strcmp(curr, ".") == 0) + continue; + if (strcmp(curr, "..") == 0) { + if (!parent->parent) + goto err_remove; + parent = parent->parent; + continue; + } + node = get_child_by_name(parent, curr); + if (!node) { + node = new_node(parent, NULL, curr); + if (!node) + goto err_remove; + } + parent = node; + } + /* Create file node. */ + node = new_node(parent, info, curr); + if (!node) + goto err_remove; +out: + kfree(filename); + return; + +err_remove: + remove_node(parent); + goto out; +} + +/* + * The profiling data set associated with this node is being unloaded. Store a + * copy of the profiling data and turn this node into a "ghost". + */ +static int ghost_node(struct gcov_node *node) +{ + node->ghost = gcov_info_dup(node->info); + if (!node->ghost) { + pr_warning("could not save data for '%s' (out of memory)\n", + node->info->filename); + return -ENOMEM; + } + node->info = NULL; + + return 0; +} + +/* + * Profiling data for this node has been loaded again. Add profiling data + * from previous instantiation and turn this node into a regular node. + */ +static void revive_node(struct gcov_node *node, struct gcov_info *info) +{ + if (gcov_info_is_compatible(node->ghost, info)) + gcov_info_add(info, node->ghost); + else { + pr_warning("discarding saved data for '%s' (version changed)\n", + info->filename); + } + gcov_info_free(node->ghost); + node->ghost = NULL; + node->info = info; +} + +/* + * Callback to create/remove profiling files when code compiled with + * -fprofile-arcs is loaded/unloaded. + */ +void gcov_event(enum gcov_action action, struct gcov_info *info) +{ + struct gcov_node *node; + + mutex_lock(&node_lock); + node = get_node_by_name(info->filename); + switch (action) { + case GCOV_ADD: + /* Add new node or revive ghost. */ + if (!node) { + add_node(info); + break; + } + if (gcov_persist) + revive_node(node, info); + else { + pr_warning("could not add '%s' (already exists)\n", + info->filename); + } + break; + case GCOV_REMOVE: + /* Remove node or turn into ghost. */ + if (!node) { + pr_warning("could not remove '%s' (not found)\n", + info->filename); + break; + } + if (gcov_persist) { + if (!ghost_node(node)) + break; + } + remove_node(node); + break; + } + mutex_unlock(&node_lock); +} + +/* Create debugfs entries. */ +static __init int gcov_fs_init(void) +{ + int rc = -EIO; + + init_node(&root_node, NULL, NULL, NULL); + /* + * /sys/kernel/debug/gcov will be parent for the reset control file + * and all profiling files. + */ + root_node.dentry = debugfs_create_dir("gcov", NULL); + if (!root_node.dentry) + goto err_remove; + /* + * Create reset file which resets all profiling counts when written + * to. + */ + reset_dentry = debugfs_create_file("reset", 0600, root_node.dentry, + NULL, &gcov_reset_fops); + if (!reset_dentry) + goto err_remove; + /* Replay previous events to get our fs hierarchy up-to-date. */ + gcov_enable_events(); + return 0; + +err_remove: + pr_err("init failed\n"); + if (root_node.dentry) + debugfs_remove(root_node.dentry); + + return rc; +} +device_initcall(gcov_fs_init); diff --git a/kernel/gcov/gcc_3_4.c b/kernel/gcov/gcc_3_4.c new file mode 100644 index 00000000000..ae5bb426003 --- /dev/null +++ b/kernel/gcov/gcc_3_4.c @@ -0,0 +1,447 @@ +/* + * This code provides functions to handle gcc's profiling data format + * introduced with gcc 3.4. Future versions of gcc may change the gcov + * format (as happened before), so all format-specific information needs + * to be kept modular and easily exchangeable. + * + * This file is based on gcc-internal definitions. Functions and data + * structures are defined to be compatible with gcc counterparts. + * For a better understanding, refer to gcc source: gcc/gcov-io.h. + * + * Copyright IBM Corp. 2009 + * Author(s): Peter Oberparleiter <oberpar@linux.vnet.ibm.com> + * + * Uses gcc-internal data definitions. + */ + +#include <linux/errno.h> +#include <linux/slab.h> +#include <linux/string.h> +#include <linux/seq_file.h> +#include <linux/vmalloc.h> +#include "gcov.h" + +/* Symbolic links to be created for each profiling data file. */ +const struct gcov_link gcov_link[] = { + { OBJ_TREE, "gcno" }, /* Link to .gcno file in $(objtree). */ + { 0, NULL}, +}; + +/* + * Determine whether a counter is active. Based on gcc magic. Doesn't change + * at run-time. + */ +static int counter_active(struct gcov_info *info, unsigned int type) +{ + return (1 << type) & info->ctr_mask; +} + +/* Determine number of active counters. Based on gcc magic. */ +static unsigned int num_counter_active(struct gcov_info *info) +{ + unsigned int i; + unsigned int result = 0; + + for (i = 0; i < GCOV_COUNTERS; i++) { + if (counter_active(info, i)) + result++; + } + return result; +} + +/** + * gcov_info_reset - reset profiling data to zero + * @info: profiling data set + */ +void gcov_info_reset(struct gcov_info *info) +{ + unsigned int active = num_counter_active(info); + unsigned int i; + + for (i = 0; i < active; i++) { + memset(info->counts[i].values, 0, + info->counts[i].num * sizeof(gcov_type)); + } +} + +/** + * gcov_info_is_compatible - check if profiling data can be added + * @info1: first profiling data set + * @info2: second profiling data set + * + * Returns non-zero if profiling data can be added, zero otherwise. + */ +int gcov_info_is_compatible(struct gcov_info *info1, struct gcov_info *info2) +{ + return (info1->stamp == info2->stamp); +} + +/** + * gcov_info_add - add up profiling data + * @dest: profiling data set to which data is added + * @source: profiling data set which is added + * + * Adds profiling counts of @source to @dest. + */ +void gcov_info_add(struct gcov_info *dest, struct gcov_info *source) +{ + unsigned int i; + unsigned int j; + + for (i = 0; i < num_counter_active(dest); i++) { + for (j = 0; j < dest->counts[i].num; j++) { + dest->counts[i].values[j] += + source->counts[i].values[j]; + } + } +} + +/* Get size of function info entry. Based on gcc magic. */ +static size_t get_fn_size(struct gcov_info *info) +{ + size_t size; + + size = sizeof(struct gcov_fn_info) + num_counter_active(info) * + sizeof(unsigned int); + if (__alignof__(struct gcov_fn_info) > sizeof(unsigned int)) + size = ALIGN(size, __alignof__(struct gcov_fn_info)); + return size; +} + +/* Get address of function info entry. Based on gcc magic. */ +static struct gcov_fn_info *get_fn_info(struct gcov_info *info, unsigned int fn) +{ + return (struct gcov_fn_info *) + ((char *) info->functions + fn * get_fn_size(info)); +} + +/** + * gcov_info_dup - duplicate profiling data set + * @info: profiling data set to duplicate + * + * Return newly allocated duplicate on success, %NULL on error. + */ +struct gcov_info *gcov_info_dup(struct gcov_info *info) +{ + struct gcov_info *dup; + unsigned int i; + unsigned int active; + + /* Duplicate gcov_info. */ + active = num_counter_active(info); + dup = kzalloc(sizeof(struct gcov_info) + + sizeof(struct gcov_ctr_info) * active, GFP_KERNEL); + if (!dup) + return NULL; + dup->version = info->version; + dup->stamp = info->stamp; + dup->n_functions = info->n_functions; + dup->ctr_mask = info->ctr_mask; + /* Duplicate filename. */ + dup->filename = kstrdup(info->filename, GFP_KERNEL); + if (!dup->filename) + goto err_free; + /* Duplicate table of functions. */ + dup->functions = kmemdup(info->functions, info->n_functions * + get_fn_size(info), GFP_KERNEL); + if (!dup->functions) + goto err_free; + /* Duplicate counter arrays. */ + for (i = 0; i < active ; i++) { + struct gcov_ctr_info *ctr = &info->counts[i]; + size_t size = ctr->num * sizeof(gcov_type); + + dup->counts[i].num = ctr->num; + dup->counts[i].merge = ctr->merge; + dup->counts[i].values = vmalloc(size); + if (!dup->counts[i].values) + goto err_free; + memcpy(dup->counts[i].values, ctr->values, size); + } + return dup; + +err_free: + gcov_info_free(dup); + return NULL; +} + +/** + * gcov_info_free - release memory for profiling data set duplicate + * @info: profiling data set duplicate to free + */ +void gcov_info_free(struct gcov_info *info) +{ + unsigned int active = num_counter_active(info); + unsigned int i; + + for (i = 0; i < active ; i++) + vfree(info->counts[i].values); + kfree(info->functions); + kfree(info->filename); + kfree(info); +} + +/** + * struct type_info - iterator helper array + * @ctr_type: counter type + * @offset: index of the first value of the current function for this type + * + * This array is needed to convert the in-memory data format into the in-file + * data format: + * + * In-memory: + * for each counter type + * for each function + * values + * + * In-file: + * for each function + * for each counter type + * values + * + * See gcc source gcc/gcov-io.h for more information on data organization. + */ +struct type_info { + int ctr_type; + unsigned int offset; +}; + +/** + * struct gcov_iterator - specifies current file position in logical records + * @info: associated profiling data + * @record: record type + * @function: function number + * @type: counter type + * @count: index into values array + * @num_types: number of counter types + * @type_info: helper array to get values-array offset for current function + */ +struct gcov_iterator { + struct gcov_info *info; + + int record; + unsigned int function; + unsigned int type; + unsigned int count; + + int num_types; + struct type_info type_info[0]; +}; + +static struct gcov_fn_info *get_func(struct gcov_iterator *iter) +{ + return get_fn_info(iter->info, iter->function); +} + +static struct type_info *get_type(struct gcov_iterator *iter) +{ + return &iter->type_info[iter->type]; +} + +/** + * gcov_iter_new - allocate and initialize profiling data iterator + * @info: profiling data set to be iterated + * + * Return file iterator on success, %NULL otherwise. + */ +struct gcov_iterator *gcov_iter_new(struct gcov_info *info) +{ + struct gcov_iterator *iter; + + iter = kzalloc(sizeof(struct gcov_iterator) + + num_counter_active(info) * sizeof(struct type_info), + GFP_KERNEL); + if (iter) + iter->info = info; + + return iter; +} + +/** + * gcov_iter_free - release memory for iterator + * @iter: file iterator to free + */ +void gcov_iter_free(struct gcov_iterator *iter) +{ + kfree(iter); +} + +/** + * gcov_iter_get_info - return profiling data set for given file iterator + * @iter: file iterator + */ +struct gcov_info *gcov_iter_get_info(struct gcov_iterator *iter) +{ + return iter->info; +} + +/** + * gcov_iter_start - reset file iterator to starting position + * @iter: file iterator + */ +void gcov_iter_start(struct gcov_iterator *iter) +{ + int i; + + iter->record = 0; + iter->function = 0; + iter->type = 0; + iter->count = 0; + iter->num_types = 0; + for (i = 0; i < GCOV_COUNTERS; i++) { + if (counter_active(iter->info, i)) { + iter->type_info[iter->num_types].ctr_type = i; + iter->type_info[iter->num_types++].offset = 0; + } + } +} + +/* Mapping of logical record number to actual file content. */ +#define RECORD_FILE_MAGIC 0 +#define RECORD_GCOV_VERSION 1 +#define RECORD_TIME_STAMP 2 +#define RECORD_FUNCTION_TAG 3 +#define RECORD_FUNCTON_TAG_LEN 4 +#define RECORD_FUNCTION_IDENT 5 +#define RECORD_FUNCTION_CHECK 6 +#define RECORD_COUNT_TAG 7 +#define RECORD_COUNT_LEN 8 +#define RECORD_COUNT 9 + +/** + * gcov_iter_next - advance file iterator to next logical record + * @iter: file iterator + * + * Return zero if new position is valid, non-zero if iterator has reached end. + */ +int gcov_iter_next(struct gcov_iterator *iter) +{ + switch (iter->record) { + case RECORD_FILE_MAGIC: + case RECORD_GCOV_VERSION: + case RECORD_FUNCTION_TAG: + case RECORD_FUNCTON_TAG_LEN: + case RECORD_FUNCTION_IDENT: + case RECORD_COUNT_TAG: + /* Advance to next record */ + iter->record++; + break; + case RECORD_COUNT: + /* Advance to next count */ + iter->count++; + /* fall through */ + case RECORD_COUNT_LEN: + if (iter->count < get_func(iter)->n_ctrs[iter->type]) { + iter->record = 9; + break; + } + /* Advance to next counter type */ + get_type(iter)->offset += iter->count; + iter->count = 0; + iter->type++; + /* fall through */ + case RECORD_FUNCTION_CHECK: + if (iter->type < iter->num_types) { + iter->record = 7; + break; + } + /* Advance to next function */ + iter->type = 0; + iter->function++; + /* fall through */ + case RECORD_TIME_STAMP: + if (iter->function < iter->info->n_functions) + iter->record = 3; + else + iter->record = -1; + break; + } + /* Check for EOF. */ + if (iter->record == -1) + return -EINVAL; + else + return 0; +} + +/** + * seq_write_gcov_u32 - write 32 bit number in gcov format to seq_file + * @seq: seq_file handle + * @v: value to be stored + * + * Number format defined by gcc: numbers are recorded in the 32 bit + * unsigned binary form of the endianness of the machine generating the + * file. + */ +static int seq_write_gcov_u32(struct seq_file *seq, u32 v) +{ + return seq_write(seq, &v, sizeof(v)); +} + +/** + * seq_write_gcov_u64 - write 64 bit number in gcov format to seq_file + * @seq: seq_file handle + * @v: value to be stored + * + * Number format defined by gcc: numbers are recorded in the 32 bit + * unsigned binary form of the endianness of the machine generating the + * file. 64 bit numbers are stored as two 32 bit numbers, the low part + * first. + */ +static int seq_write_gcov_u64(struct seq_file *seq, u64 v) +{ + u32 data[2]; + + data[0] = (v & 0xffffffffUL); + data[1] = (v >> 32); + return seq_write(seq, data, sizeof(data)); +} + +/** + * gcov_iter_write - write data for current pos to seq_file + * @iter: file iterator + * @seq: seq_file handle + * + * Return zero on success, non-zero otherwise. + */ +int gcov_iter_write(struct gcov_iterator *iter, struct seq_file *seq) +{ + int rc = -EINVAL; + + switch (iter->record) { + case RECORD_FILE_MAGIC: + rc = seq_write_gcov_u32(seq, GCOV_DATA_MAGIC); + break; + case RECORD_GCOV_VERSION: + rc = seq_write_gcov_u32(seq, iter->info->version); + break; + case RECORD_TIME_STAMP: + rc = seq_write_gcov_u32(seq, iter->info->stamp); + break; + case RECORD_FUNCTION_TAG: + rc = seq_write_gcov_u32(seq, GCOV_TAG_FUNCTION); + break; + case RECORD_FUNCTON_TAG_LEN: + rc = seq_write_gcov_u32(seq, 2); + break; + case RECORD_FUNCTION_IDENT: + rc = seq_write_gcov_u32(seq, get_func(iter)->ident); + break; + case RECORD_FUNCTION_CHECK: + rc = seq_write_gcov_u32(seq, get_func(iter)->checksum); + break; + case RECORD_COUNT_TAG: + rc = seq_write_gcov_u32(seq, + GCOV_TAG_FOR_COUNTER(get_type(iter)->ctr_type)); + break; + case RECORD_COUNT_LEN: + rc = seq_write_gcov_u32(seq, + get_func(iter)->n_ctrs[iter->type] * 2); + break; + case RECORD_COUNT: + rc = seq_write_gcov_u64(seq, + iter->info->counts[iter->type]. + values[iter->count + get_type(iter)->offset]); + break; + } + return rc; +} diff --git a/kernel/gcov/gcov.h b/kernel/gcov/gcov.h new file mode 100644 index 00000000000..060073ebf7a --- /dev/null +++ b/kernel/gcov/gcov.h @@ -0,0 +1,128 @@ +/* + * Profiling infrastructure declarations. + * + * This file is based on gcc-internal definitions. Data structures are + * defined to be compatible with gcc counterparts. For a better + * understanding, refer to gcc source: gcc/gcov-io.h. + * + * Copyright IBM Corp. 2009 + * Author(s): Peter Oberparleiter <oberpar@linux.vnet.ibm.com> + * + * Uses gcc-internal data definitions. + */ + +#ifndef GCOV_H +#define GCOV_H GCOV_H + +#include <linux/types.h> + +/* + * Profiling data types used for gcc 3.4 and above - these are defined by + * gcc and need to be kept as close to the original definition as possible to + * remain compatible. + */ +#define GCOV_COUNTERS 5 +#define GCOV_DATA_MAGIC ((unsigned int) 0x67636461) +#define GCOV_TAG_FUNCTION ((unsigned int) 0x01000000) +#define GCOV_TAG_COUNTER_BASE ((unsigned int) 0x01a10000) +#define GCOV_TAG_FOR_COUNTER(count) \ + (GCOV_TAG_COUNTER_BASE + ((unsigned int) (count) << 17)) + +#if BITS_PER_LONG >= 64 +typedef long gcov_type; +#else +typedef long long gcov_type; +#endif + +/** + * struct gcov_fn_info - profiling meta data per function + * @ident: object file-unique function identifier + * @checksum: function checksum + * @n_ctrs: number of values per counter type belonging to this function + * + * This data is generated by gcc during compilation and doesn't change + * at run-time. + */ +struct gcov_fn_info { + unsigned int ident; + unsigned int checksum; + unsigned int n_ctrs[0]; +}; + +/** + * struct gcov_ctr_info - profiling data per counter type + * @num: number of counter values for this type + * @values: array of counter values for this type + * @merge: merge function for counter values of this type (unused) + * + * This data is generated by gcc during compilation and doesn't change + * at run-time with the exception of the values array. + */ +struct gcov_ctr_info { + unsigned int num; + gcov_type *values; + void (*merge)(gcov_type *, unsigned int); +}; + +/** + * struct gcov_info - profiling data per object file + * @version: gcov version magic indicating the gcc version used for compilation + * @next: list head for a singly-linked list + * @stamp: time stamp + * @filename: name of the associated gcov data file + * @n_functions: number of instrumented functions + * @functions: function data + * @ctr_mask: mask specifying which counter types are active + * @counts: counter data per counter type + * + * This data is generated by gcc during compilation and doesn't change + * at run-time with the exception of the next pointer. + */ +struct gcov_info { + unsigned int version; + struct gcov_info *next; + unsigned int stamp; + const char *filename; + unsigned int n_functions; + const struct gcov_fn_info *functions; + unsigned int ctr_mask; + struct gcov_ctr_info counts[0]; +}; + +/* Base interface. */ +enum gcov_action { + GCOV_ADD, + GCOV_REMOVE, +}; + +void gcov_event(enum gcov_action action, struct gcov_info *info); +void gcov_enable_events(void); + +/* Iterator control. */ +struct seq_file; +struct gcov_iterator; + +struct gcov_iterator *gcov_iter_new(struct gcov_info *info); +void gcov_iter_free(struct gcov_iterator *iter); +void gcov_iter_start(struct gcov_iterator *iter); +int gcov_iter_next(struct gcov_iterator *iter); +int gcov_iter_write(struct gcov_iterator *iter, struct seq_file *seq); +struct gcov_info *gcov_iter_get_info(struct gcov_iterator *iter); + +/* gcov_info control. */ +void gcov_info_reset(struct gcov_info *info); +int gcov_info_is_compatible(struct gcov_info *info1, struct gcov_info *info2); +void gcov_info_add(struct gcov_info *dest, struct gcov_info *source); +struct gcov_info *gcov_info_dup(struct gcov_info *info); +void gcov_info_free(struct gcov_info *info); + +struct gcov_link { + enum { + OBJ_TREE, + SRC_TREE, + } dir; + const char *ext; +}; +extern const struct gcov_link gcov_link[]; + +#endif /* GCOV_H */ diff --git a/kernel/groups.c b/kernel/groups.c new file mode 100644 index 00000000000..2b45b2ee396 --- /dev/null +++ b/kernel/groups.c @@ -0,0 +1,288 @@ +/* + * Supplementary group IDs + */ +#include <linux/cred.h> +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/security.h> +#include <linux/syscalls.h> +#include <asm/uaccess.h> + +/* init to 2 - one for init_task, one to ensure it is never freed */ +struct group_info init_groups = { .usage = ATOMIC_INIT(2) }; + +struct group_info *groups_alloc(int gidsetsize) +{ + struct group_info *group_info; + int nblocks; + int i; + + nblocks = (gidsetsize + NGROUPS_PER_BLOCK - 1) / NGROUPS_PER_BLOCK; + /* Make sure we always allocate at least one indirect block pointer */ + nblocks = nblocks ? : 1; + group_info = kmalloc(sizeof(*group_info) + nblocks*sizeof(gid_t *), GFP_USER); + if (!group_info) + return NULL; + group_info->ngroups = gidsetsize; + group_info->nblocks = nblocks; + atomic_set(&group_info->usage, 1); + + if (gidsetsize <= NGROUPS_SMALL) + group_info->blocks[0] = group_info->small_block; + else { + for (i = 0; i < nblocks; i++) { + gid_t *b; + b = (void *)__get_free_page(GFP_USER); + if (!b) + goto out_undo_partial_alloc; + group_info->blocks[i] = b; + } + } + return group_info; + +out_undo_partial_alloc: + while (--i >= 0) { + free_page((unsigned long)group_info->blocks[i]); + } + kfree(group_info); + return NULL; +} + +EXPORT_SYMBOL(groups_alloc); + +void groups_free(struct group_info *group_info) +{ + if (group_info->blocks[0] != group_info->small_block) { + int i; + for (i = 0; i < group_info->nblocks; i++) + free_page((unsigned long)group_info->blocks[i]); + } + kfree(group_info); +} + +EXPORT_SYMBOL(groups_free); + +/* export the group_info to a user-space array */ +static int groups_to_user(gid_t __user *grouplist, + const struct group_info *group_info) +{ + int i; + unsigned int count = group_info->ngroups; + + for (i = 0; i < group_info->nblocks; i++) { + unsigned int cp_count = min(NGROUPS_PER_BLOCK, count); + unsigned int len = cp_count * sizeof(*grouplist); + + if (copy_to_user(grouplist, group_info->blocks[i], len)) + return -EFAULT; + + grouplist += NGROUPS_PER_BLOCK; + count -= cp_count; + } + return 0; +} + +/* fill a group_info from a user-space array - it must be allocated already */ +static int groups_from_user(struct group_info *group_info, + gid_t __user *grouplist) +{ + int i; + unsigned int count = group_info->ngroups; + + for (i = 0; i < group_info->nblocks; i++) { + unsigned int cp_count = min(NGROUPS_PER_BLOCK, count); + unsigned int len = cp_count * sizeof(*grouplist); + + if (copy_from_user(group_info->blocks[i], grouplist, len)) + return -EFAULT; + + grouplist += NGROUPS_PER_BLOCK; + count -= cp_count; + } + return 0; +} + +/* a simple Shell sort */ +static void groups_sort(struct group_info *group_info) +{ + int base, max, stride; + int gidsetsize = group_info->ngroups; + + for (stride = 1; stride < gidsetsize; stride = 3 * stride + 1) + ; /* nothing */ + stride /= 3; + + while (stride) { + max = gidsetsize - stride; + for (base = 0; base < max; base++) { + int left = base; + int right = left + stride; + gid_t tmp = GROUP_AT(group_info, right); + + while (left >= 0 && GROUP_AT(group_info, left) > tmp) { + GROUP_AT(group_info, right) = + GROUP_AT(group_info, left); + right = left; + left -= stride; + } + GROUP_AT(group_info, right) = tmp; + } + stride /= 3; + } +} + +/* a simple bsearch */ +int groups_search(const struct group_info *group_info, gid_t grp) +{ + unsigned int left, right; + + if (!group_info) + return 0; + + left = 0; + right = group_info->ngroups; + while (left < right) { + unsigned int mid = (left+right)/2; + int cmp = grp - GROUP_AT(group_info, mid); + if (cmp > 0) + left = mid + 1; + else if (cmp < 0) + right = mid; + else + return 1; + } + return 0; +} + +/** + * set_groups - Change a group subscription in a set of credentials + * @new: The newly prepared set of credentials to alter + * @group_info: The group list to install + * + * Validate a group subscription and, if valid, insert it into a set + * of credentials. + */ +int set_groups(struct cred *new, struct group_info *group_info) +{ + int retval; + + retval = security_task_setgroups(group_info); + if (retval) + return retval; + + put_group_info(new->group_info); + groups_sort(group_info); + get_group_info(group_info); + new->group_info = group_info; + return 0; +} + +EXPORT_SYMBOL(set_groups); + +/** + * set_current_groups - Change current's group subscription + * @group_info: The group list to impose + * + * Validate a group subscription and, if valid, impose it upon current's task + * security record. + */ +int set_current_groups(struct group_info *group_info) +{ + struct cred *new; + int ret; + + new = prepare_creds(); + if (!new) + return -ENOMEM; + + ret = set_groups(new, group_info); + if (ret < 0) { + abort_creds(new); + return ret; + } + + return commit_creds(new); +} + +EXPORT_SYMBOL(set_current_groups); + +SYSCALL_DEFINE2(getgroups, int, gidsetsize, gid_t __user *, grouplist) +{ + const struct cred *cred = current_cred(); + int i; + + if (gidsetsize < 0) + return -EINVAL; + + /* no need to grab task_lock here; it cannot change */ + i = cred->group_info->ngroups; + if (gidsetsize) { + if (i > gidsetsize) { + i = -EINVAL; + goto out; + } + if (groups_to_user(grouplist, cred->group_info)) { + i = -EFAULT; + goto out; + } + } +out: + return i; +} + +/* + * SMP: Our groups are copy-on-write. We can set them safely + * without another task interfering. + */ + +SYSCALL_DEFINE2(setgroups, int, gidsetsize, gid_t __user *, grouplist) +{ + struct group_info *group_info; + int retval; + + if (!capable(CAP_SETGID)) + return -EPERM; + if ((unsigned)gidsetsize > NGROUPS_MAX) + return -EINVAL; + + group_info = groups_alloc(gidsetsize); + if (!group_info) + return -ENOMEM; + retval = groups_from_user(group_info, grouplist); + if (retval) { + put_group_info(group_info); + return retval; + } + + retval = set_current_groups(group_info); + put_group_info(group_info); + + return retval; +} + +/* + * Check whether we're fsgid/egid or in the supplemental group.. + */ +int in_group_p(gid_t grp) +{ + const struct cred *cred = current_cred(); + int retval = 1; + + if (grp != cred->fsgid) + retval = groups_search(cred->group_info, grp); + return retval; +} + +EXPORT_SYMBOL(in_group_p); + +int in_egroup_p(gid_t grp) +{ + const struct cred *cred = current_cred(); + int retval = 1; + + if (grp != cred->egid) + retval = groups_search(cred->group_info, grp); + return retval; +} + +EXPORT_SYMBOL(in_egroup_p); diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c index 2dc30c59c5f..05071bf6a37 100644 --- a/kernel/hrtimer.c +++ b/kernel/hrtimer.c @@ -43,6 +43,8 @@ #include <linux/seq_file.h> #include <linux/err.h> #include <linux/debugobjects.h> +#include <linux/sched.h> +#include <linux/timer.h> #include <asm/uaccess.h> @@ -189,21 +191,65 @@ struct hrtimer_clock_base *lock_hrtimer_base(const struct hrtimer *timer, } } + +/* + * Get the preferred target CPU for NOHZ + */ +static int hrtimer_get_target(int this_cpu, int pinned) +{ +#ifdef CONFIG_NO_HZ + if (!pinned && get_sysctl_timer_migration() && idle_cpu(this_cpu)) { + int preferred_cpu = get_nohz_load_balancer(); + + if (preferred_cpu >= 0) + return preferred_cpu; + } +#endif + return this_cpu; +} + +/* + * With HIGHRES=y we do not migrate the timer when it is expiring + * before the next event on the target cpu because we cannot reprogram + * the target cpu hardware and we would cause it to fire late. + * + * Called with cpu_base->lock of target cpu held. + */ +static int +hrtimer_check_target(struct hrtimer *timer, struct hrtimer_clock_base *new_base) +{ +#ifdef CONFIG_HIGH_RES_TIMERS + ktime_t expires; + + if (!new_base->cpu_base->hres_active) + return 0; + + expires = ktime_sub(hrtimer_get_expires(timer), new_base->offset); + return expires.tv64 <= new_base->cpu_base->expires_next.tv64; +#else + return 0; +#endif +} + /* * Switch the timer base to the current CPU when possible. */ static inline struct hrtimer_clock_base * -switch_hrtimer_base(struct hrtimer *timer, struct hrtimer_clock_base *base) +switch_hrtimer_base(struct hrtimer *timer, struct hrtimer_clock_base *base, + int pinned) { struct hrtimer_clock_base *new_base; struct hrtimer_cpu_base *new_cpu_base; + int this_cpu = smp_processor_id(); + int cpu = hrtimer_get_target(this_cpu, pinned); - new_cpu_base = &__get_cpu_var(hrtimer_bases); +again: + new_cpu_base = &per_cpu(hrtimer_bases, cpu); new_base = &new_cpu_base->clock_base[base->index]; if (base != new_base) { /* - * We are trying to schedule the timer on the local CPU. + * We are trying to move timer to new_base. * However we can't change timer's base while it is running, * so we keep it on the same CPU. No hassle vs. reprogramming * the event source in the high resolution case. The softirq @@ -218,6 +264,14 @@ switch_hrtimer_base(struct hrtimer *timer, struct hrtimer_clock_base *base) timer->base = NULL; spin_unlock(&base->cpu_base->lock); spin_lock(&new_base->cpu_base->lock); + + if (cpu != this_cpu && hrtimer_check_target(timer, new_base)) { + cpu = this_cpu; + spin_unlock(&new_base->cpu_base->lock); + spin_lock(&base->cpu_base->lock); + timer->base = base; + goto again; + } timer->base = new_base; } return new_base; @@ -235,7 +289,7 @@ lock_hrtimer_base(const struct hrtimer *timer, unsigned long *flags) return base; } -# define switch_hrtimer_base(t, b) (b) +# define switch_hrtimer_base(t, b, p) (b) #endif /* !CONFIG_SMP */ @@ -332,6 +386,8 @@ ktime_t ktime_add_safe(const ktime_t lhs, const ktime_t rhs) return res; } +EXPORT_SYMBOL_GPL(ktime_add_safe); + #ifdef CONFIG_DEBUG_OBJECTS_TIMERS static struct debug_obj_descr hrtimer_debug_descr; @@ -429,6 +485,7 @@ void hrtimer_init_on_stack(struct hrtimer *timer, clockid_t clock_id, debug_object_init_on_stack(timer, &hrtimer_debug_descr); __hrtimer_init(timer, clock_id, mode); } +EXPORT_SYMBOL_GPL(hrtimer_init_on_stack); void destroy_hrtimer_on_stack(struct hrtimer *timer) { @@ -501,6 +558,13 @@ static void hrtimer_force_reprogram(struct hrtimer_cpu_base *cpu_base) continue; timer = rb_entry(base->first, struct hrtimer, node); expires = ktime_sub(hrtimer_get_expires(timer), base->offset); + /* + * clock_was_set() has changed base->offset so the + * result might be negative. Fix it up to prevent a + * false positive in clockevents_program_event() + */ + if (expires.tv64 < 0) + expires.tv64 = 0; if (expires.tv64 < cpu_base->expires_next.tv64) cpu_base->expires_next = expires; } @@ -614,7 +678,9 @@ void clock_was_set(void) */ void hres_timers_resume(void) { - /* Retrigger the CPU local events: */ + WARN_ONCE(!irqs_disabled(), + KERN_INFO "hres_timers_resume() called with IRQs enabled!"); + retrigger_next_event(NULL); } @@ -642,14 +708,20 @@ static inline void hrtimer_init_timer_hres(struct hrtimer *timer) * and expiry check is done in the hrtimer_interrupt or in the softirq. */ static inline int hrtimer_enqueue_reprogram(struct hrtimer *timer, - struct hrtimer_clock_base *base) + struct hrtimer_clock_base *base, + int wakeup) { if (base->cpu_base->hres_active && hrtimer_reprogram(timer, base)) { - spin_unlock(&base->cpu_base->lock); - raise_softirq_irqoff(HRTIMER_SOFTIRQ); - spin_lock(&base->cpu_base->lock); + if (wakeup) { + spin_unlock(&base->cpu_base->lock); + raise_softirq_irqoff(HRTIMER_SOFTIRQ); + spin_lock(&base->cpu_base->lock); + } else + __raise_softirq_irqoff(HRTIMER_SOFTIRQ); + return 1; } + return 0; } @@ -694,7 +766,8 @@ static inline int hrtimer_is_hres_enabled(void) { return 0; } static inline int hrtimer_switch_to_hres(void) { return 0; } static inline void hrtimer_force_reprogram(struct hrtimer_cpu_base *base) { } static inline int hrtimer_enqueue_reprogram(struct hrtimer *timer, - struct hrtimer_clock_base *base) + struct hrtimer_clock_base *base, + int wakeup) { return 0; } @@ -877,20 +950,9 @@ remove_hrtimer(struct hrtimer *timer, struct hrtimer_clock_base *base) return 0; } -/** - * hrtimer_start_range_ns - (re)start an hrtimer on the current CPU - * @timer: the timer to be added - * @tim: expiry time - * @delta_ns: "slack" range for the timer - * @mode: expiry mode: absolute (HRTIMER_ABS) or relative (HRTIMER_REL) - * - * Returns: - * 0 on success - * 1 when the timer was active - */ -int -hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, unsigned long delta_ns, - const enum hrtimer_mode mode) +int __hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, + unsigned long delta_ns, const enum hrtimer_mode mode, + int wakeup) { struct hrtimer_clock_base *base, *new_base; unsigned long flags; @@ -902,9 +964,9 @@ hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, unsigned long delta_n ret = remove_hrtimer(timer, base); /* Switch the timer base, if necessary: */ - new_base = switch_hrtimer_base(timer, base); + new_base = switch_hrtimer_base(timer, base, mode & HRTIMER_MODE_PINNED); - if (mode == HRTIMER_MODE_REL) { + if (mode & HRTIMER_MODE_REL) { tim = ktime_add_safe(tim, new_base->get_time()); /* * CONFIG_TIME_LOW_RES is a temporary way for architectures @@ -931,12 +993,29 @@ hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, unsigned long delta_n * XXX send_remote_softirq() ? */ if (leftmost && new_base->cpu_base == &__get_cpu_var(hrtimer_bases)) - hrtimer_enqueue_reprogram(timer, new_base); + hrtimer_enqueue_reprogram(timer, new_base, wakeup); unlock_hrtimer_base(timer, &flags); return ret; } + +/** + * hrtimer_start_range_ns - (re)start an hrtimer on the current CPU + * @timer: the timer to be added + * @tim: expiry time + * @delta_ns: "slack" range for the timer + * @mode: expiry mode: absolute (HRTIMER_ABS) or relative (HRTIMER_REL) + * + * Returns: + * 0 on success + * 1 when the timer was active + */ +int hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, + unsigned long delta_ns, const enum hrtimer_mode mode) +{ + return __hrtimer_start_range_ns(timer, tim, delta_ns, mode, 1); +} EXPORT_SYMBOL_GPL(hrtimer_start_range_ns); /** @@ -952,7 +1031,7 @@ EXPORT_SYMBOL_GPL(hrtimer_start_range_ns); int hrtimer_start(struct hrtimer *timer, ktime_t tim, const enum hrtimer_mode mode) { - return hrtimer_start_range_ns(timer, tim, 0, mode); + return __hrtimer_start_range_ns(timer, tim, 0, mode, 1); } EXPORT_SYMBOL_GPL(hrtimer_start); @@ -1156,6 +1235,29 @@ static void __run_hrtimer(struct hrtimer *timer) #ifdef CONFIG_HIGH_RES_TIMERS +static int force_clock_reprogram; + +/* + * After 5 iteration's attempts, we consider that hrtimer_interrupt() + * is hanging, which could happen with something that slows the interrupt + * such as the tracing. Then we force the clock reprogramming for each future + * hrtimer interrupts to avoid infinite loops and use the min_delta_ns + * threshold that we will overwrite. + * The next tick event will be scheduled to 3 times we currently spend on + * hrtimer_interrupt(). This gives a good compromise, the cpus will spend + * 1/4 of their time to process the hrtimer interrupts. This is enough to + * let it running without serious starvation. + */ + +static inline void +hrtimer_interrupt_hanging(struct clock_event_device *dev, + ktime_t try_time) +{ + force_clock_reprogram = 1; + dev->min_delta_ns = (unsigned long)try_time.tv64 * 3; + printk(KERN_WARNING "hrtimer: interrupt too slow, " + "forcing clock min delta to %lu ns\n", dev->min_delta_ns); +} /* * High resolution timer interrupt * Called with interrupts disabled @@ -1165,6 +1267,7 @@ void hrtimer_interrupt(struct clock_event_device *dev) struct hrtimer_cpu_base *cpu_base = &__get_cpu_var(hrtimer_bases); struct hrtimer_clock_base *base; ktime_t expires_next, now; + int nr_retries = 0; int i; BUG_ON(!cpu_base->hres_active); @@ -1172,18 +1275,30 @@ void hrtimer_interrupt(struct clock_event_device *dev) dev->next_event.tv64 = KTIME_MAX; retry: + /* 5 retries is enough to notice a hang */ + if (!(++nr_retries % 5)) + hrtimer_interrupt_hanging(dev, ktime_sub(ktime_get(), now)); + now = ktime_get(); expires_next.tv64 = KTIME_MAX; + spin_lock(&cpu_base->lock); + /* + * We set expires_next to KTIME_MAX here with cpu_base->lock + * held to prevent that a timer is enqueued in our queue via + * the migration code. This does not affect enqueueing of + * timers which run their callback and need to be requeued on + * this CPU. + */ + cpu_base->expires_next.tv64 = KTIME_MAX; + base = cpu_base->clock_base; for (i = 0; i < HRTIMER_MAX_CLOCK_BASES; i++) { ktime_t basenow; struct rb_node *node; - spin_lock(&cpu_base->lock); - basenow = ktime_add(now, base->offset); while ((node = base->first)) { @@ -1216,15 +1331,19 @@ void hrtimer_interrupt(struct clock_event_device *dev) __run_hrtimer(timer); } - spin_unlock(&cpu_base->lock); base++; } + /* + * Store the new expiry value so the migration code can verify + * against it. + */ cpu_base->expires_next = expires_next; + spin_unlock(&cpu_base->lock); /* Reprogramming necessary ? */ if (expires_next.tv64 != KTIME_MAX) { - if (tick_program_event(expires_next, 0)) + if (tick_program_event(expires_next, force_clock_reprogram)) goto retry; } } @@ -1359,6 +1478,7 @@ void hrtimer_init_sleeper(struct hrtimer_sleeper *sl, struct task_struct *task) sl->timer.function = hrtimer_wakeup; sl->task = task; } +EXPORT_SYMBOL_GPL(hrtimer_init_sleeper); static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum hrtimer_mode mode) { @@ -1578,6 +1698,10 @@ static int __cpuinit hrtimer_cpu_notify(struct notifier_block *self, break; #ifdef CONFIG_HOTPLUG_CPU + case CPU_DYING: + case CPU_DYING_FROZEN: + clockevents_notify(CLOCK_EVT_NOTIFY_CPU_DYING, &scpu); + break; case CPU_DEAD: case CPU_DEAD_FROZEN: { diff --git a/kernel/hung_task.c b/kernel/hung_task.c new file mode 100644 index 00000000000..022a4927b78 --- /dev/null +++ b/kernel/hung_task.c @@ -0,0 +1,217 @@ +/* + * Detect Hung Task + * + * kernel/hung_task.c - kernel thread for detecting tasks stuck in D state + * + */ + +#include <linux/mm.h> +#include <linux/cpu.h> +#include <linux/nmi.h> +#include <linux/init.h> +#include <linux/delay.h> +#include <linux/freezer.h> +#include <linux/kthread.h> +#include <linux/lockdep.h> +#include <linux/module.h> +#include <linux/sysctl.h> + +/* + * The number of tasks checked: + */ +unsigned long __read_mostly sysctl_hung_task_check_count = PID_MAX_LIMIT; + +/* + * Limit number of tasks checked in a batch. + * + * This value controls the preemptibility of khungtaskd since preemption + * is disabled during the critical section. It also controls the size of + * the RCU grace period. So it needs to be upper-bound. + */ +#define HUNG_TASK_BATCHING 1024 + +/* + * Zero means infinite timeout - no checking done: + */ +unsigned long __read_mostly sysctl_hung_task_timeout_secs = 120; + +unsigned long __read_mostly sysctl_hung_task_warnings = 10; + +static int __read_mostly did_panic; + +static struct task_struct *watchdog_task; + +/* + * Should we panic (and reboot, if panic_timeout= is set) when a + * hung task is detected: + */ +unsigned int __read_mostly sysctl_hung_task_panic = + CONFIG_BOOTPARAM_HUNG_TASK_PANIC_VALUE; + +static int __init hung_task_panic_setup(char *str) +{ + sysctl_hung_task_panic = simple_strtoul(str, NULL, 0); + + return 1; +} +__setup("hung_task_panic=", hung_task_panic_setup); + +static int +hung_task_panic(struct notifier_block *this, unsigned long event, void *ptr) +{ + did_panic = 1; + + return NOTIFY_DONE; +} + +static struct notifier_block panic_block = { + .notifier_call = hung_task_panic, +}; + +static void check_hung_task(struct task_struct *t, unsigned long timeout) +{ + unsigned long switch_count = t->nvcsw + t->nivcsw; + + /* + * Ensure the task is not frozen. + * Also, when a freshly created task is scheduled once, changes + * its state to TASK_UNINTERRUPTIBLE without having ever been + * switched out once, it musn't be checked. + */ + if (unlikely(t->flags & PF_FROZEN || !switch_count)) + return; + + if (switch_count != t->last_switch_count) { + t->last_switch_count = switch_count; + return; + } + if (!sysctl_hung_task_warnings) + return; + sysctl_hung_task_warnings--; + + /* + * Ok, the task did not get scheduled for more than 2 minutes, + * complain: + */ + printk(KERN_ERR "INFO: task %s:%d blocked for more than " + "%ld seconds.\n", t->comm, t->pid, timeout); + printk(KERN_ERR "\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\"" + " disables this message.\n"); + sched_show_task(t); + __debug_show_held_locks(t); + + touch_nmi_watchdog(); + + if (sysctl_hung_task_panic) + panic("hung_task: blocked tasks"); +} + +/* + * To avoid extending the RCU grace period for an unbounded amount of time, + * periodically exit the critical section and enter a new one. + * + * For preemptible RCU it is sufficient to call rcu_read_unlock in order + * exit the grace period. For classic RCU, a reschedule is required. + */ +static void rcu_lock_break(struct task_struct *g, struct task_struct *t) +{ + get_task_struct(g); + get_task_struct(t); + rcu_read_unlock(); + cond_resched(); + rcu_read_lock(); + put_task_struct(t); + put_task_struct(g); +} + +/* + * Check whether a TASK_UNINTERRUPTIBLE does not get woken up for + * a really long time (120 seconds). If that happens, print out + * a warning. + */ +static void check_hung_uninterruptible_tasks(unsigned long timeout) +{ + int max_count = sysctl_hung_task_check_count; + int batch_count = HUNG_TASK_BATCHING; + struct task_struct *g, *t; + + /* + * If the system crashed already then all bets are off, + * do not report extra hung tasks: + */ + if (test_taint(TAINT_DIE) || did_panic) + return; + + rcu_read_lock(); + do_each_thread(g, t) { + if (!--max_count) + goto unlock; + if (!--batch_count) { + batch_count = HUNG_TASK_BATCHING; + rcu_lock_break(g, t); + /* Exit if t or g was unhashed during refresh. */ + if (t->state == TASK_DEAD || g->state == TASK_DEAD) + goto unlock; + } + /* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */ + if (t->state == TASK_UNINTERRUPTIBLE) + check_hung_task(t, timeout); + } while_each_thread(g, t); + unlock: + rcu_read_unlock(); +} + +static unsigned long timeout_jiffies(unsigned long timeout) +{ + /* timeout of 0 will disable the watchdog */ + return timeout ? timeout * HZ : MAX_SCHEDULE_TIMEOUT; +} + +/* + * Process updating of timeout sysctl + */ +int proc_dohung_task_timeout_secs(struct ctl_table *table, int write, + struct file *filp, void __user *buffer, + size_t *lenp, loff_t *ppos) +{ + int ret; + + ret = proc_doulongvec_minmax(table, write, filp, buffer, lenp, ppos); + + if (ret || !write) + goto out; + + wake_up_process(watchdog_task); + + out: + return ret; +} + +/* + * kthread which checks for tasks stuck in D state + */ +static int watchdog(void *dummy) +{ + set_user_nice(current, 0); + + for ( ; ; ) { + unsigned long timeout = sysctl_hung_task_timeout_secs; + + while (schedule_timeout_interruptible(timeout_jiffies(timeout))) + timeout = sysctl_hung_task_timeout_secs; + + check_hung_uninterruptible_tasks(timeout); + } + + return 0; +} + +static int __init hung_task_init(void) +{ + atomic_notifier_chain_register(&panic_notifier_list, &panic_block); + watchdog_task = kthread_run(watchdog, NULL, "khungtaskd"); + + return 0; +} + +module_init(hung_task_init); diff --git a/kernel/irq/Makefile b/kernel/irq/Makefile index 4dd5b1edac9..7d047808419 100644 --- a/kernel/irq/Makefile +++ b/kernel/irq/Makefile @@ -3,4 +3,5 @@ obj-y := handle.o manage.o spurious.o resend.o chip.o devres.o obj-$(CONFIG_GENERIC_IRQ_PROBE) += autoprobe.o obj-$(CONFIG_PROC_FS) += proc.o obj-$(CONFIG_GENERIC_PENDING_IRQ) += migration.o -obj-$(CONFIG_NUMA_MIGRATE_IRQ_DESC) += numa_migrate.o +obj-$(CONFIG_NUMA_IRQ_DESC) += numa_migrate.o +obj-$(CONFIG_PM_SLEEP) += pm.o diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index f63c706d25e..c1660194d11 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -46,7 +46,10 @@ void dynamic_irq_init(unsigned int irq) desc->irq_count = 0; desc->irqs_unhandled = 0; #ifdef CONFIG_SMP - cpumask_setall(&desc->affinity); + cpumask_setall(desc->affinity); +#ifdef CONFIG_GENERIC_PENDING_IRQ + cpumask_clear(desc->pending_mask); +#endif #endif spin_unlock_irqrestore(&desc->lock, flags); } @@ -78,6 +81,7 @@ void dynamic_irq_cleanup(unsigned int irq) desc->handle_irq = handle_bad_irq; desc->chip = &no_irq_chip; desc->name = NULL; + clear_kstat_irqs(desc); spin_unlock_irqrestore(&desc->lock, flags); } @@ -218,6 +222,34 @@ int set_irq_chip_data(unsigned int irq, void *data) } EXPORT_SYMBOL(set_irq_chip_data); +/** + * set_irq_nested_thread - Set/Reset the IRQ_NESTED_THREAD flag of an irq + * + * @irq: Interrupt number + * @nest: 0 to clear / 1 to set the IRQ_NESTED_THREAD flag + * + * The IRQ_NESTED_THREAD flag indicates that on + * request_threaded_irq() no separate interrupt thread should be + * created for the irq as the handler are called nested in the + * context of a demultiplexing interrupt handler thread. + */ +void set_irq_nested_thread(unsigned int irq, int nest) +{ + struct irq_desc *desc = irq_to_desc(irq); + unsigned long flags; + + if (!desc) + return; + + spin_lock_irqsave(&desc->lock, flags); + if (nest) + desc->status |= IRQ_NESTED_THREAD; + else + desc->status &= ~IRQ_NESTED_THREAD; + spin_unlock_irqrestore(&desc->lock, flags); +} +EXPORT_SYMBOL_GPL(set_irq_nested_thread); + /* * default enable function */ @@ -290,10 +322,50 @@ static inline void mask_ack_irq(struct irq_desc *desc, int irq) desc->chip->mask_ack(irq); else { desc->chip->mask(irq); - desc->chip->ack(irq); + if (desc->chip->ack) + desc->chip->ack(irq); } } +/* + * handle_nested_irq - Handle a nested irq from a irq thread + * @irq: the interrupt number + * + * Handle interrupts which are nested into a threaded interrupt + * handler. The handler function is called inside the calling + * threads context. + */ +void handle_nested_irq(unsigned int irq) +{ + struct irq_desc *desc = irq_to_desc(irq); + struct irqaction *action; + irqreturn_t action_ret; + + might_sleep(); + + spin_lock_irq(&desc->lock); + + kstat_incr_irqs_this_cpu(irq, desc); + + action = desc->action; + if (unlikely(!action || (desc->status & IRQ_DISABLED))) + goto out_unlock; + + desc->status |= IRQ_INPROGRESS; + spin_unlock_irq(&desc->lock); + + action_ret = action->thread_fn(action->irq, action->dev_id); + if (!noirqdebug) + note_interrupt(irq, desc, action_ret); + + spin_lock_irq(&desc->lock); + desc->status &= ~IRQ_INPROGRESS; + +out_unlock: + spin_unlock_irq(&desc->lock); +} +EXPORT_SYMBOL_GPL(handle_nested_irq); + /** * handle_simple_irq - Simple and software-decoded IRQs. * @irq: the interrupt number @@ -354,7 +426,6 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc) spin_lock(&desc->lock); mask_ack_irq(desc, irq); - desc = irq_remap_to_desc(irq, desc); if (unlikely(desc->status & IRQ_INPROGRESS)) goto out_unlock; @@ -378,11 +449,15 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc) spin_lock(&desc->lock); desc->status &= ~IRQ_INPROGRESS; - if (!(desc->status & IRQ_DISABLED) && desc->chip->unmask) + + if (unlikely(desc->status & IRQ_ONESHOT)) + desc->status |= IRQ_MASKED; + else if (!(desc->status & IRQ_DISABLED) && desc->chip->unmask) desc->chip->unmask(irq); out_unlock: spin_unlock(&desc->lock); } +EXPORT_SYMBOL_GPL(handle_level_irq); /** * handle_fasteoi_irq - irq handler for transparent controllers @@ -432,7 +507,6 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc) desc->status &= ~IRQ_INPROGRESS; out: desc->chip->eoi(irq); - desc = irq_remap_to_desc(irq, desc); spin_unlock(&desc->lock); } @@ -469,14 +543,13 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) !desc->action)) { desc->status |= (IRQ_PENDING | IRQ_MASKED); mask_ack_irq(desc, irq); - desc = irq_remap_to_desc(irq, desc); goto out_unlock; } kstat_incr_irqs_this_cpu(irq, desc); /* Start handling the irq */ - desc->chip->ack(irq); - desc = irq_remap_to_desc(irq, desc); + if (desc->chip->ack) + desc->chip->ack(irq); /* Mark the IRQ currently in progress.*/ desc->status |= IRQ_INPROGRESS; @@ -537,10 +610,8 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc) if (!noirqdebug) note_interrupt(irq, desc, action_ret); - if (desc->chip->eoi) { + if (desc->chip->eoi) desc->chip->eoi(irq); - desc = irq_remap_to_desc(irq, desc); - } } void @@ -571,14 +642,13 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, desc->chip = &dummy_irq_chip; } + chip_bus_lock(irq, desc); spin_lock_irqsave(&desc->lock, flags); /* Uninstall? */ if (handle == handle_bad_irq) { - if (desc->chip != &no_irq_chip) { + if (desc->chip != &no_irq_chip) mask_ack_irq(desc, irq); - desc = irq_remap_to_desc(irq, desc); - } desc->status |= IRQ_DISABLED; desc->depth = 1; } @@ -592,7 +662,9 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, desc->chip->startup(irq); } spin_unlock_irqrestore(&desc->lock, flags); + chip_bus_sync_unlock(irq, desc); } +EXPORT_SYMBOL_GPL(__set_irq_handler); void set_irq_chip_and_handler(unsigned int irq, struct irq_chip *chip, diff --git a/kernel/irq/devres.c b/kernel/irq/devres.c index 38a25b8d8bf..d06df9c41cb 100644 --- a/kernel/irq/devres.c +++ b/kernel/irq/devres.c @@ -26,10 +26,12 @@ static int devm_irq_match(struct device *dev, void *res, void *data) } /** - * devm_request_irq - allocate an interrupt line for a managed device + * devm_request_threaded_irq - allocate an interrupt line for a managed device * @dev: device to request interrupt for * @irq: Interrupt line to allocate * @handler: Function to be called when the IRQ occurs + * @thread_fn: function to be called in a threaded interrupt context. NULL + * for devices which handle everything in @handler * @irqflags: Interrupt type flags * @devname: An ascii name for the claiming device * @dev_id: A cookie passed back to the handler function @@ -42,9 +44,10 @@ static int devm_irq_match(struct device *dev, void *res, void *data) * If an IRQ allocated with this function needs to be freed * separately, dev_free_irq() must be used. */ -int devm_request_irq(struct device *dev, unsigned int irq, - irq_handler_t handler, unsigned long irqflags, - const char *devname, void *dev_id) +int devm_request_threaded_irq(struct device *dev, unsigned int irq, + irq_handler_t handler, irq_handler_t thread_fn, + unsigned long irqflags, const char *devname, + void *dev_id) { struct irq_devres *dr; int rc; @@ -54,7 +57,8 @@ int devm_request_irq(struct device *dev, unsigned int irq, if (!dr) return -ENOMEM; - rc = request_irq(irq, handler, irqflags, devname, dev_id); + rc = request_threaded_irq(irq, handler, thread_fn, irqflags, devname, + dev_id); if (rc) { devres_free(dr); return rc; @@ -66,7 +70,7 @@ int devm_request_irq(struct device *dev, unsigned int irq, return 0; } -EXPORT_SYMBOL(devm_request_irq); +EXPORT_SYMBOL(devm_request_threaded_irq); /** * devm_free_irq - free an interrupt diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index c20db0be917..a81cf80554d 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -11,12 +11,15 @@ */ #include <linux/irq.h> +#include <linux/slab.h> #include <linux/module.h> #include <linux/random.h> #include <linux/interrupt.h> #include <linux/kernel_stat.h> #include <linux/rculist.h> #include <linux/hash.h> +#include <linux/bootmem.h> +#include <trace/events/irq.h> #include "internals.h" @@ -39,6 +42,18 @@ void handle_bad_irq(unsigned int irq, struct irq_desc *desc) ack_bad_irq(irq); } +#if defined(CONFIG_SMP) && defined(CONFIG_GENERIC_HARDIRQS) +static void __init init_irq_default_affinity(void) +{ + alloc_cpumask_var(&irq_default_affinity, GFP_NOWAIT); + cpumask_setall(irq_default_affinity); +} +#else +static void __init init_irq_default_affinity(void) +{ +} +#endif + /* * Linux has a controller-independent interrupt architecture. * Every controller has a 'controller-template', that is used @@ -57,6 +72,7 @@ int nr_irqs = NR_IRQS; EXPORT_SYMBOL_GPL(nr_irqs); #ifdef CONFIG_SPARSE_IRQ + static struct irq_desc irq_desc_init = { .irq = -1, .status = IRQ_DISABLED, @@ -64,44 +80,50 @@ static struct irq_desc irq_desc_init = { .handle_irq = handle_bad_irq, .depth = 1, .lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock), -#ifdef CONFIG_SMP - .affinity = CPU_MASK_ALL -#endif }; -void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) +void __ref init_kstat_irqs(struct irq_desc *desc, int node, int nr) { - unsigned long bytes; - char *ptr; - int node; - - /* Compute how many bytes we need per irq and allocate them */ - bytes = nr * sizeof(unsigned int); + void *ptr; - node = cpu_to_node(cpu); - ptr = kzalloc_node(bytes, GFP_ATOMIC, node); - printk(KERN_DEBUG " alloc kstat_irqs on cpu %d node %d\n", cpu, node); + if (slab_is_available()) + ptr = kzalloc_node(nr * sizeof(*desc->kstat_irqs), + GFP_ATOMIC, node); + else + ptr = alloc_bootmem_node(NODE_DATA(node), + nr * sizeof(*desc->kstat_irqs)); - if (ptr) - desc->kstat_irqs = (unsigned int *)ptr; + /* + * don't overwite if can not get new one + * init_copy_kstat_irqs() could still use old one + */ + if (ptr) { + printk(KERN_DEBUG " alloc kstat_irqs on node %d\n", node); + desc->kstat_irqs = ptr; + } } -static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) +static void init_one_irq_desc(int irq, struct irq_desc *desc, int node) { memcpy(desc, &irq_desc_init, sizeof(struct irq_desc)); spin_lock_init(&desc->lock); desc->irq = irq; #ifdef CONFIG_SMP - desc->cpu = cpu; + desc->node = node; #endif lockdep_set_class(&desc->lock, &irq_desc_lock_class); - init_kstat_irqs(desc, cpu, nr_cpu_ids); + init_kstat_irqs(desc, node, nr_cpu_ids); if (!desc->kstat_irqs) { printk(KERN_ERR "can not alloc kstat_irqs\n"); BUG_ON(1); } - arch_init_chip_data(desc, cpu); + if (!alloc_desc_masks(desc, node, false)) { + printk(KERN_ERR "can not alloc irq_desc cpumasks\n"); + BUG_ON(1); + } + init_desc_masks(desc); + arch_init_chip_data(desc, node); } /* @@ -109,7 +131,7 @@ static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) */ DEFINE_SPINLOCK(sparse_irq_lock); -struct irq_desc *irq_desc_ptrs[NR_IRQS] __read_mostly; +struct irq_desc **irq_desc_ptrs __read_mostly; static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_smp = { [0 ... NR_IRQS_LEGACY-1] = { @@ -119,33 +141,48 @@ static struct irq_desc irq_desc_legacy[NR_IRQS_LEGACY] __cacheline_aligned_in_sm .handle_irq = handle_bad_irq, .depth = 1, .lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock), -#ifdef CONFIG_SMP - .affinity = CPU_MASK_ALL -#endif } }; -/* FIXME: use bootmem alloc ...*/ -static unsigned int kstat_irqs_legacy[NR_IRQS_LEGACY][NR_CPUS]; +static unsigned int *kstat_irqs_legacy; int __init early_irq_init(void) { struct irq_desc *desc; int legacy_count; + int node; int i; + init_irq_default_affinity(); + + /* initialize nr_irqs based on nr_cpu_ids */ + arch_probe_nr_irqs(); + printk(KERN_INFO "NR_IRQS:%d nr_irqs:%d\n", NR_IRQS, nr_irqs); + desc = irq_desc_legacy; legacy_count = ARRAY_SIZE(irq_desc_legacy); + node = first_online_node; + + /* allocate irq_desc_ptrs array based on nr_irqs */ + irq_desc_ptrs = kcalloc(nr_irqs, sizeof(void *), GFP_NOWAIT); + + /* allocate based on nr_cpu_ids */ + kstat_irqs_legacy = kzalloc_node(NR_IRQS_LEGACY * nr_cpu_ids * + sizeof(int), GFP_NOWAIT, node); for (i = 0; i < legacy_count; i++) { desc[i].irq = i; - desc[i].kstat_irqs = kstat_irqs_legacy[i]; +#ifdef CONFIG_SMP + desc[i].node = node; +#endif + desc[i].kstat_irqs = kstat_irqs_legacy + i * nr_cpu_ids; lockdep_set_class(&desc[i].lock, &irq_desc_lock_class); - + alloc_desc_masks(&desc[i], node, true); + init_desc_masks(&desc[i]); irq_desc_ptrs[i] = desc + i; } - for (i = legacy_count; i < NR_IRQS; i++) + for (i = legacy_count; i < nr_irqs; i++) irq_desc_ptrs[i] = NULL; return arch_early_irq_init(); @@ -153,19 +190,20 @@ int __init early_irq_init(void) struct irq_desc *irq_to_desc(unsigned int irq) { - return (irq < NR_IRQS) ? irq_desc_ptrs[irq] : NULL; + if (irq_desc_ptrs && irq < nr_irqs) + return irq_desc_ptrs[irq]; + + return NULL; } -struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) +struct irq_desc * __ref irq_to_desc_alloc_node(unsigned int irq, int node) { struct irq_desc *desc; unsigned long flags; - int node; - if (irq >= NR_IRQS) { - printk(KERN_WARNING "irq >= NR_IRQS in irq_to_desc_alloc: %d %d\n", - irq, NR_IRQS); - WARN_ON(1); + if (irq >= nr_irqs) { + WARN(1, "irq (%d) >= nr_irqs (%d) in irq_to_desc_alloc\n", + irq, nr_irqs); return NULL; } @@ -180,15 +218,17 @@ struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) if (desc) goto out_unlock; - node = cpu_to_node(cpu); - desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); - printk(KERN_DEBUG " alloc irq_desc for %d on cpu %d node %d\n", - irq, cpu, node); + if (slab_is_available()) + desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); + else + desc = alloc_bootmem_node(NODE_DATA(node), sizeof(*desc)); + + printk(KERN_DEBUG " alloc irq_desc for %d on node %d\n", irq, node); if (!desc) { printk(KERN_ERR "can not alloc irq_desc\n"); BUG_ON(1); } - init_one_irq_desc(irq, desc, cpu); + init_one_irq_desc(irq, desc, node); irq_desc_ptrs[irq] = desc; @@ -207,24 +247,29 @@ struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { .handle_irq = handle_bad_irq, .depth = 1, .lock = __SPIN_LOCK_UNLOCKED(irq_desc->lock), -#ifdef CONFIG_SMP - .affinity = CPU_MASK_ALL -#endif } }; +static unsigned int kstat_irqs_all[NR_IRQS][NR_CPUS]; int __init early_irq_init(void) { struct irq_desc *desc; int count; int i; + init_irq_default_affinity(); + + printk(KERN_INFO "NR_IRQS:%d\n", NR_IRQS); + desc = irq_desc; count = ARRAY_SIZE(irq_desc); - for (i = 0; i < count; i++) + for (i = 0; i < count; i++) { desc[i].irq = i; - + alloc_desc_masks(&desc[i], 0, true); + init_desc_masks(&desc[i]); + desc[i].kstat_irqs = kstat_irqs_all[i]; + } return arch_early_irq_init(); } @@ -233,12 +278,17 @@ struct irq_desc *irq_to_desc(unsigned int irq) return (irq < NR_IRQS) ? irq_desc + irq : NULL; } -struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) +struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node) { return irq_to_desc(irq); } #endif /* !CONFIG_SPARSE_IRQ */ +void clear_kstat_irqs(struct irq_desc *desc) +{ + memset(desc->kstat_irqs, 0, nr_cpu_ids * sizeof(*(desc->kstat_irqs))); +} + /* * What should we do if we get a hw irq event on an illegal vector? * Each architecture has to answer this themself. @@ -300,6 +350,15 @@ irqreturn_t no_action(int cpl, void *dev_id) return IRQ_NONE; } +static void warn_no_thread(unsigned int irq, struct irqaction *action) +{ + if (test_and_set_bit(IRQTF_WARNED, &action->thread_flags)) + return; + + printk(KERN_WARNING "IRQ %d device %s returned IRQ_WAKE_THREAD " + "but no thread function available.", irq, action->name); +} + /** * handle_IRQ_event - irq action chain handler * @irq: the interrupt number @@ -316,9 +375,50 @@ irqreturn_t handle_IRQ_event(unsigned int irq, struct irqaction *action) local_irq_enable_in_hardirq(); do { + trace_irq_handler_entry(irq, action); ret = action->handler(irq, action->dev_id); - if (ret == IRQ_HANDLED) + trace_irq_handler_exit(irq, action, ret); + + switch (ret) { + case IRQ_WAKE_THREAD: + /* + * Set result to handled so the spurious check + * does not trigger. + */ + ret = IRQ_HANDLED; + + /* + * Catch drivers which return WAKE_THREAD but + * did not set up a thread function + */ + if (unlikely(!action->thread_fn)) { + warn_no_thread(irq, action); + break; + } + + /* + * Wake up the handler thread for this + * action. In case the thread crashed and was + * killed we just pretend that we handled the + * interrupt. The hardirq handler above has + * disabled the device interrupt, so no irq + * storm is lurking. + */ + if (likely(!test_bit(IRQTF_DIED, + &action->thread_flags))) { + set_bit(IRQTF_RUNTHREAD, &action->thread_flags); + wake_up_process(action->thread); + } + + /* Fall through to add to randomness */ + case IRQ_HANDLED: status |= action->flags; + break; + + default: + break; + } + retval |= ret; action = action->next; } while (action); @@ -331,6 +431,11 @@ irqreturn_t handle_IRQ_event(unsigned int irq, struct irqaction *action) } #ifndef CONFIG_GENERIC_HARDIRQS_NO__DO_IRQ + +#ifdef CONFIG_ENABLE_WARN_DEPRECATED +# warning __do_IRQ is deprecated. Please convert to proper flow handlers +#endif + /** * __do_IRQ - original all in one highlevel IRQ handler * @irq: the interrupt number @@ -356,11 +461,8 @@ unsigned int __do_IRQ(unsigned int irq) /* * No locking required for CPU-local interrupts: */ - if (desc->chip->ack) { + if (desc->chip->ack) desc->chip->ack(irq); - /* get new one */ - desc = irq_remap_to_desc(irq, desc); - } if (likely(!(desc->status & IRQ_DISABLED))) { action_ret = handle_IRQ_event(irq, desc->action); if (!noirqdebug) @@ -371,10 +473,8 @@ unsigned int __do_IRQ(unsigned int irq) } spin_lock(&desc->lock); - if (desc->chip->ack) { + if (desc->chip->ack) desc->chip->ack(irq); - desc = irq_remap_to_desc(irq, desc); - } /* * REPLAY is when Linux resends an IRQ that was dropped earlier * WAITING is used by probe to mark irqs that are being tested @@ -451,12 +551,10 @@ void early_init_irq_lock_class(void) } } -#ifdef CONFIG_SPARSE_IRQ unsigned int kstat_irqs_cpu(unsigned int irq, int cpu) { struct irq_desc *desc = irq_to_desc(irq); return desc ? desc->kstat_irqs[cpu] : 0; } -#endif EXPORT_SYMBOL(kstat_irqs_cpu); diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index e6d0a43cc12..1b5d742c6a7 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -12,11 +12,21 @@ extern void compat_irq_chip_set_default_handler(struct irq_desc *desc); extern int __irq_set_trigger(struct irq_desc *desc, unsigned int irq, unsigned long flags); +extern void __disable_irq(struct irq_desc *desc, unsigned int irq, bool susp); +extern void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume); extern struct lock_class_key irq_desc_lock_class; -extern void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr); +extern void init_kstat_irqs(struct irq_desc *desc, int node, int nr); +extern void clear_kstat_irqs(struct irq_desc *desc); extern spinlock_t sparse_irq_lock; + +#ifdef CONFIG_SPARSE_IRQ +/* irq_desc_ptrs allocated at boot time */ +extern struct irq_desc **irq_desc_ptrs; +#else +/* irq_desc_ptrs is a fixed size array */ extern struct irq_desc *irq_desc_ptrs[NR_IRQS]; +#endif #ifdef CONFIG_PROC_FS extern void register_irq_proc(unsigned int irq, struct irq_desc *desc); @@ -32,6 +42,21 @@ static inline void unregister_handler_proc(unsigned int irq, extern int irq_select_affinity_usr(unsigned int irq); +extern void irq_set_thread_affinity(struct irq_desc *desc); + +/* Inline functions for support of irq chips on slow busses */ +static inline void chip_bus_lock(unsigned int irq, struct irq_desc *desc) +{ + if (unlikely(desc->chip->bus_lock)) + desc->chip->bus_lock(irq); +} + +static inline void chip_bus_sync_unlock(unsigned int irq, struct irq_desc *desc) +{ + if (unlikely(desc->chip->bus_sync_unlock)) + desc->chip->bus_sync_unlock(irq); +} + /* * Debugging printout: */ diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index cd0cd8dcb34..bde4c667d24 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -8,24 +8,15 @@ */ #include <linux/irq.h> +#include <linux/kthread.h> #include <linux/module.h> #include <linux/random.h> #include <linux/interrupt.h> #include <linux/slab.h> +#include <linux/sched.h> #include "internals.h" -#ifdef CONFIG_SMP -cpumask_var_t irq_default_affinity; - -static int init_irq_default_affinity(void) -{ - alloc_cpumask_var(&irq_default_affinity, GFP_KERNEL); - cpumask_setall(irq_default_affinity); - return 0; -} -core_initcall(init_irq_default_affinity); - /** * synchronize_irq - wait for pending IRQ handlers (on other CPUs) * @irq: interrupt number to wait for @@ -61,9 +52,18 @@ void synchronize_irq(unsigned int irq) /* Oops, that failed? */ } while (status & IRQ_INPROGRESS); + + /* + * We made sure that no hardirq handler is running. Now verify + * that no threaded handlers are active. + */ + wait_event(desc->wait_for_threads, !atomic_read(&desc->threads_active)); } EXPORT_SYMBOL(synchronize_irq); +#ifdef CONFIG_SMP +cpumask_var_t irq_default_affinity; + /** * irq_can_set_affinity - Check if the affinity of a given irq can be set * @irq: Interrupt to check @@ -81,6 +81,26 @@ int irq_can_set_affinity(unsigned int irq) } /** + * irq_set_thread_affinity - Notify irq threads to adjust affinity + * @desc: irq descriptor which has affitnity changed + * + * We just set IRQTF_AFFINITY and delegate the affinity setting + * to the interrupt thread itself. We can not call + * set_cpus_allowed_ptr() here as we hold desc->lock and this + * code can be called from hard interrupt context. + */ +void irq_set_thread_affinity(struct irq_desc *desc) +{ + struct irqaction *action = desc->action; + + while (action) { + if (action->thread) + set_bit(IRQTF_AFFINITY, &action->thread_flags); + action = action->next; + } +} + +/** * irq_set_affinity - Set the irq affinity of a given irq * @irq: Interrupt to set affinity * @cpumask: cpumask @@ -97,16 +117,21 @@ int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask) spin_lock_irqsave(&desc->lock, flags); #ifdef CONFIG_GENERIC_PENDING_IRQ - if (desc->status & IRQ_MOVE_PCNTXT || desc->status & IRQ_DISABLED) { - cpumask_copy(&desc->affinity, cpumask); - desc->chip->set_affinity(irq, cpumask); - } else { + if (desc->status & IRQ_MOVE_PCNTXT) { + if (!desc->chip->set_affinity(irq, cpumask)) { + cpumask_copy(desc->affinity, cpumask); + irq_set_thread_affinity(desc); + } + } + else { desc->status |= IRQ_MOVE_PENDING; - cpumask_copy(&desc->pending_mask, cpumask); + cpumask_copy(desc->pending_mask, cpumask); } #else - cpumask_copy(&desc->affinity, cpumask); - desc->chip->set_affinity(irq, cpumask); + if (!desc->chip->set_affinity(irq, cpumask)) { + cpumask_copy(desc->affinity, cpumask); + irq_set_thread_affinity(desc); + } #endif desc->status |= IRQ_AFFINITY_SET; spin_unlock_irqrestore(&desc->lock, flags); @@ -117,7 +142,7 @@ int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask) /* * Generic version of the affinity autoselector. */ -int do_irq_select_affinity(unsigned int irq, struct irq_desc *desc) +static int setup_affinity(unsigned int irq, struct irq_desc *desc) { if (!irq_can_set_affinity(irq)) return 0; @@ -127,21 +152,21 @@ int do_irq_select_affinity(unsigned int irq, struct irq_desc *desc) * one of the targets is online. */ if (desc->status & (IRQ_AFFINITY_SET | IRQ_NO_BALANCING)) { - if (cpumask_any_and(&desc->affinity, cpu_online_mask) + if (cpumask_any_and(desc->affinity, cpu_online_mask) < nr_cpu_ids) goto set_affinity; else desc->status &= ~IRQ_AFFINITY_SET; } - cpumask_and(&desc->affinity, cpu_online_mask, irq_default_affinity); + cpumask_and(desc->affinity, cpu_online_mask, irq_default_affinity); set_affinity: - desc->chip->set_affinity(irq, &desc->affinity); + desc->chip->set_affinity(irq, desc->affinity); return 0; } #else -static inline int do_irq_select_affinity(unsigned int irq, struct irq_desc *d) +static inline int setup_affinity(unsigned int irq, struct irq_desc *d) { return irq_select_affinity(irq); } @@ -157,19 +182,35 @@ int irq_select_affinity_usr(unsigned int irq) int ret; spin_lock_irqsave(&desc->lock, flags); - ret = do_irq_select_affinity(irq, desc); + ret = setup_affinity(irq, desc); + if (!ret) + irq_set_thread_affinity(desc); spin_unlock_irqrestore(&desc->lock, flags); return ret; } #else -static inline int do_irq_select_affinity(int irq, struct irq_desc *desc) +static inline int setup_affinity(unsigned int irq, struct irq_desc *desc) { return 0; } #endif +void __disable_irq(struct irq_desc *desc, unsigned int irq, bool suspend) +{ + if (suspend) { + if (!desc->action || (desc->action->flags & IRQF_TIMER)) + return; + desc->status |= IRQ_SUSPENDED; + } + + if (!desc->depth++) { + desc->status |= IRQ_DISABLED; + desc->chip->disable(irq); + } +} + /** * disable_irq_nosync - disable an irq without waiting * @irq: Interrupt to disable @@ -189,12 +230,11 @@ void disable_irq_nosync(unsigned int irq) if (!desc) return; + chip_bus_lock(irq, desc); spin_lock_irqsave(&desc->lock, flags); - if (!desc->depth++) { - desc->status |= IRQ_DISABLED; - desc->chip->disable(irq); - } + __disable_irq(desc, irq, false); spin_unlock_irqrestore(&desc->lock, flags); + chip_bus_sync_unlock(irq, desc); } EXPORT_SYMBOL(disable_irq_nosync); @@ -223,15 +263,21 @@ void disable_irq(unsigned int irq) } EXPORT_SYMBOL(disable_irq); -static void __enable_irq(struct irq_desc *desc, unsigned int irq) +void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume) { + if (resume) + desc->status &= ~IRQ_SUSPENDED; + switch (desc->depth) { case 0: + err_out: WARN(1, KERN_WARNING "Unbalanced enable for IRQ %d\n", irq); break; case 1: { unsigned int status = desc->status & ~IRQ_DISABLED; + if (desc->status & IRQ_SUSPENDED) + goto err_out; /* Prevent probing on this irq: */ desc->status = status | IRQ_NOPROBE; check_irq_resend(desc, irq); @@ -250,7 +296,8 @@ static void __enable_irq(struct irq_desc *desc, unsigned int irq) * matches the last disable, processing of interrupts on this * IRQ line is re-enabled. * - * This function may be called from IRQ context. + * This function may be called from IRQ context only when + * desc->chip->bus_lock and desc->chip->bus_sync_unlock are NULL ! */ void enable_irq(unsigned int irq) { @@ -260,9 +307,11 @@ void enable_irq(unsigned int irq) if (!desc) return; + chip_bus_lock(irq, desc); spin_lock_irqsave(&desc->lock, flags); - __enable_irq(desc, irq); + __enable_irq(desc, irq, false); spin_unlock_irqrestore(&desc->lock, flags); + chip_bus_sync_unlock(irq, desc); } EXPORT_SYMBOL(enable_irq); @@ -393,16 +442,175 @@ int __irq_set_trigger(struct irq_desc *desc, unsigned int irq, } /* + * Default primary interrupt handler for threaded interrupts. Is + * assigned as primary handler when request_threaded_irq is called + * with handler == NULL. Useful for oneshot interrupts. + */ +static irqreturn_t irq_default_primary_handler(int irq, void *dev_id) +{ + return IRQ_WAKE_THREAD; +} + +/* + * Primary handler for nested threaded interrupts. Should never be + * called. + */ +static irqreturn_t irq_nested_primary_handler(int irq, void *dev_id) +{ + WARN(1, "Primary handler called for nested irq %d\n", irq); + return IRQ_NONE; +} + +static int irq_wait_for_interrupt(struct irqaction *action) +{ + while (!kthread_should_stop()) { + set_current_state(TASK_INTERRUPTIBLE); + + if (test_and_clear_bit(IRQTF_RUNTHREAD, + &action->thread_flags)) { + __set_current_state(TASK_RUNNING); + return 0; + } + schedule(); + } + return -1; +} + +/* + * Oneshot interrupts keep the irq line masked until the threaded + * handler finished. unmask if the interrupt has not been disabled and + * is marked MASKED. + */ +static void irq_finalize_oneshot(unsigned int irq, struct irq_desc *desc) +{ + chip_bus_lock(irq, desc); + spin_lock_irq(&desc->lock); + if (!(desc->status & IRQ_DISABLED) && (desc->status & IRQ_MASKED)) { + desc->status &= ~IRQ_MASKED; + desc->chip->unmask(irq); + } + spin_unlock_irq(&desc->lock); + chip_bus_sync_unlock(irq, desc); +} + +#ifdef CONFIG_SMP +/* + * Check whether we need to change the affinity of the interrupt thread. + */ +static void +irq_thread_check_affinity(struct irq_desc *desc, struct irqaction *action) +{ + cpumask_var_t mask; + + if (!test_and_clear_bit(IRQTF_AFFINITY, &action->thread_flags)) + return; + + /* + * In case we are out of memory we set IRQTF_AFFINITY again and + * try again next time + */ + if (!alloc_cpumask_var(&mask, GFP_KERNEL)) { + set_bit(IRQTF_AFFINITY, &action->thread_flags); + return; + } + + spin_lock_irq(&desc->lock); + cpumask_copy(mask, desc->affinity); + spin_unlock_irq(&desc->lock); + + set_cpus_allowed_ptr(current, mask); + free_cpumask_var(mask); +} +#else +static inline void +irq_thread_check_affinity(struct irq_desc *desc, struct irqaction *action) { } +#endif + +/* + * Interrupt handler thread + */ +static int irq_thread(void *data) +{ + struct sched_param param = { .sched_priority = MAX_USER_RT_PRIO/2, }; + struct irqaction *action = data; + struct irq_desc *desc = irq_to_desc(action->irq); + int wake, oneshot = desc->status & IRQ_ONESHOT; + + sched_setscheduler(current, SCHED_FIFO, ¶m); + current->irqaction = action; + + while (!irq_wait_for_interrupt(action)) { + + irq_thread_check_affinity(desc, action); + + atomic_inc(&desc->threads_active); + + spin_lock_irq(&desc->lock); + if (unlikely(desc->status & IRQ_DISABLED)) { + /* + * CHECKME: We might need a dedicated + * IRQ_THREAD_PENDING flag here, which + * retriggers the thread in check_irq_resend() + * but AFAICT IRQ_PENDING should be fine as it + * retriggers the interrupt itself --- tglx + */ + desc->status |= IRQ_PENDING; + spin_unlock_irq(&desc->lock); + } else { + spin_unlock_irq(&desc->lock); + + action->thread_fn(action->irq, action->dev_id); + + if (oneshot) + irq_finalize_oneshot(action->irq, desc); + } + + wake = atomic_dec_and_test(&desc->threads_active); + + if (wake && waitqueue_active(&desc->wait_for_threads)) + wake_up(&desc->wait_for_threads); + } + + /* + * Clear irqaction. Otherwise exit_irq_thread() would make + * fuzz about an active irq thread going into nirvana. + */ + current->irqaction = NULL; + return 0; +} + +/* + * Called from do_exit() + */ +void exit_irq_thread(void) +{ + struct task_struct *tsk = current; + + if (!tsk->irqaction) + return; + + printk(KERN_ERR + "exiting task \"%s\" (%d) is an active IRQ thread (irq %d)\n", + tsk->comm ? tsk->comm : "", tsk->pid, tsk->irqaction->irq); + + /* + * Set the THREAD DIED flag to prevent further wakeups of the + * soon to be gone threaded handler. + */ + set_bit(IRQTF_DIED, &tsk->irqaction->flags); +} + +/* * Internal function to register an irqaction - typically used to * allocate special interrupts that are part of the architecture. */ static int -__setup_irq(unsigned int irq, struct irq_desc * desc, struct irqaction *new) +__setup_irq(unsigned int irq, struct irq_desc *desc, struct irqaction *new) { - struct irqaction *old, **p; + struct irqaction *old, **old_ptr; const char *old_name = NULL; unsigned long flags; - int shared = 0; + int nested, shared = 0; int ret; if (!desc) @@ -427,12 +635,53 @@ __setup_irq(unsigned int irq, struct irq_desc * desc, struct irqaction *new) rand_initialize_irq(irq); } + /* Oneshot interrupts are not allowed with shared */ + if ((new->flags & IRQF_ONESHOT) && (new->flags & IRQF_SHARED)) + return -EINVAL; + + /* + * Check whether the interrupt nests into another interrupt + * thread. + */ + nested = desc->status & IRQ_NESTED_THREAD; + if (nested) { + if (!new->thread_fn) + return -EINVAL; + /* + * Replace the primary handler which was provided from + * the driver for non nested interrupt handling by the + * dummy function which warns when called. + */ + new->handler = irq_nested_primary_handler; + } + + /* + * Create a handler thread when a thread function is supplied + * and the interrupt does not nest into another interrupt + * thread. + */ + if (new->thread_fn && !nested) { + struct task_struct *t; + + t = kthread_create(irq_thread, new, "irq/%d-%s", irq, + new->name); + if (IS_ERR(t)) + return PTR_ERR(t); + /* + * We keep the reference to the task struct even if + * the thread dies to avoid that the interrupt code + * references an already freed task_struct. + */ + get_task_struct(t); + new->thread = t; + } + /* * The following block of code has to be executed atomically */ spin_lock_irqsave(&desc->lock, flags); - p = &desc->action; - old = *p; + old_ptr = &desc->action; + old = *old_ptr; if (old) { /* * Can't share interrupts unless both agree to and are @@ -455,8 +704,8 @@ __setup_irq(unsigned int irq, struct irq_desc * desc, struct irqaction *new) /* add new interrupt at end of irq queue */ do { - p = &old->next; - old = *p; + old_ptr = &old->next; + old = *old_ptr; } while (old); shared = 1; } @@ -464,15 +713,15 @@ __setup_irq(unsigned int irq, struct irq_desc * desc, struct irqaction *new) if (!shared) { irq_chip_set_defaults(desc->chip); + init_waitqueue_head(&desc->wait_for_threads); + /* Setup the type (level, edge polarity) if configured: */ if (new->flags & IRQF_TRIGGER_MASK) { ret = __irq_set_trigger(desc, irq, new->flags & IRQF_TRIGGER_MASK); - if (ret) { - spin_unlock_irqrestore(&desc->lock, flags); - return ret; - } + if (ret) + goto out_thread; } else compat_irq_chip_set_default_handler(desc); #if defined(CONFIG_IRQ_PER_CPU) @@ -480,9 +729,12 @@ __setup_irq(unsigned int irq, struct irq_desc * desc, struct irqaction *new) desc->status |= IRQ_PER_CPU; #endif - desc->status &= ~(IRQ_AUTODETECT | IRQ_WAITING | + desc->status &= ~(IRQ_AUTODETECT | IRQ_WAITING | IRQ_ONESHOT | IRQ_INPROGRESS | IRQ_SPURIOUS_DISABLED); + if (new->flags & IRQF_ONESHOT) + desc->status |= IRQ_ONESHOT; + if (!(desc->status & IRQ_NOAUTOEN)) { desc->depth = 0; desc->status &= ~IRQ_DISABLED; @@ -496,7 +748,7 @@ __setup_irq(unsigned int irq, struct irq_desc * desc, struct irqaction *new) desc->status |= IRQ_NO_BALANCING; /* Set default affinity mask once everything is setup */ - do_irq_select_affinity(irq, desc); + setup_affinity(irq, desc); } else if ((new->flags & IRQF_TRIGGER_MASK) && (new->flags & IRQF_TRIGGER_MASK) @@ -507,7 +759,8 @@ __setup_irq(unsigned int irq, struct irq_desc * desc, struct irqaction *new) (int)(new->flags & IRQF_TRIGGER_MASK)); } - *p = new; + new->irq = irq; + *old_ptr = new; /* Reset broken irq detection when installing new handler */ desc->irq_count = 0; @@ -519,12 +772,18 @@ __setup_irq(unsigned int irq, struct irq_desc * desc, struct irqaction *new) */ if (shared && (desc->status & IRQ_SPURIOUS_DISABLED)) { desc->status &= ~IRQ_SPURIOUS_DISABLED; - __enable_irq(desc, irq); + __enable_irq(desc, irq, false); } spin_unlock_irqrestore(&desc->lock, flags); - new->irq = irq; + /* + * Strictly no need to wake it up, but hung_task complains + * when no hard interrupt wakes the thread up. + */ + if (new->thread) + wake_up_process(new->thread); + register_irq_proc(irq, desc); new->dir = NULL; register_handler_proc(irq, new); @@ -540,8 +799,19 @@ mismatch: dump_stack(); } #endif + ret = -EBUSY; + +out_thread: spin_unlock_irqrestore(&desc->lock, flags); - return -EBUSY; + if (new->thread) { + struct task_struct *t = new->thread; + + new->thread = NULL; + if (likely(!test_bit(IRQTF_DIED, &new->thread_flags))) + kthread_stop(t); + put_task_struct(t); + } + return ret; } /** @@ -557,97 +827,144 @@ int setup_irq(unsigned int irq, struct irqaction *act) return __setup_irq(irq, desc, act); } +EXPORT_SYMBOL_GPL(setup_irq); -/** - * free_irq - free an interrupt - * @irq: Interrupt line to free - * @dev_id: Device identity to free - * - * Remove an interrupt handler. The handler is removed and if the - * interrupt line is no longer in use by any driver it is disabled. - * On a shared IRQ the caller must ensure the interrupt is disabled - * on the card it drives before calling this function. The function - * does not return until any executing interrupts for this IRQ - * have completed. - * - * This function must not be called from interrupt context. + /* + * Internal function to unregister an irqaction - used to free + * regular and special interrupts that are part of the architecture. */ -void free_irq(unsigned int irq, void *dev_id) +static struct irqaction *__free_irq(unsigned int irq, void *dev_id) { struct irq_desc *desc = irq_to_desc(irq); - struct irqaction **p; + struct irqaction *action, **action_ptr; unsigned long flags; - WARN_ON(in_interrupt()); + WARN(in_interrupt(), "Trying to free IRQ %d from IRQ context!\n", irq); if (!desc) - return; + return NULL; spin_lock_irqsave(&desc->lock, flags); - p = &desc->action; + + /* + * There can be multiple actions per IRQ descriptor, find the right + * one based on the dev_id: + */ + action_ptr = &desc->action; for (;;) { - struct irqaction *action = *p; + action = *action_ptr; + + if (!action) { + WARN(1, "Trying to free already-free IRQ %d\n", irq); + spin_unlock_irqrestore(&desc->lock, flags); - if (action) { - struct irqaction **pp = p; + return NULL; + } - p = &action->next; - if (action->dev_id != dev_id) - continue; + if (action->dev_id == dev_id) + break; + action_ptr = &action->next; + } - /* Found it - now remove it from the list of entries */ - *pp = action->next; + /* Found it - now remove it from the list of entries: */ + *action_ptr = action->next; - /* Currently used only by UML, might disappear one day.*/ + /* Currently used only by UML, might disappear one day: */ #ifdef CONFIG_IRQ_RELEASE_METHOD - if (desc->chip->release) - desc->chip->release(irq, dev_id); + if (desc->chip->release) + desc->chip->release(irq, dev_id); #endif - if (!desc->action) { - desc->status |= IRQ_DISABLED; - if (desc->chip->shutdown) - desc->chip->shutdown(irq); - else - desc->chip->disable(irq); - } - spin_unlock_irqrestore(&desc->lock, flags); - unregister_handler_proc(irq, action); + /* If this was the last handler, shut down the IRQ line: */ + if (!desc->action) { + desc->status |= IRQ_DISABLED; + if (desc->chip->shutdown) + desc->chip->shutdown(irq); + else + desc->chip->disable(irq); + } + + spin_unlock_irqrestore(&desc->lock, flags); + + unregister_handler_proc(irq, action); + + /* Make sure it's not being used on another CPU: */ + synchronize_irq(irq); - /* Make sure it's not being used on another CPU */ - synchronize_irq(irq); -#ifdef CONFIG_DEBUG_SHIRQ - /* - * It's a shared IRQ -- the driver ought to be - * prepared for it to happen even now it's - * being freed, so let's make sure.... We do - * this after actually deregistering it, to - * make sure that a 'real' IRQ doesn't run in - * parallel with our fake - */ - if (action->flags & IRQF_SHARED) { - local_irq_save(flags); - action->handler(irq, dev_id); - local_irq_restore(flags); - } -#endif - kfree(action); - return; - } - printk(KERN_ERR "Trying to free already-free IRQ %d\n", irq); #ifdef CONFIG_DEBUG_SHIRQ - dump_stack(); + /* + * It's a shared IRQ -- the driver ought to be prepared for an IRQ + * event to happen even now it's being freed, so let's make sure that + * is so by doing an extra call to the handler .... + * + * ( We do this after actually deregistering it, to make sure that a + * 'real' IRQ doesn't run in * parallel with our fake. ) + */ + if (action->flags & IRQF_SHARED) { + local_irq_save(flags); + action->handler(irq, dev_id); + local_irq_restore(flags); + } #endif - spin_unlock_irqrestore(&desc->lock, flags); - return; + + if (action->thread) { + if (!test_bit(IRQTF_DIED, &action->thread_flags)) + kthread_stop(action->thread); + put_task_struct(action->thread); } + + return action; +} + +/** + * remove_irq - free an interrupt + * @irq: Interrupt line to free + * @act: irqaction for the interrupt + * + * Used to remove interrupts statically setup by the early boot process. + */ +void remove_irq(unsigned int irq, struct irqaction *act) +{ + __free_irq(irq, act->dev_id); +} +EXPORT_SYMBOL_GPL(remove_irq); + +/** + * free_irq - free an interrupt allocated with request_irq + * @irq: Interrupt line to free + * @dev_id: Device identity to free + * + * Remove an interrupt handler. The handler is removed and if the + * interrupt line is no longer in use by any driver it is disabled. + * On a shared IRQ the caller must ensure the interrupt is disabled + * on the card it drives before calling this function. The function + * does not return until any executing interrupts for this IRQ + * have completed. + * + * This function must not be called from interrupt context. + */ +void free_irq(unsigned int irq, void *dev_id) +{ + struct irq_desc *desc = irq_to_desc(irq); + + if (!desc) + return; + + chip_bus_lock(irq, desc); + kfree(__free_irq(irq, dev_id)); + chip_bus_sync_unlock(irq, desc); } EXPORT_SYMBOL(free_irq); /** - * request_irq - allocate an interrupt line + * request_threaded_irq - allocate an interrupt line * @irq: Interrupt line to allocate - * @handler: Function to be called when the IRQ occurs + * @handler: Function to be called when the IRQ occurs. + * Primary handler for threaded interrupts + * If NULL and thread_fn != NULL the default + * primary handler is installed + * @thread_fn: Function called from the irq handler thread + * If NULL, no irq thread is created * @irqflags: Interrupt type flags * @devname: An ascii name for the claiming device * @dev_id: A cookie passed back to the handler function @@ -659,6 +976,15 @@ EXPORT_SYMBOL(free_irq); * raises, you must take care both to initialise your hardware * and to set up the interrupt handler in the right order. * + * If you want to set up a threaded irq handler for your device + * then you need to supply @handler and @thread_fn. @handler ist + * still called in hard interrupt context and has to check + * whether the interrupt originates from the device. If yes it + * needs to disable the interrupt on the device and return + * IRQ_WAKE_THREAD which will wake up the handler thread and run + * @thread_fn. This split handler design is necessary to support + * shared interrupts. + * * Dev_id must be globally unique. Normally the address of the * device data structure is used as the cookie. Since the handler * receives this value it makes sense to use it. @@ -674,8 +1000,9 @@ EXPORT_SYMBOL(free_irq); * IRQF_TRIGGER_* Specify active edge(s) or level * */ -int request_irq(unsigned int irq, irq_handler_t handler, - unsigned long irqflags, const char *devname, void *dev_id) +int request_threaded_irq(unsigned int irq, irq_handler_t handler, + irq_handler_t thread_fn, unsigned long irqflags, + const char *devname, void *dev_id) { struct irqaction *action; struct irq_desc *desc; @@ -687,11 +1014,12 @@ int request_irq(unsigned int irq, irq_handler_t handler, * the behavior is classified as "will not fix" so we need to * start nudging drivers away from using that idiom. */ - if ((irqflags & (IRQF_SHARED|IRQF_DISABLED)) - == (IRQF_SHARED|IRQF_DISABLED)) - pr_warning("IRQ %d/%s: IRQF_DISABLED is not " - "guaranteed on shared IRQs\n", - irq, devname); + if ((irqflags & (IRQF_SHARED|IRQF_DISABLED)) == + (IRQF_SHARED|IRQF_DISABLED)) { + pr_warning( + "IRQ %d/%s: IRQF_DISABLED is not guaranteed on shared IRQs\n", + irq, devname); + } #ifdef CONFIG_LOCKDEP /* @@ -714,21 +1042,27 @@ int request_irq(unsigned int irq, irq_handler_t handler, if (desc->status & IRQ_NOREQUEST) return -EINVAL; - if (!handler) - return -EINVAL; - action = kmalloc(sizeof(struct irqaction), GFP_ATOMIC); + if (!handler) { + if (!thread_fn) + return -EINVAL; + handler = irq_default_primary_handler; + } + + action = kzalloc(sizeof(struct irqaction), GFP_KERNEL); if (!action) return -ENOMEM; action->handler = handler; + action->thread_fn = thread_fn; action->flags = irqflags; - cpus_clear(action->mask); action->name = devname; - action->next = NULL; action->dev_id = dev_id; + chip_bus_lock(irq, desc); retval = __setup_irq(irq, desc, action); + chip_bus_sync_unlock(irq, desc); + if (retval) kfree(action); @@ -753,4 +1087,4 @@ int request_irq(unsigned int irq, irq_handler_t handler, #endif return retval; } -EXPORT_SYMBOL(request_irq); +EXPORT_SYMBOL(request_threaded_irq); diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index bd72329e630..fcb6c96f262 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -1,5 +1,8 @@ #include <linux/irq.h> +#include <linux/interrupt.h> + +#include "internals.h" void move_masked_irq(int irq) { @@ -18,7 +21,7 @@ void move_masked_irq(int irq) desc->status &= ~IRQ_MOVE_PENDING; - if (unlikely(cpumask_empty(&desc->pending_mask))) + if (unlikely(cpumask_empty(desc->pending_mask))) return; if (!desc->chip->set_affinity) @@ -38,13 +41,14 @@ void move_masked_irq(int irq) * For correct operation this depends on the caller * masking the irqs. */ - if (likely(cpumask_any_and(&desc->pending_mask, cpu_online_mask) - < nr_cpu_ids)) { - cpumask_and(&desc->affinity, - &desc->pending_mask, cpu_online_mask); - desc->chip->set_affinity(irq, &desc->affinity); - } - cpumask_clear(&desc->pending_mask); + if (likely(cpumask_any_and(desc->pending_mask, cpu_online_mask) + < nr_cpu_ids)) + if (!desc->chip->set_affinity(irq, desc->pending_mask)) { + cpumask_copy(desc->affinity, desc->pending_mask); + irq_set_thread_affinity(desc); + } + + cpumask_clear(desc->pending_mask); } void move_native_irq(int irq) diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index ecf765c6a77..3fd30197da2 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -15,18 +15,13 @@ static void init_copy_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc, - int cpu, int nr) + int node, int nr) { - unsigned long bytes; + init_kstat_irqs(desc, node, nr); - init_kstat_irqs(desc, cpu, nr); - - if (desc->kstat_irqs != old_desc->kstat_irqs) { - /* Compute how many bytes we need per irq and allocate them */ - bytes = nr * sizeof(unsigned int); - - memcpy(desc->kstat_irqs, old_desc->kstat_irqs, bytes); - } + if (desc->kstat_irqs != old_desc->kstat_irqs) + memcpy(desc->kstat_irqs, old_desc->kstat_irqs, + nr * sizeof(*desc->kstat_irqs)); } static void free_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc) @@ -38,30 +33,37 @@ static void free_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc) old_desc->kstat_irqs = NULL; } -static void init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, - struct irq_desc *desc, int cpu) +static bool init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, + struct irq_desc *desc, int node) { memcpy(desc, old_desc, sizeof(struct irq_desc)); + if (!alloc_desc_masks(desc, node, false)) { + printk(KERN_ERR "irq %d: can not get new irq_desc cpumask " + "for migration.\n", irq); + return false; + } spin_lock_init(&desc->lock); - desc->cpu = cpu; + desc->node = node; lockdep_set_class(&desc->lock, &irq_desc_lock_class); - init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids); - arch_init_copy_chip_data(old_desc, desc, cpu); + init_copy_kstat_irqs(old_desc, desc, node, nr_cpu_ids); + init_copy_desc_masks(old_desc, desc); + arch_init_copy_chip_data(old_desc, desc, node); + return true; } static void free_one_irq_desc(struct irq_desc *old_desc, struct irq_desc *desc) { free_kstat_irqs(old_desc, desc); + free_desc_masks(old_desc, desc); arch_free_chip_data(old_desc, desc); } static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, - int cpu) + int node) { struct irq_desc *desc; unsigned int irq; unsigned long flags; - int node; irq = old_desc->irq; @@ -71,48 +73,46 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, desc = irq_desc_ptrs[irq]; if (desc && old_desc != desc) - goto out_unlock; + goto out_unlock; - node = cpu_to_node(cpu); desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); if (!desc) { - printk(KERN_ERR "irq %d: can not get new irq_desc for migration.\n", irq); + printk(KERN_ERR "irq %d: can not get new irq_desc " + "for migration.\n", irq); + /* still use old one */ + desc = old_desc; + goto out_unlock; + } + if (!init_copy_one_irq_desc(irq, old_desc, desc, node)) { /* still use old one */ + kfree(desc); desc = old_desc; goto out_unlock; } - init_copy_one_irq_desc(irq, old_desc, desc, cpu); irq_desc_ptrs[irq] = desc; + spin_unlock_irqrestore(&sparse_irq_lock, flags); /* free the old one */ free_one_irq_desc(old_desc, desc); kfree(old_desc); + return desc; + out_unlock: spin_unlock_irqrestore(&sparse_irq_lock, flags); return desc; } -struct irq_desc *move_irq_desc(struct irq_desc *desc, int cpu) +struct irq_desc *move_irq_desc(struct irq_desc *desc, int node) { - int old_cpu; - int node, old_node; - - /* those all static, do move them */ - if (desc->irq < NR_IRQS_LEGACY) + /* those static or target node is -1, do not move them */ + if (desc->irq < NR_IRQS_LEGACY || node == -1) return desc; - old_cpu = desc->cpu; - if (old_cpu != cpu) { - node = cpu_to_node(cpu); - old_node = cpu_to_node(old_cpu); - if (old_node != node) - desc = __real_move_irq_desc(desc, cpu); - else - desc->cpu = cpu; - } + if (desc->node != node) + desc = __real_move_irq_desc(desc, node); return desc; } diff --git a/kernel/irq/pm.c b/kernel/irq/pm.c new file mode 100644 index 00000000000..a0bb09e7986 --- /dev/null +++ b/kernel/irq/pm.c @@ -0,0 +1,79 @@ +/* + * linux/kernel/irq/pm.c + * + * Copyright (C) 2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc. + * + * This file contains power management functions related to interrupts. + */ + +#include <linux/irq.h> +#include <linux/module.h> +#include <linux/interrupt.h> + +#include "internals.h" + +/** + * suspend_device_irqs - disable all currently enabled interrupt lines + * + * During system-wide suspend or hibernation device drivers need to be prevented + * from receiving interrupts and this function is provided for this purpose. + * It marks all interrupt lines in use, except for the timer ones, as disabled + * and sets the IRQ_SUSPENDED flag for each of them. + */ +void suspend_device_irqs(void) +{ + struct irq_desc *desc; + int irq; + + for_each_irq_desc(irq, desc) { + unsigned long flags; + + spin_lock_irqsave(&desc->lock, flags); + __disable_irq(desc, irq, true); + spin_unlock_irqrestore(&desc->lock, flags); + } + + for_each_irq_desc(irq, desc) + if (desc->status & IRQ_SUSPENDED) + synchronize_irq(irq); +} +EXPORT_SYMBOL_GPL(suspend_device_irqs); + +/** + * resume_device_irqs - enable interrupt lines disabled by suspend_device_irqs() + * + * Enable all interrupt lines previously disabled by suspend_device_irqs() that + * have the IRQ_SUSPENDED flag set. + */ +void resume_device_irqs(void) +{ + struct irq_desc *desc; + int irq; + + for_each_irq_desc(irq, desc) { + unsigned long flags; + + if (!(desc->status & IRQ_SUSPENDED)) + continue; + + spin_lock_irqsave(&desc->lock, flags); + __enable_irq(desc, irq, true); + spin_unlock_irqrestore(&desc->lock, flags); + } +} +EXPORT_SYMBOL_GPL(resume_device_irqs); + +/** + * check_wakeup_irqs - check if any wake-up interrupts are pending + */ +int check_wakeup_irqs(void) +{ + struct irq_desc *desc; + int irq; + + for_each_irq_desc(irq, desc) + if ((desc->status & IRQ_WAKEUP) && (desc->status & IRQ_PENDING)) + return -EBUSY; + + return 0; +} diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c index aae3f742bce..692363dd591 100644 --- a/kernel/irq/proc.c +++ b/kernel/irq/proc.c @@ -20,11 +20,11 @@ static struct proc_dir_entry *root_irq_dir; static int irq_affinity_proc_show(struct seq_file *m, void *v) { struct irq_desc *desc = irq_to_desc((long)m->private); - const struct cpumask *mask = &desc->affinity; + const struct cpumask *mask = desc->affinity; #ifdef CONFIG_GENERIC_PENDING_IRQ if (desc->status & IRQ_MOVE_PENDING) - mask = &desc->pending_mask; + mask = desc->pending_mask; #endif seq_cpumask(m, mask); seq_putc(m, '\n'); diff --git a/kernel/irq/resend.c b/kernel/irq/resend.c index 89c7117acf2..090c3763f3a 100644 --- a/kernel/irq/resend.c +++ b/kernel/irq/resend.c @@ -70,8 +70,7 @@ void check_irq_resend(struct irq_desc *desc, unsigned int irq) if ((status & (IRQ_LEVEL | IRQ_PENDING | IRQ_REPLAY)) == IRQ_PENDING) { desc->status = (status & ~IRQ_PENDING) | IRQ_REPLAY; - if (!desc->chip || !desc->chip->retrigger || - !desc->chip->retrigger(irq)) { + if (!desc->chip->retrigger || !desc->chip->retrigger(irq)) { #ifdef CONFIG_HARDIRQS_SW_RESEND /* Set it pending and activate the softirq: */ set_bit(irq, irqs_resend); diff --git a/kernel/irq/spurious.c b/kernel/irq/spurious.c index dd364c11e56..114e704760f 100644 --- a/kernel/irq/spurious.c +++ b/kernel/irq/spurious.c @@ -104,7 +104,7 @@ static int misrouted_irq(int irq) return ok; } -static void poll_spurious_irqs(unsigned long dummy) +static void poll_all_shared_irqs(void) { struct irq_desc *desc; int i; @@ -123,11 +123,23 @@ static void poll_spurious_irqs(unsigned long dummy) try_one_irq(i, desc); } +} + +static void poll_spurious_irqs(unsigned long dummy) +{ + poll_all_shared_irqs(); mod_timer(&poll_spurious_irq_timer, jiffies + POLL_SPURIOUS_IRQ_INTERVAL); } +#ifdef CONFIG_DEBUG_SHIRQ +void debug_poll_all_shared_irqs(void) +{ + poll_all_shared_irqs(); +} +#endif + /* * If 99,900 of the previous 100,000 interrupts have not been handled * then assume that the IRQ is stuck in some manner. Drop a diagnostic @@ -285,7 +297,6 @@ static int __init irqfixup_setup(char *str) __setup("irqfixup", irqfixup_setup); module_param(irqfixup, int, 0644); -MODULE_PARM_DESC("irqfixup", "0: No fixup, 1: irqfixup mode, 2: irqpoll mode"); static int __init irqpoll_setup(char *str) { diff --git a/kernel/itimer.c b/kernel/itimer.c index 6a5fe93dd8b..58762f7077e 100644 --- a/kernel/itimer.c +++ b/kernel/itimer.c @@ -62,7 +62,7 @@ int do_getitimer(int which, struct itimerval *value) struct task_cputime cputime; cputime_t utime; - thread_group_cputime(tsk, &cputime); + thread_group_cputimer(tsk, &cputime); utime = cputime.utime; if (cputime_le(cval, utime)) { /* about to fire */ cval = jiffies_to_cputime(1); @@ -82,7 +82,7 @@ int do_getitimer(int which, struct itimerval *value) struct task_cputime times; cputime_t ptime; - thread_group_cputime(tsk, ×); + thread_group_cputimer(tsk, ×); ptime = cputime_add(times.utime, times.stime); if (cputime_le(cval, ptime)) { /* about to fire */ cval = jiffies_to_cputime(1); diff --git a/kernel/kallsyms.c b/kernel/kallsyms.c index e694afa0eb8..3a29dbe7898 100644 --- a/kernel/kallsyms.c +++ b/kernel/kallsyms.c @@ -30,19 +30,24 @@ #define all_var 0 #endif -extern const unsigned long kallsyms_addresses[]; -extern const u8 kallsyms_names[]; +/* + * These will be re-linked against their real values + * during the second link stage. + */ +extern const unsigned long kallsyms_addresses[] __attribute__((weak)); +extern const u8 kallsyms_names[] __attribute__((weak)); -/* tell the compiler that the count isn't in the small data section if the arch - * has one (eg: FRV) +/* + * Tell the compiler that the count isn't in the small data section if the arch + * has one (eg: FRV). */ extern const unsigned long kallsyms_num_syms - __attribute__((__section__(".rodata"))); +__attribute__((weak, section(".rodata"))); -extern const u8 kallsyms_token_table[]; -extern const u16 kallsyms_token_index[]; +extern const u8 kallsyms_token_table[] __attribute__((weak)); +extern const u16 kallsyms_token_index[] __attribute__((weak)); -extern const unsigned long kallsyms_markers[]; +extern const unsigned long kallsyms_markers[] __attribute__((weak)); static inline int is_kernel_inittext(unsigned long addr) { @@ -74,31 +79,37 @@ static int is_ksym_addr(unsigned long addr) return is_kernel_text(addr) || is_kernel_inittext(addr); } -/* expand a compressed symbol data into the resulting uncompressed string, - given the offset to where the symbol is in the compressed stream */ +/* + * Expand a compressed symbol data into the resulting uncompressed string, + * given the offset to where the symbol is in the compressed stream. + */ static unsigned int kallsyms_expand_symbol(unsigned int off, char *result) { int len, skipped_first = 0; const u8 *tptr, *data; - /* get the compressed symbol length from the first symbol byte */ + /* Get the compressed symbol length from the first symbol byte. */ data = &kallsyms_names[off]; len = *data; data++; - /* update the offset to return the offset for the next symbol on - * the compressed stream */ + /* + * Update the offset to return the offset for the next symbol on + * the compressed stream. + */ off += len + 1; - /* for every byte on the compressed symbol data, copy the table - entry for that byte */ - while(len) { - tptr = &kallsyms_token_table[ kallsyms_token_index[*data] ]; + /* + * For every byte on the compressed symbol data, copy the table + * entry for that byte. + */ + while (len) { + tptr = &kallsyms_token_table[kallsyms_token_index[*data]]; data++; len--; while (*tptr) { - if(skipped_first) { + if (skipped_first) { *result = *tptr; result++; } else @@ -109,36 +120,46 @@ static unsigned int kallsyms_expand_symbol(unsigned int off, char *result) *result = '\0'; - /* return to offset to the next symbol */ + /* Return to offset to the next symbol. */ return off; } -/* get symbol type information. This is encoded as a single char at the - * begining of the symbol name */ +/* + * Get symbol type information. This is encoded as a single char at the + * beginning of the symbol name. + */ static char kallsyms_get_symbol_type(unsigned int off) { - /* get just the first code, look it up in the token table, and return the - * first char from this token */ - return kallsyms_token_table[ kallsyms_token_index[ kallsyms_names[off+1] ] ]; + /* + * Get just the first code, look it up in the token table, + * and return the first char from this token. + */ + return kallsyms_token_table[kallsyms_token_index[kallsyms_names[off + 1]]]; } -/* find the offset on the compressed stream given and index in the - * kallsyms array */ +/* + * Find the offset on the compressed stream given and index in the + * kallsyms array. + */ static unsigned int get_symbol_offset(unsigned long pos) { const u8 *name; int i; - /* use the closest marker we have. We have markers every 256 positions, - * so that should be close enough */ - name = &kallsyms_names[ kallsyms_markers[pos>>8] ]; + /* + * Use the closest marker we have. We have markers every 256 positions, + * so that should be close enough. + */ + name = &kallsyms_names[kallsyms_markers[pos >> 8]]; - /* sequentially scan all the symbols up to the point we're searching for. - * Every symbol is stored in a [<len>][<len> bytes of data] format, so we - * just need to add the len to the current pointer for every symbol we - * wish to skip */ - for(i = 0; i < (pos&0xFF); i++) + /* + * Sequentially scan all the symbols up to the point we're searching + * for. Every symbol is stored in a [<len>][<len> bytes of data] format, + * so we just need to add the len to the current pointer for every + * symbol we wish to skip. + */ + for (i = 0; i < (pos & 0xFF); i++) name = name + (*name) + 1; return name - kallsyms_names; @@ -160,6 +181,25 @@ unsigned long kallsyms_lookup_name(const char *name) return module_kallsyms_lookup_name(name); } +int kallsyms_on_each_symbol(int (*fn)(void *, const char *, struct module *, + unsigned long), + void *data) +{ + char namebuf[KSYM_NAME_LEN]; + unsigned long i; + unsigned int off; + int ret; + + for (i = 0, off = 0; i < kallsyms_num_syms; i++) { + off = kallsyms_expand_symbol(off, namebuf); + ret = fn(data, namebuf, NULL, kallsyms_addresses[i]); + if (ret != 0) + return ret; + } + return module_kallsyms_on_each_symbol(fn, data); +} +EXPORT_SYMBOL_GPL(kallsyms_on_each_symbol); + static unsigned long get_symbol_pos(unsigned long addr, unsigned long *symbolsize, unsigned long *offset) @@ -167,7 +207,10 @@ static unsigned long get_symbol_pos(unsigned long addr, unsigned long symbol_start = 0, symbol_end = 0; unsigned long i, low, high, mid; - /* do a binary search on the sorted kallsyms_addresses array */ + /* This kernel should never had been booted. */ + BUG_ON(!kallsyms_addresses); + + /* Do a binary search on the sorted kallsyms_addresses array. */ low = 0; high = kallsyms_num_syms; @@ -180,15 +223,15 @@ static unsigned long get_symbol_pos(unsigned long addr, } /* - * search for the first aliased symbol. Aliased - * symbols are symbols with the same address + * Search for the first aliased symbol. Aliased + * symbols are symbols with the same address. */ while (low && kallsyms_addresses[low-1] == kallsyms_addresses[low]) --low; symbol_start = kallsyms_addresses[low]; - /* Search for next non-aliased symbol */ + /* Search for next non-aliased symbol. */ for (i = low + 1; i < kallsyms_num_syms; i++) { if (kallsyms_addresses[i] > symbol_start) { symbol_end = kallsyms_addresses[i]; @@ -196,7 +239,7 @@ static unsigned long get_symbol_pos(unsigned long addr, } } - /* if we found no next symbol, we use the end of the section */ + /* If we found no next symbol, we use the end of the section. */ if (!symbol_end) { if (is_kernel_inittext(addr)) symbol_end = (unsigned long)_einittext; @@ -229,10 +272,10 @@ int kallsyms_lookup_size_offset(unsigned long addr, unsigned long *symbolsize, /* * Lookup an address - * - modname is set to NULL if it's in the kernel - * - we guarantee that the returned name is valid until we reschedule even if - * it resides in a module - * - we also guarantee that modname will be valid until rescheduled + * - modname is set to NULL if it's in the kernel. + * - We guarantee that the returned name is valid until we reschedule even if. + * It resides in a module. + * - We also guarantee that modname will be valid until rescheduled. */ const char *kallsyms_lookup(unsigned long addr, unsigned long *symbolsize, @@ -253,7 +296,7 @@ const char *kallsyms_lookup(unsigned long addr, return namebuf; } - /* see if it's in a module */ + /* See if it's in a module. */ return module_address_lookup(addr, symbolsize, offset, modname, namebuf); } @@ -271,7 +314,7 @@ int lookup_symbol_name(unsigned long addr, char *symname) kallsyms_expand_symbol(get_symbol_offset(pos), symname); return 0; } - /* see if it's in a module */ + /* See if it's in a module. */ return lookup_module_symbol_name(addr, symname); } @@ -290,7 +333,7 @@ int lookup_symbol_attrs(unsigned long addr, unsigned long *size, modname[0] = '\0'; return 0; } - /* see if it's in a module */ + /* See if it's in a module. */ return lookup_module_symbol_attrs(addr, size, offset, modname, name); } @@ -319,6 +362,7 @@ int sprint_symbol(char *buffer, unsigned long address) return len; } +EXPORT_SYMBOL_GPL(sprint_symbol); /* Look up a kernel symbol and print it to the kernel messages. */ void __print_symbol(const char *fmt, unsigned long address) @@ -329,13 +373,13 @@ void __print_symbol(const char *fmt, unsigned long address) printk(fmt, buffer); } +EXPORT_SYMBOL(__print_symbol); /* To avoid using get_symbol_offset for every symbol, we carry prefix along. */ -struct kallsym_iter -{ +struct kallsym_iter { loff_t pos; unsigned long value; - unsigned int nameoff; /* If iterating in core kernel symbols */ + unsigned int nameoff; /* If iterating in core kernel symbols. */ char type; char name[KSYM_NAME_LEN]; char module_name[MODULE_NAME_LEN]; @@ -381,7 +425,7 @@ static int update_iter(struct kallsym_iter *iter, loff_t pos) iter->pos = pos; return get_ksymbol_mod(iter); } - + /* If we're not on the desired position, reset to new position. */ if (pos != iter->pos) reset_iter(iter, pos); @@ -416,23 +460,25 @@ static int s_show(struct seq_file *m, void *p) { struct kallsym_iter *iter = m->private; - /* Some debugging symbols have no name. Ignore them. */ + /* Some debugging symbols have no name. Ignore them. */ if (!iter->name[0]) return 0; if (iter->module_name[0]) { char type; - /* Label it "global" if it is exported, - * "local" if not exported. */ + /* + * Label it "global" if it is exported, + * "local" if not exported. + */ type = iter->exported ? toupper(iter->type) : tolower(iter->type); seq_printf(m, "%0*lx %c %s\t[%s]\n", - (int)(2*sizeof(void*)), + (int)(2 * sizeof(void *)), iter->value, type, iter->name, iter->module_name); } else seq_printf(m, "%0*lx %c %s\n", - (int)(2*sizeof(void*)), + (int)(2 * sizeof(void *)), iter->value, iter->type, iter->name); return 0; } @@ -446,9 +492,11 @@ static const struct seq_operations kallsyms_op = { static int kallsyms_open(struct inode *inode, struct file *file) { - /* We keep iterator in m->private, since normal case is to + /* + * We keep iterator in m->private, since normal case is to * s_start from where we left off, so we avoid doing - * using get_symbol_offset for every symbol */ + * using get_symbol_offset for every symbol. + */ struct kallsym_iter *iter; int ret; @@ -477,7 +525,4 @@ static int __init kallsyms_init(void) proc_create("kallsyms", 0444, NULL, &kallsyms_operations); return 0; } -__initcall(kallsyms_init); - -EXPORT_SYMBOL(__print_symbol); -EXPORT_SYMBOL_GPL(sprint_symbol); +device_initcall(kallsyms_init); diff --git a/kernel/kexec.c b/kernel/kexec.c index 8a6d7b08864..f336e2107f9 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -42,7 +42,7 @@ note_buf_t* crash_notes; /* vmcoreinfo stuff */ -unsigned char vmcoreinfo_data[VMCOREINFO_BYTES]; +static unsigned char vmcoreinfo_data[VMCOREINFO_BYTES]; u32 vmcoreinfo_note[VMCOREINFO_NOTE_SIZE/4]; size_t vmcoreinfo_size; size_t vmcoreinfo_max_size = sizeof(vmcoreinfo_data); @@ -1130,7 +1130,7 @@ void crash_save_cpu(struct pt_regs *regs, int cpu) return; memset(&prstatus, 0, sizeof(prstatus)); prstatus.pr_pid = current->pid; - elf_core_copy_regs(&prstatus.pr_reg, regs); + elf_core_copy_kernel_regs(&prstatus.pr_reg, regs); buf = append_elf_note(buf, KEXEC_CORE_NOTE_NAME, NT_PRSTATUS, &prstatus, sizeof(prstatus)); final_note(buf); @@ -1228,7 +1228,7 @@ static int __init parse_crashkernel_mem(char *cmdline, } while (*cur++ == ','); if (*crash_size > 0) { - while (*cur != ' ' && *cur != '@') + while (*cur && *cur != ' ' && *cur != '@') cur++; if (*cur == '@') { cur++; @@ -1409,6 +1409,7 @@ static int __init crash_save_vmcoreinfo_init(void) VMCOREINFO_OFFSET(list_head, prev); VMCOREINFO_OFFSET(vm_struct, addr); VMCOREINFO_LENGTH(zone.free_area, MAX_ORDER); + log_buf_kexec_setup(); VMCOREINFO_LENGTH(free_area.free_list, MIGRATE_TYPES); VMCOREINFO_NUMBER(NR_FREE_PAGES); VMCOREINFO_NUMBER(PG_lru); @@ -1447,22 +1448,25 @@ int kernel_kexec(void) goto Restore_console; } suspend_console(); - error = device_suspend(PMSG_FREEZE); + error = dpm_suspend_start(PMSG_FREEZE); if (error) goto Resume_console; - error = disable_nonboot_cpus(); - if (error) - goto Resume_devices; - device_pm_lock(); - local_irq_disable(); - /* At this point, device_suspend() has been called, - * but *not* device_power_down(). We *must* - * device_power_down() now. Otherwise, drivers for + /* At this point, dpm_suspend_start() has been called, + * but *not* dpm_suspend_noirq(). We *must* call + * dpm_suspend_noirq() now. Otherwise, drivers for * some devices (e.g. interrupt controllers) become * desynchronized with the actual state of the * hardware at resume time, and evil weirdness ensues. */ - error = device_power_down(PMSG_FREEZE); + error = dpm_suspend_noirq(PMSG_FREEZE); + if (error) + goto Resume_devices; + error = disable_nonboot_cpus(); + if (error) + goto Enable_cpus; + local_irq_disable(); + /* Suspend system devices */ + error = sysdev_suspend(PMSG_FREEZE); if (error) goto Enable_irqs; } else @@ -1477,13 +1481,14 @@ int kernel_kexec(void) #ifdef CONFIG_KEXEC_JUMP if (kexec_image->preserve_context) { - device_power_up(PMSG_RESTORE); + sysdev_resume(); Enable_irqs: local_irq_enable(); - device_pm_unlock(); + Enable_cpus: enable_nonboot_cpus(); + dpm_resume_noirq(PMSG_RESTORE); Resume_devices: - device_resume(PMSG_RESTORE); + dpm_resume_end(PMSG_RESTORE); Resume_console: resume_console(); thaw_processes(); diff --git a/kernel/kfifo.c b/kernel/kfifo.c index bc41ad0f24f..26539e3228e 100644 --- a/kernel/kfifo.c +++ b/kernel/kfifo.c @@ -72,9 +72,9 @@ struct kfifo *kfifo_alloc(unsigned int size, gfp_t gfp_mask, spinlock_t *lock) /* * round up to the next power of 2, since our 'let the indices - * wrap' tachnique works only in this case. + * wrap' technique works only in this case. */ - if (size & (size - 1)) { + if (!is_power_of_2(size)) { BUG_ON(size > 0x80000000); size = roundup_pow_of_two(size); } diff --git a/kernel/kgdb.c b/kernel/kgdb.c index e4dcfb2272a..9147a3190c9 100644 --- a/kernel/kgdb.c +++ b/kernel/kgdb.c @@ -1583,8 +1583,8 @@ static void sysrq_handle_gdb(int key, struct tty_struct *tty) static struct sysrq_key_op sysrq_gdb_op = { .handler = sysrq_handle_gdb, - .help_msg = "Gdb", - .action_msg = "GDB", + .help_msg = "debug(G)", + .action_msg = "DEBUG", }; #endif diff --git a/kernel/kmod.c b/kernel/kmod.c index a27a5f64443..9fcb53a11f8 100644 --- a/kernel/kmod.c +++ b/kernel/kmod.c @@ -24,7 +24,6 @@ #include <linux/unistd.h> #include <linux/kmod.h> #include <linux/slab.h> -#include <linux/mnt_namespace.h> #include <linux/completion.h> #include <linux/file.h> #include <linux/fdtable.h> @@ -38,6 +37,8 @@ #include <linux/suspend.h> #include <asm/uaccess.h> +#include <trace/events/module.h> + extern int max_threads; static struct workqueue_struct *khelper_wq; @@ -50,7 +51,8 @@ static struct workqueue_struct *khelper_wq; char modprobe_path[KMOD_PATH_LEN] = "/sbin/modprobe"; /** - * request_module - try to load a kernel module + * __request_module - try to load a kernel module + * @wait: wait (or not) for the operation to complete * @fmt: printf style format string for the name of the module * @...: arguments as specified in the format string * @@ -63,7 +65,7 @@ char modprobe_path[KMOD_PATH_LEN] = "/sbin/modprobe"; * If module auto-loading support is disabled then this function * becomes a no-operation. */ -int request_module(const char *fmt, ...) +int __request_module(bool wait, const char *fmt, ...) { va_list args; char module_name[MODULE_NAME_LEN]; @@ -78,6 +80,10 @@ int request_module(const char *fmt, ...) #define MAX_KMOD_CONCURRENT 50 /* Completely arbitrary value - KAO */ static int kmod_loop_msg; + ret = security_kernel_module_request(); + if (ret) + return ret; + va_start(args, fmt); ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args); va_end(args); @@ -108,11 +114,14 @@ int request_module(const char *fmt, ...) return -ENOMEM; } - ret = call_usermodehelper(modprobe_path, argv, envp, 1); + trace_module_request(module_name, wait, _RET_IP_); + + ret = call_usermodehelper(modprobe_path, argv, envp, + wait ? UMH_WAIT_PROC : UMH_WAIT_EXEC); atomic_dec(&kmod_concurrent); return ret; } -EXPORT_SYMBOL(request_module); +EXPORT_SYMBOL(__request_module); #endif /* CONFIG_MODULES */ struct subprocess_info { @@ -167,7 +176,7 @@ static int ____call_usermodehelper(void *data) } /* We can run anywhere, unlike our parent keventd(). */ - set_cpus_allowed_ptr(current, CPU_MASK_ALL_PTR); + set_cpus_allowed_ptr(current, cpu_all_mask); /* * Our parent is keventd, which runs with elevated scheduling priority. @@ -368,8 +377,10 @@ struct subprocess_info *call_usermodehelper_setup(char *path, char **argv, sub_info->argv = argv; sub_info->envp = envp; sub_info->cred = prepare_usermodehelper_creds(); - if (!sub_info->cred) + if (!sub_info->cred) { + kfree(sub_info); return NULL; + } out: return sub_info; @@ -459,6 +470,7 @@ int call_usermodehelper_exec(struct subprocess_info *sub_info, int retval = 0; BUG_ON(atomic_read(&sub_info->cred->usage) != 1); + validate_creds(sub_info->cred); helper_lock(); if (sub_info->path[0] == '\0') diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 7ba8cd9845c..ef177d653b2 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -43,6 +43,7 @@ #include <linux/seq_file.h> #include <linux/debugfs.h> #include <linux/kdebug.h> +#include <linux/memory.h> #include <asm-generic/sections.h> #include <asm/cacheflush.h> @@ -67,7 +68,7 @@ static struct hlist_head kprobe_table[KPROBE_TABLE_SIZE]; static struct hlist_head kretprobe_inst_table[KPROBE_TABLE_SIZE]; /* NOTE: change this value only with kprobe_mutex held */ -static bool kprobe_enabled; +static bool kprobes_all_disarmed; static DEFINE_MUTEX(kprobe_mutex); /* Protects kprobe_table */ static DEFINE_PER_CPU(struct kprobe *, kprobe_instance) = NULL; @@ -102,7 +103,7 @@ static struct kprobe_blackpoint kprobe_blacklist[] = { #define INSNS_PER_PAGE (PAGE_SIZE/(MAX_INSN_SIZE * sizeof(kprobe_opcode_t))) struct kprobe_insn_page { - struct hlist_node hlist; + struct list_head list; kprobe_opcode_t *insns; /* Page of instruction slots */ char slot_used[INSNS_PER_PAGE]; int nused; @@ -116,7 +117,7 @@ enum kprobe_slot_state { }; static DEFINE_MUTEX(kprobe_insn_mutex); /* Protects kprobe_insn_pages */ -static struct hlist_head kprobe_insn_pages; +static LIST_HEAD(kprobe_insn_pages); static int kprobe_garbage_slots; static int collect_garbage_slots(void); @@ -151,10 +152,9 @@ loop_end: static kprobe_opcode_t __kprobes *__get_insn_slot(void) { struct kprobe_insn_page *kip; - struct hlist_node *pos; retry: - hlist_for_each_entry(kip, pos, &kprobe_insn_pages, hlist) { + list_for_each_entry(kip, &kprobe_insn_pages, list) { if (kip->nused < INSNS_PER_PAGE) { int i; for (i = 0; i < INSNS_PER_PAGE; i++) { @@ -188,8 +188,8 @@ static kprobe_opcode_t __kprobes *__get_insn_slot(void) kfree(kip); return NULL; } - INIT_HLIST_NODE(&kip->hlist); - hlist_add_head(&kip->hlist, &kprobe_insn_pages); + INIT_LIST_HEAD(&kip->list); + list_add(&kip->list, &kprobe_insn_pages); memset(kip->slot_used, SLOT_CLEAN, INSNS_PER_PAGE); kip->slot_used[0] = SLOT_USED; kip->nused = 1; @@ -218,12 +218,8 @@ static int __kprobes collect_one_slot(struct kprobe_insn_page *kip, int idx) * so as not to have to set it up again the * next time somebody inserts a probe. */ - hlist_del(&kip->hlist); - if (hlist_empty(&kprobe_insn_pages)) { - INIT_HLIST_NODE(&kip->hlist); - hlist_add_head(&kip->hlist, - &kprobe_insn_pages); - } else { + if (!list_is_singular(&kprobe_insn_pages)) { + list_del(&kip->list); module_free(NULL, kip->insns); kfree(kip); } @@ -234,18 +230,13 @@ static int __kprobes collect_one_slot(struct kprobe_insn_page *kip, int idx) static int __kprobes collect_garbage_slots(void) { - struct kprobe_insn_page *kip; - struct hlist_node *pos, *next; - int safety; + struct kprobe_insn_page *kip, *next; /* Ensure no-one is preepmted on the garbages */ - mutex_unlock(&kprobe_insn_mutex); - safety = check_safety(); - mutex_lock(&kprobe_insn_mutex); - if (safety != 0) + if (check_safety()) return -EAGAIN; - hlist_for_each_entry_safe(kip, pos, next, &kprobe_insn_pages, hlist) { + list_for_each_entry_safe(kip, next, &kprobe_insn_pages, list) { int i; if (kip->ngarbage == 0) continue; @@ -263,19 +254,17 @@ static int __kprobes collect_garbage_slots(void) void __kprobes free_insn_slot(kprobe_opcode_t * slot, int dirty) { struct kprobe_insn_page *kip; - struct hlist_node *pos; mutex_lock(&kprobe_insn_mutex); - hlist_for_each_entry(kip, pos, &kprobe_insn_pages, hlist) { + list_for_each_entry(kip, &kprobe_insn_pages, list) { if (kip->insns <= slot && slot < kip->insns + (INSNS_PER_PAGE * MAX_INSN_SIZE)) { int i = (slot - kip->insns) / MAX_INSN_SIZE; if (dirty) { kip->slot_used[i] = SLOT_DIRTY; kip->ngarbage++; - } else { + } else collect_one_slot(kip, i); - } break; } } @@ -318,6 +307,22 @@ struct kprobe __kprobes *get_kprobe(void *addr) return NULL; } +/* Arm a kprobe with text_mutex */ +static void __kprobes arm_kprobe(struct kprobe *kp) +{ + mutex_lock(&text_mutex); + arch_arm_kprobe(kp); + mutex_unlock(&text_mutex); +} + +/* Disarm a kprobe with text_mutex */ +static void __kprobes disarm_kprobe(struct kprobe *kp) +{ + mutex_lock(&text_mutex); + arch_disarm_kprobe(kp); + mutex_unlock(&text_mutex); +} + /* * Aggregate handlers for multiple kprobes support - these handlers * take care of invoking the individual kprobe handlers on p->list @@ -327,7 +332,7 @@ static int __kprobes aggr_pre_handler(struct kprobe *p, struct pt_regs *regs) struct kprobe *kp; list_for_each_entry_rcu(kp, &p->list, list) { - if (kp->pre_handler && !kprobe_gone(kp)) { + if (kp->pre_handler && likely(!kprobe_disabled(kp))) { set_kprobe_instance(kp); if (kp->pre_handler(kp, regs)) return 1; @@ -343,7 +348,7 @@ static void __kprobes aggr_post_handler(struct kprobe *p, struct pt_regs *regs, struct kprobe *kp; list_for_each_entry_rcu(kp, &p->list, list) { - if (kp->post_handler && !kprobe_gone(kp)) { + if (kp->post_handler && likely(!kprobe_disabled(kp))) { set_kprobe_instance(kp); kp->post_handler(kp, regs, flags); reset_kprobe_instance(); @@ -517,20 +522,28 @@ static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p) } /* -* Add the new probe to old_p->list. Fail if this is the +* Add the new probe to ap->list. Fail if this is the * second jprobe at the address - two jprobes can't coexist */ -static int __kprobes add_new_kprobe(struct kprobe *old_p, struct kprobe *p) +static int __kprobes add_new_kprobe(struct kprobe *ap, struct kprobe *p) { + BUG_ON(kprobe_gone(ap) || kprobe_gone(p)); if (p->break_handler) { - if (old_p->break_handler) + if (ap->break_handler) return -EEXIST; - list_add_tail_rcu(&p->list, &old_p->list); - old_p->break_handler = aggr_break_handler; + list_add_tail_rcu(&p->list, &ap->list); + ap->break_handler = aggr_break_handler; } else - list_add_rcu(&p->list, &old_p->list); - if (p->post_handler && !old_p->post_handler) - old_p->post_handler = aggr_post_handler; + list_add_rcu(&p->list, &ap->list); + if (p->post_handler && !ap->post_handler) + ap->post_handler = aggr_post_handler; + + if (kprobe_disabled(ap) && !kprobe_disabled(p)) { + ap->flags &= ~KPROBE_FLAG_DISABLED; + if (!kprobes_all_disarmed) + /* Arm the breakpoint again. */ + arm_kprobe(ap); + } return 0; } @@ -543,6 +556,7 @@ static inline void add_aggr_kprobe(struct kprobe *ap, struct kprobe *p) copy_kprobe(p, ap); flush_insn_slot(ap); ap->addr = p->addr; + ap->flags = p->flags; ap->pre_handler = aggr_pre_handler; ap->fault_handler = aggr_fault_handler; /* We don't care the kprobe which has gone. */ @@ -565,44 +579,59 @@ static int __kprobes register_aggr_kprobe(struct kprobe *old_p, struct kprobe *p) { int ret = 0; - struct kprobe *ap; + struct kprobe *ap = old_p; + + if (old_p->pre_handler != aggr_pre_handler) { + /* If old_p is not an aggr_probe, create new aggr_kprobe. */ + ap = kzalloc(sizeof(struct kprobe), GFP_KERNEL); + if (!ap) + return -ENOMEM; + add_aggr_kprobe(ap, old_p); + } - if (kprobe_gone(old_p)) { + if (kprobe_gone(ap)) { /* * Attempting to insert new probe at the same location that * had a probe in the module vaddr area which already * freed. So, the instruction slot has already been * released. We need a new slot for the new probe. */ - ret = arch_prepare_kprobe(old_p); + ret = arch_prepare_kprobe(ap); if (ret) + /* + * Even if fail to allocate new slot, don't need to + * free aggr_probe. It will be used next time, or + * freed by unregister_kprobe. + */ return ret; - } - if (old_p->pre_handler == aggr_pre_handler) { - copy_kprobe(old_p, p); - ret = add_new_kprobe(old_p, p); - ap = old_p; - } else { - ap = kzalloc(sizeof(struct kprobe), GFP_KERNEL); - if (!ap) { - if (kprobe_gone(old_p)) - arch_remove_kprobe(old_p); - return -ENOMEM; - } - add_aggr_kprobe(ap, old_p); - copy_kprobe(ap, p); - ret = add_new_kprobe(ap, p); - } - if (kprobe_gone(old_p)) { + /* - * If the old_p has gone, its breakpoint has been disarmed. - * We have to arm it again after preparing real kprobes. + * Clear gone flag to prevent allocating new slot again, and + * set disabled flag because it is not armed yet. */ - ap->flags &= ~KPROBE_FLAG_GONE; - if (kprobe_enabled) - arch_arm_kprobe(ap); + ap->flags = (ap->flags & ~KPROBE_FLAG_GONE) + | KPROBE_FLAG_DISABLED; } - return ret; + + copy_kprobe(ap, p); + return add_new_kprobe(ap, p); +} + +/* Try to disable aggr_kprobe, and return 1 if succeeded.*/ +static int __kprobes try_to_disable_aggr_kprobe(struct kprobe *p) +{ + struct kprobe *kp; + + list_for_each_entry_rcu(kp, &p->list, list) { + if (!kprobe_disabled(kp)) + /* + * There is an active probe on the list. + * We can't disable aggr_kprobe. + */ + return 0; + } + p->flags |= KPROBE_FLAG_DISABLED; + return 1; } static int __kprobes in_kprobes_functions(unsigned long addr) @@ -657,13 +686,15 @@ int __kprobes register_kprobe(struct kprobe *p) p->addr = addr; preempt_disable(); - if (!__kernel_text_address((unsigned long) p->addr) || + if (!kernel_text_address((unsigned long) p->addr) || in_kprobes_functions((unsigned long) p->addr)) { preempt_enable(); return -EINVAL; } - p->flags = 0; + /* User can pass only KPROBE_FLAG_DISABLED to register_kprobe */ + p->flags &= KPROBE_FLAG_DISABLED; + /* * Check if are we probing a module. */ @@ -699,17 +730,20 @@ int __kprobes register_kprobe(struct kprobe *p) goto out; } + mutex_lock(&text_mutex); ret = arch_prepare_kprobe(p); if (ret) - goto out; + goto out_unlock_text; INIT_HLIST_NODE(&p->hlist); hlist_add_head_rcu(&p->hlist, &kprobe_table[hash_ptr(p->addr, KPROBE_HASH_BITS)]); - if (kprobe_enabled) + if (!kprobes_all_disarmed && !kprobe_disabled(p)) arch_arm_kprobe(p); +out_unlock_text: + mutex_unlock(&text_mutex); out: mutex_unlock(&kprobe_mutex); @@ -718,26 +752,39 @@ out: return ret; } +EXPORT_SYMBOL_GPL(register_kprobe); -/* - * Unregister a kprobe without a scheduler synchronization. - */ -static int __kprobes __unregister_kprobe_top(struct kprobe *p) +/* Check passed kprobe is valid and return kprobe in kprobe_table. */ +static struct kprobe * __kprobes __get_valid_kprobe(struct kprobe *p) { struct kprobe *old_p, *list_p; old_p = get_kprobe(p->addr); if (unlikely(!old_p)) - return -EINVAL; + return NULL; if (p != old_p) { list_for_each_entry_rcu(list_p, &old_p->list, list) if (list_p == p) /* kprobe p is a valid probe */ - goto valid_p; - return -EINVAL; + goto valid; + return NULL; } -valid_p: +valid: + return old_p; +} + +/* + * Unregister a kprobe without a scheduler synchronization. + */ +static int __kprobes __unregister_kprobe_top(struct kprobe *p) +{ + struct kprobe *old_p, *list_p; + + old_p = __get_valid_kprobe(p); + if (old_p == NULL) + return -EINVAL; + if (old_p == p || (old_p->pre_handler == aggr_pre_handler && list_is_singular(&old_p->list))) { @@ -746,8 +793,8 @@ valid_p: * enabled and not gone - otherwise, the breakpoint would * already have been removed. We save on flushing icache. */ - if (kprobe_enabled && !kprobe_gone(old_p)) - arch_disarm_kprobe(p); + if (!kprobes_all_disarmed && !kprobe_disabled(old_p)) + disarm_kprobe(p); hlist_del_rcu(&old_p->hlist); } else { if (p->break_handler && !kprobe_gone(p)) @@ -761,6 +808,11 @@ valid_p: } noclean: list_del_rcu(&p->list); + if (!kprobe_disabled(old_p)) { + try_to_disable_aggr_kprobe(old_p); + if (!kprobes_all_disarmed && kprobe_disabled(old_p)) + disarm_kprobe(old_p); + } } return 0; } @@ -796,11 +848,13 @@ int __kprobes register_kprobes(struct kprobe **kps, int num) } return ret; } +EXPORT_SYMBOL_GPL(register_kprobes); void __kprobes unregister_kprobe(struct kprobe *p) { unregister_kprobes(&p, 1); } +EXPORT_SYMBOL_GPL(unregister_kprobe); void __kprobes unregister_kprobes(struct kprobe **kps, int num) { @@ -819,6 +873,7 @@ void __kprobes unregister_kprobes(struct kprobe **kps, int num) if (kps[i]->addr) __unregister_kprobe_bottom(kps[i]); } +EXPORT_SYMBOL_GPL(unregister_kprobes); static struct notifier_block kprobe_exceptions_nb = { .notifier_call = kprobe_exceptions_notify, @@ -858,16 +913,19 @@ int __kprobes register_jprobes(struct jprobe **jps, int num) } return ret; } +EXPORT_SYMBOL_GPL(register_jprobes); int __kprobes register_jprobe(struct jprobe *jp) { return register_jprobes(&jp, 1); } +EXPORT_SYMBOL_GPL(register_jprobe); void __kprobes unregister_jprobe(struct jprobe *jp) { unregister_jprobes(&jp, 1); } +EXPORT_SYMBOL_GPL(unregister_jprobe); void __kprobes unregister_jprobes(struct jprobe **jps, int num) { @@ -887,6 +945,7 @@ void __kprobes unregister_jprobes(struct jprobe **jps, int num) __unregister_kprobe_bottom(&jps[i]->kp); } } +EXPORT_SYMBOL_GPL(unregister_jprobes); #ifdef CONFIG_KRETPROBES /* @@ -912,10 +971,8 @@ static int __kprobes pre_handler_kretprobe(struct kprobe *p, ri->rp = rp; ri->task = current; - if (rp->entry_handler && rp->entry_handler(ri, regs)) { - spin_unlock_irqrestore(&rp->lock, flags); + if (rp->entry_handler && rp->entry_handler(ri, regs)) return 0; - } arch_prepare_kretprobe(ri, regs); @@ -982,6 +1039,7 @@ int __kprobes register_kretprobe(struct kretprobe *rp) free_rp_inst(rp); return ret; } +EXPORT_SYMBOL_GPL(register_kretprobe); int __kprobes register_kretprobes(struct kretprobe **rps, int num) { @@ -999,11 +1057,13 @@ int __kprobes register_kretprobes(struct kretprobe **rps, int num) } return ret; } +EXPORT_SYMBOL_GPL(register_kretprobes); void __kprobes unregister_kretprobe(struct kretprobe *rp) { unregister_kretprobes(&rp, 1); } +EXPORT_SYMBOL_GPL(unregister_kretprobe); void __kprobes unregister_kretprobes(struct kretprobe **rps, int num) { @@ -1025,24 +1085,30 @@ void __kprobes unregister_kretprobes(struct kretprobe **rps, int num) } } } +EXPORT_SYMBOL_GPL(unregister_kretprobes); #else /* CONFIG_KRETPROBES */ int __kprobes register_kretprobe(struct kretprobe *rp) { return -ENOSYS; } +EXPORT_SYMBOL_GPL(register_kretprobe); int __kprobes register_kretprobes(struct kretprobe **rps, int num) { return -ENOSYS; } +EXPORT_SYMBOL_GPL(register_kretprobes); + void __kprobes unregister_kretprobe(struct kretprobe *rp) { } +EXPORT_SYMBOL_GPL(unregister_kretprobe); void __kprobes unregister_kretprobes(struct kretprobe **rps, int num) { } +EXPORT_SYMBOL_GPL(unregister_kretprobes); static int __kprobes pre_handler_kretprobe(struct kprobe *p, struct pt_regs *regs) @@ -1056,6 +1122,7 @@ static int __kprobes pre_handler_kretprobe(struct kprobe *p, static void __kprobes kill_kprobe(struct kprobe *p) { struct kprobe *kp; + p->flags |= KPROBE_FLAG_GONE; if (p->pre_handler == aggr_pre_handler) { /* @@ -1168,8 +1235,8 @@ static int __init init_kprobes(void) } } - /* By default, kprobes are enabled */ - kprobe_enabled = true; + /* By default, kprobes are armed */ + kprobes_all_disarmed = false; err = arch_init_kprobes(); if (!err) @@ -1197,12 +1264,18 @@ static void __kprobes report_probe(struct seq_file *pi, struct kprobe *p, else kprobe_type = "k"; if (sym) - seq_printf(pi, "%p %s %s+0x%x %s %s\n", p->addr, kprobe_type, - sym, offset, (modname ? modname : " "), - (kprobe_gone(p) ? "[GONE]" : "")); + seq_printf(pi, "%p %s %s+0x%x %s %s%s\n", + p->addr, kprobe_type, sym, offset, + (modname ? modname : " "), + (kprobe_gone(p) ? "[GONE]" : ""), + ((kprobe_disabled(p) && !kprobe_gone(p)) ? + "[DISABLED]" : "")); else - seq_printf(pi, "%p %s %p %s\n", p->addr, kprobe_type, p->addr, - (kprobe_gone(p) ? "[GONE]" : "")); + seq_printf(pi, "%p %s %p %s%s\n", + p->addr, kprobe_type, p->addr, + (kprobe_gone(p) ? "[GONE]" : ""), + ((kprobe_disabled(p) && !kprobe_gone(p)) ? + "[DISABLED]" : "")); } static void __kprobes *kprobe_seq_start(struct seq_file *f, loff_t *pos) @@ -1267,7 +1340,72 @@ static struct file_operations debugfs_kprobes_operations = { .release = seq_release, }; -static void __kprobes enable_all_kprobes(void) +/* Disable one kprobe */ +int __kprobes disable_kprobe(struct kprobe *kp) +{ + int ret = 0; + struct kprobe *p; + + mutex_lock(&kprobe_mutex); + + /* Check whether specified probe is valid. */ + p = __get_valid_kprobe(kp); + if (unlikely(p == NULL)) { + ret = -EINVAL; + goto out; + } + + /* If the probe is already disabled (or gone), just return */ + if (kprobe_disabled(kp)) + goto out; + + kp->flags |= KPROBE_FLAG_DISABLED; + if (p != kp) + /* When kp != p, p is always enabled. */ + try_to_disable_aggr_kprobe(p); + + if (!kprobes_all_disarmed && kprobe_disabled(p)) + disarm_kprobe(p); +out: + mutex_unlock(&kprobe_mutex); + return ret; +} +EXPORT_SYMBOL_GPL(disable_kprobe); + +/* Enable one kprobe */ +int __kprobes enable_kprobe(struct kprobe *kp) +{ + int ret = 0; + struct kprobe *p; + + mutex_lock(&kprobe_mutex); + + /* Check whether specified probe is valid. */ + p = __get_valid_kprobe(kp); + if (unlikely(p == NULL)) { + ret = -EINVAL; + goto out; + } + + if (kprobe_gone(kp)) { + /* This kprobe has gone, we couldn't enable it. */ + ret = -EINVAL; + goto out; + } + + if (!kprobes_all_disarmed && kprobe_disabled(p)) + arm_kprobe(p); + + p->flags &= ~KPROBE_FLAG_DISABLED; + if (p != kp) + kp->flags &= ~KPROBE_FLAG_DISABLED; +out: + mutex_unlock(&kprobe_mutex); + return ret; +} +EXPORT_SYMBOL_GPL(enable_kprobe); + +static void __kprobes arm_all_kprobes(void) { struct hlist_head *head; struct hlist_node *node; @@ -1276,18 +1414,20 @@ static void __kprobes enable_all_kprobes(void) mutex_lock(&kprobe_mutex); - /* If kprobes are already enabled, just return */ - if (kprobe_enabled) + /* If kprobes are armed, just return */ + if (!kprobes_all_disarmed) goto already_enabled; + mutex_lock(&text_mutex); for (i = 0; i < KPROBE_TABLE_SIZE; i++) { head = &kprobe_table[i]; hlist_for_each_entry_rcu(p, node, head, hlist) - if (!kprobe_gone(p)) + if (!kprobe_disabled(p)) arch_arm_kprobe(p); } + mutex_unlock(&text_mutex); - kprobe_enabled = true; + kprobes_all_disarmed = false; printk(KERN_INFO "Kprobes globally enabled\n"); already_enabled: @@ -1295,7 +1435,7 @@ already_enabled: return; } -static void __kprobes disable_all_kprobes(void) +static void __kprobes disarm_all_kprobes(void) { struct hlist_head *head; struct hlist_node *node; @@ -1304,20 +1444,22 @@ static void __kprobes disable_all_kprobes(void) mutex_lock(&kprobe_mutex); - /* If kprobes are already disabled, just return */ - if (!kprobe_enabled) + /* If kprobes are already disarmed, just return */ + if (kprobes_all_disarmed) goto already_disabled; - kprobe_enabled = false; + kprobes_all_disarmed = true; printk(KERN_INFO "Kprobes globally disabled\n"); + mutex_lock(&text_mutex); for (i = 0; i < KPROBE_TABLE_SIZE; i++) { head = &kprobe_table[i]; hlist_for_each_entry_rcu(p, node, head, hlist) { - if (!arch_trampoline_kprobe(p) && !kprobe_gone(p)) + if (!arch_trampoline_kprobe(p) && !kprobe_disabled(p)) arch_disarm_kprobe(p); } } + mutex_unlock(&text_mutex); mutex_unlock(&kprobe_mutex); /* Allow all currently running kprobes to complete */ synchronize_sched(); @@ -1338,7 +1480,7 @@ static ssize_t read_enabled_file_bool(struct file *file, { char buf[3]; - if (kprobe_enabled) + if (!kprobes_all_disarmed) buf[0] = '1'; else buf[0] = '0'; @@ -1361,12 +1503,12 @@ static ssize_t write_enabled_file_bool(struct file *file, case 'y': case 'Y': case '1': - enable_all_kprobes(); + arm_all_kprobes(); break; case 'n': case 'N': case '0': - disable_all_kprobes(); + disarm_all_kprobes(); break; } @@ -1409,16 +1551,5 @@ late_initcall(debugfs_kprobe_init); module_init(init_kprobes); -EXPORT_SYMBOL_GPL(register_kprobe); -EXPORT_SYMBOL_GPL(unregister_kprobe); -EXPORT_SYMBOL_GPL(register_kprobes); -EXPORT_SYMBOL_GPL(unregister_kprobes); -EXPORT_SYMBOL_GPL(register_jprobe); -EXPORT_SYMBOL_GPL(unregister_jprobe); -EXPORT_SYMBOL_GPL(register_jprobes); -EXPORT_SYMBOL_GPL(unregister_jprobes); +/* defined in arch/.../kernel/kprobes.c */ EXPORT_SYMBOL_GPL(jprobe_return); -EXPORT_SYMBOL_GPL(register_kretprobe); -EXPORT_SYMBOL_GPL(unregister_kretprobe); -EXPORT_SYMBOL_GPL(register_kretprobes); -EXPORT_SYMBOL_GPL(unregister_kretprobes); diff --git a/kernel/kthread.c b/kernel/kthread.c index 4fbc456f393..5fe709982ca 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -9,27 +9,22 @@ #include <linux/kthread.h> #include <linux/completion.h> #include <linux/err.h> +#include <linux/cpuset.h> #include <linux/unistd.h> #include <linux/file.h> #include <linux/module.h> #include <linux/mutex.h> -#include <trace/sched.h> - -#define KTHREAD_NICE_LEVEL (-5) +#include <trace/events/sched.h> static DEFINE_SPINLOCK(kthread_create_lock); static LIST_HEAD(kthread_create_list); struct task_struct *kthreadd_task; -DEFINE_TRACE(sched_kthread_stop); -DEFINE_TRACE(sched_kthread_stop_ret); - struct kthread_create_info { /* Information passed to kthread() from kthreadd. */ int (*threadfn)(void *data); void *data; - struct completion started; /* Result passed back to kthread_create() from kthreadd. */ struct task_struct *result; @@ -38,17 +33,13 @@ struct kthread_create_info struct list_head list; }; -struct kthread_stop_info -{ - struct task_struct *k; - int err; - struct completion done; +struct kthread { + int should_stop; + struct completion exited; }; -/* Thread stopping is done by setthing this var: lock serializes - * multiple kthread_stop calls. */ -static DEFINE_MUTEX(kthread_stop_lock); -static struct kthread_stop_info kthread_stop_info; +#define to_kthread(tsk) \ + container_of((tsk)->vfork_done, struct kthread, exited) /** * kthread_should_stop - should this kthread return now? @@ -59,35 +50,35 @@ static struct kthread_stop_info kthread_stop_info; */ int kthread_should_stop(void) { - return (kthread_stop_info.k == current); + return to_kthread(current)->should_stop; } EXPORT_SYMBOL(kthread_should_stop); static int kthread(void *_create) { + /* Copy data: it's on kthread's stack */ struct kthread_create_info *create = _create; - int (*threadfn)(void *data); - void *data; - int ret = -EINTR; + int (*threadfn)(void *data) = create->threadfn; + void *data = create->data; + struct kthread self; + int ret; - /* Copy data: it's on kthread's stack */ - threadfn = create->threadfn; - data = create->data; + self.should_stop = 0; + init_completion(&self.exited); + current->vfork_done = &self.exited; /* OK, tell user we're spawned, wait for stop or wakeup */ __set_current_state(TASK_UNINTERRUPTIBLE); - complete(&create->started); + create->result = current; + complete(&create->done); schedule(); - if (!kthread_should_stop()) + ret = -EINTR; + if (!self.should_stop) ret = threadfn(data); - /* It might have exited on its own, w/o kthread_stop. Check. */ - if (kthread_should_stop()) { - kthread_stop_info.err = ret; - complete(&kthread_stop_info.done); - } - return 0; + /* we can't just return, we must preserve "self" on stack */ + do_exit(ret); } static void create_kthread(struct kthread_create_info *create) @@ -98,21 +89,8 @@ static void create_kthread(struct kthread_create_info *create) pid = kernel_thread(kthread, create, CLONE_FS | CLONE_FILES | SIGCHLD); if (pid < 0) { create->result = ERR_PTR(pid); - } else { - struct sched_param param = { .sched_priority = 0 }; - wait_for_completion(&create->started); - read_lock(&tasklist_lock); - create->result = find_task_by_pid_ns(pid, &init_pid_ns); - read_unlock(&tasklist_lock); - /* - * root may have changed our (kthreadd's) priority or CPU mask. - * The kernel thread should not inherit these properties. - */ - sched_setscheduler(create->result, SCHED_NORMAL, ¶m); - set_user_nice(create->result, KTHREAD_NICE_LEVEL); - set_cpus_allowed_ptr(create->result, CPU_MASK_ALL_PTR); + complete(&create->done); } - complete(&create->done); } /** @@ -143,7 +121,6 @@ struct task_struct *kthread_create(int (*threadfn)(void *data), create.threadfn = threadfn; create.data = data; - init_completion(&create.started); init_completion(&create.done); spin_lock(&kthread_create_lock); @@ -154,11 +131,19 @@ struct task_struct *kthread_create(int (*threadfn)(void *data), wait_for_completion(&create.done); if (!IS_ERR(create.result)) { + struct sched_param param = { .sched_priority = 0 }; va_list args; + va_start(args, namefmt); vsnprintf(create.result->comm, sizeof(create.result->comm), namefmt, args); va_end(args); + /* + * root may have changed our (kthreadd's) priority or CPU mask. + * The kernel thread should not inherit these properties. + */ + sched_setscheduler_nocheck(create.result, SCHED_NORMAL, ¶m); + set_cpus_allowed_ptr(create.result, cpu_all_mask); } return create.result; } @@ -192,40 +177,34 @@ EXPORT_SYMBOL(kthread_bind); * @k: thread created by kthread_create(). * * Sets kthread_should_stop() for @k to return true, wakes it, and - * waits for it to exit. Your threadfn() must not call do_exit() - * itself if you use this function! This can also be called after - * kthread_create() instead of calling wake_up_process(): the thread - * will exit without calling threadfn(). + * waits for it to exit. This can also be called after kthread_create() + * instead of calling wake_up_process(): the thread will exit without + * calling threadfn(). + * + * If threadfn() may call do_exit() itself, the caller must ensure + * task_struct can't go away. * * Returns the result of threadfn(), or %-EINTR if wake_up_process() * was never called. */ int kthread_stop(struct task_struct *k) { + struct kthread *kthread; int ret; - mutex_lock(&kthread_stop_lock); - - /* It could exit after stop_info.k set, but before wake_up_process. */ - get_task_struct(k); - trace_sched_kthread_stop(k); + get_task_struct(k); - /* Must init completion *before* thread sees kthread_stop_info.k */ - init_completion(&kthread_stop_info.done); - smp_wmb(); + kthread = to_kthread(k); + barrier(); /* it might have exited */ + if (k->vfork_done != NULL) { + kthread->should_stop = 1; + wake_up_process(k); + wait_for_completion(&kthread->exited); + } + ret = k->exit_code; - /* Now set kthread_should_stop() to true, and wake it up. */ - kthread_stop_info.k = k; - wake_up_process(k); put_task_struct(k); - - /* Once it dies, reset stop ptr, gather result and we're done. */ - wait_for_completion(&kthread_stop_info.done); - kthread_stop_info.k = NULL; - ret = kthread_stop_info.err; - mutex_unlock(&kthread_stop_lock); - trace_sched_kthread_stop_ret(ret); return ret; @@ -239,8 +218,8 @@ int kthreadd(void *unused) /* Setup a clean context for our children to inherit. */ set_task_comm(tsk, "kthreadd"); ignore_signals(tsk); - set_user_nice(tsk, KTHREAD_NICE_LEVEL); - set_cpus_allowed_ptr(tsk, CPU_MASK_ALL_PTR); + set_cpus_allowed_ptr(tsk, cpu_all_mask); + set_mems_allowed(node_possible_map); current->flags |= PF_NOFREEZE | PF_FREEZER_NOSIG; diff --git a/kernel/latencytop.c b/kernel/latencytop.c index 449db466bdb..ca07c5c0c91 100644 --- a/kernel/latencytop.c +++ b/kernel/latencytop.c @@ -9,6 +9,44 @@ * as published by the Free Software Foundation; version 2 * of the License. */ + +/* + * CONFIG_LATENCYTOP enables a kernel latency tracking infrastructure that is + * used by the "latencytop" userspace tool. The latency that is tracked is not + * the 'traditional' interrupt latency (which is primarily caused by something + * else consuming CPU), but instead, it is the latency an application encounters + * because the kernel sleeps on its behalf for various reasons. + * + * This code tracks 2 levels of statistics: + * 1) System level latency + * 2) Per process latency + * + * The latency is stored in fixed sized data structures in an accumulated form; + * if the "same" latency cause is hit twice, this will be tracked as one entry + * in the data structure. Both the count, total accumulated latency and maximum + * latency are tracked in this data structure. When the fixed size structure is + * full, no new causes are tracked until the buffer is flushed by writing to + * the /proc file; the userspace tool does this on a regular basis. + * + * A latency cause is identified by a stringified backtrace at the point that + * the scheduler gets invoked. The userland tool will use this string to + * identify the cause of the latency in human readable form. + * + * The information is exported via /proc/latency_stats and /proc/<pid>/latency. + * These files look like this: + * + * Latency Top version : v0.1 + * 70 59433 4897 i915_irq_wait drm_ioctl vfs_ioctl do_vfs_ioctl sys_ioctl + * | | | | + * | | | +----> the stringified backtrace + * | | +---------> The maximum latency for this entry in microseconds + * | +--------------> The accumulated latency for this entry (microseconds) + * +-------------------> The number of times this entry is hit + * + * (note: the average latency is the accumulated latency divided by the number + * of times) + */ + #include <linux/latencytop.h> #include <linux/kallsyms.h> #include <linux/seq_file.h> @@ -72,7 +110,7 @@ account_global_scheduler_latency(struct task_struct *tsk, struct latency_record firstnonnull = i; continue; } - for (q = 0 ; q < LT_BACKTRACEDEPTH ; q++) { + for (q = 0; q < LT_BACKTRACEDEPTH; q++) { unsigned long record = lat->backtrace[q]; if (latency_record[i].backtrace[q] != record) { @@ -101,31 +139,52 @@ account_global_scheduler_latency(struct task_struct *tsk, struct latency_record memcpy(&latency_record[i], lat, sizeof(struct latency_record)); } -static inline void store_stacktrace(struct task_struct *tsk, struct latency_record *lat) +/* + * Iterator to store a backtrace into a latency record entry + */ +static inline void store_stacktrace(struct task_struct *tsk, + struct latency_record *lat) { struct stack_trace trace; memset(&trace, 0, sizeof(trace)); trace.max_entries = LT_BACKTRACEDEPTH; trace.entries = &lat->backtrace[0]; - trace.skip = 0; save_stack_trace_tsk(tsk, &trace); } +/** + * __account_scheduler_latency - record an occured latency + * @tsk - the task struct of the task hitting the latency + * @usecs - the duration of the latency in microseconds + * @inter - 1 if the sleep was interruptible, 0 if uninterruptible + * + * This function is the main entry point for recording latency entries + * as called by the scheduler. + * + * This function has a few special cases to deal with normal 'non-latency' + * sleeps: specifically, interruptible sleep longer than 5 msec is skipped + * since this usually is caused by waiting for events via select() and co. + * + * Negative latencies (caused by time going backwards) are also explicitly + * skipped. + */ void __sched -account_scheduler_latency(struct task_struct *tsk, int usecs, int inter) +__account_scheduler_latency(struct task_struct *tsk, int usecs, int inter) { unsigned long flags; int i, q; struct latency_record lat; - if (!latencytop_enabled) - return; - /* Long interruptible waits are generally user requested... */ if (inter && usecs > 5000) return; + /* Negative sleeps are time going backwards */ + /* Zero-time sleeps are non-interesting */ + if (usecs <= 0) + return; + memset(&lat, 0, sizeof(lat)); lat.count = 1; lat.time = usecs; @@ -143,12 +202,12 @@ account_scheduler_latency(struct task_struct *tsk, int usecs, int inter) if (tsk->latency_record_count >= LT_SAVECOUNT) goto out_unlock; - for (i = 0; i < LT_SAVECOUNT ; i++) { + for (i = 0; i < LT_SAVECOUNT; i++) { struct latency_record *mylat; int same = 1; mylat = &tsk->latency_record[i]; - for (q = 0 ; q < LT_BACKTRACEDEPTH ; q++) { + for (q = 0; q < LT_BACKTRACEDEPTH; q++) { unsigned long record = lat.backtrace[q]; if (mylat->backtrace[q] != record) { @@ -186,7 +245,7 @@ static int lstats_show(struct seq_file *m, void *v) for (i = 0; i < MAXLR; i++) { if (latency_record[i].backtrace[0]) { int q; - seq_printf(m, "%i %li %li ", + seq_printf(m, "%i %lu %lu ", latency_record[i].count, latency_record[i].time, latency_record[i].max); @@ -223,7 +282,7 @@ static int lstats_open(struct inode *inode, struct file *filp) return single_open(filp, lstats_show, NULL); } -static struct file_operations lstats_fops = { +static const struct file_operations lstats_fops = { .open = lstats_open, .read = seq_read, .write = lstats_write, @@ -236,4 +295,4 @@ static int __init init_lstats_procfs(void) proc_create("latency_stats", 0644, NULL, &lstats_fops); return 0; } -__initcall(init_lstats_procfs); +device_initcall(init_lstats_procfs); diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 06b0c3568f0..f74d2d7aa60 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -41,11 +41,16 @@ #include <linux/utsname.h> #include <linux/hash.h> #include <linux/ftrace.h> +#include <linux/stringify.h> +#include <linux/bitops.h> #include <asm/sections.h> #include "lockdep_internals.h" +#define CREATE_TRACE_POINTS +#include <trace/events/lockdep.h> + #ifdef CONFIG_PROVE_LOCKING int prove_locking = 1; module_param(prove_locking, int, 0644); @@ -310,12 +315,14 @@ EXPORT_SYMBOL(lockdep_on); #if VERBOSE # define HARDIRQ_VERBOSE 1 # define SOFTIRQ_VERBOSE 1 +# define RECLAIM_VERBOSE 1 #else # define HARDIRQ_VERBOSE 0 # define SOFTIRQ_VERBOSE 0 +# define RECLAIM_VERBOSE 0 #endif -#if VERBOSE || HARDIRQ_VERBOSE || SOFTIRQ_VERBOSE +#if VERBOSE || HARDIRQ_VERBOSE || SOFTIRQ_VERBOSE || RECLAIM_VERBOSE /* * Quick filtering for interesting events: */ @@ -360,11 +367,21 @@ static int save_trace(struct stack_trace *trace) save_stack_trace(trace); + /* + * Some daft arches put -1 at the end to indicate its a full trace. + * + * <rant> this is buggy anyway, since it takes a whole extra entry so a + * complete trace that maxes out the entries provided will be reported + * as incomplete, friggin useless </rant> + */ + if (trace->entries[trace->nr_entries-1] == ULONG_MAX) + trace->nr_entries--; + trace->max_entries = trace->nr_entries; nr_stack_trace_entries += trace->nr_entries; - if (nr_stack_trace_entries == MAX_STACK_TRACE_ENTRIES) { + if (nr_stack_trace_entries >= MAX_STACK_TRACE_ENTRIES-1) { if (!debug_locks_off_graph_unlock()) return 0; @@ -382,20 +399,6 @@ unsigned int nr_hardirq_chains; unsigned int nr_softirq_chains; unsigned int nr_process_chains; unsigned int max_lockdep_depth; -unsigned int max_recursion_depth; - -static unsigned int lockdep_dependency_gen_id; - -static bool lockdep_dependency_visit(struct lock_class *source, - unsigned int depth) -{ - if (!depth) - lockdep_dependency_gen_id++; - if (source->dep_gen_id == lockdep_dependency_gen_id) - return true; - source->dep_gen_id = lockdep_dependency_gen_id; - return false; -} #ifdef CONFIG_DEBUG_LOCKDEP /* @@ -425,35 +428,26 @@ atomic_t redundant_softirqs_on; atomic_t redundant_softirqs_off; atomic_t nr_unused_locks; atomic_t nr_cyclic_checks; -atomic_t nr_cyclic_check_recursions; atomic_t nr_find_usage_forwards_checks; -atomic_t nr_find_usage_forwards_recursions; atomic_t nr_find_usage_backwards_checks; -atomic_t nr_find_usage_backwards_recursions; -# define debug_atomic_inc(ptr) atomic_inc(ptr) -# define debug_atomic_dec(ptr) atomic_dec(ptr) -# define debug_atomic_read(ptr) atomic_read(ptr) -#else -# define debug_atomic_inc(ptr) do { } while (0) -# define debug_atomic_dec(ptr) do { } while (0) -# define debug_atomic_read(ptr) 0 #endif /* * Locking printouts: */ +#define __USAGE(__STATE) \ + [LOCK_USED_IN_##__STATE] = "IN-"__stringify(__STATE)"-W", \ + [LOCK_ENABLED_##__STATE] = __stringify(__STATE)"-ON-W", \ + [LOCK_USED_IN_##__STATE##_READ] = "IN-"__stringify(__STATE)"-R",\ + [LOCK_ENABLED_##__STATE##_READ] = __stringify(__STATE)"-ON-R", + static const char *usage_str[] = { - [LOCK_USED] = "initial-use ", - [LOCK_USED_IN_HARDIRQ] = "in-hardirq-W", - [LOCK_USED_IN_SOFTIRQ] = "in-softirq-W", - [LOCK_ENABLED_SOFTIRQS] = "softirq-on-W", - [LOCK_ENABLED_HARDIRQS] = "hardirq-on-W", - [LOCK_USED_IN_HARDIRQ_READ] = "in-hardirq-R", - [LOCK_USED_IN_SOFTIRQ_READ] = "in-softirq-R", - [LOCK_ENABLED_SOFTIRQS_READ] = "softirq-on-R", - [LOCK_ENABLED_HARDIRQS_READ] = "hardirq-on-R", +#define LOCKDEP_STATE(__STATE) __USAGE(__STATE) +#include "lockdep_states.h" +#undef LOCKDEP_STATE + [LOCK_USED] = "INITIAL USE", }; const char * __get_key_name(struct lockdep_subclass_key *key, char *str) @@ -461,46 +455,45 @@ const char * __get_key_name(struct lockdep_subclass_key *key, char *str) return kallsyms_lookup((unsigned long)key, NULL, NULL, NULL, str); } -void -get_usage_chars(struct lock_class *class, char *c1, char *c2, char *c3, char *c4) +static inline unsigned long lock_flag(enum lock_usage_bit bit) { - *c1 = '.', *c2 = '.', *c3 = '.', *c4 = '.'; - - if (class->usage_mask & LOCKF_USED_IN_HARDIRQ) - *c1 = '+'; - else - if (class->usage_mask & LOCKF_ENABLED_HARDIRQS) - *c1 = '-'; + return 1UL << bit; +} - if (class->usage_mask & LOCKF_USED_IN_SOFTIRQ) - *c2 = '+'; - else - if (class->usage_mask & LOCKF_ENABLED_SOFTIRQS) - *c2 = '-'; +static char get_usage_char(struct lock_class *class, enum lock_usage_bit bit) +{ + char c = '.'; - if (class->usage_mask & LOCKF_ENABLED_HARDIRQS_READ) - *c3 = '-'; - if (class->usage_mask & LOCKF_USED_IN_HARDIRQ_READ) { - *c3 = '+'; - if (class->usage_mask & LOCKF_ENABLED_HARDIRQS_READ) - *c3 = '?'; + if (class->usage_mask & lock_flag(bit + 2)) + c = '+'; + if (class->usage_mask & lock_flag(bit)) { + c = '-'; + if (class->usage_mask & lock_flag(bit + 2)) + c = '?'; } - if (class->usage_mask & LOCKF_ENABLED_SOFTIRQS_READ) - *c4 = '-'; - if (class->usage_mask & LOCKF_USED_IN_SOFTIRQ_READ) { - *c4 = '+'; - if (class->usage_mask & LOCKF_ENABLED_SOFTIRQS_READ) - *c4 = '?'; - } + return c; +} + +void get_usage_chars(struct lock_class *class, char usage[LOCK_USAGE_CHARS]) +{ + int i = 0; + +#define LOCKDEP_STATE(__STATE) \ + usage[i++] = get_usage_char(class, LOCK_USED_IN_##__STATE); \ + usage[i++] = get_usage_char(class, LOCK_USED_IN_##__STATE##_READ); +#include "lockdep_states.h" +#undef LOCKDEP_STATE + + usage[i] = '\0'; } static void print_lock_name(struct lock_class *class) { - char str[KSYM_NAME_LEN], c1, c2, c3, c4; + char str[KSYM_NAME_LEN], usage[LOCK_USAGE_CHARS]; const char *name; - get_usage_chars(class, &c1, &c2, &c3, &c4); + get_usage_chars(class, usage); name = class->name; if (!name) { @@ -513,7 +506,7 @@ static void print_lock_name(struct lock_class *class) if (class->subclass) printk("/%d", class->subclass); } - printk("){%c%c%c%c}", c1, c2, c3, c4); + printk("){%s}", usage); } static void print_lockdep_cache(struct lockdep_map *lock) @@ -552,58 +545,6 @@ static void lockdep_print_held_locks(struct task_struct *curr) } } -static void print_lock_class_header(struct lock_class *class, int depth) -{ - int bit; - - printk("%*s->", depth, ""); - print_lock_name(class); - printk(" ops: %lu", class->ops); - printk(" {\n"); - - for (bit = 0; bit < LOCK_USAGE_STATES; bit++) { - if (class->usage_mask & (1 << bit)) { - int len = depth; - - len += printk("%*s %s", depth, "", usage_str[bit]); - len += printk(" at:\n"); - print_stack_trace(class->usage_traces + bit, len); - } - } - printk("%*s }\n", depth, ""); - - printk("%*s ... key at: ",depth,""); - print_ip_sym((unsigned long)class->key); -} - -/* - * printk all lock dependencies starting at <entry>: - */ -static void __used -print_lock_dependencies(struct lock_class *class, int depth) -{ - struct lock_list *entry; - - if (lockdep_dependency_visit(class, depth)) - return; - - if (DEBUG_LOCKS_WARN_ON(depth >= 20)) - return; - - print_lock_class_header(class, depth); - - list_for_each_entry(entry, &class->locks_after, entry) { - if (DEBUG_LOCKS_WARN_ON(!entry->class)) - return; - - print_lock_dependencies(entry->class, depth + 1); - - printk("%*s ... acquired at:\n",depth,""); - print_stack_trace(&entry->trace, 2); - printk("\n"); - } -} - static void print_kernel_version(void) { printk("%s %.*s\n", init_utsname()->release, @@ -796,6 +737,7 @@ register_lock_class(struct lockdep_map *lock, unsigned int subclass, int force) printk("BUG: MAX_LOCKDEP_KEYS too low!\n"); printk("turning off the locking correctness validator.\n"); + dump_stack(); return NULL; } class = lock_classes + nr_lock_classes++; @@ -859,6 +801,7 @@ static struct lock_list *alloc_list_entry(void) printk("BUG: MAX_LOCKDEP_ENTRIES too low!\n"); printk("turning off the locking correctness validator.\n"); + dump_stack(); return NULL; } return list_entries + nr_list_entries++; @@ -897,22 +840,203 @@ static int add_lock_to_list(struct lock_class *class, struct lock_class *this, } /* + * For good efficiency of modular, we use power of 2 + */ +#define MAX_CIRCULAR_QUEUE_SIZE 4096UL +#define CQ_MASK (MAX_CIRCULAR_QUEUE_SIZE-1) + +/* + * The circular_queue and helpers is used to implement the + * breadth-first search(BFS)algorithem, by which we can build + * the shortest path from the next lock to be acquired to the + * previous held lock if there is a circular between them. + */ +struct circular_queue { + unsigned long element[MAX_CIRCULAR_QUEUE_SIZE]; + unsigned int front, rear; +}; + +static struct circular_queue lock_cq; + +unsigned int max_bfs_queue_depth; + +static unsigned int lockdep_dependency_gen_id; + +static inline void __cq_init(struct circular_queue *cq) +{ + cq->front = cq->rear = 0; + lockdep_dependency_gen_id++; +} + +static inline int __cq_empty(struct circular_queue *cq) +{ + return (cq->front == cq->rear); +} + +static inline int __cq_full(struct circular_queue *cq) +{ + return ((cq->rear + 1) & CQ_MASK) == cq->front; +} + +static inline int __cq_enqueue(struct circular_queue *cq, unsigned long elem) +{ + if (__cq_full(cq)) + return -1; + + cq->element[cq->rear] = elem; + cq->rear = (cq->rear + 1) & CQ_MASK; + return 0; +} + +static inline int __cq_dequeue(struct circular_queue *cq, unsigned long *elem) +{ + if (__cq_empty(cq)) + return -1; + + *elem = cq->element[cq->front]; + cq->front = (cq->front + 1) & CQ_MASK; + return 0; +} + +static inline unsigned int __cq_get_elem_count(struct circular_queue *cq) +{ + return (cq->rear - cq->front) & CQ_MASK; +} + +static inline void mark_lock_accessed(struct lock_list *lock, + struct lock_list *parent) +{ + unsigned long nr; + + nr = lock - list_entries; + WARN_ON(nr >= nr_list_entries); + lock->parent = parent; + lock->class->dep_gen_id = lockdep_dependency_gen_id; +} + +static inline unsigned long lock_accessed(struct lock_list *lock) +{ + unsigned long nr; + + nr = lock - list_entries; + WARN_ON(nr >= nr_list_entries); + return lock->class->dep_gen_id == lockdep_dependency_gen_id; +} + +static inline struct lock_list *get_lock_parent(struct lock_list *child) +{ + return child->parent; +} + +static inline int get_lock_depth(struct lock_list *child) +{ + int depth = 0; + struct lock_list *parent; + + while ((parent = get_lock_parent(child))) { + child = parent; + depth++; + } + return depth; +} + +static int __bfs(struct lock_list *source_entry, + void *data, + int (*match)(struct lock_list *entry, void *data), + struct lock_list **target_entry, + int forward) +{ + struct lock_list *entry; + struct list_head *head; + struct circular_queue *cq = &lock_cq; + int ret = 1; + + if (match(source_entry, data)) { + *target_entry = source_entry; + ret = 0; + goto exit; + } + + if (forward) + head = &source_entry->class->locks_after; + else + head = &source_entry->class->locks_before; + + if (list_empty(head)) + goto exit; + + __cq_init(cq); + __cq_enqueue(cq, (unsigned long)source_entry); + + while (!__cq_empty(cq)) { + struct lock_list *lock; + + __cq_dequeue(cq, (unsigned long *)&lock); + + if (!lock->class) { + ret = -2; + goto exit; + } + + if (forward) + head = &lock->class->locks_after; + else + head = &lock->class->locks_before; + + list_for_each_entry(entry, head, entry) { + if (!lock_accessed(entry)) { + unsigned int cq_depth; + mark_lock_accessed(entry, lock); + if (match(entry, data)) { + *target_entry = entry; + ret = 0; + goto exit; + } + + if (__cq_enqueue(cq, (unsigned long)entry)) { + ret = -1; + goto exit; + } + cq_depth = __cq_get_elem_count(cq); + if (max_bfs_queue_depth < cq_depth) + max_bfs_queue_depth = cq_depth; + } + } + } +exit: + return ret; +} + +static inline int __bfs_forwards(struct lock_list *src_entry, + void *data, + int (*match)(struct lock_list *entry, void *data), + struct lock_list **target_entry) +{ + return __bfs(src_entry, data, match, target_entry, 1); + +} + +static inline int __bfs_backwards(struct lock_list *src_entry, + void *data, + int (*match)(struct lock_list *entry, void *data), + struct lock_list **target_entry) +{ + return __bfs(src_entry, data, match, target_entry, 0); + +} + +/* * Recursive, forwards-direction lock-dependency checking, used for * both noncyclic checking and for hardirq-unsafe/softirq-unsafe * checking. - * - * (to keep the stackframe of the recursive functions small we - * use these global variables, and we also mark various helper - * functions as noinline.) */ -static struct held_lock *check_source, *check_target; /* * Print a dependency chain entry (this is only done when a deadlock * has been detected): */ static noinline int -print_circular_bug_entry(struct lock_list *target, unsigned int depth) +print_circular_bug_entry(struct lock_list *target, int depth) { if (debug_locks_silent) return 0; @@ -929,11 +1053,13 @@ print_circular_bug_entry(struct lock_list *target, unsigned int depth) * header first: */ static noinline int -print_circular_bug_header(struct lock_list *entry, unsigned int depth) +print_circular_bug_header(struct lock_list *entry, unsigned int depth, + struct held_lock *check_src, + struct held_lock *check_tgt) { struct task_struct *curr = current; - if (!debug_locks_off_graph_unlock() || debug_locks_silent) + if (debug_locks_silent) return 0; printk("\n=======================================================\n"); @@ -942,9 +1068,9 @@ print_circular_bug_header(struct lock_list *entry, unsigned int depth) printk( "-------------------------------------------------------\n"); printk("%s/%d is trying to acquire lock:\n", curr->comm, task_pid_nr(curr)); - print_lock(check_source); + print_lock(check_src); printk("\nbut task is already holding lock:\n"); - print_lock(check_target); + print_lock(check_tgt); printk("\nwhich lock already depends on the new lock.\n\n"); printk("\nthe existing dependency chain (in reverse order) is:\n"); @@ -953,19 +1079,36 @@ print_circular_bug_header(struct lock_list *entry, unsigned int depth) return 0; } -static noinline int print_circular_bug_tail(void) +static inline int class_equal(struct lock_list *entry, void *data) +{ + return entry->class == data; +} + +static noinline int print_circular_bug(struct lock_list *this, + struct lock_list *target, + struct held_lock *check_src, + struct held_lock *check_tgt) { struct task_struct *curr = current; - struct lock_list this; + struct lock_list *parent; + int depth; - if (debug_locks_silent) + if (!debug_locks_off_graph_unlock() || debug_locks_silent) return 0; - this.class = hlock_class(check_source); - if (!save_trace(&this.trace)) + if (!save_trace(&this->trace)) return 0; - print_circular_bug_entry(&this, 0); + depth = get_lock_depth(target); + + print_circular_bug_header(target, depth, check_src, check_tgt); + + parent = get_lock_parent(target); + + while (parent) { + print_circular_bug_entry(parent, --depth); + parent = get_lock_parent(parent); + } printk("\nother info that might help us debug this:\n\n"); lockdep_print_held_locks(curr); @@ -976,73 +1119,69 @@ static noinline int print_circular_bug_tail(void) return 0; } -#define RECURSION_LIMIT 40 - -static int noinline print_infinite_recursion_bug(void) +static noinline int print_bfs_bug(int ret) { if (!debug_locks_off_graph_unlock()) return 0; - WARN_ON(1); + WARN(1, "lockdep bfs error:%d\n", ret); return 0; } -unsigned long __lockdep_count_forward_deps(struct lock_class *class, - unsigned int depth) +static int noop_count(struct lock_list *entry, void *data) { - struct lock_list *entry; - unsigned long ret = 1; + (*(unsigned long *)data)++; + return 0; +} - if (lockdep_dependency_visit(class, depth)) - return 0; +unsigned long __lockdep_count_forward_deps(struct lock_list *this) +{ + unsigned long count = 0; + struct lock_list *uninitialized_var(target_entry); - /* - * Recurse this class's dependency list: - */ - list_for_each_entry(entry, &class->locks_after, entry) - ret += __lockdep_count_forward_deps(entry->class, depth + 1); + __bfs_forwards(this, (void *)&count, noop_count, &target_entry); - return ret; + return count; } - unsigned long lockdep_count_forward_deps(struct lock_class *class) { unsigned long ret, flags; + struct lock_list this; + + this.parent = NULL; + this.class = class; local_irq_save(flags); __raw_spin_lock(&lockdep_lock); - ret = __lockdep_count_forward_deps(class, 0); + ret = __lockdep_count_forward_deps(&this); __raw_spin_unlock(&lockdep_lock); local_irq_restore(flags); return ret; } -unsigned long __lockdep_count_backward_deps(struct lock_class *class, - unsigned int depth) +unsigned long __lockdep_count_backward_deps(struct lock_list *this) { - struct lock_list *entry; - unsigned long ret = 1; + unsigned long count = 0; + struct lock_list *uninitialized_var(target_entry); - if (lockdep_dependency_visit(class, depth)) - return 0; - /* - * Recurse this class's dependency list: - */ - list_for_each_entry(entry, &class->locks_before, entry) - ret += __lockdep_count_backward_deps(entry->class, depth + 1); + __bfs_backwards(this, (void *)&count, noop_count, &target_entry); - return ret; + return count; } unsigned long lockdep_count_backward_deps(struct lock_class *class) { unsigned long ret, flags; + struct lock_list this; + + this.parent = NULL; + this.class = class; local_irq_save(flags); __raw_spin_lock(&lockdep_lock); - ret = __lockdep_count_backward_deps(class, 0); + ret = __lockdep_count_backward_deps(&this); __raw_spin_unlock(&lockdep_lock); local_irq_restore(flags); @@ -1054,29 +1193,16 @@ unsigned long lockdep_count_backward_deps(struct lock_class *class) * lead to <target>. Print an error and return 0 if it does. */ static noinline int -check_noncircular(struct lock_class *source, unsigned int depth) +check_noncircular(struct lock_list *root, struct lock_class *target, + struct lock_list **target_entry) { - struct lock_list *entry; + int result; - if (lockdep_dependency_visit(source, depth)) - return 1; + debug_atomic_inc(&nr_cyclic_checks); - debug_atomic_inc(&nr_cyclic_check_recursions); - if (depth > max_recursion_depth) - max_recursion_depth = depth; - if (depth >= RECURSION_LIMIT) - return print_infinite_recursion_bug(); - /* - * Check this lock's dependency list: - */ - list_for_each_entry(entry, &source->locks_after, entry) { - if (entry->class == hlock_class(check_target)) - return print_circular_bug_header(entry, depth+1); - debug_atomic_inc(&nr_cyclic_checks); - if (!check_noncircular(entry->class, depth+1)) - return print_circular_bug_entry(entry, depth+1); - } - return 1; + result = __bfs_forwards(root, target, class_equal, target_entry); + + return result; } #if defined(CONFIG_TRACE_IRQFLAGS) && defined(CONFIG_PROVE_LOCKING) @@ -1085,103 +1211,121 @@ check_noncircular(struct lock_class *source, unsigned int depth) * proving that two subgraphs can be connected by a new dependency * without creating any illegal irq-safe -> irq-unsafe lock dependency. */ -static enum lock_usage_bit find_usage_bit; -static struct lock_class *forwards_match, *backwards_match; + +static inline int usage_match(struct lock_list *entry, void *bit) +{ + return entry->class->usage_mask & (1 << (enum lock_usage_bit)bit); +} + + /* * Find a node in the forwards-direction dependency sub-graph starting - * at <source> that matches <find_usage_bit>. + * at @root->class that matches @bit. * - * Return 2 if such a node exists in the subgraph, and put that node - * into <forwards_match>. + * Return 0 if such a node exists in the subgraph, and put that node + * into *@target_entry. * - * Return 1 otherwise and keep <forwards_match> unchanged. - * Return 0 on error. + * Return 1 otherwise and keep *@target_entry unchanged. + * Return <0 on error. */ -static noinline int -find_usage_forwards(struct lock_class *source, unsigned int depth) +static int +find_usage_forwards(struct lock_list *root, enum lock_usage_bit bit, + struct lock_list **target_entry) { - struct lock_list *entry; - int ret; - - if (lockdep_dependency_visit(source, depth)) - return 1; - - if (depth > max_recursion_depth) - max_recursion_depth = depth; - if (depth >= RECURSION_LIMIT) - return print_infinite_recursion_bug(); + int result; debug_atomic_inc(&nr_find_usage_forwards_checks); - if (source->usage_mask & (1 << find_usage_bit)) { - forwards_match = source; - return 2; - } - /* - * Check this lock's dependency list: - */ - list_for_each_entry(entry, &source->locks_after, entry) { - debug_atomic_inc(&nr_find_usage_forwards_recursions); - ret = find_usage_forwards(entry->class, depth+1); - if (ret == 2 || ret == 0) - return ret; - } - return 1; + result = __bfs_forwards(root, (void *)bit, usage_match, target_entry); + + return result; } /* * Find a node in the backwards-direction dependency sub-graph starting - * at <source> that matches <find_usage_bit>. + * at @root->class that matches @bit. * - * Return 2 if such a node exists in the subgraph, and put that node - * into <backwards_match>. + * Return 0 if such a node exists in the subgraph, and put that node + * into *@target_entry. * - * Return 1 otherwise and keep <backwards_match> unchanged. - * Return 0 on error. + * Return 1 otherwise and keep *@target_entry unchanged. + * Return <0 on error. */ -static noinline int -find_usage_backwards(struct lock_class *source, unsigned int depth) +static int +find_usage_backwards(struct lock_list *root, enum lock_usage_bit bit, + struct lock_list **target_entry) { - struct lock_list *entry; - int ret; + int result; - if (lockdep_dependency_visit(source, depth)) - return 1; + debug_atomic_inc(&nr_find_usage_backwards_checks); - if (!__raw_spin_is_locked(&lockdep_lock)) - return DEBUG_LOCKS_WARN_ON(1); + result = __bfs_backwards(root, (void *)bit, usage_match, target_entry); - if (depth > max_recursion_depth) - max_recursion_depth = depth; - if (depth >= RECURSION_LIMIT) - return print_infinite_recursion_bug(); + return result; +} - debug_atomic_inc(&nr_find_usage_backwards_checks); - if (source->usage_mask & (1 << find_usage_bit)) { - backwards_match = source; - return 2; - } +static void print_lock_class_header(struct lock_class *class, int depth) +{ + int bit; - if (!source && debug_locks_off_graph_unlock()) { - WARN_ON(1); - return 0; - } + printk("%*s->", depth, ""); + print_lock_name(class); + printk(" ops: %lu", class->ops); + printk(" {\n"); - /* - * Check this lock's dependency list: - */ - list_for_each_entry(entry, &source->locks_before, entry) { - debug_atomic_inc(&nr_find_usage_backwards_recursions); - ret = find_usage_backwards(entry->class, depth+1); - if (ret == 2 || ret == 0) - return ret; + for (bit = 0; bit < LOCK_USAGE_STATES; bit++) { + if (class->usage_mask & (1 << bit)) { + int len = depth; + + len += printk("%*s %s", depth, "", usage_str[bit]); + len += printk(" at:\n"); + print_stack_trace(class->usage_traces + bit, len); + } } - return 1; + printk("%*s }\n", depth, ""); + + printk("%*s ... key at: ",depth,""); + print_ip_sym((unsigned long)class->key); +} + +/* + * printk the shortest lock dependencies from @start to @end in reverse order: + */ +static void __used +print_shortest_lock_dependencies(struct lock_list *leaf, + struct lock_list *root) +{ + struct lock_list *entry = leaf; + int depth; + + /*compute depth from generated tree by BFS*/ + depth = get_lock_depth(leaf); + + do { + print_lock_class_header(entry->class, depth); + printk("%*s ... acquired at:\n", depth, ""); + print_stack_trace(&entry->trace, 2); + printk("\n"); + + if (depth == 0 && (entry != root)) { + printk("lockdep:%s bad BFS generated tree\n", __func__); + break; + } + + entry = get_lock_parent(entry); + depth--; + } while (entry && (depth >= 0)); + + return; } static int print_bad_irq_dependency(struct task_struct *curr, + struct lock_list *prev_root, + struct lock_list *next_root, + struct lock_list *backwards_entry, + struct lock_list *forwards_entry, struct held_lock *prev, struct held_lock *next, enum lock_usage_bit bit1, @@ -1214,26 +1358,32 @@ print_bad_irq_dependency(struct task_struct *curr, printk("\nbut this new dependency connects a %s-irq-safe lock:\n", irqclass); - print_lock_name(backwards_match); + print_lock_name(backwards_entry->class); printk("\n... which became %s-irq-safe at:\n", irqclass); - print_stack_trace(backwards_match->usage_traces + bit1, 1); + print_stack_trace(backwards_entry->class->usage_traces + bit1, 1); printk("\nto a %s-irq-unsafe lock:\n", irqclass); - print_lock_name(forwards_match); + print_lock_name(forwards_entry->class); printk("\n... which became %s-irq-unsafe at:\n", irqclass); printk("..."); - print_stack_trace(forwards_match->usage_traces + bit2, 1); + print_stack_trace(forwards_entry->class->usage_traces + bit2, 1); printk("\nother info that might help us debug this:\n\n"); lockdep_print_held_locks(curr); - printk("\nthe %s-irq-safe lock's dependencies:\n", irqclass); - print_lock_dependencies(backwards_match, 0); + printk("\nthe dependencies between %s-irq-safe lock", irqclass); + printk(" and the holding lock:\n"); + if (!save_trace(&prev_root->trace)) + return 0; + print_shortest_lock_dependencies(backwards_entry, prev_root); - printk("\nthe %s-irq-unsafe lock's dependencies:\n", irqclass); - print_lock_dependencies(forwards_match, 0); + printk("\nthe dependencies between the lock to be acquired"); + printk(" and %s-irq-unsafe lock:\n", irqclass); + if (!save_trace(&next_root->trace)) + return 0; + print_shortest_lock_dependencies(forwards_entry, next_root); printk("\nstack backtrace:\n"); dump_stack(); @@ -1247,25 +1397,76 @@ check_usage(struct task_struct *curr, struct held_lock *prev, enum lock_usage_bit bit_forwards, const char *irqclass) { int ret; + struct lock_list this, that; + struct lock_list *uninitialized_var(target_entry); + struct lock_list *uninitialized_var(target_entry1); + + this.parent = NULL; - find_usage_bit = bit_backwards; - /* fills in <backwards_match> */ - ret = find_usage_backwards(hlock_class(prev), 0); - if (!ret || ret == 1) + this.class = hlock_class(prev); + ret = find_usage_backwards(&this, bit_backwards, &target_entry); + if (ret < 0) + return print_bfs_bug(ret); + if (ret == 1) return ret; - find_usage_bit = bit_forwards; - ret = find_usage_forwards(hlock_class(next), 0); - if (!ret || ret == 1) + that.parent = NULL; + that.class = hlock_class(next); + ret = find_usage_forwards(&that, bit_forwards, &target_entry1); + if (ret < 0) + return print_bfs_bug(ret); + if (ret == 1) return ret; - /* ret == 2 */ - return print_bad_irq_dependency(curr, prev, next, + + return print_bad_irq_dependency(curr, &this, &that, + target_entry, target_entry1, + prev, next, bit_backwards, bit_forwards, irqclass); } -static int -check_prev_add_irq(struct task_struct *curr, struct held_lock *prev, - struct held_lock *next) +static const char *state_names[] = { +#define LOCKDEP_STATE(__STATE) \ + __stringify(__STATE), +#include "lockdep_states.h" +#undef LOCKDEP_STATE +}; + +static const char *state_rnames[] = { +#define LOCKDEP_STATE(__STATE) \ + __stringify(__STATE)"-READ", +#include "lockdep_states.h" +#undef LOCKDEP_STATE +}; + +static inline const char *state_name(enum lock_usage_bit bit) +{ + return (bit & 1) ? state_rnames[bit >> 2] : state_names[bit >> 2]; +} + +static int exclusive_bit(int new_bit) +{ + /* + * USED_IN + * USED_IN_READ + * ENABLED + * ENABLED_READ + * + * bit 0 - write/read + * bit 1 - used_in/enabled + * bit 2+ state + */ + + int state = new_bit & ~3; + int dir = new_bit & 2; + + /* + * keep state, bit flip the direction and strip read. + */ + return state | (dir ^ 2); +} + +static int check_irq_usage(struct task_struct *curr, struct held_lock *prev, + struct held_lock *next, enum lock_usage_bit bit) { /* * Prove that the new dependency does not connect a hardirq-safe @@ -1273,38 +1474,34 @@ check_prev_add_irq(struct task_struct *curr, struct held_lock *prev, * the backwards-subgraph starting at <prev>, and the * forwards-subgraph starting at <next>: */ - if (!check_usage(curr, prev, next, LOCK_USED_IN_HARDIRQ, - LOCK_ENABLED_HARDIRQS, "hard")) + if (!check_usage(curr, prev, next, bit, + exclusive_bit(bit), state_name(bit))) return 0; + bit++; /* _READ */ + /* * Prove that the new dependency does not connect a hardirq-safe-read * lock with a hardirq-unsafe lock - to achieve this we search * the backwards-subgraph starting at <prev>, and the * forwards-subgraph starting at <next>: */ - if (!check_usage(curr, prev, next, LOCK_USED_IN_HARDIRQ_READ, - LOCK_ENABLED_HARDIRQS, "hard-read")) + if (!check_usage(curr, prev, next, bit, + exclusive_bit(bit), state_name(bit))) return 0; - /* - * Prove that the new dependency does not connect a softirq-safe - * lock with a softirq-unsafe lock - to achieve this we search - * the backwards-subgraph starting at <prev>, and the - * forwards-subgraph starting at <next>: - */ - if (!check_usage(curr, prev, next, LOCK_USED_IN_SOFTIRQ, - LOCK_ENABLED_SOFTIRQS, "soft")) - return 0; - /* - * Prove that the new dependency does not connect a softirq-safe-read - * lock with a softirq-unsafe lock - to achieve this we search - * the backwards-subgraph starting at <prev>, and the - * forwards-subgraph starting at <next>: - */ - if (!check_usage(curr, prev, next, LOCK_USED_IN_SOFTIRQ_READ, - LOCK_ENABLED_SOFTIRQS, "soft")) + return 1; +} + +static int +check_prev_add_irq(struct task_struct *curr, struct held_lock *prev, + struct held_lock *next) +{ +#define LOCKDEP_STATE(__STATE) \ + if (!check_irq_usage(curr, prev, next, LOCK_USED_IN_##__STATE)) \ return 0; +#include "lockdep_states.h" +#undef LOCKDEP_STATE return 1; } @@ -1435,6 +1632,8 @@ check_prev_add(struct task_struct *curr, struct held_lock *prev, { struct lock_list *entry; int ret; + struct lock_list this; + struct lock_list *uninitialized_var(target_entry); /* * Prove that the new <prev> -> <next> dependency would not @@ -1445,10 +1644,13 @@ check_prev_add(struct task_struct *curr, struct held_lock *prev, * We are using global variables to control the recursion, to * keep the stackframe size of the recursive functions low: */ - check_source = next; - check_target = prev; - if (!(check_noncircular(hlock_class(next), 0))) - return print_circular_bug_tail(); + this.class = hlock_class(next); + this.parent = NULL; + ret = check_noncircular(&this, hlock_class(prev), &target_entry); + if (unlikely(!ret)) + return print_circular_bug(&this, target_entry, next, prev); + else if (unlikely(ret < 0)) + return print_bfs_bug(ret); if (!check_prev_add_irq(curr, prev, next)) return 0; @@ -1649,6 +1851,7 @@ cache_hit: printk("BUG: MAX_LOCKDEP_CHAINS too low!\n"); printk("turning off the locking correctness validator.\n"); + dump_stack(); return 0; } chain = lock_chains + nr_lock_chains++; @@ -1846,7 +2049,8 @@ static int mark_lock(struct task_struct *curr, struct held_lock *this, * print irq inversion bug: */ static int -print_irq_inversion_bug(struct task_struct *curr, struct lock_class *other, +print_irq_inversion_bug(struct task_struct *curr, + struct lock_list *root, struct lock_list *other, struct held_lock *this, int forwards, const char *irqclass) { @@ -1861,20 +2065,19 @@ print_irq_inversion_bug(struct task_struct *curr, struct lock_class *other, curr->comm, task_pid_nr(curr)); print_lock(this); if (forwards) - printk("but this lock took another, %s-irq-unsafe lock in the past:\n", irqclass); + printk("but this lock took another, %s-unsafe lock in the past:\n", irqclass); else - printk("but this lock was taken by another, %s-irq-safe lock in the past:\n", irqclass); - print_lock_name(other); + printk("but this lock was taken by another, %s-safe lock in the past:\n", irqclass); + print_lock_name(other->class); printk("\n\nand interrupts could create inverse lock ordering between them.\n\n"); printk("\nother info that might help us debug this:\n"); lockdep_print_held_locks(curr); - printk("\nthe first lock's dependencies:\n"); - print_lock_dependencies(hlock_class(this), 0); - - printk("\nthe second lock's dependencies:\n"); - print_lock_dependencies(other, 0); + printk("\nthe shortest dependencies between 2nd lock and 1st lock:\n"); + if (!save_trace(&root->trace)) + return 0; + print_shortest_lock_dependencies(other, root); printk("\nstack backtrace:\n"); dump_stack(); @@ -1891,14 +2094,19 @@ check_usage_forwards(struct task_struct *curr, struct held_lock *this, enum lock_usage_bit bit, const char *irqclass) { int ret; - - find_usage_bit = bit; - /* fills in <forwards_match> */ - ret = find_usage_forwards(hlock_class(this), 0); - if (!ret || ret == 1) + struct lock_list root; + struct lock_list *uninitialized_var(target_entry); + + root.parent = NULL; + root.class = hlock_class(this); + ret = find_usage_forwards(&root, bit, &target_entry); + if (ret < 0) + return print_bfs_bug(ret); + if (ret == 1) return ret; - return print_irq_inversion_bug(curr, forwards_match, this, 1, irqclass); + return print_irq_inversion_bug(curr, &root, target_entry, + this, 1, irqclass); } /* @@ -1910,14 +2118,19 @@ check_usage_backwards(struct task_struct *curr, struct held_lock *this, enum lock_usage_bit bit, const char *irqclass) { int ret; - - find_usage_bit = bit; - /* fills in <backwards_match> */ - ret = find_usage_backwards(hlock_class(this), 0); - if (!ret || ret == 1) + struct lock_list root; + struct lock_list *uninitialized_var(target_entry); + + root.parent = NULL; + root.class = hlock_class(this); + ret = find_usage_backwards(&root, bit, &target_entry); + if (ret < 0) + return print_bfs_bug(ret); + if (ret == 1) return ret; - return print_irq_inversion_bug(curr, backwards_match, this, 0, irqclass); + return print_irq_inversion_bug(curr, &root, target_entry, + this, 1, irqclass); } void print_irqtrace_events(struct task_struct *curr) @@ -1933,7 +2146,7 @@ void print_irqtrace_events(struct task_struct *curr) print_ip_sym(curr->softirq_disable_ip); } -static int hardirq_verbose(struct lock_class *class) +static int HARDIRQ_verbose(struct lock_class *class) { #if HARDIRQ_VERBOSE return class_filter(class); @@ -1941,7 +2154,7 @@ static int hardirq_verbose(struct lock_class *class) return 0; } -static int softirq_verbose(struct lock_class *class) +static int SOFTIRQ_verbose(struct lock_class *class) { #if SOFTIRQ_VERBOSE return class_filter(class); @@ -1949,185 +2162,95 @@ static int softirq_verbose(struct lock_class *class) return 0; } +static int RECLAIM_FS_verbose(struct lock_class *class) +{ +#if RECLAIM_VERBOSE + return class_filter(class); +#endif + return 0; +} + #define STRICT_READ_CHECKS 1 -static int mark_lock_irq(struct task_struct *curr, struct held_lock *this, +static int (*state_verbose_f[])(struct lock_class *class) = { +#define LOCKDEP_STATE(__STATE) \ + __STATE##_verbose, +#include "lockdep_states.h" +#undef LOCKDEP_STATE +}; + +static inline int state_verbose(enum lock_usage_bit bit, + struct lock_class *class) +{ + return state_verbose_f[bit >> 2](class); +} + +typedef int (*check_usage_f)(struct task_struct *, struct held_lock *, + enum lock_usage_bit bit, const char *name); + +static int +mark_lock_irq(struct task_struct *curr, struct held_lock *this, enum lock_usage_bit new_bit) { - int ret = 1; + int excl_bit = exclusive_bit(new_bit); + int read = new_bit & 1; + int dir = new_bit & 2; - switch(new_bit) { - case LOCK_USED_IN_HARDIRQ: - if (!valid_state(curr, this, new_bit, LOCK_ENABLED_HARDIRQS)) - return 0; - if (!valid_state(curr, this, new_bit, - LOCK_ENABLED_HARDIRQS_READ)) - return 0; - /* - * just marked it hardirq-safe, check that this lock - * took no hardirq-unsafe lock in the past: - */ - if (!check_usage_forwards(curr, this, - LOCK_ENABLED_HARDIRQS, "hard")) - return 0; -#if STRICT_READ_CHECKS - /* - * just marked it hardirq-safe, check that this lock - * took no hardirq-unsafe-read lock in the past: - */ - if (!check_usage_forwards(curr, this, - LOCK_ENABLED_HARDIRQS_READ, "hard-read")) - return 0; -#endif - if (hardirq_verbose(hlock_class(this))) - ret = 2; - break; - case LOCK_USED_IN_SOFTIRQ: - if (!valid_state(curr, this, new_bit, LOCK_ENABLED_SOFTIRQS)) - return 0; - if (!valid_state(curr, this, new_bit, - LOCK_ENABLED_SOFTIRQS_READ)) - return 0; - /* - * just marked it softirq-safe, check that this lock - * took no softirq-unsafe lock in the past: - */ - if (!check_usage_forwards(curr, this, - LOCK_ENABLED_SOFTIRQS, "soft")) - return 0; -#if STRICT_READ_CHECKS - /* - * just marked it softirq-safe, check that this lock - * took no softirq-unsafe-read lock in the past: - */ - if (!check_usage_forwards(curr, this, - LOCK_ENABLED_SOFTIRQS_READ, "soft-read")) - return 0; -#endif - if (softirq_verbose(hlock_class(this))) - ret = 2; - break; - case LOCK_USED_IN_HARDIRQ_READ: - if (!valid_state(curr, this, new_bit, LOCK_ENABLED_HARDIRQS)) - return 0; - /* - * just marked it hardirq-read-safe, check that this lock - * took no hardirq-unsafe lock in the past: - */ - if (!check_usage_forwards(curr, this, - LOCK_ENABLED_HARDIRQS, "hard")) - return 0; - if (hardirq_verbose(hlock_class(this))) - ret = 2; - break; - case LOCK_USED_IN_SOFTIRQ_READ: - if (!valid_state(curr, this, new_bit, LOCK_ENABLED_SOFTIRQS)) - return 0; - /* - * just marked it softirq-read-safe, check that this lock - * took no softirq-unsafe lock in the past: - */ - if (!check_usage_forwards(curr, this, - LOCK_ENABLED_SOFTIRQS, "soft")) - return 0; - if (softirq_verbose(hlock_class(this))) - ret = 2; - break; - case LOCK_ENABLED_HARDIRQS: - if (!valid_state(curr, this, new_bit, LOCK_USED_IN_HARDIRQ)) - return 0; - if (!valid_state(curr, this, new_bit, - LOCK_USED_IN_HARDIRQ_READ)) - return 0; - /* - * just marked it hardirq-unsafe, check that no hardirq-safe - * lock in the system ever took it in the past: - */ - if (!check_usage_backwards(curr, this, - LOCK_USED_IN_HARDIRQ, "hard")) - return 0; -#if STRICT_READ_CHECKS - /* - * just marked it hardirq-unsafe, check that no - * hardirq-safe-read lock in the system ever took - * it in the past: - */ - if (!check_usage_backwards(curr, this, - LOCK_USED_IN_HARDIRQ_READ, "hard-read")) - return 0; -#endif - if (hardirq_verbose(hlock_class(this))) - ret = 2; - break; - case LOCK_ENABLED_SOFTIRQS: - if (!valid_state(curr, this, new_bit, LOCK_USED_IN_SOFTIRQ)) - return 0; - if (!valid_state(curr, this, new_bit, - LOCK_USED_IN_SOFTIRQ_READ)) - return 0; - /* - * just marked it softirq-unsafe, check that no softirq-safe - * lock in the system ever took it in the past: - */ - if (!check_usage_backwards(curr, this, - LOCK_USED_IN_SOFTIRQ, "soft")) - return 0; -#if STRICT_READ_CHECKS - /* - * just marked it softirq-unsafe, check that no - * softirq-safe-read lock in the system ever took - * it in the past: - */ - if (!check_usage_backwards(curr, this, - LOCK_USED_IN_SOFTIRQ_READ, "soft-read")) - return 0; -#endif - if (softirq_verbose(hlock_class(this))) - ret = 2; - break; - case LOCK_ENABLED_HARDIRQS_READ: - if (!valid_state(curr, this, new_bit, LOCK_USED_IN_HARDIRQ)) - return 0; -#if STRICT_READ_CHECKS - /* - * just marked it hardirq-read-unsafe, check that no - * hardirq-safe lock in the system ever took it in the past: - */ - if (!check_usage_backwards(curr, this, - LOCK_USED_IN_HARDIRQ, "hard")) - return 0; -#endif - if (hardirq_verbose(hlock_class(this))) - ret = 2; - break; - case LOCK_ENABLED_SOFTIRQS_READ: - if (!valid_state(curr, this, new_bit, LOCK_USED_IN_SOFTIRQ)) + /* + * mark USED_IN has to look forwards -- to ensure no dependency + * has ENABLED state, which would allow recursion deadlocks. + * + * mark ENABLED has to look backwards -- to ensure no dependee + * has USED_IN state, which, again, would allow recursion deadlocks. + */ + check_usage_f usage = dir ? + check_usage_backwards : check_usage_forwards; + + /* + * Validate that this particular lock does not have conflicting + * usage states. + */ + if (!valid_state(curr, this, new_bit, excl_bit)) + return 0; + + /* + * Validate that the lock dependencies don't have conflicting usage + * states. + */ + if ((!read || !dir || STRICT_READ_CHECKS) && + !usage(curr, this, excl_bit, state_name(new_bit & ~1))) + return 0; + + /* + * Check for read in write conflicts + */ + if (!read) { + if (!valid_state(curr, this, new_bit, excl_bit + 1)) return 0; -#if STRICT_READ_CHECKS - /* - * just marked it softirq-read-unsafe, check that no - * softirq-safe lock in the system ever took it in the past: - */ - if (!check_usage_backwards(curr, this, - LOCK_USED_IN_SOFTIRQ, "soft")) + + if (STRICT_READ_CHECKS && + !usage(curr, this, excl_bit + 1, + state_name(new_bit + 1))) return 0; -#endif - if (softirq_verbose(hlock_class(this))) - ret = 2; - break; - default: - WARN_ON(1); - break; } - return ret; + if (state_verbose(new_bit, hlock_class(this))) + return 2; + + return 1; } +enum mark_type { +#define LOCKDEP_STATE(__STATE) __STATE, +#include "lockdep_states.h" +#undef LOCKDEP_STATE +}; + /* * Mark all held locks with a usage bit: */ static int -mark_held_locks(struct task_struct *curr, int hardirq) +mark_held_locks(struct task_struct *curr, enum mark_type mark) { enum lock_usage_bit usage_bit; struct held_lock *hlock; @@ -2136,17 +2259,12 @@ mark_held_locks(struct task_struct *curr, int hardirq) for (i = 0; i < curr->lockdep_depth; i++) { hlock = curr->held_locks + i; - if (hardirq) { - if (hlock->read) - usage_bit = LOCK_ENABLED_HARDIRQS_READ; - else - usage_bit = LOCK_ENABLED_HARDIRQS; - } else { - if (hlock->read) - usage_bit = LOCK_ENABLED_SOFTIRQS_READ; - else - usage_bit = LOCK_ENABLED_SOFTIRQS; - } + usage_bit = 2 + (mark << 2); /* ENABLED */ + if (hlock->read) + usage_bit += 1; /* READ */ + + BUG_ON(usage_bit >= LOCK_USAGE_STATES); + if (!mark_lock(curr, hlock, usage_bit)) return 0; } @@ -2200,7 +2318,7 @@ void trace_hardirqs_on_caller(unsigned long ip) * We are going to turn hardirqs on, so set the * usage bit for all held locks: */ - if (!mark_held_locks(curr, 1)) + if (!mark_held_locks(curr, HARDIRQ)) return; /* * If we have softirqs enabled, then set the usage @@ -2208,7 +2326,7 @@ void trace_hardirqs_on_caller(unsigned long ip) * this bit from being set before) */ if (curr->softirqs_enabled) - if (!mark_held_locks(curr, 0)) + if (!mark_held_locks(curr, SOFTIRQ)) return; curr->hardirq_enable_ip = ip; @@ -2288,7 +2406,7 @@ void trace_softirqs_on(unsigned long ip) * enabled too: */ if (curr->hardirqs_enabled) - mark_held_locks(curr, 0); + mark_held_locks(curr, SOFTIRQ); } /* @@ -2317,6 +2435,48 @@ void trace_softirqs_off(unsigned long ip) debug_atomic_inc(&redundant_softirqs_off); } +static void __lockdep_trace_alloc(gfp_t gfp_mask, unsigned long flags) +{ + struct task_struct *curr = current; + + if (unlikely(!debug_locks)) + return; + + /* no reclaim without waiting on it */ + if (!(gfp_mask & __GFP_WAIT)) + return; + + /* this guy won't enter reclaim */ + if ((curr->flags & PF_MEMALLOC) && !(gfp_mask & __GFP_NOMEMALLOC)) + return; + + /* We're only interested __GFP_FS allocations for now */ + if (!(gfp_mask & __GFP_FS)) + return; + + if (DEBUG_LOCKS_WARN_ON(irqs_disabled_flags(flags))) + return; + + mark_held_locks(curr, RECLAIM_FS); +} + +static void check_flags(unsigned long flags); + +void lockdep_trace_alloc(gfp_t gfp_mask) +{ + unsigned long flags; + + if (unlikely(current->lockdep_recursion)) + return; + + raw_local_irq_save(flags); + check_flags(flags); + current->lockdep_recursion = 1; + __lockdep_trace_alloc(gfp_mask, flags); + current->lockdep_recursion = 0; + raw_local_irq_restore(flags); +} + static int mark_irqflags(struct task_struct *curr, struct held_lock *hlock) { /* @@ -2345,19 +2505,35 @@ static int mark_irqflags(struct task_struct *curr, struct held_lock *hlock) if (!hlock->hardirqs_off) { if (hlock->read) { if (!mark_lock(curr, hlock, - LOCK_ENABLED_HARDIRQS_READ)) + LOCK_ENABLED_HARDIRQ_READ)) return 0; if (curr->softirqs_enabled) if (!mark_lock(curr, hlock, - LOCK_ENABLED_SOFTIRQS_READ)) + LOCK_ENABLED_SOFTIRQ_READ)) return 0; } else { if (!mark_lock(curr, hlock, - LOCK_ENABLED_HARDIRQS)) + LOCK_ENABLED_HARDIRQ)) return 0; if (curr->softirqs_enabled) if (!mark_lock(curr, hlock, - LOCK_ENABLED_SOFTIRQS)) + LOCK_ENABLED_SOFTIRQ)) + return 0; + } + } + + /* + * We reuse the irq context infrastructure more broadly as a general + * context checking code. This tests GFP_FS recursion (a lock taken + * during reclaim for a GFP_FS allocation is held over a GFP_FS + * allocation). + */ + if (!hlock->trylock && (curr->lockdep_reclaim_gfp & __GFP_FS)) { + if (hlock->read) { + if (!mark_lock(curr, hlock, LOCK_USED_IN_RECLAIM_FS_READ)) + return 0; + } else { + if (!mark_lock(curr, hlock, LOCK_USED_IN_RECLAIM_FS)) return 0; } } @@ -2412,6 +2588,10 @@ static inline int separate_irq_context(struct task_struct *curr, return 0; } +void lockdep_trace_alloc(gfp_t gfp_mask) +{ +} + #endif /* @@ -2445,14 +2625,13 @@ static int mark_lock(struct task_struct *curr, struct held_lock *this, return 0; switch (new_bit) { - case LOCK_USED_IN_HARDIRQ: - case LOCK_USED_IN_SOFTIRQ: - case LOCK_USED_IN_HARDIRQ_READ: - case LOCK_USED_IN_SOFTIRQ_READ: - case LOCK_ENABLED_HARDIRQS: - case LOCK_ENABLED_SOFTIRQS: - case LOCK_ENABLED_HARDIRQS_READ: - case LOCK_ENABLED_SOFTIRQS_READ: +#define LOCKDEP_STATE(__STATE) \ + case LOCK_USED_IN_##__STATE: \ + case LOCK_USED_IN_##__STATE##_READ: \ + case LOCK_ENABLED_##__STATE: \ + case LOCK_ENABLED_##__STATE##_READ: +#include "lockdep_states.h" +#undef LOCKDEP_STATE ret = mark_lock_irq(curr, this, new_bit); if (!ret) return 0; @@ -2488,13 +2667,20 @@ static int mark_lock(struct task_struct *curr, struct held_lock *this, void lockdep_init_map(struct lockdep_map *lock, const char *name, struct lock_class_key *key, int subclass) { - if (unlikely(!debug_locks)) + lock->class_cache = NULL; +#ifdef CONFIG_LOCK_STAT + lock->cpu = raw_smp_processor_id(); +#endif + + if (DEBUG_LOCKS_WARN_ON(!name)) { + lock->name = "NULL"; return; + } + + lock->name = name; if (DEBUG_LOCKS_WARN_ON(!key)) return; - if (DEBUG_LOCKS_WARN_ON(!name)) - return; /* * Sanity check, the lock-class key must be persistent: */ @@ -2503,12 +2689,11 @@ void lockdep_init_map(struct lockdep_map *lock, const char *name, DEBUG_LOCKS_WARN_ON(1); return; } - lock->name = name; lock->key = key; - lock->class_cache = NULL; -#ifdef CONFIG_LOCK_STAT - lock->cpu = raw_smp_processor_id(); -#endif + + if (unlikely(!debug_locks)) + return; + if (subclass) register_lock_class(lock, subclass, 1); } @@ -2520,13 +2705,15 @@ EXPORT_SYMBOL_GPL(lockdep_init_map); */ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass, int trylock, int read, int check, int hardirqs_off, - struct lockdep_map *nest_lock, unsigned long ip) + struct lockdep_map *nest_lock, unsigned long ip, + int references) { struct task_struct *curr = current; struct lock_class *class = NULL; struct held_lock *hlock; unsigned int depth, id; int chain_head = 0; + int class_idx; u64 chain_key; if (!prove_locking) @@ -2542,6 +2729,7 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass, debug_locks_off(); printk("BUG: MAX_LOCKDEP_SUBCLASSES too low!\n"); printk("turning off the locking correctness validator.\n"); + dump_stack(); return 0; } @@ -2573,10 +2761,24 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass, if (DEBUG_LOCKS_WARN_ON(depth >= MAX_LOCK_DEPTH)) return 0; + class_idx = class - lock_classes + 1; + + if (depth) { + hlock = curr->held_locks + depth - 1; + if (hlock->class_idx == class_idx && nest_lock) { + if (hlock->references) + hlock->references++; + else + hlock->references = 2; + + return 1; + } + } + hlock = curr->held_locks + depth; if (DEBUG_LOCKS_WARN_ON(!class)) return 0; - hlock->class_idx = class - lock_classes + 1; + hlock->class_idx = class_idx; hlock->acquire_ip = ip; hlock->instance = lock; hlock->nest_lock = nest_lock; @@ -2584,6 +2786,7 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass, hlock->read = read; hlock->check = check; hlock->hardirqs_off = !!hardirqs_off; + hlock->references = references; #ifdef CONFIG_LOCK_STAT hlock->waittime_stamp = 0; hlock->holdtime_stamp = sched_clock(); @@ -2638,6 +2841,7 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass, debug_locks_off(); printk("BUG: MAX_LOCK_DEPTH too low!\n"); printk("turning off the locking correctness validator.\n"); + dump_stack(); return 0; } @@ -2691,6 +2895,30 @@ static int check_unlock(struct task_struct *curr, struct lockdep_map *lock, return 1; } +static int match_held_lock(struct held_lock *hlock, struct lockdep_map *lock) +{ + if (hlock->instance == lock) + return 1; + + if (hlock->references) { + struct lock_class *class = lock->class_cache; + + if (!class) + class = look_up_lock_class(lock, 0); + + if (DEBUG_LOCKS_WARN_ON(!class)) + return 0; + + if (DEBUG_LOCKS_WARN_ON(!hlock->nest_lock)) + return 0; + + if (hlock->class_idx == class - lock_classes + 1) + return 1; + } + + return 0; +} + static int __lock_set_class(struct lockdep_map *lock, const char *name, struct lock_class_key *key, unsigned int subclass, @@ -2714,7 +2942,7 @@ __lock_set_class(struct lockdep_map *lock, const char *name, */ if (prev_hlock && prev_hlock->irq_context != hlock->irq_context) break; - if (hlock->instance == lock) + if (match_held_lock(hlock, lock)) goto found_it; prev_hlock = hlock; } @@ -2733,7 +2961,8 @@ found_it: if (!__lock_acquire(hlock->instance, hlock_class(hlock)->subclass, hlock->trylock, hlock->read, hlock->check, hlock->hardirqs_off, - hlock->nest_lock, hlock->acquire_ip)) + hlock->nest_lock, hlock->acquire_ip, + hlock->references)) return 0; } @@ -2772,20 +3001,34 @@ lock_release_non_nested(struct task_struct *curr, */ if (prev_hlock && prev_hlock->irq_context != hlock->irq_context) break; - if (hlock->instance == lock) + if (match_held_lock(hlock, lock)) goto found_it; prev_hlock = hlock; } return print_unlock_inbalance_bug(curr, lock, ip); found_it: - lock_release_holdtime(hlock); + if (hlock->instance == lock) + lock_release_holdtime(hlock); + + if (hlock->references) { + hlock->references--; + if (hlock->references) { + /* + * We had, and after removing one, still have + * references, the current lock stack is still + * valid. We're done! + */ + return 1; + } + } /* * We have the right lock to unlock, 'hlock' points to it. * Now we remove it from the stack, and add back the other * entries (if any), recalculating the hash along the way: */ + curr->lockdep_depth = i; curr->curr_chain_key = hlock->prev_chain_key; @@ -2794,7 +3037,8 @@ found_it: if (!__lock_acquire(hlock->instance, hlock_class(hlock)->subclass, hlock->trylock, hlock->read, hlock->check, hlock->hardirqs_off, - hlock->nest_lock, hlock->acquire_ip)) + hlock->nest_lock, hlock->acquire_ip, + hlock->references)) return 0; } @@ -2824,7 +3068,7 @@ static int lock_release_nested(struct task_struct *curr, /* * Is the unlock non-nested: */ - if (hlock->instance != lock) + if (hlock->instance != lock || hlock->references) return lock_release_non_nested(curr, lock, ip); curr->lockdep_depth--; @@ -2869,6 +3113,21 @@ __lock_release(struct lockdep_map *lock, int nested, unsigned long ip) check_chain_key(curr); } +static int __lock_is_held(struct lockdep_map *lock) +{ + struct task_struct *curr = current; + int i; + + for (i = 0; i < curr->lockdep_depth; i++) { + struct held_lock *hlock = curr->held_locks + i; + + if (match_held_lock(hlock, lock)) + return 1; + } + + return 0; +} + /* * Check whether we follow the irq-flags state precisely: */ @@ -2935,6 +3194,8 @@ void lock_acquire(struct lockdep_map *lock, unsigned int subclass, { unsigned long flags; + trace_lock_acquire(lock, subclass, trylock, read, check, nest_lock, ip); + if (unlikely(current->lockdep_recursion)) return; @@ -2943,7 +3204,7 @@ void lock_acquire(struct lockdep_map *lock, unsigned int subclass, current->lockdep_recursion = 1; __lock_acquire(lock, subclass, trylock, read, check, - irqs_disabled_flags(flags), nest_lock, ip); + irqs_disabled_flags(flags), nest_lock, ip, 0); current->lockdep_recursion = 0; raw_local_irq_restore(flags); } @@ -2954,6 +3215,8 @@ void lock_release(struct lockdep_map *lock, int nested, { unsigned long flags; + trace_lock_release(lock, nested, ip); + if (unlikely(current->lockdep_recursion)) return; @@ -2966,6 +3229,36 @@ void lock_release(struct lockdep_map *lock, int nested, } EXPORT_SYMBOL_GPL(lock_release); +int lock_is_held(struct lockdep_map *lock) +{ + unsigned long flags; + int ret = 0; + + if (unlikely(current->lockdep_recursion)) + return ret; + + raw_local_irq_save(flags); + check_flags(flags); + + current->lockdep_recursion = 1; + ret = __lock_is_held(lock); + current->lockdep_recursion = 0; + raw_local_irq_restore(flags); + + return ret; +} +EXPORT_SYMBOL_GPL(lock_is_held); + +void lockdep_set_current_reclaim_state(gfp_t gfp_mask) +{ + current->lockdep_reclaim_gfp = gfp_mask; +} + +void lockdep_clear_current_reclaim_state(void) +{ + current->lockdep_reclaim_gfp = 0; +} + #ifdef CONFIG_LOCK_STAT static int print_lock_contention_bug(struct task_struct *curr, struct lockdep_map *lock, @@ -3015,7 +3308,7 @@ __lock_contended(struct lockdep_map *lock, unsigned long ip) */ if (prev_hlock && prev_hlock->irq_context != hlock->irq_context) break; - if (hlock->instance == lock) + if (match_held_lock(hlock, lock)) goto found_it; prev_hlock = hlock; } @@ -3023,6 +3316,9 @@ __lock_contended(struct lockdep_map *lock, unsigned long ip) return; found_it: + if (hlock->instance != lock) + return; + hlock->waittime_stamp = sched_clock(); contention_point = lock_point(hlock_class(hlock)->contention_point, ip); @@ -3062,7 +3358,7 @@ __lock_acquired(struct lockdep_map *lock, unsigned long ip) */ if (prev_hlock && prev_hlock->irq_context != hlock->irq_context) break; - if (hlock->instance == lock) + if (match_held_lock(hlock, lock)) goto found_it; prev_hlock = hlock; } @@ -3070,6 +3366,9 @@ __lock_acquired(struct lockdep_map *lock, unsigned long ip) return; found_it: + if (hlock->instance != lock) + return; + cpu = smp_processor_id(); if (hlock->waittime_stamp) { now = sched_clock(); @@ -3077,6 +3376,8 @@ found_it: hlock->holdtime_stamp = now; } + trace_lock_acquired(lock, ip, waittime); + stats = get_lock_stats(hlock_class(hlock)); if (waittime) { if (hlock->read) @@ -3096,6 +3397,8 @@ void lock_contended(struct lockdep_map *lock, unsigned long ip) { unsigned long flags; + trace_lock_contended(lock, ip); + if (unlikely(!lock_stat)) return; @@ -3296,7 +3599,12 @@ void __init lockdep_info(void) sizeof(struct list_head) * CLASSHASH_SIZE + sizeof(struct lock_list) * MAX_LOCKDEP_ENTRIES + sizeof(struct lock_chain) * MAX_LOCKDEP_CHAINS + - sizeof(struct list_head) * CHAINHASH_SIZE) / 1024); + sizeof(struct list_head) * CHAINHASH_SIZE +#ifdef CONFIG_PROVE_LOCKING + + sizeof(struct circular_queue) +#endif + ) / 1024 + ); printk(" per task-struct memory footprint: %lu bytes\n", sizeof(struct held_lock) * MAX_LOCK_DEPTH); diff --git a/kernel/lockdep_internals.h b/kernel/lockdep_internals.h index 56b196932c0..a2ee95ad131 100644 --- a/kernel/lockdep_internals.h +++ b/kernel/lockdep_internals.h @@ -7,6 +7,45 @@ */ /* + * Lock-class usage-state bits: + */ +enum lock_usage_bit { +#define LOCKDEP_STATE(__STATE) \ + LOCK_USED_IN_##__STATE, \ + LOCK_USED_IN_##__STATE##_READ, \ + LOCK_ENABLED_##__STATE, \ + LOCK_ENABLED_##__STATE##_READ, +#include "lockdep_states.h" +#undef LOCKDEP_STATE + LOCK_USED, + LOCK_USAGE_STATES +}; + +/* + * Usage-state bitmasks: + */ +#define __LOCKF(__STATE) LOCKF_##__STATE = (1 << LOCK_##__STATE), + +enum { +#define LOCKDEP_STATE(__STATE) \ + __LOCKF(USED_IN_##__STATE) \ + __LOCKF(USED_IN_##__STATE##_READ) \ + __LOCKF(ENABLED_##__STATE) \ + __LOCKF(ENABLED_##__STATE##_READ) +#include "lockdep_states.h" +#undef LOCKDEP_STATE + __LOCKF(USED) +}; + +#define LOCKF_ENABLED_IRQ (LOCKF_ENABLED_HARDIRQ | LOCKF_ENABLED_SOFTIRQ) +#define LOCKF_USED_IN_IRQ (LOCKF_USED_IN_HARDIRQ | LOCKF_USED_IN_SOFTIRQ) + +#define LOCKF_ENABLED_IRQ_READ \ + (LOCKF_ENABLED_HARDIRQ_READ | LOCKF_ENABLED_SOFTIRQ_READ) +#define LOCKF_USED_IN_IRQ_READ \ + (LOCKF_USED_IN_HARDIRQ_READ | LOCKF_USED_IN_SOFTIRQ_READ) + +/* * MAX_LOCKDEP_ENTRIES is the maximum number of lock dependencies * we track. * @@ -15,9 +54,9 @@ * table (if it's not there yet), and we check it for lock order * conflicts and deadlocks. */ -#define MAX_LOCKDEP_ENTRIES 8192UL +#define MAX_LOCKDEP_ENTRIES 16384UL -#define MAX_LOCKDEP_CHAINS_BITS 14 +#define MAX_LOCKDEP_CHAINS_BITS 15 #define MAX_LOCKDEP_CHAINS (1UL << MAX_LOCKDEP_CHAINS_BITS) #define MAX_LOCKDEP_CHAIN_HLOCKS (MAX_LOCKDEP_CHAINS*5) @@ -31,8 +70,10 @@ extern struct list_head all_lock_classes; extern struct lock_chain lock_chains[]; -extern void -get_usage_chars(struct lock_class *class, char *c1, char *c2, char *c3, char *c4); +#define LOCK_USAGE_CHARS (1+LOCK_USAGE_STATES/2) + +extern void get_usage_chars(struct lock_class *class, + char usage[LOCK_USAGE_CHARS]); extern const char * __get_key_name(struct lockdep_subclass_key *key, char *str); @@ -50,6 +91,8 @@ extern unsigned int nr_process_chains; extern unsigned int max_lockdep_depth; extern unsigned int max_recursion_depth; +extern unsigned int max_bfs_queue_depth; + #ifdef CONFIG_PROVE_LOCKING extern unsigned long lockdep_count_forward_deps(struct lock_class *); extern unsigned long lockdep_count_backward_deps(struct lock_class *); diff --git a/kernel/lockdep_proc.c b/kernel/lockdep_proc.c index 13716b81389..d4b3dbc79fd 100644 --- a/kernel/lockdep_proc.c +++ b/kernel/lockdep_proc.c @@ -25,38 +25,12 @@ static void *l_next(struct seq_file *m, void *v, loff_t *pos) { - struct lock_class *class; - - (*pos)++; - - if (v == SEQ_START_TOKEN) - class = m->private; - else { - class = v; - - if (class->lock_entry.next != &all_lock_classes) - class = list_entry(class->lock_entry.next, - struct lock_class, lock_entry); - else - class = NULL; - } - - return class; + return seq_list_next(v, &all_lock_classes, pos); } static void *l_start(struct seq_file *m, loff_t *pos) { - struct lock_class *class; - loff_t i = 0; - - if (*pos == 0) - return SEQ_START_TOKEN; - - list_for_each_entry(class, &all_lock_classes, lock_entry) { - if (++i == *pos) - return class; - } - return NULL; + return seq_list_start_head(&all_lock_classes, *pos); } static void l_stop(struct seq_file *m, void *v) @@ -82,11 +56,11 @@ static void print_name(struct seq_file *m, struct lock_class *class) static int l_show(struct seq_file *m, void *v) { - struct lock_class *class = v; + struct lock_class *class = list_entry(v, struct lock_class, lock_entry); struct lock_list *entry; - char c1, c2, c3, c4; + char usage[LOCK_USAGE_CHARS]; - if (v == SEQ_START_TOKEN) { + if (v == &all_lock_classes) { seq_printf(m, "all lock classes:\n"); return 0; } @@ -100,8 +74,8 @@ static int l_show(struct seq_file *m, void *v) seq_printf(m, " BD:%5ld", lockdep_count_backward_deps(class)); #endif - get_usage_chars(class, &c1, &c2, &c3, &c4); - seq_printf(m, " %c%c%c%c", c1, c2, c3, c4); + get_usage_chars(class, usage); + seq_printf(m, " %s", usage); seq_printf(m, ": "); print_name(m, class); @@ -128,17 +102,7 @@ static const struct seq_operations lockdep_ops = { static int lockdep_open(struct inode *inode, struct file *file) { - int res = seq_open(file, &lockdep_ops); - if (!res) { - struct seq_file *m = file->private_data; - - if (!list_empty(&all_lock_classes)) - m->private = list_entry(all_lock_classes.next, - struct lock_class, lock_entry); - else - m->private = NULL; - } - return res; + return seq_open(file, &lockdep_ops); } static const struct file_operations proc_lockdep_operations = { @@ -149,37 +113,23 @@ static const struct file_operations proc_lockdep_operations = { }; #ifdef CONFIG_PROVE_LOCKING -static void *lc_next(struct seq_file *m, void *v, loff_t *pos) -{ - struct lock_chain *chain; - - (*pos)++; - - if (v == SEQ_START_TOKEN) - chain = m->private; - else { - chain = v; - - if (*pos < nr_lock_chains) - chain = lock_chains + *pos; - else - chain = NULL; - } - - return chain; -} - static void *lc_start(struct seq_file *m, loff_t *pos) { if (*pos == 0) return SEQ_START_TOKEN; - if (*pos < nr_lock_chains) - return lock_chains + *pos; + if (*pos - 1 < nr_lock_chains) + return lock_chains + (*pos - 1); return NULL; } +static void *lc_next(struct seq_file *m, void *v, loff_t *pos) +{ + (*pos)++; + return lc_start(m, pos); +} + static void lc_stop(struct seq_file *m, void *v) { } @@ -220,16 +170,7 @@ static const struct seq_operations lockdep_chains_ops = { static int lockdep_chains_open(struct inode *inode, struct file *file) { - int res = seq_open(file, &lockdep_chains_ops); - if (!res) { - struct seq_file *m = file->private_data; - - if (nr_lock_chains) - m->private = lock_chains; - else - m->private = NULL; - } - return res; + return seq_open(file, &lockdep_chains_ops); } static const struct file_operations proc_lockdep_chains_operations = { @@ -258,16 +199,10 @@ static void lockdep_stats_debug_show(struct seq_file *m) debug_atomic_read(&chain_lookup_hits)); seq_printf(m, " cyclic checks: %11u\n", debug_atomic_read(&nr_cyclic_checks)); - seq_printf(m, " cyclic-check recursions: %11u\n", - debug_atomic_read(&nr_cyclic_check_recursions)); seq_printf(m, " find-mask forwards checks: %11u\n", debug_atomic_read(&nr_find_usage_forwards_checks)); - seq_printf(m, " find-mask forwards recursions: %11u\n", - debug_atomic_read(&nr_find_usage_forwards_recursions)); seq_printf(m, " find-mask backwards checks: %11u\n", debug_atomic_read(&nr_find_usage_backwards_checks)); - seq_printf(m, " find-mask backwards recursions:%11u\n", - debug_atomic_read(&nr_find_usage_backwards_recursions)); seq_printf(m, " hardirq on events: %11u\n", hi1); seq_printf(m, " hardirq off events: %11u\n", hi2); @@ -300,27 +235,27 @@ static int lockdep_stats_show(struct seq_file *m, void *v) nr_uncategorized++; if (class->usage_mask & LOCKF_USED_IN_IRQ) nr_irq_safe++; - if (class->usage_mask & LOCKF_ENABLED_IRQS) + if (class->usage_mask & LOCKF_ENABLED_IRQ) nr_irq_unsafe++; if (class->usage_mask & LOCKF_USED_IN_SOFTIRQ) nr_softirq_safe++; - if (class->usage_mask & LOCKF_ENABLED_SOFTIRQS) + if (class->usage_mask & LOCKF_ENABLED_SOFTIRQ) nr_softirq_unsafe++; if (class->usage_mask & LOCKF_USED_IN_HARDIRQ) nr_hardirq_safe++; - if (class->usage_mask & LOCKF_ENABLED_HARDIRQS) + if (class->usage_mask & LOCKF_ENABLED_HARDIRQ) nr_hardirq_unsafe++; if (class->usage_mask & LOCKF_USED_IN_IRQ_READ) nr_irq_read_safe++; - if (class->usage_mask & LOCKF_ENABLED_IRQS_READ) + if (class->usage_mask & LOCKF_ENABLED_IRQ_READ) nr_irq_read_unsafe++; if (class->usage_mask & LOCKF_USED_IN_SOFTIRQ_READ) nr_softirq_read_safe++; - if (class->usage_mask & LOCKF_ENABLED_SOFTIRQS_READ) + if (class->usage_mask & LOCKF_ENABLED_SOFTIRQ_READ) nr_softirq_read_unsafe++; if (class->usage_mask & LOCKF_USED_IN_HARDIRQ_READ) nr_hardirq_read_safe++; - if (class->usage_mask & LOCKF_ENABLED_HARDIRQS_READ) + if (class->usage_mask & LOCKF_ENABLED_HARDIRQ_READ) nr_hardirq_read_unsafe++; #ifdef CONFIG_PROVE_LOCKING @@ -409,8 +344,10 @@ static int lockdep_stats_show(struct seq_file *m, void *v) nr_unused); seq_printf(m, " max locking depth: %11u\n", max_lockdep_depth); - seq_printf(m, " max recursion depth: %11u\n", - max_recursion_depth); +#ifdef CONFIG_PROVE_LOCKING + seq_printf(m, " max bfs queue depth: %11u\n", + max_bfs_queue_depth); +#endif lockdep_stats_debug_show(m); seq_printf(m, " debug_locks: %11u\n", debug_locks); @@ -438,7 +375,6 @@ struct lock_stat_data { }; struct lock_stat_seq { - struct lock_stat_data *iter; struct lock_stat_data *iter_end; struct lock_stat_data stats[MAX_LOCKDEP_KEYS]; }; @@ -601,6 +537,10 @@ static void seq_stats(struct seq_file *m, struct lock_stat_data *data) static void seq_header(struct seq_file *m) { seq_printf(m, "lock_stat version 0.3\n"); + + if (unlikely(!debug_locks)) + seq_printf(m, "*WARNING* lock debugging disabled!! - possibly due to a lockdep warning\n"); + seq_line(m, '-', 0, 40 + 1 + 10 * (14 + 1)); seq_printf(m, "%40s %14s %14s %14s %14s %14s %14s %14s %14s " "%14s %14s\n", @@ -622,34 +562,22 @@ static void seq_header(struct seq_file *m) static void *ls_start(struct seq_file *m, loff_t *pos) { struct lock_stat_seq *data = m->private; + struct lock_stat_data *iter; if (*pos == 0) return SEQ_START_TOKEN; - data->iter = data->stats + *pos; - if (data->iter >= data->iter_end) - data->iter = NULL; + iter = data->stats + (*pos - 1); + if (iter >= data->iter_end) + iter = NULL; - return data->iter; + return iter; } static void *ls_next(struct seq_file *m, void *v, loff_t *pos) { - struct lock_stat_seq *data = m->private; - (*pos)++; - - if (v == SEQ_START_TOKEN) - data->iter = data->stats; - else { - data->iter = v; - data->iter++; - } - - if (data->iter == data->iter_end) - data->iter = NULL; - - return data->iter; + return ls_start(m, pos); } static void ls_stop(struct seq_file *m, void *v) @@ -687,7 +615,6 @@ static int lock_stat_open(struct inode *inode, struct file *file) struct lock_stat_data *iter = data->stats; struct seq_file *m = file->private_data; - data->iter = iter; list_for_each_entry(class, &all_lock_classes, lock_entry) { iter->class = class; iter->stats = lock_stats(class); @@ -695,7 +622,7 @@ static int lock_stat_open(struct inode *inode, struct file *file) } data->iter_end = iter; - sort(data->stats, data->iter_end - data->iter, + sort(data->stats, data->iter_end - data->stats, sizeof(struct lock_stat_data), lock_stat_cmp, NULL); @@ -730,7 +657,6 @@ static int lock_stat_release(struct inode *inode, struct file *file) struct seq_file *seq = file->private_data; vfree(seq->private); - seq->private = NULL; return seq_release(inode, file); } @@ -754,7 +680,8 @@ static int __init lockdep_proc_init(void) &proc_lockdep_stats_operations); #ifdef CONFIG_LOCK_STAT - proc_create("lock_stat", S_IRUSR, NULL, &proc_lock_stat_operations); + proc_create("lock_stat", S_IRUSR | S_IWUSR, NULL, + &proc_lock_stat_operations); #endif return 0; diff --git a/kernel/lockdep_states.h b/kernel/lockdep_states.h new file mode 100644 index 00000000000..995b0cc2b84 --- /dev/null +++ b/kernel/lockdep_states.h @@ -0,0 +1,9 @@ +/* + * Lockdep states, + * + * please update XXX_LOCK_USAGE_STATES in include/linux/lockdep.h whenever + * you add one, or come up with a nice dynamic solution. + */ +LOCKDEP_STATE(HARDIRQ) +LOCKDEP_STATE(SOFTIRQ) +LOCKDEP_STATE(RECLAIM_FS) diff --git a/kernel/module.c b/kernel/module.c index e8b51d41dd7..46580edff0c 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -18,6 +18,7 @@ */ #include <linux/module.h> #include <linux/moduleloader.h> +#include <linux/ftrace_event.h> #include <linux/init.h> #include <linux/kallsyms.h> #include <linux/fs.h> @@ -51,6 +52,13 @@ #include <linux/tracepoint.h> #include <linux/ftrace.h> #include <linux/async.h> +#include <linux/percpu.h> +#include <linux/kmemleak.h> + +#define CREATE_TRACE_POINTS +#include <trace/events/module.h> + +EXPORT_TRACEPOINT_SYMBOL(module_get); #if 0 #define DEBUGP printk @@ -67,15 +75,19 @@ /* List of modules, protected by module_mutex or preempt_disable * (delete uses stop_machine/add uses RCU list operations). */ -static DEFINE_MUTEX(module_mutex); +DEFINE_MUTEX(module_mutex); +EXPORT_SYMBOL_GPL(module_mutex); static LIST_HEAD(modules); +/* Block module loading/unloading? */ +int modules_disabled = 0; + /* Waiting for a module to finish initializing? */ static DECLARE_WAIT_QUEUE_HEAD(module_wq); static BLOCKING_NOTIFIER_HEAD(module_notify_list); -/* Bounds of module allocation, for speeding __module_text_address */ +/* Bounds of module allocation, for speeding __module_address */ static unsigned long module_addr_min = -1UL, module_addr_max = 0; int register_module_notifier(struct notifier_block * nb) @@ -185,17 +197,6 @@ extern const unsigned long __start___kcrctab_unused_gpl[]; #define symversion(base, idx) ((base != NULL) ? ((base) + (idx)) : NULL) #endif -struct symsearch { - const struct kernel_symbol *start, *stop; - const unsigned long *crcs; - enum { - NOT_GPL_ONLY, - GPL_ONLY, - WILL_BE_GPL_ONLY, - } licence; - bool unused; -}; - static bool each_symbol_in_section(const struct symsearch *arr, unsigned int arrsize, struct module *owner, @@ -216,10 +217,8 @@ static bool each_symbol_in_section(const struct symsearch *arr, } /* Returns true as soon as fn returns true, otherwise false. */ -static bool each_symbol(bool (*fn)(const struct symsearch *arr, - struct module *owner, - unsigned int symnum, void *data), - void *data) +bool each_symbol(bool (*fn)(const struct symsearch *arr, struct module *owner, + unsigned int symnum, void *data), void *data) { struct module *mod; const struct symsearch arr[] = { @@ -272,6 +271,7 @@ static bool each_symbol(bool (*fn)(const struct symsearch *arr, } return false; } +EXPORT_SYMBOL_GPL(each_symbol); struct find_symbol_arg { /* Input */ @@ -282,7 +282,7 @@ struct find_symbol_arg { /* Output */ struct module *owner; const unsigned long *crc; - unsigned long value; + const struct kernel_symbol *sym; }; static bool find_symbol_in_section(const struct symsearch *syms, @@ -323,17 +323,17 @@ static bool find_symbol_in_section(const struct symsearch *syms, fsa->owner = owner; fsa->crc = symversion(syms->crcs, symnum); - fsa->value = syms->start[symnum].value; + fsa->sym = &syms->start[symnum]; return true; } -/* Find a symbol, return value, (optional) crc and (optional) module - * which owns it */ -static unsigned long find_symbol(const char *name, - struct module **owner, - const unsigned long **crc, - bool gplok, - bool warn) +/* Find a symbol and return it, along with, (optional) crc and + * (optional) module which owns it */ +const struct kernel_symbol *find_symbol(const char *name, + struct module **owner, + const unsigned long **crc, + bool gplok, + bool warn) { struct find_symbol_arg fsa; @@ -346,15 +346,16 @@ static unsigned long find_symbol(const char *name, *owner = fsa.owner; if (crc) *crc = fsa.crc; - return fsa.value; + return fsa.sym; } DEBUGP("Failed to find symbol %s\n", name); - return -ENOENT; + return NULL; } +EXPORT_SYMBOL_GPL(find_symbol); /* Search for module by name: must hold module_mutex. */ -static struct module *find_module(const char *name) +struct module *find_module(const char *name) { struct module *mod; @@ -364,8 +365,37 @@ static struct module *find_module(const char *name) } return NULL; } +EXPORT_SYMBOL_GPL(find_module); #ifdef CONFIG_SMP + +#ifdef CONFIG_HAVE_DYNAMIC_PER_CPU_AREA + +static void *percpu_modalloc(unsigned long size, unsigned long align, + const char *name) +{ + void *ptr; + + if (align > PAGE_SIZE) { + printk(KERN_WARNING "%s: per-cpu alignment %li > %li\n", + name, align, PAGE_SIZE); + align = PAGE_SIZE; + } + + ptr = __alloc_reserved_percpu(size, align); + if (!ptr) + printk(KERN_WARNING + "Could not allocate %lu bytes percpu data\n", size); + return ptr; +} + +static void percpu_modfree(void *freeme) +{ + free_percpu(freeme); +} + +#else /* ... !CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */ + /* Number of blocks used and allocated. */ static unsigned int pcpu_num_used, pcpu_num_allocated; /* Size of each block. -ve means used. */ @@ -409,6 +439,7 @@ static void *percpu_modalloc(unsigned long size, unsigned long align, unsigned long extra; unsigned int i; void *ptr; + int cpu; if (align > PAGE_SIZE) { printk(KERN_WARNING "%s: per-cpu alignment %li > %li\n", @@ -438,6 +469,11 @@ static void *percpu_modalloc(unsigned long size, unsigned long align, if (!split_block(i, size)) return NULL; + /* add the per-cpu scanning areas */ + for_each_possible_cpu(cpu) + kmemleak_alloc(ptr + per_cpu_offset(cpu), size, 0, + GFP_KERNEL); + /* Mark allocated */ pcpu_size[i] = -pcpu_size[i]; return ptr; @@ -452,6 +488,7 @@ static void percpu_modfree(void *freeme) { unsigned int i; void *ptr = __per_cpu_start + block_size(pcpu_size[0]); + int cpu; /* First entry is core kernel percpu data. */ for (i = 1; i < pcpu_num_used; ptr += block_size(pcpu_size[i]), i++) { @@ -463,6 +500,10 @@ static void percpu_modfree(void *freeme) BUG(); free: + /* remove the per-cpu scanning areas */ + for_each_possible_cpu(cpu) + kmemleak_free(freeme + per_cpu_offset(cpu)); + /* Merge with previous? */ if (pcpu_size[i-1] >= 0) { pcpu_size[i-1] += pcpu_size[i]; @@ -480,21 +521,6 @@ static void percpu_modfree(void *freeme) } } -static unsigned int find_pcpusec(Elf_Ehdr *hdr, - Elf_Shdr *sechdrs, - const char *secstrings) -{ - return find_sec(hdr, sechdrs, secstrings, ".data.percpu"); -} - -static void percpu_modcopy(void *pcpudest, const void *from, unsigned long size) -{ - int cpu; - - for_each_possible_cpu(cpu) - memcpy(pcpudest + per_cpu_offset(cpu), from, size); -} - static int percpu_modinit(void) { pcpu_num_used = 2; @@ -513,7 +539,26 @@ static int percpu_modinit(void) return 0; } __initcall(percpu_modinit); + +#endif /* CONFIG_HAVE_DYNAMIC_PER_CPU_AREA */ + +static unsigned int find_pcpusec(Elf_Ehdr *hdr, + Elf_Shdr *sechdrs, + const char *secstrings) +{ + return find_sec(hdr, sechdrs, secstrings, ".data.percpu"); +} + +static void percpu_modcopy(void *pcpudest, const void *from, unsigned long size) +{ + int cpu; + + for_each_possible_cpu(cpu) + memcpy(pcpudest + per_cpu_offset(cpu), from, size); +} + #else /* ... !CONFIG_SMP */ + static inline void *percpu_modalloc(unsigned long size, unsigned long align, const char *name) { @@ -535,6 +580,7 @@ static inline void percpu_modcopy(void *pcpudst, const void *src, /* pcpusec should be 0, and size of that section should be 0. */ BUG_ON(size != 0); } + #endif /* CONFIG_SMP */ #define MODINFO_ATTR(field) \ @@ -573,13 +619,13 @@ static char last_unloaded_module[MODULE_NAME_LEN+1]; /* Init the unload section of the module. */ static void module_unload_init(struct module *mod) { - unsigned int i; + int cpu; INIT_LIST_HEAD(&mod->modules_which_use_me); - for (i = 0; i < NR_CPUS; i++) - local_set(&mod->ref[i].count, 0); + for_each_possible_cpu(cpu) + local_set(__module_ref_addr(mod, cpu), 0); /* Hold reference count during initialization. */ - local_set(&mod->ref[raw_smp_processor_id()].count, 1); + local_set(__module_ref_addr(mod, raw_smp_processor_id()), 1); /* Backwards compatibility macros put refcount during init. */ mod->waiter = current; } @@ -607,7 +653,7 @@ static int already_uses(struct module *a, struct module *b) } /* Module a uses b */ -static int use_module(struct module *a, struct module *b) +int use_module(struct module *a, struct module *b) { struct module_use *use; int no_warn, err; @@ -640,6 +686,7 @@ static int use_module(struct module *a, struct module *b) no_warn = sysfs_create_link(b->holders_dir, &a->mkobj.kobj, a->name); return 1; } +EXPORT_SYMBOL_GPL(use_module); /* Clear the unload stuff of the module. */ static void module_unload_free(struct module *mod) @@ -717,10 +764,11 @@ static int try_stop_module(struct module *mod, int flags, int *forced) unsigned int module_refcount(struct module *mod) { - unsigned int i, total = 0; + unsigned int total = 0; + int cpu; - for (i = 0; i < NR_CPUS; i++) - total += local_read(&mod->ref[i].count); + for_each_possible_cpu(cpu) + total += local_read(__module_ref_addr(mod, cpu)); return total; } EXPORT_SYMBOL(module_refcount); @@ -750,7 +798,7 @@ SYSCALL_DEFINE2(delete_module, const char __user *, name_user, char name[MODULE_NAME_LEN]; int ret, forced = 0; - if (!capable(CAP_SYS_MODULE)) + if (!capable(CAP_SYS_MODULE) || modules_disabled) return -EPERM; if (strncpy_from_user(name, name_user, MODULE_NAME_LEN-1) < 0) @@ -821,7 +869,7 @@ SYSCALL_DEFINE2(delete_module, const char __user *, name_user, mutex_lock(&module_mutex); /* Store the name of the last unloaded module for diagnostic purposes */ strlcpy(last_unloaded_module, mod->name, sizeof(last_unloaded_module)); - unregister_dynamic_debug_module(mod->name); + ddebug_remove_module(mod->name); free_module(mod); out: @@ -859,22 +907,26 @@ void __symbol_put(const char *symbol) struct module *owner; preempt_disable(); - if (IS_ERR_VALUE(find_symbol(symbol, &owner, NULL, true, false))) + if (!find_symbol(symbol, &owner, NULL, true, false)) BUG(); module_put(owner); preempt_enable(); } EXPORT_SYMBOL(__symbol_put); +/* Note this assumes addr is a function, which it currently always is. */ void symbol_put_addr(void *addr) { struct module *modaddr; + unsigned long a = (unsigned long)dereference_function_descriptor(addr); - if (core_kernel_text((unsigned long)addr)) + if (core_kernel_text(a)) return; - if (!(modaddr = module_text_address((unsigned long)addr))) - BUG(); + /* module_text_address is safe here: we're supposed to have reference + * to module from symbol_get, so it can't go away. */ + modaddr = __module_text_address(a); + BUG_ON(!modaddr); module_put(modaddr); } EXPORT_SYMBOL_GPL(symbol_put_addr); @@ -894,7 +946,9 @@ void module_put(struct module *module) { if (module) { unsigned int cpu = get_cpu(); - local_dec(&module->ref[cpu].count); + local_dec(__module_ref_addr(module, cpu)); + trace_module_put(module, _RET_IP_, + local_read(__module_ref_addr(module, cpu))); /* Maybe they're waiting for us to drop reference? */ if (unlikely(!module_is_live(module))) wake_up_process(module->waiter); @@ -914,10 +968,11 @@ static inline void module_unload_free(struct module *mod) { } -static inline int use_module(struct module *a, struct module *b) +int use_module(struct module *a, struct module *b) { return strong_try_module_get(b) == 0; } +EXPORT_SYMBOL_GPL(use_module); static inline void module_unload_init(struct module *mod) { @@ -960,12 +1015,12 @@ static struct module_attribute *modinfo_attrs[] = { static const char vermagic[] = VERMAGIC_STRING; -static int try_to_force_load(struct module *mod, const char *symname) +static int try_to_force_load(struct module *mod, const char *reason) { #ifdef CONFIG_MODULE_FORCE_LOAD if (!test_taint(TAINT_FORCED_MODULE)) - printk("%s: no version for \"%s\" found: kernel tainted.\n", - mod->name, symname); + printk(KERN_WARNING "%s: %s: kernel tainted.\n", + mod->name, reason); add_taint_module(mod, TAINT_FORCED_MODULE); return 0; #else @@ -1022,9 +1077,10 @@ static inline int check_modstruct_version(Elf_Shdr *sechdrs, { const unsigned long *crc; - if (IS_ERR_VALUE(find_symbol("struct_module", NULL, &crc, true, false))) + if (!find_symbol(MODULE_SYMBOL_PREFIX "module_layout", NULL, + &crc, true, false)) BUG(); - return check_version(sechdrs, versindex, "struct_module", mod, crc); + return check_version(sechdrs, versindex, "module_layout", mod, crc); } /* First part is kernel version, which we ignore if module has crcs. */ @@ -1063,25 +1119,25 @@ static inline int same_magic(const char *amagic, const char *bmagic, /* Resolve a symbol for this module. I.e. if we find one, record usage. Must be holding module_mutex. */ -static unsigned long resolve_symbol(Elf_Shdr *sechdrs, - unsigned int versindex, - const char *name, - struct module *mod) +static const struct kernel_symbol *resolve_symbol(Elf_Shdr *sechdrs, + unsigned int versindex, + const char *name, + struct module *mod) { struct module *owner; - unsigned long ret; + const struct kernel_symbol *sym; const unsigned long *crc; - ret = find_symbol(name, &owner, &crc, + sym = find_symbol(name, &owner, &crc, !(mod->taints & (1 << TAINT_PROPRIETARY_MODULE)), true); - if (!IS_ERR_VALUE(ret)) { - /* use_module can fail due to OOM, - or module initialization or unloading */ + /* use_module can fail due to OOM, + or module initialization or unloading */ + if (sym) { if (!check_version(sechdrs, versindex, name, mod, crc) || !use_module(mod, owner)) - ret = -EINVAL; + sym = NULL; } - return ret; + return sym; } /* @@ -1225,6 +1281,10 @@ static void add_notes_attrs(struct module *mod, unsigned int nsect, struct module_notes_attrs *notes_attrs; struct bin_attribute *nattr; + /* failed to create section attributes, so can't create notes */ + if (!mod->sect_attrs) + return; + /* Count notes sections and allocate structures. */ notes = 0; for (i = 0; i < nsect; i++) @@ -1444,6 +1504,8 @@ static int __unlink_module(void *_mod) /* Free a module, remove from lists, etc (must hold module_mutex). */ static void free_module(struct module *mod) { + trace_module_free(mod); + /* Delete from various lists */ stop_machine(__unlink_module, mod, NULL); remove_notes_attrs(mod); @@ -1456,15 +1518,18 @@ static void free_module(struct module *mod) /* Module unload stuff */ module_unload_free(mod); - /* release any pointers to mcount in this module */ - ftrace_release(mod->module_core, mod->core_size); + /* Free any allocated parameters. */ + destroy_params(mod->kp, mod->num_kp); /* This may be NULL, but that's OK */ module_free(mod, mod->module_init); kfree(mod->args); if (mod->percpu) percpu_modfree(mod->percpu); - +#if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP) + if (mod->refptr) + percpu_modfree(mod->refptr); +#endif /* Free lock-classes: */ lockdep_free_key_range(mod->module_core, mod->core_size); @@ -1475,17 +1540,15 @@ static void free_module(struct module *mod) void *__symbol_get(const char *symbol) { struct module *owner; - unsigned long value; + const struct kernel_symbol *sym; preempt_disable(); - value = find_symbol(symbol, &owner, NULL, true, true); - if (IS_ERR_VALUE(value)) - value = 0; - else if (strong_try_module_get(owner)) - value = 0; + sym = find_symbol(symbol, &owner, NULL, true, true); + if (sym && strong_try_module_get(owner)) + sym = NULL; preempt_enable(); - return (void *)value; + return sym ? (void *)sym->value : NULL; } EXPORT_SYMBOL_GPL(__symbol_get); @@ -1513,8 +1576,7 @@ static int verify_export_symbols(struct module *mod) for (i = 0; i < ARRAY_SIZE(arr); i++) { for (s = arr[i].sym; s < arr[i].sym + arr[i].num; s++) { - if (!IS_ERR_VALUE(find_symbol(s->name, &owner, - NULL, true, false))) { + if (find_symbol(s->name, &owner, NULL, true, false)) { printk(KERN_ERR "%s: exports duplicate symbol %s" " (owned by %s)\n", @@ -1538,6 +1600,7 @@ static int simplify_symbols(Elf_Shdr *sechdrs, unsigned long secbase; unsigned int i, n = sechdrs[symindex].sh_size / sizeof(Elf_Sym); int ret = 0; + const struct kernel_symbol *ksym; for (i = 1; i < n; i++) { switch (sym[i].st_shndx) { @@ -1557,13 +1620,14 @@ static int simplify_symbols(Elf_Shdr *sechdrs, break; case SHN_UNDEF: - sym[i].st_value - = resolve_symbol(sechdrs, versindex, - strtab + sym[i].st_name, mod); - + ksym = resolve_symbol(sechdrs, versindex, + strtab + sym[i].st_name, mod); /* Ok if resolved. */ - if (!IS_ERR_VALUE(sym[i].st_value)) + if (ksym) { + sym[i].st_value = ksym->value; break; + } + /* Ok if weak. */ if (ELF_ST_BIND(sym[i].st_info) == STB_WEAK) break; @@ -1638,8 +1702,7 @@ static void layout_sections(struct module *mod, if ((s->sh_flags & masks[m][0]) != masks[m][0] || (s->sh_flags & masks[m][1]) || s->sh_entsize != ~0UL - || strncmp(secstrings + s->sh_name, - ".init", 5) == 0) + || strstarts(secstrings + s->sh_name, ".init")) continue; s->sh_entsize = get_offset(mod, &mod->core_size, s, i); DEBUGP("\t%s\n", secstrings + s->sh_name); @@ -1656,8 +1719,7 @@ static void layout_sections(struct module *mod, if ((s->sh_flags & masks[m][0]) != masks[m][0] || (s->sh_flags & masks[m][1]) || s->sh_entsize != ~0UL - || strncmp(secstrings + s->sh_name, - ".init", 5) != 0) + || !strstarts(secstrings + s->sh_name, ".init")) continue; s->sh_entsize = (get_offset(mod, &mod->init_size, s, i) | INIT_OFFSET_MASK); @@ -1790,8 +1852,7 @@ static char elf_type(const Elf_Sym *sym, else return 'b'; } - if (strncmp(secstrings + sechdrs[sym->st_shndx].sh_name, - ".debug", strlen(".debug")) == 0) + if (strstarts(secstrings + sechdrs[sym->st_shndx].sh_name, ".debug")) return 'n'; return '?'; } @@ -1823,19 +1884,13 @@ static inline void add_kallsyms(struct module *mod, } #endif /* CONFIG_KALLSYMS */ -static void dynamic_printk_setup(struct mod_debug *debug, unsigned int num) +static void dynamic_debug_setup(struct _ddebug *debug, unsigned int num) { -#ifdef CONFIG_DYNAMIC_PRINTK_DEBUG - unsigned int i; - - for (i = 0; i < num; i++) { - register_dynamic_debug_module(debug[i].modname, - debug[i].type, - debug[i].logical_modname, - debug[i].flag_names, - debug[i].hash, debug[i].hash2); - } -#endif /* CONFIG_DYNAMIC_PRINTK_DEBUG */ +#ifdef CONFIG_DYNAMIC_DEBUG + if (ddebug_add_module(debug, num, debug->modname)) + printk(KERN_ERR "dynamic debug error adding module: %s\n", + debug->modname); +#endif } static void *module_alloc_update_bounds(unsigned long size) @@ -1852,6 +1907,36 @@ static void *module_alloc_update_bounds(unsigned long size) return ret; } +#ifdef CONFIG_DEBUG_KMEMLEAK +static void kmemleak_load_module(struct module *mod, Elf_Ehdr *hdr, + Elf_Shdr *sechdrs, char *secstrings) +{ + unsigned int i; + + /* only scan the sections containing data */ + kmemleak_scan_area(mod->module_core, (unsigned long)mod - + (unsigned long)mod->module_core, + sizeof(struct module), GFP_KERNEL); + + for (i = 1; i < hdr->e_shnum; i++) { + if (!(sechdrs[i].sh_flags & SHF_ALLOC)) + continue; + if (strncmp(secstrings + sechdrs[i].sh_name, ".data", 5) != 0 + && strncmp(secstrings + sechdrs[i].sh_name, ".bss", 4) != 0) + continue; + + kmemleak_scan_area(mod->module_core, sechdrs[i].sh_addr - + (unsigned long)mod->module_core, + sechdrs[i].sh_size, GFP_KERNEL); + } +} +#else +static inline void kmemleak_load_module(struct module *mod, Elf_Ehdr *hdr, + Elf_Shdr *sechdrs, char *secstrings) +{ +} +#endif + /* Allocate and load the module: note that size of section 0 is always zero, and we rely on this for optional sections. */ static noinline struct module *load_module(void __user *umod, @@ -1866,12 +1951,9 @@ static noinline struct module *load_module(void __user *umod, unsigned int symindex = 0; unsigned int strindex = 0; unsigned int modindex, versindex, infoindex, pcpuindex; - unsigned int num_kp, num_mcount; - struct kernel_param *kp; struct module *mod; long err = 0; void *percpu = NULL, *ptr = NULL; /* Stops spurious gcc warning */ - unsigned long *mseg; mm_segment_t old_fs; DEBUGP("load_module: umod=%p, len=%lu, uargs=%p\n", @@ -1884,12 +1966,6 @@ static noinline struct module *load_module(void __user *umod, if (len > 64 * 1024 * 1024 || (hdr = vmalloc(len)) == NULL) return ERR_PTR(-ENOMEM); - /* Create stop_machine threads since the error path relies on - * a non-failing stop_machine call. */ - err = stop_machine_create(); - if (err) - goto free_hdr; - if (copy_from_user(hdr, umod, len) != 0) { err = -EFAULT; goto free_hdr; @@ -1930,7 +2006,7 @@ static noinline struct module *load_module(void __user *umod, } #ifndef CONFIG_MODULE_UNLOAD /* Don't load .exit sections */ - if (strncmp(secstrings+sechdrs[i].sh_name, ".exit", 5) == 0) + if (strstarts(secstrings+sechdrs[i].sh_name, ".exit")) sechdrs[i].sh_flags &= ~(unsigned long)SHF_ALLOC; #endif } @@ -1974,7 +2050,7 @@ static noinline struct module *load_module(void __user *umod, modmagic = get_modinfo(sechdrs, infoindex, "vermagic"); /* This is allowed: modprobe --force will invalidate it. */ if (!modmagic) { - err = try_to_force_load(mod, "magic"); + err = try_to_force_load(mod, "bad vermagic"); if (err) goto free_hdr; } else if (!same_magic(modmagic, vermagic, versindex)) { @@ -2031,6 +2107,12 @@ static noinline struct module *load_module(void __user *umod, /* Do the allocs. */ ptr = module_alloc_update_bounds(mod->core_size); + /* + * The pointer to this block is stored in the module structure + * which is inside the block. Just mark it as not being a + * leak. + */ + kmemleak_not_leak(ptr); if (!ptr) { err = -ENOMEM; goto free_percpu; @@ -2039,6 +2121,13 @@ static noinline struct module *load_module(void __user *umod, mod->module_core = ptr; ptr = module_alloc_update_bounds(mod->init_size); + /* + * The pointer to this block is stored in the module structure + * which is inside the block. This block doesn't need to be + * scanned as it contains data and code that will be freed + * after the module is initialized. + */ + kmemleak_ignore(ptr); if (!ptr && mod->init_size) { err = -ENOMEM; goto free_core; @@ -2069,7 +2158,16 @@ static noinline struct module *load_module(void __user *umod, } /* Module has been moved. */ mod = (void *)sechdrs[modindex].sh_addr; + kmemleak_load_module(mod, hdr, sechdrs, secstrings); +#if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP) + mod->refptr = percpu_modalloc(sizeof(local_t), __alignof__(local_t), + mod->name); + if (!mod->refptr) { + err = -ENOMEM; + goto free_init; + } +#endif /* Now we've moved module, initialize linked lists, etc. */ module_unload_init(mod); @@ -2104,8 +2202,8 @@ static noinline struct module *load_module(void __user *umod, /* Now we've got everything in the final locations, we can * find optional sections. */ - kp = section_objs(hdr, sechdrs, secstrings, "__param", sizeof(*kp), - &num_kp); + mod->kp = section_objs(hdr, sechdrs, secstrings, "__param", + sizeof(*mod->kp), &mod->num_kp); mod->syms = section_objs(hdr, sechdrs, secstrings, "__ksymtab", sizeof(*mod->syms), &mod->num_syms); mod->crcs = section_addr(hdr, sechdrs, secstrings, "__kcrctab"); @@ -2134,6 +2232,10 @@ static noinline struct module *load_module(void __user *umod, mod->unused_gpl_crcs = section_addr(hdr, sechdrs, secstrings, "__kcrctab_unused_gpl"); #endif +#ifdef CONFIG_CONSTRUCTORS + mod->ctors = section_objs(hdr, sechdrs, secstrings, ".ctors", + sizeof(*mod->ctors), &mod->num_ctors); +#endif #ifdef CONFIG_MARKERS mod->markers = section_objs(hdr, sechdrs, secstrings, "__markers", @@ -2145,7 +2247,19 @@ static noinline struct module *load_module(void __user *umod, sizeof(*mod->tracepoints), &mod->num_tracepoints); #endif - +#ifdef CONFIG_EVENT_TRACING + mod->trace_events = section_objs(hdr, sechdrs, secstrings, + "_ftrace_events", + sizeof(*mod->trace_events), + &mod->num_trace_events); +#endif +#ifdef CONFIG_FTRACE_MCOUNT_RECORD + /* sechdrs[0].sh_size is always zero */ + mod->ftrace_callsites = section_objs(hdr, sechdrs, secstrings, + "__mcount_loc", + sizeof(*mod->ftrace_callsites), + &mod->num_ftrace_callsites); +#endif #ifdef CONFIG_MODVERSIONS if ((mod->num_syms && !mod->crcs) || (mod->num_gpl_syms && !mod->gpl_crcs) @@ -2155,8 +2269,8 @@ static noinline struct module *load_module(void __user *umod, || (mod->num_unused_gpl_syms && !mod->unused_gpl_crcs) #endif ) { - printk(KERN_WARNING "%s: No versions for exported symbols.\n", mod->name); - err = try_to_force_load(mod, "nocrc"); + err = try_to_force_load(mod, + "no versions for exported symbols"); if (err) goto cleanup; } @@ -2201,19 +2315,15 @@ static noinline struct module *load_module(void __user *umod, add_kallsyms(mod, sechdrs, symindex, strindex, secstrings); if (!mod->taints) { - struct mod_debug *debug; + struct _ddebug *debug; unsigned int num_debug; debug = section_objs(hdr, sechdrs, secstrings, "__verbose", sizeof(*debug), &num_debug); - dynamic_printk_setup(debug, num_debug); + if (debug) + dynamic_debug_setup(debug, num_debug); } - /* sechdrs[0].sh_size is always zero */ - mseg = section_objs(hdr, sechdrs, secstrings, "__mcount_loc", - sizeof(*mseg), &num_mcount); - ftrace_init_module(mod, mseg, mseg + num_mcount); - err = module_finalize(hdr, sechdrs, mod); if (err < 0) goto cleanup; @@ -2250,11 +2360,11 @@ static noinline struct module *load_module(void __user *umod, */ list_add_rcu(&mod->list, &modules); - err = parse_args(mod->name, mod->args, kp, num_kp, NULL); + err = parse_args(mod->name, mod->args, mod->kp, mod->num_kp, NULL); if (err < 0) goto unlink; - err = mod_sysfs_setup(mod, kp, num_kp); + err = mod_sysfs_setup(mod, mod->kp, mod->num_kp); if (err < 0) goto unlink; add_sect_attrs(mod, hdr->e_shnum, secstrings, sechdrs); @@ -2263,22 +2373,29 @@ static noinline struct module *load_module(void __user *umod, /* Get rid of temporary copy */ vfree(hdr); - stop_machine_destroy(); + trace_module_load(mod); + /* Done! */ return mod; unlink: - stop_machine(__unlink_module, mod, NULL); + /* Unlink carefully: kallsyms could be walking list. */ + list_del_rcu(&mod->list); + synchronize_sched(); module_arch_cleanup(mod); cleanup: kobject_del(&mod->mkobj.kobj); kobject_put(&mod->mkobj.kobj); - ftrace_release(mod->module_core, mod->core_size); free_unload: module_unload_free(mod); +#if defined(CONFIG_MODULE_UNLOAD) && defined(CONFIG_SMP) + free_init: + percpu_modfree(mod->refptr); +#endif module_free(mod, mod->module_init); free_core: module_free(mod, mod->module_core); + /* mod will be freed with core. Don't access it beyond this line! */ free_percpu: if (percpu) percpu_modfree(percpu); @@ -2286,7 +2403,6 @@ static noinline struct module *load_module(void __user *umod, kfree(args); free_hdr: vfree(hdr); - stop_machine_destroy(); return ERR_PTR(err); truncated: @@ -2295,6 +2411,17 @@ static noinline struct module *load_module(void __user *umod, goto free_hdr; } +/* Call module constructors. */ +static void do_mod_ctors(struct module *mod) +{ +#ifdef CONFIG_CONSTRUCTORS + unsigned long i; + + for (i = 0; i < mod->num_ctors; i++) + mod->ctors[i](); +#endif +} + /* This is where the real work happens */ SYSCALL_DEFINE3(init_module, void __user *, umod, unsigned long, len, const char __user *, uargs) @@ -2303,7 +2430,7 @@ SYSCALL_DEFINE3(init_module, void __user *, umod, int ret = 0; /* Must have permission */ - if (!capable(CAP_SYS_MODULE)) + if (!capable(CAP_SYS_MODULE) || modules_disabled) return -EPERM; /* Only one module load at a time, please */ @@ -2323,6 +2450,7 @@ SYSCALL_DEFINE3(init_module, void __user *, umod, blocking_notifier_call_chain(&module_notify_list, MODULE_STATE_COMING, mod); + do_mod_ctors(mod); /* Start the module */ if (mod->init != NULL) ret = do_one_initcall(mod->init); @@ -2341,9 +2469,9 @@ SYSCALL_DEFINE3(init_module, void __user *, umod, return ret; } if (ret > 0) { - printk(KERN_WARNING "%s: '%s'->init suspiciously returned %d, " - "it should follow 0/-E convention\n" - KERN_WARNING "%s: loading module anyway...\n", + printk(KERN_WARNING +"%s: '%s'->init suspiciously returned %d, it should follow 0/-E convention\n" +"%s: loading module anyway...\n", __func__, mod->name, ret, __func__); dump_stack(); @@ -2355,9 +2483,13 @@ SYSCALL_DEFINE3(init_module, void __user *, umod, blocking_notifier_call_chain(&module_notify_list, MODULE_STATE_LIVE, mod); + /* We need to finish all async code before the module init sequence is done */ + async_synchronize_full(); + mutex_lock(&module_mutex); /* Drop initial reference. */ module_put(mod); + trim_init_extable(mod); module_free(mod, mod->module_init); mod->module_init = NULL; mod->init_size = 0; @@ -2563,6 +2695,25 @@ unsigned long module_kallsyms_lookup_name(const char *name) preempt_enable(); return ret; } + +int module_kallsyms_on_each_symbol(int (*fn)(void *, const char *, + struct module *, unsigned long), + void *data) +{ + struct module *mod; + unsigned int i; + int ret; + + list_for_each_entry(mod, &modules, list) { + for (i = 0; i < mod->num_symtab; i++) { + ret = fn(data, mod->strtab + mod->symtab[i].st_name, + mod, mod->symtab[i].st_value); + if (ret != 0) + return ret; + } + } + return 0; +} #endif /* CONFIG_KALLSYMS */ static char *module_flags(struct module *mod, char *buf) @@ -2698,29 +2849,31 @@ const struct exception_table_entry *search_module_extables(unsigned long addr) } /* - * Is this a valid module address? + * is_module_address - is this address inside a module? + * @addr: the address to check. + * + * See is_module_text_address() if you simply want to see if the address + * is code (not data). */ -int is_module_address(unsigned long addr) +bool is_module_address(unsigned long addr) { - struct module *mod; + bool ret; preempt_disable(); - - list_for_each_entry_rcu(mod, &modules, list) { - if (within_module_core(addr, mod)) { - preempt_enable(); - return 1; - } - } - + ret = __module_address(addr) != NULL; preempt_enable(); - return 0; + return ret; } - -/* Is this a valid kernel address? */ -__notrace_funcgraph struct module *__module_text_address(unsigned long addr) +/* + * __module_address - get the module which contains an address. + * @addr: the address. + * + * Must be called with preempt disabled or module mutex held so that + * module doesn't get freed during this. + */ +struct module *__module_address(unsigned long addr) { struct module *mod; @@ -2728,22 +2881,51 @@ __notrace_funcgraph struct module *__module_text_address(unsigned long addr) return NULL; list_for_each_entry_rcu(mod, &modules, list) - if (within(addr, mod->module_init, mod->init_text_size) - || within(addr, mod->module_core, mod->core_text_size)) + if (within_module_core(addr, mod) + || within_module_init(addr, mod)) return mod; return NULL; } +EXPORT_SYMBOL_GPL(__module_address); -struct module *module_text_address(unsigned long addr) +/* + * is_module_text_address - is this address inside module code? + * @addr: the address to check. + * + * See is_module_address() if you simply want to see if the address is + * anywhere in a module. See kernel_text_address() for testing if an + * address corresponds to kernel or module code. + */ +bool is_module_text_address(unsigned long addr) { - struct module *mod; + bool ret; preempt_disable(); - mod = __module_text_address(addr); + ret = __module_text_address(addr) != NULL; preempt_enable(); + return ret; +} + +/* + * __module_text_address - get the module whose code contains an address. + * @addr: the address. + * + * Must be called with preempt disabled or module mutex held so that + * module doesn't get freed during this. + */ +struct module *__module_text_address(unsigned long addr) +{ + struct module *mod = __module_address(addr); + if (mod) { + /* Make sure it's within the text section. */ + if (!within(addr, mod->module_init, mod->init_text_size) + && !within(addr, mod->module_core, mod->core_text_size)) + mod = NULL; + } return mod; } +EXPORT_SYMBOL_GPL(__module_text_address); /* Don't grab lock, we're oopsing. */ void print_modules(void) @@ -2751,7 +2933,7 @@ void print_modules(void) struct module *mod; char buf[8]; - printk("Modules linked in:"); + printk(KERN_DEFAULT "Modules linked in:"); /* Most callers should already have preempt disabled, but make sure */ preempt_disable(); list_for_each_entry_rcu(mod, &modules, list) @@ -2763,9 +2945,17 @@ void print_modules(void) } #ifdef CONFIG_MODVERSIONS -/* Generate the signature for struct module here, too, for modversions. */ -void struct_module(struct module *mod) { return; } -EXPORT_SYMBOL(struct_module); +/* Generate the signature for all relevant module structures here. + * If these change, we don't want to try to parse the module. */ +void module_layout(struct module *mod, + struct modversion_info *ver, + struct kernel_param *kp, + struct kernel_symbol *ks, + struct marker *marker, + struct tracepoint *tp) +{ +} +EXPORT_SYMBOL(module_layout); #endif #ifdef CONFIG_MARKERS diff --git a/kernel/mutex-debug.c b/kernel/mutex-debug.c index 1d94160eb53..50d022e5a56 100644 --- a/kernel/mutex-debug.c +++ b/kernel/mutex-debug.c @@ -26,11 +26,6 @@ /* * Must be called with lock->wait_lock held. */ -void debug_mutex_set_owner(struct mutex *lock, struct thread_info *new_owner) -{ - lock->owner = new_owner; -} - void debug_mutex_lock_common(struct mutex *lock, struct mutex_waiter *waiter) { memset(waiter, MUTEX_DEBUG_INIT, sizeof(*waiter)); @@ -59,7 +54,6 @@ void debug_mutex_add_waiter(struct mutex *lock, struct mutex_waiter *waiter, /* Mark the current thread as blocked on the lock: */ ti->task->blocked_on = waiter; - waiter->lock = lock; } void mutex_remove_waiter(struct mutex *lock, struct mutex_waiter *waiter, @@ -82,7 +76,7 @@ void debug_mutex_unlock(struct mutex *lock) DEBUG_LOCKS_WARN_ON(lock->magic != lock); DEBUG_LOCKS_WARN_ON(lock->owner != current_thread_info()); DEBUG_LOCKS_WARN_ON(!lock->wait_list.prev && !lock->wait_list.next); - DEBUG_LOCKS_WARN_ON(lock->owner != current_thread_info()); + mutex_clear_owner(lock); } void debug_mutex_init(struct mutex *lock, const char *name, @@ -95,7 +89,6 @@ void debug_mutex_init(struct mutex *lock, const char *name, debug_check_no_locks_freed((void *)lock, sizeof(*lock)); lockdep_init_map(&lock->dep_map, name, key, 0); #endif - lock->owner = NULL; lock->magic = lock; } diff --git a/kernel/mutex-debug.h b/kernel/mutex-debug.h index babfbdfc534..6b2d735846a 100644 --- a/kernel/mutex-debug.h +++ b/kernel/mutex-debug.h @@ -13,14 +13,6 @@ /* * This must be called with lock->wait_lock held. */ -extern void -debug_mutex_set_owner(struct mutex *lock, struct thread_info *new_owner); - -static inline void debug_mutex_clear_owner(struct mutex *lock) -{ - lock->owner = NULL; -} - extern void debug_mutex_lock_common(struct mutex *lock, struct mutex_waiter *waiter); extern void debug_mutex_wake_waiter(struct mutex *lock, @@ -35,6 +27,16 @@ extern void debug_mutex_unlock(struct mutex *lock); extern void debug_mutex_init(struct mutex *lock, const char *name, struct lock_class_key *key); +static inline void mutex_set_owner(struct mutex *lock) +{ + lock->owner = current_thread_info(); +} + +static inline void mutex_clear_owner(struct mutex *lock) +{ + lock->owner = NULL; +} + #define spin_lock_mutex(lock, flags) \ do { \ struct mutex *l = container_of(lock, struct mutex, wait_lock); \ diff --git a/kernel/mutex.c b/kernel/mutex.c index 4f45d4b658e..947b3ad551f 100644 --- a/kernel/mutex.c +++ b/kernel/mutex.c @@ -10,6 +10,11 @@ * Many thanks to Arjan van de Ven, Thomas Gleixner, Steven Rostedt and * David Howells for suggestions and improvements. * + * - Adaptive spinning for mutexes by Peter Zijlstra. (Ported to mainline + * from the -rt tree, where it was originally implemented for rtmutexes + * by Steven Rostedt, based on work by Gregory Haskins, Peter Morreale + * and Sven Dietrich. + * * Also see Documentation/mutex-design.txt. */ #include <linux/mutex.h> @@ -46,6 +51,7 @@ __mutex_init(struct mutex *lock, const char *name, struct lock_class_key *key) atomic_set(&lock->count, 1); spin_lock_init(&lock->wait_lock); INIT_LIST_HEAD(&lock->wait_list); + mutex_clear_owner(lock); debug_mutex_init(lock, name, key); } @@ -83,7 +89,7 @@ __mutex_lock_slowpath(atomic_t *lock_count); * * This function is similar to (but not equivalent to) down(). */ -void inline __sched mutex_lock(struct mutex *lock) +void __sched mutex_lock(struct mutex *lock) { might_sleep(); /* @@ -91,6 +97,7 @@ void inline __sched mutex_lock(struct mutex *lock) * 'unlocked' into 'locked' state. */ __mutex_fastpath_lock(&lock->count, __mutex_lock_slowpath); + mutex_set_owner(lock); } EXPORT_SYMBOL(mutex_lock); @@ -115,6 +122,14 @@ void __sched mutex_unlock(struct mutex *lock) * The unlocking fastpath is the 0->1 transition from 'locked' * into 'unlocked' state: */ +#ifndef CONFIG_DEBUG_MUTEXES + /* + * When debugging is enabled we must not clear the owner before time, + * the slow path will always be taken, and that clears the owner field + * after verifying that it was indeed current. + */ + mutex_clear_owner(lock); +#endif __mutex_fastpath_unlock(&lock->count, __mutex_unlock_slowpath); } @@ -129,21 +144,76 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, { struct task_struct *task = current; struct mutex_waiter waiter; - unsigned int old_val; unsigned long flags; + preempt_disable(); + mutex_acquire(&lock->dep_map, subclass, 0, ip); +#if defined(CONFIG_SMP) && !defined(CONFIG_DEBUG_MUTEXES) && \ + !defined(CONFIG_HAVE_DEFAULT_NO_SPIN_MUTEXES) + /* + * Optimistic spinning. + * + * We try to spin for acquisition when we find that there are no + * pending waiters and the lock owner is currently running on a + * (different) CPU. + * + * The rationale is that if the lock owner is running, it is likely to + * release the lock soon. + * + * Since this needs the lock owner, and this mutex implementation + * doesn't track the owner atomically in the lock field, we need to + * track it non-atomically. + * + * We can't do this for DEBUG_MUTEXES because that relies on wait_lock + * to serialize everything. + */ + + for (;;) { + struct thread_info *owner; + + /* + * If there's an owner, wait for it to either + * release the lock or go to sleep. + */ + owner = ACCESS_ONCE(lock->owner); + if (owner && !mutex_spin_on_owner(lock, owner)) + break; + + if (atomic_cmpxchg(&lock->count, 1, 0) == 1) { + lock_acquired(&lock->dep_map, ip); + mutex_set_owner(lock); + preempt_enable(); + return 0; + } + + /* + * When there's no owner, we might have preempted between the + * owner acquiring the lock and setting the owner field. If + * we're an RT task that will live-lock because we won't let + * the owner complete. + */ + if (!owner && (need_resched() || rt_task(task))) + break; + + /* + * The cpu_relax() call is a compiler barrier which forces + * everything in this loop to be re-loaded. We don't need + * memory barriers as we'll eventually observe the right + * values at the cost of a few extra spins. + */ + cpu_relax(); + } +#endif spin_lock_mutex(&lock->wait_lock, flags); debug_mutex_lock_common(lock, &waiter); - mutex_acquire(&lock->dep_map, subclass, 0, ip); debug_mutex_add_waiter(lock, &waiter, task_thread_info(task)); /* add waiting tasks to the end of the waitqueue (FIFO): */ list_add_tail(&waiter.list, &lock->wait_list); waiter.task = task; - old_val = atomic_xchg(&lock->count, -1); - if (old_val == 1) + if (atomic_xchg(&lock->count, -1) == 1) goto done; lock_contended(&lock->dep_map, ip); @@ -158,8 +228,7 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, * that when we release the lock, we properly wake up the * other waiters: */ - old_val = atomic_xchg(&lock->count, -1); - if (old_val == 1) + if (atomic_xchg(&lock->count, -1) == 1) break; /* @@ -173,21 +242,24 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass, spin_unlock_mutex(&lock->wait_lock, flags); debug_mutex_free_waiter(&waiter); + preempt_enable(); return -EINTR; } __set_task_state(task, state); /* didnt get the lock, go to sleep: */ spin_unlock_mutex(&lock->wait_lock, flags); + preempt_enable_no_resched(); schedule(); + preempt_disable(); spin_lock_mutex(&lock->wait_lock, flags); } done: lock_acquired(&lock->dep_map, ip); /* got the lock - rejoice! */ - mutex_remove_waiter(lock, &waiter, task_thread_info(task)); - debug_mutex_set_owner(lock, task_thread_info(task)); + mutex_remove_waiter(lock, &waiter, current_thread_info()); + mutex_set_owner(lock); /* set it to 0 if there are no waiters left: */ if (likely(list_empty(&lock->wait_list))) @@ -196,6 +268,7 @@ done: spin_unlock_mutex(&lock->wait_lock, flags); debug_mutex_free_waiter(&waiter); + preempt_enable(); return 0; } @@ -222,7 +295,8 @@ int __sched mutex_lock_interruptible_nested(struct mutex *lock, unsigned int subclass) { might_sleep(); - return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, subclass, _RET_IP_); + return __mutex_lock_common(lock, TASK_INTERRUPTIBLE, + subclass, _RET_IP_); } EXPORT_SYMBOL_GPL(mutex_lock_interruptible_nested); @@ -260,8 +334,6 @@ __mutex_unlock_common_slowpath(atomic_t *lock_count, int nested) wake_up_process(waiter->task); } - debug_mutex_clear_owner(lock); - spin_unlock_mutex(&lock->wait_lock, flags); } @@ -298,18 +370,30 @@ __mutex_lock_interruptible_slowpath(atomic_t *lock_count); */ int __sched mutex_lock_interruptible(struct mutex *lock) { + int ret; + might_sleep(); - return __mutex_fastpath_lock_retval + ret = __mutex_fastpath_lock_retval (&lock->count, __mutex_lock_interruptible_slowpath); + if (!ret) + mutex_set_owner(lock); + + return ret; } EXPORT_SYMBOL(mutex_lock_interruptible); int __sched mutex_lock_killable(struct mutex *lock) { + int ret; + might_sleep(); - return __mutex_fastpath_lock_retval + ret = __mutex_fastpath_lock_retval (&lock->count, __mutex_lock_killable_slowpath); + if (!ret) + mutex_set_owner(lock); + + return ret; } EXPORT_SYMBOL(mutex_lock_killable); @@ -352,9 +436,10 @@ static inline int __mutex_trylock_slowpath(atomic_t *lock_count) prev = atomic_xchg(&lock->count, -1); if (likely(prev == 1)) { - debug_mutex_set_owner(lock, current_thread_info()); + mutex_set_owner(lock); mutex_acquire(&lock->dep_map, 0, 1, _RET_IP_); } + /* Set it back to 0 if there are no waiters: */ if (likely(list_empty(&lock->wait_list))) atomic_set(&lock->count, 0); @@ -380,8 +465,36 @@ static inline int __mutex_trylock_slowpath(atomic_t *lock_count) */ int __sched mutex_trylock(struct mutex *lock) { - return __mutex_fastpath_trylock(&lock->count, - __mutex_trylock_slowpath); -} + int ret; + + ret = __mutex_fastpath_trylock(&lock->count, __mutex_trylock_slowpath); + if (ret) + mutex_set_owner(lock); + return ret; +} EXPORT_SYMBOL(mutex_trylock); + +/** + * atomic_dec_and_mutex_lock - return holding mutex if we dec to 0 + * @cnt: the atomic which we are to dec + * @lock: the mutex to return holding if we dec to 0 + * + * return true and hold lock if we dec to 0, return false otherwise + */ +int atomic_dec_and_mutex_lock(atomic_t *cnt, struct mutex *lock) +{ + /* dec if we can't possibly hit 0 */ + if (atomic_add_unless(cnt, -1, 1)) + return 0; + /* we might hit 0, so take the lock */ + mutex_lock(lock); + if (!atomic_dec_and_test(cnt)) { + /* when we actually did the dec, we didn't hit 0 */ + mutex_unlock(lock); + return 0; + } + /* we hit 0, and we hold the lock */ + return 1; +} +EXPORT_SYMBOL(atomic_dec_and_mutex_lock); diff --git a/kernel/mutex.h b/kernel/mutex.h index a075dafbb29..67578ca48f9 100644 --- a/kernel/mutex.h +++ b/kernel/mutex.h @@ -16,8 +16,26 @@ #define mutex_remove_waiter(lock, waiter, ti) \ __list_del((waiter)->list.prev, (waiter)->list.next) -#define debug_mutex_set_owner(lock, new_owner) do { } while (0) -#define debug_mutex_clear_owner(lock) do { } while (0) +#ifdef CONFIG_SMP +static inline void mutex_set_owner(struct mutex *lock) +{ + lock->owner = current_thread_info(); +} + +static inline void mutex_clear_owner(struct mutex *lock) +{ + lock->owner = NULL; +} +#else +static inline void mutex_set_owner(struct mutex *lock) +{ +} + +static inline void mutex_clear_owner(struct mutex *lock) +{ +} +#endif + #define debug_mutex_wake_waiter(lock, waiter) do { } while (0) #define debug_mutex_free_waiter(waiter) do { } while (0) #define debug_mutex_add_waiter(lock, waiter, ti) do { } while (0) diff --git a/kernel/ns_cgroup.c b/kernel/ns_cgroup.c index 78bc3fdac0d..5aa854f9e5a 100644 --- a/kernel/ns_cgroup.c +++ b/kernel/ns_cgroup.c @@ -34,7 +34,7 @@ int ns_cgroup_clone(struct task_struct *task, struct pid *pid) /* * Rules: - * 1. you can only enter a cgroup which is a child of your current + * 1. you can only enter a cgroup which is a descendant of your current * cgroup * 2. you can only place another process into a cgroup if * a. you have CAP_SYS_ADMIN @@ -45,21 +45,15 @@ int ns_cgroup_clone(struct task_struct *task, struct pid *pid) static int ns_can_attach(struct cgroup_subsys *ss, struct cgroup *new_cgroup, struct task_struct *task) { - struct cgroup *orig; - if (current != task) { if (!capable(CAP_SYS_ADMIN)) return -EPERM; - if (!cgroup_is_descendant(new_cgroup)) + if (!cgroup_is_descendant(new_cgroup, current)) return -EPERM; } - if (atomic_read(&new_cgroup->count) != 0) - return -EPERM; - - orig = task_cgroup(task, ns_subsys_id); - if (orig && orig != new_cgroup->parent) + if (!cgroup_is_descendant(new_cgroup, task)) return -EPERM; return 0; @@ -77,7 +71,7 @@ static struct cgroup_subsys_state *ns_create(struct cgroup_subsys *ss, if (!capable(CAP_SYS_ADMIN)) return ERR_PTR(-EPERM); - if (!cgroup_is_descendant(cgroup)) + if (!cgroup_is_descendant(cgroup, current)) return ERR_PTR(-EPERM); ns_cgroup = kzalloc(sizeof(*ns_cgroup), GFP_KERNEL); diff --git a/kernel/nsproxy.c b/kernel/nsproxy.c index 63598dca2d0..09b4ff9711b 100644 --- a/kernel/nsproxy.c +++ b/kernel/nsproxy.c @@ -26,19 +26,14 @@ static struct kmem_cache *nsproxy_cachep; struct nsproxy init_nsproxy = INIT_NSPROXY(init_nsproxy); -/* - * creates a copy of "orig" with refcount 1. - */ -static inline struct nsproxy *clone_nsproxy(struct nsproxy *orig) +static inline struct nsproxy *create_nsproxy(void) { - struct nsproxy *ns; + struct nsproxy *nsproxy; - ns = kmem_cache_alloc(nsproxy_cachep, GFP_KERNEL); - if (ns) { - memcpy(ns, orig, sizeof(struct nsproxy)); - atomic_set(&ns->count, 1); - } - return ns; + nsproxy = kmem_cache_alloc(nsproxy_cachep, GFP_KERNEL); + if (nsproxy) + atomic_set(&nsproxy->count, 1); + return nsproxy; } /* @@ -52,7 +47,7 @@ static struct nsproxy *create_new_namespaces(unsigned long flags, struct nsproxy *new_nsp; int err; - new_nsp = clone_nsproxy(tsk->nsproxy); + new_nsp = create_nsproxy(); if (!new_nsp) return ERR_PTR(-ENOMEM); diff --git a/kernel/panic.c b/kernel/panic.c index 2a2ff36ff44..512ab73b0ca 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -8,19 +8,19 @@ * This function is used through-out the kernel (including mm and fs) * to indicate a major problem. */ +#include <linux/debug_locks.h> +#include <linux/interrupt.h> +#include <linux/kallsyms.h> +#include <linux/notifier.h> #include <linux/module.h> -#include <linux/sched.h> -#include <linux/delay.h> +#include <linux/random.h> #include <linux/reboot.h> -#include <linux/notifier.h> -#include <linux/init.h> +#include <linux/delay.h> +#include <linux/kexec.h> +#include <linux/sched.h> #include <linux/sysrq.h> -#include <linux/interrupt.h> +#include <linux/init.h> #include <linux/nmi.h> -#include <linux/kexec.h> -#include <linux/debug_locks.h> -#include <linux/random.h> -#include <linux/kallsyms.h> #include <linux/dmi.h> int panic_on_oops; @@ -52,19 +52,15 @@ EXPORT_SYMBOL(panic_blink); * * This function never returns. */ - NORET_TYPE void panic(const char * fmt, ...) { - long i; static char buf[1024]; va_list args; -#if defined(CONFIG_S390) - unsigned long caller = (unsigned long) __builtin_return_address(0); -#endif + long i; /* - * It's possible to come here directly from a panic-assertion and not - * have preempt disabled. Some functions called from here want + * It's possible to come here directly from a panic-assertion and + * not have preempt disabled. Some functions called from here want * preempt to be disabled. No point enabling it later though... */ preempt_disable(); @@ -74,7 +70,9 @@ NORET_TYPE void panic(const char * fmt, ...) vsnprintf(buf, sizeof(buf), fmt, args); va_end(args); printk(KERN_EMERG "Kernel panic - not syncing: %s\n",buf); - bust_spinlocks(0); +#ifdef CONFIG_DEBUG_BUGVERBOSE + dump_stack(); +#endif /* * If we have crashed and we have a crash kernel loaded let it handle @@ -83,14 +81,12 @@ NORET_TYPE void panic(const char * fmt, ...) */ crash_kexec(NULL); -#ifdef CONFIG_SMP /* * Note smp_send_stop is the usual smp shutdown function, which * unfortunately means it may not be hardened to work in a panic * situation. */ smp_send_stop(); -#endif atomic_notifier_call_chain(&panic_notifier_list, 0, buf); @@ -99,19 +95,21 @@ NORET_TYPE void panic(const char * fmt, ...) if (panic_timeout > 0) { /* - * Delay timeout seconds before rebooting the machine. - * We can't use the "normal" timers since we just panicked.. - */ - printk(KERN_EMERG "Rebooting in %d seconds..",panic_timeout); + * Delay timeout seconds before rebooting the machine. + * We can't use the "normal" timers since we just panicked. + */ + printk(KERN_EMERG "Rebooting in %d seconds..", panic_timeout); + for (i = 0; i < panic_timeout*1000; ) { touch_nmi_watchdog(); i += panic_blink(i); mdelay(1); i++; } - /* This will not be a clean reboot, with everything - * shutting down. But if there is a chance of - * rebooting the system it will be rebooted. + /* + * This will not be a clean reboot, with everything + * shutting down. But if there is a chance of + * rebooting the system it will be rebooted. */ emergency_restart(); } @@ -124,38 +122,44 @@ NORET_TYPE void panic(const char * fmt, ...) } #endif #if defined(CONFIG_S390) - disabled_wait(caller); + { + unsigned long caller; + + caller = (unsigned long)__builtin_return_address(0); + disabled_wait(caller); + } #endif local_irq_enable(); - for (i = 0;;) { + for (i = 0; ; ) { touch_softlockup_watchdog(); i += panic_blink(i); mdelay(1); i++; } + bust_spinlocks(0); } EXPORT_SYMBOL(panic); struct tnt { - u8 bit; - char true; - char false; + u8 bit; + char true; + char false; }; static const struct tnt tnts[] = { - { TAINT_PROPRIETARY_MODULE, 'P', 'G' }, - { TAINT_FORCED_MODULE, 'F', ' ' }, - { TAINT_UNSAFE_SMP, 'S', ' ' }, - { TAINT_FORCED_RMMOD, 'R', ' ' }, - { TAINT_MACHINE_CHECK, 'M', ' ' }, - { TAINT_BAD_PAGE, 'B', ' ' }, - { TAINT_USER, 'U', ' ' }, - { TAINT_DIE, 'D', ' ' }, - { TAINT_OVERRIDDEN_ACPI_TABLE, 'A', ' ' }, - { TAINT_WARN, 'W', ' ' }, - { TAINT_CRAP, 'C', ' ' }, + { TAINT_PROPRIETARY_MODULE, 'P', 'G' }, + { TAINT_FORCED_MODULE, 'F', ' ' }, + { TAINT_UNSAFE_SMP, 'S', ' ' }, + { TAINT_FORCED_RMMOD, 'R', ' ' }, + { TAINT_MACHINE_CHECK, 'M', ' ' }, + { TAINT_BAD_PAGE, 'B', ' ' }, + { TAINT_USER, 'U', ' ' }, + { TAINT_DIE, 'D', ' ' }, + { TAINT_OVERRIDDEN_ACPI_TABLE, 'A', ' ' }, + { TAINT_WARN, 'W', ' ' }, + { TAINT_CRAP, 'C', ' ' }, }; /** @@ -192,7 +196,8 @@ const char *print_tainted(void) *s = 0; } else snprintf(buf, sizeof(buf), "Not tainted"); - return(buf); + + return buf; } int test_taint(unsigned flag) @@ -208,7 +213,16 @@ unsigned long get_taint(void) void add_taint(unsigned flag) { - debug_locks = 0; /* can't trust the integrity of the kernel anymore */ + /* + * Can't trust the integrity of the kernel anymore. + * We don't call directly debug_locks_off() because the issue + * is not necessarily serious enough to set oops_in_progress to 1 + * Also we want to keep up lockdep for staging development and + * post-warning case. + */ + if (flag != TAINT_CRAP && flag != TAINT_WARN && __debug_locks_off()) + printk(KERN_WARNING "Disabling lock debugging due to kernel taint\n"); + set_bit(flag, &tainted_mask); } EXPORT_SYMBOL(add_taint); @@ -263,8 +277,8 @@ static void do_oops_enter_exit(void) } /* - * Return true if the calling CPU is allowed to print oops-related info. This - * is a bit racy.. + * Return true if the calling CPU is allowed to print oops-related info. + * This is a bit racy.. */ int oops_may_print(void) { @@ -273,20 +287,23 @@ int oops_may_print(void) /* * Called when the architecture enters its oops handler, before it prints - * anything. If this is the first CPU to oops, and it's oopsing the first time - * then let it proceed. + * anything. If this is the first CPU to oops, and it's oopsing the first + * time then let it proceed. * - * This is all enabled by the pause_on_oops kernel boot option. We do all this - * to ensure that oopses don't scroll off the screen. It has the side-effect - * of preventing later-oopsing CPUs from mucking up the display, too. + * This is all enabled by the pause_on_oops kernel boot option. We do all + * this to ensure that oopses don't scroll off the screen. It has the + * side-effect of preventing later-oopsing CPUs from mucking up the display, + * too. * - * It turns out that the CPU which is allowed to print ends up pausing for the - * right duration, whereas all the other CPUs pause for twice as long: once in - * oops_enter(), once in oops_exit(). + * It turns out that the CPU which is allowed to print ends up pausing for + * the right duration, whereas all the other CPUs pause for twice as long: + * once in oops_enter(), once in oops_exit(). */ void oops_enter(void) { - debug_locks_off(); /* can't trust the integrity of the kernel anymore */ + tracing_off(); + /* can't trust the integrity of the kernel anymore: */ + debug_locks_off(); do_oops_enter_exit(); } @@ -324,46 +341,61 @@ void oops_exit(void) } #ifdef WANT_WARN_ON_SLOWPATH -void warn_slowpath(const char *file, int line, const char *fmt, ...) -{ +struct slowpath_args { + const char *fmt; va_list args; - char function[KSYM_SYMBOL_LEN]; - unsigned long caller = (unsigned long)__builtin_return_address(0); - const char *board; +}; - sprint_symbol(function, caller); +static void warn_slowpath_common(const char *file, int line, void *caller, struct slowpath_args *args) +{ + const char *board; printk(KERN_WARNING "------------[ cut here ]------------\n"); - printk(KERN_WARNING "WARNING: at %s:%d %s()\n", file, - line, function); + printk(KERN_WARNING "WARNING: at %s:%d %pS()\n", file, line, caller); board = dmi_get_system_info(DMI_PRODUCT_NAME); if (board) printk(KERN_WARNING "Hardware name: %s\n", board); - if (fmt) { - va_start(args, fmt); - vprintk(fmt, args); - va_end(args); - } + if (args) + vprintk(args->fmt, args->args); print_modules(); dump_stack(); print_oops_end_marker(); add_taint(TAINT_WARN); } -EXPORT_SYMBOL(warn_slowpath); + +void warn_slowpath_fmt(const char *file, int line, const char *fmt, ...) +{ + struct slowpath_args args; + + args.fmt = fmt; + va_start(args.args, fmt); + warn_slowpath_common(file, line, __builtin_return_address(0), &args); + va_end(args.args); +} +EXPORT_SYMBOL(warn_slowpath_fmt); + +void warn_slowpath_null(const char *file, int line) +{ + warn_slowpath_common(file, line, __builtin_return_address(0), NULL); +} +EXPORT_SYMBOL(warn_slowpath_null); #endif #ifdef CONFIG_CC_STACKPROTECTOR + /* * Called when gcc's -fstack-protector feature is used, and * gcc detects corruption of the on-stack canary value */ void __stack_chk_fail(void) { - panic("stack-protector: Kernel stack is corrupted"); + panic("stack-protector: Kernel stack is corrupted in: %p\n", + __builtin_return_address(0)); } EXPORT_SYMBOL(__stack_chk_fail); + #endif core_param(panic, panic_timeout, int, 0644); diff --git a/kernel/params.c b/kernel/params.c index a1e3025b19a..7f6912ced2b 100644 --- a/kernel/params.c +++ b/kernel/params.c @@ -217,7 +217,19 @@ int param_set_charp(const char *val, struct kernel_param *kp) return -ENOSPC; } - *(char **)kp->arg = (char *)val; + if (kp->flags & KPARAM_KMALLOCED) + kfree(*(char **)kp->arg); + + /* This is a hack. We can't need to strdup in early boot, and we + * don't need to; this mangled commandline is preserved. */ + if (slab_is_available()) { + kp->flags |= KPARAM_KMALLOCED; + *(char **)kp->arg = kstrdup(val, GFP_KERNEL); + if (!kp->arg) + return -ENOMEM; + } else + *(const char **)kp->arg = val; + return 0; } @@ -226,44 +238,63 @@ int param_get_charp(char *buffer, struct kernel_param *kp) return sprintf(buffer, "%s", *((char **)kp->arg)); } +/* Actually could be a bool or an int, for historical reasons. */ int param_set_bool(const char *val, struct kernel_param *kp) { + bool v; + /* No equals means "set"... */ if (!val) val = "1"; /* One of =[yYnN01] */ switch (val[0]) { case 'y': case 'Y': case '1': - *(int *)kp->arg = 1; - return 0; + v = true; + break; case 'n': case 'N': case '0': - *(int *)kp->arg = 0; - return 0; + v = false; + break; + default: + return -EINVAL; } - return -EINVAL; + + if (kp->flags & KPARAM_ISBOOL) + *(bool *)kp->arg = v; + else + *(int *)kp->arg = v; + return 0; } int param_get_bool(char *buffer, struct kernel_param *kp) { + bool val; + if (kp->flags & KPARAM_ISBOOL) + val = *(bool *)kp->arg; + else + val = *(int *)kp->arg; + /* Y and N chosen as being relatively non-coder friendly */ - return sprintf(buffer, "%c", (*(int *)kp->arg) ? 'Y' : 'N'); + return sprintf(buffer, "%c", val ? 'Y' : 'N'); } +/* This one must be bool. */ int param_set_invbool(const char *val, struct kernel_param *kp) { - int boolval, ret; + int ret; + bool boolval; struct kernel_param dummy; dummy.arg = &boolval; + dummy.flags = KPARAM_ISBOOL; ret = param_set_bool(val, &dummy); if (ret == 0) - *(int *)kp->arg = !boolval; + *(bool *)kp->arg = !boolval; return ret; } int param_get_invbool(char *buffer, struct kernel_param *kp) { - return sprintf(buffer, "%c", (*(int *)kp->arg) ? 'N' : 'Y'); + return sprintf(buffer, "%c", (*(bool *)kp->arg) ? 'N' : 'Y'); } /* We break the rule and mangle the string. */ @@ -571,6 +602,15 @@ void module_param_sysfs_remove(struct module *mod) } #endif +void destroy_params(const struct kernel_param *params, unsigned num) +{ + unsigned int i; + + for (i = 0; i < num; i++) + if (params[i].flags & KPARAM_KMALLOCED) + kfree(*(char **)params[i].arg); +} + static void __init kernel_add_sysfs_param(const char *name, struct kernel_param *kparam, unsigned int name_skip) diff --git a/kernel/perf_counter.c b/kernel/perf_counter.c new file mode 100644 index 00000000000..e0d91fdf0c3 --- /dev/null +++ b/kernel/perf_counter.c @@ -0,0 +1,4962 @@ +/* + * Performance counter core code + * + * Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de> + * Copyright (C) 2008-2009 Red Hat, Inc., Ingo Molnar + * Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * Copyright © 2009 Paul Mackerras, IBM Corp. <paulus@au1.ibm.com> + * + * For licensing details see kernel-base/COPYING + */ + +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/cpu.h> +#include <linux/smp.h> +#include <linux/file.h> +#include <linux/poll.h> +#include <linux/sysfs.h> +#include <linux/dcache.h> +#include <linux/percpu.h> +#include <linux/ptrace.h> +#include <linux/vmstat.h> +#include <linux/hardirq.h> +#include <linux/rculist.h> +#include <linux/uaccess.h> +#include <linux/syscalls.h> +#include <linux/anon_inodes.h> +#include <linux/kernel_stat.h> +#include <linux/perf_counter.h> + +#include <asm/irq_regs.h> + +/* + * Each CPU has a list of per CPU counters: + */ +DEFINE_PER_CPU(struct perf_cpu_context, perf_cpu_context); + +int perf_max_counters __read_mostly = 1; +static int perf_reserved_percpu __read_mostly; +static int perf_overcommit __read_mostly = 1; + +static atomic_t nr_counters __read_mostly; +static atomic_t nr_mmap_counters __read_mostly; +static atomic_t nr_comm_counters __read_mostly; +static atomic_t nr_task_counters __read_mostly; + +/* + * perf counter paranoia level: + * -1 - not paranoid at all + * 0 - disallow raw tracepoint access for unpriv + * 1 - disallow cpu counters for unpriv + * 2 - disallow kernel profiling for unpriv + */ +int sysctl_perf_counter_paranoid __read_mostly = 1; + +static inline bool perf_paranoid_tracepoint_raw(void) +{ + return sysctl_perf_counter_paranoid > -1; +} + +static inline bool perf_paranoid_cpu(void) +{ + return sysctl_perf_counter_paranoid > 0; +} + +static inline bool perf_paranoid_kernel(void) +{ + return sysctl_perf_counter_paranoid > 1; +} + +int sysctl_perf_counter_mlock __read_mostly = 512; /* 'free' kb per user */ + +/* + * max perf counter sample rate + */ +int sysctl_perf_counter_sample_rate __read_mostly = 100000; + +static atomic64_t perf_counter_id; + +/* + * Lock for (sysadmin-configurable) counter reservations: + */ +static DEFINE_SPINLOCK(perf_resource_lock); + +/* + * Architecture provided APIs - weak aliases: + */ +extern __weak const struct pmu *hw_perf_counter_init(struct perf_counter *counter) +{ + return NULL; +} + +void __weak hw_perf_disable(void) { barrier(); } +void __weak hw_perf_enable(void) { barrier(); } + +void __weak hw_perf_counter_setup(int cpu) { barrier(); } +void __weak hw_perf_counter_setup_online(int cpu) { barrier(); } + +int __weak +hw_perf_group_sched_in(struct perf_counter *group_leader, + struct perf_cpu_context *cpuctx, + struct perf_counter_context *ctx, int cpu) +{ + return 0; +} + +void __weak perf_counter_print_debug(void) { } + +static DEFINE_PER_CPU(int, disable_count); + +void __perf_disable(void) +{ + __get_cpu_var(disable_count)++; +} + +bool __perf_enable(void) +{ + return !--__get_cpu_var(disable_count); +} + +void perf_disable(void) +{ + __perf_disable(); + hw_perf_disable(); +} + +void perf_enable(void) +{ + if (__perf_enable()) + hw_perf_enable(); +} + +static void get_ctx(struct perf_counter_context *ctx) +{ + WARN_ON(!atomic_inc_not_zero(&ctx->refcount)); +} + +static void free_ctx(struct rcu_head *head) +{ + struct perf_counter_context *ctx; + + ctx = container_of(head, struct perf_counter_context, rcu_head); + kfree(ctx); +} + +static void put_ctx(struct perf_counter_context *ctx) +{ + if (atomic_dec_and_test(&ctx->refcount)) { + if (ctx->parent_ctx) + put_ctx(ctx->parent_ctx); + if (ctx->task) + put_task_struct(ctx->task); + call_rcu(&ctx->rcu_head, free_ctx); + } +} + +static void unclone_ctx(struct perf_counter_context *ctx) +{ + if (ctx->parent_ctx) { + put_ctx(ctx->parent_ctx); + ctx->parent_ctx = NULL; + } +} + +/* + * If we inherit counters we want to return the parent counter id + * to userspace. + */ +static u64 primary_counter_id(struct perf_counter *counter) +{ + u64 id = counter->id; + + if (counter->parent) + id = counter->parent->id; + + return id; +} + +/* + * Get the perf_counter_context for a task and lock it. + * This has to cope with with the fact that until it is locked, + * the context could get moved to another task. + */ +static struct perf_counter_context * +perf_lock_task_context(struct task_struct *task, unsigned long *flags) +{ + struct perf_counter_context *ctx; + + rcu_read_lock(); + retry: + ctx = rcu_dereference(task->perf_counter_ctxp); + if (ctx) { + /* + * If this context is a clone of another, it might + * get swapped for another underneath us by + * perf_counter_task_sched_out, though the + * rcu_read_lock() protects us from any context + * getting freed. Lock the context and check if it + * got swapped before we could get the lock, and retry + * if so. If we locked the right context, then it + * can't get swapped on us any more. + */ + spin_lock_irqsave(&ctx->lock, *flags); + if (ctx != rcu_dereference(task->perf_counter_ctxp)) { + spin_unlock_irqrestore(&ctx->lock, *flags); + goto retry; + } + + if (!atomic_inc_not_zero(&ctx->refcount)) { + spin_unlock_irqrestore(&ctx->lock, *flags); + ctx = NULL; + } + } + rcu_read_unlock(); + return ctx; +} + +/* + * Get the context for a task and increment its pin_count so it + * can't get swapped to another task. This also increments its + * reference count so that the context can't get freed. + */ +static struct perf_counter_context *perf_pin_task_context(struct task_struct *task) +{ + struct perf_counter_context *ctx; + unsigned long flags; + + ctx = perf_lock_task_context(task, &flags); + if (ctx) { + ++ctx->pin_count; + spin_unlock_irqrestore(&ctx->lock, flags); + } + return ctx; +} + +static void perf_unpin_context(struct perf_counter_context *ctx) +{ + unsigned long flags; + + spin_lock_irqsave(&ctx->lock, flags); + --ctx->pin_count; + spin_unlock_irqrestore(&ctx->lock, flags); + put_ctx(ctx); +} + +/* + * Add a counter from the lists for its context. + * Must be called with ctx->mutex and ctx->lock held. + */ +static void +list_add_counter(struct perf_counter *counter, struct perf_counter_context *ctx) +{ + struct perf_counter *group_leader = counter->group_leader; + + /* + * Depending on whether it is a standalone or sibling counter, + * add it straight to the context's counter list, or to the group + * leader's sibling list: + */ + if (group_leader == counter) + list_add_tail(&counter->list_entry, &ctx->counter_list); + else { + list_add_tail(&counter->list_entry, &group_leader->sibling_list); + group_leader->nr_siblings++; + } + + list_add_rcu(&counter->event_entry, &ctx->event_list); + ctx->nr_counters++; + if (counter->attr.inherit_stat) + ctx->nr_stat++; +} + +/* + * Remove a counter from the lists for its context. + * Must be called with ctx->mutex and ctx->lock held. + */ +static void +list_del_counter(struct perf_counter *counter, struct perf_counter_context *ctx) +{ + struct perf_counter *sibling, *tmp; + + if (list_empty(&counter->list_entry)) + return; + ctx->nr_counters--; + if (counter->attr.inherit_stat) + ctx->nr_stat--; + + list_del_init(&counter->list_entry); + list_del_rcu(&counter->event_entry); + + if (counter->group_leader != counter) + counter->group_leader->nr_siblings--; + + /* + * If this was a group counter with sibling counters then + * upgrade the siblings to singleton counters by adding them + * to the context list directly: + */ + list_for_each_entry_safe(sibling, tmp, + &counter->sibling_list, list_entry) { + + list_move_tail(&sibling->list_entry, &ctx->counter_list); + sibling->group_leader = sibling; + } +} + +static void +counter_sched_out(struct perf_counter *counter, + struct perf_cpu_context *cpuctx, + struct perf_counter_context *ctx) +{ + if (counter->state != PERF_COUNTER_STATE_ACTIVE) + return; + + counter->state = PERF_COUNTER_STATE_INACTIVE; + if (counter->pending_disable) { + counter->pending_disable = 0; + counter->state = PERF_COUNTER_STATE_OFF; + } + counter->tstamp_stopped = ctx->time; + counter->pmu->disable(counter); + counter->oncpu = -1; + + if (!is_software_counter(counter)) + cpuctx->active_oncpu--; + ctx->nr_active--; + if (counter->attr.exclusive || !cpuctx->active_oncpu) + cpuctx->exclusive = 0; +} + +static void +group_sched_out(struct perf_counter *group_counter, + struct perf_cpu_context *cpuctx, + struct perf_counter_context *ctx) +{ + struct perf_counter *counter; + + if (group_counter->state != PERF_COUNTER_STATE_ACTIVE) + return; + + counter_sched_out(group_counter, cpuctx, ctx); + + /* + * Schedule out siblings (if any): + */ + list_for_each_entry(counter, &group_counter->sibling_list, list_entry) + counter_sched_out(counter, cpuctx, ctx); + + if (group_counter->attr.exclusive) + cpuctx->exclusive = 0; +} + +/* + * Cross CPU call to remove a performance counter + * + * We disable the counter on the hardware level first. After that we + * remove it from the context list. + */ +static void __perf_counter_remove_from_context(void *info) +{ + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_counter *counter = info; + struct perf_counter_context *ctx = counter->ctx; + + /* + * If this is a task context, we need to check whether it is + * the current task context of this cpu. If not it has been + * scheduled out before the smp call arrived. + */ + if (ctx->task && cpuctx->task_ctx != ctx) + return; + + spin_lock(&ctx->lock); + /* + * Protect the list operation against NMI by disabling the + * counters on a global level. + */ + perf_disable(); + + counter_sched_out(counter, cpuctx, ctx); + + list_del_counter(counter, ctx); + + if (!ctx->task) { + /* + * Allow more per task counters with respect to the + * reservation: + */ + cpuctx->max_pertask = + min(perf_max_counters - ctx->nr_counters, + perf_max_counters - perf_reserved_percpu); + } + + perf_enable(); + spin_unlock(&ctx->lock); +} + + +/* + * Remove the counter from a task's (or a CPU's) list of counters. + * + * Must be called with ctx->mutex held. + * + * CPU counters are removed with a smp call. For task counters we only + * call when the task is on a CPU. + * + * If counter->ctx is a cloned context, callers must make sure that + * every task struct that counter->ctx->task could possibly point to + * remains valid. This is OK when called from perf_release since + * that only calls us on the top-level context, which can't be a clone. + * When called from perf_counter_exit_task, it's OK because the + * context has been detached from its task. + */ +static void perf_counter_remove_from_context(struct perf_counter *counter) +{ + struct perf_counter_context *ctx = counter->ctx; + struct task_struct *task = ctx->task; + + if (!task) { + /* + * Per cpu counters are removed via an smp call and + * the removal is always sucessful. + */ + smp_call_function_single(counter->cpu, + __perf_counter_remove_from_context, + counter, 1); + return; + } + +retry: + task_oncpu_function_call(task, __perf_counter_remove_from_context, + counter); + + spin_lock_irq(&ctx->lock); + /* + * If the context is active we need to retry the smp call. + */ + if (ctx->nr_active && !list_empty(&counter->list_entry)) { + spin_unlock_irq(&ctx->lock); + goto retry; + } + + /* + * The lock prevents that this context is scheduled in so we + * can remove the counter safely, if the call above did not + * succeed. + */ + if (!list_empty(&counter->list_entry)) { + list_del_counter(counter, ctx); + } + spin_unlock_irq(&ctx->lock); +} + +static inline u64 perf_clock(void) +{ + return cpu_clock(smp_processor_id()); +} + +/* + * Update the record of the current time in a context. + */ +static void update_context_time(struct perf_counter_context *ctx) +{ + u64 now = perf_clock(); + + ctx->time += now - ctx->timestamp; + ctx->timestamp = now; +} + +/* + * Update the total_time_enabled and total_time_running fields for a counter. + */ +static void update_counter_times(struct perf_counter *counter) +{ + struct perf_counter_context *ctx = counter->ctx; + u64 run_end; + + if (counter->state < PERF_COUNTER_STATE_INACTIVE || + counter->group_leader->state < PERF_COUNTER_STATE_INACTIVE) + return; + + counter->total_time_enabled = ctx->time - counter->tstamp_enabled; + + if (counter->state == PERF_COUNTER_STATE_INACTIVE) + run_end = counter->tstamp_stopped; + else + run_end = ctx->time; + + counter->total_time_running = run_end - counter->tstamp_running; +} + +/* + * Update total_time_enabled and total_time_running for all counters in a group. + */ +static void update_group_times(struct perf_counter *leader) +{ + struct perf_counter *counter; + + update_counter_times(leader); + list_for_each_entry(counter, &leader->sibling_list, list_entry) + update_counter_times(counter); +} + +/* + * Cross CPU call to disable a performance counter + */ +static void __perf_counter_disable(void *info) +{ + struct perf_counter *counter = info; + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_counter_context *ctx = counter->ctx; + + /* + * If this is a per-task counter, need to check whether this + * counter's task is the current task on this cpu. + */ + if (ctx->task && cpuctx->task_ctx != ctx) + return; + + spin_lock(&ctx->lock); + + /* + * If the counter is on, turn it off. + * If it is in error state, leave it in error state. + */ + if (counter->state >= PERF_COUNTER_STATE_INACTIVE) { + update_context_time(ctx); + update_group_times(counter); + if (counter == counter->group_leader) + group_sched_out(counter, cpuctx, ctx); + else + counter_sched_out(counter, cpuctx, ctx); + counter->state = PERF_COUNTER_STATE_OFF; + } + + spin_unlock(&ctx->lock); +} + +/* + * Disable a counter. + * + * If counter->ctx is a cloned context, callers must make sure that + * every task struct that counter->ctx->task could possibly point to + * remains valid. This condition is satisifed when called through + * perf_counter_for_each_child or perf_counter_for_each because they + * hold the top-level counter's child_mutex, so any descendant that + * goes to exit will block in sync_child_counter. + * When called from perf_pending_counter it's OK because counter->ctx + * is the current context on this CPU and preemption is disabled, + * hence we can't get into perf_counter_task_sched_out for this context. + */ +static void perf_counter_disable(struct perf_counter *counter) +{ + struct perf_counter_context *ctx = counter->ctx; + struct task_struct *task = ctx->task; + + if (!task) { + /* + * Disable the counter on the cpu that it's on + */ + smp_call_function_single(counter->cpu, __perf_counter_disable, + counter, 1); + return; + } + + retry: + task_oncpu_function_call(task, __perf_counter_disable, counter); + + spin_lock_irq(&ctx->lock); + /* + * If the counter is still active, we need to retry the cross-call. + */ + if (counter->state == PERF_COUNTER_STATE_ACTIVE) { + spin_unlock_irq(&ctx->lock); + goto retry; + } + + /* + * Since we have the lock this context can't be scheduled + * in, so we can change the state safely. + */ + if (counter->state == PERF_COUNTER_STATE_INACTIVE) { + update_group_times(counter); + counter->state = PERF_COUNTER_STATE_OFF; + } + + spin_unlock_irq(&ctx->lock); +} + +static int +counter_sched_in(struct perf_counter *counter, + struct perf_cpu_context *cpuctx, + struct perf_counter_context *ctx, + int cpu) +{ + if (counter->state <= PERF_COUNTER_STATE_OFF) + return 0; + + counter->state = PERF_COUNTER_STATE_ACTIVE; + counter->oncpu = cpu; /* TODO: put 'cpu' into cpuctx->cpu */ + /* + * The new state must be visible before we turn it on in the hardware: + */ + smp_wmb(); + + if (counter->pmu->enable(counter)) { + counter->state = PERF_COUNTER_STATE_INACTIVE; + counter->oncpu = -1; + return -EAGAIN; + } + + counter->tstamp_running += ctx->time - counter->tstamp_stopped; + + if (!is_software_counter(counter)) + cpuctx->active_oncpu++; + ctx->nr_active++; + + if (counter->attr.exclusive) + cpuctx->exclusive = 1; + + return 0; +} + +static int +group_sched_in(struct perf_counter *group_counter, + struct perf_cpu_context *cpuctx, + struct perf_counter_context *ctx, + int cpu) +{ + struct perf_counter *counter, *partial_group; + int ret; + + if (group_counter->state == PERF_COUNTER_STATE_OFF) + return 0; + + ret = hw_perf_group_sched_in(group_counter, cpuctx, ctx, cpu); + if (ret) + return ret < 0 ? ret : 0; + + if (counter_sched_in(group_counter, cpuctx, ctx, cpu)) + return -EAGAIN; + + /* + * Schedule in siblings as one group (if any): + */ + list_for_each_entry(counter, &group_counter->sibling_list, list_entry) { + if (counter_sched_in(counter, cpuctx, ctx, cpu)) { + partial_group = counter; + goto group_error; + } + } + + return 0; + +group_error: + /* + * Groups can be scheduled in as one unit only, so undo any + * partial group before returning: + */ + list_for_each_entry(counter, &group_counter->sibling_list, list_entry) { + if (counter == partial_group) + break; + counter_sched_out(counter, cpuctx, ctx); + } + counter_sched_out(group_counter, cpuctx, ctx); + + return -EAGAIN; +} + +/* + * Return 1 for a group consisting entirely of software counters, + * 0 if the group contains any hardware counters. + */ +static int is_software_only_group(struct perf_counter *leader) +{ + struct perf_counter *counter; + + if (!is_software_counter(leader)) + return 0; + + list_for_each_entry(counter, &leader->sibling_list, list_entry) + if (!is_software_counter(counter)) + return 0; + + return 1; +} + +/* + * Work out whether we can put this counter group on the CPU now. + */ +static int group_can_go_on(struct perf_counter *counter, + struct perf_cpu_context *cpuctx, + int can_add_hw) +{ + /* + * Groups consisting entirely of software counters can always go on. + */ + if (is_software_only_group(counter)) + return 1; + /* + * If an exclusive group is already on, no other hardware + * counters can go on. + */ + if (cpuctx->exclusive) + return 0; + /* + * If this group is exclusive and there are already + * counters on the CPU, it can't go on. + */ + if (counter->attr.exclusive && cpuctx->active_oncpu) + return 0; + /* + * Otherwise, try to add it if all previous groups were able + * to go on. + */ + return can_add_hw; +} + +static void add_counter_to_ctx(struct perf_counter *counter, + struct perf_counter_context *ctx) +{ + list_add_counter(counter, ctx); + counter->tstamp_enabled = ctx->time; + counter->tstamp_running = ctx->time; + counter->tstamp_stopped = ctx->time; +} + +/* + * Cross CPU call to install and enable a performance counter + * + * Must be called with ctx->mutex held + */ +static void __perf_install_in_context(void *info) +{ + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_counter *counter = info; + struct perf_counter_context *ctx = counter->ctx; + struct perf_counter *leader = counter->group_leader; + int cpu = smp_processor_id(); + int err; + + /* + * If this is a task context, we need to check whether it is + * the current task context of this cpu. If not it has been + * scheduled out before the smp call arrived. + * Or possibly this is the right context but it isn't + * on this cpu because it had no counters. + */ + if (ctx->task && cpuctx->task_ctx != ctx) { + if (cpuctx->task_ctx || ctx->task != current) + return; + cpuctx->task_ctx = ctx; + } + + spin_lock(&ctx->lock); + ctx->is_active = 1; + update_context_time(ctx); + + /* + * Protect the list operation against NMI by disabling the + * counters on a global level. NOP for non NMI based counters. + */ + perf_disable(); + + add_counter_to_ctx(counter, ctx); + + /* + * Don't put the counter on if it is disabled or if + * it is in a group and the group isn't on. + */ + if (counter->state != PERF_COUNTER_STATE_INACTIVE || + (leader != counter && leader->state != PERF_COUNTER_STATE_ACTIVE)) + goto unlock; + + /* + * An exclusive counter can't go on if there are already active + * hardware counters, and no hardware counter can go on if there + * is already an exclusive counter on. + */ + if (!group_can_go_on(counter, cpuctx, 1)) + err = -EEXIST; + else + err = counter_sched_in(counter, cpuctx, ctx, cpu); + + if (err) { + /* + * This counter couldn't go on. If it is in a group + * then we have to pull the whole group off. + * If the counter group is pinned then put it in error state. + */ + if (leader != counter) + group_sched_out(leader, cpuctx, ctx); + if (leader->attr.pinned) { + update_group_times(leader); + leader->state = PERF_COUNTER_STATE_ERROR; + } + } + + if (!err && !ctx->task && cpuctx->max_pertask) + cpuctx->max_pertask--; + + unlock: + perf_enable(); + + spin_unlock(&ctx->lock); +} + +/* + * Attach a performance counter to a context + * + * First we add the counter to the list with the hardware enable bit + * in counter->hw_config cleared. + * + * If the counter is attached to a task which is on a CPU we use a smp + * call to enable it in the task context. The task might have been + * scheduled away, but we check this in the smp call again. + * + * Must be called with ctx->mutex held. + */ +static void +perf_install_in_context(struct perf_counter_context *ctx, + struct perf_counter *counter, + int cpu) +{ + struct task_struct *task = ctx->task; + + if (!task) { + /* + * Per cpu counters are installed via an smp call and + * the install is always sucessful. + */ + smp_call_function_single(cpu, __perf_install_in_context, + counter, 1); + return; + } + +retry: + task_oncpu_function_call(task, __perf_install_in_context, + counter); + + spin_lock_irq(&ctx->lock); + /* + * we need to retry the smp call. + */ + if (ctx->is_active && list_empty(&counter->list_entry)) { + spin_unlock_irq(&ctx->lock); + goto retry; + } + + /* + * The lock prevents that this context is scheduled in so we + * can add the counter safely, if it the call above did not + * succeed. + */ + if (list_empty(&counter->list_entry)) + add_counter_to_ctx(counter, ctx); + spin_unlock_irq(&ctx->lock); +} + +/* + * Put a counter into inactive state and update time fields. + * Enabling the leader of a group effectively enables all + * the group members that aren't explicitly disabled, so we + * have to update their ->tstamp_enabled also. + * Note: this works for group members as well as group leaders + * since the non-leader members' sibling_lists will be empty. + */ +static void __perf_counter_mark_enabled(struct perf_counter *counter, + struct perf_counter_context *ctx) +{ + struct perf_counter *sub; + + counter->state = PERF_COUNTER_STATE_INACTIVE; + counter->tstamp_enabled = ctx->time - counter->total_time_enabled; + list_for_each_entry(sub, &counter->sibling_list, list_entry) + if (sub->state >= PERF_COUNTER_STATE_INACTIVE) + sub->tstamp_enabled = + ctx->time - sub->total_time_enabled; +} + +/* + * Cross CPU call to enable a performance counter + */ +static void __perf_counter_enable(void *info) +{ + struct perf_counter *counter = info; + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_counter_context *ctx = counter->ctx; + struct perf_counter *leader = counter->group_leader; + int err; + + /* + * If this is a per-task counter, need to check whether this + * counter's task is the current task on this cpu. + */ + if (ctx->task && cpuctx->task_ctx != ctx) { + if (cpuctx->task_ctx || ctx->task != current) + return; + cpuctx->task_ctx = ctx; + } + + spin_lock(&ctx->lock); + ctx->is_active = 1; + update_context_time(ctx); + + if (counter->state >= PERF_COUNTER_STATE_INACTIVE) + goto unlock; + __perf_counter_mark_enabled(counter, ctx); + + /* + * If the counter is in a group and isn't the group leader, + * then don't put it on unless the group is on. + */ + if (leader != counter && leader->state != PERF_COUNTER_STATE_ACTIVE) + goto unlock; + + if (!group_can_go_on(counter, cpuctx, 1)) { + err = -EEXIST; + } else { + perf_disable(); + if (counter == leader) + err = group_sched_in(counter, cpuctx, ctx, + smp_processor_id()); + else + err = counter_sched_in(counter, cpuctx, ctx, + smp_processor_id()); + perf_enable(); + } + + if (err) { + /* + * If this counter can't go on and it's part of a + * group, then the whole group has to come off. + */ + if (leader != counter) + group_sched_out(leader, cpuctx, ctx); + if (leader->attr.pinned) { + update_group_times(leader); + leader->state = PERF_COUNTER_STATE_ERROR; + } + } + + unlock: + spin_unlock(&ctx->lock); +} + +/* + * Enable a counter. + * + * If counter->ctx is a cloned context, callers must make sure that + * every task struct that counter->ctx->task could possibly point to + * remains valid. This condition is satisfied when called through + * perf_counter_for_each_child or perf_counter_for_each as described + * for perf_counter_disable. + */ +static void perf_counter_enable(struct perf_counter *counter) +{ + struct perf_counter_context *ctx = counter->ctx; + struct task_struct *task = ctx->task; + + if (!task) { + /* + * Enable the counter on the cpu that it's on + */ + smp_call_function_single(counter->cpu, __perf_counter_enable, + counter, 1); + return; + } + + spin_lock_irq(&ctx->lock); + if (counter->state >= PERF_COUNTER_STATE_INACTIVE) + goto out; + + /* + * If the counter is in error state, clear that first. + * That way, if we see the counter in error state below, we + * know that it has gone back into error state, as distinct + * from the task having been scheduled away before the + * cross-call arrived. + */ + if (counter->state == PERF_COUNTER_STATE_ERROR) + counter->state = PERF_COUNTER_STATE_OFF; + + retry: + spin_unlock_irq(&ctx->lock); + task_oncpu_function_call(task, __perf_counter_enable, counter); + + spin_lock_irq(&ctx->lock); + + /* + * If the context is active and the counter is still off, + * we need to retry the cross-call. + */ + if (ctx->is_active && counter->state == PERF_COUNTER_STATE_OFF) + goto retry; + + /* + * Since we have the lock this context can't be scheduled + * in, so we can change the state safely. + */ + if (counter->state == PERF_COUNTER_STATE_OFF) + __perf_counter_mark_enabled(counter, ctx); + + out: + spin_unlock_irq(&ctx->lock); +} + +static int perf_counter_refresh(struct perf_counter *counter, int refresh) +{ + /* + * not supported on inherited counters + */ + if (counter->attr.inherit) + return -EINVAL; + + atomic_add(refresh, &counter->event_limit); + perf_counter_enable(counter); + + return 0; +} + +void __perf_counter_sched_out(struct perf_counter_context *ctx, + struct perf_cpu_context *cpuctx) +{ + struct perf_counter *counter; + + spin_lock(&ctx->lock); + ctx->is_active = 0; + if (likely(!ctx->nr_counters)) + goto out; + update_context_time(ctx); + + perf_disable(); + if (ctx->nr_active) { + list_for_each_entry(counter, &ctx->counter_list, list_entry) { + if (counter != counter->group_leader) + counter_sched_out(counter, cpuctx, ctx); + else + group_sched_out(counter, cpuctx, ctx); + } + } + perf_enable(); + out: + spin_unlock(&ctx->lock); +} + +/* + * Test whether two contexts are equivalent, i.e. whether they + * have both been cloned from the same version of the same context + * and they both have the same number of enabled counters. + * If the number of enabled counters is the same, then the set + * of enabled counters should be the same, because these are both + * inherited contexts, therefore we can't access individual counters + * in them directly with an fd; we can only enable/disable all + * counters via prctl, or enable/disable all counters in a family + * via ioctl, which will have the same effect on both contexts. + */ +static int context_equiv(struct perf_counter_context *ctx1, + struct perf_counter_context *ctx2) +{ + return ctx1->parent_ctx && ctx1->parent_ctx == ctx2->parent_ctx + && ctx1->parent_gen == ctx2->parent_gen + && !ctx1->pin_count && !ctx2->pin_count; +} + +static void __perf_counter_read(void *counter); + +static void __perf_counter_sync_stat(struct perf_counter *counter, + struct perf_counter *next_counter) +{ + u64 value; + + if (!counter->attr.inherit_stat) + return; + + /* + * Update the counter value, we cannot use perf_counter_read() + * because we're in the middle of a context switch and have IRQs + * disabled, which upsets smp_call_function_single(), however + * we know the counter must be on the current CPU, therefore we + * don't need to use it. + */ + switch (counter->state) { + case PERF_COUNTER_STATE_ACTIVE: + __perf_counter_read(counter); + break; + + case PERF_COUNTER_STATE_INACTIVE: + update_counter_times(counter); + break; + + default: + break; + } + + /* + * In order to keep per-task stats reliable we need to flip the counter + * values when we flip the contexts. + */ + value = atomic64_read(&next_counter->count); + value = atomic64_xchg(&counter->count, value); + atomic64_set(&next_counter->count, value); + + swap(counter->total_time_enabled, next_counter->total_time_enabled); + swap(counter->total_time_running, next_counter->total_time_running); + + /* + * Since we swizzled the values, update the user visible data too. + */ + perf_counter_update_userpage(counter); + perf_counter_update_userpage(next_counter); +} + +#define list_next_entry(pos, member) \ + list_entry(pos->member.next, typeof(*pos), member) + +static void perf_counter_sync_stat(struct perf_counter_context *ctx, + struct perf_counter_context *next_ctx) +{ + struct perf_counter *counter, *next_counter; + + if (!ctx->nr_stat) + return; + + counter = list_first_entry(&ctx->event_list, + struct perf_counter, event_entry); + + next_counter = list_first_entry(&next_ctx->event_list, + struct perf_counter, event_entry); + + while (&counter->event_entry != &ctx->event_list && + &next_counter->event_entry != &next_ctx->event_list) { + + __perf_counter_sync_stat(counter, next_counter); + + counter = list_next_entry(counter, event_entry); + next_counter = list_next_entry(next_counter, event_entry); + } +} + +/* + * Called from scheduler to remove the counters of the current task, + * with interrupts disabled. + * + * We stop each counter and update the counter value in counter->count. + * + * This does not protect us against NMI, but disable() + * sets the disabled bit in the control field of counter _before_ + * accessing the counter control register. If a NMI hits, then it will + * not restart the counter. + */ +void perf_counter_task_sched_out(struct task_struct *task, + struct task_struct *next, int cpu) +{ + struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct perf_counter_context *ctx = task->perf_counter_ctxp; + struct perf_counter_context *next_ctx; + struct perf_counter_context *parent; + struct pt_regs *regs; + int do_switch = 1; + + regs = task_pt_regs(task); + perf_swcounter_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, 1, regs, 0); + + if (likely(!ctx || !cpuctx->task_ctx)) + return; + + update_context_time(ctx); + + rcu_read_lock(); + parent = rcu_dereference(ctx->parent_ctx); + next_ctx = next->perf_counter_ctxp; + if (parent && next_ctx && + rcu_dereference(next_ctx->parent_ctx) == parent) { + /* + * Looks like the two contexts are clones, so we might be + * able to optimize the context switch. We lock both + * contexts and check that they are clones under the + * lock (including re-checking that neither has been + * uncloned in the meantime). It doesn't matter which + * order we take the locks because no other cpu could + * be trying to lock both of these tasks. + */ + spin_lock(&ctx->lock); + spin_lock_nested(&next_ctx->lock, SINGLE_DEPTH_NESTING); + if (context_equiv(ctx, next_ctx)) { + /* + * XXX do we need a memory barrier of sorts + * wrt to rcu_dereference() of perf_counter_ctxp + */ + task->perf_counter_ctxp = next_ctx; + next->perf_counter_ctxp = ctx; + ctx->task = next; + next_ctx->task = task; + do_switch = 0; + + perf_counter_sync_stat(ctx, next_ctx); + } + spin_unlock(&next_ctx->lock); + spin_unlock(&ctx->lock); + } + rcu_read_unlock(); + + if (do_switch) { + __perf_counter_sched_out(ctx, cpuctx); + cpuctx->task_ctx = NULL; + } +} + +/* + * Called with IRQs disabled + */ +static void __perf_counter_task_sched_out(struct perf_counter_context *ctx) +{ + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + + if (!cpuctx->task_ctx) + return; + + if (WARN_ON_ONCE(ctx != cpuctx->task_ctx)) + return; + + __perf_counter_sched_out(ctx, cpuctx); + cpuctx->task_ctx = NULL; +} + +/* + * Called with IRQs disabled + */ +static void perf_counter_cpu_sched_out(struct perf_cpu_context *cpuctx) +{ + __perf_counter_sched_out(&cpuctx->ctx, cpuctx); +} + +static void +__perf_counter_sched_in(struct perf_counter_context *ctx, + struct perf_cpu_context *cpuctx, int cpu) +{ + struct perf_counter *counter; + int can_add_hw = 1; + + spin_lock(&ctx->lock); + ctx->is_active = 1; + if (likely(!ctx->nr_counters)) + goto out; + + ctx->timestamp = perf_clock(); + + perf_disable(); + + /* + * First go through the list and put on any pinned groups + * in order to give them the best chance of going on. + */ + list_for_each_entry(counter, &ctx->counter_list, list_entry) { + if (counter->state <= PERF_COUNTER_STATE_OFF || + !counter->attr.pinned) + continue; + if (counter->cpu != -1 && counter->cpu != cpu) + continue; + + if (counter != counter->group_leader) + counter_sched_in(counter, cpuctx, ctx, cpu); + else { + if (group_can_go_on(counter, cpuctx, 1)) + group_sched_in(counter, cpuctx, ctx, cpu); + } + + /* + * If this pinned group hasn't been scheduled, + * put it in error state. + */ + if (counter->state == PERF_COUNTER_STATE_INACTIVE) { + update_group_times(counter); + counter->state = PERF_COUNTER_STATE_ERROR; + } + } + + list_for_each_entry(counter, &ctx->counter_list, list_entry) { + /* + * Ignore counters in OFF or ERROR state, and + * ignore pinned counters since we did them already. + */ + if (counter->state <= PERF_COUNTER_STATE_OFF || + counter->attr.pinned) + continue; + + /* + * Listen to the 'cpu' scheduling filter constraint + * of counters: + */ + if (counter->cpu != -1 && counter->cpu != cpu) + continue; + + if (counter != counter->group_leader) { + if (counter_sched_in(counter, cpuctx, ctx, cpu)) + can_add_hw = 0; + } else { + if (group_can_go_on(counter, cpuctx, can_add_hw)) { + if (group_sched_in(counter, cpuctx, ctx, cpu)) + can_add_hw = 0; + } + } + } + perf_enable(); + out: + spin_unlock(&ctx->lock); +} + +/* + * Called from scheduler to add the counters of the current task + * with interrupts disabled. + * + * We restore the counter value and then enable it. + * + * This does not protect us against NMI, but enable() + * sets the enabled bit in the control field of counter _before_ + * accessing the counter control register. If a NMI hits, then it will + * keep the counter running. + */ +void perf_counter_task_sched_in(struct task_struct *task, int cpu) +{ + struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct perf_counter_context *ctx = task->perf_counter_ctxp; + + if (likely(!ctx)) + return; + if (cpuctx->task_ctx == ctx) + return; + __perf_counter_sched_in(ctx, cpuctx, cpu); + cpuctx->task_ctx = ctx; +} + +static void perf_counter_cpu_sched_in(struct perf_cpu_context *cpuctx, int cpu) +{ + struct perf_counter_context *ctx = &cpuctx->ctx; + + __perf_counter_sched_in(ctx, cpuctx, cpu); +} + +#define MAX_INTERRUPTS (~0ULL) + +static void perf_log_throttle(struct perf_counter *counter, int enable); + +static void perf_adjust_period(struct perf_counter *counter, u64 events) +{ + struct hw_perf_counter *hwc = &counter->hw; + u64 period, sample_period; + s64 delta; + + events *= hwc->sample_period; + period = div64_u64(events, counter->attr.sample_freq); + + delta = (s64)(period - hwc->sample_period); + delta = (delta + 7) / 8; /* low pass filter */ + + sample_period = hwc->sample_period + delta; + + if (!sample_period) + sample_period = 1; + + hwc->sample_period = sample_period; +} + +static void perf_ctx_adjust_freq(struct perf_counter_context *ctx) +{ + struct perf_counter *counter; + struct hw_perf_counter *hwc; + u64 interrupts, freq; + + spin_lock(&ctx->lock); + list_for_each_entry(counter, &ctx->counter_list, list_entry) { + if (counter->state != PERF_COUNTER_STATE_ACTIVE) + continue; + + hwc = &counter->hw; + + interrupts = hwc->interrupts; + hwc->interrupts = 0; + + /* + * unthrottle counters on the tick + */ + if (interrupts == MAX_INTERRUPTS) { + perf_log_throttle(counter, 1); + counter->pmu->unthrottle(counter); + interrupts = 2*sysctl_perf_counter_sample_rate/HZ; + } + + if (!counter->attr.freq || !counter->attr.sample_freq) + continue; + + /* + * if the specified freq < HZ then we need to skip ticks + */ + if (counter->attr.sample_freq < HZ) { + freq = counter->attr.sample_freq; + + hwc->freq_count += freq; + hwc->freq_interrupts += interrupts; + + if (hwc->freq_count < HZ) + continue; + + interrupts = hwc->freq_interrupts; + hwc->freq_interrupts = 0; + hwc->freq_count -= HZ; + } else + freq = HZ; + + perf_adjust_period(counter, freq * interrupts); + + /* + * In order to avoid being stalled by an (accidental) huge + * sample period, force reset the sample period if we didn't + * get any events in this freq period. + */ + if (!interrupts) { + perf_disable(); + counter->pmu->disable(counter); + atomic64_set(&hwc->period_left, 0); + counter->pmu->enable(counter); + perf_enable(); + } + } + spin_unlock(&ctx->lock); +} + +/* + * Round-robin a context's counters: + */ +static void rotate_ctx(struct perf_counter_context *ctx) +{ + struct perf_counter *counter; + + if (!ctx->nr_counters) + return; + + spin_lock(&ctx->lock); + /* + * Rotate the first entry last (works just fine for group counters too): + */ + perf_disable(); + list_for_each_entry(counter, &ctx->counter_list, list_entry) { + list_move_tail(&counter->list_entry, &ctx->counter_list); + break; + } + perf_enable(); + + spin_unlock(&ctx->lock); +} + +void perf_counter_task_tick(struct task_struct *curr, int cpu) +{ + struct perf_cpu_context *cpuctx; + struct perf_counter_context *ctx; + + if (!atomic_read(&nr_counters)) + return; + + cpuctx = &per_cpu(perf_cpu_context, cpu); + ctx = curr->perf_counter_ctxp; + + perf_ctx_adjust_freq(&cpuctx->ctx); + if (ctx) + perf_ctx_adjust_freq(ctx); + + perf_counter_cpu_sched_out(cpuctx); + if (ctx) + __perf_counter_task_sched_out(ctx); + + rotate_ctx(&cpuctx->ctx); + if (ctx) + rotate_ctx(ctx); + + perf_counter_cpu_sched_in(cpuctx, cpu); + if (ctx) + perf_counter_task_sched_in(curr, cpu); +} + +/* + * Enable all of a task's counters that have been marked enable-on-exec. + * This expects task == current. + */ +static void perf_counter_enable_on_exec(struct task_struct *task) +{ + struct perf_counter_context *ctx; + struct perf_counter *counter; + unsigned long flags; + int enabled = 0; + + local_irq_save(flags); + ctx = task->perf_counter_ctxp; + if (!ctx || !ctx->nr_counters) + goto out; + + __perf_counter_task_sched_out(ctx); + + spin_lock(&ctx->lock); + + list_for_each_entry(counter, &ctx->counter_list, list_entry) { + if (!counter->attr.enable_on_exec) + continue; + counter->attr.enable_on_exec = 0; + if (counter->state >= PERF_COUNTER_STATE_INACTIVE) + continue; + __perf_counter_mark_enabled(counter, ctx); + enabled = 1; + } + + /* + * Unclone this context if we enabled any counter. + */ + if (enabled) + unclone_ctx(ctx); + + spin_unlock(&ctx->lock); + + perf_counter_task_sched_in(task, smp_processor_id()); + out: + local_irq_restore(flags); +} + +/* + * Cross CPU call to read the hardware counter + */ +static void __perf_counter_read(void *info) +{ + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_counter *counter = info; + struct perf_counter_context *ctx = counter->ctx; + unsigned long flags; + + /* + * If this is a task context, we need to check whether it is + * the current task context of this cpu. If not it has been + * scheduled out before the smp call arrived. In that case + * counter->count would have been updated to a recent sample + * when the counter was scheduled out. + */ + if (ctx->task && cpuctx->task_ctx != ctx) + return; + + local_irq_save(flags); + if (ctx->is_active) + update_context_time(ctx); + counter->pmu->read(counter); + update_counter_times(counter); + local_irq_restore(flags); +} + +static u64 perf_counter_read(struct perf_counter *counter) +{ + /* + * If counter is enabled and currently active on a CPU, update the + * value in the counter structure: + */ + if (counter->state == PERF_COUNTER_STATE_ACTIVE) { + smp_call_function_single(counter->oncpu, + __perf_counter_read, counter, 1); + } else if (counter->state == PERF_COUNTER_STATE_INACTIVE) { + update_counter_times(counter); + } + + return atomic64_read(&counter->count); +} + +/* + * Initialize the perf_counter context in a task_struct: + */ +static void +__perf_counter_init_context(struct perf_counter_context *ctx, + struct task_struct *task) +{ + memset(ctx, 0, sizeof(*ctx)); + spin_lock_init(&ctx->lock); + mutex_init(&ctx->mutex); + INIT_LIST_HEAD(&ctx->counter_list); + INIT_LIST_HEAD(&ctx->event_list); + atomic_set(&ctx->refcount, 1); + ctx->task = task; +} + +static struct perf_counter_context *find_get_context(pid_t pid, int cpu) +{ + struct perf_counter_context *ctx; + struct perf_cpu_context *cpuctx; + struct task_struct *task; + unsigned long flags; + int err; + + /* + * If cpu is not a wildcard then this is a percpu counter: + */ + if (cpu != -1) { + /* Must be root to operate on a CPU counter: */ + if (perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN)) + return ERR_PTR(-EACCES); + + if (cpu < 0 || cpu > num_possible_cpus()) + return ERR_PTR(-EINVAL); + + /* + * We could be clever and allow to attach a counter to an + * offline CPU and activate it when the CPU comes up, but + * that's for later. + */ + if (!cpu_isset(cpu, cpu_online_map)) + return ERR_PTR(-ENODEV); + + cpuctx = &per_cpu(perf_cpu_context, cpu); + ctx = &cpuctx->ctx; + get_ctx(ctx); + + return ctx; + } + + rcu_read_lock(); + if (!pid) + task = current; + else + task = find_task_by_vpid(pid); + if (task) + get_task_struct(task); + rcu_read_unlock(); + + if (!task) + return ERR_PTR(-ESRCH); + + /* + * Can't attach counters to a dying task. + */ + err = -ESRCH; + if (task->flags & PF_EXITING) + goto errout; + + /* Reuse ptrace permission checks for now. */ + err = -EACCES; + if (!ptrace_may_access(task, PTRACE_MODE_READ)) + goto errout; + + retry: + ctx = perf_lock_task_context(task, &flags); + if (ctx) { + unclone_ctx(ctx); + spin_unlock_irqrestore(&ctx->lock, flags); + } + + if (!ctx) { + ctx = kmalloc(sizeof(struct perf_counter_context), GFP_KERNEL); + err = -ENOMEM; + if (!ctx) + goto errout; + __perf_counter_init_context(ctx, task); + get_ctx(ctx); + if (cmpxchg(&task->perf_counter_ctxp, NULL, ctx)) { + /* + * We raced with some other task; use + * the context they set. + */ + kfree(ctx); + goto retry; + } + get_task_struct(task); + } + + put_task_struct(task); + return ctx; + + errout: + put_task_struct(task); + return ERR_PTR(err); +} + +static void free_counter_rcu(struct rcu_head *head) +{ + struct perf_counter *counter; + + counter = container_of(head, struct perf_counter, rcu_head); + if (counter->ns) + put_pid_ns(counter->ns); + kfree(counter); +} + +static void perf_pending_sync(struct perf_counter *counter); + +static void free_counter(struct perf_counter *counter) +{ + perf_pending_sync(counter); + + if (!counter->parent) { + atomic_dec(&nr_counters); + if (counter->attr.mmap) + atomic_dec(&nr_mmap_counters); + if (counter->attr.comm) + atomic_dec(&nr_comm_counters); + if (counter->attr.task) + atomic_dec(&nr_task_counters); + } + + if (counter->output) { + fput(counter->output->filp); + counter->output = NULL; + } + + if (counter->destroy) + counter->destroy(counter); + + put_ctx(counter->ctx); + call_rcu(&counter->rcu_head, free_counter_rcu); +} + +/* + * Called when the last reference to the file is gone. + */ +static int perf_release(struct inode *inode, struct file *file) +{ + struct perf_counter *counter = file->private_data; + struct perf_counter_context *ctx = counter->ctx; + + file->private_data = NULL; + + WARN_ON_ONCE(ctx->parent_ctx); + mutex_lock(&ctx->mutex); + perf_counter_remove_from_context(counter); + mutex_unlock(&ctx->mutex); + + mutex_lock(&counter->owner->perf_counter_mutex); + list_del_init(&counter->owner_entry); + mutex_unlock(&counter->owner->perf_counter_mutex); + put_task_struct(counter->owner); + + free_counter(counter); + + return 0; +} + +static int perf_counter_read_size(struct perf_counter *counter) +{ + int entry = sizeof(u64); /* value */ + int size = 0; + int nr = 1; + + if (counter->attr.read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) + size += sizeof(u64); + + if (counter->attr.read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) + size += sizeof(u64); + + if (counter->attr.read_format & PERF_FORMAT_ID) + entry += sizeof(u64); + + if (counter->attr.read_format & PERF_FORMAT_GROUP) { + nr += counter->group_leader->nr_siblings; + size += sizeof(u64); + } + + size += entry * nr; + + return size; +} + +static u64 perf_counter_read_value(struct perf_counter *counter) +{ + struct perf_counter *child; + u64 total = 0; + + total += perf_counter_read(counter); + list_for_each_entry(child, &counter->child_list, child_list) + total += perf_counter_read(child); + + return total; +} + +static int perf_counter_read_entry(struct perf_counter *counter, + u64 read_format, char __user *buf) +{ + int n = 0, count = 0; + u64 values[2]; + + values[n++] = perf_counter_read_value(counter); + if (read_format & PERF_FORMAT_ID) + values[n++] = primary_counter_id(counter); + + count = n * sizeof(u64); + + if (copy_to_user(buf, values, count)) + return -EFAULT; + + return count; +} + +static int perf_counter_read_group(struct perf_counter *counter, + u64 read_format, char __user *buf) +{ + struct perf_counter *leader = counter->group_leader, *sub; + int n = 0, size = 0, err = -EFAULT; + u64 values[3]; + + values[n++] = 1 + leader->nr_siblings; + if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) { + values[n++] = leader->total_time_enabled + + atomic64_read(&leader->child_total_time_enabled); + } + if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) { + values[n++] = leader->total_time_running + + atomic64_read(&leader->child_total_time_running); + } + + size = n * sizeof(u64); + + if (copy_to_user(buf, values, size)) + return -EFAULT; + + err = perf_counter_read_entry(leader, read_format, buf + size); + if (err < 0) + return err; + + size += err; + + list_for_each_entry(sub, &leader->sibling_list, list_entry) { + err = perf_counter_read_entry(sub, read_format, + buf + size); + if (err < 0) + return err; + + size += err; + } + + return size; +} + +static int perf_counter_read_one(struct perf_counter *counter, + u64 read_format, char __user *buf) +{ + u64 values[4]; + int n = 0; + + values[n++] = perf_counter_read_value(counter); + if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) { + values[n++] = counter->total_time_enabled + + atomic64_read(&counter->child_total_time_enabled); + } + if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) { + values[n++] = counter->total_time_running + + atomic64_read(&counter->child_total_time_running); + } + if (read_format & PERF_FORMAT_ID) + values[n++] = primary_counter_id(counter); + + if (copy_to_user(buf, values, n * sizeof(u64))) + return -EFAULT; + + return n * sizeof(u64); +} + +/* + * Read the performance counter - simple non blocking version for now + */ +static ssize_t +perf_read_hw(struct perf_counter *counter, char __user *buf, size_t count) +{ + u64 read_format = counter->attr.read_format; + int ret; + + /* + * Return end-of-file for a read on a counter that is in + * error state (i.e. because it was pinned but it couldn't be + * scheduled on to the CPU at some point). + */ + if (counter->state == PERF_COUNTER_STATE_ERROR) + return 0; + + if (count < perf_counter_read_size(counter)) + return -ENOSPC; + + WARN_ON_ONCE(counter->ctx->parent_ctx); + mutex_lock(&counter->child_mutex); + if (read_format & PERF_FORMAT_GROUP) + ret = perf_counter_read_group(counter, read_format, buf); + else + ret = perf_counter_read_one(counter, read_format, buf); + mutex_unlock(&counter->child_mutex); + + return ret; +} + +static ssize_t +perf_read(struct file *file, char __user *buf, size_t count, loff_t *ppos) +{ + struct perf_counter *counter = file->private_data; + + return perf_read_hw(counter, buf, count); +} + +static unsigned int perf_poll(struct file *file, poll_table *wait) +{ + struct perf_counter *counter = file->private_data; + struct perf_mmap_data *data; + unsigned int events = POLL_HUP; + + rcu_read_lock(); + data = rcu_dereference(counter->data); + if (data) + events = atomic_xchg(&data->poll, 0); + rcu_read_unlock(); + + poll_wait(file, &counter->waitq, wait); + + return events; +} + +static void perf_counter_reset(struct perf_counter *counter) +{ + (void)perf_counter_read(counter); + atomic64_set(&counter->count, 0); + perf_counter_update_userpage(counter); +} + +/* + * Holding the top-level counter's child_mutex means that any + * descendant process that has inherited this counter will block + * in sync_child_counter if it goes to exit, thus satisfying the + * task existence requirements of perf_counter_enable/disable. + */ +static void perf_counter_for_each_child(struct perf_counter *counter, + void (*func)(struct perf_counter *)) +{ + struct perf_counter *child; + + WARN_ON_ONCE(counter->ctx->parent_ctx); + mutex_lock(&counter->child_mutex); + func(counter); + list_for_each_entry(child, &counter->child_list, child_list) + func(child); + mutex_unlock(&counter->child_mutex); +} + +static void perf_counter_for_each(struct perf_counter *counter, + void (*func)(struct perf_counter *)) +{ + struct perf_counter_context *ctx = counter->ctx; + struct perf_counter *sibling; + + WARN_ON_ONCE(ctx->parent_ctx); + mutex_lock(&ctx->mutex); + counter = counter->group_leader; + + perf_counter_for_each_child(counter, func); + func(counter); + list_for_each_entry(sibling, &counter->sibling_list, list_entry) + perf_counter_for_each_child(counter, func); + mutex_unlock(&ctx->mutex); +} + +static int perf_counter_period(struct perf_counter *counter, u64 __user *arg) +{ + struct perf_counter_context *ctx = counter->ctx; + unsigned long size; + int ret = 0; + u64 value; + + if (!counter->attr.sample_period) + return -EINVAL; + + size = copy_from_user(&value, arg, sizeof(value)); + if (size != sizeof(value)) + return -EFAULT; + + if (!value) + return -EINVAL; + + spin_lock_irq(&ctx->lock); + if (counter->attr.freq) { + if (value > sysctl_perf_counter_sample_rate) { + ret = -EINVAL; + goto unlock; + } + + counter->attr.sample_freq = value; + } else { + counter->attr.sample_period = value; + counter->hw.sample_period = value; + } +unlock: + spin_unlock_irq(&ctx->lock); + + return ret; +} + +int perf_counter_set_output(struct perf_counter *counter, int output_fd); + +static long perf_ioctl(struct file *file, unsigned int cmd, unsigned long arg) +{ + struct perf_counter *counter = file->private_data; + void (*func)(struct perf_counter *); + u32 flags = arg; + + switch (cmd) { + case PERF_COUNTER_IOC_ENABLE: + func = perf_counter_enable; + break; + case PERF_COUNTER_IOC_DISABLE: + func = perf_counter_disable; + break; + case PERF_COUNTER_IOC_RESET: + func = perf_counter_reset; + break; + + case PERF_COUNTER_IOC_REFRESH: + return perf_counter_refresh(counter, arg); + + case PERF_COUNTER_IOC_PERIOD: + return perf_counter_period(counter, (u64 __user *)arg); + + case PERF_COUNTER_IOC_SET_OUTPUT: + return perf_counter_set_output(counter, arg); + + default: + return -ENOTTY; + } + + if (flags & PERF_IOC_FLAG_GROUP) + perf_counter_for_each(counter, func); + else + perf_counter_for_each_child(counter, func); + + return 0; +} + +int perf_counter_task_enable(void) +{ + struct perf_counter *counter; + + mutex_lock(¤t->perf_counter_mutex); + list_for_each_entry(counter, ¤t->perf_counter_list, owner_entry) + perf_counter_for_each_child(counter, perf_counter_enable); + mutex_unlock(¤t->perf_counter_mutex); + + return 0; +} + +int perf_counter_task_disable(void) +{ + struct perf_counter *counter; + + mutex_lock(¤t->perf_counter_mutex); + list_for_each_entry(counter, ¤t->perf_counter_list, owner_entry) + perf_counter_for_each_child(counter, perf_counter_disable); + mutex_unlock(¤t->perf_counter_mutex); + + return 0; +} + +#ifndef PERF_COUNTER_INDEX_OFFSET +# define PERF_COUNTER_INDEX_OFFSET 0 +#endif + +static int perf_counter_index(struct perf_counter *counter) +{ + if (counter->state != PERF_COUNTER_STATE_ACTIVE) + return 0; + + return counter->hw.idx + 1 - PERF_COUNTER_INDEX_OFFSET; +} + +/* + * Callers need to ensure there can be no nesting of this function, otherwise + * the seqlock logic goes bad. We can not serialize this because the arch + * code calls this from NMI context. + */ +void perf_counter_update_userpage(struct perf_counter *counter) +{ + struct perf_counter_mmap_page *userpg; + struct perf_mmap_data *data; + + rcu_read_lock(); + data = rcu_dereference(counter->data); + if (!data) + goto unlock; + + userpg = data->user_page; + + /* + * Disable preemption so as to not let the corresponding user-space + * spin too long if we get preempted. + */ + preempt_disable(); + ++userpg->lock; + barrier(); + userpg->index = perf_counter_index(counter); + userpg->offset = atomic64_read(&counter->count); + if (counter->state == PERF_COUNTER_STATE_ACTIVE) + userpg->offset -= atomic64_read(&counter->hw.prev_count); + + userpg->time_enabled = counter->total_time_enabled + + atomic64_read(&counter->child_total_time_enabled); + + userpg->time_running = counter->total_time_running + + atomic64_read(&counter->child_total_time_running); + + barrier(); + ++userpg->lock; + preempt_enable(); +unlock: + rcu_read_unlock(); +} + +static int perf_mmap_fault(struct vm_area_struct *vma, struct vm_fault *vmf) +{ + struct perf_counter *counter = vma->vm_file->private_data; + struct perf_mmap_data *data; + int ret = VM_FAULT_SIGBUS; + + if (vmf->flags & FAULT_FLAG_MKWRITE) { + if (vmf->pgoff == 0) + ret = 0; + return ret; + } + + rcu_read_lock(); + data = rcu_dereference(counter->data); + if (!data) + goto unlock; + + if (vmf->pgoff == 0) { + vmf->page = virt_to_page(data->user_page); + } else { + int nr = vmf->pgoff - 1; + + if ((unsigned)nr > data->nr_pages) + goto unlock; + + if (vmf->flags & FAULT_FLAG_WRITE) + goto unlock; + + vmf->page = virt_to_page(data->data_pages[nr]); + } + + get_page(vmf->page); + vmf->page->mapping = vma->vm_file->f_mapping; + vmf->page->index = vmf->pgoff; + + ret = 0; +unlock: + rcu_read_unlock(); + + return ret; +} + +static int perf_mmap_data_alloc(struct perf_counter *counter, int nr_pages) +{ + struct perf_mmap_data *data; + unsigned long size; + int i; + + WARN_ON(atomic_read(&counter->mmap_count)); + + size = sizeof(struct perf_mmap_data); + size += nr_pages * sizeof(void *); + + data = kzalloc(size, GFP_KERNEL); + if (!data) + goto fail; + + data->user_page = (void *)get_zeroed_page(GFP_KERNEL); + if (!data->user_page) + goto fail_user_page; + + for (i = 0; i < nr_pages; i++) { + data->data_pages[i] = (void *)get_zeroed_page(GFP_KERNEL); + if (!data->data_pages[i]) + goto fail_data_pages; + } + + data->nr_pages = nr_pages; + atomic_set(&data->lock, -1); + + rcu_assign_pointer(counter->data, data); + + return 0; + +fail_data_pages: + for (i--; i >= 0; i--) + free_page((unsigned long)data->data_pages[i]); + + free_page((unsigned long)data->user_page); + +fail_user_page: + kfree(data); + +fail: + return -ENOMEM; +} + +static void perf_mmap_free_page(unsigned long addr) +{ + struct page *page = virt_to_page((void *)addr); + + page->mapping = NULL; + __free_page(page); +} + +static void __perf_mmap_data_free(struct rcu_head *rcu_head) +{ + struct perf_mmap_data *data; + int i; + + data = container_of(rcu_head, struct perf_mmap_data, rcu_head); + + perf_mmap_free_page((unsigned long)data->user_page); + for (i = 0; i < data->nr_pages; i++) + perf_mmap_free_page((unsigned long)data->data_pages[i]); + + kfree(data); +} + +static void perf_mmap_data_free(struct perf_counter *counter) +{ + struct perf_mmap_data *data = counter->data; + + WARN_ON(atomic_read(&counter->mmap_count)); + + rcu_assign_pointer(counter->data, NULL); + call_rcu(&data->rcu_head, __perf_mmap_data_free); +} + +static void perf_mmap_open(struct vm_area_struct *vma) +{ + struct perf_counter *counter = vma->vm_file->private_data; + + atomic_inc(&counter->mmap_count); +} + +static void perf_mmap_close(struct vm_area_struct *vma) +{ + struct perf_counter *counter = vma->vm_file->private_data; + + WARN_ON_ONCE(counter->ctx->parent_ctx); + if (atomic_dec_and_mutex_lock(&counter->mmap_count, &counter->mmap_mutex)) { + struct user_struct *user = current_user(); + + atomic_long_sub(counter->data->nr_pages + 1, &user->locked_vm); + vma->vm_mm->locked_vm -= counter->data->nr_locked; + perf_mmap_data_free(counter); + mutex_unlock(&counter->mmap_mutex); + } +} + +static struct vm_operations_struct perf_mmap_vmops = { + .open = perf_mmap_open, + .close = perf_mmap_close, + .fault = perf_mmap_fault, + .page_mkwrite = perf_mmap_fault, +}; + +static int perf_mmap(struct file *file, struct vm_area_struct *vma) +{ + struct perf_counter *counter = file->private_data; + unsigned long user_locked, user_lock_limit; + struct user_struct *user = current_user(); + unsigned long locked, lock_limit; + unsigned long vma_size; + unsigned long nr_pages; + long user_extra, extra; + int ret = 0; + + if (!(vma->vm_flags & VM_SHARED)) + return -EINVAL; + + vma_size = vma->vm_end - vma->vm_start; + nr_pages = (vma_size / PAGE_SIZE) - 1; + + /* + * If we have data pages ensure they're a power-of-two number, so we + * can do bitmasks instead of modulo. + */ + if (nr_pages != 0 && !is_power_of_2(nr_pages)) + return -EINVAL; + + if (vma_size != PAGE_SIZE * (1 + nr_pages)) + return -EINVAL; + + if (vma->vm_pgoff != 0) + return -EINVAL; + + WARN_ON_ONCE(counter->ctx->parent_ctx); + mutex_lock(&counter->mmap_mutex); + if (counter->output) { + ret = -EINVAL; + goto unlock; + } + + if (atomic_inc_not_zero(&counter->mmap_count)) { + if (nr_pages != counter->data->nr_pages) + ret = -EINVAL; + goto unlock; + } + + user_extra = nr_pages + 1; + user_lock_limit = sysctl_perf_counter_mlock >> (PAGE_SHIFT - 10); + + /* + * Increase the limit linearly with more CPUs: + */ + user_lock_limit *= num_online_cpus(); + + user_locked = atomic_long_read(&user->locked_vm) + user_extra; + + extra = 0; + if (user_locked > user_lock_limit) + extra = user_locked - user_lock_limit; + + lock_limit = current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur; + lock_limit >>= PAGE_SHIFT; + locked = vma->vm_mm->locked_vm + extra; + + if ((locked > lock_limit) && !capable(CAP_IPC_LOCK)) { + ret = -EPERM; + goto unlock; + } + + WARN_ON(counter->data); + ret = perf_mmap_data_alloc(counter, nr_pages); + if (ret) + goto unlock; + + atomic_set(&counter->mmap_count, 1); + atomic_long_add(user_extra, &user->locked_vm); + vma->vm_mm->locked_vm += extra; + counter->data->nr_locked = extra; + if (vma->vm_flags & VM_WRITE) + counter->data->writable = 1; + +unlock: + mutex_unlock(&counter->mmap_mutex); + + vma->vm_flags |= VM_RESERVED; + vma->vm_ops = &perf_mmap_vmops; + + return ret; +} + +static int perf_fasync(int fd, struct file *filp, int on) +{ + struct inode *inode = filp->f_path.dentry->d_inode; + struct perf_counter *counter = filp->private_data; + int retval; + + mutex_lock(&inode->i_mutex); + retval = fasync_helper(fd, filp, on, &counter->fasync); + mutex_unlock(&inode->i_mutex); + + if (retval < 0) + return retval; + + return 0; +} + +static const struct file_operations perf_fops = { + .release = perf_release, + .read = perf_read, + .poll = perf_poll, + .unlocked_ioctl = perf_ioctl, + .compat_ioctl = perf_ioctl, + .mmap = perf_mmap, + .fasync = perf_fasync, +}; + +/* + * Perf counter wakeup + * + * If there's data, ensure we set the poll() state and publish everything + * to user-space before waking everybody up. + */ + +void perf_counter_wakeup(struct perf_counter *counter) +{ + wake_up_all(&counter->waitq); + + if (counter->pending_kill) { + kill_fasync(&counter->fasync, SIGIO, counter->pending_kill); + counter->pending_kill = 0; + } +} + +/* + * Pending wakeups + * + * Handle the case where we need to wakeup up from NMI (or rq->lock) context. + * + * The NMI bit means we cannot possibly take locks. Therefore, maintain a + * single linked list and use cmpxchg() to add entries lockless. + */ + +static void perf_pending_counter(struct perf_pending_entry *entry) +{ + struct perf_counter *counter = container_of(entry, + struct perf_counter, pending); + + if (counter->pending_disable) { + counter->pending_disable = 0; + __perf_counter_disable(counter); + } + + if (counter->pending_wakeup) { + counter->pending_wakeup = 0; + perf_counter_wakeup(counter); + } +} + +#define PENDING_TAIL ((struct perf_pending_entry *)-1UL) + +static DEFINE_PER_CPU(struct perf_pending_entry *, perf_pending_head) = { + PENDING_TAIL, +}; + +static void perf_pending_queue(struct perf_pending_entry *entry, + void (*func)(struct perf_pending_entry *)) +{ + struct perf_pending_entry **head; + + if (cmpxchg(&entry->next, NULL, PENDING_TAIL) != NULL) + return; + + entry->func = func; + + head = &get_cpu_var(perf_pending_head); + + do { + entry->next = *head; + } while (cmpxchg(head, entry->next, entry) != entry->next); + + set_perf_counter_pending(); + + put_cpu_var(perf_pending_head); +} + +static int __perf_pending_run(void) +{ + struct perf_pending_entry *list; + int nr = 0; + + list = xchg(&__get_cpu_var(perf_pending_head), PENDING_TAIL); + while (list != PENDING_TAIL) { + void (*func)(struct perf_pending_entry *); + struct perf_pending_entry *entry = list; + + list = list->next; + + func = entry->func; + entry->next = NULL; + /* + * Ensure we observe the unqueue before we issue the wakeup, + * so that we won't be waiting forever. + * -- see perf_not_pending(). + */ + smp_wmb(); + + func(entry); + nr++; + } + + return nr; +} + +static inline int perf_not_pending(struct perf_counter *counter) +{ + /* + * If we flush on whatever cpu we run, there is a chance we don't + * need to wait. + */ + get_cpu(); + __perf_pending_run(); + put_cpu(); + + /* + * Ensure we see the proper queue state before going to sleep + * so that we do not miss the wakeup. -- see perf_pending_handle() + */ + smp_rmb(); + return counter->pending.next == NULL; +} + +static void perf_pending_sync(struct perf_counter *counter) +{ + wait_event(counter->waitq, perf_not_pending(counter)); +} + +void perf_counter_do_pending(void) +{ + __perf_pending_run(); +} + +/* + * Callchain support -- arch specific + */ + +__weak struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) +{ + return NULL; +} + +/* + * Output + */ + +struct perf_output_handle { + struct perf_counter *counter; + struct perf_mmap_data *data; + unsigned long head; + unsigned long offset; + int nmi; + int sample; + int locked; + unsigned long flags; +}; + +static bool perf_output_space(struct perf_mmap_data *data, + unsigned int offset, unsigned int head) +{ + unsigned long tail; + unsigned long mask; + + if (!data->writable) + return true; + + mask = (data->nr_pages << PAGE_SHIFT) - 1; + /* + * Userspace could choose to issue a mb() before updating the tail + * pointer. So that all reads will be completed before the write is + * issued. + */ + tail = ACCESS_ONCE(data->user_page->data_tail); + smp_rmb(); + + offset = (offset - tail) & mask; + head = (head - tail) & mask; + + if ((int)(head - offset) < 0) + return false; + + return true; +} + +static void perf_output_wakeup(struct perf_output_handle *handle) +{ + atomic_set(&handle->data->poll, POLL_IN); + + if (handle->nmi) { + handle->counter->pending_wakeup = 1; + perf_pending_queue(&handle->counter->pending, + perf_pending_counter); + } else + perf_counter_wakeup(handle->counter); +} + +/* + * Curious locking construct. + * + * We need to ensure a later event doesn't publish a head when a former + * event isn't done writing. However since we need to deal with NMIs we + * cannot fully serialize things. + * + * What we do is serialize between CPUs so we only have to deal with NMI + * nesting on a single CPU. + * + * We only publish the head (and generate a wakeup) when the outer-most + * event completes. + */ +static void perf_output_lock(struct perf_output_handle *handle) +{ + struct perf_mmap_data *data = handle->data; + int cpu; + + handle->locked = 0; + + local_irq_save(handle->flags); + cpu = smp_processor_id(); + + if (in_nmi() && atomic_read(&data->lock) == cpu) + return; + + while (atomic_cmpxchg(&data->lock, -1, cpu) != -1) + cpu_relax(); + + handle->locked = 1; +} + +static void perf_output_unlock(struct perf_output_handle *handle) +{ + struct perf_mmap_data *data = handle->data; + unsigned long head; + int cpu; + + data->done_head = data->head; + + if (!handle->locked) + goto out; + +again: + /* + * The xchg implies a full barrier that ensures all writes are done + * before we publish the new head, matched by a rmb() in userspace when + * reading this position. + */ + while ((head = atomic_long_xchg(&data->done_head, 0))) + data->user_page->data_head = head; + + /* + * NMI can happen here, which means we can miss a done_head update. + */ + + cpu = atomic_xchg(&data->lock, -1); + WARN_ON_ONCE(cpu != smp_processor_id()); + + /* + * Therefore we have to validate we did not indeed do so. + */ + if (unlikely(atomic_long_read(&data->done_head))) { + /* + * Since we had it locked, we can lock it again. + */ + while (atomic_cmpxchg(&data->lock, -1, cpu) != -1) + cpu_relax(); + + goto again; + } + + if (atomic_xchg(&data->wakeup, 0)) + perf_output_wakeup(handle); +out: + local_irq_restore(handle->flags); +} + +static void perf_output_copy(struct perf_output_handle *handle, + const void *buf, unsigned int len) +{ + unsigned int pages_mask; + unsigned int offset; + unsigned int size; + void **pages; + + offset = handle->offset; + pages_mask = handle->data->nr_pages - 1; + pages = handle->data->data_pages; + + do { + unsigned int page_offset; + int nr; + + nr = (offset >> PAGE_SHIFT) & pages_mask; + page_offset = offset & (PAGE_SIZE - 1); + size = min_t(unsigned int, PAGE_SIZE - page_offset, len); + + memcpy(pages[nr] + page_offset, buf, size); + + len -= size; + buf += size; + offset += size; + } while (len); + + handle->offset = offset; + + /* + * Check we didn't copy past our reservation window, taking the + * possible unsigned int wrap into account. + */ + WARN_ON_ONCE(((long)(handle->head - handle->offset)) < 0); +} + +#define perf_output_put(handle, x) \ + perf_output_copy((handle), &(x), sizeof(x)) + +static int perf_output_begin(struct perf_output_handle *handle, + struct perf_counter *counter, unsigned int size, + int nmi, int sample) +{ + struct perf_counter *output_counter; + struct perf_mmap_data *data; + unsigned int offset, head; + int have_lost; + struct { + struct perf_event_header header; + u64 id; + u64 lost; + } lost_event; + + rcu_read_lock(); + /* + * For inherited counters we send all the output towards the parent. + */ + if (counter->parent) + counter = counter->parent; + + output_counter = rcu_dereference(counter->output); + if (output_counter) + counter = output_counter; + + data = rcu_dereference(counter->data); + if (!data) + goto out; + + handle->data = data; + handle->counter = counter; + handle->nmi = nmi; + handle->sample = sample; + + if (!data->nr_pages) + goto fail; + + have_lost = atomic_read(&data->lost); + if (have_lost) + size += sizeof(lost_event); + + perf_output_lock(handle); + + do { + offset = head = atomic_long_read(&data->head); + head += size; + if (unlikely(!perf_output_space(data, offset, head))) + goto fail; + } while (atomic_long_cmpxchg(&data->head, offset, head) != offset); + + handle->offset = offset; + handle->head = head; + + if ((offset >> PAGE_SHIFT) != (head >> PAGE_SHIFT)) + atomic_set(&data->wakeup, 1); + + if (have_lost) { + lost_event.header.type = PERF_EVENT_LOST; + lost_event.header.misc = 0; + lost_event.header.size = sizeof(lost_event); + lost_event.id = counter->id; + lost_event.lost = atomic_xchg(&data->lost, 0); + + perf_output_put(handle, lost_event); + } + + return 0; + +fail: + atomic_inc(&data->lost); + perf_output_unlock(handle); +out: + rcu_read_unlock(); + + return -ENOSPC; +} + +static void perf_output_end(struct perf_output_handle *handle) +{ + struct perf_counter *counter = handle->counter; + struct perf_mmap_data *data = handle->data; + + int wakeup_events = counter->attr.wakeup_events; + + if (handle->sample && wakeup_events) { + int events = atomic_inc_return(&data->events); + if (events >= wakeup_events) { + atomic_sub(wakeup_events, &data->events); + atomic_set(&data->wakeup, 1); + } + } + + perf_output_unlock(handle); + rcu_read_unlock(); +} + +static u32 perf_counter_pid(struct perf_counter *counter, struct task_struct *p) +{ + /* + * only top level counters have the pid namespace they were created in + */ + if (counter->parent) + counter = counter->parent; + + return task_tgid_nr_ns(p, counter->ns); +} + +static u32 perf_counter_tid(struct perf_counter *counter, struct task_struct *p) +{ + /* + * only top level counters have the pid namespace they were created in + */ + if (counter->parent) + counter = counter->parent; + + return task_pid_nr_ns(p, counter->ns); +} + +static void perf_output_read_one(struct perf_output_handle *handle, + struct perf_counter *counter) +{ + u64 read_format = counter->attr.read_format; + u64 values[4]; + int n = 0; + + values[n++] = atomic64_read(&counter->count); + if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) { + values[n++] = counter->total_time_enabled + + atomic64_read(&counter->child_total_time_enabled); + } + if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) { + values[n++] = counter->total_time_running + + atomic64_read(&counter->child_total_time_running); + } + if (read_format & PERF_FORMAT_ID) + values[n++] = primary_counter_id(counter); + + perf_output_copy(handle, values, n * sizeof(u64)); +} + +/* + * XXX PERF_FORMAT_GROUP vs inherited counters seems difficult. + */ +static void perf_output_read_group(struct perf_output_handle *handle, + struct perf_counter *counter) +{ + struct perf_counter *leader = counter->group_leader, *sub; + u64 read_format = counter->attr.read_format; + u64 values[5]; + int n = 0; + + values[n++] = 1 + leader->nr_siblings; + + if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) + values[n++] = leader->total_time_enabled; + + if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) + values[n++] = leader->total_time_running; + + if (leader != counter) + leader->pmu->read(leader); + + values[n++] = atomic64_read(&leader->count); + if (read_format & PERF_FORMAT_ID) + values[n++] = primary_counter_id(leader); + + perf_output_copy(handle, values, n * sizeof(u64)); + + list_for_each_entry(sub, &leader->sibling_list, list_entry) { + n = 0; + + if (sub != counter) + sub->pmu->read(sub); + + values[n++] = atomic64_read(&sub->count); + if (read_format & PERF_FORMAT_ID) + values[n++] = primary_counter_id(sub); + + perf_output_copy(handle, values, n * sizeof(u64)); + } +} + +static void perf_output_read(struct perf_output_handle *handle, + struct perf_counter *counter) +{ + if (counter->attr.read_format & PERF_FORMAT_GROUP) + perf_output_read_group(handle, counter); + else + perf_output_read_one(handle, counter); +} + +void perf_counter_output(struct perf_counter *counter, int nmi, + struct perf_sample_data *data) +{ + int ret; + u64 sample_type = counter->attr.sample_type; + struct perf_output_handle handle; + struct perf_event_header header; + u64 ip; + struct { + u32 pid, tid; + } tid_entry; + struct perf_callchain_entry *callchain = NULL; + int callchain_size = 0; + u64 time; + struct { + u32 cpu, reserved; + } cpu_entry; + + header.type = PERF_EVENT_SAMPLE; + header.size = sizeof(header); + + header.misc = 0; + header.misc |= perf_misc_flags(data->regs); + + if (sample_type & PERF_SAMPLE_IP) { + ip = perf_instruction_pointer(data->regs); + header.size += sizeof(ip); + } + + if (sample_type & PERF_SAMPLE_TID) { + /* namespace issues */ + tid_entry.pid = perf_counter_pid(counter, current); + tid_entry.tid = perf_counter_tid(counter, current); + + header.size += sizeof(tid_entry); + } + + if (sample_type & PERF_SAMPLE_TIME) { + /* + * Maybe do better on x86 and provide cpu_clock_nmi() + */ + time = sched_clock(); + + header.size += sizeof(u64); + } + + if (sample_type & PERF_SAMPLE_ADDR) + header.size += sizeof(u64); + + if (sample_type & PERF_SAMPLE_ID) + header.size += sizeof(u64); + + if (sample_type & PERF_SAMPLE_STREAM_ID) + header.size += sizeof(u64); + + if (sample_type & PERF_SAMPLE_CPU) { + header.size += sizeof(cpu_entry); + + cpu_entry.cpu = raw_smp_processor_id(); + cpu_entry.reserved = 0; + } + + if (sample_type & PERF_SAMPLE_PERIOD) + header.size += sizeof(u64); + + if (sample_type & PERF_SAMPLE_READ) + header.size += perf_counter_read_size(counter); + + if (sample_type & PERF_SAMPLE_CALLCHAIN) { + callchain = perf_callchain(data->regs); + + if (callchain) { + callchain_size = (1 + callchain->nr) * sizeof(u64); + header.size += callchain_size; + } else + header.size += sizeof(u64); + } + + if (sample_type & PERF_SAMPLE_RAW) { + int size = sizeof(u32); + + if (data->raw) + size += data->raw->size; + else + size += sizeof(u32); + + WARN_ON_ONCE(size & (sizeof(u64)-1)); + header.size += size; + } + + ret = perf_output_begin(&handle, counter, header.size, nmi, 1); + if (ret) + return; + + perf_output_put(&handle, header); + + if (sample_type & PERF_SAMPLE_IP) + perf_output_put(&handle, ip); + + if (sample_type & PERF_SAMPLE_TID) + perf_output_put(&handle, tid_entry); + + if (sample_type & PERF_SAMPLE_TIME) + perf_output_put(&handle, time); + + if (sample_type & PERF_SAMPLE_ADDR) + perf_output_put(&handle, data->addr); + + if (sample_type & PERF_SAMPLE_ID) { + u64 id = primary_counter_id(counter); + + perf_output_put(&handle, id); + } + + if (sample_type & PERF_SAMPLE_STREAM_ID) + perf_output_put(&handle, counter->id); + + if (sample_type & PERF_SAMPLE_CPU) + perf_output_put(&handle, cpu_entry); + + if (sample_type & PERF_SAMPLE_PERIOD) + perf_output_put(&handle, data->period); + + if (sample_type & PERF_SAMPLE_READ) + perf_output_read(&handle, counter); + + if (sample_type & PERF_SAMPLE_CALLCHAIN) { + if (callchain) + perf_output_copy(&handle, callchain, callchain_size); + else { + u64 nr = 0; + perf_output_put(&handle, nr); + } + } + + if (sample_type & PERF_SAMPLE_RAW) { + if (data->raw) { + perf_output_put(&handle, data->raw->size); + perf_output_copy(&handle, data->raw->data, data->raw->size); + } else { + struct { + u32 size; + u32 data; + } raw = { + .size = sizeof(u32), + .data = 0, + }; + perf_output_put(&handle, raw); + } + } + + perf_output_end(&handle); +} + +/* + * read event + */ + +struct perf_read_event { + struct perf_event_header header; + + u32 pid; + u32 tid; +}; + +static void +perf_counter_read_event(struct perf_counter *counter, + struct task_struct *task) +{ + struct perf_output_handle handle; + struct perf_read_event event = { + .header = { + .type = PERF_EVENT_READ, + .misc = 0, + .size = sizeof(event) + perf_counter_read_size(counter), + }, + .pid = perf_counter_pid(counter, task), + .tid = perf_counter_tid(counter, task), + }; + int ret; + + ret = perf_output_begin(&handle, counter, event.header.size, 0, 0); + if (ret) + return; + + perf_output_put(&handle, event); + perf_output_read(&handle, counter); + + perf_output_end(&handle); +} + +/* + * task tracking -- fork/exit + * + * enabled by: attr.comm | attr.mmap | attr.task + */ + +struct perf_task_event { + struct task_struct *task; + struct perf_counter_context *task_ctx; + + struct { + struct perf_event_header header; + + u32 pid; + u32 ppid; + u32 tid; + u32 ptid; + } event; +}; + +static void perf_counter_task_output(struct perf_counter *counter, + struct perf_task_event *task_event) +{ + struct perf_output_handle handle; + int size = task_event->event.header.size; + struct task_struct *task = task_event->task; + int ret = perf_output_begin(&handle, counter, size, 0, 0); + + if (ret) + return; + + task_event->event.pid = perf_counter_pid(counter, task); + task_event->event.ppid = perf_counter_pid(counter, current); + + task_event->event.tid = perf_counter_tid(counter, task); + task_event->event.ptid = perf_counter_tid(counter, current); + + perf_output_put(&handle, task_event->event); + perf_output_end(&handle); +} + +static int perf_counter_task_match(struct perf_counter *counter) +{ + if (counter->attr.comm || counter->attr.mmap || counter->attr.task) + return 1; + + return 0; +} + +static void perf_counter_task_ctx(struct perf_counter_context *ctx, + struct perf_task_event *task_event) +{ + struct perf_counter *counter; + + if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list)) + return; + + rcu_read_lock(); + list_for_each_entry_rcu(counter, &ctx->event_list, event_entry) { + if (perf_counter_task_match(counter)) + perf_counter_task_output(counter, task_event); + } + rcu_read_unlock(); +} + +static void perf_counter_task_event(struct perf_task_event *task_event) +{ + struct perf_cpu_context *cpuctx; + struct perf_counter_context *ctx = task_event->task_ctx; + + cpuctx = &get_cpu_var(perf_cpu_context); + perf_counter_task_ctx(&cpuctx->ctx, task_event); + put_cpu_var(perf_cpu_context); + + rcu_read_lock(); + if (!ctx) + ctx = rcu_dereference(task_event->task->perf_counter_ctxp); + if (ctx) + perf_counter_task_ctx(ctx, task_event); + rcu_read_unlock(); +} + +static void perf_counter_task(struct task_struct *task, + struct perf_counter_context *task_ctx, + int new) +{ + struct perf_task_event task_event; + + if (!atomic_read(&nr_comm_counters) && + !atomic_read(&nr_mmap_counters) && + !atomic_read(&nr_task_counters)) + return; + + task_event = (struct perf_task_event){ + .task = task, + .task_ctx = task_ctx, + .event = { + .header = { + .type = new ? PERF_EVENT_FORK : PERF_EVENT_EXIT, + .misc = 0, + .size = sizeof(task_event.event), + }, + /* .pid */ + /* .ppid */ + /* .tid */ + /* .ptid */ + }, + }; + + perf_counter_task_event(&task_event); +} + +void perf_counter_fork(struct task_struct *task) +{ + perf_counter_task(task, NULL, 1); +} + +/* + * comm tracking + */ + +struct perf_comm_event { + struct task_struct *task; + char *comm; + int comm_size; + + struct { + struct perf_event_header header; + + u32 pid; + u32 tid; + } event; +}; + +static void perf_counter_comm_output(struct perf_counter *counter, + struct perf_comm_event *comm_event) +{ + struct perf_output_handle handle; + int size = comm_event->event.header.size; + int ret = perf_output_begin(&handle, counter, size, 0, 0); + + if (ret) + return; + + comm_event->event.pid = perf_counter_pid(counter, comm_event->task); + comm_event->event.tid = perf_counter_tid(counter, comm_event->task); + + perf_output_put(&handle, comm_event->event); + perf_output_copy(&handle, comm_event->comm, + comm_event->comm_size); + perf_output_end(&handle); +} + +static int perf_counter_comm_match(struct perf_counter *counter) +{ + if (counter->attr.comm) + return 1; + + return 0; +} + +static void perf_counter_comm_ctx(struct perf_counter_context *ctx, + struct perf_comm_event *comm_event) +{ + struct perf_counter *counter; + + if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list)) + return; + + rcu_read_lock(); + list_for_each_entry_rcu(counter, &ctx->event_list, event_entry) { + if (perf_counter_comm_match(counter)) + perf_counter_comm_output(counter, comm_event); + } + rcu_read_unlock(); +} + +static void perf_counter_comm_event(struct perf_comm_event *comm_event) +{ + struct perf_cpu_context *cpuctx; + struct perf_counter_context *ctx; + unsigned int size; + char comm[TASK_COMM_LEN]; + + memset(comm, 0, sizeof(comm)); + strncpy(comm, comm_event->task->comm, sizeof(comm)); + size = ALIGN(strlen(comm)+1, sizeof(u64)); + + comm_event->comm = comm; + comm_event->comm_size = size; + + comm_event->event.header.size = sizeof(comm_event->event) + size; + + cpuctx = &get_cpu_var(perf_cpu_context); + perf_counter_comm_ctx(&cpuctx->ctx, comm_event); + put_cpu_var(perf_cpu_context); + + rcu_read_lock(); + /* + * doesn't really matter which of the child contexts the + * events ends up in. + */ + ctx = rcu_dereference(current->perf_counter_ctxp); + if (ctx) + perf_counter_comm_ctx(ctx, comm_event); + rcu_read_unlock(); +} + +void perf_counter_comm(struct task_struct *task) +{ + struct perf_comm_event comm_event; + + if (task->perf_counter_ctxp) + perf_counter_enable_on_exec(task); + + if (!atomic_read(&nr_comm_counters)) + return; + + comm_event = (struct perf_comm_event){ + .task = task, + /* .comm */ + /* .comm_size */ + .event = { + .header = { + .type = PERF_EVENT_COMM, + .misc = 0, + /* .size */ + }, + /* .pid */ + /* .tid */ + }, + }; + + perf_counter_comm_event(&comm_event); +} + +/* + * mmap tracking + */ + +struct perf_mmap_event { + struct vm_area_struct *vma; + + const char *file_name; + int file_size; + + struct { + struct perf_event_header header; + + u32 pid; + u32 tid; + u64 start; + u64 len; + u64 pgoff; + } event; +}; + +static void perf_counter_mmap_output(struct perf_counter *counter, + struct perf_mmap_event *mmap_event) +{ + struct perf_output_handle handle; + int size = mmap_event->event.header.size; + int ret = perf_output_begin(&handle, counter, size, 0, 0); + + if (ret) + return; + + mmap_event->event.pid = perf_counter_pid(counter, current); + mmap_event->event.tid = perf_counter_tid(counter, current); + + perf_output_put(&handle, mmap_event->event); + perf_output_copy(&handle, mmap_event->file_name, + mmap_event->file_size); + perf_output_end(&handle); +} + +static int perf_counter_mmap_match(struct perf_counter *counter, + struct perf_mmap_event *mmap_event) +{ + if (counter->attr.mmap) + return 1; + + return 0; +} + +static void perf_counter_mmap_ctx(struct perf_counter_context *ctx, + struct perf_mmap_event *mmap_event) +{ + struct perf_counter *counter; + + if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list)) + return; + + rcu_read_lock(); + list_for_each_entry_rcu(counter, &ctx->event_list, event_entry) { + if (perf_counter_mmap_match(counter, mmap_event)) + perf_counter_mmap_output(counter, mmap_event); + } + rcu_read_unlock(); +} + +static void perf_counter_mmap_event(struct perf_mmap_event *mmap_event) +{ + struct perf_cpu_context *cpuctx; + struct perf_counter_context *ctx; + struct vm_area_struct *vma = mmap_event->vma; + struct file *file = vma->vm_file; + unsigned int size; + char tmp[16]; + char *buf = NULL; + const char *name; + + memset(tmp, 0, sizeof(tmp)); + + if (file) { + /* + * d_path works from the end of the buffer backwards, so we + * need to add enough zero bytes after the string to handle + * the 64bit alignment we do later. + */ + buf = kzalloc(PATH_MAX + sizeof(u64), GFP_KERNEL); + if (!buf) { + name = strncpy(tmp, "//enomem", sizeof(tmp)); + goto got_name; + } + name = d_path(&file->f_path, buf, PATH_MAX); + if (IS_ERR(name)) { + name = strncpy(tmp, "//toolong", sizeof(tmp)); + goto got_name; + } + } else { + if (arch_vma_name(mmap_event->vma)) { + name = strncpy(tmp, arch_vma_name(mmap_event->vma), + sizeof(tmp)); + goto got_name; + } + + if (!vma->vm_mm) { + name = strncpy(tmp, "[vdso]", sizeof(tmp)); + goto got_name; + } + + name = strncpy(tmp, "//anon", sizeof(tmp)); + goto got_name; + } + +got_name: + size = ALIGN(strlen(name)+1, sizeof(u64)); + + mmap_event->file_name = name; + mmap_event->file_size = size; + + mmap_event->event.header.size = sizeof(mmap_event->event) + size; + + cpuctx = &get_cpu_var(perf_cpu_context); + perf_counter_mmap_ctx(&cpuctx->ctx, mmap_event); + put_cpu_var(perf_cpu_context); + + rcu_read_lock(); + /* + * doesn't really matter which of the child contexts the + * events ends up in. + */ + ctx = rcu_dereference(current->perf_counter_ctxp); + if (ctx) + perf_counter_mmap_ctx(ctx, mmap_event); + rcu_read_unlock(); + + kfree(buf); +} + +void __perf_counter_mmap(struct vm_area_struct *vma) +{ + struct perf_mmap_event mmap_event; + + if (!atomic_read(&nr_mmap_counters)) + return; + + mmap_event = (struct perf_mmap_event){ + .vma = vma, + /* .file_name */ + /* .file_size */ + .event = { + .header = { + .type = PERF_EVENT_MMAP, + .misc = 0, + /* .size */ + }, + /* .pid */ + /* .tid */ + .start = vma->vm_start, + .len = vma->vm_end - vma->vm_start, + .pgoff = vma->vm_pgoff, + }, + }; + + perf_counter_mmap_event(&mmap_event); +} + +/* + * IRQ throttle logging + */ + +static void perf_log_throttle(struct perf_counter *counter, int enable) +{ + struct perf_output_handle handle; + int ret; + + struct { + struct perf_event_header header; + u64 time; + u64 id; + u64 stream_id; + } throttle_event = { + .header = { + .type = PERF_EVENT_THROTTLE, + .misc = 0, + .size = sizeof(throttle_event), + }, + .time = sched_clock(), + .id = primary_counter_id(counter), + .stream_id = counter->id, + }; + + if (enable) + throttle_event.header.type = PERF_EVENT_UNTHROTTLE; + + ret = perf_output_begin(&handle, counter, sizeof(throttle_event), 1, 0); + if (ret) + return; + + perf_output_put(&handle, throttle_event); + perf_output_end(&handle); +} + +/* + * Generic counter overflow handling, sampling. + */ + +int perf_counter_overflow(struct perf_counter *counter, int nmi, + struct perf_sample_data *data) +{ + int events = atomic_read(&counter->event_limit); + int throttle = counter->pmu->unthrottle != NULL; + struct hw_perf_counter *hwc = &counter->hw; + int ret = 0; + + if (!throttle) { + hwc->interrupts++; + } else { + if (hwc->interrupts != MAX_INTERRUPTS) { + hwc->interrupts++; + if (HZ * hwc->interrupts > + (u64)sysctl_perf_counter_sample_rate) { + hwc->interrupts = MAX_INTERRUPTS; + perf_log_throttle(counter, 0); + ret = 1; + } + } else { + /* + * Keep re-disabling counters even though on the previous + * pass we disabled it - just in case we raced with a + * sched-in and the counter got enabled again: + */ + ret = 1; + } + } + + if (counter->attr.freq) { + u64 now = sched_clock(); + s64 delta = now - hwc->freq_stamp; + + hwc->freq_stamp = now; + + if (delta > 0 && delta < TICK_NSEC) + perf_adjust_period(counter, NSEC_PER_SEC / (int)delta); + } + + /* + * XXX event_limit might not quite work as expected on inherited + * counters + */ + + counter->pending_kill = POLL_IN; + if (events && atomic_dec_and_test(&counter->event_limit)) { + ret = 1; + counter->pending_kill = POLL_HUP; + if (nmi) { + counter->pending_disable = 1; + perf_pending_queue(&counter->pending, + perf_pending_counter); + } else + perf_counter_disable(counter); + } + + perf_counter_output(counter, nmi, data); + return ret; +} + +/* + * Generic software counter infrastructure + */ + +/* + * We directly increment counter->count and keep a second value in + * counter->hw.period_left to count intervals. This period counter + * is kept in the range [-sample_period, 0] so that we can use the + * sign as trigger. + */ + +static u64 perf_swcounter_set_period(struct perf_counter *counter) +{ + struct hw_perf_counter *hwc = &counter->hw; + u64 period = hwc->last_period; + u64 nr, offset; + s64 old, val; + + hwc->last_period = hwc->sample_period; + +again: + old = val = atomic64_read(&hwc->period_left); + if (val < 0) + return 0; + + nr = div64_u64(period + val, period); + offset = nr * period; + val -= offset; + if (atomic64_cmpxchg(&hwc->period_left, old, val) != old) + goto again; + + return nr; +} + +static void perf_swcounter_overflow(struct perf_counter *counter, + int nmi, struct perf_sample_data *data) +{ + struct hw_perf_counter *hwc = &counter->hw; + u64 overflow; + + data->period = counter->hw.last_period; + overflow = perf_swcounter_set_period(counter); + + if (hwc->interrupts == MAX_INTERRUPTS) + return; + + for (; overflow; overflow--) { + if (perf_counter_overflow(counter, nmi, data)) { + /* + * We inhibit the overflow from happening when + * hwc->interrupts == MAX_INTERRUPTS. + */ + break; + } + } +} + +static void perf_swcounter_unthrottle(struct perf_counter *counter) +{ + /* + * Nothing to do, we already reset hwc->interrupts. + */ +} + +static void perf_swcounter_add(struct perf_counter *counter, u64 nr, + int nmi, struct perf_sample_data *data) +{ + struct hw_perf_counter *hwc = &counter->hw; + + atomic64_add(nr, &counter->count); + + if (!hwc->sample_period) + return; + + if (!data->regs) + return; + + if (!atomic64_add_negative(nr, &hwc->period_left)) + perf_swcounter_overflow(counter, nmi, data); +} + +static int perf_swcounter_is_counting(struct perf_counter *counter) +{ + /* + * The counter is active, we're good! + */ + if (counter->state == PERF_COUNTER_STATE_ACTIVE) + return 1; + + /* + * The counter is off/error, not counting. + */ + if (counter->state != PERF_COUNTER_STATE_INACTIVE) + return 0; + + /* + * The counter is inactive, if the context is active + * we're part of a group that didn't make it on the 'pmu', + * not counting. + */ + if (counter->ctx->is_active) + return 0; + + /* + * We're inactive and the context is too, this means the + * task is scheduled out, we're counting events that happen + * to us, like migration events. + */ + return 1; +} + +static int perf_swcounter_match(struct perf_counter *counter, + enum perf_type_id type, + u32 event, struct pt_regs *regs) +{ + if (!perf_swcounter_is_counting(counter)) + return 0; + + if (counter->attr.type != type) + return 0; + if (counter->attr.config != event) + return 0; + + if (regs) { + if (counter->attr.exclude_user && user_mode(regs)) + return 0; + + if (counter->attr.exclude_kernel && !user_mode(regs)) + return 0; + } + + return 1; +} + +static void perf_swcounter_ctx_event(struct perf_counter_context *ctx, + enum perf_type_id type, + u32 event, u64 nr, int nmi, + struct perf_sample_data *data) +{ + struct perf_counter *counter; + + if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list)) + return; + + rcu_read_lock(); + list_for_each_entry_rcu(counter, &ctx->event_list, event_entry) { + if (perf_swcounter_match(counter, type, event, data->regs)) + perf_swcounter_add(counter, nr, nmi, data); + } + rcu_read_unlock(); +} + +static int *perf_swcounter_recursion_context(struct perf_cpu_context *cpuctx) +{ + if (in_nmi()) + return &cpuctx->recursion[3]; + + if (in_irq()) + return &cpuctx->recursion[2]; + + if (in_softirq()) + return &cpuctx->recursion[1]; + + return &cpuctx->recursion[0]; +} + +static void do_perf_swcounter_event(enum perf_type_id type, u32 event, + u64 nr, int nmi, + struct perf_sample_data *data) +{ + struct perf_cpu_context *cpuctx = &get_cpu_var(perf_cpu_context); + int *recursion = perf_swcounter_recursion_context(cpuctx); + struct perf_counter_context *ctx; + + if (*recursion) + goto out; + + (*recursion)++; + barrier(); + + perf_swcounter_ctx_event(&cpuctx->ctx, type, event, + nr, nmi, data); + rcu_read_lock(); + /* + * doesn't really matter which of the child contexts the + * events ends up in. + */ + ctx = rcu_dereference(current->perf_counter_ctxp); + if (ctx) + perf_swcounter_ctx_event(ctx, type, event, nr, nmi, data); + rcu_read_unlock(); + + barrier(); + (*recursion)--; + +out: + put_cpu_var(perf_cpu_context); +} + +void __perf_swcounter_event(u32 event, u64 nr, int nmi, + struct pt_regs *regs, u64 addr) +{ + struct perf_sample_data data = { + .regs = regs, + .addr = addr, + }; + + do_perf_swcounter_event(PERF_TYPE_SOFTWARE, event, nr, nmi, &data); +} + +static void perf_swcounter_read(struct perf_counter *counter) +{ +} + +static int perf_swcounter_enable(struct perf_counter *counter) +{ + struct hw_perf_counter *hwc = &counter->hw; + + if (hwc->sample_period) { + hwc->last_period = hwc->sample_period; + perf_swcounter_set_period(counter); + } + return 0; +} + +static void perf_swcounter_disable(struct perf_counter *counter) +{ +} + +static const struct pmu perf_ops_generic = { + .enable = perf_swcounter_enable, + .disable = perf_swcounter_disable, + .read = perf_swcounter_read, + .unthrottle = perf_swcounter_unthrottle, +}; + +/* + * hrtimer based swcounter callback + */ + +static enum hrtimer_restart perf_swcounter_hrtimer(struct hrtimer *hrtimer) +{ + enum hrtimer_restart ret = HRTIMER_RESTART; + struct perf_sample_data data; + struct perf_counter *counter; + u64 period; + + counter = container_of(hrtimer, struct perf_counter, hw.hrtimer); + counter->pmu->read(counter); + + data.addr = 0; + data.regs = get_irq_regs(); + /* + * In case we exclude kernel IPs or are somehow not in interrupt + * context, provide the next best thing, the user IP. + */ + if ((counter->attr.exclude_kernel || !data.regs) && + !counter->attr.exclude_user) + data.regs = task_pt_regs(current); + + if (data.regs) { + if (perf_counter_overflow(counter, 0, &data)) + ret = HRTIMER_NORESTART; + } + + period = max_t(u64, 10000, counter->hw.sample_period); + hrtimer_forward_now(hrtimer, ns_to_ktime(period)); + + return ret; +} + +/* + * Software counter: cpu wall time clock + */ + +static void cpu_clock_perf_counter_update(struct perf_counter *counter) +{ + int cpu = raw_smp_processor_id(); + s64 prev; + u64 now; + + now = cpu_clock(cpu); + prev = atomic64_read(&counter->hw.prev_count); + atomic64_set(&counter->hw.prev_count, now); + atomic64_add(now - prev, &counter->count); +} + +static int cpu_clock_perf_counter_enable(struct perf_counter *counter) +{ + struct hw_perf_counter *hwc = &counter->hw; + int cpu = raw_smp_processor_id(); + + atomic64_set(&hwc->prev_count, cpu_clock(cpu)); + hrtimer_init(&hwc->hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + hwc->hrtimer.function = perf_swcounter_hrtimer; + if (hwc->sample_period) { + u64 period = max_t(u64, 10000, hwc->sample_period); + __hrtimer_start_range_ns(&hwc->hrtimer, + ns_to_ktime(period), 0, + HRTIMER_MODE_REL, 0); + } + + return 0; +} + +static void cpu_clock_perf_counter_disable(struct perf_counter *counter) +{ + if (counter->hw.sample_period) + hrtimer_cancel(&counter->hw.hrtimer); + cpu_clock_perf_counter_update(counter); +} + +static void cpu_clock_perf_counter_read(struct perf_counter *counter) +{ + cpu_clock_perf_counter_update(counter); +} + +static const struct pmu perf_ops_cpu_clock = { + .enable = cpu_clock_perf_counter_enable, + .disable = cpu_clock_perf_counter_disable, + .read = cpu_clock_perf_counter_read, +}; + +/* + * Software counter: task time clock + */ + +static void task_clock_perf_counter_update(struct perf_counter *counter, u64 now) +{ + u64 prev; + s64 delta; + + prev = atomic64_xchg(&counter->hw.prev_count, now); + delta = now - prev; + atomic64_add(delta, &counter->count); +} + +static int task_clock_perf_counter_enable(struct perf_counter *counter) +{ + struct hw_perf_counter *hwc = &counter->hw; + u64 now; + + now = counter->ctx->time; + + atomic64_set(&hwc->prev_count, now); + hrtimer_init(&hwc->hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + hwc->hrtimer.function = perf_swcounter_hrtimer; + if (hwc->sample_period) { + u64 period = max_t(u64, 10000, hwc->sample_period); + __hrtimer_start_range_ns(&hwc->hrtimer, + ns_to_ktime(period), 0, + HRTIMER_MODE_REL, 0); + } + + return 0; +} + +static void task_clock_perf_counter_disable(struct perf_counter *counter) +{ + if (counter->hw.sample_period) + hrtimer_cancel(&counter->hw.hrtimer); + task_clock_perf_counter_update(counter, counter->ctx->time); + +} + +static void task_clock_perf_counter_read(struct perf_counter *counter) +{ + u64 time; + + if (!in_nmi()) { + update_context_time(counter->ctx); + time = counter->ctx->time; + } else { + u64 now = perf_clock(); + u64 delta = now - counter->ctx->timestamp; + time = counter->ctx->time + delta; + } + + task_clock_perf_counter_update(counter, time); +} + +static const struct pmu perf_ops_task_clock = { + .enable = task_clock_perf_counter_enable, + .disable = task_clock_perf_counter_disable, + .read = task_clock_perf_counter_read, +}; + +#ifdef CONFIG_EVENT_PROFILE +void perf_tpcounter_event(int event_id, u64 addr, u64 count, void *record, + int entry_size) +{ + struct perf_raw_record raw = { + .size = entry_size, + .data = record, + }; + + struct perf_sample_data data = { + .regs = get_irq_regs(), + .addr = addr, + .raw = &raw, + }; + + if (!data.regs) + data.regs = task_pt_regs(current); + + do_perf_swcounter_event(PERF_TYPE_TRACEPOINT, event_id, count, 1, &data); +} +EXPORT_SYMBOL_GPL(perf_tpcounter_event); + +extern int ftrace_profile_enable(int); +extern void ftrace_profile_disable(int); + +static void tp_perf_counter_destroy(struct perf_counter *counter) +{ + ftrace_profile_disable(counter->attr.config); +} + +static const struct pmu *tp_perf_counter_init(struct perf_counter *counter) +{ + /* + * Raw tracepoint data is a severe data leak, only allow root to + * have these. + */ + if ((counter->attr.sample_type & PERF_SAMPLE_RAW) && + perf_paranoid_tracepoint_raw() && + !capable(CAP_SYS_ADMIN)) + return ERR_PTR(-EPERM); + + if (ftrace_profile_enable(counter->attr.config)) + return NULL; + + counter->destroy = tp_perf_counter_destroy; + + return &perf_ops_generic; +} +#else +static const struct pmu *tp_perf_counter_init(struct perf_counter *counter) +{ + return NULL; +} +#endif + +atomic_t perf_swcounter_enabled[PERF_COUNT_SW_MAX]; + +static void sw_perf_counter_destroy(struct perf_counter *counter) +{ + u64 event = counter->attr.config; + + WARN_ON(counter->parent); + + atomic_dec(&perf_swcounter_enabled[event]); +} + +static const struct pmu *sw_perf_counter_init(struct perf_counter *counter) +{ + const struct pmu *pmu = NULL; + u64 event = counter->attr.config; + + /* + * Software counters (currently) can't in general distinguish + * between user, kernel and hypervisor events. + * However, context switches and cpu migrations are considered + * to be kernel events, and page faults are never hypervisor + * events. + */ + switch (event) { + case PERF_COUNT_SW_CPU_CLOCK: + pmu = &perf_ops_cpu_clock; + + break; + case PERF_COUNT_SW_TASK_CLOCK: + /* + * If the user instantiates this as a per-cpu counter, + * use the cpu_clock counter instead. + */ + if (counter->ctx->task) + pmu = &perf_ops_task_clock; + else + pmu = &perf_ops_cpu_clock; + + break; + case PERF_COUNT_SW_PAGE_FAULTS: + case PERF_COUNT_SW_PAGE_FAULTS_MIN: + case PERF_COUNT_SW_PAGE_FAULTS_MAJ: + case PERF_COUNT_SW_CONTEXT_SWITCHES: + case PERF_COUNT_SW_CPU_MIGRATIONS: + if (!counter->parent) { + atomic_inc(&perf_swcounter_enabled[event]); + counter->destroy = sw_perf_counter_destroy; + } + pmu = &perf_ops_generic; + break; + } + + return pmu; +} + +/* + * Allocate and initialize a counter structure + */ +static struct perf_counter * +perf_counter_alloc(struct perf_counter_attr *attr, + int cpu, + struct perf_counter_context *ctx, + struct perf_counter *group_leader, + struct perf_counter *parent_counter, + gfp_t gfpflags) +{ + const struct pmu *pmu; + struct perf_counter *counter; + struct hw_perf_counter *hwc; + long err; + + counter = kzalloc(sizeof(*counter), gfpflags); + if (!counter) + return ERR_PTR(-ENOMEM); + + /* + * Single counters are their own group leaders, with an + * empty sibling list: + */ + if (!group_leader) + group_leader = counter; + + mutex_init(&counter->child_mutex); + INIT_LIST_HEAD(&counter->child_list); + + INIT_LIST_HEAD(&counter->list_entry); + INIT_LIST_HEAD(&counter->event_entry); + INIT_LIST_HEAD(&counter->sibling_list); + init_waitqueue_head(&counter->waitq); + + mutex_init(&counter->mmap_mutex); + + counter->cpu = cpu; + counter->attr = *attr; + counter->group_leader = group_leader; + counter->pmu = NULL; + counter->ctx = ctx; + counter->oncpu = -1; + + counter->parent = parent_counter; + + counter->ns = get_pid_ns(current->nsproxy->pid_ns); + counter->id = atomic64_inc_return(&perf_counter_id); + + counter->state = PERF_COUNTER_STATE_INACTIVE; + + if (attr->disabled) + counter->state = PERF_COUNTER_STATE_OFF; + + pmu = NULL; + + hwc = &counter->hw; + hwc->sample_period = attr->sample_period; + if (attr->freq && attr->sample_freq) + hwc->sample_period = 1; + hwc->last_period = hwc->sample_period; + + atomic64_set(&hwc->period_left, hwc->sample_period); + + /* + * we currently do not support PERF_FORMAT_GROUP on inherited counters + */ + if (attr->inherit && (attr->read_format & PERF_FORMAT_GROUP)) + goto done; + + switch (attr->type) { + case PERF_TYPE_RAW: + case PERF_TYPE_HARDWARE: + case PERF_TYPE_HW_CACHE: + pmu = hw_perf_counter_init(counter); + break; + + case PERF_TYPE_SOFTWARE: + pmu = sw_perf_counter_init(counter); + break; + + case PERF_TYPE_TRACEPOINT: + pmu = tp_perf_counter_init(counter); + break; + + default: + break; + } +done: + err = 0; + if (!pmu) + err = -EINVAL; + else if (IS_ERR(pmu)) + err = PTR_ERR(pmu); + + if (err) { + if (counter->ns) + put_pid_ns(counter->ns); + kfree(counter); + return ERR_PTR(err); + } + + counter->pmu = pmu; + + if (!counter->parent) { + atomic_inc(&nr_counters); + if (counter->attr.mmap) + atomic_inc(&nr_mmap_counters); + if (counter->attr.comm) + atomic_inc(&nr_comm_counters); + if (counter->attr.task) + atomic_inc(&nr_task_counters); + } + + return counter; +} + +static int perf_copy_attr(struct perf_counter_attr __user *uattr, + struct perf_counter_attr *attr) +{ + int ret; + u32 size; + + if (!access_ok(VERIFY_WRITE, uattr, PERF_ATTR_SIZE_VER0)) + return -EFAULT; + + /* + * zero the full structure, so that a short copy will be nice. + */ + memset(attr, 0, sizeof(*attr)); + + ret = get_user(size, &uattr->size); + if (ret) + return ret; + + if (size > PAGE_SIZE) /* silly large */ + goto err_size; + + if (!size) /* abi compat */ + size = PERF_ATTR_SIZE_VER0; + + if (size < PERF_ATTR_SIZE_VER0) + goto err_size; + + /* + * If we're handed a bigger struct than we know of, + * ensure all the unknown bits are 0. + */ + if (size > sizeof(*attr)) { + unsigned long val; + unsigned long __user *addr; + unsigned long __user *end; + + addr = PTR_ALIGN((void __user *)uattr + sizeof(*attr), + sizeof(unsigned long)); + end = PTR_ALIGN((void __user *)uattr + size, + sizeof(unsigned long)); + + for (; addr < end; addr += sizeof(unsigned long)) { + ret = get_user(val, addr); + if (ret) + return ret; + if (val) + goto err_size; + } + } + + ret = copy_from_user(attr, uattr, size); + if (ret) + return -EFAULT; + + /* + * If the type exists, the corresponding creation will verify + * the attr->config. + */ + if (attr->type >= PERF_TYPE_MAX) + return -EINVAL; + + if (attr->__reserved_1 || attr->__reserved_2 || attr->__reserved_3) + return -EINVAL; + + if (attr->sample_type & ~(PERF_SAMPLE_MAX-1)) + return -EINVAL; + + if (attr->read_format & ~(PERF_FORMAT_MAX-1)) + return -EINVAL; + +out: + return ret; + +err_size: + put_user(sizeof(*attr), &uattr->size); + ret = -E2BIG; + goto out; +} + +int perf_counter_set_output(struct perf_counter *counter, int output_fd) +{ + struct perf_counter *output_counter = NULL; + struct file *output_file = NULL; + struct perf_counter *old_output; + int fput_needed = 0; + int ret = -EINVAL; + + if (!output_fd) + goto set; + + output_file = fget_light(output_fd, &fput_needed); + if (!output_file) + return -EBADF; + + if (output_file->f_op != &perf_fops) + goto out; + + output_counter = output_file->private_data; + + /* Don't chain output fds */ + if (output_counter->output) + goto out; + + /* Don't set an output fd when we already have an output channel */ + if (counter->data) + goto out; + + atomic_long_inc(&output_file->f_count); + +set: + mutex_lock(&counter->mmap_mutex); + old_output = counter->output; + rcu_assign_pointer(counter->output, output_counter); + mutex_unlock(&counter->mmap_mutex); + + if (old_output) { + /* + * we need to make sure no existing perf_output_*() + * is still referencing this counter. + */ + synchronize_rcu(); + fput(old_output->filp); + } + + ret = 0; +out: + fput_light(output_file, fput_needed); + return ret; +} + +/** + * sys_perf_counter_open - open a performance counter, associate it to a task/cpu + * + * @attr_uptr: event type attributes for monitoring/sampling + * @pid: target pid + * @cpu: target cpu + * @group_fd: group leader counter fd + */ +SYSCALL_DEFINE5(perf_counter_open, + struct perf_counter_attr __user *, attr_uptr, + pid_t, pid, int, cpu, int, group_fd, unsigned long, flags) +{ + struct perf_counter *counter, *group_leader; + struct perf_counter_attr attr; + struct perf_counter_context *ctx; + struct file *counter_file = NULL; + struct file *group_file = NULL; + int fput_needed = 0; + int fput_needed2 = 0; + int err; + + /* for future expandability... */ + if (flags & ~(PERF_FLAG_FD_NO_GROUP | PERF_FLAG_FD_OUTPUT)) + return -EINVAL; + + err = perf_copy_attr(attr_uptr, &attr); + if (err) + return err; + + if (!attr.exclude_kernel) { + if (perf_paranoid_kernel() && !capable(CAP_SYS_ADMIN)) + return -EACCES; + } + + if (attr.freq) { + if (attr.sample_freq > sysctl_perf_counter_sample_rate) + return -EINVAL; + } + + /* + * Get the target context (task or percpu): + */ + ctx = find_get_context(pid, cpu); + if (IS_ERR(ctx)) + return PTR_ERR(ctx); + + /* + * Look up the group leader (we will attach this counter to it): + */ + group_leader = NULL; + if (group_fd != -1 && !(flags & PERF_FLAG_FD_NO_GROUP)) { + err = -EINVAL; + group_file = fget_light(group_fd, &fput_needed); + if (!group_file) + goto err_put_context; + if (group_file->f_op != &perf_fops) + goto err_put_context; + + group_leader = group_file->private_data; + /* + * Do not allow a recursive hierarchy (this new sibling + * becoming part of another group-sibling): + */ + if (group_leader->group_leader != group_leader) + goto err_put_context; + /* + * Do not allow to attach to a group in a different + * task or CPU context: + */ + if (group_leader->ctx != ctx) + goto err_put_context; + /* + * Only a group leader can be exclusive or pinned + */ + if (attr.exclusive || attr.pinned) + goto err_put_context; + } + + counter = perf_counter_alloc(&attr, cpu, ctx, group_leader, + NULL, GFP_KERNEL); + err = PTR_ERR(counter); + if (IS_ERR(counter)) + goto err_put_context; + + err = anon_inode_getfd("[perf_counter]", &perf_fops, counter, 0); + if (err < 0) + goto err_free_put_context; + + counter_file = fget_light(err, &fput_needed2); + if (!counter_file) + goto err_free_put_context; + + if (flags & PERF_FLAG_FD_OUTPUT) { + err = perf_counter_set_output(counter, group_fd); + if (err) + goto err_fput_free_put_context; + } + + counter->filp = counter_file; + WARN_ON_ONCE(ctx->parent_ctx); + mutex_lock(&ctx->mutex); + perf_install_in_context(ctx, counter, cpu); + ++ctx->generation; + mutex_unlock(&ctx->mutex); + + counter->owner = current; + get_task_struct(current); + mutex_lock(¤t->perf_counter_mutex); + list_add_tail(&counter->owner_entry, ¤t->perf_counter_list); + mutex_unlock(¤t->perf_counter_mutex); + +err_fput_free_put_context: + fput_light(counter_file, fput_needed2); + +err_free_put_context: + if (err < 0) + kfree(counter); + +err_put_context: + if (err < 0) + put_ctx(ctx); + + fput_light(group_file, fput_needed); + + return err; +} + +/* + * inherit a counter from parent task to child task: + */ +static struct perf_counter * +inherit_counter(struct perf_counter *parent_counter, + struct task_struct *parent, + struct perf_counter_context *parent_ctx, + struct task_struct *child, + struct perf_counter *group_leader, + struct perf_counter_context *child_ctx) +{ + struct perf_counter *child_counter; + + /* + * Instead of creating recursive hierarchies of counters, + * we link inherited counters back to the original parent, + * which has a filp for sure, which we use as the reference + * count: + */ + if (parent_counter->parent) + parent_counter = parent_counter->parent; + + child_counter = perf_counter_alloc(&parent_counter->attr, + parent_counter->cpu, child_ctx, + group_leader, parent_counter, + GFP_KERNEL); + if (IS_ERR(child_counter)) + return child_counter; + get_ctx(child_ctx); + + /* + * Make the child state follow the state of the parent counter, + * not its attr.disabled bit. We hold the parent's mutex, + * so we won't race with perf_counter_{en, dis}able_family. + */ + if (parent_counter->state >= PERF_COUNTER_STATE_INACTIVE) + child_counter->state = PERF_COUNTER_STATE_INACTIVE; + else + child_counter->state = PERF_COUNTER_STATE_OFF; + + if (parent_counter->attr.freq) + child_counter->hw.sample_period = parent_counter->hw.sample_period; + + /* + * Link it up in the child's context: + */ + add_counter_to_ctx(child_counter, child_ctx); + + /* + * Get a reference to the parent filp - we will fput it + * when the child counter exits. This is safe to do because + * we are in the parent and we know that the filp still + * exists and has a nonzero count: + */ + atomic_long_inc(&parent_counter->filp->f_count); + + /* + * Link this into the parent counter's child list + */ + WARN_ON_ONCE(parent_counter->ctx->parent_ctx); + mutex_lock(&parent_counter->child_mutex); + list_add_tail(&child_counter->child_list, &parent_counter->child_list); + mutex_unlock(&parent_counter->child_mutex); + + return child_counter; +} + +static int inherit_group(struct perf_counter *parent_counter, + struct task_struct *parent, + struct perf_counter_context *parent_ctx, + struct task_struct *child, + struct perf_counter_context *child_ctx) +{ + struct perf_counter *leader; + struct perf_counter *sub; + struct perf_counter *child_ctr; + + leader = inherit_counter(parent_counter, parent, parent_ctx, + child, NULL, child_ctx); + if (IS_ERR(leader)) + return PTR_ERR(leader); + list_for_each_entry(sub, &parent_counter->sibling_list, list_entry) { + child_ctr = inherit_counter(sub, parent, parent_ctx, + child, leader, child_ctx); + if (IS_ERR(child_ctr)) + return PTR_ERR(child_ctr); + } + return 0; +} + +static void sync_child_counter(struct perf_counter *child_counter, + struct task_struct *child) +{ + struct perf_counter *parent_counter = child_counter->parent; + u64 child_val; + + if (child_counter->attr.inherit_stat) + perf_counter_read_event(child_counter, child); + + child_val = atomic64_read(&child_counter->count); + + /* + * Add back the child's count to the parent's count: + */ + atomic64_add(child_val, &parent_counter->count); + atomic64_add(child_counter->total_time_enabled, + &parent_counter->child_total_time_enabled); + atomic64_add(child_counter->total_time_running, + &parent_counter->child_total_time_running); + + /* + * Remove this counter from the parent's list + */ + WARN_ON_ONCE(parent_counter->ctx->parent_ctx); + mutex_lock(&parent_counter->child_mutex); + list_del_init(&child_counter->child_list); + mutex_unlock(&parent_counter->child_mutex); + + /* + * Release the parent counter, if this was the last + * reference to it. + */ + fput(parent_counter->filp); +} + +static void +__perf_counter_exit_task(struct perf_counter *child_counter, + struct perf_counter_context *child_ctx, + struct task_struct *child) +{ + struct perf_counter *parent_counter; + + update_counter_times(child_counter); + perf_counter_remove_from_context(child_counter); + + parent_counter = child_counter->parent; + /* + * It can happen that parent exits first, and has counters + * that are still around due to the child reference. These + * counters need to be zapped - but otherwise linger. + */ + if (parent_counter) { + sync_child_counter(child_counter, child); + free_counter(child_counter); + } +} + +/* + * When a child task exits, feed back counter values to parent counters. + */ +void perf_counter_exit_task(struct task_struct *child) +{ + struct perf_counter *child_counter, *tmp; + struct perf_counter_context *child_ctx; + unsigned long flags; + + if (likely(!child->perf_counter_ctxp)) { + perf_counter_task(child, NULL, 0); + return; + } + + local_irq_save(flags); + /* + * We can't reschedule here because interrupts are disabled, + * and either child is current or it is a task that can't be + * scheduled, so we are now safe from rescheduling changing + * our context. + */ + child_ctx = child->perf_counter_ctxp; + __perf_counter_task_sched_out(child_ctx); + + /* + * Take the context lock here so that if find_get_context is + * reading child->perf_counter_ctxp, we wait until it has + * incremented the context's refcount before we do put_ctx below. + */ + spin_lock(&child_ctx->lock); + child->perf_counter_ctxp = NULL; + /* + * If this context is a clone; unclone it so it can't get + * swapped to another process while we're removing all + * the counters from it. + */ + unclone_ctx(child_ctx); + spin_unlock_irqrestore(&child_ctx->lock, flags); + + /* + * Report the task dead after unscheduling the counters so that we + * won't get any samples after PERF_EVENT_EXIT. We can however still + * get a few PERF_EVENT_READ events. + */ + perf_counter_task(child, child_ctx, 0); + + /* + * We can recurse on the same lock type through: + * + * __perf_counter_exit_task() + * sync_child_counter() + * fput(parent_counter->filp) + * perf_release() + * mutex_lock(&ctx->mutex) + * + * But since its the parent context it won't be the same instance. + */ + mutex_lock_nested(&child_ctx->mutex, SINGLE_DEPTH_NESTING); + +again: + list_for_each_entry_safe(child_counter, tmp, &child_ctx->counter_list, + list_entry) + __perf_counter_exit_task(child_counter, child_ctx, child); + + /* + * If the last counter was a group counter, it will have appended all + * its siblings to the list, but we obtained 'tmp' before that which + * will still point to the list head terminating the iteration. + */ + if (!list_empty(&child_ctx->counter_list)) + goto again; + + mutex_unlock(&child_ctx->mutex); + + put_ctx(child_ctx); +} + +/* + * free an unexposed, unused context as created by inheritance by + * init_task below, used by fork() in case of fail. + */ +void perf_counter_free_task(struct task_struct *task) +{ + struct perf_counter_context *ctx = task->perf_counter_ctxp; + struct perf_counter *counter, *tmp; + + if (!ctx) + return; + + mutex_lock(&ctx->mutex); +again: + list_for_each_entry_safe(counter, tmp, &ctx->counter_list, list_entry) { + struct perf_counter *parent = counter->parent; + + if (WARN_ON_ONCE(!parent)) + continue; + + mutex_lock(&parent->child_mutex); + list_del_init(&counter->child_list); + mutex_unlock(&parent->child_mutex); + + fput(parent->filp); + + list_del_counter(counter, ctx); + free_counter(counter); + } + + if (!list_empty(&ctx->counter_list)) + goto again; + + mutex_unlock(&ctx->mutex); + + put_ctx(ctx); +} + +/* + * Initialize the perf_counter context in task_struct + */ +int perf_counter_init_task(struct task_struct *child) +{ + struct perf_counter_context *child_ctx, *parent_ctx; + struct perf_counter_context *cloned_ctx; + struct perf_counter *counter; + struct task_struct *parent = current; + int inherited_all = 1; + int ret = 0; + + child->perf_counter_ctxp = NULL; + + mutex_init(&child->perf_counter_mutex); + INIT_LIST_HEAD(&child->perf_counter_list); + + if (likely(!parent->perf_counter_ctxp)) + return 0; + + /* + * This is executed from the parent task context, so inherit + * counters that have been marked for cloning. + * First allocate and initialize a context for the child. + */ + + child_ctx = kmalloc(sizeof(struct perf_counter_context), GFP_KERNEL); + if (!child_ctx) + return -ENOMEM; + + __perf_counter_init_context(child_ctx, child); + child->perf_counter_ctxp = child_ctx; + get_task_struct(child); + + /* + * If the parent's context is a clone, pin it so it won't get + * swapped under us. + */ + parent_ctx = perf_pin_task_context(parent); + + /* + * No need to check if parent_ctx != NULL here; since we saw + * it non-NULL earlier, the only reason for it to become NULL + * is if we exit, and since we're currently in the middle of + * a fork we can't be exiting at the same time. + */ + + /* + * Lock the parent list. No need to lock the child - not PID + * hashed yet and not running, so nobody can access it. + */ + mutex_lock(&parent_ctx->mutex); + + /* + * We dont have to disable NMIs - we are only looking at + * the list, not manipulating it: + */ + list_for_each_entry_rcu(counter, &parent_ctx->event_list, event_entry) { + if (counter != counter->group_leader) + continue; + + if (!counter->attr.inherit) { + inherited_all = 0; + continue; + } + + ret = inherit_group(counter, parent, parent_ctx, + child, child_ctx); + if (ret) { + inherited_all = 0; + break; + } + } + + if (inherited_all) { + /* + * Mark the child context as a clone of the parent + * context, or of whatever the parent is a clone of. + * Note that if the parent is a clone, it could get + * uncloned at any point, but that doesn't matter + * because the list of counters and the generation + * count can't have changed since we took the mutex. + */ + cloned_ctx = rcu_dereference(parent_ctx->parent_ctx); + if (cloned_ctx) { + child_ctx->parent_ctx = cloned_ctx; + child_ctx->parent_gen = parent_ctx->parent_gen; + } else { + child_ctx->parent_ctx = parent_ctx; + child_ctx->parent_gen = parent_ctx->generation; + } + get_ctx(child_ctx->parent_ctx); + } + + mutex_unlock(&parent_ctx->mutex); + + perf_unpin_context(parent_ctx); + + return ret; +} + +static void __cpuinit perf_counter_init_cpu(int cpu) +{ + struct perf_cpu_context *cpuctx; + + cpuctx = &per_cpu(perf_cpu_context, cpu); + __perf_counter_init_context(&cpuctx->ctx, NULL); + + spin_lock(&perf_resource_lock); + cpuctx->max_pertask = perf_max_counters - perf_reserved_percpu; + spin_unlock(&perf_resource_lock); + + hw_perf_counter_setup(cpu); +} + +#ifdef CONFIG_HOTPLUG_CPU +static void __perf_counter_exit_cpu(void *info) +{ + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_counter_context *ctx = &cpuctx->ctx; + struct perf_counter *counter, *tmp; + + list_for_each_entry_safe(counter, tmp, &ctx->counter_list, list_entry) + __perf_counter_remove_from_context(counter); +} +static void perf_counter_exit_cpu(int cpu) +{ + struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct perf_counter_context *ctx = &cpuctx->ctx; + + mutex_lock(&ctx->mutex); + smp_call_function_single(cpu, __perf_counter_exit_cpu, NULL, 1); + mutex_unlock(&ctx->mutex); +} +#else +static inline void perf_counter_exit_cpu(int cpu) { } +#endif + +static int __cpuinit +perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) +{ + unsigned int cpu = (long)hcpu; + + switch (action) { + + case CPU_UP_PREPARE: + case CPU_UP_PREPARE_FROZEN: + perf_counter_init_cpu(cpu); + break; + + case CPU_ONLINE: + case CPU_ONLINE_FROZEN: + hw_perf_counter_setup_online(cpu); + break; + + case CPU_DOWN_PREPARE: + case CPU_DOWN_PREPARE_FROZEN: + perf_counter_exit_cpu(cpu); + break; + + default: + break; + } + + return NOTIFY_OK; +} + +/* + * This has to have a higher priority than migration_notifier in sched.c. + */ +static struct notifier_block __cpuinitdata perf_cpu_nb = { + .notifier_call = perf_cpu_notify, + .priority = 20, +}; + +void __init perf_counter_init(void) +{ + perf_cpu_notify(&perf_cpu_nb, (unsigned long)CPU_UP_PREPARE, + (void *)(long)smp_processor_id()); + perf_cpu_notify(&perf_cpu_nb, (unsigned long)CPU_ONLINE, + (void *)(long)smp_processor_id()); + register_cpu_notifier(&perf_cpu_nb); +} + +static ssize_t perf_show_reserve_percpu(struct sysdev_class *class, char *buf) +{ + return sprintf(buf, "%d\n", perf_reserved_percpu); +} + +static ssize_t +perf_set_reserve_percpu(struct sysdev_class *class, + const char *buf, + size_t count) +{ + struct perf_cpu_context *cpuctx; + unsigned long val; + int err, cpu, mpt; + + err = strict_strtoul(buf, 10, &val); + if (err) + return err; + if (val > perf_max_counters) + return -EINVAL; + + spin_lock(&perf_resource_lock); + perf_reserved_percpu = val; + for_each_online_cpu(cpu) { + cpuctx = &per_cpu(perf_cpu_context, cpu); + spin_lock_irq(&cpuctx->ctx.lock); + mpt = min(perf_max_counters - cpuctx->ctx.nr_counters, + perf_max_counters - perf_reserved_percpu); + cpuctx->max_pertask = mpt; + spin_unlock_irq(&cpuctx->ctx.lock); + } + spin_unlock(&perf_resource_lock); + + return count; +} + +static ssize_t perf_show_overcommit(struct sysdev_class *class, char *buf) +{ + return sprintf(buf, "%d\n", perf_overcommit); +} + +static ssize_t +perf_set_overcommit(struct sysdev_class *class, const char *buf, size_t count) +{ + unsigned long val; + int err; + + err = strict_strtoul(buf, 10, &val); + if (err) + return err; + if (val > 1) + return -EINVAL; + + spin_lock(&perf_resource_lock); + perf_overcommit = val; + spin_unlock(&perf_resource_lock); + + return count; +} + +static SYSDEV_CLASS_ATTR( + reserve_percpu, + 0644, + perf_show_reserve_percpu, + perf_set_reserve_percpu + ); + +static SYSDEV_CLASS_ATTR( + overcommit, + 0644, + perf_show_overcommit, + perf_set_overcommit + ); + +static struct attribute *perfclass_attrs[] = { + &attr_reserve_percpu.attr, + &attr_overcommit.attr, + NULL +}; + +static struct attribute_group perfclass_attr_group = { + .attrs = perfclass_attrs, + .name = "perf_counters", +}; + +static int __init perf_counter_sysfs_init(void) +{ + return sysfs_create_group(&cpu_sysdev_class.kset.kobj, + &perfclass_attr_group); +} +device_initcall(perf_counter_sysfs_init); diff --git a/kernel/pid.c b/kernel/pid.c index 1b3586fe753..31310b5d3f5 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -378,31 +378,22 @@ EXPORT_SYMBOL(pid_task); /* * Must be called under rcu_read_lock() or with tasklist_lock read-held. */ -struct task_struct *find_task_by_pid_type_ns(int type, int nr, - struct pid_namespace *ns) +struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns) { - return pid_task(find_pid_ns(nr, ns), type); + return pid_task(find_pid_ns(nr, ns), PIDTYPE_PID); } -EXPORT_SYMBOL(find_task_by_pid_type_ns); - struct task_struct *find_task_by_vpid(pid_t vnr) { - return find_task_by_pid_type_ns(PIDTYPE_PID, vnr, - current->nsproxy->pid_ns); -} -EXPORT_SYMBOL(find_task_by_vpid); - -struct task_struct *find_task_by_pid_ns(pid_t nr, struct pid_namespace *ns) -{ - return find_task_by_pid_type_ns(PIDTYPE_PID, nr, ns); + return find_task_by_pid_ns(vnr, current->nsproxy->pid_ns); } -EXPORT_SYMBOL(find_task_by_pid_ns); struct pid *get_task_pid(struct task_struct *task, enum pid_type type) { struct pid *pid; rcu_read_lock(); + if (type != PIDTYPE_PID) + task = task->group_leader; pid = get_pid(task->pids[type].pid); rcu_read_unlock(); return pid; @@ -450,11 +441,24 @@ pid_t pid_vnr(struct pid *pid) } EXPORT_SYMBOL_GPL(pid_vnr); -pid_t task_pid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) +pid_t __task_pid_nr_ns(struct task_struct *task, enum pid_type type, + struct pid_namespace *ns) { - return pid_nr_ns(task_pid(tsk), ns); + pid_t nr = 0; + + rcu_read_lock(); + if (!ns) + ns = current->nsproxy->pid_ns; + if (likely(pid_alive(task))) { + if (type != PIDTYPE_PID) + task = task->group_leader; + nr = pid_nr_ns(task->pids[type].pid, ns); + } + rcu_read_unlock(); + + return nr; } -EXPORT_SYMBOL(task_pid_nr_ns); +EXPORT_SYMBOL(__task_pid_nr_ns); pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) { @@ -462,18 +466,6 @@ pid_t task_tgid_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) } EXPORT_SYMBOL(task_tgid_nr_ns); -pid_t task_pgrp_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) -{ - return pid_nr_ns(task_pgrp(tsk), ns); -} -EXPORT_SYMBOL(task_pgrp_nr_ns); - -pid_t task_session_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) -{ - return pid_nr_ns(task_session(tsk), ns); -} -EXPORT_SYMBOL(task_session_nr_ns); - struct pid_namespace *task_active_pid_ns(struct task_struct *tsk) { return ns_of_pid(task_pid(tsk)); diff --git a/kernel/pid_namespace.c b/kernel/pid_namespace.c index fab8ea86fac..821722ae58a 100644 --- a/kernel/pid_namespace.c +++ b/kernel/pid_namespace.c @@ -67,9 +67,10 @@ err_alloc: return NULL; } -static struct pid_namespace *create_pid_namespace(unsigned int level) +static struct pid_namespace *create_pid_namespace(struct pid_namespace *parent_pid_ns) { struct pid_namespace *ns; + unsigned int level = parent_pid_ns->level + 1; int i; ns = kmem_cache_zalloc(pid_ns_cachep, GFP_KERNEL); @@ -86,6 +87,7 @@ static struct pid_namespace *create_pid_namespace(unsigned int level) kref_init(&ns->kref); ns->level = level; + ns->parent = get_pid_ns(parent_pid_ns); set_bit(0, ns->pidmap[0].page); atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1); @@ -114,25 +116,11 @@ static void destroy_pid_namespace(struct pid_namespace *ns) struct pid_namespace *copy_pid_ns(unsigned long flags, struct pid_namespace *old_ns) { - struct pid_namespace *new_ns; - - BUG_ON(!old_ns); - new_ns = get_pid_ns(old_ns); if (!(flags & CLONE_NEWPID)) - goto out; - - new_ns = ERR_PTR(-EINVAL); + return get_pid_ns(old_ns); if (flags & CLONE_THREAD) - goto out_put; - - new_ns = create_pid_namespace(old_ns->level + 1); - if (!IS_ERR(new_ns)) - new_ns->parent = get_pid_ns(old_ns); - -out_put: - put_pid_ns(old_ns); -out: - return new_ns; + return ERR_PTR(-EINVAL); + return create_pid_namespace(old_ns); } void free_pid_ns(struct kref *kref) @@ -152,6 +140,7 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) { int nr; int rc; + struct task_struct *task; /* * The last thread in the cgroup-init thread group is terminating. @@ -169,7 +158,19 @@ void zap_pid_ns_processes(struct pid_namespace *pid_ns) read_lock(&tasklist_lock); nr = next_pidmap(pid_ns, 1); while (nr > 0) { - kill_proc_info(SIGKILL, SEND_SIG_PRIV, nr); + rcu_read_lock(); + + /* + * Use force_sig() since it clears SIGNAL_UNKILLABLE ensuring + * any nested-container's init processes don't ignore the + * signal + */ + task = pid_task(find_vpid(nr), PIDTYPE_PID); + if (task) + force_sig(SIGKILL, task); + + rcu_read_unlock(); + nr = next_pidmap(pid_ns, nr); } read_unlock(&tasklist_lock); diff --git a/kernel/posix-cpu-timers.c b/kernel/posix-cpu-timers.c index 157de3a4783..e33a21cb940 100644 --- a/kernel/posix-cpu-timers.c +++ b/kernel/posix-cpu-timers.c @@ -10,76 +10,6 @@ #include <linux/kernel_stat.h> /* - * Allocate the thread_group_cputime structure appropriately and fill in the - * current values of the fields. Called from copy_signal() via - * thread_group_cputime_clone_thread() when adding a second or subsequent - * thread to a thread group. Assumes interrupts are enabled when called. - */ -int thread_group_cputime_alloc(struct task_struct *tsk) -{ - struct signal_struct *sig = tsk->signal; - struct task_cputime *cputime; - - /* - * If we have multiple threads and we don't already have a - * per-CPU task_cputime struct (checked in the caller), allocate - * one and fill it in with the times accumulated so far. We may - * race with another thread so recheck after we pick up the sighand - * lock. - */ - cputime = alloc_percpu(struct task_cputime); - if (cputime == NULL) - return -ENOMEM; - spin_lock_irq(&tsk->sighand->siglock); - if (sig->cputime.totals) { - spin_unlock_irq(&tsk->sighand->siglock); - free_percpu(cputime); - return 0; - } - sig->cputime.totals = cputime; - cputime = per_cpu_ptr(sig->cputime.totals, smp_processor_id()); - cputime->utime = tsk->utime; - cputime->stime = tsk->stime; - cputime->sum_exec_runtime = tsk->se.sum_exec_runtime; - spin_unlock_irq(&tsk->sighand->siglock); - return 0; -} - -/** - * thread_group_cputime - Sum the thread group time fields across all CPUs. - * - * @tsk: The task we use to identify the thread group. - * @times: task_cputime structure in which we return the summed fields. - * - * Walk the list of CPUs to sum the per-CPU time fields in the thread group - * time structure. - */ -void thread_group_cputime( - struct task_struct *tsk, - struct task_cputime *times) -{ - struct task_cputime *totals, *tot; - int i; - - totals = tsk->signal->cputime.totals; - if (!totals) { - times->utime = tsk->utime; - times->stime = tsk->stime; - times->sum_exec_runtime = tsk->se.sum_exec_runtime; - return; - } - - times->stime = times->utime = cputime_zero; - times->sum_exec_runtime = 0; - for_each_possible_cpu(i) { - tot = per_cpu_ptr(totals, i); - times->utime = cputime_add(times->utime, tot->utime); - times->stime = cputime_add(times->stime, tot->stime); - times->sum_exec_runtime += tot->sum_exec_runtime; - } -} - -/* * Called after updating RLIMIT_CPU to set timer expiration if necessary. */ void update_rlimit_cpu(unsigned long rlim_new) @@ -88,7 +18,7 @@ void update_rlimit_cpu(unsigned long rlim_new) cputime = secs_to_cputime(rlim_new); if (cputime_eq(current->signal->it_prof_expires, cputime_zero) || - cputime_lt(current->signal->it_prof_expires, cputime)) { + cputime_gt(current->signal->it_prof_expires, cputime)) { spin_lock_irq(¤t->sighand->siglock); set_process_cpu_timer(current, CPUCLOCK_PROF, &cputime, NULL); spin_unlock_irq(¤t->sighand->siglock); @@ -294,12 +224,77 @@ static int cpu_clock_sample(const clockid_t which_clock, struct task_struct *p, cpu->cpu = virt_ticks(p); break; case CPUCLOCK_SCHED: - cpu->sched = p->se.sum_exec_runtime + task_delta_exec(p); + cpu->sched = task_sched_runtime(p); break; } return 0; } +void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times) +{ + struct sighand_struct *sighand; + struct signal_struct *sig; + struct task_struct *t; + + *times = INIT_CPUTIME; + + rcu_read_lock(); + sighand = rcu_dereference(tsk->sighand); + if (!sighand) + goto out; + + sig = tsk->signal; + + t = tsk; + do { + times->utime = cputime_add(times->utime, t->utime); + times->stime = cputime_add(times->stime, t->stime); + times->sum_exec_runtime += t->se.sum_exec_runtime; + + t = next_thread(t); + } while (t != tsk); + + times->utime = cputime_add(times->utime, sig->utime); + times->stime = cputime_add(times->stime, sig->stime); + times->sum_exec_runtime += sig->sum_sched_runtime; +out: + rcu_read_unlock(); +} + +static void update_gt_cputime(struct task_cputime *a, struct task_cputime *b) +{ + if (cputime_gt(b->utime, a->utime)) + a->utime = b->utime; + + if (cputime_gt(b->stime, a->stime)) + a->stime = b->stime; + + if (b->sum_exec_runtime > a->sum_exec_runtime) + a->sum_exec_runtime = b->sum_exec_runtime; +} + +void thread_group_cputimer(struct task_struct *tsk, struct task_cputime *times) +{ + struct thread_group_cputimer *cputimer = &tsk->signal->cputimer; + struct task_cputime sum; + unsigned long flags; + + spin_lock_irqsave(&cputimer->lock, flags); + if (!cputimer->running) { + cputimer->running = 1; + /* + * The POSIX timer interface allows for absolute time expiry + * values through the TIMER_ABSTIME flag, therefore we have + * to synchronize the timer to the clock every time we start + * it. + */ + thread_group_cputime(tsk, &sum); + update_gt_cputime(&cputimer->cputime, &sum); + } + *times = cputimer->cputime; + spin_unlock_irqrestore(&cputimer->lock, flags); +} + /* * Sample a process (thread group) clock for the given group_leader task. * Must be called with tasklist_lock held for reading. @@ -310,18 +305,19 @@ static int cpu_clock_sample_group(const clockid_t which_clock, { struct task_cputime cputime; - thread_group_cputime(p, &cputime); switch (CPUCLOCK_WHICH(which_clock)) { default: return -EINVAL; case CPUCLOCK_PROF: + thread_group_cputime(p, &cputime); cpu->cpu = cputime_add(cputime.utime, cputime.stime); break; case CPUCLOCK_VIRT: + thread_group_cputime(p, &cputime); cpu->cpu = cputime.utime; break; case CPUCLOCK_SCHED: - cpu->sched = cputime.sum_exec_runtime + task_delta_exec(p); + cpu->sched = thread_group_sched_runtime(p); break; } return 0; @@ -525,11 +521,12 @@ void posix_cpu_timers_exit(struct task_struct *tsk) } void posix_cpu_timers_exit_group(struct task_struct *tsk) { - struct task_cputime cputime; + struct signal_struct *const sig = tsk->signal; - thread_group_cputime(tsk, &cputime); cleanup_timers(tsk->signal->cpu_timers, - cputime.utime, cputime.stime, cputime.sum_exec_runtime); + cputime_add(tsk->utime, sig->utime), + cputime_add(tsk->stime, sig->stime), + tsk->se.sum_exec_runtime + sig->sum_sched_runtime); } static void clear_dead_task(struct k_itimer *timer, union cpu_time_count now) @@ -686,6 +683,33 @@ static void cpu_timer_fire(struct k_itimer *timer) } /* + * Sample a process (thread group) timer for the given group_leader task. + * Must be called with tasklist_lock held for reading. + */ +static int cpu_timer_sample_group(const clockid_t which_clock, + struct task_struct *p, + union cpu_time_count *cpu) +{ + struct task_cputime cputime; + + thread_group_cputimer(p, &cputime); + switch (CPUCLOCK_WHICH(which_clock)) { + default: + return -EINVAL; + case CPUCLOCK_PROF: + cpu->cpu = cputime_add(cputime.utime, cputime.stime); + break; + case CPUCLOCK_VIRT: + cpu->cpu = cputime.utime; + break; + case CPUCLOCK_SCHED: + cpu->sched = cputime.sum_exec_runtime + task_delta_exec(p); + break; + } + return 0; +} + +/* * Guts of sys_timer_settime for CPU timers. * This is called with the timer locked and interrupts disabled. * If we return TIMER_RETRY, it's necessary to release the timer's lock @@ -746,7 +770,7 @@ int posix_cpu_timer_set(struct k_itimer *timer, int flags, if (CPUCLOCK_PERTHREAD(timer->it_clock)) { cpu_clock_sample(timer->it_clock, p, &val); } else { - cpu_clock_sample_group(timer->it_clock, p, &val); + cpu_timer_sample_group(timer->it_clock, p, &val); } if (old) { @@ -894,7 +918,7 @@ void posix_cpu_timer_get(struct k_itimer *timer, struct itimerspec *itp) read_unlock(&tasklist_lock); goto dead; } else { - cpu_clock_sample_group(timer->it_clock, p, &now); + cpu_timer_sample_group(timer->it_clock, p, &now); clear_dead = (unlikely(p->exit_state) && thread_group_empty(p)); } @@ -1034,6 +1058,19 @@ static void check_thread_timers(struct task_struct *tsk, } } +static void stop_process_timers(struct task_struct *tsk) +{ + struct thread_group_cputimer *cputimer = &tsk->signal->cputimer; + unsigned long flags; + + if (!cputimer->running) + return; + + spin_lock_irqsave(&cputimer->lock, flags); + cputimer->running = 0; + spin_unlock_irqrestore(&cputimer->lock, flags); +} + /* * Check for any per-thread CPU timers that have fired and move them * off the tsk->*_timers list onto the firing list. Per-thread timers @@ -1057,13 +1094,15 @@ static void check_process_timers(struct task_struct *tsk, sig->rlim[RLIMIT_CPU].rlim_cur == RLIM_INFINITY && list_empty(&timers[CPUCLOCK_VIRT]) && cputime_eq(sig->it_virt_expires, cputime_zero) && - list_empty(&timers[CPUCLOCK_SCHED])) + list_empty(&timers[CPUCLOCK_SCHED])) { + stop_process_timers(tsk); return; + } /* * Collect the current process totals. */ - thread_group_cputime(tsk, &cputime); + thread_group_cputimer(tsk, &cputime); utime = cputime.utime; ptime = cputime_add(utime, cputime.stime); sum_sched_runtime = cputime.sum_exec_runtime; @@ -1234,7 +1273,7 @@ void posix_cpu_timer_schedule(struct k_itimer *timer) clear_dead_task(timer, now); goto out_unlock; } - cpu_clock_sample_group(timer->it_clock, p, &now); + cpu_timer_sample_group(timer->it_clock, p, &now); bump_cpu_timer(timer, now); /* Leave the tasklist_lock locked for the call below. */ } @@ -1329,11 +1368,12 @@ static inline int fastpath_timer_check(struct task_struct *tsk) if (!task_cputime_zero(&sig->cputime_expires)) { struct task_cputime group_sample; - thread_group_cputime(tsk, &group_sample); + thread_group_cputimer(tsk, &group_sample); if (task_cputime_expired(&group_sample, &sig->cputime_expires)) return 1; } - return 0; + + return sig->rlim[RLIMIT_CPU].rlim_cur != RLIM_INFINITY; } /* @@ -1381,19 +1421,19 @@ void run_posix_cpu_timers(struct task_struct *tsk) * timer call will interfere. */ list_for_each_entry_safe(timer, next, &firing, it.cpu.entry) { - int firing; + int cpu_firing; + spin_lock(&timer->it_lock); list_del_init(&timer->it.cpu.entry); - firing = timer->it.cpu.firing; + cpu_firing = timer->it.cpu.firing; timer->it.cpu.firing = 0; /* * The firing flag is -1 if we collided with a reset * of the timer, which already reported this * almost-firing as an overrun. So don't generate an event. */ - if (likely(firing >= 0)) { + if (likely(cpu_firing >= 0)) cpu_timer_fire(timer); - } spin_unlock(&timer->it_lock); } } @@ -1411,7 +1451,7 @@ void set_process_cpu_timer(struct task_struct *tsk, unsigned int clock_idx, struct list_head *head; BUG_ON(clock_idx == CPUCLOCK_SCHED); - cpu_clock_sample_group(clock_idx, tsk, &now); + cpu_timer_sample_group(clock_idx, tsk, &now); if (oldval) { if (!cputime_eq(*oldval, cputime_zero)) { diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c index 052ec4d195c..d089d052c4a 100644 --- a/kernel/posix-timers.c +++ b/kernel/posix-timers.c @@ -202,6 +202,12 @@ static int no_timer_create(struct k_itimer *new_timer) return -EOPNOTSUPP; } +static int no_nsleep(const clockid_t which_clock, int flags, + struct timespec *tsave, struct timespec __user *rmtp) +{ + return -EOPNOTSUPP; +} + /* * Return nonzero if we know a priori this clockid_t value is bogus. */ @@ -254,6 +260,7 @@ static __init int init_posix_timers(void) .clock_get = posix_get_monotonic_raw, .clock_set = do_posix_clock_nosettime, .timer_create = no_timer_create, + .nsleep = no_nsleep, }; register_posix_clock(CLOCK_REALTIME, &clock_realtime); diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 23bd4daeb96..72067cbdb37 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -116,9 +116,13 @@ config SUSPEND_FREEZER Turning OFF this setting is NOT recommended! If in doubt, say Y. +config HIBERNATION_NVS + bool + config HIBERNATION bool "Hibernation (aka 'suspend to disk')" depends on PM && SWAP && ARCH_HIBERNATION_POSSIBLE + select HIBERNATION_NVS if HAS_IOMEM ---help--- Enable the suspend to disk (STD) functionality, which is usually called "hibernation" in user interfaces. STD checkpoints the diff --git a/kernel/power/Makefile b/kernel/power/Makefile index d7a10167a25..c3b81c30e5d 100644 --- a/kernel/power/Makefile +++ b/kernel/power/Makefile @@ -3,9 +3,12 @@ ifeq ($(CONFIG_PM_DEBUG),y) EXTRA_CFLAGS += -DDEBUG endif -obj-y := main.o +obj-$(CONFIG_PM) += main.o obj-$(CONFIG_PM_SLEEP) += console.o obj-$(CONFIG_FREEZER) += process.o -obj-$(CONFIG_HIBERNATION) += swsusp.o disk.o snapshot.o swap.o user.o +obj-$(CONFIG_SUSPEND) += suspend.o +obj-$(CONFIG_PM_TEST_SUSPEND) += suspend_test.o +obj-$(CONFIG_HIBERNATION) += swsusp.o hibernate.o snapshot.o swap.o user.o +obj-$(CONFIG_HIBERNATION_NVS) += hibernate_nvs.o obj-$(CONFIG_MAGIC_SYSRQ) += poweroff.o diff --git a/kernel/power/console.c b/kernel/power/console.c index b8628be2a46..a3961b205de 100644 --- a/kernel/power/console.c +++ b/kernel/power/console.c @@ -78,6 +78,12 @@ void pm_restore_console(void) } set_console(orig_fgconsole); release_console_sem(); + + if (vt_waitactive(orig_fgconsole)) { + pr_debug("Resume: Can't switch VCs."); + return; + } + kmsg_redirect = orig_kmsg; } #endif diff --git a/kernel/power/disk.c b/kernel/power/hibernate.c index 45e8541ab7e..81d2e746489 100644 --- a/kernel/power/disk.c +++ b/kernel/power/hibernate.c @@ -1,12 +1,12 @@ /* - * kernel/power/disk.c - Suspend-to-disk support. + * kernel/power/hibernate.c - Hibernation (a.k.a suspend-to-disk) support. * * Copyright (c) 2003 Patrick Mochel * Copyright (c) 2003 Open Source Development Lab * Copyright (c) 2004 Pavel Machek <pavel@suse.cz> + * Copyright (c) 2009 Rafael J. Wysocki, Novell Inc. * * This file is released under the GPLv2. - * */ #include <linux/suspend.h> @@ -22,6 +22,8 @@ #include <linux/console.h> #include <linux/cpu.h> #include <linux/freezer.h> +#include <scsi/scsi_scan.h> +#include <asm/suspend.h> #include "power.h" @@ -71,6 +73,14 @@ void hibernation_set_ops(struct platform_hibernation_ops *ops) mutex_unlock(&pm_mutex); } +static bool entering_platform_hibernation; + +bool system_entering_hibernation(void) +{ + return entering_platform_hibernation; +} +EXPORT_SYMBOL(system_entering_hibernation); + #ifdef CONFIG_PM_DEBUG static void hibernation_debug_sleep(void) { @@ -205,18 +215,34 @@ static int create_image(int platform_mode) if (error) return error; - device_pm_lock(); - local_irq_disable(); - /* At this point, device_suspend() has been called, but *not* - * device_power_down(). We *must* call device_power_down() now. + /* At this point, dpm_suspend_start() has been called, but *not* + * dpm_suspend_noirq(). We *must* call dpm_suspend_noirq() now. * Otherwise, drivers for some devices (e.g. interrupt controllers) * become desynchronized with the actual state of the hardware * at resume time, and evil weirdness ensues. */ - error = device_power_down(PMSG_FREEZE); + error = dpm_suspend_noirq(PMSG_FREEZE); if (error) { printk(KERN_ERR "PM: Some devices failed to power down, " "aborting hibernation\n"); + return error; + } + + error = platform_pre_snapshot(platform_mode); + if (error || hibernation_test(TEST_PLATFORM)) + goto Platform_finish; + + error = disable_nonboot_cpus(); + if (error || hibernation_test(TEST_CPUS) + || hibernation_testmode(HIBERNATION_TEST)) + goto Enable_cpus; + + local_irq_disable(); + + error = sysdev_suspend(PMSG_FREEZE); + if (error) { + printk(KERN_ERR "PM: Some system devices failed to power down, " + "aborting hibernation\n"); goto Enable_irqs; } @@ -233,15 +259,25 @@ static int create_image(int platform_mode) restore_processor_state(); if (!in_suspend) platform_leave(platform_mode); + Power_up: - /* NOTE: device_power_up() is just a resume() for devices + sysdev_resume(); + /* NOTE: dpm_resume_noirq() is just a resume() for devices * that suspended with irqs off ... no overall powerup. */ - device_power_up(in_suspend ? - (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); + Enable_irqs: local_irq_enable(); - device_pm_unlock(); + + Enable_cpus: + enable_nonboot_cpus(); + + Platform_finish: + platform_finish(platform_mode); + + dpm_resume_noirq(in_suspend ? + (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); + return error; } @@ -249,7 +285,7 @@ static int create_image(int platform_mode) * hibernation_snapshot - quiesce devices and create the hibernation * snapshot image. * @platform_mode - if set, use the platform driver, if available, to - * prepare the platform frimware for the power transition. + * prepare the platform firmware for the power transition. * * Must be called with pm_mutex held */ @@ -268,34 +304,18 @@ int hibernation_snapshot(int platform_mode) goto Close; suspend_console(); - error = device_suspend(PMSG_FREEZE); + error = dpm_suspend_start(PMSG_FREEZE); if (error) goto Recover_platform; if (hibernation_test(TEST_DEVICES)) goto Recover_platform; - error = platform_pre_snapshot(platform_mode); - if (error || hibernation_test(TEST_PLATFORM)) - goto Finish; - - error = disable_nonboot_cpus(); - if (!error) { - if (hibernation_test(TEST_CPUS)) - goto Enable_cpus; - - if (hibernation_testmode(HIBERNATION_TEST)) - goto Enable_cpus; + error = create_image(platform_mode); + /* Control returns here after successful restore */ - error = create_image(platform_mode); - /* Control returns here after successful restore */ - } - Enable_cpus: - enable_nonboot_cpus(); - Finish: - platform_finish(platform_mode); Resume_devices: - device_resume(in_suspend ? + dpm_resume_end(in_suspend ? (error ? PMSG_RECOVER : PMSG_THAW) : PMSG_RESTORE); resume_console(); Close: @@ -315,18 +335,31 @@ int hibernation_snapshot(int platform_mode) * kernel. */ -static int resume_target_kernel(void) +static int resume_target_kernel(bool platform_mode) { int error; - device_pm_lock(); - local_irq_disable(); - error = device_power_down(PMSG_QUIESCE); + error = dpm_suspend_noirq(PMSG_QUIESCE); if (error) { printk(KERN_ERR "PM: Some devices failed to power down, " "aborting resume\n"); - goto Enable_irqs; + return error; } + + error = platform_pre_restore(platform_mode); + if (error) + goto Cleanup; + + error = disable_nonboot_cpus(); + if (error) + goto Enable_cpus; + + local_irq_disable(); + + error = sysdev_suspend(PMSG_QUIESCE); + if (error) + goto Enable_irqs; + /* We'll ignore saved state, but this gets preempt count (etc) right */ save_processor_state(); error = restore_highmem(); @@ -349,10 +382,20 @@ static int resume_target_kernel(void) swsusp_free(); restore_processor_state(); touch_softlockup_watchdog(); - device_power_up(PMSG_RECOVER); + + sysdev_resume(); + Enable_irqs: local_irq_enable(); - device_pm_unlock(); + + Enable_cpus: + enable_nonboot_cpus(); + + Cleanup: + platform_restore_cleanup(platform_mode); + + dpm_resume_noirq(PMSG_RECOVER); + return error; } @@ -360,7 +403,7 @@ static int resume_target_kernel(void) * hibernation_restore - quiesce devices and restore the hibernation * snapshot image. If successful, control returns in hibernation_snaphot() * @platform_mode - if set, use the platform driver, if available, to - * prepare the platform frimware for the transition. + * prepare the platform firmware for the transition. * * Must be called with pm_mutex held */ @@ -371,20 +414,11 @@ int hibernation_restore(int platform_mode) pm_prepare_console(); suspend_console(); - error = device_suspend(PMSG_QUIESCE); - if (error) - goto Finish; - - error = platform_pre_restore(platform_mode); + error = dpm_suspend_start(PMSG_QUIESCE); if (!error) { - error = disable_nonboot_cpus(); - if (!error) - error = resume_target_kernel(); - enable_nonboot_cpus(); + error = resume_target_kernel(platform_mode); + dpm_resume_end(PMSG_RECOVER); } - platform_restore_cleanup(platform_mode); - device_resume(PMSG_RECOVER); - Finish: resume_console(); pm_restore_console(); return error; @@ -411,44 +445,50 @@ int hibernation_platform_enter(void) if (error) goto Close; + entering_platform_hibernation = true; suspend_console(); - error = device_suspend(PMSG_HIBERNATE); + error = dpm_suspend_start(PMSG_HIBERNATE); if (error) { if (hibernation_ops->recover) hibernation_ops->recover(); goto Resume_devices; } - error = hibernation_ops->prepare(); + error = dpm_suspend_noirq(PMSG_HIBERNATE); if (error) goto Resume_devices; + error = hibernation_ops->prepare(); + if (error) + goto Platofrm_finish; + error = disable_nonboot_cpus(); if (error) - goto Finish; + goto Platofrm_finish; - device_pm_lock(); local_irq_disable(); - error = device_power_down(PMSG_HIBERNATE); - if (!error) { - hibernation_ops->enter(); - /* We should never get here */ - while (1); - } - local_irq_enable(); - device_pm_unlock(); + sysdev_suspend(PMSG_HIBERNATE); + hibernation_ops->enter(); + /* We should never get here */ + while (1); /* * We don't need to reenable the nonboot CPUs or resume consoles, since * the system is going to be halted anyway. */ - Finish: + Platofrm_finish: hibernation_ops->finish(); + + dpm_suspend_noirq(PMSG_RESTORE); + Resume_devices: - device_resume(PMSG_RESTORE); + entering_platform_hibernation = false; + dpm_resume_end(PMSG_RESTORE); resume_console(); + Close: hibernation_ops->end(); + return error; } @@ -585,6 +625,12 @@ static int software_resume(void) unsigned int flags; /* + * If the user said "noresume".. bail out early. + */ + if (noresume) + return 0; + + /* * name_to_dev_t() below takes a sysfs buffer mutex when sysfs * is configured into the kernel. Since the regular hibernate * trigger path is via sysfs which takes a buffer mutex before @@ -595,28 +641,43 @@ static int software_resume(void) * here to avoid lockdep complaining. */ mutex_lock_nested(&pm_mutex, SINGLE_DEPTH_NESTING); - if (!swsusp_resume_device) { - if (!strlen(resume_file)) { - mutex_unlock(&pm_mutex); - return -ENOENT; - } - swsusp_resume_device = name_to_dev_t(resume_file); - pr_debug("PM: Resume from partition %s\n", resume_file); - } else { - pr_debug("PM: Resume from partition %d:%d\n", - MAJOR(swsusp_resume_device), - MINOR(swsusp_resume_device)); + + if (swsusp_resume_device) + goto Check_image; + + if (!strlen(resume_file)) { + error = -ENOENT; + goto Unlock; } - if (noresume) { - /** - * FIXME: If noresume is specified, we need to find the - * partition and reset it back to normal swap space. + pr_debug("PM: Checking image partition %s\n", resume_file); + + /* Check if the device is there */ + swsusp_resume_device = name_to_dev_t(resume_file); + if (!swsusp_resume_device) { + /* + * Some device discovery might still be in progress; we need + * to wait for this to finish. */ - mutex_unlock(&pm_mutex); - return 0; + wait_for_device_probe(); + /* + * We can't depend on SCSI devices being available after loading + * one of their modules until scsi_complete_async_scans() is + * called and the resume device usually is a SCSI one. + */ + scsi_complete_async_scans(); + + swsusp_resume_device = name_to_dev_t(resume_file); + if (!swsusp_resume_device) { + error = -ENODEV; + goto Unlock; + } } + Check_image: + pr_debug("PM: Resume from partition %d:%d\n", + MAJOR(swsusp_resume_device), MINOR(swsusp_resume_device)); + pr_debug("PM: Checking hibernation image.\n"); error = swsusp_check(); if (error) diff --git a/kernel/power/hibernate_nvs.c b/kernel/power/hibernate_nvs.c new file mode 100644 index 00000000000..39ac698ef83 --- /dev/null +++ b/kernel/power/hibernate_nvs.c @@ -0,0 +1,135 @@ +/* + * linux/kernel/power/hibernate_nvs.c - Routines for handling NVS memory + * + * Copyright (C) 2008,2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc. + * + * This file is released under the GPLv2. + */ + +#include <linux/io.h> +#include <linux/kernel.h> +#include <linux/list.h> +#include <linux/mm.h> +#include <linux/suspend.h> + +/* + * Platforms, like ACPI, may want us to save some memory used by them during + * hibernation and to restore the contents of this memory during the subsequent + * resume. The code below implements a mechanism allowing us to do that. + */ + +struct nvs_page { + unsigned long phys_start; + unsigned int size; + void *kaddr; + void *data; + struct list_head node; +}; + +static LIST_HEAD(nvs_list); + +/** + * hibernate_nvs_register - register platform NVS memory region to save + * @start - physical address of the region + * @size - size of the region + * + * The NVS region need not be page-aligned (both ends) and we arrange + * things so that the data from page-aligned addresses in this region will + * be copied into separate RAM pages. + */ +int hibernate_nvs_register(unsigned long start, unsigned long size) +{ + struct nvs_page *entry, *next; + + while (size > 0) { + unsigned int nr_bytes; + + entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL); + if (!entry) + goto Error; + + list_add_tail(&entry->node, &nvs_list); + entry->phys_start = start; + nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK); + entry->size = (size < nr_bytes) ? size : nr_bytes; + + start += entry->size; + size -= entry->size; + } + return 0; + + Error: + list_for_each_entry_safe(entry, next, &nvs_list, node) { + list_del(&entry->node); + kfree(entry); + } + return -ENOMEM; +} + +/** + * hibernate_nvs_free - free data pages allocated for saving NVS regions + */ +void hibernate_nvs_free(void) +{ + struct nvs_page *entry; + + list_for_each_entry(entry, &nvs_list, node) + if (entry->data) { + free_page((unsigned long)entry->data); + entry->data = NULL; + if (entry->kaddr) { + iounmap(entry->kaddr); + entry->kaddr = NULL; + } + } +} + +/** + * hibernate_nvs_alloc - allocate memory necessary for saving NVS regions + */ +int hibernate_nvs_alloc(void) +{ + struct nvs_page *entry; + + list_for_each_entry(entry, &nvs_list, node) { + entry->data = (void *)__get_free_page(GFP_KERNEL); + if (!entry->data) { + hibernate_nvs_free(); + return -ENOMEM; + } + } + return 0; +} + +/** + * hibernate_nvs_save - save NVS memory regions + */ +void hibernate_nvs_save(void) +{ + struct nvs_page *entry; + + printk(KERN_INFO "PM: Saving platform NVS memory\n"); + + list_for_each_entry(entry, &nvs_list, node) + if (entry->data) { + entry->kaddr = ioremap(entry->phys_start, entry->size); + memcpy(entry->data, entry->kaddr, entry->size); + } +} + +/** + * hibernate_nvs_restore - restore NVS memory regions + * + * This function is going to be called with interrupts disabled, so it + * cannot iounmap the virtual addresses used to access the NVS region. + */ +void hibernate_nvs_restore(void) +{ + struct nvs_page *entry; + + printk(KERN_INFO "PM: Restoring platform NVS memory\n"); + + list_for_each_entry(entry, &nvs_list, node) + if (entry->data) + memcpy(entry->kaddr, entry->data, entry->size); +} diff --git a/kernel/power/main.c b/kernel/power/main.c index 23998887397..f710e36930c 100644 --- a/kernel/power/main.c +++ b/kernel/power/main.c @@ -8,20 +8,9 @@ * */ -#include <linux/module.h> -#include <linux/suspend.h> #include <linux/kobject.h> #include <linux/string.h> -#include <linux/delay.h> -#include <linux/errno.h> -#include <linux/kmod.h> -#include <linux/init.h> -#include <linux/console.h> -#include <linux/cpu.h> #include <linux/resume-trace.h> -#include <linux/freezer.h> -#include <linux/vmstat.h> -#include <linux/syscalls.h> #include "power.h" @@ -57,16 +46,6 @@ int pm_notifier_call_chain(unsigned long val) #ifdef CONFIG_PM_DEBUG int pm_test_level = TEST_NONE; -static int suspend_test(int level) -{ - if (pm_test_level == level) { - printk(KERN_INFO "suspend debug: Waiting for 5 seconds.\n"); - mdelay(5000); - return 1; - } - return 0; -} - static const char * const pm_tests[__TEST_AFTER_LAST] = { [TEST_NONE] = "none", [TEST_CORE] = "core", @@ -125,347 +104,10 @@ static ssize_t pm_test_store(struct kobject *kobj, struct kobj_attribute *attr, } power_attr(pm_test); -#else /* !CONFIG_PM_DEBUG */ -static inline int suspend_test(int level) { return 0; } -#endif /* !CONFIG_PM_DEBUG */ +#endif /* CONFIG_PM_DEBUG */ #endif /* CONFIG_PM_SLEEP */ -#ifdef CONFIG_SUSPEND - -#ifdef CONFIG_PM_TEST_SUSPEND - -/* - * We test the system suspend code by setting an RTC wakealarm a short - * time in the future, then suspending. Suspending the devices won't - * normally take long ... some systems only need a few milliseconds. - * - * The time it takes is system-specific though, so when we test this - * during system bootup we allow a LOT of time. - */ -#define TEST_SUSPEND_SECONDS 5 - -static unsigned long suspend_test_start_time; - -static void suspend_test_start(void) -{ - /* FIXME Use better timebase than "jiffies", ideally a clocksource. - * What we want is a hardware counter that will work correctly even - * during the irqs-are-off stages of the suspend/resume cycle... - */ - suspend_test_start_time = jiffies; -} - -static void suspend_test_finish(const char *label) -{ - long nj = jiffies - suspend_test_start_time; - unsigned msec; - - msec = jiffies_to_msecs(abs(nj)); - pr_info("PM: %s took %d.%03d seconds\n", label, - msec / 1000, msec % 1000); - - /* Warning on suspend means the RTC alarm period needs to be - * larger -- the system was sooo slooowwww to suspend that the - * alarm (should have) fired before the system went to sleep! - * - * Warning on either suspend or resume also means the system - * has some performance issues. The stack dump of a WARN_ON - * is more likely to get the right attention than a printk... - */ - WARN(msec > (TEST_SUSPEND_SECONDS * 1000), "Component: %s\n", label); -} - -#else - -static void suspend_test_start(void) -{ -} - -static void suspend_test_finish(const char *label) -{ -} - -#endif - -/* This is just an arbitrary number */ -#define FREE_PAGE_NUMBER (100) - -static struct platform_suspend_ops *suspend_ops; - -/** - * suspend_set_ops - Set the global suspend method table. - * @ops: Pointer to ops structure. - */ - -void suspend_set_ops(struct platform_suspend_ops *ops) -{ - mutex_lock(&pm_mutex); - suspend_ops = ops; - mutex_unlock(&pm_mutex); -} - -/** - * suspend_valid_only_mem - generic memory-only valid callback - * - * Platform drivers that implement mem suspend only and only need - * to check for that in their .valid callback can use this instead - * of rolling their own .valid callback. - */ -int suspend_valid_only_mem(suspend_state_t state) -{ - return state == PM_SUSPEND_MEM; -} - -/** - * suspend_prepare - Do prep work before entering low-power state. - * - * This is common code that is called for each state that we're entering. - * Run suspend notifiers, allocate a console and stop all processes. - */ -static int suspend_prepare(void) -{ - int error; - unsigned int free_pages; - - if (!suspend_ops || !suspend_ops->enter) - return -EPERM; - - pm_prepare_console(); - - error = pm_notifier_call_chain(PM_SUSPEND_PREPARE); - if (error) - goto Finish; - - error = usermodehelper_disable(); - if (error) - goto Finish; - - if (suspend_freeze_processes()) { - error = -EAGAIN; - goto Thaw; - } - - free_pages = global_page_state(NR_FREE_PAGES); - if (free_pages < FREE_PAGE_NUMBER) { - pr_debug("PM: free some memory\n"); - shrink_all_memory(FREE_PAGE_NUMBER - free_pages); - if (nr_free_pages() < FREE_PAGE_NUMBER) { - error = -ENOMEM; - printk(KERN_ERR "PM: No enough memory\n"); - } - } - if (!error) - return 0; - - Thaw: - suspend_thaw_processes(); - usermodehelper_enable(); - Finish: - pm_notifier_call_chain(PM_POST_SUSPEND); - pm_restore_console(); - return error; -} - -/* default implementation */ -void __attribute__ ((weak)) arch_suspend_disable_irqs(void) -{ - local_irq_disable(); -} - -/* default implementation */ -void __attribute__ ((weak)) arch_suspend_enable_irqs(void) -{ - local_irq_enable(); -} - -/** - * suspend_enter - enter the desired system sleep state. - * @state: state to enter - * - * This function should be called after devices have been suspended. - */ -static int suspend_enter(suspend_state_t state) -{ - int error = 0; - - device_pm_lock(); - arch_suspend_disable_irqs(); - BUG_ON(!irqs_disabled()); - - if ((error = device_power_down(PMSG_SUSPEND))) { - printk(KERN_ERR "PM: Some devices failed to power down\n"); - goto Done; - } - - if (!suspend_test(TEST_CORE)) - error = suspend_ops->enter(state); - - device_power_up(PMSG_RESUME); - Done: - arch_suspend_enable_irqs(); - BUG_ON(irqs_disabled()); - device_pm_unlock(); - return error; -} - -/** - * suspend_devices_and_enter - suspend devices and enter the desired system - * sleep state. - * @state: state to enter - */ -int suspend_devices_and_enter(suspend_state_t state) -{ - int error; - - if (!suspend_ops) - return -ENOSYS; - - if (suspend_ops->begin) { - error = suspend_ops->begin(state); - if (error) - goto Close; - } - suspend_console(); - suspend_test_start(); - error = device_suspend(PMSG_SUSPEND); - if (error) { - printk(KERN_ERR "PM: Some devices failed to suspend\n"); - goto Recover_platform; - } - suspend_test_finish("suspend devices"); - if (suspend_test(TEST_DEVICES)) - goto Recover_platform; - - if (suspend_ops->prepare) { - error = suspend_ops->prepare(); - if (error) - goto Resume_devices; - } - - if (suspend_test(TEST_PLATFORM)) - goto Finish; - - error = disable_nonboot_cpus(); - if (!error && !suspend_test(TEST_CPUS)) - suspend_enter(state); - - enable_nonboot_cpus(); - Finish: - if (suspend_ops->finish) - suspend_ops->finish(); - Resume_devices: - suspend_test_start(); - device_resume(PMSG_RESUME); - suspend_test_finish("resume devices"); - resume_console(); - Close: - if (suspend_ops->end) - suspend_ops->end(); - return error; - - Recover_platform: - if (suspend_ops->recover) - suspend_ops->recover(); - goto Resume_devices; -} - -/** - * suspend_finish - Do final work before exiting suspend sequence. - * - * Call platform code to clean up, restart processes, and free the - * console that we've allocated. This is not called for suspend-to-disk. - */ -static void suspend_finish(void) -{ - suspend_thaw_processes(); - usermodehelper_enable(); - pm_notifier_call_chain(PM_POST_SUSPEND); - pm_restore_console(); -} - - - - -static const char * const pm_states[PM_SUSPEND_MAX] = { - [PM_SUSPEND_STANDBY] = "standby", - [PM_SUSPEND_MEM] = "mem", -}; - -static inline int valid_state(suspend_state_t state) -{ - /* All states need lowlevel support and need to be valid - * to the lowlevel implementation, no valid callback - * implies that none are valid. */ - if (!suspend_ops || !suspend_ops->valid || !suspend_ops->valid(state)) - return 0; - return 1; -} - - -/** - * enter_state - Do common work of entering low-power state. - * @state: pm_state structure for state we're entering. - * - * Make sure we're the only ones trying to enter a sleep state. Fail - * if someone has beat us to it, since we don't want anything weird to - * happen when we wake up. - * Then, do the setup for suspend, enter the state, and cleaup (after - * we've woken up). - */ -static int enter_state(suspend_state_t state) -{ - int error; - - if (!valid_state(state)) - return -ENODEV; - - if (!mutex_trylock(&pm_mutex)) - return -EBUSY; - - printk(KERN_INFO "PM: Syncing filesystems ... "); - sys_sync(); - printk("done.\n"); - - pr_debug("PM: Preparing system for %s sleep\n", pm_states[state]); - error = suspend_prepare(); - if (error) - goto Unlock; - - if (suspend_test(TEST_FREEZER)) - goto Finish; - - pr_debug("PM: Entering %s sleep\n", pm_states[state]); - error = suspend_devices_and_enter(state); - - Finish: - pr_debug("PM: Finishing wakeup.\n"); - suspend_finish(); - Unlock: - mutex_unlock(&pm_mutex); - return error; -} - - -/** - * pm_suspend - Externally visible function for suspending system. - * @state: Enumerated value of state to enter. - * - * Determine whether or not value is within range, get state - * structure, and enter (above). - */ - -int pm_suspend(suspend_state_t state) -{ - if (state > PM_SUSPEND_ON && state <= PM_SUSPEND_MAX) - return enter_state(state); - return -EINVAL; -} - -EXPORT_SYMBOL(pm_suspend); - -#endif /* CONFIG_SUSPEND */ - struct kobject *power_kobj; /** @@ -478,7 +120,6 @@ struct kobject *power_kobj; * store() accepts one of those strings, translates it into the * proper enumerated value, and initiates a suspend transition. */ - static ssize_t state_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { @@ -576,7 +217,6 @@ static struct attribute_group attr_group = { .attrs = g, }; - static int __init pm_init(void) { power_kobj = kobject_create_and_add("power", NULL); @@ -586,144 +226,3 @@ static int __init pm_init(void) } core_initcall(pm_init); - - -#ifdef CONFIG_PM_TEST_SUSPEND - -#include <linux/rtc.h> - -/* - * To test system suspend, we need a hands-off mechanism to resume the - * system. RTCs wake alarms are a common self-contained mechanism. - */ - -static void __init test_wakealarm(struct rtc_device *rtc, suspend_state_t state) -{ - static char err_readtime[] __initdata = - KERN_ERR "PM: can't read %s time, err %d\n"; - static char err_wakealarm [] __initdata = - KERN_ERR "PM: can't set %s wakealarm, err %d\n"; - static char err_suspend[] __initdata = - KERN_ERR "PM: suspend test failed, error %d\n"; - static char info_test[] __initdata = - KERN_INFO "PM: test RTC wakeup from '%s' suspend\n"; - - unsigned long now; - struct rtc_wkalrm alm; - int status; - - /* this may fail if the RTC hasn't been initialized */ - status = rtc_read_time(rtc, &alm.time); - if (status < 0) { - printk(err_readtime, dev_name(&rtc->dev), status); - return; - } - rtc_tm_to_time(&alm.time, &now); - - memset(&alm, 0, sizeof alm); - rtc_time_to_tm(now + TEST_SUSPEND_SECONDS, &alm.time); - alm.enabled = true; - - status = rtc_set_alarm(rtc, &alm); - if (status < 0) { - printk(err_wakealarm, dev_name(&rtc->dev), status); - return; - } - - if (state == PM_SUSPEND_MEM) { - printk(info_test, pm_states[state]); - status = pm_suspend(state); - if (status == -ENODEV) - state = PM_SUSPEND_STANDBY; - } - if (state == PM_SUSPEND_STANDBY) { - printk(info_test, pm_states[state]); - status = pm_suspend(state); - } - if (status < 0) - printk(err_suspend, status); - - /* Some platforms can't detect that the alarm triggered the - * wakeup, or (accordingly) disable it after it afterwards. - * It's supposed to give oneshot behavior; cope. - */ - alm.enabled = false; - rtc_set_alarm(rtc, &alm); -} - -static int __init has_wakealarm(struct device *dev, void *name_ptr) -{ - struct rtc_device *candidate = to_rtc_device(dev); - - if (!candidate->ops->set_alarm) - return 0; - if (!device_may_wakeup(candidate->dev.parent)) - return 0; - - *(const char **)name_ptr = dev_name(dev); - return 1; -} - -/* - * Kernel options like "test_suspend=mem" force suspend/resume sanity tests - * at startup time. They're normally disabled, for faster boot and because - * we can't know which states really work on this particular system. - */ -static suspend_state_t test_state __initdata = PM_SUSPEND_ON; - -static char warn_bad_state[] __initdata = - KERN_WARNING "PM: can't test '%s' suspend state\n"; - -static int __init setup_test_suspend(char *value) -{ - unsigned i; - - /* "=mem" ==> "mem" */ - value++; - for (i = 0; i < PM_SUSPEND_MAX; i++) { - if (!pm_states[i]) - continue; - if (strcmp(pm_states[i], value) != 0) - continue; - test_state = (__force suspend_state_t) i; - return 0; - } - printk(warn_bad_state, value); - return 0; -} -__setup("test_suspend", setup_test_suspend); - -static int __init test_suspend(void) -{ - static char warn_no_rtc[] __initdata = - KERN_WARNING "PM: no wakealarm-capable RTC driver is ready\n"; - - char *pony = NULL; - struct rtc_device *rtc = NULL; - - /* PM is initialized by now; is that state testable? */ - if (test_state == PM_SUSPEND_ON) - goto done; - if (!valid_state(test_state)) { - printk(warn_bad_state, pm_states[test_state]); - goto done; - } - - /* RTCs have initialized by now too ... can we use one? */ - class_find_device(rtc_class, NULL, &pony, has_wakealarm); - if (pony) - rtc = rtc_class_open(pony); - if (!rtc) { - printk(warn_no_rtc); - goto done; - } - - /* go for it */ - test_wakealarm(rtc, test_state); - rtc_class_close(rtc); -done: - return 0; -} -late_initcall(test_suspend); - -#endif /* CONFIG_PM_TEST_SUSPEND */ diff --git a/kernel/power/power.h b/kernel/power/power.h index 46b5ec7a3af..26d5a26f82e 100644 --- a/kernel/power/power.h +++ b/kernel/power/power.h @@ -45,7 +45,7 @@ static inline char *check_image_kernel(struct swsusp_info *info) */ #define SPARE_PAGES ((1024 * 1024) >> PAGE_SHIFT) -/* kernel/power/disk.c */ +/* kernel/power/hibernate.c */ extern int hibernation_snapshot(int platform_mode); extern int hibernation_restore(int platform_mode); extern int hibernation_platform_enter(void); @@ -74,7 +74,7 @@ extern asmlinkage int swsusp_arch_resume(void); extern int create_basic_memory_bitmaps(void); extern void free_basic_memory_bitmaps(void); -extern unsigned int count_data_pages(void); +extern int swsusp_shrink_memory(void); /** * Auxiliary structure used for reading the snapshot image data and @@ -147,9 +147,8 @@ extern int swsusp_swap_in_use(void); */ #define SF_PLATFORM_MODE 1 -/* kernel/power/disk.c */ +/* kernel/power/hibernate.c */ extern int swsusp_check(void); -extern int swsusp_shrink_memory(void); extern void swsusp_free(void); extern int swsusp_read(unsigned int *flags_p); extern int swsusp_write(unsigned int flags); @@ -161,22 +160,36 @@ extern void swsusp_show_speed(struct timeval *, struct timeval *, unsigned int, char *); #ifdef CONFIG_SUSPEND -/* kernel/power/main.c */ +/* kernel/power/suspend.c */ +extern const char *const pm_states[]; + +extern bool valid_state(suspend_state_t state); extern int suspend_devices_and_enter(suspend_state_t state); +extern int enter_state(suspend_state_t state); #else /* !CONFIG_SUSPEND */ static inline int suspend_devices_and_enter(suspend_state_t state) { return -ENOSYS; } +static inline int enter_state(suspend_state_t state) { return -ENOSYS; } +static inline bool valid_state(suspend_state_t state) { return false; } #endif /* !CONFIG_SUSPEND */ +#ifdef CONFIG_PM_TEST_SUSPEND +/* kernel/power/suspend_test.c */ +extern void suspend_test_start(void); +extern void suspend_test_finish(const char *label); +#else /* !CONFIG_PM_TEST_SUSPEND */ +static inline void suspend_test_start(void) {} +static inline void suspend_test_finish(const char *label) {} +#endif /* !CONFIG_PM_TEST_SUSPEND */ + #ifdef CONFIG_PM_SLEEP /* kernel/power/main.c */ extern int pm_notifier_call_chain(unsigned long val); #endif #ifdef CONFIG_HIGHMEM -unsigned int count_highmem_pages(void); int restore_highmem(void); #else static inline unsigned int count_highmem_pages(void) { return 0; } diff --git a/kernel/power/poweroff.c b/kernel/power/poweroff.c index 97890831e1b..e8b33700627 100644 --- a/kernel/power/poweroff.c +++ b/kernel/power/poweroff.c @@ -34,7 +34,7 @@ static struct sysrq_key_op sysrq_poweroff_op = { .handler = handle_poweroff, .help_msg = "powerOff", .action_msg = "Power Off", - .enable_mask = SYSRQ_ENABLE_BOOT, + .enable_mask = SYSRQ_ENABLE_BOOT, }; static int pm_sysrq_init(void) diff --git a/kernel/power/process.c b/kernel/power/process.c index ca634019497..da2072d7381 100644 --- a/kernel/power/process.c +++ b/kernel/power/process.c @@ -117,9 +117,12 @@ int freeze_processes(void) if (error) goto Exit; printk("done."); + + oom_killer_disable(); Exit: BUG_ON(in_atomic()); printk("\n"); + return error; } @@ -145,6 +148,8 @@ static void thaw_tasks(bool nosig_only) void thaw_processes(void) { + oom_killer_enable(); + printk("Restarting tasks ... "); thaw_tasks(true); thaw_tasks(false); diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index f5fc2d7680f..523a451b45d 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -39,6 +39,14 @@ static int swsusp_page_is_free(struct page *); static void swsusp_set_page_forbidden(struct page *); static void swsusp_unset_page_forbidden(struct page *); +/* + * Preferred image size in bytes (tunable via /sys/power/image_size). + * When it is set to N, swsusp will do its best to ensure the image + * size will not exceed N bytes, but if that is impossible, it will + * try to create the smallest image possible. + */ +unsigned long image_size = 500 * 1024 * 1024; + /* List of PBEs needed for restoring the pages that were allocated before * the suspend and included in the suspend image, but have also been * allocated by the "resume" kernel, so their contents cannot be written @@ -321,13 +329,10 @@ static int create_mem_extents(struct list_head *list, gfp_t gfp_mask) INIT_LIST_HEAD(list); - for_each_zone(zone) { + for_each_populated_zone(zone) { unsigned long zone_start, zone_end; struct mem_extent *ext, *cur, *aux; - if (!populated_zone(zone)) - continue; - zone_start = zone->zone_start_pfn; zone_end = zone->zone_start_pfn + zone->spanned_pages; @@ -804,8 +809,8 @@ static unsigned int count_free_highmem_pages(void) struct zone *zone; unsigned int cnt = 0; - for_each_zone(zone) - if (populated_zone(zone) && is_highmem(zone)) + for_each_populated_zone(zone) + if (is_highmem(zone)) cnt += zone_page_state(zone, NR_FREE_PAGES); return cnt; @@ -843,7 +848,7 @@ static struct page *saveable_highmem_page(struct zone *zone, unsigned long pfn) * pages. */ -unsigned int count_highmem_pages(void) +static unsigned int count_highmem_pages(void) { struct zone *zone; unsigned int n = 0; @@ -905,7 +910,7 @@ static struct page *saveable_page(struct zone *zone, unsigned long pfn) * pages. */ -unsigned int count_data_pages(void) +static unsigned int count_data_pages(void) { struct zone *zone; unsigned long pfn, max_zone_pfn; @@ -1061,6 +1066,74 @@ void swsusp_free(void) buffer = NULL; } +/** + * swsusp_shrink_memory - Try to free as much memory as needed + * + * ... but do not OOM-kill anyone + * + * Notice: all userland should be stopped before it is called, or + * livelock is possible. + */ + +#define SHRINK_BITE 10000 +static inline unsigned long __shrink_memory(long tmp) +{ + if (tmp > SHRINK_BITE) + tmp = SHRINK_BITE; + return shrink_all_memory(tmp); +} + +int swsusp_shrink_memory(void) +{ + long tmp; + struct zone *zone; + unsigned long pages = 0; + unsigned int i = 0; + char *p = "-\\|/"; + struct timeval start, stop; + + printk(KERN_INFO "PM: Shrinking memory... "); + do_gettimeofday(&start); + do { + long size, highmem_size; + + highmem_size = count_highmem_pages(); + size = count_data_pages() + PAGES_FOR_IO + SPARE_PAGES; + tmp = size; + size += highmem_size; + for_each_populated_zone(zone) { + tmp += snapshot_additional_pages(zone); + if (is_highmem(zone)) { + highmem_size -= + zone_page_state(zone, NR_FREE_PAGES); + } else { + tmp -= zone_page_state(zone, NR_FREE_PAGES); + tmp += zone->lowmem_reserve[ZONE_NORMAL]; + } + } + + if (highmem_size < 0) + highmem_size = 0; + + tmp += highmem_size; + if (tmp > 0) { + tmp = __shrink_memory(tmp); + if (!tmp) + return -ENOMEM; + pages += tmp; + } else if (size > image_size / PAGE_SIZE) { + tmp = __shrink_memory(size - (image_size / PAGE_SIZE)); + pages += tmp; + } + printk("\b%c", p[i++%4]); + } while (tmp > 0); + do_gettimeofday(&stop); + printk("\bdone (%lu pages freed)\n", pages); + swsusp_show_speed(&start, &stop, pages, "Freed"); + + return 0; +} + #ifdef CONFIG_HIGHMEM /** * count_pages_for_highmem - compute the number of non-highmem pages diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c new file mode 100644 index 00000000000..6f10dfc2d3e --- /dev/null +++ b/kernel/power/suspend.c @@ -0,0 +1,300 @@ +/* + * kernel/power/suspend.c - Suspend to RAM and standby functionality. + * + * Copyright (c) 2003 Patrick Mochel + * Copyright (c) 2003 Open Source Development Lab + * Copyright (c) 2009 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc. + * + * This file is released under the GPLv2. + */ + +#include <linux/string.h> +#include <linux/delay.h> +#include <linux/errno.h> +#include <linux/init.h> +#include <linux/console.h> +#include <linux/cpu.h> +#include <linux/syscalls.h> + +#include "power.h" + +const char *const pm_states[PM_SUSPEND_MAX] = { + [PM_SUSPEND_STANDBY] = "standby", + [PM_SUSPEND_MEM] = "mem", +}; + +static struct platform_suspend_ops *suspend_ops; + +/** + * suspend_set_ops - Set the global suspend method table. + * @ops: Pointer to ops structure. + */ +void suspend_set_ops(struct platform_suspend_ops *ops) +{ + mutex_lock(&pm_mutex); + suspend_ops = ops; + mutex_unlock(&pm_mutex); +} + +bool valid_state(suspend_state_t state) +{ + /* + * All states need lowlevel support and need to be valid to the lowlevel + * implementation, no valid callback implies that none are valid. + */ + return suspend_ops && suspend_ops->valid && suspend_ops->valid(state); +} + +/** + * suspend_valid_only_mem - generic memory-only valid callback + * + * Platform drivers that implement mem suspend only and only need + * to check for that in their .valid callback can use this instead + * of rolling their own .valid callback. + */ +int suspend_valid_only_mem(suspend_state_t state) +{ + return state == PM_SUSPEND_MEM; +} + +static int suspend_test(int level) +{ +#ifdef CONFIG_PM_DEBUG + if (pm_test_level == level) { + printk(KERN_INFO "suspend debug: Waiting for 5 seconds.\n"); + mdelay(5000); + return 1; + } +#endif /* !CONFIG_PM_DEBUG */ + return 0; +} + +/** + * suspend_prepare - Do prep work before entering low-power state. + * + * This is common code that is called for each state that we're entering. + * Run suspend notifiers, allocate a console and stop all processes. + */ +static int suspend_prepare(void) +{ + int error; + + if (!suspend_ops || !suspend_ops->enter) + return -EPERM; + + pm_prepare_console(); + + error = pm_notifier_call_chain(PM_SUSPEND_PREPARE); + if (error) + goto Finish; + + error = usermodehelper_disable(); + if (error) + goto Finish; + + error = suspend_freeze_processes(); + if (!error) + return 0; + + suspend_thaw_processes(); + usermodehelper_enable(); + Finish: + pm_notifier_call_chain(PM_POST_SUSPEND); + pm_restore_console(); + return error; +} + +/* default implementation */ +void __attribute__ ((weak)) arch_suspend_disable_irqs(void) +{ + local_irq_disable(); +} + +/* default implementation */ +void __attribute__ ((weak)) arch_suspend_enable_irqs(void) +{ + local_irq_enable(); +} + +/** + * suspend_enter - enter the desired system sleep state. + * @state: state to enter + * + * This function should be called after devices have been suspended. + */ +static int suspend_enter(suspend_state_t state) +{ + int error; + + if (suspend_ops->prepare) { + error = suspend_ops->prepare(); + if (error) + return error; + } + + error = dpm_suspend_noirq(PMSG_SUSPEND); + if (error) { + printk(KERN_ERR "PM: Some devices failed to power down\n"); + goto Platfrom_finish; + } + + if (suspend_ops->prepare_late) { + error = suspend_ops->prepare_late(); + if (error) + goto Power_up_devices; + } + + if (suspend_test(TEST_PLATFORM)) + goto Platform_wake; + + error = disable_nonboot_cpus(); + if (error || suspend_test(TEST_CPUS)) + goto Enable_cpus; + + arch_suspend_disable_irqs(); + BUG_ON(!irqs_disabled()); + + error = sysdev_suspend(PMSG_SUSPEND); + if (!error) { + if (!suspend_test(TEST_CORE)) + error = suspend_ops->enter(state); + sysdev_resume(); + } + + arch_suspend_enable_irqs(); + BUG_ON(irqs_disabled()); + + Enable_cpus: + enable_nonboot_cpus(); + + Platform_wake: + if (suspend_ops->wake) + suspend_ops->wake(); + + Power_up_devices: + dpm_resume_noirq(PMSG_RESUME); + + Platfrom_finish: + if (suspend_ops->finish) + suspend_ops->finish(); + + return error; +} + +/** + * suspend_devices_and_enter - suspend devices and enter the desired system + * sleep state. + * @state: state to enter + */ +int suspend_devices_and_enter(suspend_state_t state) +{ + int error; + + if (!suspend_ops) + return -ENOSYS; + + if (suspend_ops->begin) { + error = suspend_ops->begin(state); + if (error) + goto Close; + } + suspend_console(); + suspend_test_start(); + error = dpm_suspend_start(PMSG_SUSPEND); + if (error) { + printk(KERN_ERR "PM: Some devices failed to suspend\n"); + goto Recover_platform; + } + suspend_test_finish("suspend devices"); + if (suspend_test(TEST_DEVICES)) + goto Recover_platform; + + suspend_enter(state); + + Resume_devices: + suspend_test_start(); + dpm_resume_end(PMSG_RESUME); + suspend_test_finish("resume devices"); + resume_console(); + Close: + if (suspend_ops->end) + suspend_ops->end(); + return error; + + Recover_platform: + if (suspend_ops->recover) + suspend_ops->recover(); + goto Resume_devices; +} + +/** + * suspend_finish - Do final work before exiting suspend sequence. + * + * Call platform code to clean up, restart processes, and free the + * console that we've allocated. This is not called for suspend-to-disk. + */ +static void suspend_finish(void) +{ + suspend_thaw_processes(); + usermodehelper_enable(); + pm_notifier_call_chain(PM_POST_SUSPEND); + pm_restore_console(); +} + +/** + * enter_state - Do common work of entering low-power state. + * @state: pm_state structure for state we're entering. + * + * Make sure we're the only ones trying to enter a sleep state. Fail + * if someone has beat us to it, since we don't want anything weird to + * happen when we wake up. + * Then, do the setup for suspend, enter the state, and cleaup (after + * we've woken up). + */ +int enter_state(suspend_state_t state) +{ + int error; + + if (!valid_state(state)) + return -ENODEV; + + if (!mutex_trylock(&pm_mutex)) + return -EBUSY; + + printk(KERN_INFO "PM: Syncing filesystems ... "); + sys_sync(); + printk("done.\n"); + + pr_debug("PM: Preparing system for %s sleep\n", pm_states[state]); + error = suspend_prepare(); + if (error) + goto Unlock; + + if (suspend_test(TEST_FREEZER)) + goto Finish; + + pr_debug("PM: Entering %s sleep\n", pm_states[state]); + error = suspend_devices_and_enter(state); + + Finish: + pr_debug("PM: Finishing wakeup.\n"); + suspend_finish(); + Unlock: + mutex_unlock(&pm_mutex); + return error; +} + +/** + * pm_suspend - Externally visible function for suspending system. + * @state: Enumerated value of state to enter. + * + * Determine whether or not value is within range, get state + * structure, and enter (above). + */ +int pm_suspend(suspend_state_t state) +{ + if (state > PM_SUSPEND_ON && state <= PM_SUSPEND_MAX) + return enter_state(state); + return -EINVAL; +} +EXPORT_SYMBOL(pm_suspend); diff --git a/kernel/power/suspend_test.c b/kernel/power/suspend_test.c new file mode 100644 index 00000000000..17d8bb1acf9 --- /dev/null +++ b/kernel/power/suspend_test.c @@ -0,0 +1,187 @@ +/* + * kernel/power/suspend_test.c - Suspend to RAM and standby test facility. + * + * Copyright (c) 2009 Pavel Machek <pavel@ucw.cz> + * + * This file is released under the GPLv2. + */ + +#include <linux/init.h> +#include <linux/rtc.h> + +#include "power.h" + +/* + * We test the system suspend code by setting an RTC wakealarm a short + * time in the future, then suspending. Suspending the devices won't + * normally take long ... some systems only need a few milliseconds. + * + * The time it takes is system-specific though, so when we test this + * during system bootup we allow a LOT of time. + */ +#define TEST_SUSPEND_SECONDS 5 + +static unsigned long suspend_test_start_time; + +void suspend_test_start(void) +{ + /* FIXME Use better timebase than "jiffies", ideally a clocksource. + * What we want is a hardware counter that will work correctly even + * during the irqs-are-off stages of the suspend/resume cycle... + */ + suspend_test_start_time = jiffies; +} + +void suspend_test_finish(const char *label) +{ + long nj = jiffies - suspend_test_start_time; + unsigned msec; + + msec = jiffies_to_msecs(abs(nj)); + pr_info("PM: %s took %d.%03d seconds\n", label, + msec / 1000, msec % 1000); + + /* Warning on suspend means the RTC alarm period needs to be + * larger -- the system was sooo slooowwww to suspend that the + * alarm (should have) fired before the system went to sleep! + * + * Warning on either suspend or resume also means the system + * has some performance issues. The stack dump of a WARN_ON + * is more likely to get the right attention than a printk... + */ + WARN(msec > (TEST_SUSPEND_SECONDS * 1000), "Component: %s\n", label); +} + +/* + * To test system suspend, we need a hands-off mechanism to resume the + * system. RTCs wake alarms are a common self-contained mechanism. + */ + +static void __init test_wakealarm(struct rtc_device *rtc, suspend_state_t state) +{ + static char err_readtime[] __initdata = + KERN_ERR "PM: can't read %s time, err %d\n"; + static char err_wakealarm [] __initdata = + KERN_ERR "PM: can't set %s wakealarm, err %d\n"; + static char err_suspend[] __initdata = + KERN_ERR "PM: suspend test failed, error %d\n"; + static char info_test[] __initdata = + KERN_INFO "PM: test RTC wakeup from '%s' suspend\n"; + + unsigned long now; + struct rtc_wkalrm alm; + int status; + + /* this may fail if the RTC hasn't been initialized */ + status = rtc_read_time(rtc, &alm.time); + if (status < 0) { + printk(err_readtime, dev_name(&rtc->dev), status); + return; + } + rtc_tm_to_time(&alm.time, &now); + + memset(&alm, 0, sizeof alm); + rtc_time_to_tm(now + TEST_SUSPEND_SECONDS, &alm.time); + alm.enabled = true; + + status = rtc_set_alarm(rtc, &alm); + if (status < 0) { + printk(err_wakealarm, dev_name(&rtc->dev), status); + return; + } + + if (state == PM_SUSPEND_MEM) { + printk(info_test, pm_states[state]); + status = pm_suspend(state); + if (status == -ENODEV) + state = PM_SUSPEND_STANDBY; + } + if (state == PM_SUSPEND_STANDBY) { + printk(info_test, pm_states[state]); + status = pm_suspend(state); + } + if (status < 0) + printk(err_suspend, status); + + /* Some platforms can't detect that the alarm triggered the + * wakeup, or (accordingly) disable it after it afterwards. + * It's supposed to give oneshot behavior; cope. + */ + alm.enabled = false; + rtc_set_alarm(rtc, &alm); +} + +static int __init has_wakealarm(struct device *dev, void *name_ptr) +{ + struct rtc_device *candidate = to_rtc_device(dev); + + if (!candidate->ops->set_alarm) + return 0; + if (!device_may_wakeup(candidate->dev.parent)) + return 0; + + *(const char **)name_ptr = dev_name(dev); + return 1; +} + +/* + * Kernel options like "test_suspend=mem" force suspend/resume sanity tests + * at startup time. They're normally disabled, for faster boot and because + * we can't know which states really work on this particular system. + */ +static suspend_state_t test_state __initdata = PM_SUSPEND_ON; + +static char warn_bad_state[] __initdata = + KERN_WARNING "PM: can't test '%s' suspend state\n"; + +static int __init setup_test_suspend(char *value) +{ + unsigned i; + + /* "=mem" ==> "mem" */ + value++; + for (i = 0; i < PM_SUSPEND_MAX; i++) { + if (!pm_states[i]) + continue; + if (strcmp(pm_states[i], value) != 0) + continue; + test_state = (__force suspend_state_t) i; + return 0; + } + printk(warn_bad_state, value); + return 0; +} +__setup("test_suspend", setup_test_suspend); + +static int __init test_suspend(void) +{ + static char warn_no_rtc[] __initdata = + KERN_WARNING "PM: no wakealarm-capable RTC driver is ready\n"; + + char *pony = NULL; + struct rtc_device *rtc = NULL; + + /* PM is initialized by now; is that state testable? */ + if (test_state == PM_SUSPEND_ON) + goto done; + if (!valid_state(test_state)) { + printk(warn_bad_state, pm_states[test_state]); + goto done; + } + + /* RTCs have initialized by now too ... can we use one? */ + class_find_device(rtc_class, NULL, &pony, has_wakealarm); + if (pony) + rtc = rtc_class_open(pony); + if (!rtc) { + printk(warn_no_rtc); + goto done; + } + + /* go for it */ + test_wakealarm(rtc, test_state); + rtc_class_close(rtc); +done: + return 0; +} +late_initcall(test_suspend); diff --git a/kernel/power/swap.c b/kernel/power/swap.c index 6da14358537..8ba052c86d4 100644 --- a/kernel/power/swap.c +++ b/kernel/power/swap.c @@ -60,11 +60,10 @@ static struct block_device *resume_bdev; static int submit(int rw, pgoff_t page_off, struct page *page, struct bio **bio_chain) { + const int bio_rw = rw | (1 << BIO_RW_SYNCIO) | (1 << BIO_RW_UNPLUG); struct bio *bio; bio = bio_alloc(__GFP_WAIT | __GFP_HIGH, 1); - if (!bio) - return -ENOMEM; bio->bi_sector = page_off * (PAGE_SIZE >> 9); bio->bi_bdev = resume_bdev; bio->bi_end_io = end_swap_bio_read; @@ -80,7 +79,7 @@ static int submit(int rw, pgoff_t page_off, struct page *page, bio_get(bio); if (bio_chain == NULL) { - submit_bio(rw | (1 << BIO_RW_SYNC), bio); + submit_bio(bio_rw, bio); wait_on_page_locked(page); if (rw == READ) bio_set_pages_dirty(bio); @@ -90,7 +89,7 @@ static int submit(int rw, pgoff_t page_off, struct page *page, get_page(page); /* These pages are freed later */ bio->bi_private = *bio_chain; *bio_chain = bio; - submit_bio(rw | (1 << BIO_RW_SYNC), bio); + submit_bio(bio_rw, bio); } return 0; } diff --git a/kernel/power/swsusp.c b/kernel/power/swsusp.c index a92c9145155..6a07f4dbf2f 100644 --- a/kernel/power/swsusp.c +++ b/kernel/power/swsusp.c @@ -51,17 +51,10 @@ #include <linux/highmem.h> #include <linux/time.h> #include <linux/rbtree.h> +#include <linux/io.h> #include "power.h" -/* - * Preferred image size in bytes (tunable via /sys/power/image_size). - * When it is set to N, swsusp will do its best to ensure the image - * size will not exceed N bytes, but if that is impossible, it will - * try to create the smallest image possible. - */ -unsigned long image_size = 500 * 1024 * 1024; - int in_suspend __nosavedata = 0; /** @@ -193,194 +186,3 @@ void swsusp_show_speed(struct timeval *start, struct timeval *stop, centisecs / 100, centisecs % 100, kps / 1000, (kps % 1000) / 10); } - -/** - * swsusp_shrink_memory - Try to free as much memory as needed - * - * ... but do not OOM-kill anyone - * - * Notice: all userland should be stopped before it is called, or - * livelock is possible. - */ - -#define SHRINK_BITE 10000 -static inline unsigned long __shrink_memory(long tmp) -{ - if (tmp > SHRINK_BITE) - tmp = SHRINK_BITE; - return shrink_all_memory(tmp); -} - -int swsusp_shrink_memory(void) -{ - long tmp; - struct zone *zone; - unsigned long pages = 0; - unsigned int i = 0; - char *p = "-\\|/"; - struct timeval start, stop; - - printk(KERN_INFO "PM: Shrinking memory... "); - do_gettimeofday(&start); - do { - long size, highmem_size; - - highmem_size = count_highmem_pages(); - size = count_data_pages() + PAGES_FOR_IO + SPARE_PAGES; - tmp = size; - size += highmem_size; - for_each_zone (zone) - if (populated_zone(zone)) { - tmp += snapshot_additional_pages(zone); - if (is_highmem(zone)) { - highmem_size -= - zone_page_state(zone, NR_FREE_PAGES); - } else { - tmp -= zone_page_state(zone, NR_FREE_PAGES); - tmp += zone->lowmem_reserve[ZONE_NORMAL]; - } - } - - if (highmem_size < 0) - highmem_size = 0; - - tmp += highmem_size; - if (tmp > 0) { - tmp = __shrink_memory(tmp); - if (!tmp) - return -ENOMEM; - pages += tmp; - } else if (size > image_size / PAGE_SIZE) { - tmp = __shrink_memory(size - (image_size / PAGE_SIZE)); - pages += tmp; - } - printk("\b%c", p[i++%4]); - } while (tmp > 0); - do_gettimeofday(&stop); - printk("\bdone (%lu pages freed)\n", pages); - swsusp_show_speed(&start, &stop, pages, "Freed"); - - return 0; -} - -/* - * Platforms, like ACPI, may want us to save some memory used by them during - * hibernation and to restore the contents of this memory during the subsequent - * resume. The code below implements a mechanism allowing us to do that. - */ - -struct nvs_page { - unsigned long phys_start; - unsigned int size; - void *kaddr; - void *data; - struct list_head node; -}; - -static LIST_HEAD(nvs_list); - -/** - * hibernate_nvs_register - register platform NVS memory region to save - * @start - physical address of the region - * @size - size of the region - * - * The NVS region need not be page-aligned (both ends) and we arrange - * things so that the data from page-aligned addresses in this region will - * be copied into separate RAM pages. - */ -int hibernate_nvs_register(unsigned long start, unsigned long size) -{ - struct nvs_page *entry, *next; - - while (size > 0) { - unsigned int nr_bytes; - - entry = kzalloc(sizeof(struct nvs_page), GFP_KERNEL); - if (!entry) - goto Error; - - list_add_tail(&entry->node, &nvs_list); - entry->phys_start = start; - nr_bytes = PAGE_SIZE - (start & ~PAGE_MASK); - entry->size = (size < nr_bytes) ? size : nr_bytes; - - start += entry->size; - size -= entry->size; - } - return 0; - - Error: - list_for_each_entry_safe(entry, next, &nvs_list, node) { - list_del(&entry->node); - kfree(entry); - } - return -ENOMEM; -} - -/** - * hibernate_nvs_free - free data pages allocated for saving NVS regions - */ -void hibernate_nvs_free(void) -{ - struct nvs_page *entry; - - list_for_each_entry(entry, &nvs_list, node) - if (entry->data) { - free_page((unsigned long)entry->data); - entry->data = NULL; - if (entry->kaddr) { - iounmap(entry->kaddr); - entry->kaddr = NULL; - } - } -} - -/** - * hibernate_nvs_alloc - allocate memory necessary for saving NVS regions - */ -int hibernate_nvs_alloc(void) -{ - struct nvs_page *entry; - - list_for_each_entry(entry, &nvs_list, node) { - entry->data = (void *)__get_free_page(GFP_KERNEL); - if (!entry->data) { - hibernate_nvs_free(); - return -ENOMEM; - } - } - return 0; -} - -/** - * hibernate_nvs_save - save NVS memory regions - */ -void hibernate_nvs_save(void) -{ - struct nvs_page *entry; - - printk(KERN_INFO "PM: Saving platform NVS memory\n"); - - list_for_each_entry(entry, &nvs_list, node) - if (entry->data) { - entry->kaddr = ioremap(entry->phys_start, entry->size); - memcpy(entry->data, entry->kaddr, entry->size); - } -} - -/** - * hibernate_nvs_restore - restore NVS memory regions - * - * This function is going to be called with interrupts disabled, so it - * cannot iounmap the virtual addresses used to access the NVS region. - */ -void hibernate_nvs_restore(void) -{ - struct nvs_page *entry; - - printk(KERN_INFO "PM: Restoring platform NVS memory\n"); - - list_for_each_entry(entry, &nvs_list, node) - if (entry->data) - memcpy(entry->kaddr, entry->data, entry->size); -} diff --git a/kernel/power/user.c b/kernel/power/user.c index 005b93d839b..bf0014d6a5f 100644 --- a/kernel/power/user.c +++ b/kernel/power/user.c @@ -23,7 +23,7 @@ #include <linux/console.h> #include <linux/cpu.h> #include <linux/freezer.h> -#include <linux/smp_lock.h> +#include <scsi/scsi_scan.h> #include <asm/uaccess.h> @@ -92,18 +92,26 @@ static int snapshot_open(struct inode *inode, struct file *filp) filp->private_data = data; memset(&data->handle, 0, sizeof(struct snapshot_handle)); if ((filp->f_flags & O_ACCMODE) == O_RDONLY) { + /* Hibernating. The image device should be accessible. */ data->swap = swsusp_resume_device ? swap_type_of(swsusp_resume_device, 0, NULL) : -1; data->mode = O_RDONLY; - error = pm_notifier_call_chain(PM_RESTORE_PREPARE); + error = pm_notifier_call_chain(PM_HIBERNATION_PREPARE); if (error) - pm_notifier_call_chain(PM_POST_RESTORE); + pm_notifier_call_chain(PM_POST_HIBERNATION); } else { + /* + * Resuming. We may need to wait for the image device to + * appear. + */ + wait_for_device_probe(); + scsi_complete_async_scans(); + data->swap = -1; data->mode = O_WRONLY; - error = pm_notifier_call_chain(PM_HIBERNATION_PREPARE); + error = pm_notifier_call_chain(PM_RESTORE_PREPARE); if (error) - pm_notifier_call_chain(PM_POST_HIBERNATION); + pm_notifier_call_chain(PM_POST_RESTORE); } if (error) atomic_inc(&snapshot_device_available); diff --git a/kernel/printk.c b/kernel/printk.c index 69188f226a9..e10d193a833 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -32,10 +32,17 @@ #include <linux/security.h> #include <linux/bootmem.h> #include <linux/syscalls.h> +#include <linux/kexec.h> #include <asm/uaccess.h> /* + * for_each_console() allows you to iterate on each console + */ +#define for_each_console(con) \ + for (con = console_drivers; con != NULL; con = con->next) + +/* * Architectures can override it: */ void asmlinkage __attribute__((weak)) early_printk(const char *fmt, ...) @@ -60,6 +67,8 @@ int console_printk[4] = { DEFAULT_CONSOLE_LOGLEVEL, /* default_console_loglevel */ }; +static int saved_console_loglevel = -1; + /* * Low level drivers may need that to know if they can schedule in * their unblank() callback or not. So let's export it. @@ -73,7 +82,6 @@ EXPORT_SYMBOL(oops_in_progress); * driver system. */ static DECLARE_MUTEX(console_sem); -static DECLARE_MUTEX(secondary_console_sem); struct console *console_drivers; EXPORT_SYMBOL_GPL(console_drivers); @@ -136,6 +144,24 @@ static char *log_buf = __log_buf; static int log_buf_len = __LOG_BUF_LEN; static unsigned logged_chars; /* Number of chars produced since last read+clear operation */ +#ifdef CONFIG_KEXEC +/* + * This appends the listed symbols to /proc/vmcoreinfo + * + * /proc/vmcoreinfo is used by various utiilties, like crash and makedumpfile to + * obtain access to symbols that are otherwise very difficult to locate. These + * symbols are specifically used so that utilities can access and extract the + * dmesg log from a vmcore file after a crash. + */ +void log_buf_kexec_setup(void) +{ + VMCOREINFO_SYMBOL(log_buf); + VMCOREINFO_SYMBOL(log_end); + VMCOREINFO_SYMBOL(log_buf_len); + VMCOREINFO_SYMBOL(logged_chars); +} +#endif + static int __init log_buf_len_setup(char *str) { unsigned size = memparse(str, &str); @@ -354,10 +380,15 @@ int do_syslog(int type, char __user *buf, int len) logged_chars = 0; break; case 6: /* Disable logging to console */ + if (saved_console_loglevel == -1) + saved_console_loglevel = console_loglevel; console_loglevel = minimum_console_loglevel; break; case 7: /* Enable logging to console */ - console_loglevel = default_console_loglevel; + if (saved_console_loglevel != -1) { + console_loglevel = saved_console_loglevel; + saved_console_loglevel = -1; + } break; case 8: /* Set level of messages printed to console */ error = -EINVAL; @@ -366,6 +397,8 @@ int do_syslog(int type, char __user *buf, int len) if (len < minimum_console_loglevel) len = minimum_console_loglevel; console_loglevel = len; + /* Implicitly re-enable logging to console */ + saved_console_loglevel = -1; error = 0; break; case 9: /* Number of chars in the log buffer */ @@ -394,7 +427,7 @@ static void __call_console_drivers(unsigned start, unsigned end) { struct console *con; - for (con = console_drivers; con; con = con->next) { + for_each_console(con) { if ((con->flags & CON_ENABLED) && con->write && (cpu_online(smp_processor_id()) || (con->flags & CON_ANYTIME))) @@ -526,7 +559,7 @@ static int have_callable_console(void) { struct console *con; - for (con = console_drivers; con; con = con->next) + for_each_console(con) if (con->flags & CON_ANYTIME) return 1; @@ -669,20 +702,35 @@ asmlinkage int vprintk(const char *fmt, va_list args) sizeof(printk_buf) - printed_len, fmt, args); + p = printk_buf; + + /* Do we have a loglevel in the string? */ + if (p[0] == '<') { + unsigned char c = p[1]; + if (c && p[2] == '>') { + switch (c) { + case '0' ... '7': /* loglevel */ + current_log_level = c - '0'; + /* Fallthrough - make sure we're on a new line */ + case 'd': /* KERN_DEFAULT */ + if (!new_text_line) { + emit_log_char('\n'); + new_text_line = 1; + } + /* Fallthrough - skip the loglevel */ + case 'c': /* KERN_CONT */ + p += 3; + break; + } + } + } + /* * Copy the output into log_buf. If the caller didn't provide * appropriate log level tags, we insert them here */ - for (p = printk_buf; *p; p++) { + for ( ; *p; p++) { if (new_text_line) { - /* If a token, set current_log_level and skip over */ - if (p[0] == '<' && p[1] >= '0' && p[1] <= '7' && - p[2] == '>') { - current_log_level = p[1] - '0'; - p += 3; - printed_len -= 3; - } - /* Always output the token */ emit_log_char('<'); emit_log_char(current_log_level + '0'); @@ -891,12 +939,14 @@ void suspend_console(void) printk("Suspending console(s) (use no_console_suspend to debug)\n"); acquire_console_sem(); console_suspended = 1; + up(&console_sem); } void resume_console(void) { if (!console_suspend_enabled) return; + down(&console_sem); console_suspended = 0; release_console_sem(); } @@ -912,11 +962,9 @@ void resume_console(void) void acquire_console_sem(void) { BUG_ON(in_interrupt()); - if (console_suspended) { - down(&secondary_console_sem); - return; - } down(&console_sem); + if (console_suspended) + return; console_locked = 1; console_may_schedule = 1; } @@ -926,6 +974,10 @@ int try_acquire_console_sem(void) { if (down_trylock(&console_sem)) return -1; + if (console_suspended) { + up(&console_sem); + return -1; + } console_locked = 1; console_may_schedule = 0; return 0; @@ -979,7 +1031,7 @@ void release_console_sem(void) unsigned wake_klogd = 0; if (console_suspended) { - up(&secondary_console_sem); + up(&console_sem); return; } @@ -1045,7 +1097,7 @@ void console_unblank(void) console_locked = 1; console_may_schedule = 0; - for (c = console_drivers; c != NULL; c = c->next) + for_each_console(c) if ((c->flags & CON_ENABLED) && c->unblank) c->unblank(); release_console_sem(); @@ -1060,7 +1112,7 @@ struct tty_driver *console_device(int *index) struct tty_driver *driver = NULL; acquire_console_sem(); - for (c = console_drivers; c != NULL; c = c->next) { + for_each_console(c) { if (!c->device) continue; driver = c->device(c, index); @@ -1097,25 +1149,49 @@ EXPORT_SYMBOL(console_start); * to register the console printing procedure with printk() and to * print any messages that were printed by the kernel before the * console driver was initialized. + * + * This can happen pretty early during the boot process (because of + * early_printk) - sometimes before setup_arch() completes - be careful + * of what kernel features are used - they may not be initialised yet. + * + * There are two types of consoles - bootconsoles (early_printk) and + * "real" consoles (everything which is not a bootconsole) which are + * handled differently. + * - Any number of bootconsoles can be registered at any time. + * - As soon as a "real" console is registered, all bootconsoles + * will be unregistered automatically. + * - Once a "real" console is registered, any attempt to register a + * bootconsoles will be rejected */ -void register_console(struct console *console) +void register_console(struct console *newcon) { int i; unsigned long flags; - struct console *bootconsole = NULL; + struct console *bcon = NULL; - if (console_drivers) { - if (console->flags & CON_BOOT) - return; - if (console_drivers->flags & CON_BOOT) - bootconsole = console_drivers; + /* + * before we register a new CON_BOOT console, make sure we don't + * already have a valid console + */ + if (console_drivers && newcon->flags & CON_BOOT) { + /* find the last or real console */ + for_each_console(bcon) { + if (!(bcon->flags & CON_BOOT)) { + printk(KERN_INFO "Too late to register bootconsole %s%d\n", + newcon->name, newcon->index); + return; + } + } } - if (preferred_console < 0 || bootconsole || !console_drivers) + if (console_drivers && console_drivers->flags & CON_BOOT) + bcon = console_drivers; + + if (preferred_console < 0 || bcon || !console_drivers) preferred_console = selected_console; - if (console->early_setup) - console->early_setup(); + if (newcon->early_setup) + newcon->early_setup(); /* * See if we want to use this console driver. If we @@ -1123,13 +1199,13 @@ void register_console(struct console *console) * that registers here. */ if (preferred_console < 0) { - if (console->index < 0) - console->index = 0; - if (console->setup == NULL || - console->setup(console, NULL) == 0) { - console->flags |= CON_ENABLED; - if (console->device) { - console->flags |= CON_CONSDEV; + if (newcon->index < 0) + newcon->index = 0; + if (newcon->setup == NULL || + newcon->setup(newcon, NULL) == 0) { + newcon->flags |= CON_ENABLED; + if (newcon->device) { + newcon->flags |= CON_CONSDEV; preferred_console = 0; } } @@ -1141,64 +1217,62 @@ void register_console(struct console *console) */ for (i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++) { - if (strcmp(console_cmdline[i].name, console->name) != 0) + if (strcmp(console_cmdline[i].name, newcon->name) != 0) continue; - if (console->index >= 0 && - console->index != console_cmdline[i].index) + if (newcon->index >= 0 && + newcon->index != console_cmdline[i].index) continue; - if (console->index < 0) - console->index = console_cmdline[i].index; + if (newcon->index < 0) + newcon->index = console_cmdline[i].index; #ifdef CONFIG_A11Y_BRAILLE_CONSOLE if (console_cmdline[i].brl_options) { - console->flags |= CON_BRL; - braille_register_console(console, + newcon->flags |= CON_BRL; + braille_register_console(newcon, console_cmdline[i].index, console_cmdline[i].options, console_cmdline[i].brl_options); return; } #endif - if (console->setup && - console->setup(console, console_cmdline[i].options) != 0) + if (newcon->setup && + newcon->setup(newcon, console_cmdline[i].options) != 0) break; - console->flags |= CON_ENABLED; - console->index = console_cmdline[i].index; + newcon->flags |= CON_ENABLED; + newcon->index = console_cmdline[i].index; if (i == selected_console) { - console->flags |= CON_CONSDEV; + newcon->flags |= CON_CONSDEV; preferred_console = selected_console; } break; } - if (!(console->flags & CON_ENABLED)) + if (!(newcon->flags & CON_ENABLED)) return; - if (bootconsole && (console->flags & CON_CONSDEV)) { - printk(KERN_INFO "console handover: boot [%s%d] -> real [%s%d]\n", - bootconsole->name, bootconsole->index, - console->name, console->index); - unregister_console(bootconsole); - console->flags &= ~CON_PRINTBUFFER; - } else { - printk(KERN_INFO "console [%s%d] enabled\n", - console->name, console->index); - } + /* + * If we have a bootconsole, and are switching to a real console, + * don't print everything out again, since when the boot console, and + * the real console are the same physical device, it's annoying to + * see the beginning boot messages twice + */ + if (bcon && ((newcon->flags & (CON_CONSDEV | CON_BOOT)) == CON_CONSDEV)) + newcon->flags &= ~CON_PRINTBUFFER; /* * Put this console in the list - keep the * preferred driver at the head of the list. */ acquire_console_sem(); - if ((console->flags & CON_CONSDEV) || console_drivers == NULL) { - console->next = console_drivers; - console_drivers = console; - if (console->next) - console->next->flags &= ~CON_CONSDEV; + if ((newcon->flags & CON_CONSDEV) || console_drivers == NULL) { + newcon->next = console_drivers; + console_drivers = newcon; + if (newcon->next) + newcon->next->flags &= ~CON_CONSDEV; } else { - console->next = console_drivers->next; - console_drivers->next = console; + newcon->next = console_drivers->next; + console_drivers->next = newcon; } - if (console->flags & CON_PRINTBUFFER) { + if (newcon->flags & CON_PRINTBUFFER) { /* * release_console_sem() will print out the buffered messages * for us. @@ -1208,6 +1282,28 @@ void register_console(struct console *console) spin_unlock_irqrestore(&logbuf_lock, flags); } release_console_sem(); + + /* + * By unregistering the bootconsoles after we enable the real console + * we get the "console xxx enabled" message on all the consoles - + * boot consoles, real consoles, etc - this is to ensure that end + * users know there might be something in the kernel's log buffer that + * went to the bootconsole (that they do not see on the real console) + */ + if (bcon && ((newcon->flags & (CON_CONSDEV | CON_BOOT)) == CON_CONSDEV)) { + /* we need to iterate through twice, to make sure we print + * everything out, before we unregister the console(s) + */ + printk(KERN_INFO "console [%s%d] enabled, bootconsole disabled\n", + newcon->name, newcon->index); + for_each_console(bcon) + if (bcon->flags & CON_BOOT) + unregister_console(bcon); + } else { + printk(KERN_INFO "%sconsole [%s%d] enabled\n", + (newcon->flags & CON_BOOT) ? "boot" : "" , + newcon->name, newcon->index); + } } EXPORT_SYMBOL(register_console); @@ -1250,11 +1346,13 @@ EXPORT_SYMBOL(unregister_console); static int __init disable_boot_consoles(void) { - if (console_drivers != NULL) { - if (console_drivers->flags & CON_BOOT) { + struct console *con; + + for_each_console(con) { + if (con->flags & CON_BOOT) { printk(KERN_INFO "turn off boot console %s%d\n", - console_drivers->name, console_drivers->index); - return unregister_console(console_drivers); + con->name, con->index); + unregister_console(con); } } return 0; @@ -1289,8 +1387,11 @@ EXPORT_SYMBOL(printk_ratelimit); bool printk_timed_ratelimit(unsigned long *caller_jiffies, unsigned int interval_msecs) { - if (*caller_jiffies == 0 || time_after(jiffies, *caller_jiffies)) { - *caller_jiffies = jiffies + msecs_to_jiffies(interval_msecs); + if (*caller_jiffies == 0 + || !time_in_range(jiffies, *caller_jiffies, + *caller_jiffies + + msecs_to_jiffies(interval_msecs))) { + *caller_jiffies = jiffies; return true; } return false; diff --git a/kernel/profile.c b/kernel/profile.c index 784933acf5b..419250ebec4 100644 --- a/kernel/profile.c +++ b/kernel/profile.c @@ -111,20 +111,18 @@ int __ref profile_init(void) /* only text is profiled */ prof_len = (_etext - _stext) >> prof_shift; buffer_bytes = prof_len*sizeof(atomic_t); - if (!slab_is_available()) { - prof_buffer = alloc_bootmem(buffer_bytes); - alloc_bootmem_cpumask_var(&prof_cpu_mask); - return 0; - } if (!alloc_cpumask_var(&prof_cpu_mask, GFP_KERNEL)) return -ENOMEM; - prof_buffer = kzalloc(buffer_bytes, GFP_KERNEL); + cpumask_copy(prof_cpu_mask, cpu_possible_mask); + + prof_buffer = kzalloc(buffer_bytes, GFP_KERNEL|__GFP_NOWARN); if (prof_buffer) return 0; - prof_buffer = alloc_pages_exact(buffer_bytes, GFP_KERNEL|__GFP_ZERO); + prof_buffer = alloc_pages_exact(buffer_bytes, + GFP_KERNEL|__GFP_ZERO|__GFP_NOWARN); if (prof_buffer) return 0; @@ -368,7 +366,7 @@ static int __cpuinit profile_cpu_callback(struct notifier_block *info, node = cpu_to_node(cpu); per_cpu(cpu_profile_flip, cpu) = 0; if (!per_cpu(cpu_profile_hits, cpu)[1]) { - page = alloc_pages_node(node, + page = alloc_pages_exact_node(node, GFP_KERNEL | __GFP_ZERO, 0); if (!page) @@ -376,7 +374,7 @@ static int __cpuinit profile_cpu_callback(struct notifier_block *info, per_cpu(cpu_profile_hits, cpu)[1] = page_address(page); } if (!per_cpu(cpu_profile_hits, cpu)[0]) { - page = alloc_pages_node(node, + page = alloc_pages_exact_node(node, GFP_KERNEL | __GFP_ZERO, 0); if (!page) @@ -567,14 +565,14 @@ static int create_hash_tables(void) int node = cpu_to_node(cpu); struct page *page; - page = alloc_pages_node(node, + page = alloc_pages_exact_node(node, GFP_KERNEL | __GFP_ZERO | GFP_THISNODE, 0); if (!page) goto out_cleanup; per_cpu(cpu_profile_hits, cpu)[1] = (struct profile_hit *)page_address(page); - page = alloc_pages_node(node, + page = alloc_pages_exact_node(node, GFP_KERNEL | __GFP_ZERO | GFP_THISNODE, 0); if (!page) diff --git a/kernel/ptrace.c b/kernel/ptrace.c index c9cf48b21f0..307c285af59 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -21,22 +21,10 @@ #include <linux/audit.h> #include <linux/pid_namespace.h> #include <linux/syscalls.h> - -#include <asm/pgtable.h> -#include <asm/uaccess.h> +#include <linux/uaccess.h> /* - * Initialize a new task whose father had been ptraced. - * - * Called from copy_process(). - */ -void ptrace_fork(struct task_struct *child, unsigned long clone_flags) -{ - arch_ptrace_fork(child, clone_flags); -} - -/* * ptrace a task: make the debugger its new parent and * move it to the ptrace list. * @@ -48,7 +36,7 @@ void __ptrace_link(struct task_struct *child, struct task_struct *new_parent) list_add(&child->ptrace_entry, &new_parent->ptraced); child->parent = new_parent; } - + /* * Turn a tracing stop into a normal stop now, since with no tracer there * would be no way to wake it up with SIGCONT or SIGKILL. If there was a @@ -60,11 +48,15 @@ static void ptrace_untrace(struct task_struct *child) { spin_lock(&child->sighand->siglock); if (task_is_traced(child)) { - if (child->signal->flags & SIGNAL_STOP_STOPPED) { + /* + * If the group stop is completed or in progress, + * this thread was already counted as stopped. + */ + if (child->signal->flags & SIGNAL_STOP_STOPPED || + child->signal->group_stop_count) __set_task_state(child, TASK_STOPPED); - } else { + else signal_wake_up(child, 1); - } } spin_unlock(&child->sighand->siglock); } @@ -160,7 +152,7 @@ int __ptrace_may_access(struct task_struct *task, unsigned int mode) if (!dumpable && !capable(CAP_SYS_PTRACE)) return -EPERM; - return security_ptrace_may_access(task, mode); + return security_ptrace_access_check(task, mode); } bool ptrace_may_access(struct task_struct *task, unsigned int mode) @@ -169,84 +161,140 @@ bool ptrace_may_access(struct task_struct *task, unsigned int mode) task_lock(task); err = __ptrace_may_access(task, mode); task_unlock(task); - return (!err ? true : false); + return !err; } int ptrace_attach(struct task_struct *task) { int retval; - unsigned long flags; audit_ptrace(task); retval = -EPERM; + if (unlikely(task->flags & PF_KTHREAD)) + goto out; if (same_thread_group(task, current)) goto out; - /* Protect exec's credential calculations against our interference; - * SUID, SGID and LSM creds get determined differently under ptrace. + /* + * Protect exec's credential calculations against our interference; + * interference; SUID, SGID and LSM creds get determined differently + * under ptrace. */ - retval = mutex_lock_interruptible(¤t->cred_exec_mutex); - if (retval < 0) + retval = -ERESTARTNOINTR; + if (mutex_lock_interruptible(&task->cred_guard_mutex)) goto out; - retval = -EPERM; -repeat: - /* - * Nasty, nasty. - * - * We want to hold both the task-lock and the - * tasklist_lock for writing at the same time. - * But that's against the rules (tasklist_lock - * is taken for reading by interrupts on other - * cpu's that may have task_lock). - */ task_lock(task); - if (!write_trylock_irqsave(&tasklist_lock, flags)) { - task_unlock(task); - do { - cpu_relax(); - } while (!write_can_lock(&tasklist_lock)); - goto repeat; - } - - if (!task->mm) - goto bad; - /* the same process cannot be attached many times */ - if (task->ptrace & PT_PTRACED) - goto bad; retval = __ptrace_may_access(task, PTRACE_MODE_ATTACH); + task_unlock(task); if (retval) - goto bad; + goto unlock_creds; - /* Go */ - task->ptrace |= PT_PTRACED; + write_lock_irq(&tasklist_lock); + retval = -EPERM; + if (unlikely(task->exit_state)) + goto unlock_tasklist; + if (task->ptrace) + goto unlock_tasklist; + + task->ptrace = PT_PTRACED; if (capable(CAP_SYS_PTRACE)) task->ptrace |= PT_PTRACE_CAP; __ptrace_link(task, current); - send_sig_info(SIGSTOP, SEND_SIG_FORCED, task); -bad: - write_unlock_irqrestore(&tasklist_lock, flags); - task_unlock(task); - mutex_unlock(¤t->cred_exec_mutex); + + retval = 0; +unlock_tasklist: + write_unlock_irq(&tasklist_lock); +unlock_creds: + mutex_unlock(&task->cred_guard_mutex); out: return retval; } -static inline void __ptrace_detach(struct task_struct *child, unsigned int data) +/** + * ptrace_traceme -- helper for PTRACE_TRACEME + * + * Performs checks and sets PT_PTRACED. + * Should be used by all ptrace implementations for PTRACE_TRACEME. + */ +int ptrace_traceme(void) { - child->exit_code = data; - /* .. re-parent .. */ - __ptrace_unlink(child); - /* .. and wake it up. */ - if (child->exit_state != EXIT_ZOMBIE) - wake_up_process(child); + int ret = -EPERM; + + write_lock_irq(&tasklist_lock); + /* Are we already being traced? */ + if (!current->ptrace) { + ret = security_ptrace_traceme(current->parent); + /* + * Check PF_EXITING to ensure ->real_parent has not passed + * exit_ptrace(). Otherwise we don't report the error but + * pretend ->real_parent untraces us right after return. + */ + if (!ret && !(current->real_parent->flags & PF_EXITING)) { + current->ptrace = PT_PTRACED; + __ptrace_link(current, current->real_parent); + } + } + write_unlock_irq(&tasklist_lock); + + return ret; +} + +/* + * Called with irqs disabled, returns true if childs should reap themselves. + */ +static int ignoring_children(struct sighand_struct *sigh) +{ + int ret; + spin_lock(&sigh->siglock); + ret = (sigh->action[SIGCHLD-1].sa.sa_handler == SIG_IGN) || + (sigh->action[SIGCHLD-1].sa.sa_flags & SA_NOCLDWAIT); + spin_unlock(&sigh->siglock); + return ret; +} + +/* + * Called with tasklist_lock held for writing. + * Unlink a traced task, and clean it up if it was a traced zombie. + * Return true if it needs to be reaped with release_task(). + * (We can't call release_task() here because we already hold tasklist_lock.) + * + * If it's a zombie, our attachedness prevented normal parent notification + * or self-reaping. Do notification now if it would have happened earlier. + * If it should reap itself, return true. + * + * If it's our own child, there is no notification to do. + * But if our normal children self-reap, then this child + * was prevented by ptrace and we must reap it now. + */ +static bool __ptrace_detach(struct task_struct *tracer, struct task_struct *p) +{ + __ptrace_unlink(p); + + if (p->exit_state == EXIT_ZOMBIE) { + if (!task_detached(p) && thread_group_empty(p)) { + if (!same_thread_group(p->real_parent, tracer)) + do_notify_parent(p, p->exit_signal); + else if (ignoring_children(tracer->sighand)) + p->exit_signal = -1; + } + if (task_detached(p)) { + /* Mark it as in the process of being reaped. */ + p->exit_state = EXIT_DEAD; + return true; + } + } + + return false; } int ptrace_detach(struct task_struct *child, unsigned int data) { + bool dead = false; + if (!valid_signal(data)) return -EIO; @@ -255,14 +303,47 @@ int ptrace_detach(struct task_struct *child, unsigned int data) clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE); write_lock_irq(&tasklist_lock); - /* protect against de_thread()->release_task() */ - if (child->ptrace) - __ptrace_detach(child, data); + /* + * This child can be already killed. Make sure de_thread() or + * our sub-thread doing do_wait() didn't do release_task() yet. + */ + if (child->ptrace) { + child->exit_code = data; + dead = __ptrace_detach(current, child); + if (!child->exit_state) + wake_up_process(child); + } write_unlock_irq(&tasklist_lock); + if (unlikely(dead)) + release_task(child); + return 0; } +/* + * Detach all tasks we were using ptrace on. + */ +void exit_ptrace(struct task_struct *tracer) +{ + struct task_struct *p, *n; + LIST_HEAD(ptrace_dead); + + write_lock_irq(&tasklist_lock); + list_for_each_entry_safe(p, n, &tracer->ptraced, ptrace_entry) { + if (__ptrace_detach(tracer, p)) + list_add(&p->ptrace_entry, &ptrace_dead); + } + write_unlock_irq(&tasklist_lock); + + BUG_ON(!list_empty(&tracer->ptraced)); + + list_for_each_entry_safe(p, n, &ptrace_dead, ptrace_entry) { + list_del_init(&p->ptrace_entry); + release_task(p); + } +} + int ptrace_readdata(struct task_struct *tsk, unsigned long src, char __user *dst, int len) { int copied = 0; @@ -283,7 +364,7 @@ int ptrace_readdata(struct task_struct *tsk, unsigned long src, char __user *dst copied += retval; src += retval; dst += retval; - len -= retval; + len -= retval; } return copied; } @@ -308,7 +389,7 @@ int ptrace_writedata(struct task_struct *tsk, char __user *src, unsigned long ds copied += retval; src += retval; dst += retval; - len -= retval; + len -= retval; } return copied; } @@ -343,37 +424,33 @@ static int ptrace_setoptions(struct task_struct *child, long data) static int ptrace_getsiginfo(struct task_struct *child, siginfo_t *info) { + unsigned long flags; int error = -ESRCH; - read_lock(&tasklist_lock); - if (likely(child->sighand != NULL)) { + if (lock_task_sighand(child, &flags)) { error = -EINVAL; - spin_lock_irq(&child->sighand->siglock); if (likely(child->last_siginfo != NULL)) { *info = *child->last_siginfo; error = 0; } - spin_unlock_irq(&child->sighand->siglock); + unlock_task_sighand(child, &flags); } - read_unlock(&tasklist_lock); return error; } static int ptrace_setsiginfo(struct task_struct *child, const siginfo_t *info) { + unsigned long flags; int error = -ESRCH; - read_lock(&tasklist_lock); - if (likely(child->sighand != NULL)) { + if (lock_task_sighand(child, &flags)) { error = -EINVAL; - spin_lock_irq(&child->sighand->siglock); if (likely(child->last_siginfo != NULL)) { *child->last_siginfo = *info; error = 0; } - spin_unlock_irq(&child->sighand->siglock); + unlock_task_sighand(child, &flags); } - read_unlock(&tasklist_lock); return error; } @@ -421,9 +498,9 @@ static int ptrace_resume(struct task_struct *child, long request, long data) if (unlikely(!arch_has_single_step())) return -EIO; user_enable_single_step(child); - } - else + } else { user_disable_single_step(child); + } child->exit_code = data; wake_up_process(child); @@ -500,71 +577,16 @@ int ptrace_request(struct task_struct *child, long request, return ret; } -/** - * ptrace_traceme -- helper for PTRACE_TRACEME - * - * Performs checks and sets PT_PTRACED. - * Should be used by all ptrace implementations for PTRACE_TRACEME. - */ -int ptrace_traceme(void) -{ - int ret = -EPERM; - - /* - * Are we already being traced? - */ -repeat: - task_lock(current); - if (!(current->ptrace & PT_PTRACED)) { - /* - * See ptrace_attach() comments about the locking here. - */ - unsigned long flags; - if (!write_trylock_irqsave(&tasklist_lock, flags)) { - task_unlock(current); - do { - cpu_relax(); - } while (!write_can_lock(&tasklist_lock)); - goto repeat; - } - - ret = security_ptrace_traceme(current->parent); - - /* - * Set the ptrace bit in the process ptrace flags. - * Then link us on our parent's ptraced list. - */ - if (!ret) { - current->ptrace |= PT_PTRACED; - __ptrace_link(current, current->real_parent); - } - - write_unlock_irqrestore(&tasklist_lock, flags); - } - task_unlock(current); - return ret; -} - -/** - * ptrace_get_task_struct -- grab a task struct reference for ptrace - * @pid: process id to grab a task_struct reference of - * - * This function is a helper for ptrace implementations. It checks - * permissions and then grabs a task struct for use of the actual - * ptrace implementation. - * - * Returns the task_struct for @pid or an ERR_PTR() on failure. - */ -struct task_struct *ptrace_get_task_struct(pid_t pid) +static struct task_struct *ptrace_get_task_struct(pid_t pid) { struct task_struct *child; - read_lock(&tasklist_lock); + rcu_read_lock(); child = find_task_by_vpid(pid); if (child) get_task_struct(child); + rcu_read_unlock(); - read_unlock(&tasklist_lock); if (!child) return ERR_PTR(-ESRCH); return child; @@ -612,8 +634,6 @@ SYSCALL_DEFINE4(ptrace, long, request, long, pid, long, addr, long, data) goto out_put_task_struct; ret = arch_ptrace(child, request, addr, data); - if (ret < 0) - goto out_put_task_struct; out_put_task_struct: put_task_struct(child); diff --git a/kernel/rcuclassic.c b/kernel/rcuclassic.c deleted file mode 100644 index 490934fc7ac..00000000000 --- a/kernel/rcuclassic.c +++ /dev/null @@ -1,788 +0,0 @@ -/* - * Read-Copy Update mechanism for mutual exclusion - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright IBM Corporation, 2001 - * - * Authors: Dipankar Sarma <dipankar@in.ibm.com> - * Manfred Spraul <manfred@colorfullife.com> - * - * Based on the original work by Paul McKenney <paulmck@us.ibm.com> - * and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen. - * Papers: - * http://www.rdrop.com/users/paulmck/paper/rclockpdcsproof.pdf - * http://lse.sourceforge.net/locking/rclock_OLS.2001.05.01c.sc.pdf (OLS2001) - * - * For detailed explanation of Read-Copy Update mechanism see - - * Documentation/RCU - * - */ -#include <linux/types.h> -#include <linux/kernel.h> -#include <linux/init.h> -#include <linux/spinlock.h> -#include <linux/smp.h> -#include <linux/rcupdate.h> -#include <linux/interrupt.h> -#include <linux/sched.h> -#include <asm/atomic.h> -#include <linux/bitops.h> -#include <linux/module.h> -#include <linux/completion.h> -#include <linux/moduleparam.h> -#include <linux/percpu.h> -#include <linux/notifier.h> -#include <linux/cpu.h> -#include <linux/mutex.h> -#include <linux/time.h> - -#ifdef CONFIG_DEBUG_LOCK_ALLOC -static struct lock_class_key rcu_lock_key; -struct lockdep_map rcu_lock_map = - STATIC_LOCKDEP_MAP_INIT("rcu_read_lock", &rcu_lock_key); -EXPORT_SYMBOL_GPL(rcu_lock_map); -#endif - - -/* Definition for rcupdate control block. */ -static struct rcu_ctrlblk rcu_ctrlblk = { - .cur = -300, - .completed = -300, - .pending = -300, - .lock = __SPIN_LOCK_UNLOCKED(&rcu_ctrlblk.lock), - .cpumask = CPU_BITS_NONE, -}; -static struct rcu_ctrlblk rcu_bh_ctrlblk = { - .cur = -300, - .completed = -300, - .pending = -300, - .lock = __SPIN_LOCK_UNLOCKED(&rcu_bh_ctrlblk.lock), - .cpumask = CPU_BITS_NONE, -}; - -DEFINE_PER_CPU(struct rcu_data, rcu_data) = { 0L }; -DEFINE_PER_CPU(struct rcu_data, rcu_bh_data) = { 0L }; - -static int blimit = 10; -static int qhimark = 10000; -static int qlowmark = 100; - -#ifdef CONFIG_SMP -static void force_quiescent_state(struct rcu_data *rdp, - struct rcu_ctrlblk *rcp) -{ - int cpu; - unsigned long flags; - - set_need_resched(); - spin_lock_irqsave(&rcp->lock, flags); - if (unlikely(!rcp->signaled)) { - rcp->signaled = 1; - /* - * Don't send IPI to itself. With irqs disabled, - * rdp->cpu is the current cpu. - * - * cpu_online_mask is updated by the _cpu_down() - * using __stop_machine(). Since we're in irqs disabled - * section, __stop_machine() is not exectuting, hence - * the cpu_online_mask is stable. - * - * However, a cpu might have been offlined _just_ before - * we disabled irqs while entering here. - * And rcu subsystem might not yet have handled the CPU_DEAD - * notification, leading to the offlined cpu's bit - * being set in the rcp->cpumask. - * - * Hence cpumask = (rcp->cpumask & cpu_online_mask) to prevent - * sending smp_reschedule() to an offlined CPU. - */ - for_each_cpu_and(cpu, - to_cpumask(rcp->cpumask), cpu_online_mask) { - if (cpu != rdp->cpu) - smp_send_reschedule(cpu); - } - } - spin_unlock_irqrestore(&rcp->lock, flags); -} -#else -static inline void force_quiescent_state(struct rcu_data *rdp, - struct rcu_ctrlblk *rcp) -{ - set_need_resched(); -} -#endif - -static void __call_rcu(struct rcu_head *head, struct rcu_ctrlblk *rcp, - struct rcu_data *rdp) -{ - long batch; - - head->next = NULL; - smp_mb(); /* Read of rcu->cur must happen after any change by caller. */ - - /* - * Determine the batch number of this callback. - * - * Using ACCESS_ONCE to avoid the following error when gcc eliminates - * local variable "batch" and emits codes like this: - * 1) rdp->batch = rcp->cur + 1 # gets old value - * ...... - * 2)rcu_batch_after(rcp->cur + 1, rdp->batch) # gets new value - * then [*nxttail[0], *nxttail[1]) may contain callbacks - * that batch# = rdp->batch, see the comment of struct rcu_data. - */ - batch = ACCESS_ONCE(rcp->cur) + 1; - - if (rdp->nxtlist && rcu_batch_after(batch, rdp->batch)) { - /* process callbacks */ - rdp->nxttail[0] = rdp->nxttail[1]; - rdp->nxttail[1] = rdp->nxttail[2]; - if (rcu_batch_after(batch - 1, rdp->batch)) - rdp->nxttail[0] = rdp->nxttail[2]; - } - - rdp->batch = batch; - *rdp->nxttail[2] = head; - rdp->nxttail[2] = &head->next; - - if (unlikely(++rdp->qlen > qhimark)) { - rdp->blimit = INT_MAX; - force_quiescent_state(rdp, &rcu_ctrlblk); - } -} - -#ifdef CONFIG_RCU_CPU_STALL_DETECTOR - -static void record_gp_stall_check_time(struct rcu_ctrlblk *rcp) -{ - rcp->gp_start = jiffies; - rcp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_CHECK; -} - -static void print_other_cpu_stall(struct rcu_ctrlblk *rcp) -{ - int cpu; - long delta; - unsigned long flags; - - /* Only let one CPU complain about others per time interval. */ - - spin_lock_irqsave(&rcp->lock, flags); - delta = jiffies - rcp->jiffies_stall; - if (delta < 2 || rcp->cur != rcp->completed) { - spin_unlock_irqrestore(&rcp->lock, flags); - return; - } - rcp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; - spin_unlock_irqrestore(&rcp->lock, flags); - - /* OK, time to rat on our buddy... */ - - printk(KERN_ERR "INFO: RCU detected CPU stalls:"); - for_each_possible_cpu(cpu) { - if (cpumask_test_cpu(cpu, to_cpumask(rcp->cpumask))) - printk(" %d", cpu); - } - printk(" (detected by %d, t=%ld jiffies)\n", - smp_processor_id(), (long)(jiffies - rcp->gp_start)); -} - -static void print_cpu_stall(struct rcu_ctrlblk *rcp) -{ - unsigned long flags; - - printk(KERN_ERR "INFO: RCU detected CPU %d stall (t=%lu/%lu jiffies)\n", - smp_processor_id(), jiffies, - jiffies - rcp->gp_start); - dump_stack(); - spin_lock_irqsave(&rcp->lock, flags); - if ((long)(jiffies - rcp->jiffies_stall) >= 0) - rcp->jiffies_stall = - jiffies + RCU_SECONDS_TILL_STALL_RECHECK; - spin_unlock_irqrestore(&rcp->lock, flags); - set_need_resched(); /* kick ourselves to get things going. */ -} - -static void check_cpu_stall(struct rcu_ctrlblk *rcp) -{ - long delta; - - delta = jiffies - rcp->jiffies_stall; - if (cpumask_test_cpu(smp_processor_id(), to_cpumask(rcp->cpumask)) && - delta >= 0) { - - /* We haven't checked in, so go dump stack. */ - print_cpu_stall(rcp); - - } else if (rcp->cur != rcp->completed && delta >= 2) { - - /* They had two seconds to dump stack, so complain. */ - print_other_cpu_stall(rcp); - } -} - -#else /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ - -static void record_gp_stall_check_time(struct rcu_ctrlblk *rcp) -{ -} - -static inline void check_cpu_stall(struct rcu_ctrlblk *rcp) -{ -} - -#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ - -/** - * call_rcu - Queue an RCU callback for invocation after a grace period. - * @head: structure to be used for queueing the RCU updates. - * @func: actual update function to be invoked after the grace period - * - * The update function will be invoked some time after a full grace - * period elapses, in other words after all currently executing RCU - * read-side critical sections have completed. RCU read-side critical - * sections are delimited by rcu_read_lock() and rcu_read_unlock(), - * and may be nested. - */ -void call_rcu(struct rcu_head *head, - void (*func)(struct rcu_head *rcu)) -{ - unsigned long flags; - - head->func = func; - local_irq_save(flags); - __call_rcu(head, &rcu_ctrlblk, &__get_cpu_var(rcu_data)); - local_irq_restore(flags); -} -EXPORT_SYMBOL_GPL(call_rcu); - -/** - * call_rcu_bh - Queue an RCU for invocation after a quicker grace period. - * @head: structure to be used for queueing the RCU updates. - * @func: actual update function to be invoked after the grace period - * - * The update function will be invoked some time after a full grace - * period elapses, in other words after all currently executing RCU - * read-side critical sections have completed. call_rcu_bh() assumes - * that the read-side critical sections end on completion of a softirq - * handler. This means that read-side critical sections in process - * context must not be interrupted by softirqs. This interface is to be - * used when most of the read-side critical sections are in softirq context. - * RCU read-side critical sections are delimited by rcu_read_lock() and - * rcu_read_unlock(), * if in interrupt context or rcu_read_lock_bh() - * and rcu_read_unlock_bh(), if in process context. These may be nested. - */ -void call_rcu_bh(struct rcu_head *head, - void (*func)(struct rcu_head *rcu)) -{ - unsigned long flags; - - head->func = func; - local_irq_save(flags); - __call_rcu(head, &rcu_bh_ctrlblk, &__get_cpu_var(rcu_bh_data)); - local_irq_restore(flags); -} -EXPORT_SYMBOL_GPL(call_rcu_bh); - -/* - * Return the number of RCU batches processed thus far. Useful - * for debug and statistics. - */ -long rcu_batches_completed(void) -{ - return rcu_ctrlblk.completed; -} -EXPORT_SYMBOL_GPL(rcu_batches_completed); - -/* - * Return the number of RCU batches processed thus far. Useful - * for debug and statistics. - */ -long rcu_batches_completed_bh(void) -{ - return rcu_bh_ctrlblk.completed; -} -EXPORT_SYMBOL_GPL(rcu_batches_completed_bh); - -/* Raises the softirq for processing rcu_callbacks. */ -static inline void raise_rcu_softirq(void) -{ - raise_softirq(RCU_SOFTIRQ); -} - -/* - * Invoke the completed RCU callbacks. They are expected to be in - * a per-cpu list. - */ -static void rcu_do_batch(struct rcu_data *rdp) -{ - unsigned long flags; - struct rcu_head *next, *list; - int count = 0; - - list = rdp->donelist; - while (list) { - next = list->next; - prefetch(next); - list->func(list); - list = next; - if (++count >= rdp->blimit) - break; - } - rdp->donelist = list; - - local_irq_save(flags); - rdp->qlen -= count; - local_irq_restore(flags); - if (rdp->blimit == INT_MAX && rdp->qlen <= qlowmark) - rdp->blimit = blimit; - - if (!rdp->donelist) - rdp->donetail = &rdp->donelist; - else - raise_rcu_softirq(); -} - -/* - * Grace period handling: - * The grace period handling consists out of two steps: - * - A new grace period is started. - * This is done by rcu_start_batch. The start is not broadcasted to - * all cpus, they must pick this up by comparing rcp->cur with - * rdp->quiescbatch. All cpus are recorded in the - * rcu_ctrlblk.cpumask bitmap. - * - All cpus must go through a quiescent state. - * Since the start of the grace period is not broadcasted, at least two - * calls to rcu_check_quiescent_state are required: - * The first call just notices that a new grace period is running. The - * following calls check if there was a quiescent state since the beginning - * of the grace period. If so, it updates rcu_ctrlblk.cpumask. If - * the bitmap is empty, then the grace period is completed. - * rcu_check_quiescent_state calls rcu_start_batch(0) to start the next grace - * period (if necessary). - */ - -/* - * Register a new batch of callbacks, and start it up if there is currently no - * active batch and the batch to be registered has not already occurred. - * Caller must hold rcu_ctrlblk.lock. - */ -static void rcu_start_batch(struct rcu_ctrlblk *rcp) -{ - if (rcp->cur != rcp->pending && - rcp->completed == rcp->cur) { - rcp->cur++; - record_gp_stall_check_time(rcp); - - /* - * Accessing nohz_cpu_mask before incrementing rcp->cur needs a - * Barrier Otherwise it can cause tickless idle CPUs to be - * included in rcp->cpumask, which will extend graceperiods - * unnecessarily. - */ - smp_mb(); - cpumask_andnot(to_cpumask(rcp->cpumask), - cpu_online_mask, nohz_cpu_mask); - - rcp->signaled = 0; - } -} - -/* - * cpu went through a quiescent state since the beginning of the grace period. - * Clear it from the cpu mask and complete the grace period if it was the last - * cpu. Start another grace period if someone has further entries pending - */ -static void cpu_quiet(int cpu, struct rcu_ctrlblk *rcp) -{ - cpumask_clear_cpu(cpu, to_cpumask(rcp->cpumask)); - if (cpumask_empty(to_cpumask(rcp->cpumask))) { - /* batch completed ! */ - rcp->completed = rcp->cur; - rcu_start_batch(rcp); - } -} - -/* - * Check if the cpu has gone through a quiescent state (say context - * switch). If so and if it already hasn't done so in this RCU - * quiescent cycle, then indicate that it has done so. - */ -static void rcu_check_quiescent_state(struct rcu_ctrlblk *rcp, - struct rcu_data *rdp) -{ - unsigned long flags; - - if (rdp->quiescbatch != rcp->cur) { - /* start new grace period: */ - rdp->qs_pending = 1; - rdp->passed_quiesc = 0; - rdp->quiescbatch = rcp->cur; - return; - } - - /* Grace period already completed for this cpu? - * qs_pending is checked instead of the actual bitmap to avoid - * cacheline trashing. - */ - if (!rdp->qs_pending) - return; - - /* - * Was there a quiescent state since the beginning of the grace - * period? If no, then exit and wait for the next call. - */ - if (!rdp->passed_quiesc) - return; - rdp->qs_pending = 0; - - spin_lock_irqsave(&rcp->lock, flags); - /* - * rdp->quiescbatch/rcp->cur and the cpu bitmap can come out of sync - * during cpu startup. Ignore the quiescent state. - */ - if (likely(rdp->quiescbatch == rcp->cur)) - cpu_quiet(rdp->cpu, rcp); - - spin_unlock_irqrestore(&rcp->lock, flags); -} - - -#ifdef CONFIG_HOTPLUG_CPU - -/* warning! helper for rcu_offline_cpu. do not use elsewhere without reviewing - * locking requirements, the list it's pulling from has to belong to a cpu - * which is dead and hence not processing interrupts. - */ -static void rcu_move_batch(struct rcu_data *this_rdp, struct rcu_head *list, - struct rcu_head **tail, long batch) -{ - unsigned long flags; - - if (list) { - local_irq_save(flags); - this_rdp->batch = batch; - *this_rdp->nxttail[2] = list; - this_rdp->nxttail[2] = tail; - local_irq_restore(flags); - } -} - -static void __rcu_offline_cpu(struct rcu_data *this_rdp, - struct rcu_ctrlblk *rcp, struct rcu_data *rdp) -{ - unsigned long flags; - - /* - * if the cpu going offline owns the grace period - * we can block indefinitely waiting for it, so flush - * it here - */ - spin_lock_irqsave(&rcp->lock, flags); - if (rcp->cur != rcp->completed) - cpu_quiet(rdp->cpu, rcp); - rcu_move_batch(this_rdp, rdp->donelist, rdp->donetail, rcp->cur + 1); - rcu_move_batch(this_rdp, rdp->nxtlist, rdp->nxttail[2], rcp->cur + 1); - spin_unlock(&rcp->lock); - - this_rdp->qlen += rdp->qlen; - local_irq_restore(flags); -} - -static void rcu_offline_cpu(int cpu) -{ - struct rcu_data *this_rdp = &get_cpu_var(rcu_data); - struct rcu_data *this_bh_rdp = &get_cpu_var(rcu_bh_data); - - __rcu_offline_cpu(this_rdp, &rcu_ctrlblk, - &per_cpu(rcu_data, cpu)); - __rcu_offline_cpu(this_bh_rdp, &rcu_bh_ctrlblk, - &per_cpu(rcu_bh_data, cpu)); - put_cpu_var(rcu_data); - put_cpu_var(rcu_bh_data); -} - -#else - -static void rcu_offline_cpu(int cpu) -{ -} - -#endif - -/* - * This does the RCU processing work from softirq context. - */ -static void __rcu_process_callbacks(struct rcu_ctrlblk *rcp, - struct rcu_data *rdp) -{ - unsigned long flags; - long completed_snap; - - if (rdp->nxtlist) { - local_irq_save(flags); - completed_snap = ACCESS_ONCE(rcp->completed); - - /* - * move the other grace-period-completed entries to - * [rdp->nxtlist, *rdp->nxttail[0]) temporarily - */ - if (!rcu_batch_before(completed_snap, rdp->batch)) - rdp->nxttail[0] = rdp->nxttail[1] = rdp->nxttail[2]; - else if (!rcu_batch_before(completed_snap, rdp->batch - 1)) - rdp->nxttail[0] = rdp->nxttail[1]; - - /* - * the grace period for entries in - * [rdp->nxtlist, *rdp->nxttail[0]) has completed and - * move these entries to donelist - */ - if (rdp->nxttail[0] != &rdp->nxtlist) { - *rdp->donetail = rdp->nxtlist; - rdp->donetail = rdp->nxttail[0]; - rdp->nxtlist = *rdp->nxttail[0]; - *rdp->donetail = NULL; - - if (rdp->nxttail[1] == rdp->nxttail[0]) - rdp->nxttail[1] = &rdp->nxtlist; - if (rdp->nxttail[2] == rdp->nxttail[0]) - rdp->nxttail[2] = &rdp->nxtlist; - rdp->nxttail[0] = &rdp->nxtlist; - } - - local_irq_restore(flags); - - if (rcu_batch_after(rdp->batch, rcp->pending)) { - unsigned long flags2; - - /* and start it/schedule start if it's a new batch */ - spin_lock_irqsave(&rcp->lock, flags2); - if (rcu_batch_after(rdp->batch, rcp->pending)) { - rcp->pending = rdp->batch; - rcu_start_batch(rcp); - } - spin_unlock_irqrestore(&rcp->lock, flags2); - } - } - - rcu_check_quiescent_state(rcp, rdp); - if (rdp->donelist) - rcu_do_batch(rdp); -} - -static void rcu_process_callbacks(struct softirq_action *unused) -{ - /* - * Memory references from any prior RCU read-side critical sections - * executed by the interrupted code must be see before any RCU - * grace-period manupulations below. - */ - - smp_mb(); /* See above block comment. */ - - __rcu_process_callbacks(&rcu_ctrlblk, &__get_cpu_var(rcu_data)); - __rcu_process_callbacks(&rcu_bh_ctrlblk, &__get_cpu_var(rcu_bh_data)); - - /* - * Memory references from any later RCU read-side critical sections - * executed by the interrupted code must be see after any RCU - * grace-period manupulations above. - */ - - smp_mb(); /* See above block comment. */ -} - -static int __rcu_pending(struct rcu_ctrlblk *rcp, struct rcu_data *rdp) -{ - /* Check for CPU stalls, if enabled. */ - check_cpu_stall(rcp); - - if (rdp->nxtlist) { - long completed_snap = ACCESS_ONCE(rcp->completed); - - /* - * This cpu has pending rcu entries and the grace period - * for them has completed. - */ - if (!rcu_batch_before(completed_snap, rdp->batch)) - return 1; - if (!rcu_batch_before(completed_snap, rdp->batch - 1) && - rdp->nxttail[0] != rdp->nxttail[1]) - return 1; - if (rdp->nxttail[0] != &rdp->nxtlist) - return 1; - - /* - * This cpu has pending rcu entries and the new batch - * for then hasn't been started nor scheduled start - */ - if (rcu_batch_after(rdp->batch, rcp->pending)) - return 1; - } - - /* This cpu has finished callbacks to invoke */ - if (rdp->donelist) - return 1; - - /* The rcu core waits for a quiescent state from the cpu */ - if (rdp->quiescbatch != rcp->cur || rdp->qs_pending) - return 1; - - /* nothing to do */ - return 0; -} - -/* - * Check to see if there is any immediate RCU-related work to be done - * by the current CPU, returning 1 if so. This function is part of the - * RCU implementation; it is -not- an exported member of the RCU API. - */ -int rcu_pending(int cpu) -{ - return __rcu_pending(&rcu_ctrlblk, &per_cpu(rcu_data, cpu)) || - __rcu_pending(&rcu_bh_ctrlblk, &per_cpu(rcu_bh_data, cpu)); -} - -/* - * Check to see if any future RCU-related work will need to be done - * by the current CPU, even if none need be done immediately, returning - * 1 if so. This function is part of the RCU implementation; it is -not- - * an exported member of the RCU API. - */ -int rcu_needs_cpu(int cpu) -{ - struct rcu_data *rdp = &per_cpu(rcu_data, cpu); - struct rcu_data *rdp_bh = &per_cpu(rcu_bh_data, cpu); - - return !!rdp->nxtlist || !!rdp_bh->nxtlist || rcu_pending(cpu); -} - -/* - * Top-level function driving RCU grace-period detection, normally - * invoked from the scheduler-clock interrupt. This function simply - * increments counters that are read only from softirq by this same - * CPU, so there are no memory barriers required. - */ -void rcu_check_callbacks(int cpu, int user) -{ - if (user || - (idle_cpu(cpu) && !in_softirq() && - hardirq_count() <= (1 << HARDIRQ_SHIFT))) { - - /* - * Get here if this CPU took its interrupt from user - * mode or from the idle loop, and if this is not a - * nested interrupt. In this case, the CPU is in - * a quiescent state, so count it. - * - * Also do a memory barrier. This is needed to handle - * the case where writes from a preempt-disable section - * of code get reordered into schedule() by this CPU's - * write buffer. The memory barrier makes sure that - * the rcu_qsctr_inc() and rcu_bh_qsctr_inc() are see - * by other CPUs to happen after any such write. - */ - - smp_mb(); /* See above block comment. */ - rcu_qsctr_inc(cpu); - rcu_bh_qsctr_inc(cpu); - - } else if (!in_softirq()) { - - /* - * Get here if this CPU did not take its interrupt from - * softirq, in other words, if it is not interrupting - * a rcu_bh read-side critical section. This is an _bh - * critical section, so count it. The memory barrier - * is needed for the same reason as is the above one. - */ - - smp_mb(); /* See above block comment. */ - rcu_bh_qsctr_inc(cpu); - } - raise_rcu_softirq(); -} - -static void rcu_init_percpu_data(int cpu, struct rcu_ctrlblk *rcp, - struct rcu_data *rdp) -{ - unsigned long flags; - - spin_lock_irqsave(&rcp->lock, flags); - memset(rdp, 0, sizeof(*rdp)); - rdp->nxttail[0] = rdp->nxttail[1] = rdp->nxttail[2] = &rdp->nxtlist; - rdp->donetail = &rdp->donelist; - rdp->quiescbatch = rcp->completed; - rdp->qs_pending = 0; - rdp->cpu = cpu; - rdp->blimit = blimit; - spin_unlock_irqrestore(&rcp->lock, flags); -} - -static void __cpuinit rcu_online_cpu(int cpu) -{ - struct rcu_data *rdp = &per_cpu(rcu_data, cpu); - struct rcu_data *bh_rdp = &per_cpu(rcu_bh_data, cpu); - - rcu_init_percpu_data(cpu, &rcu_ctrlblk, rdp); - rcu_init_percpu_data(cpu, &rcu_bh_ctrlblk, bh_rdp); - open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); -} - -static int __cpuinit rcu_cpu_notify(struct notifier_block *self, - unsigned long action, void *hcpu) -{ - long cpu = (long)hcpu; - - switch (action) { - case CPU_UP_PREPARE: - case CPU_UP_PREPARE_FROZEN: - rcu_online_cpu(cpu); - break; - case CPU_DEAD: - case CPU_DEAD_FROZEN: - rcu_offline_cpu(cpu); - break; - default: - break; - } - return NOTIFY_OK; -} - -static struct notifier_block __cpuinitdata rcu_nb = { - .notifier_call = rcu_cpu_notify, -}; - -/* - * Initializes rcu mechanism. Assumed to be called early. - * That is before local timer(SMP) or jiffie timer (uniproc) is setup. - * Note that rcu_qsctr and friends are implicitly - * initialized due to the choice of ``0'' for RCU_CTR_INVALID. - */ -void __init __rcu_init(void) -{ -#ifdef CONFIG_RCU_CPU_STALL_DETECTOR - printk(KERN_INFO "RCU-based detection of stalled CPUs is enabled.\n"); -#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ - rcu_cpu_notify(&rcu_nb, CPU_UP_PREPARE, - (void *)(long)smp_processor_id()); - /* Register notifier for non-boot CPUs */ - register_cpu_notifier(&rcu_nb); -} - -module_param(blimit, int, 0); -module_param(qhimark, int, 0); -module_param(qlowmark, int, 0); diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c index d92a76a881a..bd5d5c8e514 100644 --- a/kernel/rcupdate.c +++ b/kernel/rcupdate.c @@ -44,6 +44,7 @@ #include <linux/cpu.h> #include <linux/mutex.h> #include <linux/module.h> +#include <linux/kernel_stat.h> enum rcu_barrier { RCU_BARRIER_STD, @@ -55,6 +56,11 @@ static DEFINE_PER_CPU(struct rcu_head, rcu_barrier_head) = {NULL}; static atomic_t rcu_barrier_cpu_count; static DEFINE_MUTEX(rcu_barrier_mutex); static struct completion rcu_barrier_completion; +int rcu_scheduler_active __read_mostly; + +static atomic_t rcu_migrate_type_count = ATOMIC_INIT(0); +static struct rcu_head rcu_migrate_head[3]; +static DECLARE_WAIT_QUEUE_HEAD(rcu_migrate_wq); /* * Awaken the corresponding synchronize_rcu() instance now that a @@ -80,6 +86,10 @@ void wakeme_after_rcu(struct rcu_head *head) void synchronize_rcu(void) { struct rcu_synchronize rcu; + + if (rcu_blocking_is_gp()) + return; + init_completion(&rcu.completion); /* Will wake me after RCU finished. */ call_rcu(&rcu.head, wakeme_after_rcu); @@ -88,6 +98,30 @@ void synchronize_rcu(void) } EXPORT_SYMBOL_GPL(synchronize_rcu); +/** + * synchronize_rcu_bh - wait until an rcu_bh grace period has elapsed. + * + * Control will return to the caller some time after a full rcu_bh grace + * period has elapsed, in other words after all currently executing rcu_bh + * read-side critical sections have completed. RCU read-side critical + * sections are delimited by rcu_read_lock_bh() and rcu_read_unlock_bh(), + * and may be nested. + */ +void synchronize_rcu_bh(void) +{ + struct rcu_synchronize rcu; + + if (rcu_blocking_is_gp()) + return; + + init_completion(&rcu.completion); + /* Will wake me after RCU finished. */ + call_rcu_bh(&rcu.head, wakeme_after_rcu); + /* Wait for it. */ + wait_for_completion(&rcu.completion); +} +EXPORT_SYMBOL_GPL(synchronize_rcu_bh); + static void rcu_barrier_callback(struct rcu_head *notused) { if (atomic_dec_and_test(&rcu_barrier_cpu_count)) @@ -116,6 +150,12 @@ static void rcu_barrier_func(void *type) } } +static inline void wait_migrated_callbacks(void) +{ + wait_event(rcu_migrate_wq, !atomic_read(&rcu_migrate_type_count)); + smp_mb(); /* In case we didn't sleep. */ +} + /* * Orchestrate the specified type of RCU barrier, waiting for all * RCU callbacks of the specified type to complete. @@ -141,6 +181,7 @@ static void _rcu_barrier(enum rcu_barrier type) complete(&rcu_barrier_completion); wait_for_completion(&rcu_barrier_completion); mutex_unlock(&rcu_barrier_mutex); + wait_migrated_callbacks(); } /** @@ -170,8 +211,61 @@ void rcu_barrier_sched(void) } EXPORT_SYMBOL_GPL(rcu_barrier_sched); +static void rcu_migrate_callback(struct rcu_head *notused) +{ + if (atomic_dec_and_test(&rcu_migrate_type_count)) + wake_up(&rcu_migrate_wq); +} + +extern int rcu_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu); + +static int __cpuinit rcu_barrier_cpu_hotplug(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + rcu_cpu_notify(self, action, hcpu); + if (action == CPU_DYING) { + /* + * preempt_disable() in on_each_cpu() prevents stop_machine(), + * so when "on_each_cpu(rcu_barrier_func, (void *)type, 1);" + * returns, all online cpus have queued rcu_barrier_func(), + * and the dead cpu(if it exist) queues rcu_migrate_callback()s. + * + * These callbacks ensure _rcu_barrier() waits for all + * RCU callbacks of the specified type to complete. + */ + atomic_set(&rcu_migrate_type_count, 3); + call_rcu_bh(rcu_migrate_head, rcu_migrate_callback); + call_rcu_sched(rcu_migrate_head + 1, rcu_migrate_callback); + call_rcu(rcu_migrate_head + 2, rcu_migrate_callback); + } else if (action == CPU_DOWN_PREPARE) { + /* Don't need to wait until next removal operation. */ + /* rcu_migrate_head is protected by cpu_add_remove_lock */ + wait_migrated_callbacks(); + } + + return NOTIFY_OK; +} + void __init rcu_init(void) { + int i; + __rcu_init(); + cpu_notifier(rcu_barrier_cpu_hotplug, 0); + + /* + * We don't need protection against CPU-hotplug here because + * this is called early in boot, before either interrupts + * or the scheduler are operational. + */ + for_each_online_cpu(i) + rcu_barrier_cpu_hotplug(NULL, CPU_UP_PREPARE, (void *)(long)i); } +void rcu_scheduler_starting(void) +{ + WARN_ON(num_online_cpus() != 1); + WARN_ON(nr_context_switches() > 0); + rcu_scheduler_active = 1; +} diff --git a/kernel/rcupreempt.c b/kernel/rcupreempt.c deleted file mode 100644 index 33cfc50781f..00000000000 --- a/kernel/rcupreempt.c +++ /dev/null @@ -1,1502 +0,0 @@ -/* - * Read-Copy Update mechanism for mutual exclusion, realtime implementation - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright IBM Corporation, 2006 - * - * Authors: Paul E. McKenney <paulmck@us.ibm.com> - * With thanks to Esben Nielsen, Bill Huey, and Ingo Molnar - * for pushing me away from locks and towards counters, and - * to Suparna Bhattacharya for pushing me completely away - * from atomic instructions on the read side. - * - * - Added handling of Dynamic Ticks - * Copyright 2007 - Paul E. Mckenney <paulmck@us.ibm.com> - * - Steven Rostedt <srostedt@redhat.com> - * - * Papers: http://www.rdrop.com/users/paulmck/RCU - * - * Design Document: http://lwn.net/Articles/253651/ - * - * For detailed explanation of Read-Copy Update mechanism see - - * Documentation/RCU/ *.txt - * - */ -#include <linux/types.h> -#include <linux/kernel.h> -#include <linux/init.h> -#include <linux/spinlock.h> -#include <linux/smp.h> -#include <linux/rcupdate.h> -#include <linux/interrupt.h> -#include <linux/sched.h> -#include <asm/atomic.h> -#include <linux/bitops.h> -#include <linux/module.h> -#include <linux/kthread.h> -#include <linux/completion.h> -#include <linux/moduleparam.h> -#include <linux/percpu.h> -#include <linux/notifier.h> -#include <linux/cpu.h> -#include <linux/random.h> -#include <linux/delay.h> -#include <linux/cpumask.h> -#include <linux/rcupreempt_trace.h> -#include <asm/byteorder.h> - -/* - * PREEMPT_RCU data structures. - */ - -/* - * GP_STAGES specifies the number of times the state machine has - * to go through the all the rcu_try_flip_states (see below) - * in a single Grace Period. - * - * GP in GP_STAGES stands for Grace Period ;) - */ -#define GP_STAGES 2 -struct rcu_data { - spinlock_t lock; /* Protect rcu_data fields. */ - long completed; /* Number of last completed batch. */ - int waitlistcount; - struct rcu_head *nextlist; - struct rcu_head **nexttail; - struct rcu_head *waitlist[GP_STAGES]; - struct rcu_head **waittail[GP_STAGES]; - struct rcu_head *donelist; /* from waitlist & waitschedlist */ - struct rcu_head **donetail; - long rcu_flipctr[2]; - struct rcu_head *nextschedlist; - struct rcu_head **nextschedtail; - struct rcu_head *waitschedlist; - struct rcu_head **waitschedtail; - int rcu_sched_sleeping; -#ifdef CONFIG_RCU_TRACE - struct rcupreempt_trace trace; -#endif /* #ifdef CONFIG_RCU_TRACE */ -}; - -/* - * States for rcu_try_flip() and friends. - */ - -enum rcu_try_flip_states { - - /* - * Stay here if nothing is happening. Flip the counter if somthing - * starts happening. Denoted by "I" - */ - rcu_try_flip_idle_state, - - /* - * Wait here for all CPUs to notice that the counter has flipped. This - * prevents the old set of counters from ever being incremented once - * we leave this state, which in turn is necessary because we cannot - * test any individual counter for zero -- we can only check the sum. - * Denoted by "A". - */ - rcu_try_flip_waitack_state, - - /* - * Wait here for the sum of the old per-CPU counters to reach zero. - * Denoted by "Z". - */ - rcu_try_flip_waitzero_state, - - /* - * Wait here for each of the other CPUs to execute a memory barrier. - * This is necessary to ensure that these other CPUs really have - * completed executing their RCU read-side critical sections, despite - * their CPUs wildly reordering memory. Denoted by "M". - */ - rcu_try_flip_waitmb_state, -}; - -/* - * States for rcu_ctrlblk.rcu_sched_sleep. - */ - -enum rcu_sched_sleep_states { - rcu_sched_not_sleeping, /* Not sleeping, callbacks need GP. */ - rcu_sched_sleep_prep, /* Thinking of sleeping, rechecking. */ - rcu_sched_sleeping, /* Sleeping, awaken if GP needed. */ -}; - -struct rcu_ctrlblk { - spinlock_t fliplock; /* Protect state-machine transitions. */ - long completed; /* Number of last completed batch. */ - enum rcu_try_flip_states rcu_try_flip_state; /* The current state of - the rcu state machine */ - spinlock_t schedlock; /* Protect rcu_sched sleep state. */ - enum rcu_sched_sleep_states sched_sleep; /* rcu_sched state. */ - wait_queue_head_t sched_wq; /* Place for rcu_sched to sleep. */ -}; - -static DEFINE_PER_CPU(struct rcu_data, rcu_data); -static struct rcu_ctrlblk rcu_ctrlblk = { - .fliplock = __SPIN_LOCK_UNLOCKED(rcu_ctrlblk.fliplock), - .completed = 0, - .rcu_try_flip_state = rcu_try_flip_idle_state, - .schedlock = __SPIN_LOCK_UNLOCKED(rcu_ctrlblk.schedlock), - .sched_sleep = rcu_sched_not_sleeping, - .sched_wq = __WAIT_QUEUE_HEAD_INITIALIZER(rcu_ctrlblk.sched_wq), -}; - -static struct task_struct *rcu_sched_grace_period_task; - -#ifdef CONFIG_RCU_TRACE -static char *rcu_try_flip_state_names[] = - { "idle", "waitack", "waitzero", "waitmb" }; -#endif /* #ifdef CONFIG_RCU_TRACE */ - -static DECLARE_BITMAP(rcu_cpu_online_map, NR_CPUS) __read_mostly - = CPU_BITS_NONE; - -/* - * Enum and per-CPU flag to determine when each CPU has seen - * the most recent counter flip. - */ - -enum rcu_flip_flag_values { - rcu_flip_seen, /* Steady/initial state, last flip seen. */ - /* Only GP detector can update. */ - rcu_flipped /* Flip just completed, need confirmation. */ - /* Only corresponding CPU can update. */ -}; -static DEFINE_PER_CPU_SHARED_ALIGNED(enum rcu_flip_flag_values, rcu_flip_flag) - = rcu_flip_seen; - -/* - * Enum and per-CPU flag to determine when each CPU has executed the - * needed memory barrier to fence in memory references from its last RCU - * read-side critical section in the just-completed grace period. - */ - -enum rcu_mb_flag_values { - rcu_mb_done, /* Steady/initial state, no mb()s required. */ - /* Only GP detector can update. */ - rcu_mb_needed /* Flip just completed, need an mb(). */ - /* Only corresponding CPU can update. */ -}; -static DEFINE_PER_CPU_SHARED_ALIGNED(enum rcu_mb_flag_values, rcu_mb_flag) - = rcu_mb_done; - -/* - * RCU_DATA_ME: find the current CPU's rcu_data structure. - * RCU_DATA_CPU: find the specified CPU's rcu_data structure. - */ -#define RCU_DATA_ME() (&__get_cpu_var(rcu_data)) -#define RCU_DATA_CPU(cpu) (&per_cpu(rcu_data, cpu)) - -/* - * Helper macro for tracing when the appropriate rcu_data is not - * cached in a local variable, but where the CPU number is so cached. - */ -#define RCU_TRACE_CPU(f, cpu) RCU_TRACE(f, &(RCU_DATA_CPU(cpu)->trace)); - -/* - * Helper macro for tracing when the appropriate rcu_data is not - * cached in a local variable. - */ -#define RCU_TRACE_ME(f) RCU_TRACE(f, &(RCU_DATA_ME()->trace)); - -/* - * Helper macro for tracing when the appropriate rcu_data is pointed - * to by a local variable. - */ -#define RCU_TRACE_RDP(f, rdp) RCU_TRACE(f, &((rdp)->trace)); - -#define RCU_SCHED_BATCH_TIME (HZ / 50) - -/* - * Return the number of RCU batches processed thus far. Useful - * for debug and statistics. - */ -long rcu_batches_completed(void) -{ - return rcu_ctrlblk.completed; -} -EXPORT_SYMBOL_GPL(rcu_batches_completed); - -void __rcu_read_lock(void) -{ - int idx; - struct task_struct *t = current; - int nesting; - - nesting = ACCESS_ONCE(t->rcu_read_lock_nesting); - if (nesting != 0) { - - /* An earlier rcu_read_lock() covers us, just count it. */ - - t->rcu_read_lock_nesting = nesting + 1; - - } else { - unsigned long flags; - - /* - * We disable interrupts for the following reasons: - * - If we get scheduling clock interrupt here, and we - * end up acking the counter flip, it's like a promise - * that we will never increment the old counter again. - * Thus we will break that promise if that - * scheduling clock interrupt happens between the time - * we pick the .completed field and the time that we - * increment our counter. - * - * - We don't want to be preempted out here. - * - * NMIs can still occur, of course, and might themselves - * contain rcu_read_lock(). - */ - - local_irq_save(flags); - - /* - * Outermost nesting of rcu_read_lock(), so increment - * the current counter for the current CPU. Use volatile - * casts to prevent the compiler from reordering. - */ - - idx = ACCESS_ONCE(rcu_ctrlblk.completed) & 0x1; - ACCESS_ONCE(RCU_DATA_ME()->rcu_flipctr[idx])++; - - /* - * Now that the per-CPU counter has been incremented, we - * are protected from races with rcu_read_lock() invoked - * from NMI handlers on this CPU. We can therefore safely - * increment the nesting counter, relieving further NMIs - * of the need to increment the per-CPU counter. - */ - - ACCESS_ONCE(t->rcu_read_lock_nesting) = nesting + 1; - - /* - * Now that we have preventing any NMIs from storing - * to the ->rcu_flipctr_idx, we can safely use it to - * remember which counter to decrement in the matching - * rcu_read_unlock(). - */ - - ACCESS_ONCE(t->rcu_flipctr_idx) = idx; - local_irq_restore(flags); - } -} -EXPORT_SYMBOL_GPL(__rcu_read_lock); - -void __rcu_read_unlock(void) -{ - int idx; - struct task_struct *t = current; - int nesting; - - nesting = ACCESS_ONCE(t->rcu_read_lock_nesting); - if (nesting > 1) { - - /* - * We are still protected by the enclosing rcu_read_lock(), - * so simply decrement the counter. - */ - - t->rcu_read_lock_nesting = nesting - 1; - - } else { - unsigned long flags; - - /* - * Disable local interrupts to prevent the grace-period - * detection state machine from seeing us half-done. - * NMIs can still occur, of course, and might themselves - * contain rcu_read_lock() and rcu_read_unlock(). - */ - - local_irq_save(flags); - - /* - * Outermost nesting of rcu_read_unlock(), so we must - * decrement the current counter for the current CPU. - * This must be done carefully, because NMIs can - * occur at any point in this code, and any rcu_read_lock() - * and rcu_read_unlock() pairs in the NMI handlers - * must interact non-destructively with this code. - * Lots of volatile casts, and -very- careful ordering. - * - * Changes to this code, including this one, must be - * inspected, validated, and tested extremely carefully!!! - */ - - /* - * First, pick up the index. - */ - - idx = ACCESS_ONCE(t->rcu_flipctr_idx); - - /* - * Now that we have fetched the counter index, it is - * safe to decrement the per-task RCU nesting counter. - * After this, any interrupts or NMIs will increment and - * decrement the per-CPU counters. - */ - ACCESS_ONCE(t->rcu_read_lock_nesting) = nesting - 1; - - /* - * It is now safe to decrement this task's nesting count. - * NMIs that occur after this statement will route their - * rcu_read_lock() calls through this "else" clause, and - * will thus start incrementing the per-CPU counter on - * their own. They will also clobber ->rcu_flipctr_idx, - * but that is OK, since we have already fetched it. - */ - - ACCESS_ONCE(RCU_DATA_ME()->rcu_flipctr[idx])--; - local_irq_restore(flags); - } -} -EXPORT_SYMBOL_GPL(__rcu_read_unlock); - -/* - * If a global counter flip has occurred since the last time that we - * advanced callbacks, advance them. Hardware interrupts must be - * disabled when calling this function. - */ -static void __rcu_advance_callbacks(struct rcu_data *rdp) -{ - int cpu; - int i; - int wlc = 0; - - if (rdp->completed != rcu_ctrlblk.completed) { - if (rdp->waitlist[GP_STAGES - 1] != NULL) { - *rdp->donetail = rdp->waitlist[GP_STAGES - 1]; - rdp->donetail = rdp->waittail[GP_STAGES - 1]; - RCU_TRACE_RDP(rcupreempt_trace_move2done, rdp); - } - for (i = GP_STAGES - 2; i >= 0; i--) { - if (rdp->waitlist[i] != NULL) { - rdp->waitlist[i + 1] = rdp->waitlist[i]; - rdp->waittail[i + 1] = rdp->waittail[i]; - wlc++; - } else { - rdp->waitlist[i + 1] = NULL; - rdp->waittail[i + 1] = - &rdp->waitlist[i + 1]; - } - } - if (rdp->nextlist != NULL) { - rdp->waitlist[0] = rdp->nextlist; - rdp->waittail[0] = rdp->nexttail; - wlc++; - rdp->nextlist = NULL; - rdp->nexttail = &rdp->nextlist; - RCU_TRACE_RDP(rcupreempt_trace_move2wait, rdp); - } else { - rdp->waitlist[0] = NULL; - rdp->waittail[0] = &rdp->waitlist[0]; - } - rdp->waitlistcount = wlc; - rdp->completed = rcu_ctrlblk.completed; - } - - /* - * Check to see if this CPU needs to report that it has seen - * the most recent counter flip, thereby declaring that all - * subsequent rcu_read_lock() invocations will respect this flip. - */ - - cpu = raw_smp_processor_id(); - if (per_cpu(rcu_flip_flag, cpu) == rcu_flipped) { - smp_mb(); /* Subsequent counter accesses must see new value */ - per_cpu(rcu_flip_flag, cpu) = rcu_flip_seen; - smp_mb(); /* Subsequent RCU read-side critical sections */ - /* seen -after- acknowledgement. */ - } -} - -DEFINE_PER_CPU_SHARED_ALIGNED(struct rcu_dyntick_sched, rcu_dyntick_sched) = { - .dynticks = 1, -}; - -#ifdef CONFIG_NO_HZ -static DEFINE_PER_CPU(int, rcu_update_flag); - -/** - * rcu_irq_enter - Called from Hard irq handlers and NMI/SMI. - * - * If the CPU was idle with dynamic ticks active, this updates the - * rcu_dyntick_sched.dynticks to let the RCU handling know that the - * CPU is active. - */ -void rcu_irq_enter(void) -{ - int cpu = smp_processor_id(); - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - if (per_cpu(rcu_update_flag, cpu)) - per_cpu(rcu_update_flag, cpu)++; - - /* - * Only update if we are coming from a stopped ticks mode - * (rcu_dyntick_sched.dynticks is even). - */ - if (!in_interrupt() && - (rdssp->dynticks & 0x1) == 0) { - /* - * The following might seem like we could have a race - * with NMI/SMIs. But this really isn't a problem. - * Here we do a read/modify/write, and the race happens - * when an NMI/SMI comes in after the read and before - * the write. But NMI/SMIs will increment this counter - * twice before returning, so the zero bit will not - * be corrupted by the NMI/SMI which is the most important - * part. - * - * The only thing is that we would bring back the counter - * to a postion that it was in during the NMI/SMI. - * But the zero bit would be set, so the rest of the - * counter would again be ignored. - * - * On return from the IRQ, the counter may have the zero - * bit be 0 and the counter the same as the return from - * the NMI/SMI. If the state machine was so unlucky to - * see that, it still doesn't matter, since all - * RCU read-side critical sections on this CPU would - * have already completed. - */ - rdssp->dynticks++; - /* - * The following memory barrier ensures that any - * rcu_read_lock() primitives in the irq handler - * are seen by other CPUs to follow the above - * increment to rcu_dyntick_sched.dynticks. This is - * required in order for other CPUs to correctly - * determine when it is safe to advance the RCU - * grace-period state machine. - */ - smp_mb(); /* see above block comment. */ - /* - * Since we can't determine the dynamic tick mode from - * the rcu_dyntick_sched.dynticks after this routine, - * we use a second flag to acknowledge that we came - * from an idle state with ticks stopped. - */ - per_cpu(rcu_update_flag, cpu)++; - /* - * If we take an NMI/SMI now, they will also increment - * the rcu_update_flag, and will not update the - * rcu_dyntick_sched.dynticks on exit. That is for - * this IRQ to do. - */ - } -} - -/** - * rcu_irq_exit - Called from exiting Hard irq context. - * - * If the CPU was idle with dynamic ticks active, update the - * rcu_dyntick_sched.dynticks to put let the RCU handling be - * aware that the CPU is going back to idle with no ticks. - */ -void rcu_irq_exit(void) -{ - int cpu = smp_processor_id(); - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - /* - * rcu_update_flag is set if we interrupted the CPU - * when it was idle with ticks stopped. - * Once this occurs, we keep track of interrupt nesting - * because a NMI/SMI could also come in, and we still - * only want the IRQ that started the increment of the - * rcu_dyntick_sched.dynticks to be the one that modifies - * it on exit. - */ - if (per_cpu(rcu_update_flag, cpu)) { - if (--per_cpu(rcu_update_flag, cpu)) - return; - - /* This must match the interrupt nesting */ - WARN_ON(in_interrupt()); - - /* - * If an NMI/SMI happens now we are still - * protected by the rcu_dyntick_sched.dynticks being odd. - */ - - /* - * The following memory barrier ensures that any - * rcu_read_unlock() primitives in the irq handler - * are seen by other CPUs to preceed the following - * increment to rcu_dyntick_sched.dynticks. This - * is required in order for other CPUs to determine - * when it is safe to advance the RCU grace-period - * state machine. - */ - smp_mb(); /* see above block comment. */ - rdssp->dynticks++; - WARN_ON(rdssp->dynticks & 0x1); - } -} - -void rcu_nmi_enter(void) -{ - rcu_irq_enter(); -} - -void rcu_nmi_exit(void) -{ - rcu_irq_exit(); -} - -static void dyntick_save_progress_counter(int cpu) -{ - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - rdssp->dynticks_snap = rdssp->dynticks; -} - -static inline int -rcu_try_flip_waitack_needed(int cpu) -{ - long curr; - long snap; - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - curr = rdssp->dynticks; - snap = rdssp->dynticks_snap; - smp_mb(); /* force ordering with cpu entering/leaving dynticks. */ - - /* - * If the CPU remained in dynticks mode for the entire time - * and didn't take any interrupts, NMIs, SMIs, or whatever, - * then it cannot be in the middle of an rcu_read_lock(), so - * the next rcu_read_lock() it executes must use the new value - * of the counter. So we can safely pretend that this CPU - * already acknowledged the counter. - */ - - if ((curr == snap) && ((curr & 0x1) == 0)) - return 0; - - /* - * If the CPU passed through or entered a dynticks idle phase with - * no active irq handlers, then, as above, we can safely pretend - * that this CPU already acknowledged the counter. - */ - - if ((curr - snap) > 2 || (curr & 0x1) == 0) - return 0; - - /* We need this CPU to explicitly acknowledge the counter flip. */ - - return 1; -} - -static inline int -rcu_try_flip_waitmb_needed(int cpu) -{ - long curr; - long snap; - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - curr = rdssp->dynticks; - snap = rdssp->dynticks_snap; - smp_mb(); /* force ordering with cpu entering/leaving dynticks. */ - - /* - * If the CPU remained in dynticks mode for the entire time - * and didn't take any interrupts, NMIs, SMIs, or whatever, - * then it cannot have executed an RCU read-side critical section - * during that time, so there is no need for it to execute a - * memory barrier. - */ - - if ((curr == snap) && ((curr & 0x1) == 0)) - return 0; - - /* - * If the CPU either entered or exited an outermost interrupt, - * SMI, NMI, or whatever handler, then we know that it executed - * a memory barrier when doing so. So we don't need another one. - */ - if (curr != snap) - return 0; - - /* We need the CPU to execute a memory barrier. */ - - return 1; -} - -static void dyntick_save_progress_counter_sched(int cpu) -{ - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - rdssp->sched_dynticks_snap = rdssp->dynticks; -} - -static int rcu_qsctr_inc_needed_dyntick(int cpu) -{ - long curr; - long snap; - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - curr = rdssp->dynticks; - snap = rdssp->sched_dynticks_snap; - smp_mb(); /* force ordering with cpu entering/leaving dynticks. */ - - /* - * If the CPU remained in dynticks mode for the entire time - * and didn't take any interrupts, NMIs, SMIs, or whatever, - * then it cannot be in the middle of an rcu_read_lock(), so - * the next rcu_read_lock() it executes must use the new value - * of the counter. Therefore, this CPU has been in a quiescent - * state the entire time, and we don't need to wait for it. - */ - - if ((curr == snap) && ((curr & 0x1) == 0)) - return 0; - - /* - * If the CPU passed through or entered a dynticks idle phase with - * no active irq handlers, then, as above, this CPU has already - * passed through a quiescent state. - */ - - if ((curr - snap) > 2 || (snap & 0x1) == 0) - return 0; - - /* We need this CPU to go through a quiescent state. */ - - return 1; -} - -#else /* !CONFIG_NO_HZ */ - -# define dyntick_save_progress_counter(cpu) do { } while (0) -# define rcu_try_flip_waitack_needed(cpu) (1) -# define rcu_try_flip_waitmb_needed(cpu) (1) - -# define dyntick_save_progress_counter_sched(cpu) do { } while (0) -# define rcu_qsctr_inc_needed_dyntick(cpu) (1) - -#endif /* CONFIG_NO_HZ */ - -static void save_qsctr_sched(int cpu) -{ - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - rdssp->sched_qs_snap = rdssp->sched_qs; -} - -static inline int rcu_qsctr_inc_needed(int cpu) -{ - struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); - - /* - * If there has been a quiescent state, no more need to wait - * on this CPU. - */ - - if (rdssp->sched_qs != rdssp->sched_qs_snap) { - smp_mb(); /* force ordering with cpu entering schedule(). */ - return 0; - } - - /* We need this CPU to go through a quiescent state. */ - - return 1; -} - -/* - * Get here when RCU is idle. Decide whether we need to - * move out of idle state, and return non-zero if so. - * "Straightforward" approach for the moment, might later - * use callback-list lengths, grace-period duration, or - * some such to determine when to exit idle state. - * Might also need a pre-idle test that does not acquire - * the lock, but let's get the simple case working first... - */ - -static int -rcu_try_flip_idle(void) -{ - int cpu; - - RCU_TRACE_ME(rcupreempt_trace_try_flip_i1); - if (!rcu_pending(smp_processor_id())) { - RCU_TRACE_ME(rcupreempt_trace_try_flip_ie1); - return 0; - } - - /* - * Do the flip. - */ - - RCU_TRACE_ME(rcupreempt_trace_try_flip_g1); - rcu_ctrlblk.completed++; /* stands in for rcu_try_flip_g2 */ - - /* - * Need a memory barrier so that other CPUs see the new - * counter value before they see the subsequent change of all - * the rcu_flip_flag instances to rcu_flipped. - */ - - smp_mb(); /* see above block comment. */ - - /* Now ask each CPU for acknowledgement of the flip. */ - - for_each_cpu(cpu, to_cpumask(rcu_cpu_online_map)) { - per_cpu(rcu_flip_flag, cpu) = rcu_flipped; - dyntick_save_progress_counter(cpu); - } - - return 1; -} - -/* - * Wait for CPUs to acknowledge the flip. - */ - -static int -rcu_try_flip_waitack(void) -{ - int cpu; - - RCU_TRACE_ME(rcupreempt_trace_try_flip_a1); - for_each_cpu(cpu, to_cpumask(rcu_cpu_online_map)) - if (rcu_try_flip_waitack_needed(cpu) && - per_cpu(rcu_flip_flag, cpu) != rcu_flip_seen) { - RCU_TRACE_ME(rcupreempt_trace_try_flip_ae1); - return 0; - } - - /* - * Make sure our checks above don't bleed into subsequent - * waiting for the sum of the counters to reach zero. - */ - - smp_mb(); /* see above block comment. */ - RCU_TRACE_ME(rcupreempt_trace_try_flip_a2); - return 1; -} - -/* - * Wait for collective ``last'' counter to reach zero, - * then tell all CPUs to do an end-of-grace-period memory barrier. - */ - -static int -rcu_try_flip_waitzero(void) -{ - int cpu; - int lastidx = !(rcu_ctrlblk.completed & 0x1); - int sum = 0; - - /* Check to see if the sum of the "last" counters is zero. */ - - RCU_TRACE_ME(rcupreempt_trace_try_flip_z1); - for_each_cpu(cpu, to_cpumask(rcu_cpu_online_map)) - sum += RCU_DATA_CPU(cpu)->rcu_flipctr[lastidx]; - if (sum != 0) { - RCU_TRACE_ME(rcupreempt_trace_try_flip_ze1); - return 0; - } - - /* - * This ensures that the other CPUs see the call for - * memory barriers -after- the sum to zero has been - * detected here - */ - smp_mb(); /* ^^^^^^^^^^^^ */ - - /* Call for a memory barrier from each CPU. */ - for_each_cpu(cpu, to_cpumask(rcu_cpu_online_map)) { - per_cpu(rcu_mb_flag, cpu) = rcu_mb_needed; - dyntick_save_progress_counter(cpu); - } - - RCU_TRACE_ME(rcupreempt_trace_try_flip_z2); - return 1; -} - -/* - * Wait for all CPUs to do their end-of-grace-period memory barrier. - * Return 0 once all CPUs have done so. - */ - -static int -rcu_try_flip_waitmb(void) -{ - int cpu; - - RCU_TRACE_ME(rcupreempt_trace_try_flip_m1); - for_each_cpu(cpu, to_cpumask(rcu_cpu_online_map)) - if (rcu_try_flip_waitmb_needed(cpu) && - per_cpu(rcu_mb_flag, cpu) != rcu_mb_done) { - RCU_TRACE_ME(rcupreempt_trace_try_flip_me1); - return 0; - } - - smp_mb(); /* Ensure that the above checks precede any following flip. */ - RCU_TRACE_ME(rcupreempt_trace_try_flip_m2); - return 1; -} - -/* - * Attempt a single flip of the counters. Remember, a single flip does - * -not- constitute a grace period. Instead, the interval between - * at least GP_STAGES consecutive flips is a grace period. - * - * If anyone is nuts enough to run this CONFIG_PREEMPT_RCU implementation - * on a large SMP, they might want to use a hierarchical organization of - * the per-CPU-counter pairs. - */ -static void rcu_try_flip(void) -{ - unsigned long flags; - - RCU_TRACE_ME(rcupreempt_trace_try_flip_1); - if (unlikely(!spin_trylock_irqsave(&rcu_ctrlblk.fliplock, flags))) { - RCU_TRACE_ME(rcupreempt_trace_try_flip_e1); - return; - } - - /* - * Take the next transition(s) through the RCU grace-period - * flip-counter state machine. - */ - - switch (rcu_ctrlblk.rcu_try_flip_state) { - case rcu_try_flip_idle_state: - if (rcu_try_flip_idle()) - rcu_ctrlblk.rcu_try_flip_state = - rcu_try_flip_waitack_state; - break; - case rcu_try_flip_waitack_state: - if (rcu_try_flip_waitack()) - rcu_ctrlblk.rcu_try_flip_state = - rcu_try_flip_waitzero_state; - break; - case rcu_try_flip_waitzero_state: - if (rcu_try_flip_waitzero()) - rcu_ctrlblk.rcu_try_flip_state = - rcu_try_flip_waitmb_state; - break; - case rcu_try_flip_waitmb_state: - if (rcu_try_flip_waitmb()) - rcu_ctrlblk.rcu_try_flip_state = - rcu_try_flip_idle_state; - } - spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, flags); -} - -/* - * Check to see if this CPU needs to do a memory barrier in order to - * ensure that any prior RCU read-side critical sections have committed - * their counter manipulations and critical-section memory references - * before declaring the grace period to be completed. - */ -static void rcu_check_mb(int cpu) -{ - if (per_cpu(rcu_mb_flag, cpu) == rcu_mb_needed) { - smp_mb(); /* Ensure RCU read-side accesses are visible. */ - per_cpu(rcu_mb_flag, cpu) = rcu_mb_done; - } -} - -void rcu_check_callbacks(int cpu, int user) -{ - unsigned long flags; - struct rcu_data *rdp = RCU_DATA_CPU(cpu); - - /* - * If this CPU took its interrupt from user mode or from the - * idle loop, and this is not a nested interrupt, then - * this CPU has to have exited all prior preept-disable - * sections of code. So increment the counter to note this. - * - * The memory barrier is needed to handle the case where - * writes from a preempt-disable section of code get reordered - * into schedule() by this CPU's write buffer. So the memory - * barrier makes sure that the rcu_qsctr_inc() is seen by other - * CPUs to happen after any such write. - */ - - if (user || - (idle_cpu(cpu) && !in_softirq() && - hardirq_count() <= (1 << HARDIRQ_SHIFT))) { - smp_mb(); /* Guard against aggressive schedule(). */ - rcu_qsctr_inc(cpu); - } - - rcu_check_mb(cpu); - if (rcu_ctrlblk.completed == rdp->completed) - rcu_try_flip(); - spin_lock_irqsave(&rdp->lock, flags); - RCU_TRACE_RDP(rcupreempt_trace_check_callbacks, rdp); - __rcu_advance_callbacks(rdp); - if (rdp->donelist == NULL) { - spin_unlock_irqrestore(&rdp->lock, flags); - } else { - spin_unlock_irqrestore(&rdp->lock, flags); - raise_softirq(RCU_SOFTIRQ); - } -} - -/* - * Needed by dynticks, to make sure all RCU processing has finished - * when we go idle: - */ -void rcu_advance_callbacks(int cpu, int user) -{ - unsigned long flags; - struct rcu_data *rdp = RCU_DATA_CPU(cpu); - - if (rcu_ctrlblk.completed == rdp->completed) { - rcu_try_flip(); - if (rcu_ctrlblk.completed == rdp->completed) - return; - } - spin_lock_irqsave(&rdp->lock, flags); - RCU_TRACE_RDP(rcupreempt_trace_check_callbacks, rdp); - __rcu_advance_callbacks(rdp); - spin_unlock_irqrestore(&rdp->lock, flags); -} - -#ifdef CONFIG_HOTPLUG_CPU -#define rcu_offline_cpu_enqueue(srclist, srctail, dstlist, dsttail) do { \ - *dsttail = srclist; \ - if (srclist != NULL) { \ - dsttail = srctail; \ - srclist = NULL; \ - srctail = &srclist;\ - } \ - } while (0) - -void rcu_offline_cpu(int cpu) -{ - int i; - struct rcu_head *list = NULL; - unsigned long flags; - struct rcu_data *rdp = RCU_DATA_CPU(cpu); - struct rcu_head *schedlist = NULL; - struct rcu_head **schedtail = &schedlist; - struct rcu_head **tail = &list; - - /* - * Remove all callbacks from the newly dead CPU, retaining order. - * Otherwise rcu_barrier() will fail - */ - - spin_lock_irqsave(&rdp->lock, flags); - rcu_offline_cpu_enqueue(rdp->donelist, rdp->donetail, list, tail); - for (i = GP_STAGES - 1; i >= 0; i--) - rcu_offline_cpu_enqueue(rdp->waitlist[i], rdp->waittail[i], - list, tail); - rcu_offline_cpu_enqueue(rdp->nextlist, rdp->nexttail, list, tail); - rcu_offline_cpu_enqueue(rdp->waitschedlist, rdp->waitschedtail, - schedlist, schedtail); - rcu_offline_cpu_enqueue(rdp->nextschedlist, rdp->nextschedtail, - schedlist, schedtail); - rdp->rcu_sched_sleeping = 0; - spin_unlock_irqrestore(&rdp->lock, flags); - rdp->waitlistcount = 0; - - /* Disengage the newly dead CPU from the grace-period computation. */ - - spin_lock_irqsave(&rcu_ctrlblk.fliplock, flags); - rcu_check_mb(cpu); - if (per_cpu(rcu_flip_flag, cpu) == rcu_flipped) { - smp_mb(); /* Subsequent counter accesses must see new value */ - per_cpu(rcu_flip_flag, cpu) = rcu_flip_seen; - smp_mb(); /* Subsequent RCU read-side critical sections */ - /* seen -after- acknowledgement. */ - } - - RCU_DATA_ME()->rcu_flipctr[0] += RCU_DATA_CPU(cpu)->rcu_flipctr[0]; - RCU_DATA_ME()->rcu_flipctr[1] += RCU_DATA_CPU(cpu)->rcu_flipctr[1]; - - RCU_DATA_CPU(cpu)->rcu_flipctr[0] = 0; - RCU_DATA_CPU(cpu)->rcu_flipctr[1] = 0; - - cpumask_clear_cpu(cpu, to_cpumask(rcu_cpu_online_map)); - - spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, flags); - - /* - * Place the removed callbacks on the current CPU's queue. - * Make them all start a new grace period: simple approach, - * in theory could starve a given set of callbacks, but - * you would need to be doing some serious CPU hotplugging - * to make this happen. If this becomes a problem, adding - * a synchronize_rcu() to the hotplug path would be a simple - * fix. - */ - - local_irq_save(flags); /* disable preempt till we know what lock. */ - rdp = RCU_DATA_ME(); - spin_lock(&rdp->lock); - *rdp->nexttail = list; - if (list) - rdp->nexttail = tail; - *rdp->nextschedtail = schedlist; - if (schedlist) - rdp->nextschedtail = schedtail; - spin_unlock_irqrestore(&rdp->lock, flags); -} - -#else /* #ifdef CONFIG_HOTPLUG_CPU */ - -void rcu_offline_cpu(int cpu) -{ -} - -#endif /* #else #ifdef CONFIG_HOTPLUG_CPU */ - -void __cpuinit rcu_online_cpu(int cpu) -{ - unsigned long flags; - struct rcu_data *rdp; - - spin_lock_irqsave(&rcu_ctrlblk.fliplock, flags); - cpumask_set_cpu(cpu, to_cpumask(rcu_cpu_online_map)); - spin_unlock_irqrestore(&rcu_ctrlblk.fliplock, flags); - - /* - * The rcu_sched grace-period processing might have bypassed - * this CPU, given that it was not in the rcu_cpu_online_map - * when the grace-period scan started. This means that the - * grace-period task might sleep. So make sure that if this - * should happen, the first callback posted to this CPU will - * wake up the grace-period task if need be. - */ - - rdp = RCU_DATA_CPU(cpu); - spin_lock_irqsave(&rdp->lock, flags); - rdp->rcu_sched_sleeping = 1; - spin_unlock_irqrestore(&rdp->lock, flags); -} - -static void rcu_process_callbacks(struct softirq_action *unused) -{ - unsigned long flags; - struct rcu_head *next, *list; - struct rcu_data *rdp; - - local_irq_save(flags); - rdp = RCU_DATA_ME(); - spin_lock(&rdp->lock); - list = rdp->donelist; - if (list == NULL) { - spin_unlock_irqrestore(&rdp->lock, flags); - return; - } - rdp->donelist = NULL; - rdp->donetail = &rdp->donelist; - RCU_TRACE_RDP(rcupreempt_trace_done_remove, rdp); - spin_unlock_irqrestore(&rdp->lock, flags); - while (list) { - next = list->next; - list->func(list); - list = next; - RCU_TRACE_ME(rcupreempt_trace_invoke); - } -} - -void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) -{ - unsigned long flags; - struct rcu_data *rdp; - - head->func = func; - head->next = NULL; - local_irq_save(flags); - rdp = RCU_DATA_ME(); - spin_lock(&rdp->lock); - __rcu_advance_callbacks(rdp); - *rdp->nexttail = head; - rdp->nexttail = &head->next; - RCU_TRACE_RDP(rcupreempt_trace_next_add, rdp); - spin_unlock_irqrestore(&rdp->lock, flags); -} -EXPORT_SYMBOL_GPL(call_rcu); - -void call_rcu_sched(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) -{ - unsigned long flags; - struct rcu_data *rdp; - int wake_gp = 0; - - head->func = func; - head->next = NULL; - local_irq_save(flags); - rdp = RCU_DATA_ME(); - spin_lock(&rdp->lock); - *rdp->nextschedtail = head; - rdp->nextschedtail = &head->next; - if (rdp->rcu_sched_sleeping) { - - /* Grace-period processing might be sleeping... */ - - rdp->rcu_sched_sleeping = 0; - wake_gp = 1; - } - spin_unlock_irqrestore(&rdp->lock, flags); - if (wake_gp) { - - /* Wake up grace-period processing, unless someone beat us. */ - - spin_lock_irqsave(&rcu_ctrlblk.schedlock, flags); - if (rcu_ctrlblk.sched_sleep != rcu_sched_sleeping) - wake_gp = 0; - rcu_ctrlblk.sched_sleep = rcu_sched_not_sleeping; - spin_unlock_irqrestore(&rcu_ctrlblk.schedlock, flags); - if (wake_gp) - wake_up_interruptible(&rcu_ctrlblk.sched_wq); - } -} -EXPORT_SYMBOL_GPL(call_rcu_sched); - -/* - * Wait until all currently running preempt_disable() code segments - * (including hardware-irq-disable segments) complete. Note that - * in -rt this does -not- necessarily result in all currently executing - * interrupt -handlers- having completed. - */ -void __synchronize_sched(void) -{ - struct rcu_synchronize rcu; - - init_completion(&rcu.completion); - /* Will wake me after RCU finished. */ - call_rcu_sched(&rcu.head, wakeme_after_rcu); - /* Wait for it. */ - wait_for_completion(&rcu.completion); -} -EXPORT_SYMBOL_GPL(__synchronize_sched); - -/* - * kthread function that manages call_rcu_sched grace periods. - */ -static int rcu_sched_grace_period(void *arg) -{ - int couldsleep; /* might sleep after current pass. */ - int couldsleepnext = 0; /* might sleep after next pass. */ - int cpu; - unsigned long flags; - struct rcu_data *rdp; - int ret; - - /* - * Each pass through the following loop handles one - * rcu_sched grace period cycle. - */ - do { - /* Save each CPU's current state. */ - - for_each_online_cpu(cpu) { - dyntick_save_progress_counter_sched(cpu); - save_qsctr_sched(cpu); - } - - /* - * Sleep for about an RCU grace-period's worth to - * allow better batching and to consume less CPU. - */ - schedule_timeout_interruptible(RCU_SCHED_BATCH_TIME); - - /* - * If there was nothing to do last time, prepare to - * sleep at the end of the current grace period cycle. - */ - couldsleep = couldsleepnext; - couldsleepnext = 1; - if (couldsleep) { - spin_lock_irqsave(&rcu_ctrlblk.schedlock, flags); - rcu_ctrlblk.sched_sleep = rcu_sched_sleep_prep; - spin_unlock_irqrestore(&rcu_ctrlblk.schedlock, flags); - } - - /* - * Wait on each CPU in turn to have either visited - * a quiescent state or been in dynticks-idle mode. - */ - for_each_online_cpu(cpu) { - while (rcu_qsctr_inc_needed(cpu) && - rcu_qsctr_inc_needed_dyntick(cpu)) { - /* resched_cpu(cpu); @@@ */ - schedule_timeout_interruptible(1); - } - } - - /* Advance callbacks for each CPU. */ - - for_each_online_cpu(cpu) { - - rdp = RCU_DATA_CPU(cpu); - spin_lock_irqsave(&rdp->lock, flags); - - /* - * We are running on this CPU irq-disabled, so no - * CPU can go offline until we re-enable irqs. - * The current CPU might have already gone - * offline (between the for_each_offline_cpu and - * the spin_lock_irqsave), but in that case all its - * callback lists will be empty, so no harm done. - * - * Advance the callbacks! We share normal RCU's - * donelist, since callbacks are invoked the - * same way in either case. - */ - if (rdp->waitschedlist != NULL) { - *rdp->donetail = rdp->waitschedlist; - rdp->donetail = rdp->waitschedtail; - - /* - * Next rcu_check_callbacks() will - * do the required raise_softirq(). - */ - } - if (rdp->nextschedlist != NULL) { - rdp->waitschedlist = rdp->nextschedlist; - rdp->waitschedtail = rdp->nextschedtail; - couldsleep = 0; - couldsleepnext = 0; - } else { - rdp->waitschedlist = NULL; - rdp->waitschedtail = &rdp->waitschedlist; - } - rdp->nextschedlist = NULL; - rdp->nextschedtail = &rdp->nextschedlist; - - /* Mark sleep intention. */ - - rdp->rcu_sched_sleeping = couldsleep; - - spin_unlock_irqrestore(&rdp->lock, flags); - } - - /* If we saw callbacks on the last scan, go deal with them. */ - - if (!couldsleep) - continue; - - /* Attempt to block... */ - - spin_lock_irqsave(&rcu_ctrlblk.schedlock, flags); - if (rcu_ctrlblk.sched_sleep != rcu_sched_sleep_prep) { - - /* - * Someone posted a callback after we scanned. - * Go take care of it. - */ - spin_unlock_irqrestore(&rcu_ctrlblk.schedlock, flags); - couldsleepnext = 0; - continue; - } - - /* Block until the next person posts a callback. */ - - rcu_ctrlblk.sched_sleep = rcu_sched_sleeping; - spin_unlock_irqrestore(&rcu_ctrlblk.schedlock, flags); - ret = 0; - __wait_event_interruptible(rcu_ctrlblk.sched_wq, - rcu_ctrlblk.sched_sleep != rcu_sched_sleeping, - ret); - - /* - * Signals would prevent us from sleeping, and we cannot - * do much with them in any case. So flush them. - */ - if (ret) - flush_signals(current); - couldsleepnext = 0; - - } while (!kthread_should_stop()); - - return (0); -} - -/* - * Check to see if any future RCU-related work will need to be done - * by the current CPU, even if none need be done immediately, returning - * 1 if so. Assumes that notifiers would take care of handling any - * outstanding requests from the RCU core. - * - * This function is part of the RCU implementation; it is -not- - * an exported member of the RCU API. - */ -int rcu_needs_cpu(int cpu) -{ - struct rcu_data *rdp = RCU_DATA_CPU(cpu); - - return (rdp->donelist != NULL || - !!rdp->waitlistcount || - rdp->nextlist != NULL || - rdp->nextschedlist != NULL || - rdp->waitschedlist != NULL); -} - -int rcu_pending(int cpu) -{ - struct rcu_data *rdp = RCU_DATA_CPU(cpu); - - /* The CPU has at least one callback queued somewhere. */ - - if (rdp->donelist != NULL || - !!rdp->waitlistcount || - rdp->nextlist != NULL || - rdp->nextschedlist != NULL || - rdp->waitschedlist != NULL) - return 1; - - /* The RCU core needs an acknowledgement from this CPU. */ - - if ((per_cpu(rcu_flip_flag, cpu) == rcu_flipped) || - (per_cpu(rcu_mb_flag, cpu) == rcu_mb_needed)) - return 1; - - /* This CPU has fallen behind the global grace-period number. */ - - if (rdp->completed != rcu_ctrlblk.completed) - return 1; - - /* Nothing needed from this CPU. */ - - return 0; -} - -static int __cpuinit rcu_cpu_notify(struct notifier_block *self, - unsigned long action, void *hcpu) -{ - long cpu = (long)hcpu; - - switch (action) { - case CPU_UP_PREPARE: - case CPU_UP_PREPARE_FROZEN: - rcu_online_cpu(cpu); - break; - case CPU_UP_CANCELED: - case CPU_UP_CANCELED_FROZEN: - case CPU_DEAD: - case CPU_DEAD_FROZEN: - rcu_offline_cpu(cpu); - break; - default: - break; - } - return NOTIFY_OK; -} - -static struct notifier_block __cpuinitdata rcu_nb = { - .notifier_call = rcu_cpu_notify, -}; - -void __init __rcu_init(void) -{ - int cpu; - int i; - struct rcu_data *rdp; - - printk(KERN_NOTICE "Preemptible RCU implementation.\n"); - for_each_possible_cpu(cpu) { - rdp = RCU_DATA_CPU(cpu); - spin_lock_init(&rdp->lock); - rdp->completed = 0; - rdp->waitlistcount = 0; - rdp->nextlist = NULL; - rdp->nexttail = &rdp->nextlist; - for (i = 0; i < GP_STAGES; i++) { - rdp->waitlist[i] = NULL; - rdp->waittail[i] = &rdp->waitlist[i]; - } - rdp->donelist = NULL; - rdp->donetail = &rdp->donelist; - rdp->rcu_flipctr[0] = 0; - rdp->rcu_flipctr[1] = 0; - rdp->nextschedlist = NULL; - rdp->nextschedtail = &rdp->nextschedlist; - rdp->waitschedlist = NULL; - rdp->waitschedtail = &rdp->waitschedlist; - rdp->rcu_sched_sleeping = 0; - } - register_cpu_notifier(&rcu_nb); - - /* - * We don't need protection against CPU-Hotplug here - * since - * a) If a CPU comes online while we are iterating over the - * cpu_online_mask below, we would only end up making a - * duplicate call to rcu_online_cpu() which sets the corresponding - * CPU's mask in the rcu_cpu_online_map. - * - * b) A CPU cannot go offline at this point in time since the user - * does not have access to the sysfs interface, nor do we - * suspend the system. - */ - for_each_online_cpu(cpu) - rcu_cpu_notify(&rcu_nb, CPU_UP_PREPARE, (void *)(long) cpu); - - open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); -} - -/* - * Late-boot-time RCU initialization that must wait until after scheduler - * has been initialized. - */ -void __init rcu_init_sched(void) -{ - rcu_sched_grace_period_task = kthread_run(rcu_sched_grace_period, - NULL, - "rcu_sched_grace_period"); - WARN_ON(IS_ERR(rcu_sched_grace_period_task)); -} - -#ifdef CONFIG_RCU_TRACE -long *rcupreempt_flipctr(int cpu) -{ - return &RCU_DATA_CPU(cpu)->rcu_flipctr[0]; -} -EXPORT_SYMBOL_GPL(rcupreempt_flipctr); - -int rcupreempt_flip_flag(int cpu) -{ - return per_cpu(rcu_flip_flag, cpu); -} -EXPORT_SYMBOL_GPL(rcupreempt_flip_flag); - -int rcupreempt_mb_flag(int cpu) -{ - return per_cpu(rcu_mb_flag, cpu); -} -EXPORT_SYMBOL_GPL(rcupreempt_mb_flag); - -char *rcupreempt_try_flip_state_name(void) -{ - return rcu_try_flip_state_names[rcu_ctrlblk.rcu_try_flip_state]; -} -EXPORT_SYMBOL_GPL(rcupreempt_try_flip_state_name); - -struct rcupreempt_trace *rcupreempt_trace_cpu(int cpu) -{ - struct rcu_data *rdp = RCU_DATA_CPU(cpu); - - return &rdp->trace; -} -EXPORT_SYMBOL_GPL(rcupreempt_trace_cpu); - -#endif /* #ifdef RCU_TRACE */ diff --git a/kernel/rcupreempt_trace.c b/kernel/rcupreempt_trace.c deleted file mode 100644 index 7c2665cac17..00000000000 --- a/kernel/rcupreempt_trace.c +++ /dev/null @@ -1,334 +0,0 @@ -/* - * Read-Copy Update tracing for realtime implementation - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. - * - * Copyright IBM Corporation, 2006 - * - * Papers: http://www.rdrop.com/users/paulmck/RCU - * - * For detailed explanation of Read-Copy Update mechanism see - - * Documentation/RCU/ *.txt - * - */ -#include <linux/types.h> -#include <linux/kernel.h> -#include <linux/init.h> -#include <linux/spinlock.h> -#include <linux/smp.h> -#include <linux/rcupdate.h> -#include <linux/interrupt.h> -#include <linux/sched.h> -#include <asm/atomic.h> -#include <linux/bitops.h> -#include <linux/module.h> -#include <linux/completion.h> -#include <linux/moduleparam.h> -#include <linux/percpu.h> -#include <linux/notifier.h> -#include <linux/cpu.h> -#include <linux/mutex.h> -#include <linux/rcupreempt_trace.h> -#include <linux/debugfs.h> - -static struct mutex rcupreempt_trace_mutex; -static char *rcupreempt_trace_buf; -#define RCUPREEMPT_TRACE_BUF_SIZE 4096 - -void rcupreempt_trace_move2done(struct rcupreempt_trace *trace) -{ - trace->done_length += trace->wait_length; - trace->done_add += trace->wait_length; - trace->wait_length = 0; -} -void rcupreempt_trace_move2wait(struct rcupreempt_trace *trace) -{ - trace->wait_length += trace->next_length; - trace->wait_add += trace->next_length; - trace->next_length = 0; -} -void rcupreempt_trace_try_flip_1(struct rcupreempt_trace *trace) -{ - atomic_inc(&trace->rcu_try_flip_1); -} -void rcupreempt_trace_try_flip_e1(struct rcupreempt_trace *trace) -{ - atomic_inc(&trace->rcu_try_flip_e1); -} -void rcupreempt_trace_try_flip_i1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_i1++; -} -void rcupreempt_trace_try_flip_ie1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_ie1++; -} -void rcupreempt_trace_try_flip_g1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_g1++; -} -void rcupreempt_trace_try_flip_a1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_a1++; -} -void rcupreempt_trace_try_flip_ae1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_ae1++; -} -void rcupreempt_trace_try_flip_a2(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_a2++; -} -void rcupreempt_trace_try_flip_z1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_z1++; -} -void rcupreempt_trace_try_flip_ze1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_ze1++; -} -void rcupreempt_trace_try_flip_z2(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_z2++; -} -void rcupreempt_trace_try_flip_m1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_m1++; -} -void rcupreempt_trace_try_flip_me1(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_me1++; -} -void rcupreempt_trace_try_flip_m2(struct rcupreempt_trace *trace) -{ - trace->rcu_try_flip_m2++; -} -void rcupreempt_trace_check_callbacks(struct rcupreempt_trace *trace) -{ - trace->rcu_check_callbacks++; -} -void rcupreempt_trace_done_remove(struct rcupreempt_trace *trace) -{ - trace->done_remove += trace->done_length; - trace->done_length = 0; -} -void rcupreempt_trace_invoke(struct rcupreempt_trace *trace) -{ - atomic_inc(&trace->done_invoked); -} -void rcupreempt_trace_next_add(struct rcupreempt_trace *trace) -{ - trace->next_add++; - trace->next_length++; -} - -static void rcupreempt_trace_sum(struct rcupreempt_trace *sp) -{ - struct rcupreempt_trace *cp; - int cpu; - - memset(sp, 0, sizeof(*sp)); - for_each_possible_cpu(cpu) { - cp = rcupreempt_trace_cpu(cpu); - sp->next_length += cp->next_length; - sp->next_add += cp->next_add; - sp->wait_length += cp->wait_length; - sp->wait_add += cp->wait_add; - sp->done_length += cp->done_length; - sp->done_add += cp->done_add; - sp->done_remove += cp->done_remove; - atomic_add(atomic_read(&cp->done_invoked), &sp->done_invoked); - sp->rcu_check_callbacks += cp->rcu_check_callbacks; - atomic_add(atomic_read(&cp->rcu_try_flip_1), - &sp->rcu_try_flip_1); - atomic_add(atomic_read(&cp->rcu_try_flip_e1), - &sp->rcu_try_flip_e1); - sp->rcu_try_flip_i1 += cp->rcu_try_flip_i1; - sp->rcu_try_flip_ie1 += cp->rcu_try_flip_ie1; - sp->rcu_try_flip_g1 += cp->rcu_try_flip_g1; - sp->rcu_try_flip_a1 += cp->rcu_try_flip_a1; - sp->rcu_try_flip_ae1 += cp->rcu_try_flip_ae1; - sp->rcu_try_flip_a2 += cp->rcu_try_flip_a2; - sp->rcu_try_flip_z1 += cp->rcu_try_flip_z1; - sp->rcu_try_flip_ze1 += cp->rcu_try_flip_ze1; - sp->rcu_try_flip_z2 += cp->rcu_try_flip_z2; - sp->rcu_try_flip_m1 += cp->rcu_try_flip_m1; - sp->rcu_try_flip_me1 += cp->rcu_try_flip_me1; - sp->rcu_try_flip_m2 += cp->rcu_try_flip_m2; - } -} - -static ssize_t rcustats_read(struct file *filp, char __user *buffer, - size_t count, loff_t *ppos) -{ - struct rcupreempt_trace trace; - ssize_t bcount; - int cnt = 0; - - rcupreempt_trace_sum(&trace); - mutex_lock(&rcupreempt_trace_mutex); - snprintf(&rcupreempt_trace_buf[cnt], RCUPREEMPT_TRACE_BUF_SIZE - cnt, - "ggp=%ld rcc=%ld\n", - rcu_batches_completed(), - trace.rcu_check_callbacks); - snprintf(&rcupreempt_trace_buf[cnt], RCUPREEMPT_TRACE_BUF_SIZE - cnt, - "na=%ld nl=%ld wa=%ld wl=%ld da=%ld dl=%ld dr=%ld di=%d\n" - "1=%d e1=%d i1=%ld ie1=%ld g1=%ld a1=%ld ae1=%ld a2=%ld\n" - "z1=%ld ze1=%ld z2=%ld m1=%ld me1=%ld m2=%ld\n", - - trace.next_add, trace.next_length, - trace.wait_add, trace.wait_length, - trace.done_add, trace.done_length, - trace.done_remove, atomic_read(&trace.done_invoked), - atomic_read(&trace.rcu_try_flip_1), - atomic_read(&trace.rcu_try_flip_e1), - trace.rcu_try_flip_i1, trace.rcu_try_flip_ie1, - trace.rcu_try_flip_g1, - trace.rcu_try_flip_a1, trace.rcu_try_flip_ae1, - trace.rcu_try_flip_a2, - trace.rcu_try_flip_z1, trace.rcu_try_flip_ze1, - trace.rcu_try_flip_z2, - trace.rcu_try_flip_m1, trace.rcu_try_flip_me1, - trace.rcu_try_flip_m2); - bcount = simple_read_from_buffer(buffer, count, ppos, - rcupreempt_trace_buf, strlen(rcupreempt_trace_buf)); - mutex_unlock(&rcupreempt_trace_mutex); - return bcount; -} - -static ssize_t rcugp_read(struct file *filp, char __user *buffer, - size_t count, loff_t *ppos) -{ - long oldgp = rcu_batches_completed(); - ssize_t bcount; - - mutex_lock(&rcupreempt_trace_mutex); - synchronize_rcu(); - snprintf(rcupreempt_trace_buf, RCUPREEMPT_TRACE_BUF_SIZE, - "oldggp=%ld newggp=%ld\n", oldgp, rcu_batches_completed()); - bcount = simple_read_from_buffer(buffer, count, ppos, - rcupreempt_trace_buf, strlen(rcupreempt_trace_buf)); - mutex_unlock(&rcupreempt_trace_mutex); - return bcount; -} - -static ssize_t rcuctrs_read(struct file *filp, char __user *buffer, - size_t count, loff_t *ppos) -{ - int cnt = 0; - int cpu; - int f = rcu_batches_completed() & 0x1; - ssize_t bcount; - - mutex_lock(&rcupreempt_trace_mutex); - - cnt += snprintf(&rcupreempt_trace_buf[cnt], RCUPREEMPT_TRACE_BUF_SIZE, - "CPU last cur F M\n"); - for_each_online_cpu(cpu) { - long *flipctr = rcupreempt_flipctr(cpu); - cnt += snprintf(&rcupreempt_trace_buf[cnt], - RCUPREEMPT_TRACE_BUF_SIZE - cnt, - "%3d %4ld %3ld %d %d\n", - cpu, - flipctr[!f], - flipctr[f], - rcupreempt_flip_flag(cpu), - rcupreempt_mb_flag(cpu)); - } - cnt += snprintf(&rcupreempt_trace_buf[cnt], - RCUPREEMPT_TRACE_BUF_SIZE - cnt, - "ggp = %ld, state = %s\n", - rcu_batches_completed(), - rcupreempt_try_flip_state_name()); - cnt += snprintf(&rcupreempt_trace_buf[cnt], - RCUPREEMPT_TRACE_BUF_SIZE - cnt, - "\n"); - bcount = simple_read_from_buffer(buffer, count, ppos, - rcupreempt_trace_buf, strlen(rcupreempt_trace_buf)); - mutex_unlock(&rcupreempt_trace_mutex); - return bcount; -} - -static struct file_operations rcustats_fops = { - .owner = THIS_MODULE, - .read = rcustats_read, -}; - -static struct file_operations rcugp_fops = { - .owner = THIS_MODULE, - .read = rcugp_read, -}; - -static struct file_operations rcuctrs_fops = { - .owner = THIS_MODULE, - .read = rcuctrs_read, -}; - -static struct dentry *rcudir, *statdir, *ctrsdir, *gpdir; -static int rcupreempt_debugfs_init(void) -{ - rcudir = debugfs_create_dir("rcu", NULL); - if (!rcudir) - goto out; - statdir = debugfs_create_file("rcustats", 0444, rcudir, - NULL, &rcustats_fops); - if (!statdir) - goto free_out; - - gpdir = debugfs_create_file("rcugp", 0444, rcudir, NULL, &rcugp_fops); - if (!gpdir) - goto free_out; - - ctrsdir = debugfs_create_file("rcuctrs", 0444, rcudir, - NULL, &rcuctrs_fops); - if (!ctrsdir) - goto free_out; - return 0; -free_out: - if (statdir) - debugfs_remove(statdir); - if (gpdir) - debugfs_remove(gpdir); - debugfs_remove(rcudir); -out: - return 1; -} - -static int __init rcupreempt_trace_init(void) -{ - int ret; - - mutex_init(&rcupreempt_trace_mutex); - rcupreempt_trace_buf = kmalloc(RCUPREEMPT_TRACE_BUF_SIZE, GFP_KERNEL); - if (!rcupreempt_trace_buf) - return 1; - ret = rcupreempt_debugfs_init(); - if (ret) - kfree(rcupreempt_trace_buf); - return ret; -} - -static void __exit rcupreempt_trace_cleanup(void) -{ - debugfs_remove(statdir); - debugfs_remove(gpdir); - debugfs_remove(ctrsdir); - debugfs_remove(rcudir); - kfree(rcupreempt_trace_buf); -} - - -module_init(rcupreempt_trace_init); -module_exit(rcupreempt_trace_cleanup); diff --git a/kernel/rcutorture.c b/kernel/rcutorture.c index 7c4142a79f0..b33db539a8a 100644 --- a/kernel/rcutorture.c +++ b/kernel/rcutorture.c @@ -126,6 +126,7 @@ static atomic_t n_rcu_torture_mberror; static atomic_t n_rcu_torture_error; static long n_rcu_torture_timers = 0; static struct list_head rcu_torture_removed; +static cpumask_var_t shuffle_tmp_mask; static int stutter_pause_test = 0; @@ -256,14 +257,14 @@ struct rcu_torture_ops { void (*init)(void); void (*cleanup)(void); int (*readlock)(void); - void (*readdelay)(struct rcu_random_state *rrsp); + void (*read_delay)(struct rcu_random_state *rrsp); void (*readunlock)(int idx); int (*completed)(void); - void (*deferredfree)(struct rcu_torture *p); + void (*deferred_free)(struct rcu_torture *p); void (*sync)(void); void (*cb_barrier)(void); int (*stats)(char *page); - int irqcapable; + int irq_capable; char *name; }; static struct rcu_torture_ops *cur_ops = NULL; @@ -319,7 +320,7 @@ rcu_torture_cb(struct rcu_head *p) rp->rtort_mbtest = 0; rcu_torture_free(rp); } else - cur_ops->deferredfree(rp); + cur_ops->deferred_free(rp); } static void rcu_torture_deferred_free(struct rcu_torture *p) @@ -328,18 +329,18 @@ static void rcu_torture_deferred_free(struct rcu_torture *p) } static struct rcu_torture_ops rcu_ops = { - .init = NULL, - .cleanup = NULL, - .readlock = rcu_torture_read_lock, - .readdelay = rcu_read_delay, - .readunlock = rcu_torture_read_unlock, - .completed = rcu_torture_completed, - .deferredfree = rcu_torture_deferred_free, - .sync = synchronize_rcu, - .cb_barrier = rcu_barrier, - .stats = NULL, - .irqcapable = 1, - .name = "rcu" + .init = NULL, + .cleanup = NULL, + .readlock = rcu_torture_read_lock, + .read_delay = rcu_read_delay, + .readunlock = rcu_torture_read_unlock, + .completed = rcu_torture_completed, + .deferred_free = rcu_torture_deferred_free, + .sync = synchronize_rcu, + .cb_barrier = rcu_barrier, + .stats = NULL, + .irq_capable = 1, + .name = "rcu" }; static void rcu_sync_torture_deferred_free(struct rcu_torture *p) @@ -369,18 +370,18 @@ static void rcu_sync_torture_init(void) } static struct rcu_torture_ops rcu_sync_ops = { - .init = rcu_sync_torture_init, - .cleanup = NULL, - .readlock = rcu_torture_read_lock, - .readdelay = rcu_read_delay, - .readunlock = rcu_torture_read_unlock, - .completed = rcu_torture_completed, - .deferredfree = rcu_sync_torture_deferred_free, - .sync = synchronize_rcu, - .cb_barrier = NULL, - .stats = NULL, - .irqcapable = 1, - .name = "rcu_sync" + .init = rcu_sync_torture_init, + .cleanup = NULL, + .readlock = rcu_torture_read_lock, + .read_delay = rcu_read_delay, + .readunlock = rcu_torture_read_unlock, + .completed = rcu_torture_completed, + .deferred_free = rcu_sync_torture_deferred_free, + .sync = synchronize_rcu, + .cb_barrier = NULL, + .stats = NULL, + .irq_capable = 1, + .name = "rcu_sync" }; /* @@ -431,33 +432,33 @@ static void rcu_bh_torture_synchronize(void) } static struct rcu_torture_ops rcu_bh_ops = { - .init = NULL, - .cleanup = NULL, - .readlock = rcu_bh_torture_read_lock, - .readdelay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = rcu_bh_torture_read_unlock, - .completed = rcu_bh_torture_completed, - .deferredfree = rcu_bh_torture_deferred_free, - .sync = rcu_bh_torture_synchronize, - .cb_barrier = rcu_barrier_bh, - .stats = NULL, - .irqcapable = 1, - .name = "rcu_bh" + .init = NULL, + .cleanup = NULL, + .readlock = rcu_bh_torture_read_lock, + .read_delay = rcu_read_delay, /* just reuse rcu's version. */ + .readunlock = rcu_bh_torture_read_unlock, + .completed = rcu_bh_torture_completed, + .deferred_free = rcu_bh_torture_deferred_free, + .sync = rcu_bh_torture_synchronize, + .cb_barrier = rcu_barrier_bh, + .stats = NULL, + .irq_capable = 1, + .name = "rcu_bh" }; static struct rcu_torture_ops rcu_bh_sync_ops = { - .init = rcu_sync_torture_init, - .cleanup = NULL, - .readlock = rcu_bh_torture_read_lock, - .readdelay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = rcu_bh_torture_read_unlock, - .completed = rcu_bh_torture_completed, - .deferredfree = rcu_sync_torture_deferred_free, - .sync = rcu_bh_torture_synchronize, - .cb_barrier = NULL, - .stats = NULL, - .irqcapable = 1, - .name = "rcu_bh_sync" + .init = rcu_sync_torture_init, + .cleanup = NULL, + .readlock = rcu_bh_torture_read_lock, + .read_delay = rcu_read_delay, /* just reuse rcu's version. */ + .readunlock = rcu_bh_torture_read_unlock, + .completed = rcu_bh_torture_completed, + .deferred_free = rcu_sync_torture_deferred_free, + .sync = rcu_bh_torture_synchronize, + .cb_barrier = NULL, + .stats = NULL, + .irq_capable = 1, + .name = "rcu_bh_sync" }; /* @@ -529,17 +530,17 @@ static int srcu_torture_stats(char *page) } static struct rcu_torture_ops srcu_ops = { - .init = srcu_torture_init, - .cleanup = srcu_torture_cleanup, - .readlock = srcu_torture_read_lock, - .readdelay = srcu_read_delay, - .readunlock = srcu_torture_read_unlock, - .completed = srcu_torture_completed, - .deferredfree = rcu_sync_torture_deferred_free, - .sync = srcu_torture_synchronize, - .cb_barrier = NULL, - .stats = srcu_torture_stats, - .name = "srcu" + .init = srcu_torture_init, + .cleanup = srcu_torture_cleanup, + .readlock = srcu_torture_read_lock, + .read_delay = srcu_read_delay, + .readunlock = srcu_torture_read_unlock, + .completed = srcu_torture_completed, + .deferred_free = rcu_sync_torture_deferred_free, + .sync = srcu_torture_synchronize, + .cb_barrier = NULL, + .stats = srcu_torture_stats, + .name = "srcu" }; /* @@ -573,32 +574,49 @@ static void sched_torture_synchronize(void) } static struct rcu_torture_ops sched_ops = { - .init = rcu_sync_torture_init, - .cleanup = NULL, - .readlock = sched_torture_read_lock, - .readdelay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = sched_torture_read_unlock, - .completed = sched_torture_completed, - .deferredfree = rcu_sched_torture_deferred_free, - .sync = sched_torture_synchronize, - .cb_barrier = rcu_barrier_sched, - .stats = NULL, - .irqcapable = 1, - .name = "sched" + .init = rcu_sync_torture_init, + .cleanup = NULL, + .readlock = sched_torture_read_lock, + .read_delay = rcu_read_delay, /* just reuse rcu's version. */ + .readunlock = sched_torture_read_unlock, + .completed = sched_torture_completed, + .deferred_free = rcu_sched_torture_deferred_free, + .sync = sched_torture_synchronize, + .cb_barrier = rcu_barrier_sched, + .stats = NULL, + .irq_capable = 1, + .name = "sched" }; static struct rcu_torture_ops sched_ops_sync = { - .init = rcu_sync_torture_init, - .cleanup = NULL, - .readlock = sched_torture_read_lock, - .readdelay = rcu_read_delay, /* just reuse rcu's version. */ - .readunlock = sched_torture_read_unlock, - .completed = sched_torture_completed, - .deferredfree = rcu_sync_torture_deferred_free, - .sync = sched_torture_synchronize, - .cb_barrier = NULL, - .stats = NULL, - .name = "sched_sync" + .init = rcu_sync_torture_init, + .cleanup = NULL, + .readlock = sched_torture_read_lock, + .read_delay = rcu_read_delay, /* just reuse rcu's version. */ + .readunlock = sched_torture_read_unlock, + .completed = sched_torture_completed, + .deferred_free = rcu_sync_torture_deferred_free, + .sync = sched_torture_synchronize, + .cb_barrier = NULL, + .stats = NULL, + .name = "sched_sync" +}; + +extern int rcu_expedited_torture_stats(char *page); + +static struct rcu_torture_ops sched_expedited_ops = { + .init = rcu_sync_torture_init, + .cleanup = NULL, + .readlock = sched_torture_read_lock, + .read_delay = rcu_read_delay, /* just reuse rcu's version. */ + .readunlock = sched_torture_read_unlock, + .completed = sched_torture_completed, + .deferred_free = rcu_sync_torture_deferred_free, + .sync = synchronize_sched_expedited, + .cb_barrier = NULL, + .stats = rcu_expedited_torture_stats, + .irq_capable = 1, + .name = "sched_expedited" }; /* @@ -634,7 +652,7 @@ rcu_torture_writer(void *arg) i = RCU_TORTURE_PIPE_LEN; atomic_inc(&rcu_torture_wcount[i]); old_rp->rtort_pipe_count++; - cur_ops->deferredfree(old_rp); + cur_ops->deferred_free(old_rp); } rcu_torture_current_version++; oldbatch = cur_ops->completed(); @@ -699,7 +717,7 @@ static void rcu_torture_timer(unsigned long unused) if (p->rtort_mbtest == 0) atomic_inc(&n_rcu_torture_mberror); spin_lock(&rand_lock); - cur_ops->readdelay(&rand); + cur_ops->read_delay(&rand); n_rcu_torture_timers++; spin_unlock(&rand_lock); preempt_disable(); @@ -737,11 +755,11 @@ rcu_torture_reader(void *arg) VERBOSE_PRINTK_STRING("rcu_torture_reader task started"); set_user_nice(current, 19); - if (irqreader && cur_ops->irqcapable) + if (irqreader && cur_ops->irq_capable) setup_timer_on_stack(&t, rcu_torture_timer, 0); do { - if (irqreader && cur_ops->irqcapable) { + if (irqreader && cur_ops->irq_capable) { if (!timer_pending(&t)) mod_timer(&t, 1); } @@ -756,7 +774,7 @@ rcu_torture_reader(void *arg) } if (p->rtort_mbtest == 0) atomic_inc(&n_rcu_torture_mberror); - cur_ops->readdelay(&rand); + cur_ops->read_delay(&rand); preempt_disable(); pipe_count = p->rtort_pipe_count; if (pipe_count > RCU_TORTURE_PIPE_LEN) { @@ -777,7 +795,7 @@ rcu_torture_reader(void *arg) } while (!kthread_should_stop() && fullstop == FULLSTOP_DONTSTOP); VERBOSE_PRINTK_STRING("rcu_torture_reader task stopping"); rcutorture_shutdown_absorb("rcu_torture_reader"); - if (irqreader && cur_ops->irqcapable) + if (irqreader && cur_ops->irq_capable) del_timer_sync(&t); while (!kthread_should_stop()) schedule_timeout_uninterruptible(1); @@ -889,10 +907,9 @@ static int rcu_idle_cpu; /* Force all torture tasks off this CPU */ */ static void rcu_torture_shuffle_tasks(void) { - cpumask_t tmp_mask; int i; - cpus_setall(tmp_mask); + cpumask_setall(shuffle_tmp_mask); get_online_cpus(); /* No point in shuffling if there is only one online CPU (ex: UP) */ @@ -902,29 +919,29 @@ static void rcu_torture_shuffle_tasks(void) } if (rcu_idle_cpu != -1) - cpu_clear(rcu_idle_cpu, tmp_mask); + cpumask_clear_cpu(rcu_idle_cpu, shuffle_tmp_mask); - set_cpus_allowed_ptr(current, &tmp_mask); + set_cpus_allowed_ptr(current, shuffle_tmp_mask); if (reader_tasks) { for (i = 0; i < nrealreaders; i++) if (reader_tasks[i]) set_cpus_allowed_ptr(reader_tasks[i], - &tmp_mask); + shuffle_tmp_mask); } if (fakewriter_tasks) { for (i = 0; i < nfakewriters; i++) if (fakewriter_tasks[i]) set_cpus_allowed_ptr(fakewriter_tasks[i], - &tmp_mask); + shuffle_tmp_mask); } if (writer_task) - set_cpus_allowed_ptr(writer_task, &tmp_mask); + set_cpus_allowed_ptr(writer_task, shuffle_tmp_mask); if (stats_task) - set_cpus_allowed_ptr(stats_task, &tmp_mask); + set_cpus_allowed_ptr(stats_task, shuffle_tmp_mask); if (rcu_idle_cpu == -1) rcu_idle_cpu = num_online_cpus() - 1; @@ -1012,6 +1029,7 @@ rcu_torture_cleanup(void) if (shuffler_task) { VERBOSE_PRINTK_STRING("Stopping rcu_torture_shuffle task"); kthread_stop(shuffler_task); + free_cpumask_var(shuffle_tmp_mask); } shuffler_task = NULL; @@ -1077,6 +1095,7 @@ rcu_torture_init(void) int firsterr = 0; static struct rcu_torture_ops *torture_ops[] = { &rcu_ops, &rcu_sync_ops, &rcu_bh_ops, &rcu_bh_sync_ops, + &sched_expedited_ops, &srcu_ops, &sched_ops, &sched_ops_sync, }; mutex_lock(&fullstop_mutex); @@ -1190,10 +1209,18 @@ rcu_torture_init(void) } if (test_no_idle_hz) { rcu_idle_cpu = num_online_cpus() - 1; + + if (!alloc_cpumask_var(&shuffle_tmp_mask, GFP_KERNEL)) { + firsterr = -ENOMEM; + VERBOSE_PRINTK_ERRSTRING("Failed to alloc mask"); + goto unwind; + } + /* Create the shuffler thread */ shuffler_task = kthread_run(rcu_torture_shuffle, NULL, "rcu_torture_shuffle"); if (IS_ERR(shuffler_task)) { + free_cpumask_var(shuffle_tmp_mask); firsterr = PTR_ERR(shuffler_task); VERBOSE_PRINTK_ERRSTRING("Failed to create shuffler"); shuffler_task = NULL; diff --git a/kernel/rcutree.c b/kernel/rcutree.c index f2d8638e6c6..6b11b07cfe7 100644 --- a/kernel/rcutree.c +++ b/kernel/rcutree.c @@ -35,6 +35,7 @@ #include <linux/rcupdate.h> #include <linux/interrupt.h> #include <linux/sched.h> +#include <linux/nmi.h> #include <asm/atomic.h> #include <linux/bitops.h> #include <linux/module.h> @@ -46,6 +47,8 @@ #include <linux/mutex.h> #include <linux/time.h> +#include "rcutree.h" + #ifdef CONFIG_DEBUG_LOCK_ALLOC static struct lock_class_key rcu_lock_key; struct lockdep_map rcu_lock_map = @@ -72,12 +75,61 @@ EXPORT_SYMBOL_GPL(rcu_lock_map); .n_force_qs_ngp = 0, \ } -struct rcu_state rcu_state = RCU_STATE_INITIALIZER(rcu_state); -DEFINE_PER_CPU(struct rcu_data, rcu_data); +struct rcu_state rcu_sched_state = RCU_STATE_INITIALIZER(rcu_sched_state); +DEFINE_PER_CPU(struct rcu_data, rcu_sched_data); struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh_state); DEFINE_PER_CPU(struct rcu_data, rcu_bh_data); +extern long rcu_batches_completed_sched(void); +static struct rcu_node *rcu_get_root(struct rcu_state *rsp); +static void cpu_quiet_msk(unsigned long mask, struct rcu_state *rsp, + struct rcu_node *rnp, unsigned long flags); +static void cpu_quiet_msk_finish(struct rcu_state *rsp, unsigned long flags); +#ifdef CONFIG_HOTPLUG_CPU +static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp); +#endif /* #ifdef CONFIG_HOTPLUG_CPU */ +static void __rcu_process_callbacks(struct rcu_state *rsp, + struct rcu_data *rdp); +static void __call_rcu(struct rcu_head *head, + void (*func)(struct rcu_head *rcu), + struct rcu_state *rsp); +static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp); +static void __cpuinit rcu_init_percpu_data(int cpu, struct rcu_state *rsp, + int preemptable); + +#include "rcutree_plugin.h" + +/* + * Note a quiescent state. Because we do not need to know + * how many quiescent states passed, just if there was at least + * one since the start of the grace period, this just sets a flag. + */ +void rcu_sched_qs(int cpu) +{ + unsigned long flags; + struct rcu_data *rdp; + + local_irq_save(flags); + rdp = &per_cpu(rcu_sched_data, cpu); + rdp->passed_quiesc = 1; + rdp->passed_quiesc_completed = rdp->completed; + rcu_preempt_qs(cpu); + local_irq_restore(flags); +} + +void rcu_bh_qs(int cpu) +{ + unsigned long flags; + struct rcu_data *rdp; + + local_irq_save(flags); + rdp = &per_cpu(rcu_bh_data, cpu); + rdp->passed_quiesc = 1; + rdp->passed_quiesc_completed = rdp->completed; + local_irq_restore(flags); +} + #ifdef CONFIG_NO_HZ DEFINE_PER_CPU(struct rcu_dynticks, rcu_dynticks) = { .dynticks_nesting = 1, @@ -90,15 +142,16 @@ static int qhimark = 10000; /* If this many pending, ignore blimit. */ static int qlowmark = 100; /* Once only this many pending, use blimit. */ static void force_quiescent_state(struct rcu_state *rsp, int relaxed); +static int rcu_pending(int cpu); /* - * Return the number of RCU batches processed thus far for debug & stats. + * Return the number of RCU-sched batches processed thus far for debug & stats. */ -long rcu_batches_completed(void) +long rcu_batches_completed_sched(void) { - return rcu_state.completed; + return rcu_sched_state.completed; } -EXPORT_SYMBOL_GPL(rcu_batches_completed); +EXPORT_SYMBOL_GPL(rcu_batches_completed_sched); /* * Return the number of RCU BH batches processed thus far for debug & stats. @@ -161,6 +214,10 @@ static int rcu_implicit_offline_qs(struct rcu_data *rdp) return 1; } + /* If preemptable RCU, no point in sending reschedule IPI. */ + if (rdp->preemptable) + return 0; + /* The CPU is online, so send it a reschedule IPI. */ if (rdp->cpu != smp_processor_id()) smp_send_reschedule(rdp->cpu); @@ -173,7 +230,6 @@ static int rcu_implicit_offline_qs(struct rcu_data *rdp) #endif /* #ifdef CONFIG_SMP */ #ifdef CONFIG_NO_HZ -static DEFINE_RATELIMIT_STATE(rcu_rs, 10 * HZ, 5); /** * rcu_enter_nohz - inform RCU that current CPU is entering nohz @@ -193,7 +249,7 @@ void rcu_enter_nohz(void) rdtp = &__get_cpu_var(rcu_dynticks); rdtp->dynticks++; rdtp->dynticks_nesting--; - WARN_ON_RATELIMIT(rdtp->dynticks & 0x1, &rcu_rs); + WARN_ON_ONCE(rdtp->dynticks & 0x1); local_irq_restore(flags); } @@ -212,7 +268,7 @@ void rcu_exit_nohz(void) rdtp = &__get_cpu_var(rcu_dynticks); rdtp->dynticks++; rdtp->dynticks_nesting++; - WARN_ON_RATELIMIT(!(rdtp->dynticks & 0x1), &rcu_rs); + WARN_ON_ONCE(!(rdtp->dynticks & 0x1)); local_irq_restore(flags); smp_mb(); /* CPUs seeing ++ must see later RCU read-side crit sects */ } @@ -231,7 +287,7 @@ void rcu_nmi_enter(void) if (rdtp->dynticks & 0x1) return; rdtp->dynticks_nmi++; - WARN_ON_RATELIMIT(!(rdtp->dynticks_nmi & 0x1), &rcu_rs); + WARN_ON_ONCE(!(rdtp->dynticks_nmi & 0x1)); smp_mb(); /* CPUs seeing ++ must see later RCU read-side crit sects */ } @@ -250,7 +306,7 @@ void rcu_nmi_exit(void) return; smp_mb(); /* CPUs seeing ++ must see prior RCU read-side crit sects */ rdtp->dynticks_nmi++; - WARN_ON_RATELIMIT(rdtp->dynticks_nmi & 0x1, &rcu_rs); + WARN_ON_ONCE(rdtp->dynticks_nmi & 0x1); } /** @@ -266,7 +322,7 @@ void rcu_irq_enter(void) if (rdtp->dynticks_nesting++) return; rdtp->dynticks++; - WARN_ON_RATELIMIT(!(rdtp->dynticks & 0x1), &rcu_rs); + WARN_ON_ONCE(!(rdtp->dynticks & 0x1)); smp_mb(); /* CPUs seeing ++ must see later RCU read-side crit sects */ } @@ -285,10 +341,10 @@ void rcu_irq_exit(void) return; smp_mb(); /* CPUs seeing ++ must see prior RCU read-side crit sects */ rdtp->dynticks++; - WARN_ON_RATELIMIT(rdtp->dynticks & 0x1, &rcu_rs); + WARN_ON_ONCE(rdtp->dynticks & 0x1); /* If the interrupt queued a callback, get out of dyntick mode. */ - if (__get_cpu_var(rcu_data).nxtlist || + if (__get_cpu_var(rcu_sched_data).nxtlist || __get_cpu_var(rcu_bh_data).nxtlist) set_need_resched(); } @@ -441,6 +497,7 @@ static void print_other_cpu_stall(struct rcu_state *rsp) printk(KERN_ERR "INFO: RCU detected CPU stalls:"); for (; rnp_cur < rnp_end; rnp_cur++) { + rcu_print_task_stall(rnp); if (rnp_cur->qsmask == 0) continue; for (cpu = 0; cpu <= rnp_cur->grphi - rnp_cur->grplo; cpu++) @@ -449,6 +506,8 @@ static void print_other_cpu_stall(struct rcu_state *rsp) } printk(" (detected by %d, t=%ld jiffies)\n", smp_processor_id(), (long)(jiffies - rsp->gp_start)); + trigger_all_cpu_backtrace(); + force_quiescent_state(rsp, 0); /* Kick them all. */ } @@ -459,12 +518,14 @@ static void print_cpu_stall(struct rcu_state *rsp) printk(KERN_ERR "INFO: RCU detected CPU %d stall (t=%lu jiffies)\n", smp_processor_id(), jiffies - rsp->gp_start); - dump_stack(); + trigger_all_cpu_backtrace(); + spin_lock_irqsave(&rnp->lock, flags); if ((long)(jiffies - rsp->jiffies_stall) >= 0) rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; spin_unlock_irqrestore(&rnp->lock, flags); + set_need_resched(); /* kick ourselves to get things going. */ } @@ -510,8 +571,6 @@ static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) rdp->qs_pending = 1; rdp->passed_quiesc = 0; rdp->gpnum = rsp->gpnum; - rdp->n_rcu_pending_force_qs = rdp->n_rcu_pending + - RCU_JIFFIES_TILL_FORCE_QS; } /* @@ -558,8 +617,6 @@ rcu_start_gp(struct rcu_state *rsp, unsigned long flags) rsp->gpnum++; rsp->signaled = RCU_GP_INIT; /* Hold off force_quiescent_state. */ rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; - rdp->n_rcu_pending_force_qs = rdp->n_rcu_pending + - RCU_JIFFIES_TILL_FORCE_QS; record_gp_stall_check_time(rsp); dyntick_record_completed(rsp, rsp->completed - 1); note_new_gpnum(rsp, rdp); @@ -658,6 +715,19 @@ rcu_process_gp_end(struct rcu_state *rsp, struct rcu_data *rdp) } /* + * Clean up after the prior grace period and let rcu_start_gp() start up + * the next grace period if one is needed. Note that the caller must + * hold rnp->lock, as required by rcu_start_gp(), which will release it. + */ +static void cpu_quiet_msk_finish(struct rcu_state *rsp, unsigned long flags) + __releases(rnp->lock) +{ + rsp->completed = rsp->gpnum; + rcu_process_gp_end(rsp, rsp->rda[smp_processor_id()]); + rcu_start_gp(rsp, flags); /* releases root node's rnp->lock. */ +} + +/* * Similar to cpu_quiet(), for which it is a helper function. Allows * a group of CPUs to be quieted at one go, though all the CPUs in the * group must be represented by the same leaf rcu_node structure. @@ -678,7 +748,7 @@ cpu_quiet_msk(unsigned long mask, struct rcu_state *rsp, struct rcu_node *rnp, return; } rnp->qsmask &= ~mask; - if (rnp->qsmask != 0) { + if (rnp->qsmask != 0 || rcu_preempted_readers(rnp)) { /* Other bits still set at this level, so done. */ spin_unlock_irqrestore(&rnp->lock, flags); @@ -698,14 +768,10 @@ cpu_quiet_msk(unsigned long mask, struct rcu_state *rsp, struct rcu_node *rnp, /* * Get here if we are the last CPU to pass through a quiescent - * state for this grace period. Clean up and let rcu_start_gp() - * start up the next grace period if one is needed. Note that - * we still hold rnp->lock, as required by rcu_start_gp(), which - * will release it. + * state for this grace period. Invoke cpu_quiet_msk_finish() + * to clean up and start the next grace period if one is needed. */ - rsp->completed = rsp->gpnum; - rcu_process_gp_end(rsp, rsp->rda[smp_processor_id()]); - rcu_start_gp(rsp, flags); /* releases rnp->lock. */ + cpu_quiet_msk_finish(rsp, flags); /* releases rnp->lock. */ } /* @@ -812,11 +878,12 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->qsmaskinit &= ~mask; if (rnp->qsmaskinit != 0) { - spin_unlock(&rnp->lock); /* irqs already disabled. */ + spin_unlock(&rnp->lock); /* irqs remain disabled. */ break; } + rcu_preempt_offline_tasks(rsp, rnp); mask = rnp->grpmask; - spin_unlock(&rnp->lock); /* irqs already disabled. */ + spin_unlock(&rnp->lock); /* irqs remain disabled. */ rnp = rnp->parent; } while (rnp != NULL); lastcomp = rsp->completed; @@ -829,7 +896,7 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) /* * Move callbacks from the outgoing CPU to the running CPU. * Note that the outgoing CPU is now quiscent, so it is now - * (uncharacteristically) safe to access it rcu_data structure. + * (uncharacteristically) safe to access its rcu_data structure. * Note also that we must carefully retain the order of the * outgoing CPU's callbacks in order for rcu_barrier() to work * correctly. Finally, note that we start all the callbacks @@ -860,8 +927,9 @@ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) */ static void rcu_offline_cpu(int cpu) { - __rcu_offline_cpu(cpu, &rcu_state); + __rcu_offline_cpu(cpu, &rcu_sched_state); __rcu_offline_cpu(cpu, &rcu_bh_state); + rcu_preempt_offline_cpu(cpu); } #else /* #ifdef CONFIG_HOTPLUG_CPU */ @@ -947,25 +1015,26 @@ static void rcu_do_batch(struct rcu_data *rdp) */ void rcu_check_callbacks(int cpu, int user) { + if (!rcu_pending(cpu)) + return; /* if nothing for RCU to do. */ if (user || - (idle_cpu(cpu) && !in_softirq() && - hardirq_count() <= (1 << HARDIRQ_SHIFT))) { + (idle_cpu(cpu) && rcu_scheduler_active && + !in_softirq() && hardirq_count() <= (1 << HARDIRQ_SHIFT))) { /* * Get here if this CPU took its interrupt from user * mode or from the idle loop, and if this is not a * nested interrupt. In this case, the CPU is in - * a quiescent state, so count it. + * a quiescent state, so note it. * * No memory barrier is required here because both - * rcu_qsctr_inc() and rcu_bh_qsctr_inc() reference - * only CPU-local variables that other CPUs neither - * access nor modify, at least not while the corresponding - * CPU is online. + * rcu_sched_qs() and rcu_bh_qs() reference only CPU-local + * variables that other CPUs neither access nor modify, + * at least not while the corresponding CPU is online. */ - rcu_qsctr_inc(cpu); - rcu_bh_qsctr_inc(cpu); + rcu_sched_qs(cpu); + rcu_bh_qs(cpu); } else if (!in_softirq()) { @@ -973,11 +1042,12 @@ void rcu_check_callbacks(int cpu, int user) * Get here if this CPU did not take its interrupt from * softirq, in other words, if it is not interrupting * a rcu_bh read-side critical section. This is an _bh - * critical section, so count it. + * critical section, so note it. */ - rcu_bh_qsctr_inc(cpu); + rcu_bh_qs(cpu); } + rcu_preempt_check_callbacks(cpu); raise_softirq(RCU_SOFTIRQ); } @@ -1035,7 +1105,6 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) { unsigned long flags; long lastcomp; - struct rcu_data *rdp = rsp->rda[smp_processor_id()]; struct rcu_node *rnp = rcu_get_root(rsp); u8 signaled; @@ -1046,16 +1115,13 @@ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) return; /* Someone else is already on the job. */ } if (relaxed && - (long)(rsp->jiffies_force_qs - jiffies) >= 0 && - (rdp->n_rcu_pending_force_qs - rdp->n_rcu_pending) >= 0) + (long)(rsp->jiffies_force_qs - jiffies) >= 0) goto unlock_ret; /* no emergency and done recently. */ rsp->n_force_qs++; spin_lock(&rnp->lock); lastcomp = rsp->completed; signaled = rsp->signaled; rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; - rdp->n_rcu_pending_force_qs = rdp->n_rcu_pending + - RCU_JIFFIES_TILL_FORCE_QS; if (lastcomp == rsp->gpnum) { rsp->n_force_qs_ngp++; spin_unlock(&rnp->lock); @@ -1120,12 +1186,13 @@ __rcu_process_callbacks(struct rcu_state *rsp, struct rcu_data *rdp) { unsigned long flags; + WARN_ON_ONCE(rdp->beenonline == 0); + /* * If an RCU GP has gone long enough, go check for dyntick * idle CPUs and, if needed, send resched IPIs. */ - if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0 || - (rdp->n_rcu_pending_force_qs - rdp->n_rcu_pending) < 0) + if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0) force_quiescent_state(rsp, 1); /* @@ -1159,8 +1226,10 @@ static void rcu_process_callbacks(struct softirq_action *unused) */ smp_mb(); /* See above block comment. */ - __rcu_process_callbacks(&rcu_state, &__get_cpu_var(rcu_data)); + __rcu_process_callbacks(&rcu_sched_state, + &__get_cpu_var(rcu_sched_data)); __rcu_process_callbacks(&rcu_bh_state, &__get_cpu_var(rcu_bh_data)); + rcu_preempt_process_callbacks(); /* * Memory references from any later RCU read-side critical sections @@ -1210,20 +1279,19 @@ __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), if (unlikely(++rdp->qlen > qhimark)) { rdp->blimit = LONG_MAX; force_quiescent_state(rsp, 0); - } else if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0 || - (rdp->n_rcu_pending_force_qs - rdp->n_rcu_pending) < 0) + } else if ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0) force_quiescent_state(rsp, 1); local_irq_restore(flags); } /* - * Queue an RCU callback for invocation after a grace period. + * Queue an RCU-sched callback for invocation after a grace period. */ -void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) +void call_rcu_sched(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) { - __call_rcu(head, func, &rcu_state); + __call_rcu(head, func, &rcu_sched_state); } -EXPORT_SYMBOL_GPL(call_rcu); +EXPORT_SYMBOL_GPL(call_rcu_sched); /* * Queue an RCU for invocation after a quicker grace period. @@ -1249,32 +1317,44 @@ static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp) check_cpu_stall(rsp, rdp); /* Is the RCU core waiting for a quiescent state from this CPU? */ - if (rdp->qs_pending) + if (rdp->qs_pending) { + rdp->n_rp_qs_pending++; return 1; + } /* Does this CPU have callbacks ready to invoke? */ - if (cpu_has_callbacks_ready_to_invoke(rdp)) + if (cpu_has_callbacks_ready_to_invoke(rdp)) { + rdp->n_rp_cb_ready++; return 1; + } /* Has RCU gone idle with this CPU needing another grace period? */ - if (cpu_needs_another_gp(rsp, rdp)) + if (cpu_needs_another_gp(rsp, rdp)) { + rdp->n_rp_cpu_needs_gp++; return 1; + } /* Has another RCU grace period completed? */ - if (ACCESS_ONCE(rsp->completed) != rdp->completed) /* outside of lock */ + if (ACCESS_ONCE(rsp->completed) != rdp->completed) { /* outside lock */ + rdp->n_rp_gp_completed++; return 1; + } /* Has a new RCU grace period started? */ - if (ACCESS_ONCE(rsp->gpnum) != rdp->gpnum) /* outside of lock */ + if (ACCESS_ONCE(rsp->gpnum) != rdp->gpnum) { /* outside lock */ + rdp->n_rp_gp_started++; return 1; + } /* Has an RCU GP gone long enough to send resched IPIs &c? */ if (ACCESS_ONCE(rsp->completed) != ACCESS_ONCE(rsp->gpnum) && - ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0 || - (rdp->n_rcu_pending_force_qs - rdp->n_rcu_pending) < 0)) + ((long)(ACCESS_ONCE(rsp->jiffies_force_qs) - jiffies) < 0)) { + rdp->n_rp_need_fqs++; return 1; + } /* nothing to do */ + rdp->n_rp_need_nothing++; return 0; } @@ -1283,10 +1363,11 @@ static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp) * by the current CPU, returning 1 if so. This function is part of the * RCU implementation; it is -not- an exported member of the RCU API. */ -int rcu_pending(int cpu) +static int rcu_pending(int cpu) { - return __rcu_pending(&rcu_state, &per_cpu(rcu_data, cpu)) || - __rcu_pending(&rcu_bh_state, &per_cpu(rcu_bh_data, cpu)); + return __rcu_pending(&rcu_sched_state, &per_cpu(rcu_sched_data, cpu)) || + __rcu_pending(&rcu_bh_state, &per_cpu(rcu_bh_data, cpu)) || + rcu_preempt_pending(cpu); } /* @@ -1298,27 +1379,46 @@ int rcu_pending(int cpu) int rcu_needs_cpu(int cpu) { /* RCU callbacks either ready or pending? */ - return per_cpu(rcu_data, cpu).nxtlist || - per_cpu(rcu_bh_data, cpu).nxtlist; + return per_cpu(rcu_sched_data, cpu).nxtlist || + per_cpu(rcu_bh_data, cpu).nxtlist || + rcu_preempt_needs_cpu(cpu); } /* - * Initialize a CPU's per-CPU RCU data. We take this "scorched earth" - * approach so that we don't have to worry about how long the CPU has - * been gone, or whether it ever was online previously. We do trust the - * ->mynode field, as it is constant for a given struct rcu_data and - * initialized during early boot. - * - * Note that only one online or offline event can be happening at a given - * time. Note also that we can accept some slop in the rsp->completed - * access due to the fact that this CPU cannot possibly have any RCU - * callbacks in flight yet. + * Do boot-time initialization of a CPU's per-CPU RCU data. */ -static void -rcu_init_percpu_data(int cpu, struct rcu_state *rsp) +static void __init +rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) { unsigned long flags; int i; + struct rcu_data *rdp = rsp->rda[cpu]; + struct rcu_node *rnp = rcu_get_root(rsp); + + /* Set up local state, ensuring consistent view of global state. */ + spin_lock_irqsave(&rnp->lock, flags); + rdp->grpmask = 1UL << (cpu - rdp->mynode->grplo); + rdp->nxtlist = NULL; + for (i = 0; i < RCU_NEXT_SIZE; i++) + rdp->nxttail[i] = &rdp->nxtlist; + rdp->qlen = 0; +#ifdef CONFIG_NO_HZ + rdp->dynticks = &per_cpu(rcu_dynticks, cpu); +#endif /* #ifdef CONFIG_NO_HZ */ + rdp->cpu = cpu; + spin_unlock_irqrestore(&rnp->lock, flags); +} + +/* + * Initialize a CPU's per-CPU RCU data. Note that only one online or + * offline event can be happening at a given time. Note also that we + * can accept some slop in the rsp->completed access due to the fact + * that this CPU cannot possibly have any RCU callbacks in flight yet. + */ +static void __cpuinit +rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptable) +{ + unsigned long flags; long lastcomp; unsigned long mask; struct rcu_data *rdp = rsp->rda[cpu]; @@ -1332,17 +1432,9 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp) rdp->passed_quiesc = 0; /* We could be racing with new GP, */ rdp->qs_pending = 1; /* so set up to respond to current GP. */ rdp->beenonline = 1; /* We have now been online. */ + rdp->preemptable = preemptable; rdp->passed_quiesc_completed = lastcomp - 1; - rdp->grpmask = 1UL << (cpu - rdp->mynode->grplo); - rdp->nxtlist = NULL; - for (i = 0; i < RCU_NEXT_SIZE; i++) - rdp->nxttail[i] = &rdp->nxtlist; - rdp->qlen = 0; rdp->blimit = blimit; -#ifdef CONFIG_NO_HZ - rdp->dynticks = &per_cpu(rcu_dynticks, cpu); -#endif /* #ifdef CONFIG_NO_HZ */ - rdp->cpu = cpu; spin_unlock(&rnp->lock); /* irqs remain disabled. */ /* @@ -1383,16 +1475,16 @@ rcu_init_percpu_data(int cpu, struct rcu_state *rsp) static void __cpuinit rcu_online_cpu(int cpu) { - rcu_init_percpu_data(cpu, &rcu_state); - rcu_init_percpu_data(cpu, &rcu_bh_state); - open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); + rcu_init_percpu_data(cpu, &rcu_sched_state, 0); + rcu_init_percpu_data(cpu, &rcu_bh_state, 0); + rcu_preempt_init_percpu_data(cpu); } /* - * Handle CPU online/offline notifcation events. + * Handle CPU online/offline notification events. */ -static int __cpuinit rcu_cpu_notify(struct notifier_block *self, - unsigned long action, void *hcpu) +int __cpuinit rcu_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu) { long cpu = (long)hcpu; @@ -1464,6 +1556,7 @@ static void __init rcu_init_one(struct rcu_state *rsp) rnp = rsp->level[i]; for (j = 0; j < rsp->levelcnt[i]; j++, rnp++) { spin_lock_init(&rnp->lock); + rnp->gpnum = 0; rnp->qsmask = 0; rnp->qsmaskinit = 0; rnp->grplo = j * cpustride; @@ -1481,16 +1574,20 @@ static void __init rcu_init_one(struct rcu_state *rsp) j / rsp->levelspread[i - 1]; } rnp->level = i; + INIT_LIST_HEAD(&rnp->blocked_tasks[0]); + INIT_LIST_HEAD(&rnp->blocked_tasks[1]); } } } /* - * Helper macro for __rcu_init(). To be used nowhere else! - * Assigns leaf node pointers into each CPU's rcu_data structure. + * Helper macro for __rcu_init() and __rcu_init_preempt(). To be used + * nowhere else! Assigns leaf node pointers into each CPU's rcu_data + * structure. */ -#define RCU_DATA_PTR_INIT(rsp, rcu_data) \ +#define RCU_INIT_FLAVOR(rsp, rcu_data) \ do { \ + rcu_init_one(rsp); \ rnp = (rsp)->level[NUM_RCU_LVLS - 1]; \ j = 0; \ for_each_possible_cpu(i) { \ @@ -1498,33 +1595,43 @@ do { \ j++; \ per_cpu(rcu_data, i).mynode = &rnp[j]; \ (rsp)->rda[i] = &per_cpu(rcu_data, i); \ + rcu_boot_init_percpu_data(i, rsp); \ } \ } while (0) -static struct notifier_block __cpuinitdata rcu_nb = { - .notifier_call = rcu_cpu_notify, -}; +#ifdef CONFIG_TREE_PREEMPT_RCU + +void __init __rcu_init_preempt(void) +{ + int i; /* All used by RCU_INIT_FLAVOR(). */ + int j; + struct rcu_node *rnp; + + RCU_INIT_FLAVOR(&rcu_preempt_state, rcu_preempt_data); +} + +#else /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + +void __init __rcu_init_preempt(void) +{ +} + +#endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */ void __init __rcu_init(void) { - int i; /* All used by RCU_DATA_PTR_INIT(). */ + int i; /* All used by RCU_INIT_FLAVOR(). */ int j; struct rcu_node *rnp; - printk(KERN_WARNING "Experimental hierarchical RCU implementation.\n"); + rcu_bootup_announce(); #ifdef CONFIG_RCU_CPU_STALL_DETECTOR printk(KERN_INFO "RCU-based detection of stalled CPUs is enabled.\n"); #endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ - rcu_init_one(&rcu_state); - RCU_DATA_PTR_INIT(&rcu_state, rcu_data); - rcu_init_one(&rcu_bh_state); - RCU_DATA_PTR_INIT(&rcu_bh_state, rcu_bh_data); - - for_each_online_cpu(i) - rcu_cpu_notify(&rcu_nb, CPU_UP_PREPARE, (void *)(long)i); - /* Register notifier for non-boot CPUs */ - register_cpu_notifier(&rcu_nb); - printk(KERN_WARNING "Experimental hierarchical RCU init done.\n"); + RCU_INIT_FLAVOR(&rcu_sched_state, rcu_sched_data); + RCU_INIT_FLAVOR(&rcu_bh_state, rcu_bh_data); + __rcu_init_preempt(); + open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); } module_param(blimit, int, 0); diff --git a/kernel/rcutree.h b/kernel/rcutree.h new file mode 100644 index 00000000000..bf8a6f9f134 --- /dev/null +++ b/kernel/rcutree.h @@ -0,0 +1,259 @@ +/* + * Read-Copy Update mechanism for mutual exclusion (tree-based version) + * Internal non-public definitions. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright IBM Corporation, 2008 + * + * Author: Ingo Molnar <mingo@elte.hu> + * Paul E. McKenney <paulmck@linux.vnet.ibm.com> + */ + +#include <linux/cache.h> +#include <linux/spinlock.h> +#include <linux/threads.h> +#include <linux/cpumask.h> +#include <linux/seqlock.h> + +/* + * Define shape of hierarchy based on NR_CPUS and CONFIG_RCU_FANOUT. + * In theory, it should be possible to add more levels straightforwardly. + * In practice, this has not been tested, so there is probably some + * bug somewhere. + */ +#define MAX_RCU_LVLS 3 +#define RCU_FANOUT (CONFIG_RCU_FANOUT) +#define RCU_FANOUT_SQ (RCU_FANOUT * RCU_FANOUT) +#define RCU_FANOUT_CUBE (RCU_FANOUT_SQ * RCU_FANOUT) + +#if NR_CPUS <= RCU_FANOUT +# define NUM_RCU_LVLS 1 +# define NUM_RCU_LVL_0 1 +# define NUM_RCU_LVL_1 (NR_CPUS) +# define NUM_RCU_LVL_2 0 +# define NUM_RCU_LVL_3 0 +#elif NR_CPUS <= RCU_FANOUT_SQ +# define NUM_RCU_LVLS 2 +# define NUM_RCU_LVL_0 1 +# define NUM_RCU_LVL_1 (((NR_CPUS) + RCU_FANOUT - 1) / RCU_FANOUT) +# define NUM_RCU_LVL_2 (NR_CPUS) +# define NUM_RCU_LVL_3 0 +#elif NR_CPUS <= RCU_FANOUT_CUBE +# define NUM_RCU_LVLS 3 +# define NUM_RCU_LVL_0 1 +# define NUM_RCU_LVL_1 (((NR_CPUS) + RCU_FANOUT_SQ - 1) / RCU_FANOUT_SQ) +# define NUM_RCU_LVL_2 (((NR_CPUS) + (RCU_FANOUT) - 1) / (RCU_FANOUT)) +# define NUM_RCU_LVL_3 NR_CPUS +#else +# error "CONFIG_RCU_FANOUT insufficient for NR_CPUS" +#endif /* #if (NR_CPUS) <= RCU_FANOUT */ + +#define RCU_SUM (NUM_RCU_LVL_0 + NUM_RCU_LVL_1 + NUM_RCU_LVL_2 + NUM_RCU_LVL_3) +#define NUM_RCU_NODES (RCU_SUM - NR_CPUS) + +/* + * Dynticks per-CPU state. + */ +struct rcu_dynticks { + int dynticks_nesting; /* Track nesting level, sort of. */ + int dynticks; /* Even value for dynticks-idle, else odd. */ + int dynticks_nmi; /* Even value for either dynticks-idle or */ + /* not in nmi handler, else odd. So this */ + /* remains even for nmi from irq handler. */ +}; + +/* + * Definition for node within the RCU grace-period-detection hierarchy. + */ +struct rcu_node { + spinlock_t lock; + long gpnum; /* Current grace period for this node. */ + /* This will either be equal to or one */ + /* behind the root rcu_node's gpnum. */ + unsigned long qsmask; /* CPUs or groups that need to switch in */ + /* order for current grace period to proceed.*/ + unsigned long qsmaskinit; + /* Per-GP initialization for qsmask. */ + unsigned long grpmask; /* Mask to apply to parent qsmask. */ + int grplo; /* lowest-numbered CPU or group here. */ + int grphi; /* highest-numbered CPU or group here. */ + u8 grpnum; /* CPU/group number for next level up. */ + u8 level; /* root is at level 0. */ + struct rcu_node *parent; + struct list_head blocked_tasks[2]; + /* Tasks blocked in RCU read-side critsect. */ +} ____cacheline_internodealigned_in_smp; + +/* Index values for nxttail array in struct rcu_data. */ +#define RCU_DONE_TAIL 0 /* Also RCU_WAIT head. */ +#define RCU_WAIT_TAIL 1 /* Also RCU_NEXT_READY head. */ +#define RCU_NEXT_READY_TAIL 2 /* Also RCU_NEXT head. */ +#define RCU_NEXT_TAIL 3 +#define RCU_NEXT_SIZE 4 + +/* Per-CPU data for read-copy update. */ +struct rcu_data { + /* 1) quiescent-state and grace-period handling : */ + long completed; /* Track rsp->completed gp number */ + /* in order to detect GP end. */ + long gpnum; /* Highest gp number that this CPU */ + /* is aware of having started. */ + long passed_quiesc_completed; + /* Value of completed at time of qs. */ + bool passed_quiesc; /* User-mode/idle loop etc. */ + bool qs_pending; /* Core waits for quiesc state. */ + bool beenonline; /* CPU online at least once. */ + bool preemptable; /* Preemptable RCU? */ + struct rcu_node *mynode; /* This CPU's leaf of hierarchy */ + unsigned long grpmask; /* Mask to apply to leaf qsmask. */ + + /* 2) batch handling */ + /* + * If nxtlist is not NULL, it is partitioned as follows. + * Any of the partitions might be empty, in which case the + * pointer to that partition will be equal to the pointer for + * the following partition. When the list is empty, all of + * the nxttail elements point to nxtlist, which is NULL. + * + * [*nxttail[RCU_NEXT_READY_TAIL], NULL = *nxttail[RCU_NEXT_TAIL]): + * Entries that might have arrived after current GP ended + * [*nxttail[RCU_WAIT_TAIL], *nxttail[RCU_NEXT_READY_TAIL]): + * Entries known to have arrived before current GP ended + * [*nxttail[RCU_DONE_TAIL], *nxttail[RCU_WAIT_TAIL]): + * Entries that batch # <= ->completed - 1: waiting for current GP + * [nxtlist, *nxttail[RCU_DONE_TAIL]): + * Entries that batch # <= ->completed + * The grace period for these entries has completed, and + * the other grace-period-completed entries may be moved + * here temporarily in rcu_process_callbacks(). + */ + struct rcu_head *nxtlist; + struct rcu_head **nxttail[RCU_NEXT_SIZE]; + long qlen; /* # of queued callbacks */ + long blimit; /* Upper limit on a processed batch */ + +#ifdef CONFIG_NO_HZ + /* 3) dynticks interface. */ + struct rcu_dynticks *dynticks; /* Shared per-CPU dynticks state. */ + int dynticks_snap; /* Per-GP tracking for dynticks. */ + int dynticks_nmi_snap; /* Per-GP tracking for dynticks_nmi. */ +#endif /* #ifdef CONFIG_NO_HZ */ + + /* 4) reasons this CPU needed to be kicked by force_quiescent_state */ +#ifdef CONFIG_NO_HZ + unsigned long dynticks_fqs; /* Kicked due to dynticks idle. */ +#endif /* #ifdef CONFIG_NO_HZ */ + unsigned long offline_fqs; /* Kicked due to being offline. */ + unsigned long resched_ipi; /* Sent a resched IPI. */ + + /* 5) __rcu_pending() statistics. */ + long n_rcu_pending; /* rcu_pending() calls since boot. */ + long n_rp_qs_pending; + long n_rp_cb_ready; + long n_rp_cpu_needs_gp; + long n_rp_gp_completed; + long n_rp_gp_started; + long n_rp_need_fqs; + long n_rp_need_nothing; + + int cpu; +}; + +/* Values for signaled field in struct rcu_state. */ +#define RCU_GP_INIT 0 /* Grace period being initialized. */ +#define RCU_SAVE_DYNTICK 1 /* Need to scan dyntick state. */ +#define RCU_FORCE_QS 2 /* Need to force quiescent state. */ +#ifdef CONFIG_NO_HZ +#define RCU_SIGNAL_INIT RCU_SAVE_DYNTICK +#else /* #ifdef CONFIG_NO_HZ */ +#define RCU_SIGNAL_INIT RCU_FORCE_QS +#endif /* #else #ifdef CONFIG_NO_HZ */ + +#define RCU_JIFFIES_TILL_FORCE_QS 3 /* for rsp->jiffies_force_qs */ +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR +#define RCU_SECONDS_TILL_STALL_CHECK (10 * HZ) /* for rsp->jiffies_stall */ +#define RCU_SECONDS_TILL_STALL_RECHECK (30 * HZ) /* for rsp->jiffies_stall */ +#define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time */ + /* to take at least one */ + /* scheduling clock irq */ + /* before ratting on them. */ + +#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ + +/* + * RCU global state, including node hierarchy. This hierarchy is + * represented in "heap" form in a dense array. The root (first level) + * of the hierarchy is in ->node[0] (referenced by ->level[0]), the second + * level in ->node[1] through ->node[m] (->node[1] referenced by ->level[1]), + * and the third level in ->node[m+1] and following (->node[m+1] referenced + * by ->level[2]). The number of levels is determined by the number of + * CPUs and by CONFIG_RCU_FANOUT. Small systems will have a "hierarchy" + * consisting of a single rcu_node. + */ +struct rcu_state { + struct rcu_node node[NUM_RCU_NODES]; /* Hierarchy. */ + struct rcu_node *level[NUM_RCU_LVLS]; /* Hierarchy levels. */ + u32 levelcnt[MAX_RCU_LVLS + 1]; /* # nodes in each level. */ + u8 levelspread[NUM_RCU_LVLS]; /* kids/node in each level. */ + struct rcu_data *rda[NR_CPUS]; /* array of rdp pointers. */ + + /* The following fields are guarded by the root rcu_node's lock. */ + + u8 signaled ____cacheline_internodealigned_in_smp; + /* Force QS state. */ + long gpnum; /* Current gp number. */ + long completed; /* # of last completed gp. */ + spinlock_t onofflock; /* exclude on/offline and */ + /* starting new GP. */ + spinlock_t fqslock; /* Only one task forcing */ + /* quiescent states. */ + unsigned long jiffies_force_qs; /* Time at which to invoke */ + /* force_quiescent_state(). */ + unsigned long n_force_qs; /* Number of calls to */ + /* force_quiescent_state(). */ + unsigned long n_force_qs_lh; /* ~Number of calls leaving */ + /* due to lock unavailable. */ + unsigned long n_force_qs_ngp; /* Number of calls leaving */ + /* due to no GP active. */ +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR + unsigned long gp_start; /* Time at which GP started, */ + /* but in jiffies. */ + unsigned long jiffies_stall; /* Time at which to check */ + /* for CPU stalls. */ +#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ +#ifdef CONFIG_NO_HZ + long dynticks_completed; /* Value of completed @ snap. */ +#endif /* #ifdef CONFIG_NO_HZ */ +}; + +#ifdef RCU_TREE_NONCORE + +/* + * RCU implementation internal declarations: + */ +extern struct rcu_state rcu_sched_state; +DECLARE_PER_CPU(struct rcu_data, rcu_sched_data); + +extern struct rcu_state rcu_bh_state; +DECLARE_PER_CPU(struct rcu_data, rcu_bh_data); + +#ifdef CONFIG_TREE_PREEMPT_RCU +extern struct rcu_state rcu_preempt_state; +DECLARE_PER_CPU(struct rcu_data, rcu_preempt_data); +#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + +#endif /* #ifdef RCU_TREE_NONCORE */ + diff --git a/kernel/rcutree_plugin.h b/kernel/rcutree_plugin.h new file mode 100644 index 00000000000..47789369ea5 --- /dev/null +++ b/kernel/rcutree_plugin.h @@ -0,0 +1,532 @@ +/* + * Read-Copy Update mechanism for mutual exclusion (tree-based version) + * Internal non-public definitions that provide either classic + * or preemptable semantics. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright Red Hat, 2009 + * Copyright IBM Corporation, 2009 + * + * Author: Ingo Molnar <mingo@elte.hu> + * Paul E. McKenney <paulmck@linux.vnet.ibm.com> + */ + + +#ifdef CONFIG_TREE_PREEMPT_RCU + +struct rcu_state rcu_preempt_state = RCU_STATE_INITIALIZER(rcu_preempt_state); +DEFINE_PER_CPU(struct rcu_data, rcu_preempt_data); + +/* + * Tell them what RCU they are running. + */ +static inline void rcu_bootup_announce(void) +{ + printk(KERN_INFO + "Experimental preemptable hierarchical RCU implementation.\n"); +} + +/* + * Return the number of RCU-preempt batches processed thus far + * for debug and statistics. + */ +long rcu_batches_completed_preempt(void) +{ + return rcu_preempt_state.completed; +} +EXPORT_SYMBOL_GPL(rcu_batches_completed_preempt); + +/* + * Return the number of RCU batches processed thus far for debug & stats. + */ +long rcu_batches_completed(void) +{ + return rcu_batches_completed_preempt(); +} +EXPORT_SYMBOL_GPL(rcu_batches_completed); + +/* + * Record a preemptable-RCU quiescent state for the specified CPU. Note + * that this just means that the task currently running on the CPU is + * not in a quiescent state. There might be any number of tasks blocked + * while in an RCU read-side critical section. + */ +static void rcu_preempt_qs_record(int cpu) +{ + struct rcu_data *rdp = &per_cpu(rcu_preempt_data, cpu); + rdp->passed_quiesc = 1; + rdp->passed_quiesc_completed = rdp->completed; +} + +/* + * We have entered the scheduler or are between softirqs in ksoftirqd. + * If we are in an RCU read-side critical section, we need to reflect + * that in the state of the rcu_node structure corresponding to this CPU. + * Caller must disable hardirqs. + */ +static void rcu_preempt_qs(int cpu) +{ + struct task_struct *t = current; + int phase; + struct rcu_data *rdp; + struct rcu_node *rnp; + + if (t->rcu_read_lock_nesting && + (t->rcu_read_unlock_special & RCU_READ_UNLOCK_BLOCKED) == 0) { + + /* Possibly blocking in an RCU read-side critical section. */ + rdp = rcu_preempt_state.rda[cpu]; + rnp = rdp->mynode; + spin_lock(&rnp->lock); + t->rcu_read_unlock_special |= RCU_READ_UNLOCK_BLOCKED; + t->rcu_blocked_node = rnp; + + /* + * If this CPU has already checked in, then this task + * will hold up the next grace period rather than the + * current grace period. Queue the task accordingly. + * If the task is queued for the current grace period + * (i.e., this CPU has not yet passed through a quiescent + * state for the current grace period), then as long + * as that task remains queued, the current grace period + * cannot end. + */ + phase = !(rnp->qsmask & rdp->grpmask) ^ (rnp->gpnum & 0x1); + list_add(&t->rcu_node_entry, &rnp->blocked_tasks[phase]); + smp_mb(); /* Ensure later ctxt swtch seen after above. */ + spin_unlock(&rnp->lock); + } + + /* + * Either we were not in an RCU read-side critical section to + * begin with, or we have now recorded that critical section + * globally. Either way, we can now note a quiescent state + * for this CPU. Again, if we were in an RCU read-side critical + * section, and if that critical section was blocking the current + * grace period, then the fact that the task has been enqueued + * means that we continue to block the current grace period. + */ + rcu_preempt_qs_record(cpu); + t->rcu_read_unlock_special &= ~(RCU_READ_UNLOCK_NEED_QS | + RCU_READ_UNLOCK_GOT_QS); +} + +/* + * Tree-preemptable RCU implementation for rcu_read_lock(). + * Just increment ->rcu_read_lock_nesting, shared state will be updated + * if we block. + */ +void __rcu_read_lock(void) +{ + ACCESS_ONCE(current->rcu_read_lock_nesting)++; + barrier(); /* needed if we ever invoke rcu_read_lock in rcutree.c */ +} +EXPORT_SYMBOL_GPL(__rcu_read_lock); + +static void rcu_read_unlock_special(struct task_struct *t) +{ + int empty; + unsigned long flags; + unsigned long mask; + struct rcu_node *rnp; + int special; + + /* NMI handlers cannot block and cannot safely manipulate state. */ + if (in_nmi()) + return; + + local_irq_save(flags); + + /* + * If RCU core is waiting for this CPU to exit critical section, + * let it know that we have done so. + */ + special = t->rcu_read_unlock_special; + if (special & RCU_READ_UNLOCK_NEED_QS) { + t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_NEED_QS; + t->rcu_read_unlock_special |= RCU_READ_UNLOCK_GOT_QS; + } + + /* Hardware IRQ handlers cannot block. */ + if (in_irq()) { + local_irq_restore(flags); + return; + } + + /* Clean up if blocked during RCU read-side critical section. */ + if (special & RCU_READ_UNLOCK_BLOCKED) { + t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_BLOCKED; + + /* + * Remove this task from the list it blocked on. The + * task can migrate while we acquire the lock, but at + * most one time. So at most two passes through loop. + */ + for (;;) { + rnp = t->rcu_blocked_node; + spin_lock(&rnp->lock); + if (rnp == t->rcu_blocked_node) + break; + spin_unlock(&rnp->lock); + } + empty = list_empty(&rnp->blocked_tasks[rnp->gpnum & 0x1]); + list_del_init(&t->rcu_node_entry); + t->rcu_blocked_node = NULL; + + /* + * If this was the last task on the current list, and if + * we aren't waiting on any CPUs, report the quiescent state. + * Note that both cpu_quiet_msk_finish() and cpu_quiet_msk() + * drop rnp->lock and restore irq. + */ + if (!empty && rnp->qsmask == 0 && + list_empty(&rnp->blocked_tasks[rnp->gpnum & 0x1])) { + t->rcu_read_unlock_special &= + ~(RCU_READ_UNLOCK_NEED_QS | + RCU_READ_UNLOCK_GOT_QS); + if (rnp->parent == NULL) { + /* Only one rcu_node in the tree. */ + cpu_quiet_msk_finish(&rcu_preempt_state, flags); + return; + } + /* Report up the rest of the hierarchy. */ + mask = rnp->grpmask; + spin_unlock_irqrestore(&rnp->lock, flags); + rnp = rnp->parent; + spin_lock_irqsave(&rnp->lock, flags); + cpu_quiet_msk(mask, &rcu_preempt_state, rnp, flags); + return; + } + spin_unlock(&rnp->lock); + } + local_irq_restore(flags); +} + +/* + * Tree-preemptable RCU implementation for rcu_read_unlock(). + * Decrement ->rcu_read_lock_nesting. If the result is zero (outermost + * rcu_read_unlock()) and ->rcu_read_unlock_special is non-zero, then + * invoke rcu_read_unlock_special() to clean up after a context switch + * in an RCU read-side critical section and other special cases. + */ +void __rcu_read_unlock(void) +{ + struct task_struct *t = current; + + barrier(); /* needed if we ever invoke rcu_read_unlock in rcutree.c */ + if (--ACCESS_ONCE(t->rcu_read_lock_nesting) == 0 && + unlikely(ACCESS_ONCE(t->rcu_read_unlock_special))) + rcu_read_unlock_special(t); +} +EXPORT_SYMBOL_GPL(__rcu_read_unlock); + +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR + +/* + * Scan the current list of tasks blocked within RCU read-side critical + * sections, printing out the tid of each. + */ +static void rcu_print_task_stall(struct rcu_node *rnp) +{ + unsigned long flags; + struct list_head *lp; + int phase = rnp->gpnum & 0x1; + struct task_struct *t; + + if (!list_empty(&rnp->blocked_tasks[phase])) { + spin_lock_irqsave(&rnp->lock, flags); + phase = rnp->gpnum & 0x1; /* re-read under lock. */ + lp = &rnp->blocked_tasks[phase]; + list_for_each_entry(t, lp, rcu_node_entry) + printk(" P%d", t->pid); + spin_unlock_irqrestore(&rnp->lock, flags); + } +} + +#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ + +/* + * Check for preempted RCU readers for the specified rcu_node structure. + * If the caller needs a reliable answer, it must hold the rcu_node's + * >lock. + */ +static int rcu_preempted_readers(struct rcu_node *rnp) +{ + return !list_empty(&rnp->blocked_tasks[rnp->gpnum & 0x1]); +} + +#ifdef CONFIG_HOTPLUG_CPU + +/* + * Handle tasklist migration for case in which all CPUs covered by the + * specified rcu_node have gone offline. Move them up to the root + * rcu_node. The reason for not just moving them to the immediate + * parent is to remove the need for rcu_read_unlock_special() to + * make more than two attempts to acquire the target rcu_node's lock. + * + * The caller must hold rnp->lock with irqs disabled. + */ +static void rcu_preempt_offline_tasks(struct rcu_state *rsp, + struct rcu_node *rnp) +{ + int i; + struct list_head *lp; + struct list_head *lp_root; + struct rcu_node *rnp_root = rcu_get_root(rsp); + struct task_struct *tp; + + if (rnp == rnp_root) { + WARN_ONCE(1, "Last CPU thought to be offlined?"); + return; /* Shouldn't happen: at least one CPU online. */ + } + + /* + * Move tasks up to root rcu_node. Rely on the fact that the + * root rcu_node can be at most one ahead of the rest of the + * rcu_nodes in terms of gp_num value. This fact allows us to + * move the blocked_tasks[] array directly, element by element. + */ + for (i = 0; i < 2; i++) { + lp = &rnp->blocked_tasks[i]; + lp_root = &rnp_root->blocked_tasks[i]; + while (!list_empty(lp)) { + tp = list_entry(lp->next, typeof(*tp), rcu_node_entry); + spin_lock(&rnp_root->lock); /* irqs already disabled */ + list_del(&tp->rcu_node_entry); + tp->rcu_blocked_node = rnp_root; + list_add(&tp->rcu_node_entry, lp_root); + spin_unlock(&rnp_root->lock); /* irqs remain disabled */ + } + } +} + +/* + * Do CPU-offline processing for preemptable RCU. + */ +static void rcu_preempt_offline_cpu(int cpu) +{ + __rcu_offline_cpu(cpu, &rcu_preempt_state); +} + +#endif /* #ifdef CONFIG_HOTPLUG_CPU */ + +/* + * Check for a quiescent state from the current CPU. When a task blocks, + * the task is recorded in the corresponding CPU's rcu_node structure, + * which is checked elsewhere. + * + * Caller must disable hard irqs. + */ +static void rcu_preempt_check_callbacks(int cpu) +{ + struct task_struct *t = current; + + if (t->rcu_read_lock_nesting == 0) { + t->rcu_read_unlock_special &= + ~(RCU_READ_UNLOCK_NEED_QS | RCU_READ_UNLOCK_GOT_QS); + rcu_preempt_qs_record(cpu); + return; + } + if (per_cpu(rcu_preempt_data, cpu).qs_pending) { + if (t->rcu_read_unlock_special & RCU_READ_UNLOCK_GOT_QS) { + rcu_preempt_qs_record(cpu); + t->rcu_read_unlock_special &= ~RCU_READ_UNLOCK_GOT_QS; + } else if (!(t->rcu_read_unlock_special & + RCU_READ_UNLOCK_NEED_QS)) { + t->rcu_read_unlock_special |= RCU_READ_UNLOCK_NEED_QS; + } + } +} + +/* + * Process callbacks for preemptable RCU. + */ +static void rcu_preempt_process_callbacks(void) +{ + __rcu_process_callbacks(&rcu_preempt_state, + &__get_cpu_var(rcu_preempt_data)); +} + +/* + * Queue a preemptable-RCU callback for invocation after a grace period. + */ +void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) +{ + __call_rcu(head, func, &rcu_preempt_state); +} +EXPORT_SYMBOL_GPL(call_rcu); + +/* + * Check to see if there is any immediate preemptable-RCU-related work + * to be done. + */ +static int rcu_preempt_pending(int cpu) +{ + return __rcu_pending(&rcu_preempt_state, + &per_cpu(rcu_preempt_data, cpu)); +} + +/* + * Does preemptable RCU need the CPU to stay out of dynticks mode? + */ +static int rcu_preempt_needs_cpu(int cpu) +{ + return !!per_cpu(rcu_preempt_data, cpu).nxtlist; +} + +/* + * Initialize preemptable RCU's per-CPU data. + */ +static void __cpuinit rcu_preempt_init_percpu_data(int cpu) +{ + rcu_init_percpu_data(cpu, &rcu_preempt_state, 1); +} + +/* + * Check for a task exiting while in a preemptable-RCU read-side + * critical section, clean up if so. No need to issue warnings, + * as debug_check_no_locks_held() already does this if lockdep + * is enabled. + */ +void exit_rcu(void) +{ + struct task_struct *t = current; + + if (t->rcu_read_lock_nesting == 0) + return; + t->rcu_read_lock_nesting = 1; + rcu_read_unlock(); +} + +#else /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + +/* + * Tell them what RCU they are running. + */ +static inline void rcu_bootup_announce(void) +{ + printk(KERN_INFO "Hierarchical RCU implementation.\n"); +} + +/* + * Return the number of RCU batches processed thus far for debug & stats. + */ +long rcu_batches_completed(void) +{ + return rcu_batches_completed_sched(); +} +EXPORT_SYMBOL_GPL(rcu_batches_completed); + +/* + * Because preemptable RCU does not exist, we never have to check for + * CPUs being in quiescent states. + */ +static void rcu_preempt_qs(int cpu) +{ +} + +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR + +/* + * Because preemptable RCU does not exist, we never have to check for + * tasks blocked within RCU read-side critical sections. + */ +static void rcu_print_task_stall(struct rcu_node *rnp) +{ +} + +#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ + +/* + * Because preemptable RCU does not exist, there are never any preempted + * RCU readers. + */ +static int rcu_preempted_readers(struct rcu_node *rnp) +{ + return 0; +} + +#ifdef CONFIG_HOTPLUG_CPU + +/* + * Because preemptable RCU does not exist, it never needs to migrate + * tasks that were blocked within RCU read-side critical sections. + */ +static void rcu_preempt_offline_tasks(struct rcu_state *rsp, + struct rcu_node *rnp) +{ +} + +/* + * Because preemptable RCU does not exist, it never needs CPU-offline + * processing. + */ +static void rcu_preempt_offline_cpu(int cpu) +{ +} + +#endif /* #ifdef CONFIG_HOTPLUG_CPU */ + +/* + * Because preemptable RCU does not exist, it never has any callbacks + * to check. + */ +void rcu_preempt_check_callbacks(int cpu) +{ +} + +/* + * Because preemptable RCU does not exist, it never has any callbacks + * to process. + */ +void rcu_preempt_process_callbacks(void) +{ +} + +/* + * In classic RCU, call_rcu() is just call_rcu_sched(). + */ +void call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) +{ + call_rcu_sched(head, func); +} +EXPORT_SYMBOL_GPL(call_rcu); + +/* + * Because preemptable RCU does not exist, it never has any work to do. + */ +static int rcu_preempt_pending(int cpu) +{ + return 0; +} + +/* + * Because preemptable RCU does not exist, it never needs any CPU. + */ +static int rcu_preempt_needs_cpu(int cpu) +{ + return 0; +} + +/* + * Because preemptable RCU does not exist, there is no per-CPU + * data to initialize. + */ +static void __cpuinit rcu_preempt_init_percpu_data(int cpu) +{ +} + +#endif /* #else #ifdef CONFIG_TREE_PREEMPT_RCU */ diff --git a/kernel/rcutree_trace.c b/kernel/rcutree_trace.c index d6db3e83782..0ea1bff6972 100644 --- a/kernel/rcutree_trace.c +++ b/kernel/rcutree_trace.c @@ -43,18 +43,19 @@ #include <linux/debugfs.h> #include <linux/seq_file.h> +#define RCU_TREE_NONCORE +#include "rcutree.h" + static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp) { if (!rdp->beenonline) return; - seq_printf(m, "%3d%cc=%ld g=%ld pq=%d pqc=%ld qp=%d rpfq=%ld rp=%x", + seq_printf(m, "%3d%cc=%ld g=%ld pq=%d pqc=%ld qp=%d", rdp->cpu, cpu_is_offline(rdp->cpu) ? '!' : ' ', rdp->completed, rdp->gpnum, rdp->passed_quiesc, rdp->passed_quiesc_completed, - rdp->qs_pending, - rdp->n_rcu_pending_force_qs - rdp->n_rcu_pending, - (int)(rdp->n_rcu_pending & 0xffff)); + rdp->qs_pending); #ifdef CONFIG_NO_HZ seq_printf(m, " dt=%d/%d dn=%d df=%lu", rdp->dynticks->dynticks, @@ -76,8 +77,12 @@ static void print_one_rcu_data(struct seq_file *m, struct rcu_data *rdp) static int show_rcudata(struct seq_file *m, void *unused) { - seq_puts(m, "rcu:\n"); - PRINT_RCU_DATA(rcu_data, print_one_rcu_data, m); +#ifdef CONFIG_TREE_PREEMPT_RCU + seq_puts(m, "rcu_preempt:\n"); + PRINT_RCU_DATA(rcu_preempt_data, print_one_rcu_data, m); +#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + seq_puts(m, "rcu_sched:\n"); + PRINT_RCU_DATA(rcu_sched_data, print_one_rcu_data, m); seq_puts(m, "rcu_bh:\n"); PRINT_RCU_DATA(rcu_bh_data, print_one_rcu_data, m); return 0; @@ -100,14 +105,12 @@ static void print_one_rcu_data_csv(struct seq_file *m, struct rcu_data *rdp) { if (!rdp->beenonline) return; - seq_printf(m, "%d,%s,%ld,%ld,%d,%ld,%d,%ld,%ld", + seq_printf(m, "%d,%s,%ld,%ld,%d,%ld,%d", rdp->cpu, - cpu_is_offline(rdp->cpu) ? "\"Y\"" : "\"N\"", + cpu_is_offline(rdp->cpu) ? "\"N\"" : "\"Y\"", rdp->completed, rdp->gpnum, rdp->passed_quiesc, rdp->passed_quiesc_completed, - rdp->qs_pending, - rdp->n_rcu_pending_force_qs - rdp->n_rcu_pending, - rdp->n_rcu_pending); + rdp->qs_pending); #ifdef CONFIG_NO_HZ seq_printf(m, ",%d,%d,%d,%lu", rdp->dynticks->dynticks, @@ -121,13 +124,17 @@ static void print_one_rcu_data_csv(struct seq_file *m, struct rcu_data *rdp) static int show_rcudata_csv(struct seq_file *m, void *unused) { - seq_puts(m, "\"CPU\",\"Online?\",\"c\",\"g\",\"pq\",\"pqc\",\"pq\",\"rpfq\",\"rp\","); + seq_puts(m, "\"CPU\",\"Online?\",\"c\",\"g\",\"pq\",\"pqc\",\"pq\","); #ifdef CONFIG_NO_HZ seq_puts(m, "\"dt\",\"dt nesting\",\"dn\",\"df\","); #endif /* #ifdef CONFIG_NO_HZ */ seq_puts(m, "\"of\",\"ri\",\"ql\",\"b\"\n"); - seq_puts(m, "\"rcu:\"\n"); - PRINT_RCU_DATA(rcu_data, print_one_rcu_data_csv, m); +#ifdef CONFIG_TREE_PREEMPT_RCU + seq_puts(m, "\"rcu_preempt:\"\n"); + PRINT_RCU_DATA(rcu_preempt_data, print_one_rcu_data_csv, m); +#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + seq_puts(m, "\"rcu_sched:\"\n"); + PRINT_RCU_DATA(rcu_sched_data, print_one_rcu_data_csv, m); seq_puts(m, "\"rcu_bh:\"\n"); PRINT_RCU_DATA(rcu_bh_data, print_one_rcu_data_csv, m); return 0; @@ -173,8 +180,12 @@ static void print_one_rcu_state(struct seq_file *m, struct rcu_state *rsp) static int show_rcuhier(struct seq_file *m, void *unused) { - seq_puts(m, "rcu:\n"); - print_one_rcu_state(m, &rcu_state); +#ifdef CONFIG_TREE_PREEMPT_RCU + seq_puts(m, "rcu_preempt:\n"); + print_one_rcu_state(m, &rcu_preempt_state); +#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + seq_puts(m, "rcu_sched:\n"); + print_one_rcu_state(m, &rcu_sched_state); seq_puts(m, "rcu_bh:\n"); print_one_rcu_state(m, &rcu_bh_state); return 0; @@ -195,8 +206,12 @@ static struct file_operations rcuhier_fops = { static int show_rcugp(struct seq_file *m, void *unused) { - seq_printf(m, "rcu: completed=%ld gpnum=%ld\n", - rcu_state.completed, rcu_state.gpnum); +#ifdef CONFIG_TREE_PREEMPT_RCU + seq_printf(m, "rcu_preempt: completed=%ld gpnum=%ld\n", + rcu_preempt_state.completed, rcu_preempt_state.gpnum); +#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + seq_printf(m, "rcu_sched: completed=%ld gpnum=%ld\n", + rcu_sched_state.completed, rcu_sched_state.gpnum); seq_printf(m, "rcu_bh: completed=%ld gpnum=%ld\n", rcu_bh_state.completed, rcu_bh_state.gpnum); return 0; @@ -215,51 +230,102 @@ static struct file_operations rcugp_fops = { .release = single_release, }; -static struct dentry *rcudir, *datadir, *datadir_csv, *hierdir, *gpdir; +static void print_one_rcu_pending(struct seq_file *m, struct rcu_data *rdp) +{ + seq_printf(m, "%3d%cnp=%ld " + "qsp=%ld cbr=%ld cng=%ld gpc=%ld gps=%ld nf=%ld nn=%ld\n", + rdp->cpu, + cpu_is_offline(rdp->cpu) ? '!' : ' ', + rdp->n_rcu_pending, + rdp->n_rp_qs_pending, + rdp->n_rp_cb_ready, + rdp->n_rp_cpu_needs_gp, + rdp->n_rp_gp_completed, + rdp->n_rp_gp_started, + rdp->n_rp_need_fqs, + rdp->n_rp_need_nothing); +} + +static void print_rcu_pendings(struct seq_file *m, struct rcu_state *rsp) +{ + int cpu; + struct rcu_data *rdp; + + for_each_possible_cpu(cpu) { + rdp = rsp->rda[cpu]; + if (rdp->beenonline) + print_one_rcu_pending(m, rdp); + } +} + +static int show_rcu_pending(struct seq_file *m, void *unused) +{ +#ifdef CONFIG_TREE_PREEMPT_RCU + seq_puts(m, "rcu_preempt:\n"); + print_rcu_pendings(m, &rcu_preempt_state); +#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */ + seq_puts(m, "rcu_sched:\n"); + print_rcu_pendings(m, &rcu_sched_state); + seq_puts(m, "rcu_bh:\n"); + print_rcu_pendings(m, &rcu_bh_state); + return 0; +} + +static int rcu_pending_open(struct inode *inode, struct file *file) +{ + return single_open(file, show_rcu_pending, NULL); +} + +static struct file_operations rcu_pending_fops = { + .owner = THIS_MODULE, + .open = rcu_pending_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, +}; + +static struct dentry *rcudir; + static int __init rcuclassic_trace_init(void) { + struct dentry *retval; + rcudir = debugfs_create_dir("rcu", NULL); if (!rcudir) - goto out; + goto free_out; - datadir = debugfs_create_file("rcudata", 0444, rcudir, + retval = debugfs_create_file("rcudata", 0444, rcudir, NULL, &rcudata_fops); - if (!datadir) + if (!retval) goto free_out; - datadir_csv = debugfs_create_file("rcudata.csv", 0444, rcudir, + retval = debugfs_create_file("rcudata.csv", 0444, rcudir, NULL, &rcudata_csv_fops); - if (!datadir_csv) + if (!retval) goto free_out; - gpdir = debugfs_create_file("rcugp", 0444, rcudir, NULL, &rcugp_fops); - if (!gpdir) + retval = debugfs_create_file("rcugp", 0444, rcudir, NULL, &rcugp_fops); + if (!retval) goto free_out; - hierdir = debugfs_create_file("rcuhier", 0444, rcudir, + retval = debugfs_create_file("rcuhier", 0444, rcudir, NULL, &rcuhier_fops); - if (!hierdir) + if (!retval) + goto free_out; + + retval = debugfs_create_file("rcu_pending", 0444, rcudir, + NULL, &rcu_pending_fops); + if (!retval) goto free_out; return 0; free_out: - if (datadir) - debugfs_remove(datadir); - if (datadir_csv) - debugfs_remove(datadir_csv); - if (gpdir) - debugfs_remove(gpdir); - debugfs_remove(rcudir); -out: + debugfs_remove_recursive(rcudir); return 1; } static void __exit rcuclassic_trace_cleanup(void) { - debugfs_remove(datadir); - debugfs_remove(datadir_csv); - debugfs_remove(gpdir); - debugfs_remove(hierdir); - debugfs_remove(rcudir); + debugfs_remove_recursive(rcudir); } diff --git a/kernel/relay.c b/kernel/relay.c index 09ac2008f77..bc188549788 100644 --- a/kernel/relay.c +++ b/kernel/relay.c @@ -663,8 +663,10 @@ int relay_late_setup_files(struct rchan *chan, mutex_lock(&relay_channels_mutex); /* Is chan already set up? */ - if (unlikely(chan->has_base_filename)) + if (unlikely(chan->has_base_filename)) { + mutex_unlock(&relay_channels_mutex); return -EEXIST; + } chan->has_base_filename = 1; chan->parent = parent; curr_cpu = get_cpu(); @@ -675,9 +677,7 @@ int relay_late_setup_files(struct rchan *chan, */ for_each_online_cpu(i) { if (unlikely(!chan->buf[i])) { - printk(KERN_ERR "relay_late_setup_files: CPU %u " - "has no buffer, it must have!\n", i); - BUG(); + WARN_ONCE(1, KERN_ERR "CPU has no buffer!\n"); err = -EINVAL; break; } @@ -748,7 +748,7 @@ size_t relay_switch_subbuf(struct rchan_buf *buf, size_t length) * from the scheduler (trying to re-grab * rq->lock), so defer it. */ - __mod_timer(&buf->timer, jiffies + 1); + mod_timer(&buf->timer, jiffies + 1); } old = buf->data; @@ -795,13 +795,15 @@ void relay_subbufs_consumed(struct rchan *chan, if (!chan) return; - if (cpu >= NR_CPUS || !chan->buf[cpu]) + if (cpu >= NR_CPUS || !chan->buf[cpu] || + subbufs_consumed > chan->n_subbufs) return; buf = chan->buf[cpu]; - buf->subbufs_consumed += subbufs_consumed; - if (buf->subbufs_consumed > buf->subbufs_produced) + if (subbufs_consumed > buf->subbufs_produced - buf->subbufs_consumed) buf->subbufs_consumed = buf->subbufs_produced; + else + buf->subbufs_consumed += subbufs_consumed; } EXPORT_SYMBOL_GPL(relay_subbufs_consumed); diff --git a/kernel/res_counter.c b/kernel/res_counter.c index bf8e7534c80..e1338f07431 100644 --- a/kernel/res_counter.c +++ b/kernel/res_counter.c @@ -18,7 +18,7 @@ void res_counter_init(struct res_counter *counter, struct res_counter *parent) { spin_lock_init(&counter->lock); - counter->limit = (unsigned long long)LLONG_MAX; + counter->limit = RESOURCE_MAX; counter->parent = parent; } @@ -133,6 +133,16 @@ int res_counter_memparse_write_strategy(const char *buf, unsigned long long *res) { char *end; + + /* return RESOURCE_MAX(unlimited) if "-1" is specified */ + if (*buf == '-') { + *res = simple_strtoull(buf + 1, &end, 10); + if (*res != 1 || *end != '\0') + return -EINVAL; + *res = RESOURCE_MAX; + return 0; + } + /* FIXME - make memparse() take const char* args */ *res = memparse((char *)buf, &end); if (*end != '\0') diff --git a/kernel/resource.c b/kernel/resource.c index fd5d7d574bb..78b087221c1 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -533,43 +533,21 @@ static void __init __reserve_region_with_split(struct resource *root, res->end = end; res->flags = IORESOURCE_BUSY; - for (;;) { - conflict = __request_resource(parent, res); - if (!conflict) - break; - if (conflict != parent) { - parent = conflict; - if (!(conflict->flags & IORESOURCE_BUSY)) - continue; - } - - /* Uhhuh, that didn't work out.. */ - kfree(res); - res = NULL; - break; - } - - if (!res) { - /* failed, split and try again */ - - /* conflict covered whole area */ - if (conflict->start <= start && conflict->end >= end) - return; + conflict = __request_resource(parent, res); + if (!conflict) + return; - if (conflict->start > start) - __reserve_region_with_split(root, start, conflict->start-1, name); - if (!(conflict->flags & IORESOURCE_BUSY)) { - resource_size_t common_start, common_end; + /* failed, split and try again */ + kfree(res); - common_start = max(conflict->start, start); - common_end = min(conflict->end, end); - if (common_start < common_end) - __reserve_region_with_split(root, common_start, common_end, name); - } - if (conflict->end < end) - __reserve_region_with_split(root, conflict->end+1, end, name); - } + /* conflict covered whole area */ + if (conflict->start <= start && conflict->end >= end) + return; + if (conflict->start > start) + __reserve_region_with_split(root, start, conflict->start-1, name); + if (conflict->end < end) + __reserve_region_with_split(root, conflict->end+1, end, name); } void __init reserve_region_with_split(struct resource *root, @@ -809,7 +787,7 @@ static int __init reserve_setup(char *str) static struct resource reserve[MAXRESERVE]; for (;;) { - int io_start, io_num; + unsigned int io_start, io_num; int x = reserved; if (get_option (&str, &io_start) != 2) diff --git a/kernel/rtmutex.c b/kernel/rtmutex.c index 69d9cb921ff..29bd4baf9e7 100644 --- a/kernel/rtmutex.c +++ b/kernel/rtmutex.c @@ -300,7 +300,8 @@ static int rt_mutex_adjust_prio_chain(struct task_struct *task, * assigned pending owner [which might not have taken the * lock yet]: */ -static inline int try_to_steal_lock(struct rt_mutex *lock) +static inline int try_to_steal_lock(struct rt_mutex *lock, + struct task_struct *task) { struct task_struct *pendowner = rt_mutex_owner(lock); struct rt_mutex_waiter *next; @@ -309,11 +310,11 @@ static inline int try_to_steal_lock(struct rt_mutex *lock) if (!rt_mutex_owner_pending(lock)) return 0; - if (pendowner == current) + if (pendowner == task) return 1; spin_lock_irqsave(&pendowner->pi_lock, flags); - if (current->prio >= pendowner->prio) { + if (task->prio >= pendowner->prio) { spin_unlock_irqrestore(&pendowner->pi_lock, flags); return 0; } @@ -338,21 +339,21 @@ static inline int try_to_steal_lock(struct rt_mutex *lock) * We are going to steal the lock and a waiter was * enqueued on the pending owners pi_waiters queue. So * we have to enqueue this waiter into - * current->pi_waiters list. This covers the case, - * where current is boosted because it holds another + * task->pi_waiters list. This covers the case, + * where task is boosted because it holds another * lock and gets unboosted because the booster is * interrupted, so we would delay a waiter with higher - * priority as current->normal_prio. + * priority as task->normal_prio. * * Note: in the rare case of a SCHED_OTHER task changing * its priority and thus stealing the lock, next->task - * might be current: + * might be task: */ - if (likely(next->task != current)) { - spin_lock_irqsave(¤t->pi_lock, flags); - plist_add(&next->pi_list_entry, ¤t->pi_waiters); - __rt_mutex_adjust_prio(current); - spin_unlock_irqrestore(¤t->pi_lock, flags); + if (likely(next->task != task)) { + spin_lock_irqsave(&task->pi_lock, flags); + plist_add(&next->pi_list_entry, &task->pi_waiters); + __rt_mutex_adjust_prio(task); + spin_unlock_irqrestore(&task->pi_lock, flags); } return 1; } @@ -389,7 +390,7 @@ static int try_to_take_rt_mutex(struct rt_mutex *lock) */ mark_rt_mutex_waiters(lock); - if (rt_mutex_owner(lock) && !try_to_steal_lock(lock)) + if (rt_mutex_owner(lock) && !try_to_steal_lock(lock, current)) return 0; /* We got the lock. */ @@ -411,6 +412,7 @@ static int try_to_take_rt_mutex(struct rt_mutex *lock) */ static int task_blocks_on_rt_mutex(struct rt_mutex *lock, struct rt_mutex_waiter *waiter, + struct task_struct *task, int detect_deadlock) { struct task_struct *owner = rt_mutex_owner(lock); @@ -418,21 +420,21 @@ static int task_blocks_on_rt_mutex(struct rt_mutex *lock, unsigned long flags; int chain_walk = 0, res; - spin_lock_irqsave(¤t->pi_lock, flags); - __rt_mutex_adjust_prio(current); - waiter->task = current; + spin_lock_irqsave(&task->pi_lock, flags); + __rt_mutex_adjust_prio(task); + waiter->task = task; waiter->lock = lock; - plist_node_init(&waiter->list_entry, current->prio); - plist_node_init(&waiter->pi_list_entry, current->prio); + plist_node_init(&waiter->list_entry, task->prio); + plist_node_init(&waiter->pi_list_entry, task->prio); /* Get the top priority waiter on the lock */ if (rt_mutex_has_waiters(lock)) top_waiter = rt_mutex_top_waiter(lock); plist_add(&waiter->list_entry, &lock->wait_list); - current->pi_blocked_on = waiter; + task->pi_blocked_on = waiter; - spin_unlock_irqrestore(¤t->pi_lock, flags); + spin_unlock_irqrestore(&task->pi_lock, flags); if (waiter == rt_mutex_top_waiter(lock)) { spin_lock_irqsave(&owner->pi_lock, flags); @@ -460,7 +462,7 @@ static int task_blocks_on_rt_mutex(struct rt_mutex *lock, spin_unlock(&lock->wait_lock); res = rt_mutex_adjust_prio_chain(owner, detect_deadlock, lock, waiter, - current); + task); spin_lock(&lock->wait_lock); @@ -605,37 +607,25 @@ void rt_mutex_adjust_pi(struct task_struct *task) rt_mutex_adjust_prio_chain(task, 0, NULL, NULL, task); } -/* - * Slow path lock function: +/** + * __rt_mutex_slowlock() - Perform the wait-wake-try-to-take loop + * @lock: the rt_mutex to take + * @state: the state the task should block in (TASK_INTERRUPTIBLE + * or TASK_UNINTERRUPTIBLE) + * @timeout: the pre-initialized and started timer, or NULL for none + * @waiter: the pre-initialized rt_mutex_waiter + * @detect_deadlock: passed to task_blocks_on_rt_mutex + * + * lock->wait_lock must be held by the caller. */ static int __sched -rt_mutex_slowlock(struct rt_mutex *lock, int state, - struct hrtimer_sleeper *timeout, - int detect_deadlock) +__rt_mutex_slowlock(struct rt_mutex *lock, int state, + struct hrtimer_sleeper *timeout, + struct rt_mutex_waiter *waiter, + int detect_deadlock) { - struct rt_mutex_waiter waiter; int ret = 0; - debug_rt_mutex_init_waiter(&waiter); - waiter.task = NULL; - - spin_lock(&lock->wait_lock); - - /* Try to acquire the lock again: */ - if (try_to_take_rt_mutex(lock)) { - spin_unlock(&lock->wait_lock); - return 0; - } - - set_current_state(state); - - /* Setup the timer, when timeout != NULL */ - if (unlikely(timeout)) { - hrtimer_start_expires(&timeout->timer, HRTIMER_MODE_ABS); - if (!hrtimer_active(&timeout->timer)) - timeout->task = NULL; - } - for (;;) { /* Try to acquire the lock: */ if (try_to_take_rt_mutex(lock)) @@ -656,19 +646,19 @@ rt_mutex_slowlock(struct rt_mutex *lock, int state, } /* - * waiter.task is NULL the first time we come here and + * waiter->task is NULL the first time we come here and * when we have been woken up by the previous owner * but the lock got stolen by a higher prio task. */ - if (!waiter.task) { - ret = task_blocks_on_rt_mutex(lock, &waiter, + if (!waiter->task) { + ret = task_blocks_on_rt_mutex(lock, waiter, current, detect_deadlock); /* * If we got woken up by the owner then start loop * all over without going into schedule to try * to get the lock now: */ - if (unlikely(!waiter.task)) { + if (unlikely(!waiter->task)) { /* * Reset the return value. We might * have returned with -EDEADLK and the @@ -684,15 +674,52 @@ rt_mutex_slowlock(struct rt_mutex *lock, int state, spin_unlock(&lock->wait_lock); - debug_rt_mutex_print_deadlock(&waiter); + debug_rt_mutex_print_deadlock(waiter); - if (waiter.task) + if (waiter->task) schedule_rt_mutex(lock); spin_lock(&lock->wait_lock); set_current_state(state); } + return ret; +} + +/* + * Slow path lock function: + */ +static int __sched +rt_mutex_slowlock(struct rt_mutex *lock, int state, + struct hrtimer_sleeper *timeout, + int detect_deadlock) +{ + struct rt_mutex_waiter waiter; + int ret = 0; + + debug_rt_mutex_init_waiter(&waiter); + waiter.task = NULL; + + spin_lock(&lock->wait_lock); + + /* Try to acquire the lock again: */ + if (try_to_take_rt_mutex(lock)) { + spin_unlock(&lock->wait_lock); + return 0; + } + + set_current_state(state); + + /* Setup the timer, when timeout != NULL */ + if (unlikely(timeout)) { + hrtimer_start_expires(&timeout->timer, HRTIMER_MODE_ABS); + if (!hrtimer_active(&timeout->timer)) + timeout->task = NULL; + } + + ret = __rt_mutex_slowlock(lock, state, timeout, &waiter, + detect_deadlock); + set_current_state(TASK_RUNNING); if (unlikely(waiter.task)) @@ -864,9 +891,9 @@ int __sched rt_mutex_lock_interruptible(struct rt_mutex *lock, EXPORT_SYMBOL_GPL(rt_mutex_lock_interruptible); /** - * rt_mutex_lock_interruptible_ktime - lock a rt_mutex interruptible - * the timeout structure is provided - * by the caller + * rt_mutex_timed_lock - lock a rt_mutex interruptible + * the timeout structure is provided + * by the caller * * @lock: the rt_mutex to be locked * @timeout: timeout structure or NULL (no timeout) @@ -875,7 +902,7 @@ EXPORT_SYMBOL_GPL(rt_mutex_lock_interruptible); * Returns: * 0 on success * -EINTR when interrupted by a signal - * -ETIMEOUT when the timeout expired + * -ETIMEDOUT when the timeout expired * -EDEADLK when the lock would deadlock (when deadlock detection is on) */ int @@ -913,7 +940,7 @@ void __sched rt_mutex_unlock(struct rt_mutex *lock) } EXPORT_SYMBOL_GPL(rt_mutex_unlock); -/*** +/** * rt_mutex_destroy - mark a mutex unusable * @lock: the mutex to be destroyed * @@ -986,6 +1013,57 @@ void rt_mutex_proxy_unlock(struct rt_mutex *lock, } /** + * rt_mutex_start_proxy_lock() - Start lock acquisition for another task + * @lock: the rt_mutex to take + * @waiter: the pre-initialized rt_mutex_waiter + * @task: the task to prepare + * @detect_deadlock: perform deadlock detection (1) or not (0) + * + * Returns: + * 0 - task blocked on lock + * 1 - acquired the lock for task, caller should wake it up + * <0 - error + * + * Special API call for FUTEX_REQUEUE_PI support. + */ +int rt_mutex_start_proxy_lock(struct rt_mutex *lock, + struct rt_mutex_waiter *waiter, + struct task_struct *task, int detect_deadlock) +{ + int ret; + + spin_lock(&lock->wait_lock); + + mark_rt_mutex_waiters(lock); + + if (!rt_mutex_owner(lock) || try_to_steal_lock(lock, task)) { + /* We got the lock for task. */ + debug_rt_mutex_lock(lock); + rt_mutex_set_owner(lock, task, 0); + spin_unlock(&lock->wait_lock); + rt_mutex_deadlock_account_lock(lock, task); + return 1; + } + + ret = task_blocks_on_rt_mutex(lock, waiter, task, detect_deadlock); + + if (ret && !waiter->task) { + /* + * Reset the return value. We might have + * returned with -EDEADLK and the owner + * released the lock while we were walking the + * pi chain. Let the waiter sort it out. + */ + ret = 0; + } + spin_unlock(&lock->wait_lock); + + debug_rt_mutex_print_deadlock(waiter); + + return ret; +} + +/** * rt_mutex_next_owner - return the next owner of the lock * * @lock: the rt lock query @@ -1004,3 +1082,57 @@ struct task_struct *rt_mutex_next_owner(struct rt_mutex *lock) return rt_mutex_top_waiter(lock)->task; } + +/** + * rt_mutex_finish_proxy_lock() - Complete lock acquisition + * @lock: the rt_mutex we were woken on + * @to: the timeout, null if none. hrtimer should already have + * been started. + * @waiter: the pre-initialized rt_mutex_waiter + * @detect_deadlock: perform deadlock detection (1) or not (0) + * + * Complete the lock acquisition started our behalf by another thread. + * + * Returns: + * 0 - success + * <0 - error, one of -EINTR, -ETIMEDOUT, or -EDEADLK + * + * Special API call for PI-futex requeue support + */ +int rt_mutex_finish_proxy_lock(struct rt_mutex *lock, + struct hrtimer_sleeper *to, + struct rt_mutex_waiter *waiter, + int detect_deadlock) +{ + int ret; + + spin_lock(&lock->wait_lock); + + set_current_state(TASK_INTERRUPTIBLE); + + ret = __rt_mutex_slowlock(lock, TASK_INTERRUPTIBLE, to, waiter, + detect_deadlock); + + set_current_state(TASK_RUNNING); + + if (unlikely(waiter->task)) + remove_waiter(lock, waiter); + + /* + * try_to_take_rt_mutex() sets the waiter bit unconditionally. We might + * have to fix that up. + */ + fixup_rt_mutex_waiters(lock); + + spin_unlock(&lock->wait_lock); + + /* + * Readjust priority, when we did not get the lock. We might have been + * the pending owner and boosted. Since we did not take the lock, the + * PI boost has to go. + */ + if (unlikely(ret)) + rt_mutex_adjust_prio(current); + + return ret; +} diff --git a/kernel/rtmutex_common.h b/kernel/rtmutex_common.h index e124bf5800e..97a2f81866a 100644 --- a/kernel/rtmutex_common.h +++ b/kernel/rtmutex_common.h @@ -120,6 +120,14 @@ extern void rt_mutex_init_proxy_locked(struct rt_mutex *lock, struct task_struct *proxy_owner); extern void rt_mutex_proxy_unlock(struct rt_mutex *lock, struct task_struct *proxy_owner); +extern int rt_mutex_start_proxy_lock(struct rt_mutex *lock, + struct rt_mutex_waiter *waiter, + struct task_struct *task, + int detect_deadlock); +extern int rt_mutex_finish_proxy_lock(struct rt_mutex *lock, + struct hrtimer_sleeper *to, + struct rt_mutex_waiter *waiter, + int detect_deadlock); #ifdef CONFIG_DEBUG_RT_MUTEXES # include "rtmutex-debug.h" diff --git a/kernel/sched.c b/kernel/sched.c index 52bbf1c842a..e27a53685ed 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -39,6 +39,7 @@ #include <linux/completion.h> #include <linux/kernel_stat.h> #include <linux/debug_locks.h> +#include <linux/perf_counter.h> #include <linux/security.h> #include <linux/notifier.h> #include <linux/profile.h> @@ -63,22 +64,22 @@ #include <linux/tsacct_kern.h> #include <linux/kprobes.h> #include <linux/delayacct.h> -#include <linux/reciprocal_div.h> #include <linux/unistd.h> #include <linux/pagemap.h> #include <linux/hrtimer.h> #include <linux/tick.h> -#include <linux/bootmem.h> #include <linux/debugfs.h> #include <linux/ctype.h> #include <linux/ftrace.h> -#include <trace/sched.h> #include <asm/tlb.h> #include <asm/irq_regs.h> #include "sched_cpupri.h" +#define CREATE_TRACE_POINTS +#include <trace/events/sched.h> + /* * Convert user-nice values [ -20 ... 0 ... 19 ] * to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ], @@ -118,36 +119,8 @@ */ #define RUNTIME_INF ((u64)~0ULL) -DEFINE_TRACE(sched_wait_task); -DEFINE_TRACE(sched_wakeup); -DEFINE_TRACE(sched_wakeup_new); -DEFINE_TRACE(sched_switch); -DEFINE_TRACE(sched_migrate_task); - -#ifdef CONFIG_SMP - static void double_rq_lock(struct rq *rq1, struct rq *rq2); -/* - * Divide a load by a sched group cpu_power : (load / sg->__cpu_power) - * Since cpu_power is a 'constant', we can use a reciprocal divide. - */ -static inline u32 sg_div_cpu_power(const struct sched_group *sg, u32 load) -{ - return reciprocal_divide(load, sg->reciprocal_cpu_power); -} - -/* - * Each time a sched group cpu_power is changed, - * we must compute its reciprocal value - */ -static inline void sg_inc_cpu_power(struct sched_group *sg, u32 val) -{ - sg->__cpu_power += val; - sg->reciprocal_cpu_power = reciprocal_value(sg->__cpu_power); -} -#endif - static inline int rt_policy(int policy) { if (unlikely(policy == SCHED_FIFO || policy == SCHED_RR)) @@ -223,7 +196,7 @@ static void start_rt_bandwidth(struct rt_bandwidth *rt_b) { ktime_t now; - if (rt_bandwidth_enabled() && rt_b->rt_runtime == RUNTIME_INF) + if (!rt_bandwidth_enabled() || rt_b->rt_runtime == RUNTIME_INF) return; if (hrtimer_active(&rt_b->rt_period_timer)) @@ -231,13 +204,20 @@ static void start_rt_bandwidth(struct rt_bandwidth *rt_b) spin_lock(&rt_b->rt_runtime_lock); for (;;) { + unsigned long delta; + ktime_t soft, hard; + if (hrtimer_active(&rt_b->rt_period_timer)) break; now = hrtimer_cb_get_time(&rt_b->rt_period_timer); hrtimer_forward(&rt_b->rt_period_timer, now, rt_b->rt_period); - hrtimer_start_expires(&rt_b->rt_period_timer, - HRTIMER_MODE_ABS); + + soft = hrtimer_get_softexpires(&rt_b->rt_period_timer); + hard = hrtimer_get_expires(&rt_b->rt_period_timer); + delta = ktime_to_ns(ktime_sub(hard, soft)); + __hrtimer_start_range_ns(&rt_b->rt_period_timer, soft, delta, + HRTIMER_MODE_ABS_PINNED, 0); } spin_unlock(&rt_b->rt_runtime_lock); } @@ -306,8 +286,8 @@ void set_tg_uid(struct user_struct *user) /* * Root task group. - * Every UID task group (including init_task_group aka UID-0) will - * be a child to this group. + * Every UID task group (including init_task_group aka UID-0) will + * be a child to this group. */ struct task_group root_task_group; @@ -315,7 +295,7 @@ struct task_group root_task_group; /* Default task group's sched entity on each cpu */ static DEFINE_PER_CPU(struct sched_entity, init_sched_entity); /* Default task group's cfs_rq on each cpu */ -static DEFINE_PER_CPU(struct cfs_rq, init_cfs_rq) ____cacheline_aligned_in_smp; +static DEFINE_PER_CPU(struct cfs_rq, init_tg_cfs_rq) ____cacheline_aligned_in_smp; #endif /* CONFIG_FAIR_GROUP_SCHED */ #ifdef CONFIG_RT_GROUP_SCHED @@ -331,6 +311,13 @@ static DEFINE_PER_CPU(struct rt_rq, init_rt_rq) ____cacheline_aligned_in_smp; */ static DEFINE_SPINLOCK(task_group_lock); +#ifdef CONFIG_SMP +static int root_task_group_empty(void) +{ + return list_empty(&root_task_group.children); +} +#endif + #ifdef CONFIG_FAIR_GROUP_SCHED #ifdef CONFIG_USER_SCHED # define INIT_TASK_GROUP_LOAD (2*NICE_0_LOAD) @@ -391,6 +378,13 @@ static inline void set_task_rq(struct task_struct *p, unsigned int cpu) #else +#ifdef CONFIG_SMP +static int root_task_group_empty(void) +{ + return 1; +} +#endif + static inline void set_task_rq(struct task_struct *p, unsigned int cpu) { } static inline struct task_group *task_group(struct task_struct *p) { @@ -467,11 +461,18 @@ struct rt_rq { struct rt_prio_array active; unsigned long rt_nr_running; #if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED - int highest_prio; /* highest queued rt task prio */ + struct { + int curr; /* highest queued rt task prio */ +#ifdef CONFIG_SMP + int next; /* next highest */ +#endif + } highest_prio; #endif #ifdef CONFIG_SMP unsigned long rt_nr_migratory; + unsigned long rt_nr_total; int overloaded; + struct plist_head pushable_tasks; #endif int rt_throttled; u64 rt_time; @@ -549,7 +550,6 @@ struct rq { unsigned long nr_running; #define CPU_LOAD_IDX_MAX 5 unsigned long cpu_load[CPU_LOAD_IDX_MAX]; - unsigned char idle_at_tick; #ifdef CONFIG_NO_HZ unsigned long last_tick_seen; unsigned char in_nohz_recently; @@ -558,6 +558,7 @@ struct rq { struct load_weight load; unsigned long nr_load_updates; u64 nr_switches; + u64 nr_migrations_in; struct cfs_rq cfs; struct rt_rq rt; @@ -590,7 +591,9 @@ struct rq { struct root_domain *rd; struct sched_domain *sd; + unsigned char idle_at_tick; /* For active balancing */ + int post_schedule; int active_balance; int push_cpu; /* cpu of this runqueue: */ @@ -601,8 +604,15 @@ struct rq { struct task_struct *migration_thread; struct list_head migration_queue; + + u64 rt_avg; + u64 age_stamp; #endif + /* calc_load related fields */ + unsigned long calc_load_update; + long calc_load_active; + #ifdef CONFIG_SCHED_HRTICK #ifdef CONFIG_SMP int hrtick_csd_pending; @@ -618,9 +628,6 @@ struct rq { /* could above be rq->cfs_rq.exec_clock + rq->rt_rq.rt_runtime ? */ /* sys_sched_yield() stats */ - unsigned int yld_exp_empty; - unsigned int yld_act_empty; - unsigned int yld_both_empty; unsigned int yld_count; /* schedule() stats */ @@ -667,8 +674,9 @@ static inline int cpu_of(struct rq *rq) #define this_rq() (&__get_cpu_var(runqueues)) #define task_rq(p) cpu_rq(task_cpu(p)) #define cpu_curr(cpu) (cpu_rq(cpu)->curr) +#define raw_rq() (&__raw_get_cpu_var(runqueues)) -static inline void update_rq_clock(struct rq *rq) +inline void update_rq_clock(struct rq *rq) { rq->clock = sched_clock_cpu(cpu_of(rq)); } @@ -835,6 +843,14 @@ unsigned int sysctl_sched_shares_ratelimit = 250000; unsigned int sysctl_sched_shares_thresh = 4; /* + * period over which we average the RT time consumption, measured + * in ms. + * + * default: 1s + */ +const_debug unsigned int sysctl_sched_time_avg = MSEC_PER_SEC; + +/* * period over which we measure -rt task cpu usage in us. * default: 1s */ @@ -1093,7 +1109,7 @@ static void hrtick_start(struct rq *rq, u64 delay) if (rq == this_rq()) { hrtimer_restart(timer); } else if (!rq->hrtick_csd_pending) { - __smp_call_function_single(cpu_of(rq), &rq->hrtick_csd); + __smp_call_function_single(cpu_of(rq), &rq->hrtick_csd, 0); rq->hrtick_csd_pending = 1; } } @@ -1129,7 +1145,8 @@ static __init void init_hrtick(void) */ static void hrtick_start(struct rq *rq, u64 delay) { - hrtimer_start(&rq->hrtick_timer, ns_to_ktime(delay), HRTIMER_MODE_REL); + __hrtimer_start_range_ns(&rq->hrtick_timer, ns_to_ktime(delay), 0, + HRTIMER_MODE_REL_PINNED, 0); } static inline void init_hrtick(void) @@ -1183,10 +1200,10 @@ static void resched_task(struct task_struct *p) assert_spin_locked(&task_rq(p)->lock); - if (unlikely(test_tsk_thread_flag(p, TIF_NEED_RESCHED))) + if (test_tsk_need_resched(p)) return; - set_tsk_thread_flag(p, TIF_NEED_RESCHED); + set_tsk_need_resched(p); cpu = task_cpu(p); if (cpu == smp_processor_id()) @@ -1242,7 +1259,7 @@ void wake_up_idle_cpu(int cpu) * lockless. The worst case is that the other CPU runs the * idle task through an additional NOOP schedule() */ - set_tsk_thread_flag(rq->idle, TIF_NEED_RESCHED); + set_tsk_need_resched(rq->idle); /* NEED_RESCHED must be visible before we test polling */ smp_mb(); @@ -1251,12 +1268,37 @@ void wake_up_idle_cpu(int cpu) } #endif /* CONFIG_NO_HZ */ +static u64 sched_avg_period(void) +{ + return (u64)sysctl_sched_time_avg * NSEC_PER_MSEC / 2; +} + +static void sched_avg_update(struct rq *rq) +{ + s64 period = sched_avg_period(); + + while ((s64)(rq->clock - rq->age_stamp) > period) { + rq->age_stamp += period; + rq->rt_avg /= 2; + } +} + +static void sched_rt_avg_update(struct rq *rq, u64 rt_delta) +{ + rq->rt_avg += rt_delta; + sched_avg_update(rq); +} + #else /* !CONFIG_SMP */ static void resched_task(struct task_struct *p) { assert_spin_locked(&task_rq(p)->lock); set_tsk_need_resched(p); } + +static void sched_rt_avg_update(struct rq *rq, u64 rt_delta) +{ +} #endif /* CONFIG_SMP */ #if BITS_PER_LONG == 32 @@ -1393,10 +1435,22 @@ iter_move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, struct rq_iterator *iterator); #endif +/* Time spent by the tasks of the cpu accounting group executing in ... */ +enum cpuacct_stat_index { + CPUACCT_STAT_USER, /* ... user mode */ + CPUACCT_STAT_SYSTEM, /* ... kernel mode */ + + CPUACCT_STAT_NSTATS, +}; + #ifdef CONFIG_CGROUP_CPUACCT static void cpuacct_charge(struct task_struct *tsk, u64 cputime); +static void cpuacct_update_stats(struct task_struct *tsk, + enum cpuacct_stat_index idx, cputime_t val); #else static inline void cpuacct_charge(struct task_struct *tsk, u64 cputime) {} +static inline void cpuacct_update_stats(struct task_struct *tsk, + enum cpuacct_stat_index idx, cputime_t val) {} #endif static inline void inc_cpu_load(struct rq *rq, unsigned long load) @@ -1474,28 +1528,35 @@ static unsigned long cpu_avg_load_per_task(int cpu) #ifdef CONFIG_FAIR_GROUP_SCHED +struct update_shares_data { + unsigned long rq_weight[NR_CPUS]; +}; + +static DEFINE_PER_CPU(struct update_shares_data, update_shares_data); + static void __set_se_shares(struct sched_entity *se, unsigned long shares); /* * Calculate and set the cpu's group shares. */ -static void -update_group_shares_cpu(struct task_group *tg, int cpu, - unsigned long sd_shares, unsigned long sd_rq_weight) +static void update_group_shares_cpu(struct task_group *tg, int cpu, + unsigned long sd_shares, + unsigned long sd_rq_weight, + struct update_shares_data *usd) { - unsigned long shares; - unsigned long rq_weight; - - if (!tg->se[cpu]) - return; + unsigned long shares, rq_weight; + int boost = 0; - rq_weight = tg->cfs_rq[cpu]->rq_weight; + rq_weight = usd->rq_weight[cpu]; + if (!rq_weight) { + boost = 1; + rq_weight = NICE_0_LOAD; + } /* - * \Sum shares * rq_weight - * shares = ----------------------- - * \Sum rq_weight - * + * \Sum_j shares_j * rq_weight_i + * shares_i = ----------------------------- + * \Sum_j rq_weight_j */ shares = (sd_shares * rq_weight) / sd_rq_weight; shares = clamp_t(unsigned long, shares, MIN_SHARES, MAX_SHARES); @@ -1506,8 +1567,8 @@ update_group_shares_cpu(struct task_group *tg, int cpu, unsigned long flags; spin_lock_irqsave(&rq->lock, flags); - tg->cfs_rq[cpu]->shares = shares; - + tg->cfs_rq[cpu]->rq_weight = boost ? 0 : rq_weight; + tg->cfs_rq[cpu]->shares = boost ? 0 : shares; __set_se_shares(tg->se[cpu], shares); spin_unlock_irqrestore(&rq->lock, flags); } @@ -1520,22 +1581,30 @@ update_group_shares_cpu(struct task_group *tg, int cpu, */ static int tg_shares_up(struct task_group *tg, void *data) { - unsigned long weight, rq_weight = 0; - unsigned long shares = 0; + unsigned long weight, rq_weight = 0, shares = 0; + struct update_shares_data *usd; struct sched_domain *sd = data; + unsigned long flags; int i; + if (!tg->se[0]) + return 0; + + local_irq_save(flags); + usd = &__get_cpu_var(update_shares_data); + for_each_cpu(i, sched_domain_span(sd)) { + weight = tg->cfs_rq[i]->load.weight; + usd->rq_weight[i] = weight; + /* * If there are currently no tasks on the cpu pretend there * is one of average load so that when a new task gets to * run here it will not get delayed by group starvation. */ - weight = tg->cfs_rq[i]->load.weight; if (!weight) weight = NICE_0_LOAD; - tg->cfs_rq[i]->rq_weight = weight; rq_weight += weight; shares += tg->cfs_rq[i]->shares; } @@ -1547,7 +1616,9 @@ static int tg_shares_up(struct task_group *tg, void *data) shares = tg->shares; for_each_cpu(i, sched_domain_span(sd)) - update_group_shares_cpu(tg, i, shares, rq_weight); + update_group_shares_cpu(tg, i, shares, rq_weight, usd); + + local_irq_restore(flags); return 0; } @@ -1577,8 +1648,14 @@ static int tg_load_down(struct task_group *tg, void *data) static void update_shares(struct sched_domain *sd) { - u64 now = cpu_clock(raw_smp_processor_id()); - s64 elapsed = now - sd->last_update; + s64 elapsed; + u64 now; + + if (root_task_group_empty()) + return; + + now = cpu_clock(raw_smp_processor_id()); + elapsed = now - sd->last_update; if (elapsed >= (s64)(u64)sysctl_sched_shares_ratelimit) { sd->last_update = now; @@ -1588,6 +1665,9 @@ static void update_shares(struct sched_domain *sd) static void update_shares_locked(struct rq *rq, struct sched_domain *sd) { + if (root_task_group_empty()) + return; + spin_unlock(&rq->lock); update_shares(sd); spin_lock(&rq->lock); @@ -1595,6 +1675,9 @@ static void update_shares_locked(struct rq *rq, struct sched_domain *sd) static void update_h_load(long cpu) { + if (root_task_group_empty()) + return; + walk_tg_tree(tg_load_down, tg_nop, (void *)cpu); } @@ -1610,21 +1693,42 @@ static inline void update_shares_locked(struct rq *rq, struct sched_domain *sd) #endif +#ifdef CONFIG_PREEMPT + /* - * double_lock_balance - lock the busiest runqueue, this_rq is locked already. + * fair double_lock_balance: Safely acquires both rq->locks in a fair + * way at the expense of forcing extra atomic operations in all + * invocations. This assures that the double_lock is acquired using the + * same underlying policy as the spinlock_t on this architecture, which + * reduces latency compared to the unfair variant below. However, it + * also adds more overhead and therefore may reduce throughput. */ -static int double_lock_balance(struct rq *this_rq, struct rq *busiest) +static inline int _double_lock_balance(struct rq *this_rq, struct rq *busiest) + __releases(this_rq->lock) + __acquires(busiest->lock) + __acquires(this_rq->lock) +{ + spin_unlock(&this_rq->lock); + double_rq_lock(this_rq, busiest); + + return 1; +} + +#else +/* + * Unfair double_lock_balance: Optimizes throughput at the expense of + * latency by eliminating extra atomic operations when the locks are + * already in proper order on entry. This favors lower cpu-ids and will + * grant the double lock to lower cpus over higher ids under contention, + * regardless of entry order into the function. + */ +static int _double_lock_balance(struct rq *this_rq, struct rq *busiest) __releases(this_rq->lock) __acquires(busiest->lock) __acquires(this_rq->lock) { int ret = 0; - if (unlikely(!irqs_disabled())) { - /* printk() doesn't work good under rq->lock */ - spin_unlock(&this_rq->lock); - BUG_ON(1); - } if (unlikely(!spin_trylock(&busiest->lock))) { if (busiest < this_rq) { spin_unlock(&this_rq->lock); @@ -1637,6 +1741,22 @@ static int double_lock_balance(struct rq *this_rq, struct rq *busiest) return ret; } +#endif /* CONFIG_PREEMPT */ + +/* + * double_lock_balance - lock the busiest runqueue, this_rq is locked already. + */ +static int double_lock_balance(struct rq *this_rq, struct rq *busiest) +{ + if (unlikely(!irqs_disabled())) { + /* printk() doesn't work good under rq->lock */ + spin_unlock(&this_rq->lock); + BUG_ON(1); + } + + return _double_lock_balance(this_rq, busiest); +} + static inline void double_unlock_balance(struct rq *this_rq, struct rq *busiest) __releases(busiest->lock) { @@ -1654,6 +1774,8 @@ static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares) } #endif +static void calc_load_account_active(struct rq *this_rq); + #include "sched_stats.h" #include "sched_idletask.c" #include "sched_fair.c" @@ -1705,6 +1827,9 @@ static void update_avg(u64 *avg, u64 sample) static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup) { + if (wakeup) + p->se.start_runtime = p->se.sum_exec_runtime; + sched_info_queued(p); p->sched_class->enqueue_task(rq, p, wakeup); p->se.on_rq = 1; @@ -1712,10 +1837,15 @@ static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup) static void dequeue_task(struct rq *rq, struct task_struct *p, int sleep) { - if (sleep && p->se.last_wakeup) { - update_avg(&p->se.avg_overlap, - p->se.sum_exec_runtime - p->se.last_wakeup); - p->se.last_wakeup = 0; + if (sleep) { + if (p->se.last_wakeup) { + update_avg(&p->se.avg_overlap, + p->se.sum_exec_runtime - p->se.last_wakeup); + p->se.last_wakeup = 0; + } else { + update_avg(&p->se.avg_wakeup, + sysctl_sched_wakeup_granularity); + } } sched_info_dequeued(p); @@ -1876,7 +2006,7 @@ void set_task_cpu(struct task_struct *p, unsigned int new_cpu) clock_offset = old_rq->clock - new_rq->clock; - trace_sched_migrate_task(p, task_cpu(p), new_cpu); + trace_sched_migrate_task(p, new_cpu); #ifdef CONFIG_SCHEDSTATS if (p->se.wait_start) @@ -1885,12 +2015,17 @@ void set_task_cpu(struct task_struct *p, unsigned int new_cpu) p->se.sleep_start -= clock_offset; if (p->se.block_start) p->se.block_start -= clock_offset; +#endif if (old_cpu != new_cpu) { - schedstat_inc(p, se.nr_migrations); + p->se.nr_migrations++; + new_rq->nr_migrations_in++; +#ifdef CONFIG_SCHEDSTATS if (task_hot(p, old_rq->clock, NULL)) schedstat_inc(p, se.nr_forced2_migrations); - } #endif + perf_swcounter_event(PERF_COUNT_SW_CPU_MIGRATIONS, + 1, 1, NULL, 0); + } p->se.vruntime -= old_cfsrq->min_vruntime - new_cfsrq->min_vruntime; @@ -1933,6 +2068,49 @@ migrate_task(struct task_struct *p, int dest_cpu, struct migration_req *req) } /* + * wait_task_context_switch - wait for a thread to complete at least one + * context switch. + * + * @p must not be current. + */ +void wait_task_context_switch(struct task_struct *p) +{ + unsigned long nvcsw, nivcsw, flags; + int running; + struct rq *rq; + + nvcsw = p->nvcsw; + nivcsw = p->nivcsw; + for (;;) { + /* + * The runqueue is assigned before the actual context + * switch. We need to take the runqueue lock. + * + * We could check initially without the lock but it is + * very likely that we need to take the lock in every + * iteration. + */ + rq = task_rq_lock(p, &flags); + running = task_running(rq, p); + task_rq_unlock(rq, &flags); + + if (likely(!running)) + break; + /* + * The switch count is incremented before the actual + * context switch. We thus wait for two switches to be + * sure at least one completed. + */ + if ((p->nvcsw - nvcsw) > 1) + break; + if ((p->nivcsw - nivcsw) > 1) + break; + + cpu_relax(); + } +} + +/* * wait_task_inactive - wait for a thread to unschedule. * * If @match_state is nonzero, it's the @p->state value just checked and @@ -2017,7 +2195,7 @@ unsigned long wait_task_inactive(struct task_struct *p, long match_state) * it must be off the runqueue _entirely_, and not * preempted! * - * So if it wa still runnable (but just not actively + * So if it was still runnable (but just not actively * running right now), it's preempted, and we should * yield - it could be a while. */ @@ -2060,6 +2238,7 @@ void kick_process(struct task_struct *p) smp_send_reschedule(cpu); preempt_enable(); } +EXPORT_SYMBOL_GPL(kick_process); /* * Return a low guess at the load of a migration-source cpu weighted @@ -2133,8 +2312,7 @@ find_idlest_group(struct sched_domain *sd, struct task_struct *p, int this_cpu) } /* Adjust by relative CPU power of the group */ - avg_load = sg_div_cpu_power(group, - avg_load * SCHED_LOAD_SCALE); + avg_load = (avg_load * SCHED_LOAD_SCALE) / group->cpu_power; if (local_group) { this_load = avg_load; @@ -2242,6 +2420,27 @@ static int sched_balance_self(int cpu, int flag) #endif /* CONFIG_SMP */ +/** + * task_oncpu_function_call - call a function on the cpu on which a task runs + * @p: the task to evaluate + * @func: the function to be called + * @info: the function call argument + * + * Calls the function @func when the task is currently running. This might + * be on the current CPU, which just calls the function directly + */ +void task_oncpu_function_call(struct task_struct *p, + void (*func) (void *info), void *info) +{ + int cpu; + + preempt_disable(); + cpu = task_cpu(p); + if (task_curr(p)) + smp_call_function_single(cpu, func, info, 1); + preempt_enable(); +} + /*** * try_to_wake_up - wake up a thread * @p: the to-be-woken-up thread @@ -2267,7 +2466,7 @@ static int try_to_wake_up(struct task_struct *p, unsigned int state, int sync) sync = 0; #ifdef CONFIG_SMP - if (sched_feat(LB_WAKEUP_UPDATE)) { + if (sched_feat(LB_WAKEUP_UPDATE) && !root_task_group_empty()) { struct sched_domain *sd; this_cpu = raw_smp_processor_id(); @@ -2345,6 +2544,22 @@ out_activate: activate_task(rq, p, 1); success = 1; + /* + * Only attribute actual wakeups done by this task. + */ + if (!in_interrupt()) { + struct sched_entity *se = ¤t->se; + u64 sample = se->sum_exec_runtime; + + if (se->last_wakeup) + sample -= se->last_wakeup; + else + sample -= se->start_runtime; + update_avg(&se->avg_wakeup, sample); + + se->last_wakeup = se->sum_exec_runtime; + } + out_running: trace_sched_wakeup(rq, p, success); check_preempt_curr(rq, p, sync); @@ -2355,13 +2570,22 @@ out_running: p->sched_class->task_wake_up(rq, p); #endif out: - current->se.last_wakeup = current->se.sum_exec_runtime; - task_rq_unlock(rq, &flags); return success; } +/** + * wake_up_process - Wake up a specific process + * @p: The process to be woken up. + * + * Attempt to wake up the nominated process and move it to the set of runnable + * processes. Returns 1 if the process was woken up, 0 if it was already + * running. + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. + */ int wake_up_process(struct task_struct *p) { return try_to_wake_up(p, TASK_ALL, 0); @@ -2384,19 +2608,44 @@ static void __sched_fork(struct task_struct *p) p->se.exec_start = 0; p->se.sum_exec_runtime = 0; p->se.prev_sum_exec_runtime = 0; + p->se.nr_migrations = 0; p->se.last_wakeup = 0; p->se.avg_overlap = 0; + p->se.start_runtime = 0; + p->se.avg_wakeup = sysctl_sched_wakeup_granularity; #ifdef CONFIG_SCHEDSTATS - p->se.wait_start = 0; - p->se.sum_sleep_runtime = 0; - p->se.sleep_start = 0; - p->se.block_start = 0; - p->se.sleep_max = 0; - p->se.block_max = 0; - p->se.exec_max = 0; - p->se.slice_max = 0; - p->se.wait_max = 0; + p->se.wait_start = 0; + p->se.wait_max = 0; + p->se.wait_count = 0; + p->se.wait_sum = 0; + + p->se.sleep_start = 0; + p->se.sleep_max = 0; + p->se.sum_sleep_runtime = 0; + + p->se.block_start = 0; + p->se.block_max = 0; + p->se.exec_max = 0; + p->se.slice_max = 0; + + p->se.nr_migrations_cold = 0; + p->se.nr_failed_migrations_affine = 0; + p->se.nr_failed_migrations_running = 0; + p->se.nr_failed_migrations_hot = 0; + p->se.nr_forced_migrations = 0; + p->se.nr_forced2_migrations = 0; + + p->se.nr_wakeups = 0; + p->se.nr_wakeups_sync = 0; + p->se.nr_wakeups_migrate = 0; + p->se.nr_wakeups_local = 0; + p->se.nr_wakeups_remote = 0; + p->se.nr_wakeups_affine = 0; + p->se.nr_wakeups_affine_attempts = 0; + p->se.nr_wakeups_passive = 0; + p->se.nr_wakeups_idle = 0; + #endif INIT_LIST_HEAD(&p->rt.run_list); @@ -2431,9 +2680,32 @@ void sched_fork(struct task_struct *p, int clone_flags) set_task_cpu(p, cpu); /* - * Make sure we do not leak PI boosting priority to the child: + * Make sure we do not leak PI boosting priority to the child. */ p->prio = current->normal_prio; + + /* + * Revert to default priority/policy on fork if requested. + */ + if (unlikely(p->sched_reset_on_fork)) { + if (p->policy == SCHED_FIFO || p->policy == SCHED_RR) + p->policy = SCHED_NORMAL; + + if (p->normal_prio < DEFAULT_PRIO) + p->prio = DEFAULT_PRIO; + + if (PRIO_TO_NICE(p->static_prio) < 0) { + p->static_prio = NICE_TO_PRIO(0); + set_load_weight(p); + } + + /* + * We don't need the reset flag anymore after the fork. It has + * fulfilled its duty: + */ + p->sched_reset_on_fork = 0; + } + if (!rt_prio(p->prio)) p->sched_class = &fair_sched_class; @@ -2448,6 +2720,8 @@ void sched_fork(struct task_struct *p, int clone_flags) /* Want to start with kernel preemption disabled. */ task_thread_info(p)->preempt_count = 1; #endif + plist_node_init(&p->pushable_tasks, MAX_PRIO); + put_cpu(); } @@ -2491,7 +2765,7 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags) #ifdef CONFIG_PREEMPT_NOTIFIERS /** - * preempt_notifier_register - tell me when current is being being preempted & rescheduled + * preempt_notifier_register - tell me when current is being preempted & rescheduled * @notifier: notifier struct to register */ void preempt_notifier_register(struct preempt_notifier *notifier) @@ -2604,11 +2878,8 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev) */ prev_state = prev->state; finish_arch_switch(prev); + perf_counter_task_sched_in(current, cpu_of(rq)); finish_lock_switch(rq, prev); -#ifdef CONFIG_SMP - if (current->sched_class->post_schedule) - current->sched_class->post_schedule(rq); -#endif fire_sched_in_preempt_notifiers(current); if (mm) @@ -2623,6 +2894,42 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev) } } +#ifdef CONFIG_SMP + +/* assumes rq->lock is held */ +static inline void pre_schedule(struct rq *rq, struct task_struct *prev) +{ + if (prev->sched_class->pre_schedule) + prev->sched_class->pre_schedule(rq, prev); +} + +/* rq->lock is NOT held, but preemption is disabled */ +static inline void post_schedule(struct rq *rq) +{ + if (rq->post_schedule) { + unsigned long flags; + + spin_lock_irqsave(&rq->lock, flags); + if (rq->curr->sched_class->post_schedule) + rq->curr->sched_class->post_schedule(rq); + spin_unlock_irqrestore(&rq->lock, flags); + + rq->post_schedule = 0; + } +} + +#else + +static inline void pre_schedule(struct rq *rq, struct task_struct *p) +{ +} + +static inline void post_schedule(struct rq *rq) +{ +} + +#endif + /** * schedule_tail - first thing a freshly forked thread must call. * @prev: the thread we just switched away from. @@ -2633,6 +2940,13 @@ asmlinkage void schedule_tail(struct task_struct *prev) struct rq *rq = this_rq(); finish_task_switch(rq, prev); + + /* + * FIXME: do we need to worry about rq being invalidated by the + * task_switch? + */ + post_schedule(rq); + #ifdef __ARCH_WANT_UNLOCKED_CTXSW /* In this case, finish_task_switch does not reenable preemption */ preempt_enable(); @@ -2660,7 +2974,7 @@ context_switch(struct rq *rq, struct task_struct *prev, * combine the page table reload and the switch backend into * one hypercall. */ - arch_enter_lazy_cpu_mode(); + arch_start_context_switch(prev); if (unlikely(!mm)) { next->active_mm = oldmm; @@ -2750,19 +3064,81 @@ unsigned long nr_iowait(void) return sum; } -unsigned long nr_active(void) +/* Variables and functions for calc_load */ +static atomic_long_t calc_load_tasks; +static unsigned long calc_load_update; +unsigned long avenrun[3]; +EXPORT_SYMBOL(avenrun); + +/** + * get_avenrun - get the load average array + * @loads: pointer to dest load array + * @offset: offset to add + * @shift: shift count to shift the result left + * + * These values are estimates at best, so no need for locking. + */ +void get_avenrun(unsigned long *loads, unsigned long offset, int shift) +{ + loads[0] = (avenrun[0] + offset) << shift; + loads[1] = (avenrun[1] + offset) << shift; + loads[2] = (avenrun[2] + offset) << shift; +} + +static unsigned long +calc_load(unsigned long load, unsigned long exp, unsigned long active) { - unsigned long i, running = 0, uninterruptible = 0; + load *= exp; + load += active * (FIXED_1 - exp); + return load >> FSHIFT; +} - for_each_online_cpu(i) { - running += cpu_rq(i)->nr_running; - uninterruptible += cpu_rq(i)->nr_uninterruptible; - } +/* + * calc_load - update the avenrun load estimates 10 ticks after the + * CPUs have updated calc_load_tasks. + */ +void calc_global_load(void) +{ + unsigned long upd = calc_load_update + 10; + long active; + + if (time_before(jiffies, upd)) + return; + + active = atomic_long_read(&calc_load_tasks); + active = active > 0 ? active * FIXED_1 : 0; + + avenrun[0] = calc_load(avenrun[0], EXP_1, active); + avenrun[1] = calc_load(avenrun[1], EXP_5, active); + avenrun[2] = calc_load(avenrun[2], EXP_15, active); + + calc_load_update += LOAD_FREQ; +} + +/* + * Either called from update_cpu_load() or from a cpu going idle + */ +static void calc_load_account_active(struct rq *this_rq) +{ + long nr_active, delta; - if (unlikely((long)uninterruptible < 0)) - uninterruptible = 0; + nr_active = this_rq->nr_running; + nr_active += (long) this_rq->nr_uninterruptible; - return running + uninterruptible; + if (nr_active != this_rq->calc_load_active) { + delta = nr_active - this_rq->calc_load_active; + this_rq->calc_load_active = nr_active; + atomic_long_add(delta, &calc_load_tasks); + } +} + +/* + * Externally visible per-cpu scheduler statistics: + * cpu_nr_migrations(cpu) - number of migrations into that cpu + */ +u64 cpu_nr_migrations(int cpu) +{ + return cpu_rq(cpu)->nr_migrations_in; } /* @@ -2793,6 +3169,11 @@ static void update_cpu_load(struct rq *this_rq) new_load += scale-1; this_rq->cpu_load[i] = (old_load*(scale-1) + new_load) >> i; } + + if (time_after_eq(jiffies, this_rq->calc_load_update)) { + this_rq->calc_load_update += LOAD_FREQ; + calc_load_account_active(this_rq); + } } #ifdef CONFIG_SMP @@ -2913,6 +3294,7 @@ int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned) { + int tsk_cache_hot = 0; /* * We do not migrate tasks that are: * 1) running (obviously), or @@ -2936,10 +3318,11 @@ int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, * 2) too many balance attempts have failed. */ - if (!task_hot(p, rq->clock, sd) || - sd->nr_balance_failed > sd->cache_nice_tries) { + tsk_cache_hot = task_hot(p, rq->clock, sd); + if (!tsk_cache_hot || + sd->nr_balance_failed > sd->cache_nice_tries) { #ifdef CONFIG_SCHEDSTATS - if (task_hot(p, rq->clock, sd)) { + if (tsk_cache_hot) { schedstat_inc(sd, lb_hot_gained[idle]); schedstat_inc(p, se.nr_forced_migrations); } @@ -2947,7 +3330,7 @@ int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, return 1; } - if (task_hot(p, rq->clock, sd)) { + if (tsk_cache_hot) { schedstat_inc(p, se.nr_failed_migrations_hot); return 0; } @@ -2987,6 +3370,16 @@ next: pulled++; rem_load_move -= p->se.load.weight; +#ifdef CONFIG_PREEMPT + /* + * NEWIDLE balancing is a source of latency, so preemptible kernels + * will stop after the first task is pulled to minimize the critical + * section. + */ + if (idle == CPU_NEWLY_IDLE) + goto out; +#endif + /* * We only want to steal up to the prescribed amount of weighted load. */ @@ -3033,9 +3426,15 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, sd, idle, all_pinned, &this_best_prio); class = class->next; +#ifdef CONFIG_PREEMPT + /* + * NEWIDLE balancing is a source of latency, so preemptible + * kernels will stop after the first task is pulled to minimize + * the critical section. + */ if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) break; - +#endif } while (class && max_load_move > total_load_moved); return total_load_moved > 0; @@ -3079,252 +3478,572 @@ static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, { const struct sched_class *class; - for (class = sched_class_highest; class; class = class->next) + for_each_class(class) { if (class->move_one_task(this_rq, this_cpu, busiest, sd, idle)) return 1; + } return 0; } - +/********** Helpers for find_busiest_group ************************/ /* - * find_busiest_group finds and returns the busiest CPU group within the - * domain. It calculates and returns the amount of weighted load which - * should be moved to restore balance via the imbalance parameter. + * sd_lb_stats - Structure to store the statistics of a sched_domain + * during load balancing. */ -static struct sched_group * -find_busiest_group(struct sched_domain *sd, int this_cpu, - unsigned long *imbalance, enum cpu_idle_type idle, - int *sd_idle, const struct cpumask *cpus, int *balance) -{ - struct sched_group *busiest = NULL, *this = NULL, *group = sd->groups; - unsigned long max_load, avg_load, total_load, this_load, total_pwr; - unsigned long max_pull; - unsigned long busiest_load_per_task, busiest_nr_running; - unsigned long this_load_per_task, this_nr_running; - int load_idx, group_imb = 0; +struct sd_lb_stats { + struct sched_group *busiest; /* Busiest group in this sd */ + struct sched_group *this; /* Local group in this sd */ + unsigned long total_load; /* Total load of all groups in sd */ + unsigned long total_pwr; /* Total power of all groups in sd */ + unsigned long avg_load; /* Average load across all groups in sd */ + + /** Statistics of this group */ + unsigned long this_load; + unsigned long this_load_per_task; + unsigned long this_nr_running; + + /* Statistics of the busiest group */ + unsigned long max_load; + unsigned long busiest_load_per_task; + unsigned long busiest_nr_running; + + int group_imb; /* Is there imbalance in this sd */ #if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) - int power_savings_balance = 1; - unsigned long leader_nr_running = 0, min_load_per_task = 0; - unsigned long min_nr_running = ULONG_MAX; - struct sched_group *group_min = NULL, *group_leader = NULL; + int power_savings_balance; /* Is powersave balance needed for this sd */ + struct sched_group *group_min; /* Least loaded group in sd */ + struct sched_group *group_leader; /* Group which relieves group_min */ + unsigned long min_load_per_task; /* load_per_task in group_min */ + unsigned long leader_nr_running; /* Nr running of group_leader */ + unsigned long min_nr_running; /* Nr running of group_min */ #endif +}; - max_load = this_load = total_load = total_pwr = 0; - busiest_load_per_task = busiest_nr_running = 0; - this_load_per_task = this_nr_running = 0; +/* + * sg_lb_stats - stats of a sched_group required for load_balancing + */ +struct sg_lb_stats { + unsigned long avg_load; /*Avg load across the CPUs of the group */ + unsigned long group_load; /* Total load over the CPUs of the group */ + unsigned long sum_nr_running; /* Nr tasks running in the group */ + unsigned long sum_weighted_load; /* Weighted load of group's tasks */ + unsigned long group_capacity; + int group_imb; /* Is there an imbalance in the group ? */ +}; + +/** + * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. + * @group: The group whose first cpu is to be returned. + */ +static inline unsigned int group_first_cpu(struct sched_group *group) +{ + return cpumask_first(sched_group_cpus(group)); +} + +/** + * get_sd_load_idx - Obtain the load index for a given sched domain. + * @sd: The sched_domain whose load_idx is to be obtained. + * @idle: The Idle status of the CPU for whose sd load_icx is obtained. + */ +static inline int get_sd_load_idx(struct sched_domain *sd, + enum cpu_idle_type idle) +{ + int load_idx; - if (idle == CPU_NOT_IDLE) + switch (idle) { + case CPU_NOT_IDLE: load_idx = sd->busy_idx; - else if (idle == CPU_NEWLY_IDLE) + break; + + case CPU_NEWLY_IDLE: load_idx = sd->newidle_idx; - else + break; + default: load_idx = sd->idle_idx; + break; + } - do { - unsigned long load, group_capacity, max_cpu_load, min_cpu_load; - int local_group; - int i; - int __group_imb = 0; - unsigned int balance_cpu = -1, first_idle_cpu = 0; - unsigned long sum_nr_running, sum_weighted_load; - unsigned long sum_avg_load_per_task; - unsigned long avg_load_per_task; + return load_idx; +} - local_group = cpumask_test_cpu(this_cpu, - sched_group_cpus(group)); - if (local_group) - balance_cpu = cpumask_first(sched_group_cpus(group)); +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * init_sd_power_savings_stats - Initialize power savings statistics for + * the given sched_domain, during load balancing. + * + * @sd: Sched domain whose power-savings statistics are to be initialized. + * @sds: Variable containing the statistics for sd. + * @idle: Idle status of the CPU at which we're performing load-balancing. + */ +static inline void init_sd_power_savings_stats(struct sched_domain *sd, + struct sd_lb_stats *sds, enum cpu_idle_type idle) +{ + /* + * Busy processors will not participate in power savings + * balance. + */ + if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) + sds->power_savings_balance = 0; + else { + sds->power_savings_balance = 1; + sds->min_nr_running = ULONG_MAX; + sds->leader_nr_running = 0; + } +} - /* Tally up the load of all CPUs in the group */ - sum_weighted_load = sum_nr_running = avg_load = 0; - sum_avg_load_per_task = avg_load_per_task = 0; +/** + * update_sd_power_savings_stats - Update the power saving stats for a + * sched_domain while performing load balancing. + * + * @group: sched_group belonging to the sched_domain under consideration. + * @sds: Variable containing the statistics of the sched_domain + * @local_group: Does group contain the CPU for which we're performing + * load balancing ? + * @sgs: Variable containing the statistics of the group. + */ +static inline void update_sd_power_savings_stats(struct sched_group *group, + struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) +{ - max_cpu_load = 0; - min_cpu_load = ~0UL; + if (!sds->power_savings_balance) + return; - for_each_cpu_and(i, sched_group_cpus(group), cpus) { - struct rq *rq = cpu_rq(i); + /* + * If the local group is idle or completely loaded + * no need to do power savings balance at this domain + */ + if (local_group && (sds->this_nr_running >= sgs->group_capacity || + !sds->this_nr_running)) + sds->power_savings_balance = 0; - if (*sd_idle && rq->nr_running) - *sd_idle = 0; + /* + * If a group is already running at full capacity or idle, + * don't include that group in power savings calculations + */ + if (!sds->power_savings_balance || + sgs->sum_nr_running >= sgs->group_capacity || + !sgs->sum_nr_running) + return; - /* Bias balancing toward cpus of our domain */ - if (local_group) { - if (idle_cpu(i) && !first_idle_cpu) { - first_idle_cpu = 1; - balance_cpu = i; - } + /* + * Calculate the group which has the least non-idle load. + * This is the group from where we need to pick up the load + * for saving power + */ + if ((sgs->sum_nr_running < sds->min_nr_running) || + (sgs->sum_nr_running == sds->min_nr_running && + group_first_cpu(group) > group_first_cpu(sds->group_min))) { + sds->group_min = group; + sds->min_nr_running = sgs->sum_nr_running; + sds->min_load_per_task = sgs->sum_weighted_load / + sgs->sum_nr_running; + } - load = target_load(i, load_idx); - } else { - load = source_load(i, load_idx); - if (load > max_cpu_load) - max_cpu_load = load; - if (min_cpu_load > load) - min_cpu_load = load; - } + /* + * Calculate the group which is almost near its + * capacity but still has some space to pick up some load + * from other group and save more power + */ + if (sgs->sum_nr_running + 1 > sgs->group_capacity) + return; - avg_load += load; - sum_nr_running += rq->nr_running; - sum_weighted_load += weighted_cpuload(i); + if (sgs->sum_nr_running > sds->leader_nr_running || + (sgs->sum_nr_running == sds->leader_nr_running && + group_first_cpu(group) < group_first_cpu(sds->group_leader))) { + sds->group_leader = group; + sds->leader_nr_running = sgs->sum_nr_running; + } +} - sum_avg_load_per_task += cpu_avg_load_per_task(i); - } +/** + * check_power_save_busiest_group - see if there is potential for some power-savings balance + * @sds: Variable containing the statistics of the sched_domain + * under consideration. + * @this_cpu: Cpu at which we're currently performing load-balancing. + * @imbalance: Variable to store the imbalance. + * + * Description: + * Check if we have potential to perform some power-savings balance. + * If yes, set the busiest group to be the least loaded group in the + * sched_domain, so that it's CPUs can be put to idle. + * + * Returns 1 if there is potential to perform power-savings balance. + * Else returns 0. + */ +static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + if (!sds->power_savings_balance) + return 0; - /* - * First idle cpu or the first cpu(busiest) in this sched group - * is eligible for doing load balancing at this and above - * domains. In the newly idle case, we will allow all the cpu's - * to do the newly idle load balance. - */ - if (idle != CPU_NEWLY_IDLE && local_group && - balance_cpu != this_cpu && balance) { - *balance = 0; - goto ret; - } + if (sds->this != sds->group_leader || + sds->group_leader == sds->group_min) + return 0; - total_load += avg_load; - total_pwr += group->__cpu_power; + *imbalance = sds->min_load_per_task; + sds->busiest = sds->group_min; - /* Adjust by relative CPU power of the group */ - avg_load = sg_div_cpu_power(group, - avg_load * SCHED_LOAD_SCALE); + if (sched_mc_power_savings >= POWERSAVINGS_BALANCE_WAKEUP) { + cpu_rq(this_cpu)->rd->sched_mc_preferred_wakeup_cpu = + group_first_cpu(sds->group_leader); + } + return 1; - /* - * Consider the group unbalanced when the imbalance is larger - * than the average weight of two tasks. - * - * APZ: with cgroup the avg task weight can vary wildly and - * might not be a suitable number - should we keep a - * normalized nr_running number somewhere that negates - * the hierarchy? - */ - avg_load_per_task = sg_div_cpu_power(group, - sum_avg_load_per_task * SCHED_LOAD_SCALE); +} +#else /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ +static inline void init_sd_power_savings_stats(struct sched_domain *sd, + struct sd_lb_stats *sds, enum cpu_idle_type idle) +{ + return; +} - if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) - __group_imb = 1; +static inline void update_sd_power_savings_stats(struct sched_group *group, + struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) +{ + return; +} - group_capacity = group->__cpu_power / SCHED_LOAD_SCALE; +static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + return 0; +} +#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ +unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) +{ + unsigned long weight = cpumask_weight(sched_domain_span(sd)); + unsigned long smt_gain = sd->smt_gain; + + smt_gain /= weight; + + return smt_gain; +} + +unsigned long scale_rt_power(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + u64 total, available; + + sched_avg_update(rq); + + total = sched_avg_period() + (rq->clock - rq->age_stamp); + available = total - rq->rt_avg; + + if (unlikely((s64)total < SCHED_LOAD_SCALE)) + total = SCHED_LOAD_SCALE; + + total >>= SCHED_LOAD_SHIFT; + + return div_u64(available, total); +} + +static void update_cpu_power(struct sched_domain *sd, int cpu) +{ + unsigned long weight = cpumask_weight(sched_domain_span(sd)); + unsigned long power = SCHED_LOAD_SCALE; + struct sched_group *sdg = sd->groups; + + /* here we could scale based on cpufreq */ + + if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { + power *= arch_scale_smt_power(sd, cpu); + power >>= SCHED_LOAD_SHIFT; + } + + power *= scale_rt_power(cpu); + power >>= SCHED_LOAD_SHIFT; + + if (!power) + power = 1; + + sdg->cpu_power = power; +} + +static void update_group_power(struct sched_domain *sd, int cpu) +{ + struct sched_domain *child = sd->child; + struct sched_group *group, *sdg = sd->groups; + unsigned long power; + + if (!child) { + update_cpu_power(sd, cpu); + return; + } + + power = 0; + + group = child->groups; + do { + power += group->cpu_power; + group = group->next; + } while (group != child->groups); + + sdg->cpu_power = power; +} + +/** + * update_sg_lb_stats - Update sched_group's statistics for load balancing. + * @group: sched_group whose statistics are to be updated. + * @this_cpu: Cpu for which load balance is currently performed. + * @idle: Idle status of this_cpu + * @load_idx: Load index of sched_domain of this_cpu for load calc. + * @sd_idle: Idle status of the sched_domain containing group. + * @local_group: Does group contain this_cpu. + * @cpus: Set of cpus considered for load balancing. + * @balance: Should we balance. + * @sgs: variable to hold the statistics for this group. + */ +static inline void update_sg_lb_stats(struct sched_domain *sd, + struct sched_group *group, int this_cpu, + enum cpu_idle_type idle, int load_idx, int *sd_idle, + int local_group, const struct cpumask *cpus, + int *balance, struct sg_lb_stats *sgs) +{ + unsigned long load, max_cpu_load, min_cpu_load; + int i; + unsigned int balance_cpu = -1, first_idle_cpu = 0; + unsigned long sum_avg_load_per_task; + unsigned long avg_load_per_task; + + if (local_group) { + balance_cpu = group_first_cpu(group); + if (balance_cpu == this_cpu) + update_group_power(sd, this_cpu); + } + + /* Tally up the load of all CPUs in the group */ + sum_avg_load_per_task = avg_load_per_task = 0; + max_cpu_load = 0; + min_cpu_load = ~0UL; + + for_each_cpu_and(i, sched_group_cpus(group), cpus) { + struct rq *rq = cpu_rq(i); + + if (*sd_idle && rq->nr_running) + *sd_idle = 0; + + /* Bias balancing toward cpus of our domain */ if (local_group) { - this_load = avg_load; - this = group; - this_nr_running = sum_nr_running; - this_load_per_task = sum_weighted_load; - } else if (avg_load > max_load && - (sum_nr_running > group_capacity || __group_imb)) { - max_load = avg_load; - busiest = group; - busiest_nr_running = sum_nr_running; - busiest_load_per_task = sum_weighted_load; - group_imb = __group_imb; + if (idle_cpu(i) && !first_idle_cpu) { + first_idle_cpu = 1; + balance_cpu = i; + } + + load = target_load(i, load_idx); + } else { + load = source_load(i, load_idx); + if (load > max_cpu_load) + max_cpu_load = load; + if (min_cpu_load > load) + min_cpu_load = load; } -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) - /* - * Busy processors will not participate in power savings - * balance. - */ - if (idle == CPU_NOT_IDLE || - !(sd->flags & SD_POWERSAVINGS_BALANCE)) - goto group_next; + sgs->group_load += load; + sgs->sum_nr_running += rq->nr_running; + sgs->sum_weighted_load += weighted_cpuload(i); - /* - * If the local group is idle or completely loaded - * no need to do power savings balance at this domain - */ - if (local_group && (this_nr_running >= group_capacity || - !this_nr_running)) - power_savings_balance = 0; + sum_avg_load_per_task += cpu_avg_load_per_task(i); + } - /* - * If a group is already running at full capacity or idle, - * don't include that group in power savings calculations - */ - if (!power_savings_balance || sum_nr_running >= group_capacity - || !sum_nr_running) - goto group_next; + /* + * First idle cpu or the first cpu(busiest) in this sched group + * is eligible for doing load balancing at this and above + * domains. In the newly idle case, we will allow all the cpu's + * to do the newly idle load balance. + */ + if (idle != CPU_NEWLY_IDLE && local_group && + balance_cpu != this_cpu && balance) { + *balance = 0; + return; + } - /* - * Calculate the group which has the least non-idle load. - * This is the group from where we need to pick up the load - * for saving power - */ - if ((sum_nr_running < min_nr_running) || - (sum_nr_running == min_nr_running && - cpumask_first(sched_group_cpus(group)) > - cpumask_first(sched_group_cpus(group_min)))) { - group_min = group; - min_nr_running = sum_nr_running; - min_load_per_task = sum_weighted_load / - sum_nr_running; - } + /* Adjust by relative CPU power of the group */ + sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; + + + /* + * Consider the group unbalanced when the imbalance is larger + * than the average weight of two tasks. + * + * APZ: with cgroup the avg task weight can vary wildly and + * might not be a suitable number - should we keep a + * normalized nr_running number somewhere that negates + * the hierarchy? + */ + avg_load_per_task = (sum_avg_load_per_task * SCHED_LOAD_SCALE) / + group->cpu_power; + + if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task) + sgs->group_imb = 1; + + sgs->group_capacity = + DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); +} + +/** + * update_sd_lb_stats - Update sched_group's statistics for load balancing. + * @sd: sched_domain whose statistics are to be updated. + * @this_cpu: Cpu for which load balance is currently performed. + * @idle: Idle status of this_cpu + * @sd_idle: Idle status of the sched_domain containing group. + * @cpus: Set of cpus considered for load balancing. + * @balance: Should we balance. + * @sds: variable to hold the statistics for this sched_domain. + */ +static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, + enum cpu_idle_type idle, int *sd_idle, + const struct cpumask *cpus, int *balance, + struct sd_lb_stats *sds) +{ + struct sched_domain *child = sd->child; + struct sched_group *group = sd->groups; + struct sg_lb_stats sgs; + int load_idx, prefer_sibling = 0; + + if (child && child->flags & SD_PREFER_SIBLING) + prefer_sibling = 1; + + init_sd_power_savings_stats(sd, sds, idle); + load_idx = get_sd_load_idx(sd, idle); + + do { + int local_group; + + local_group = cpumask_test_cpu(this_cpu, + sched_group_cpus(group)); + memset(&sgs, 0, sizeof(sgs)); + update_sg_lb_stats(sd, group, this_cpu, idle, load_idx, sd_idle, + local_group, cpus, balance, &sgs); + + if (local_group && balance && !(*balance)) + return; + + sds->total_load += sgs.group_load; + sds->total_pwr += group->cpu_power; /* - * Calculate the group which is almost near its - * capacity but still has some space to pick up some load - * from other group and save more power + * In case the child domain prefers tasks go to siblings + * first, lower the group capacity to one so that we'll try + * and move all the excess tasks away. */ - if (sum_nr_running <= group_capacity - 1) { - if (sum_nr_running > leader_nr_running || - (sum_nr_running == leader_nr_running && - cpumask_first(sched_group_cpus(group)) < - cpumask_first(sched_group_cpus(group_leader)))) { - group_leader = group; - leader_nr_running = sum_nr_running; - } + if (prefer_sibling) + sgs.group_capacity = min(sgs.group_capacity, 1UL); + + if (local_group) { + sds->this_load = sgs.avg_load; + sds->this = group; + sds->this_nr_running = sgs.sum_nr_running; + sds->this_load_per_task = sgs.sum_weighted_load; + } else if (sgs.avg_load > sds->max_load && + (sgs.sum_nr_running > sgs.group_capacity || + sgs.group_imb)) { + sds->max_load = sgs.avg_load; + sds->busiest = group; + sds->busiest_nr_running = sgs.sum_nr_running; + sds->busiest_load_per_task = sgs.sum_weighted_load; + sds->group_imb = sgs.group_imb; } -group_next: -#endif + + update_sd_power_savings_stats(group, sds, local_group, &sgs); group = group->next; } while (group != sd->groups); +} - if (!busiest || this_load >= max_load || busiest_nr_running == 0) - goto out_balanced; - - avg_load = (SCHED_LOAD_SCALE * total_load) / total_pwr; - - if (this_load >= avg_load || - 100*max_load <= sd->imbalance_pct*this_load) - goto out_balanced; +/** + * fix_small_imbalance - Calculate the minor imbalance that exists + * amongst the groups of a sched_domain, during + * load balancing. + * @sds: Statistics of the sched_domain whose imbalance is to be calculated. + * @this_cpu: The cpu at whose sched_domain we're performing load-balance. + * @imbalance: Variable to store the imbalance. + */ +static inline void fix_small_imbalance(struct sd_lb_stats *sds, + int this_cpu, unsigned long *imbalance) +{ + unsigned long tmp, pwr_now = 0, pwr_move = 0; + unsigned int imbn = 2; + + if (sds->this_nr_running) { + sds->this_load_per_task /= sds->this_nr_running; + if (sds->busiest_load_per_task > + sds->this_load_per_task) + imbn = 1; + } else + sds->this_load_per_task = + cpu_avg_load_per_task(this_cpu); - busiest_load_per_task /= busiest_nr_running; - if (group_imb) - busiest_load_per_task = min(busiest_load_per_task, avg_load); + if (sds->max_load - sds->this_load + sds->busiest_load_per_task >= + sds->busiest_load_per_task * imbn) { + *imbalance = sds->busiest_load_per_task; + return; + } /* - * We're trying to get all the cpus to the average_load, so we don't - * want to push ourselves above the average load, nor do we wish to - * reduce the max loaded cpu below the average load, as either of these - * actions would just result in more rebalancing later, and ping-pong - * tasks around. Thus we look for the minimum possible imbalance. - * Negative imbalances (*we* are more loaded than anyone else) will - * be counted as no imbalance for these purposes -- we can't fix that - * by pulling tasks to us. Be careful of negative numbers as they'll - * appear as very large values with unsigned longs. + * OK, we don't have enough imbalance to justify moving tasks, + * however we may be able to increase total CPU power used by + * moving them. */ - if (max_load <= busiest_load_per_task) - goto out_balanced; + pwr_now += sds->busiest->cpu_power * + min(sds->busiest_load_per_task, sds->max_load); + pwr_now += sds->this->cpu_power * + min(sds->this_load_per_task, sds->this_load); + pwr_now /= SCHED_LOAD_SCALE; + + /* Amount of load we'd subtract */ + tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / + sds->busiest->cpu_power; + if (sds->max_load > tmp) + pwr_move += sds->busiest->cpu_power * + min(sds->busiest_load_per_task, sds->max_load - tmp); + + /* Amount of load we'd add */ + if (sds->max_load * sds->busiest->cpu_power < + sds->busiest_load_per_task * SCHED_LOAD_SCALE) + tmp = (sds->max_load * sds->busiest->cpu_power) / + sds->this->cpu_power; + else + tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / + sds->this->cpu_power; + pwr_move += sds->this->cpu_power * + min(sds->this_load_per_task, sds->this_load + tmp); + pwr_move /= SCHED_LOAD_SCALE; + + /* Move if we gain throughput */ + if (pwr_move > pwr_now) + *imbalance = sds->busiest_load_per_task; +} + +/** + * calculate_imbalance - Calculate the amount of imbalance present within the + * groups of a given sched_domain during load balance. + * @sds: statistics of the sched_domain whose imbalance is to be calculated. + * @this_cpu: Cpu for which currently load balance is being performed. + * @imbalance: The variable to store the imbalance. + */ +static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, + unsigned long *imbalance) +{ + unsigned long max_pull; /* * In the presence of smp nice balancing, certain scenarios can have * max load less than avg load(as we skip the groups at or below * its cpu_power, while calculating max_load..) */ - if (max_load < avg_load) { + if (sds->max_load < sds->avg_load) { *imbalance = 0; - goto small_imbalance; + return fix_small_imbalance(sds, this_cpu, imbalance); } /* Don't want to pull so many tasks that a group would go idle */ - max_pull = min(max_load - avg_load, max_load - busiest_load_per_task); + max_pull = min(sds->max_load - sds->avg_load, + sds->max_load - sds->busiest_load_per_task); /* How much load to actually move to equalise the imbalance */ - *imbalance = min(max_pull * busiest->__cpu_power, - (avg_load - this_load) * this->__cpu_power) + *imbalance = min(max_pull * sds->busiest->cpu_power, + (sds->avg_load - sds->this_load) * sds->this->cpu_power) / SCHED_LOAD_SCALE; /* @@ -3333,83 +4052,135 @@ group_next: * a think about bumping its value to force at least one task to be * moved */ - if (*imbalance < busiest_load_per_task) { - unsigned long tmp, pwr_now, pwr_move; - unsigned int imbn; - -small_imbalance: - pwr_move = pwr_now = 0; - imbn = 2; - if (this_nr_running) { - this_load_per_task /= this_nr_running; - if (busiest_load_per_task > this_load_per_task) - imbn = 1; - } else - this_load_per_task = cpu_avg_load_per_task(this_cpu); + if (*imbalance < sds->busiest_load_per_task) + return fix_small_imbalance(sds, this_cpu, imbalance); - if (max_load - this_load + busiest_load_per_task >= - busiest_load_per_task * imbn) { - *imbalance = busiest_load_per_task; - return busiest; - } +} +/******* find_busiest_group() helpers end here *********************/ - /* - * OK, we don't have enough imbalance to justify moving tasks, - * however we may be able to increase total CPU power used by - * moving them. - */ +/** + * find_busiest_group - Returns the busiest group within the sched_domain + * if there is an imbalance. If there isn't an imbalance, and + * the user has opted for power-savings, it returns a group whose + * CPUs can be put to idle by rebalancing those tasks elsewhere, if + * such a group exists. + * + * Also calculates the amount of weighted load which should be moved + * to restore balance. + * + * @sd: The sched_domain whose busiest group is to be returned. + * @this_cpu: The cpu for which load balancing is currently being performed. + * @imbalance: Variable which stores amount of weighted load which should + * be moved to restore balance/put a group to idle. + * @idle: The idle status of this_cpu. + * @sd_idle: The idleness of sd + * @cpus: The set of CPUs under consideration for load-balancing. + * @balance: Pointer to a variable indicating if this_cpu + * is the appropriate cpu to perform load balancing at this_level. + * + * Returns: - the busiest group if imbalance exists. + * - If no imbalance and user has opted for power-savings balance, + * return the least loaded group whose CPUs can be + * put to idle by rebalancing its tasks onto our group. + */ +static struct sched_group * +find_busiest_group(struct sched_domain *sd, int this_cpu, + unsigned long *imbalance, enum cpu_idle_type idle, + int *sd_idle, const struct cpumask *cpus, int *balance) +{ + struct sd_lb_stats sds; - pwr_now += busiest->__cpu_power * - min(busiest_load_per_task, max_load); - pwr_now += this->__cpu_power * - min(this_load_per_task, this_load); - pwr_now /= SCHED_LOAD_SCALE; - - /* Amount of load we'd subtract */ - tmp = sg_div_cpu_power(busiest, - busiest_load_per_task * SCHED_LOAD_SCALE); - if (max_load > tmp) - pwr_move += busiest->__cpu_power * - min(busiest_load_per_task, max_load - tmp); - - /* Amount of load we'd add */ - if (max_load * busiest->__cpu_power < - busiest_load_per_task * SCHED_LOAD_SCALE) - tmp = sg_div_cpu_power(this, - max_load * busiest->__cpu_power); - else - tmp = sg_div_cpu_power(this, - busiest_load_per_task * SCHED_LOAD_SCALE); - pwr_move += this->__cpu_power * - min(this_load_per_task, this_load + tmp); - pwr_move /= SCHED_LOAD_SCALE; + memset(&sds, 0, sizeof(sds)); - /* Move if we gain throughput */ - if (pwr_move > pwr_now) - *imbalance = busiest_load_per_task; - } + /* + * Compute the various statistics relavent for load balancing at + * this level. + */ + update_sd_lb_stats(sd, this_cpu, idle, sd_idle, cpus, + balance, &sds); + + /* Cases where imbalance does not exist from POV of this_cpu */ + /* 1) this_cpu is not the appropriate cpu to perform load balancing + * at this level. + * 2) There is no busy sibling group to pull from. + * 3) This group is the busiest group. + * 4) This group is more busy than the avg busieness at this + * sched_domain. + * 5) The imbalance is within the specified limit. + * 6) Any rebalance would lead to ping-pong + */ + if (balance && !(*balance)) + goto ret; - return busiest; + if (!sds.busiest || sds.busiest_nr_running == 0) + goto out_balanced; -out_balanced: -#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) - if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) - goto ret; + if (sds.this_load >= sds.max_load) + goto out_balanced; - if (this == group_leader && group_leader != group_min) { - *imbalance = min_load_per_task; - if (sched_mc_power_savings >= POWERSAVINGS_BALANCE_WAKEUP) { - cpu_rq(this_cpu)->rd->sched_mc_preferred_wakeup_cpu = - cpumask_first(sched_group_cpus(group_leader)); - } - return group_min; - } -#endif + sds.avg_load = (SCHED_LOAD_SCALE * sds.total_load) / sds.total_pwr; + + if (sds.this_load >= sds.avg_load) + goto out_balanced; + + if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) + goto out_balanced; + + sds.busiest_load_per_task /= sds.busiest_nr_running; + if (sds.group_imb) + sds.busiest_load_per_task = + min(sds.busiest_load_per_task, sds.avg_load); + + /* + * We're trying to get all the cpus to the average_load, so we don't + * want to push ourselves above the average load, nor do we wish to + * reduce the max loaded cpu below the average load, as either of these + * actions would just result in more rebalancing later, and ping-pong + * tasks around. Thus we look for the minimum possible imbalance. + * Negative imbalances (*we* are more loaded than anyone else) will + * be counted as no imbalance for these purposes -- we can't fix that + * by pulling tasks to us. Be careful of negative numbers as they'll + * appear as very large values with unsigned longs. + */ + if (sds.max_load <= sds.busiest_load_per_task) + goto out_balanced; + + /* Looks like there is an imbalance. Compute it */ + calculate_imbalance(&sds, this_cpu, imbalance); + return sds.busiest; + +out_balanced: + /* + * There is no obvious imbalance. But check if we can do some balancing + * to save power. + */ + if (check_power_save_busiest_group(&sds, this_cpu, imbalance)) + return sds.busiest; ret: *imbalance = 0; return NULL; } +static struct sched_group *group_of(int cpu) +{ + struct sched_domain *sd = rcu_dereference(cpu_rq(cpu)->sd); + + if (!sd) + return NULL; + + return sd->groups; +} + +static unsigned long power_of(int cpu) +{ + struct sched_group *group = group_of(cpu); + + if (!group) + return SCHED_LOAD_SCALE; + + return group->cpu_power; +} + /* * find_busiest_queue - find the busiest runqueue among the cpus in group. */ @@ -3422,15 +4193,18 @@ find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, int i; for_each_cpu(i, sched_group_cpus(group)) { + unsigned long power = power_of(i); + unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); unsigned long wl; if (!cpumask_test_cpu(i, cpus)) continue; rq = cpu_rq(i); - wl = weighted_cpuload(i); + wl = weighted_cpuload(i) * SCHED_LOAD_SCALE; + wl /= power; - if (rq->nr_running == 1 && wl > imbalance) + if (capacity && rq->nr_running == 1 && wl > imbalance) continue; if (wl > max_load) { @@ -3448,19 +4222,23 @@ find_busiest_queue(struct sched_group *group, enum cpu_idle_type idle, */ #define MAX_PINNED_INTERVAL 512 +/* Working cpumask for load_balance and load_balance_newidle. */ +static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); + /* * Check this_cpu to ensure it is balanced within domain. Attempt to move * tasks if there is an imbalance. */ static int load_balance(int this_cpu, struct rq *this_rq, struct sched_domain *sd, enum cpu_idle_type idle, - int *balance, struct cpumask *cpus) + int *balance) { int ld_moved, all_pinned = 0, active_balance = 0, sd_idle = 0; struct sched_group *group; unsigned long imbalance; struct rq *busiest; unsigned long flags; + struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); cpumask_setall(cpus); @@ -3615,8 +4393,7 @@ out: * this_rq is locked. */ static int -load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd, - struct cpumask *cpus) +load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd) { struct sched_group *group; struct rq *busiest = NULL; @@ -3624,6 +4401,7 @@ load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd, int ld_moved = 0; int sd_idle = 0; int all_pinned = 0; + struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); cpumask_setall(cpus); @@ -3764,10 +4542,6 @@ static void idle_balance(int this_cpu, struct rq *this_rq) struct sched_domain *sd; int pulled_task = 0; unsigned long next_balance = jiffies + HZ; - cpumask_var_t tmpmask; - - if (!alloc_cpumask_var(&tmpmask, GFP_ATOMIC)) - return; for_each_domain(this_cpu, sd) { unsigned long interval; @@ -3778,7 +4552,7 @@ static void idle_balance(int this_cpu, struct rq *this_rq) if (sd->flags & SD_BALANCE_NEWIDLE) /* If we've pulled tasks over stop searching: */ pulled_task = load_balance_newidle(this_cpu, this_rq, - sd, tmpmask); + sd); interval = msecs_to_jiffies(sd->balance_interval); if (time_after(next_balance, sd->last_balance + interval)) @@ -3793,7 +4567,6 @@ static void idle_balance(int this_cpu, struct rq *this_rq) */ this_rq->next_balance = next_balance; } - free_cpumask_var(tmpmask); } /* @@ -3851,10 +4624,131 @@ static void active_load_balance(struct rq *busiest_rq, int busiest_cpu) static struct { atomic_t load_balancer; cpumask_var_t cpu_mask; + cpumask_var_t ilb_grp_nohz_mask; } nohz ____cacheline_aligned = { .load_balancer = ATOMIC_INIT(-1), }; +int get_nohz_load_balancer(void) +{ + return atomic_read(&nohz.load_balancer); +} + +#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) +/** + * lowest_flag_domain - Return lowest sched_domain containing flag. + * @cpu: The cpu whose lowest level of sched domain is to + * be returned. + * @flag: The flag to check for the lowest sched_domain + * for the given cpu. + * + * Returns the lowest sched_domain of a cpu which contains the given flag. + */ +static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) +{ + struct sched_domain *sd; + + for_each_domain(cpu, sd) + if (sd && (sd->flags & flag)) + break; + + return sd; +} + +/** + * for_each_flag_domain - Iterates over sched_domains containing the flag. + * @cpu: The cpu whose domains we're iterating over. + * @sd: variable holding the value of the power_savings_sd + * for cpu. + * @flag: The flag to filter the sched_domains to be iterated. + * + * Iterates over all the scheduler domains for a given cpu that has the 'flag' + * set, starting from the lowest sched_domain to the highest. + */ +#define for_each_flag_domain(cpu, sd, flag) \ + for (sd = lowest_flag_domain(cpu, flag); \ + (sd && (sd->flags & flag)); sd = sd->parent) + +/** + * is_semi_idle_group - Checks if the given sched_group is semi-idle. + * @ilb_group: group to be checked for semi-idleness + * + * Returns: 1 if the group is semi-idle. 0 otherwise. + * + * We define a sched_group to be semi idle if it has atleast one idle-CPU + * and atleast one non-idle CPU. This helper function checks if the given + * sched_group is semi-idle or not. + */ +static inline int is_semi_idle_group(struct sched_group *ilb_group) +{ + cpumask_and(nohz.ilb_grp_nohz_mask, nohz.cpu_mask, + sched_group_cpus(ilb_group)); + + /* + * A sched_group is semi-idle when it has atleast one busy cpu + * and atleast one idle cpu. + */ + if (cpumask_empty(nohz.ilb_grp_nohz_mask)) + return 0; + + if (cpumask_equal(nohz.ilb_grp_nohz_mask, sched_group_cpus(ilb_group))) + return 0; + + return 1; +} +/** + * find_new_ilb - Finds the optimum idle load balancer for nomination. + * @cpu: The cpu which is nominating a new idle_load_balancer. + * + * Returns: Returns the id of the idle load balancer if it exists, + * Else, returns >= nr_cpu_ids. + * + * This algorithm picks the idle load balancer such that it belongs to a + * semi-idle powersavings sched_domain. The idea is to try and avoid + * completely idle packages/cores just for the purpose of idle load balancing + * when there are other idle cpu's which are better suited for that job. + */ +static int find_new_ilb(int cpu) +{ + struct sched_domain *sd; + struct sched_group *ilb_group; + + /* + * Have idle load balancer selection from semi-idle packages only + * when power-aware load balancing is enabled + */ + if (!(sched_smt_power_savings || sched_mc_power_savings)) + goto out_done; + + /* + * Optimize for the case when we have no idle CPUs or only one + * idle CPU. Don't walk the sched_domain hierarchy in such cases + */ + if (cpumask_weight(nohz.cpu_mask) < 2) + goto out_done; + + for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { + ilb_group = sd->groups; + + do { + if (is_semi_idle_group(ilb_group)) + return cpumask_first(nohz.ilb_grp_nohz_mask); + + ilb_group = ilb_group->next; + + } while (ilb_group != sd->groups); + } + +out_done: + return cpumask_first(nohz.cpu_mask); +} +#else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ +static inline int find_new_ilb(int call_cpu) +{ + return cpumask_first(nohz.cpu_mask); +} +#endif + /* * This routine will try to nominate the ilb (idle load balancing) * owner among the cpus whose ticks are stopped. ilb owner will do the idle @@ -3880,19 +4774,24 @@ int select_nohz_load_balancer(int stop_tick) int cpu = smp_processor_id(); if (stop_tick) { - cpumask_set_cpu(cpu, nohz.cpu_mask); cpu_rq(cpu)->in_nohz_recently = 1; - /* - * If we are going offline and still the leader, give up! - */ - if (!cpu_active(cpu) && - atomic_read(&nohz.load_balancer) == cpu) { + if (!cpu_active(cpu)) { + if (atomic_read(&nohz.load_balancer) != cpu) + return 0; + + /* + * If we are going offline and still the leader, + * give up! + */ if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu) BUG(); + return 0; } + cpumask_set_cpu(cpu, nohz.cpu_mask); + /* time for ilb owner also to sleep */ if (cpumask_weight(nohz.cpu_mask) == num_online_cpus()) { if (atomic_read(&nohz.load_balancer) == cpu) @@ -3904,8 +4803,24 @@ int select_nohz_load_balancer(int stop_tick) /* make me the ilb owner */ if (atomic_cmpxchg(&nohz.load_balancer, -1, cpu) == -1) return 1; - } else if (atomic_read(&nohz.load_balancer) == cpu) + } else if (atomic_read(&nohz.load_balancer) == cpu) { + int new_ilb; + + if (!(sched_smt_power_savings || + sched_mc_power_savings)) + return 1; + /* + * Check to see if there is a more power-efficient + * ilb. + */ + new_ilb = find_new_ilb(cpu); + if (new_ilb < nr_cpu_ids && new_ilb != cpu) { + atomic_set(&nohz.load_balancer, -1); + resched_cpu(new_ilb); + return 0; + } return 1; + } } else { if (!cpumask_test_cpu(cpu, nohz.cpu_mask)) return 0; @@ -3938,11 +4853,6 @@ static void rebalance_domains(int cpu, enum cpu_idle_type idle) unsigned long next_balance = jiffies + 60*HZ; int update_next_balance = 0; int need_serialize; - cpumask_var_t tmp; - - /* Fails alloc? Rebalancing probably not a priority right now. */ - if (!alloc_cpumask_var(&tmp, GFP_ATOMIC)) - return; for_each_domain(cpu, sd) { if (!(sd->flags & SD_LOAD_BALANCE)) @@ -3967,7 +4877,7 @@ static void rebalance_domains(int cpu, enum cpu_idle_type idle) } if (time_after_eq(jiffies, sd->last_balance + interval)) { - if (load_balance(cpu, rq, sd, idle, &balance, tmp)) { + if (load_balance(cpu, rq, sd, idle, &balance)) { /* * We've pulled tasks over so either we're no * longer idle, or one of our SMT siblings is @@ -4001,8 +4911,6 @@ out: */ if (likely(update_next_balance)) rq->next_balance = next_balance; - - free_cpumask_var(tmp); } /* @@ -4052,6 +4960,11 @@ static void run_rebalance_domains(struct softirq_action *h) #endif } +static inline int on_null_domain(int cpu) +{ + return !rcu_dereference(cpu_rq(cpu)->sd); +} + /* * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. * @@ -4076,15 +4989,7 @@ static inline void trigger_load_balance(struct rq *rq, int cpu) } if (atomic_read(&nohz.load_balancer) == -1) { - /* - * simple selection for now: Nominate the - * first cpu in the nohz list to be the next - * ilb owner. - * - * TBD: Traverse the sched domains and nominate - * the nearest cpu in the nohz.cpu_mask. - */ - int ilb = cpumask_first(nohz.cpu_mask); + int ilb = find_new_ilb(cpu); if (ilb < nr_cpu_ids) resched_cpu(ilb); @@ -4109,7 +5014,9 @@ static inline void trigger_load_balance(struct rq *rq, int cpu) cpumask_test_cpu(cpu, nohz.cpu_mask)) return; #endif - if (time_after_eq(jiffies, rq->next_balance)) + /* Don't need to rebalance while attached to NULL domain */ + if (time_after_eq(jiffies, rq->next_balance) && + likely(!on_null_domain(cpu))) raise_softirq(SCHED_SOFTIRQ); } @@ -4129,9 +5036,25 @@ DEFINE_PER_CPU(struct kernel_stat, kstat); EXPORT_PER_CPU_SYMBOL(kstat); /* - * Return any ns on the sched_clock that have not yet been banked in + * Return any ns on the sched_clock that have not yet been accounted in * @p in case that task is currently running. + * + * Called with task_rq_lock() held on @rq. */ +static u64 do_task_delta_exec(struct task_struct *p, struct rq *rq) +{ + u64 ns = 0; + + if (task_current(rq, p)) { + update_rq_clock(rq); + ns = rq->clock - p->se.exec_start; + if ((s64)ns < 0) + ns = 0; + } + + return ns; +} + unsigned long long task_delta_exec(struct task_struct *p) { unsigned long flags; @@ -4139,16 +5062,49 @@ unsigned long long task_delta_exec(struct task_struct *p) u64 ns = 0; rq = task_rq_lock(p, &flags); + ns = do_task_delta_exec(p, rq); + task_rq_unlock(rq, &flags); - if (task_current(rq, p)) { - u64 delta_exec; + return ns; +} - update_rq_clock(rq); - delta_exec = rq->clock - p->se.exec_start; - if ((s64)delta_exec > 0) - ns = delta_exec; - } +/* + * Return accounted runtime for the task. + * In case the task is currently running, return the runtime plus current's + * pending runtime that have not been accounted yet. + */ +unsigned long long task_sched_runtime(struct task_struct *p) +{ + unsigned long flags; + struct rq *rq; + u64 ns = 0; + rq = task_rq_lock(p, &flags); + ns = p->se.sum_exec_runtime + do_task_delta_exec(p, rq); + task_rq_unlock(rq, &flags); + + return ns; +} + +/* + * Return sum_exec_runtime for the thread group. + * In case the task is currently running, return the sum plus current's + * pending runtime that have not been accounted yet. + * + * Note that the thread group might have other running tasks as well, + * so the return value not includes other pending runtime that other + * running tasks might have. + */ +unsigned long long thread_group_sched_runtime(struct task_struct *p) +{ + struct task_cputime totals; + unsigned long flags; + struct rq *rq; + u64 ns; + + rq = task_rq_lock(p, &flags); + thread_group_cputime(p, &totals); + ns = totals.sum_exec_runtime + do_task_delta_exec(p, rq); task_rq_unlock(rq, &flags); return ns; @@ -4177,6 +5133,8 @@ void account_user_time(struct task_struct *p, cputime_t cputime, cpustat->nice = cputime64_add(cpustat->nice, tmp); else cpustat->user = cputime64_add(cpustat->user, tmp); + + cpuacct_update_stats(p, CPUACCT_STAT_USER, cputime); /* Account for user time used */ acct_update_integrals(p); } @@ -4238,6 +5196,8 @@ void account_system_time(struct task_struct *p, int hardirq_offset, else cpustat->system = cputime64_add(cpustat->system, tmp); + cpuacct_update_stats(p, CPUACCT_STAT_SYSTEM, cputime); + /* Account for system time used */ acct_update_integrals(p); } @@ -4285,7 +5245,7 @@ void account_process_tick(struct task_struct *p, int user_tick) if (user_tick) account_user_time(p, one_jiffy, one_jiffy_scaled); - else if (p != rq->idle) + else if ((p != rq->idle) || (irq_count() != HARDIRQ_OFFSET)) account_system_time(p, HARDIRQ_OFFSET, one_jiffy, one_jiffy_scaled); else @@ -4393,16 +5353,15 @@ void scheduler_tick(void) curr->sched_class->task_tick(rq, curr, 0); spin_unlock(&rq->lock); + perf_counter_task_tick(curr, cpu); + #ifdef CONFIG_SMP rq->idle_at_tick = idle_cpu(cpu); trigger_load_balance(rq, cpu); #endif } -#if defined(CONFIG_PREEMPT) && (defined(CONFIG_DEBUG_PREEMPT) || \ - defined(CONFIG_PREEMPT_TRACER)) - -static inline unsigned long get_parent_ip(unsigned long addr) +notrace unsigned long get_parent_ip(unsigned long addr) { if (in_lock_functions(addr)) { addr = CALLER_ADDR2; @@ -4412,6 +5371,9 @@ static inline unsigned long get_parent_ip(unsigned long addr) return addr; } +#if defined(CONFIG_PREEMPT) && (defined(CONFIG_DEBUG_PREEMPT) || \ + defined(CONFIG_PREEMPT_TRACER)) + void __kprobes add_preempt_count(int val) { #ifdef CONFIG_DEBUG_PREEMPT @@ -4503,11 +5465,33 @@ static inline void schedule_debug(struct task_struct *prev) #endif } +static void put_prev_task(struct rq *rq, struct task_struct *prev) +{ + if (prev->state == TASK_RUNNING) { + u64 runtime = prev->se.sum_exec_runtime; + + runtime -= prev->se.prev_sum_exec_runtime; + runtime = min_t(u64, runtime, 2*sysctl_sched_migration_cost); + + /* + * In order to avoid avg_overlap growing stale when we are + * indeed overlapping and hence not getting put to sleep, grow + * the avg_overlap on preemption. + * + * We use the average preemption runtime because that + * correlates to the amount of cache footprint a task can + * build up. + */ + update_avg(&prev->se.avg_overlap, runtime); + } + prev->sched_class->put_prev_task(rq, prev); +} + /* * Pick up the highest-prio task: */ static inline struct task_struct * -pick_next_task(struct rq *rq, struct task_struct *prev) +pick_next_task(struct rq *rq) { const struct sched_class *class; struct task_struct *p; @@ -4549,7 +5533,7 @@ need_resched: preempt_disable(); cpu = smp_processor_id(); rq = cpu_rq(cpu); - rcu_qsctr_inc(cpu); + rcu_sched_qs(cpu); prev = rq->curr; switch_count = &prev->nivcsw; @@ -4573,19 +5557,17 @@ need_resched_nonpreemptible: switch_count = &prev->nvcsw; } -#ifdef CONFIG_SMP - if (prev->sched_class->pre_schedule) - prev->sched_class->pre_schedule(rq, prev); -#endif + pre_schedule(rq, prev); if (unlikely(!rq->nr_running)) idle_balance(cpu, rq); - prev->sched_class->put_prev_task(rq, prev); - next = pick_next_task(rq, prev); + put_prev_task(rq, prev); + next = pick_next_task(rq); if (likely(prev != next)) { sched_info_switch(prev, next); + perf_counter_task_sched_out(prev, next, cpu); rq->nr_switches++; rq->curr = next; @@ -4601,15 +5583,78 @@ need_resched_nonpreemptible: } else spin_unlock_irq(&rq->lock); + post_schedule(rq); + if (unlikely(reacquire_kernel_lock(current) < 0)) goto need_resched_nonpreemptible; preempt_enable_no_resched(); - if (unlikely(test_thread_flag(TIF_NEED_RESCHED))) + if (need_resched()) goto need_resched; } EXPORT_SYMBOL(schedule); +#ifdef CONFIG_SMP +/* + * Look out! "owner" is an entirely speculative pointer + * access and not reliable. + */ +int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner) +{ + unsigned int cpu; + struct rq *rq; + + if (!sched_feat(OWNER_SPIN)) + return 0; + +#ifdef CONFIG_DEBUG_PAGEALLOC + /* + * Need to access the cpu field knowing that + * DEBUG_PAGEALLOC could have unmapped it if + * the mutex owner just released it and exited. + */ + if (probe_kernel_address(&owner->cpu, cpu)) + goto out; +#else + cpu = owner->cpu; +#endif + + /* + * Even if the access succeeded (likely case), + * the cpu field may no longer be valid. + */ + if (cpu >= nr_cpumask_bits) + goto out; + + /* + * We need to validate that we can do a + * get_cpu() and that we have the percpu area. + */ + if (!cpu_online(cpu)) + goto out; + + rq = cpu_rq(cpu); + + for (;;) { + /* + * Owner changed, break to re-assess state. + */ + if (lock->owner != owner) + break; + + /* + * Is that owner really running on that cpu? + */ + if (task_thread_info(rq->curr) != owner || need_resched()) + return 0; + + cpu_relax(); + } +out: + return 1; +} +#endif + #ifdef CONFIG_PREEMPT /* * this is the entry point to schedule() from in-kernel preemption @@ -4637,7 +5682,7 @@ asmlinkage void __sched preempt_schedule(void) * between schedule and now. */ barrier(); - } while (unlikely(test_thread_flag(TIF_NEED_RESCHED))); + } while (need_resched()); } EXPORT_SYMBOL(preempt_schedule); @@ -4666,7 +5711,7 @@ asmlinkage void __sched preempt_schedule_irq(void) * between schedule and now. */ barrier(); - } while (unlikely(test_thread_flag(TIF_NEED_RESCHED))); + } while (need_resched()); } #endif /* CONFIG_PREEMPT */ @@ -4688,7 +5733,7 @@ EXPORT_SYMBOL(default_wake_function); * zero in this (rare) case, and we handle it by continuing to scan the queue. */ static void __wake_up_common(wait_queue_head_t *q, unsigned int mode, - int nr_exclusive, int sync, void *key) + int nr_exclusive, int sync, void *key) { wait_queue_t *curr, *next; @@ -4707,6 +5752,9 @@ static void __wake_up_common(wait_queue_head_t *q, unsigned int mode, * @mode: which threads * @nr_exclusive: how many wake-one or wake-many threads to wake up * @key: is directly passed to the wakeup function + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. */ void __wake_up(wait_queue_head_t *q, unsigned int mode, int nr_exclusive, void *key) @@ -4727,11 +5775,17 @@ void __wake_up_locked(wait_queue_head_t *q, unsigned int mode) __wake_up_common(q, mode, 1, 0, NULL); } +void __wake_up_locked_key(wait_queue_head_t *q, unsigned int mode, void *key) +{ + __wake_up_common(q, mode, 1, 0, key); +} + /** - * __wake_up_sync - wake up threads blocked on a waitqueue. + * __wake_up_sync_key - wake up threads blocked on a waitqueue. * @q: the waitqueue * @mode: which threads * @nr_exclusive: how many wake-one or wake-many threads to wake up + * @key: opaque value to be passed to wakeup targets * * The sync wakeup differs that the waker knows that it will schedule * away soon, so while the target thread will be woken up, it will not @@ -4739,9 +5793,12 @@ void __wake_up_locked(wait_queue_head_t *q, unsigned int mode) * with each other. This can prevent needless bouncing between CPUs. * * On UP it can prevent extra preemption. + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. */ -void -__wake_up_sync(wait_queue_head_t *q, unsigned int mode, int nr_exclusive) +void __wake_up_sync_key(wait_queue_head_t *q, unsigned int mode, + int nr_exclusive, void *key) { unsigned long flags; int sync = 1; @@ -4753,9 +5810,18 @@ __wake_up_sync(wait_queue_head_t *q, unsigned int mode, int nr_exclusive) sync = 0; spin_lock_irqsave(&q->lock, flags); - __wake_up_common(q, mode, nr_exclusive, sync, NULL); + __wake_up_common(q, mode, nr_exclusive, sync, key); spin_unlock_irqrestore(&q->lock, flags); } +EXPORT_SYMBOL_GPL(__wake_up_sync_key); + +/* + * __wake_up_sync - see __wake_up_sync_key() + */ +void __wake_up_sync(wait_queue_head_t *q, unsigned int mode, int nr_exclusive) +{ + __wake_up_sync_key(q, mode, nr_exclusive, NULL); +} EXPORT_SYMBOL_GPL(__wake_up_sync); /* For internal use only */ /** @@ -4766,6 +5832,9 @@ EXPORT_SYMBOL_GPL(__wake_up_sync); /* For internal use only */ * awakened in the same order in which they were queued. * * See also complete_all(), wait_for_completion() and related routines. + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. */ void complete(struct completion *x) { @@ -4783,6 +5852,9 @@ EXPORT_SYMBOL(complete); * @x: holds the state of this particular completion * * This will wake up all threads waiting on this particular completion event. + * + * It may be assumed that this function implies a write memory barrier before + * changing the task state if and only if any tasks are woken up. */ void complete_all(struct completion *x) { @@ -5140,7 +6212,7 @@ SYSCALL_DEFINE1(nice, int, increment) if (increment > 40) increment = 40; - nice = PRIO_TO_NICE(current->static_prio) + increment; + nice = TASK_NICE(current) + increment; if (nice < -20) nice = -20; if (nice > 19) @@ -5258,17 +6330,25 @@ static int __sched_setscheduler(struct task_struct *p, int policy, unsigned long flags; const struct sched_class *prev_class = p->sched_class; struct rq *rq; + int reset_on_fork; /* may grab non-irq protected spin_locks */ BUG_ON(in_interrupt()); recheck: /* double check policy once rq lock held */ - if (policy < 0) + if (policy < 0) { + reset_on_fork = p->sched_reset_on_fork; policy = oldpolicy = p->policy; - else if (policy != SCHED_FIFO && policy != SCHED_RR && - policy != SCHED_NORMAL && policy != SCHED_BATCH && - policy != SCHED_IDLE) - return -EINVAL; + } else { + reset_on_fork = !!(policy & SCHED_RESET_ON_FORK); + policy &= ~SCHED_RESET_ON_FORK; + + if (policy != SCHED_FIFO && policy != SCHED_RR && + policy != SCHED_NORMAL && policy != SCHED_BATCH && + policy != SCHED_IDLE) + return -EINVAL; + } + /* * Valid priorities for SCHED_FIFO and SCHED_RR are * 1..MAX_USER_RT_PRIO-1, valid priority for SCHED_NORMAL, @@ -5312,6 +6392,10 @@ recheck: /* can't change other user's priorities */ if (!check_same_owner(p)) return -EPERM; + + /* Normal users shall not reset the sched_reset_on_fork flag */ + if (p->sched_reset_on_fork && !reset_on_fork) + return -EPERM; } if (user) { @@ -5355,6 +6439,8 @@ recheck: if (running) p->sched_class->put_prev_task(rq, p); + p->sched_reset_on_fork = reset_on_fork; + oldprio = p->prio; __setscheduler(rq, p, policy, param->sched_priority); @@ -5471,14 +6557,15 @@ SYSCALL_DEFINE1(sched_getscheduler, pid_t, pid) if (p) { retval = security_task_getscheduler(p); if (!retval) - retval = p->policy; + retval = p->policy + | (p->sched_reset_on_fork ? SCHED_RESET_ON_FORK : 0); } read_unlock(&tasklist_lock); return retval; } /** - * sys_sched_getscheduler - get the RT priority of a thread + * sys_sched_getparam - get the RT priority of a thread * @pid: the pid in question. * @param: structure containing the RT priority. */ @@ -5699,27 +6786,21 @@ SYSCALL_DEFINE0(sched_yield) return 0; } +static inline int should_resched(void) +{ + return need_resched() && !(preempt_count() & PREEMPT_ACTIVE); +} + static void __cond_resched(void) { -#ifdef CONFIG_DEBUG_SPINLOCK_SLEEP - __might_sleep(__FILE__, __LINE__); -#endif - /* - * The BKS might be reacquired before we have dropped - * PREEMPT_ACTIVE, which could trigger a second - * cond_resched() call. - */ - do { - add_preempt_count(PREEMPT_ACTIVE); - schedule(); - sub_preempt_count(PREEMPT_ACTIVE); - } while (need_resched()); + add_preempt_count(PREEMPT_ACTIVE); + schedule(); + sub_preempt_count(PREEMPT_ACTIVE); } int __sched _cond_resched(void) { - if (need_resched() && !(preempt_count() & PREEMPT_ACTIVE) && - system_state == SYSTEM_RUNNING) { + if (should_resched()) { __cond_resched(); return 1; } @@ -5728,21 +6809,23 @@ int __sched _cond_resched(void) EXPORT_SYMBOL(_cond_resched); /* - * cond_resched_lock() - if a reschedule is pending, drop the given lock, + * __cond_resched_lock() - if a reschedule is pending, drop the given lock, * call schedule, and on return reacquire the lock. * * This works OK both with and without CONFIG_PREEMPT. We do strange low-level * operations here to prevent schedule() from being called twice (once via * spin_unlock(), once by hand). */ -int cond_resched_lock(spinlock_t *lock) +int __cond_resched_lock(spinlock_t *lock) { - int resched = need_resched() && system_state == SYSTEM_RUNNING; + int resched = should_resched(); int ret = 0; + lockdep_assert_held(lock); + if (spin_needbreak(lock) || resched) { spin_unlock(lock); - if (resched && need_resched()) + if (resched) __cond_resched(); else cpu_relax(); @@ -5751,13 +6834,13 @@ int cond_resched_lock(spinlock_t *lock) } return ret; } -EXPORT_SYMBOL(cond_resched_lock); +EXPORT_SYMBOL(__cond_resched_lock); -int __sched cond_resched_softirq(void) +int __sched __cond_resched_softirq(void) { BUG_ON(!in_softirq()); - if (need_resched() && system_state == SYSTEM_RUNNING) { + if (should_resched()) { local_bh_enable(); __cond_resched(); local_bh_disable(); @@ -5765,7 +6848,7 @@ int __sched cond_resched_softirq(void) } return 0; } -EXPORT_SYMBOL(cond_resched_softirq); +EXPORT_SYMBOL(__cond_resched_softirq); /** * yield - yield the current processor to other threads. @@ -5789,11 +6872,13 @@ EXPORT_SYMBOL(yield); */ void __sched io_schedule(void) { - struct rq *rq = &__raw_get_cpu_var(runqueues); + struct rq *rq = raw_rq(); delayacct_blkio_start(); atomic_inc(&rq->nr_iowait); + current->in_iowait = 1; schedule(); + current->in_iowait = 0; atomic_dec(&rq->nr_iowait); delayacct_blkio_end(); } @@ -5801,12 +6886,14 @@ EXPORT_SYMBOL(io_schedule); long __sched io_schedule_timeout(long timeout) { - struct rq *rq = &__raw_get_cpu_var(runqueues); + struct rq *rq = raw_rq(); long ret; delayacct_blkio_start(); atomic_inc(&rq->nr_iowait); + current->in_iowait = 1; ret = schedule_timeout(timeout); + current->in_iowait = 0; atomic_dec(&rq->nr_iowait); delayacct_blkio_end(); return ret; @@ -5939,15 +7026,11 @@ void sched_show_task(struct task_struct *p) printk(KERN_CONT " %016lx ", thread_saved_pc(p)); #endif #ifdef CONFIG_DEBUG_STACK_USAGE - { - unsigned long *n = end_of_stack(p); - while (!*n) - n++; - free = (unsigned long)n - (unsigned long)end_of_stack(p); - } + free = stack_not_used(p); #endif - printk(KERN_CONT "%5lu %5d %6d\n", free, - task_pid_nr(p), task_pid_nr(p->real_parent)); + printk(KERN_CONT "%5lu %5d %6d 0x%08lx\n", free, + task_pid_nr(p), task_pid_nr(p->real_parent), + (unsigned long)task_thread_info(p)->flags); show_stack(p, NULL); } @@ -6127,8 +7210,12 @@ int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask) if (migrate_task(p, cpumask_any_and(cpu_online_mask, new_mask), &req)) { /* Need help from migration thread: drop lock and wait. */ + struct task_struct *mt = rq->migration_thread; + + get_task_struct(mt); task_rq_unlock(rq, &flags); wake_up_process(rq->migration_thread); + put_task_struct(mt); wait_for_completion(&req.done); tlb_migrate_finish(p->mm); return 0; @@ -6186,6 +7273,11 @@ fail: return ret; } +#define RCU_MIGRATION_IDLE 0 +#define RCU_MIGRATION_NEED_QS 1 +#define RCU_MIGRATION_GOT_QS 2 +#define RCU_MIGRATION_MUST_SYNC 3 + /* * migration_thread - this is a highprio system thread that performs * thread migration by bumping thread off CPU then 'pushing' onto @@ -6193,6 +7285,7 @@ fail: */ static int migration_thread(void *data) { + int badcpu; int cpu = (long)data; struct rq *rq; @@ -6208,7 +7301,7 @@ static int migration_thread(void *data) if (cpu_is_offline(cpu)) { spin_unlock_irq(&rq->lock); - goto wait_to_die; + break; } if (rq->active_balance) { @@ -6227,23 +7320,23 @@ static int migration_thread(void *data) req = list_entry(head->next, struct migration_req, list); list_del_init(head->next); - spin_unlock(&rq->lock); - __migrate_task(req->task, cpu, req->dest_cpu); + if (req->task != NULL) { + spin_unlock(&rq->lock); + __migrate_task(req->task, cpu, req->dest_cpu); + } else if (likely(cpu == (badcpu = smp_processor_id()))) { + req->dest_cpu = RCU_MIGRATION_GOT_QS; + spin_unlock(&rq->lock); + } else { + req->dest_cpu = RCU_MIGRATION_MUST_SYNC; + spin_unlock(&rq->lock); + WARN_ONCE(1, "migration_thread() on CPU %d, expected %d\n", badcpu, cpu); + } local_irq_enable(); complete(&req->done); } __set_current_state(TASK_RUNNING); - return 0; -wait_to_die: - /* Wait for kthread_stop */ - set_current_state(TASK_INTERRUPTIBLE); - while (!kthread_should_stop()) { - schedule(); - set_current_state(TASK_INTERRUPTIBLE); - } - __set_current_state(TASK_RUNNING); return 0; } @@ -6418,7 +7511,7 @@ static void migrate_dead_tasks(unsigned int dead_cpu) if (!rq->nr_running) break; update_rq_clock(rq); - next = pick_next_task(rq, rq->curr); + next = pick_next_task(rq); if (!next) break; next->sched_class->put_prev_task(rq, next); @@ -6426,6 +7519,15 @@ static void migrate_dead_tasks(unsigned int dead_cpu) } } + +/* + * remove the tasks which were accounted by rq from calc_load_tasks. + */ +static void calc_global_load_remove(struct rq *rq) +{ + atomic_long_sub(rq->calc_load_active, &calc_load_tasks); + rq->calc_load_active = 0; +} #endif /* CONFIG_HOTPLUG_CPU */ #if defined(CONFIG_SCHED_DEBUG) && defined(CONFIG_SYSCTL) @@ -6649,7 +7751,9 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) rq = task_rq_lock(p, &flags); __setscheduler(rq, p, SCHED_FIFO, MAX_RT_PRIO-1); task_rq_unlock(rq, &flags); + get_task_struct(p); cpu_rq(cpu)->migration_thread = p; + rq->calc_load_update = calc_load_update; break; case CPU_ONLINE: @@ -6677,6 +7781,7 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) kthread_bind(cpu_rq(cpu)->migration_thread, cpumask_any(cpu_online_mask)); kthread_stop(cpu_rq(cpu)->migration_thread); + put_task_struct(cpu_rq(cpu)->migration_thread); cpu_rq(cpu)->migration_thread = NULL; break; @@ -6686,6 +7791,7 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) migrate_live_tasks(cpu); rq = cpu_rq(cpu); kthread_stop(rq->migration_thread); + put_task_struct(rq->migration_thread); rq->migration_thread = NULL; /* Idle task back to normal (off runqueue, low prio) */ spin_lock_irq(&rq->lock); @@ -6699,7 +7805,7 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) cpuset_unlock(); migrate_nr_uninterruptible(rq); BUG_ON(rq->nr_running != 0); - + calc_global_load_remove(rq); /* * No need to migrate the tasks: it was best-effort if * they didn't take sched_hotcpu_mutex. Just wake up @@ -6735,8 +7841,10 @@ migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu) return NOTIFY_OK; } -/* Register at highest priority so that task migration (migrate_all_tasks) - * happens before everything else. +/* + * Register at high priority so that task migration (migrate_all_tasks) + * happens before everything else. This has to be lower priority than + * the notifier in the perf_counter subsystem, though. */ static struct notifier_block __cpuinitdata migration_notifier = { .notifier_call = migration_call, @@ -6754,7 +7862,7 @@ static int __init migration_init(void) migration_call(&migration_notifier, CPU_ONLINE, cpu); register_cpu_notifier(&migration_notifier); - return err; + return 0; } early_initcall(migration_init); #endif @@ -6801,7 +7909,7 @@ static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level, break; } - if (!group->__cpu_power) { + if (!group->cpu_power) { printk(KERN_CONT "\n"); printk(KERN_ERR "ERROR: domain->cpu_power not " "set\n"); @@ -6823,7 +7931,12 @@ static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level, cpumask_or(groupmask, groupmask, sched_group_cpus(group)); cpulist_scnprintf(str, sizeof(str), sched_group_cpus(group)); + printk(KERN_CONT " %s", str); + if (group->cpu_power != SCHED_LOAD_SCALE) { + printk(KERN_CONT " (cpu_power = %d)", + group->cpu_power); + } group = group->next; } while (group != sd->groups); @@ -6939,52 +8052,58 @@ static void free_rootdomain(struct root_domain *rd) static void rq_attach_root(struct rq *rq, struct root_domain *rd) { + struct root_domain *old_rd = NULL; unsigned long flags; spin_lock_irqsave(&rq->lock, flags); if (rq->rd) { - struct root_domain *old_rd = rq->rd; + old_rd = rq->rd; if (cpumask_test_cpu(rq->cpu, old_rd->online)) set_rq_offline(rq); cpumask_clear_cpu(rq->cpu, old_rd->span); - if (atomic_dec_and_test(&old_rd->refcount)) - free_rootdomain(old_rd); + /* + * If we dont want to free the old_rt yet then + * set old_rd to NULL to skip the freeing later + * in this function: + */ + if (!atomic_dec_and_test(&old_rd->refcount)) + old_rd = NULL; } atomic_inc(&rd->refcount); rq->rd = rd; cpumask_set_cpu(rq->cpu, rd->span); - if (cpumask_test_cpu(rq->cpu, cpu_online_mask)) + if (cpumask_test_cpu(rq->cpu, cpu_active_mask)) set_rq_online(rq); spin_unlock_irqrestore(&rq->lock, flags); + + if (old_rd) + free_rootdomain(old_rd); } -static int __init_refok init_rootdomain(struct root_domain *rd, bool bootmem) +static int init_rootdomain(struct root_domain *rd, bool bootmem) { + gfp_t gfp = GFP_KERNEL; + memset(rd, 0, sizeof(*rd)); - if (bootmem) { - alloc_bootmem_cpumask_var(&def_root_domain.span); - alloc_bootmem_cpumask_var(&def_root_domain.online); - alloc_bootmem_cpumask_var(&def_root_domain.rto_mask); - cpupri_init(&rd->cpupri, true); - return 0; - } + if (bootmem) + gfp = GFP_NOWAIT; - if (!alloc_cpumask_var(&rd->span, GFP_KERNEL)) + if (!alloc_cpumask_var(&rd->span, gfp)) goto out; - if (!alloc_cpumask_var(&rd->online, GFP_KERNEL)) + if (!alloc_cpumask_var(&rd->online, gfp)) goto free_span; - if (!alloc_cpumask_var(&rd->rto_mask, GFP_KERNEL)) + if (!alloc_cpumask_var(&rd->rto_mask, gfp)) goto free_online; - if (cpupri_init(&rd->cpupri, false) != 0) + if (cpupri_init(&rd->cpupri, bootmem) != 0) goto free_rto_mask; return 0; @@ -7101,7 +8220,7 @@ init_sched_build_groups(const struct cpumask *span, continue; cpumask_clear(sched_group_cpus(sg)); - sg->__cpu_power = 0; + sg->cpu_power = 0; for_each_cpu(j, span) { if (group_fn(j, cpu_map, NULL, tmpmask) != group) @@ -7195,8 +8314,9 @@ int sched_smt_power_savings = 0, sched_mc_power_savings = 0; /* * The cpus mask in sched_group and sched_domain hangs off the end. - * FIXME: use cpumask_var_t or dynamic percpu alloc to avoid wasting space - * for nr_cpu_ids < CONFIG_NR_CPUS. + * + * ( See the the comments in include/linux/sched.h:struct sched_group + * and struct sched_domain. ) */ struct static_sched_group { struct sched_group sg; @@ -7208,6 +8328,39 @@ struct static_sched_domain { DECLARE_BITMAP(span, CONFIG_NR_CPUS); }; +struct s_data { +#ifdef CONFIG_NUMA + int sd_allnodes; + cpumask_var_t domainspan; + cpumask_var_t covered; + cpumask_var_t notcovered; +#endif + cpumask_var_t nodemask; + cpumask_var_t this_sibling_map; + cpumask_var_t this_core_map; + cpumask_var_t send_covered; + cpumask_var_t tmpmask; + struct sched_group **sched_group_nodes; + struct root_domain *rd; +}; + +enum s_alloc { + sa_sched_groups = 0, + sa_rootdomain, + sa_tmpmask, + sa_send_covered, + sa_this_core_map, + sa_this_sibling_map, + sa_nodemask, + sa_sched_group_nodes, +#ifdef CONFIG_NUMA + sa_notcovered, + sa_covered, + sa_domainspan, +#endif + sa_none, +}; + /* * SMT sched-domains: */ @@ -7240,7 +8393,7 @@ cpu_to_core_group(int cpu, const struct cpumask *cpu_map, { int group; - cpumask_and(mask, &per_cpu(cpu_sibling_map, cpu), cpu_map); + cpumask_and(mask, topology_thread_cpumask(cpu), cpu_map); group = cpumask_first(mask); if (sg) *sg = &per_cpu(sched_group_core, group).sg; @@ -7269,7 +8422,7 @@ cpu_to_phys_group(int cpu, const struct cpumask *cpu_map, cpumask_and(mask, cpu_coregroup_mask(cpu), cpu_map); group = cpumask_first(mask); #elif defined(CONFIG_SCHED_SMT) - cpumask_and(mask, &per_cpu(cpu_sibling_map, cpu), cpu_map); + cpumask_and(mask, topology_thread_cpumask(cpu), cpu_map); group = cpumask_first(mask); #else group = cpu; @@ -7317,7 +8470,7 @@ static void init_numa_sched_groups_power(struct sched_group *group_head) struct sched_domain *sd; sd = &per_cpu(phys_domains, j).sd; - if (j != cpumask_first(sched_group_cpus(sd->groups))) { + if (j != group_first_cpu(sd->groups)) { /* * Only add "power" once for each * physical package. @@ -7325,11 +8478,76 @@ static void init_numa_sched_groups_power(struct sched_group *group_head) continue; } - sg_inc_cpu_power(sg, sd->groups->__cpu_power); + sg->cpu_power += sd->groups->cpu_power; } sg = sg->next; } while (sg != group_head); } + +static int build_numa_sched_groups(struct s_data *d, + const struct cpumask *cpu_map, int num) +{ + struct sched_domain *sd; + struct sched_group *sg, *prev; + int n, j; + + cpumask_clear(d->covered); + cpumask_and(d->nodemask, cpumask_of_node(num), cpu_map); + if (cpumask_empty(d->nodemask)) { + d->sched_group_nodes[num] = NULL; + goto out; + } + + sched_domain_node_span(num, d->domainspan); + cpumask_and(d->domainspan, d->domainspan, cpu_map); + + sg = kmalloc_node(sizeof(struct sched_group) + cpumask_size(), + GFP_KERNEL, num); + if (!sg) { + printk(KERN_WARNING "Can not alloc domain group for node %d\n", + num); + return -ENOMEM; + } + d->sched_group_nodes[num] = sg; + + for_each_cpu(j, d->nodemask) { + sd = &per_cpu(node_domains, j).sd; + sd->groups = sg; + } + + sg->cpu_power = 0; + cpumask_copy(sched_group_cpus(sg), d->nodemask); + sg->next = sg; + cpumask_or(d->covered, d->covered, d->nodemask); + + prev = sg; + for (j = 0; j < nr_node_ids; j++) { + n = (num + j) % nr_node_ids; + cpumask_complement(d->notcovered, d->covered); + cpumask_and(d->tmpmask, d->notcovered, cpu_map); + cpumask_and(d->tmpmask, d->tmpmask, d->domainspan); + if (cpumask_empty(d->tmpmask)) + break; + cpumask_and(d->tmpmask, d->tmpmask, cpumask_of_node(n)); + if (cpumask_empty(d->tmpmask)) + continue; + sg = kmalloc_node(sizeof(struct sched_group) + cpumask_size(), + GFP_KERNEL, num); + if (!sg) { + printk(KERN_WARNING + "Can not alloc domain group for node %d\n", j); + return -ENOMEM; + } + sg->cpu_power = 0; + cpumask_copy(sched_group_cpus(sg), d->tmpmask); + sg->next = prev->next; + cpumask_or(d->covered, d->covered, d->tmpmask); + prev->next = sg; + prev = sg; + } +out: + return 0; +} #endif /* CONFIG_NUMA */ #ifdef CONFIG_NUMA @@ -7383,45 +8601,47 @@ static void free_sched_groups(const struct cpumask *cpu_map, * there are asymmetries in the topology. If there are asymmetries, group * having more cpu_power will pickup more load compared to the group having * less cpu_power. - * - * cpu_power will be a multiple of SCHED_LOAD_SCALE. This multiple represents - * the maximum number of tasks a group can handle in the presence of other idle - * or lightly loaded groups in the same sched domain. */ static void init_sched_groups_power(int cpu, struct sched_domain *sd) { struct sched_domain *child; struct sched_group *group; + long power; + int weight; WARN_ON(!sd || !sd->groups); - if (cpu != cpumask_first(sched_group_cpus(sd->groups))) + if (cpu != group_first_cpu(sd->groups)) return; child = sd->child; - sd->groups->__cpu_power = 0; + sd->groups->cpu_power = 0; - /* - * For perf policy, if the groups in child domain share resources - * (for example cores sharing some portions of the cache hierarchy - * or SMT), then set this domain groups cpu_power such that each group - * can handle only one task, when there are other idle groups in the - * same sched domain. - */ - if (!child || (!(sd->flags & SD_POWERSAVINGS_BALANCE) && - (child->flags & - (SD_SHARE_CPUPOWER | SD_SHARE_PKG_RESOURCES)))) { - sg_inc_cpu_power(sd->groups, SCHED_LOAD_SCALE); + if (!child) { + power = SCHED_LOAD_SCALE; + weight = cpumask_weight(sched_domain_span(sd)); + /* + * SMT siblings share the power of a single core. + * Usually multiple threads get a better yield out of + * that one core than a single thread would have, + * reflect that in sd->smt_gain. + */ + if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { + power *= sd->smt_gain; + power /= weight; + power >>= SCHED_LOAD_SHIFT; + } + sd->groups->cpu_power += power; return; } /* - * add cpu_power of each child group to this groups cpu_power + * Add cpu_power of each child group to this groups cpu_power. */ group = child->groups; do { - sg_inc_cpu_power(sd->groups, group->__cpu_power); + sd->groups->cpu_power += group->cpu_power; group = group->next; } while (group != child->groups); } @@ -7495,280 +8715,285 @@ static void set_domain_attribute(struct sched_domain *sd, } } -/* - * Build sched domains for a given set of cpus and attach the sched domains - * to the individual cpus - */ -static int __build_sched_domains(const struct cpumask *cpu_map, - struct sched_domain_attr *attr) -{ - int i, err = -ENOMEM; - struct root_domain *rd; - cpumask_var_t nodemask, this_sibling_map, this_core_map, send_covered, - tmpmask; +static void __free_domain_allocs(struct s_data *d, enum s_alloc what, + const struct cpumask *cpu_map) +{ + switch (what) { + case sa_sched_groups: + free_sched_groups(cpu_map, d->tmpmask); /* fall through */ + d->sched_group_nodes = NULL; + case sa_rootdomain: + free_rootdomain(d->rd); /* fall through */ + case sa_tmpmask: + free_cpumask_var(d->tmpmask); /* fall through */ + case sa_send_covered: + free_cpumask_var(d->send_covered); /* fall through */ + case sa_this_core_map: + free_cpumask_var(d->this_core_map); /* fall through */ + case sa_this_sibling_map: + free_cpumask_var(d->this_sibling_map); /* fall through */ + case sa_nodemask: + free_cpumask_var(d->nodemask); /* fall through */ + case sa_sched_group_nodes: #ifdef CONFIG_NUMA - cpumask_var_t domainspan, covered, notcovered; - struct sched_group **sched_group_nodes = NULL; - int sd_allnodes = 0; - - if (!alloc_cpumask_var(&domainspan, GFP_KERNEL)) - goto out; - if (!alloc_cpumask_var(&covered, GFP_KERNEL)) - goto free_domainspan; - if (!alloc_cpumask_var(¬covered, GFP_KERNEL)) - goto free_covered; -#endif - - if (!alloc_cpumask_var(&nodemask, GFP_KERNEL)) - goto free_notcovered; - if (!alloc_cpumask_var(&this_sibling_map, GFP_KERNEL)) - goto free_nodemask; - if (!alloc_cpumask_var(&this_core_map, GFP_KERNEL)) - goto free_this_sibling_map; - if (!alloc_cpumask_var(&send_covered, GFP_KERNEL)) - goto free_this_core_map; - if (!alloc_cpumask_var(&tmpmask, GFP_KERNEL)) - goto free_send_covered; + kfree(d->sched_group_nodes); /* fall through */ + case sa_notcovered: + free_cpumask_var(d->notcovered); /* fall through */ + case sa_covered: + free_cpumask_var(d->covered); /* fall through */ + case sa_domainspan: + free_cpumask_var(d->domainspan); /* fall through */ +#endif + case sa_none: + break; + } +} +static enum s_alloc __visit_domain_allocation_hell(struct s_data *d, + const struct cpumask *cpu_map) +{ #ifdef CONFIG_NUMA - /* - * Allocate the per-node list of sched groups - */ - sched_group_nodes = kcalloc(nr_node_ids, sizeof(struct sched_group *), - GFP_KERNEL); - if (!sched_group_nodes) { + if (!alloc_cpumask_var(&d->domainspan, GFP_KERNEL)) + return sa_none; + if (!alloc_cpumask_var(&d->covered, GFP_KERNEL)) + return sa_domainspan; + if (!alloc_cpumask_var(&d->notcovered, GFP_KERNEL)) + return sa_covered; + /* Allocate the per-node list of sched groups */ + d->sched_group_nodes = kcalloc(nr_node_ids, + sizeof(struct sched_group *), GFP_KERNEL); + if (!d->sched_group_nodes) { printk(KERN_WARNING "Can not alloc sched group node list\n"); - goto free_tmpmask; - } -#endif - - rd = alloc_rootdomain(); - if (!rd) { + return sa_notcovered; + } + sched_group_nodes_bycpu[cpumask_first(cpu_map)] = d->sched_group_nodes; +#endif + if (!alloc_cpumask_var(&d->nodemask, GFP_KERNEL)) + return sa_sched_group_nodes; + if (!alloc_cpumask_var(&d->this_sibling_map, GFP_KERNEL)) + return sa_nodemask; + if (!alloc_cpumask_var(&d->this_core_map, GFP_KERNEL)) + return sa_this_sibling_map; + if (!alloc_cpumask_var(&d->send_covered, GFP_KERNEL)) + return sa_this_core_map; + if (!alloc_cpumask_var(&d->tmpmask, GFP_KERNEL)) + return sa_send_covered; + d->rd = alloc_rootdomain(); + if (!d->rd) { printk(KERN_WARNING "Cannot alloc root domain\n"); - goto free_sched_groups; + return sa_tmpmask; } + return sa_rootdomain; +} +static struct sched_domain *__build_numa_sched_domains(struct s_data *d, + const struct cpumask *cpu_map, struct sched_domain_attr *attr, int i) +{ + struct sched_domain *sd = NULL; #ifdef CONFIG_NUMA - sched_group_nodes_bycpu[cpumask_first(cpu_map)] = sched_group_nodes; -#endif - - /* - * Set up domains for cpus specified by the cpu_map. - */ - for_each_cpu(i, cpu_map) { - struct sched_domain *sd = NULL, *p; - - cpumask_and(nodemask, cpumask_of_node(cpu_to_node(i)), cpu_map); - -#ifdef CONFIG_NUMA - if (cpumask_weight(cpu_map) > - SD_NODES_PER_DOMAIN*cpumask_weight(nodemask)) { - sd = &per_cpu(allnodes_domains, i).sd; - SD_INIT(sd, ALLNODES); - set_domain_attribute(sd, attr); - cpumask_copy(sched_domain_span(sd), cpu_map); - cpu_to_allnodes_group(i, cpu_map, &sd->groups, tmpmask); - p = sd; - sd_allnodes = 1; - } else - p = NULL; + struct sched_domain *parent; - sd = &per_cpu(node_domains, i).sd; - SD_INIT(sd, NODE); + d->sd_allnodes = 0; + if (cpumask_weight(cpu_map) > + SD_NODES_PER_DOMAIN * cpumask_weight(d->nodemask)) { + sd = &per_cpu(allnodes_domains, i).sd; + SD_INIT(sd, ALLNODES); set_domain_attribute(sd, attr); - sched_domain_node_span(cpu_to_node(i), sched_domain_span(sd)); - sd->parent = p; - if (p) - p->child = sd; - cpumask_and(sched_domain_span(sd), - sched_domain_span(sd), cpu_map); + cpumask_copy(sched_domain_span(sd), cpu_map); + cpu_to_allnodes_group(i, cpu_map, &sd->groups, d->tmpmask); + d->sd_allnodes = 1; + } + parent = sd; + + sd = &per_cpu(node_domains, i).sd; + SD_INIT(sd, NODE); + set_domain_attribute(sd, attr); + sched_domain_node_span(cpu_to_node(i), sched_domain_span(sd)); + sd->parent = parent; + if (parent) + parent->child = sd; + cpumask_and(sched_domain_span(sd), sched_domain_span(sd), cpu_map); #endif + return sd; +} - p = sd; - sd = &per_cpu(phys_domains, i).sd; - SD_INIT(sd, CPU); - set_domain_attribute(sd, attr); - cpumask_copy(sched_domain_span(sd), nodemask); - sd->parent = p; - if (p) - p->child = sd; - cpu_to_phys_group(i, cpu_map, &sd->groups, tmpmask); +static struct sched_domain *__build_cpu_sched_domain(struct s_data *d, + const struct cpumask *cpu_map, struct sched_domain_attr *attr, + struct sched_domain *parent, int i) +{ + struct sched_domain *sd; + sd = &per_cpu(phys_domains, i).sd; + SD_INIT(sd, CPU); + set_domain_attribute(sd, attr); + cpumask_copy(sched_domain_span(sd), d->nodemask); + sd->parent = parent; + if (parent) + parent->child = sd; + cpu_to_phys_group(i, cpu_map, &sd->groups, d->tmpmask); + return sd; +} +static struct sched_domain *__build_mc_sched_domain(struct s_data *d, + const struct cpumask *cpu_map, struct sched_domain_attr *attr, + struct sched_domain *parent, int i) +{ + struct sched_domain *sd = parent; #ifdef CONFIG_SCHED_MC - p = sd; - sd = &per_cpu(core_domains, i).sd; - SD_INIT(sd, MC); - set_domain_attribute(sd, attr); - cpumask_and(sched_domain_span(sd), cpu_map, - cpu_coregroup_mask(i)); - sd->parent = p; - p->child = sd; - cpu_to_core_group(i, cpu_map, &sd->groups, tmpmask); + sd = &per_cpu(core_domains, i).sd; + SD_INIT(sd, MC); + set_domain_attribute(sd, attr); + cpumask_and(sched_domain_span(sd), cpu_map, cpu_coregroup_mask(i)); + sd->parent = parent; + parent->child = sd; + cpu_to_core_group(i, cpu_map, &sd->groups, d->tmpmask); #endif + return sd; +} +static struct sched_domain *__build_smt_sched_domain(struct s_data *d, + const struct cpumask *cpu_map, struct sched_domain_attr *attr, + struct sched_domain *parent, int i) +{ + struct sched_domain *sd = parent; #ifdef CONFIG_SCHED_SMT - p = sd; - sd = &per_cpu(cpu_domains, i).sd; - SD_INIT(sd, SIBLING); - set_domain_attribute(sd, attr); - cpumask_and(sched_domain_span(sd), - &per_cpu(cpu_sibling_map, i), cpu_map); - sd->parent = p; - p->child = sd; - cpu_to_cpu_group(i, cpu_map, &sd->groups, tmpmask); + sd = &per_cpu(cpu_domains, i).sd; + SD_INIT(sd, SIBLING); + set_domain_attribute(sd, attr); + cpumask_and(sched_domain_span(sd), cpu_map, topology_thread_cpumask(i)); + sd->parent = parent; + parent->child = sd; + cpu_to_cpu_group(i, cpu_map, &sd->groups, d->tmpmask); #endif - } + return sd; +} +static void build_sched_groups(struct s_data *d, enum sched_domain_level l, + const struct cpumask *cpu_map, int cpu) +{ + switch (l) { #ifdef CONFIG_SCHED_SMT - /* Set up CPU (sibling) groups */ - for_each_cpu(i, cpu_map) { - cpumask_and(this_sibling_map, - &per_cpu(cpu_sibling_map, i), cpu_map); - if (i != cpumask_first(this_sibling_map)) - continue; - - init_sched_build_groups(this_sibling_map, cpu_map, - &cpu_to_cpu_group, - send_covered, tmpmask); - } + case SD_LV_SIBLING: /* set up CPU (sibling) groups */ + cpumask_and(d->this_sibling_map, cpu_map, + topology_thread_cpumask(cpu)); + if (cpu == cpumask_first(d->this_sibling_map)) + init_sched_build_groups(d->this_sibling_map, cpu_map, + &cpu_to_cpu_group, + d->send_covered, d->tmpmask); + break; #endif - #ifdef CONFIG_SCHED_MC - /* Set up multi-core groups */ - for_each_cpu(i, cpu_map) { - cpumask_and(this_core_map, cpu_coregroup_mask(i), cpu_map); - if (i != cpumask_first(this_core_map)) - continue; - - init_sched_build_groups(this_core_map, cpu_map, - &cpu_to_core_group, - send_covered, tmpmask); - } + case SD_LV_MC: /* set up multi-core groups */ + cpumask_and(d->this_core_map, cpu_map, cpu_coregroup_mask(cpu)); + if (cpu == cpumask_first(d->this_core_map)) + init_sched_build_groups(d->this_core_map, cpu_map, + &cpu_to_core_group, + d->send_covered, d->tmpmask); + break; #endif - - /* Set up physical groups */ - for (i = 0; i < nr_node_ids; i++) { - cpumask_and(nodemask, cpumask_of_node(i), cpu_map); - if (cpumask_empty(nodemask)) - continue; - - init_sched_build_groups(nodemask, cpu_map, - &cpu_to_phys_group, - send_covered, tmpmask); - } - + case SD_LV_CPU: /* set up physical groups */ + cpumask_and(d->nodemask, cpumask_of_node(cpu), cpu_map); + if (!cpumask_empty(d->nodemask)) + init_sched_build_groups(d->nodemask, cpu_map, + &cpu_to_phys_group, + d->send_covered, d->tmpmask); + break; #ifdef CONFIG_NUMA - /* Set up node groups */ - if (sd_allnodes) { - init_sched_build_groups(cpu_map, cpu_map, - &cpu_to_allnodes_group, - send_covered, tmpmask); + case SD_LV_ALLNODES: + init_sched_build_groups(cpu_map, cpu_map, &cpu_to_allnodes_group, + d->send_covered, d->tmpmask); + break; +#endif + default: + break; } +} - for (i = 0; i < nr_node_ids; i++) { - /* Set up node groups */ - struct sched_group *sg, *prev; - int j; - - cpumask_clear(covered); - cpumask_and(nodemask, cpumask_of_node(i), cpu_map); - if (cpumask_empty(nodemask)) { - sched_group_nodes[i] = NULL; - continue; - } +/* + * Build sched domains for a given set of cpus and attach the sched domains + * to the individual cpus + */ +static int __build_sched_domains(const struct cpumask *cpu_map, + struct sched_domain_attr *attr) +{ + enum s_alloc alloc_state = sa_none; + struct s_data d; + struct sched_domain *sd; + int i; +#ifdef CONFIG_NUMA + d.sd_allnodes = 0; +#endif - sched_domain_node_span(i, domainspan); - cpumask_and(domainspan, domainspan, cpu_map); + alloc_state = __visit_domain_allocation_hell(&d, cpu_map); + if (alloc_state != sa_rootdomain) + goto error; + alloc_state = sa_sched_groups; - sg = kmalloc_node(sizeof(struct sched_group) + cpumask_size(), - GFP_KERNEL, i); - if (!sg) { - printk(KERN_WARNING "Can not alloc domain group for " - "node %d\n", i); - goto error; - } - sched_group_nodes[i] = sg; - for_each_cpu(j, nodemask) { - struct sched_domain *sd; + /* + * Set up domains for cpus specified by the cpu_map. + */ + for_each_cpu(i, cpu_map) { + cpumask_and(d.nodemask, cpumask_of_node(cpu_to_node(i)), + cpu_map); - sd = &per_cpu(node_domains, j).sd; - sd->groups = sg; - } - sg->__cpu_power = 0; - cpumask_copy(sched_group_cpus(sg), nodemask); - sg->next = sg; - cpumask_or(covered, covered, nodemask); - prev = sg; + sd = __build_numa_sched_domains(&d, cpu_map, attr, i); + sd = __build_cpu_sched_domain(&d, cpu_map, attr, sd, i); + sd = __build_mc_sched_domain(&d, cpu_map, attr, sd, i); + sd = __build_smt_sched_domain(&d, cpu_map, attr, sd, i); + } - for (j = 0; j < nr_node_ids; j++) { - int n = (i + j) % nr_node_ids; + for_each_cpu(i, cpu_map) { + build_sched_groups(&d, SD_LV_SIBLING, cpu_map, i); + build_sched_groups(&d, SD_LV_MC, cpu_map, i); + } - cpumask_complement(notcovered, covered); - cpumask_and(tmpmask, notcovered, cpu_map); - cpumask_and(tmpmask, tmpmask, domainspan); - if (cpumask_empty(tmpmask)) - break; + /* Set up physical groups */ + for (i = 0; i < nr_node_ids; i++) + build_sched_groups(&d, SD_LV_CPU, cpu_map, i); - cpumask_and(tmpmask, tmpmask, cpumask_of_node(n)); - if (cpumask_empty(tmpmask)) - continue; +#ifdef CONFIG_NUMA + /* Set up node groups */ + if (d.sd_allnodes) + build_sched_groups(&d, SD_LV_ALLNODES, cpu_map, 0); - sg = kmalloc_node(sizeof(struct sched_group) + - cpumask_size(), - GFP_KERNEL, i); - if (!sg) { - printk(KERN_WARNING - "Can not alloc domain group for node %d\n", j); - goto error; - } - sg->__cpu_power = 0; - cpumask_copy(sched_group_cpus(sg), tmpmask); - sg->next = prev->next; - cpumask_or(covered, covered, tmpmask); - prev->next = sg; - prev = sg; - } - } + for (i = 0; i < nr_node_ids; i++) + if (build_numa_sched_groups(&d, cpu_map, i)) + goto error; #endif /* Calculate CPU power for physical packages and nodes */ #ifdef CONFIG_SCHED_SMT for_each_cpu(i, cpu_map) { - struct sched_domain *sd = &per_cpu(cpu_domains, i).sd; - + sd = &per_cpu(cpu_domains, i).sd; init_sched_groups_power(i, sd); } #endif #ifdef CONFIG_SCHED_MC for_each_cpu(i, cpu_map) { - struct sched_domain *sd = &per_cpu(core_domains, i).sd; - + sd = &per_cpu(core_domains, i).sd; init_sched_groups_power(i, sd); } #endif for_each_cpu(i, cpu_map) { - struct sched_domain *sd = &per_cpu(phys_domains, i).sd; - + sd = &per_cpu(phys_domains, i).sd; init_sched_groups_power(i, sd); } #ifdef CONFIG_NUMA for (i = 0; i < nr_node_ids; i++) - init_numa_sched_groups_power(sched_group_nodes[i]); + init_numa_sched_groups_power(d.sched_group_nodes[i]); - if (sd_allnodes) { + if (d.sd_allnodes) { struct sched_group *sg; cpu_to_allnodes_group(cpumask_first(cpu_map), cpu_map, &sg, - tmpmask); + d.tmpmask); init_numa_sched_groups_power(sg); } #endif /* Attach the domains */ for_each_cpu(i, cpu_map) { - struct sched_domain *sd; #ifdef CONFIG_SCHED_SMT sd = &per_cpu(cpu_domains, i).sd; #elif defined(CONFIG_SCHED_MC) @@ -7776,44 +9001,16 @@ static int __build_sched_domains(const struct cpumask *cpu_map, #else sd = &per_cpu(phys_domains, i).sd; #endif - cpu_attach_domain(sd, rd, i); + cpu_attach_domain(sd, d.rd, i); } - err = 0; - -free_tmpmask: - free_cpumask_var(tmpmask); -free_send_covered: - free_cpumask_var(send_covered); -free_this_core_map: - free_cpumask_var(this_core_map); -free_this_sibling_map: - free_cpumask_var(this_sibling_map); -free_nodemask: - free_cpumask_var(nodemask); -free_notcovered: -#ifdef CONFIG_NUMA - free_cpumask_var(notcovered); -free_covered: - free_cpumask_var(covered); -free_domainspan: - free_cpumask_var(domainspan); -out: -#endif - return err; - -free_sched_groups: -#ifdef CONFIG_NUMA - kfree(sched_group_nodes); -#endif - goto free_tmpmask; + d.sched_group_nodes = NULL; /* don't free this we still need it */ + __free_domain_allocs(&d, sa_tmpmask, cpu_map); + return 0; -#ifdef CONFIG_NUMA error: - free_sched_groups(cpu_map, tmpmask); - free_rootdomain(rd); - goto free_tmpmask; -#endif + __free_domain_allocs(&d, alloc_state, cpu_map); + return -ENOMEM; } static int build_sched_domains(const struct cpumask *cpu_map) @@ -8173,6 +9370,8 @@ void __init sched_init_smp(void) } #endif /* CONFIG_SMP */ +const_debug unsigned int sysctl_timer_migration = 1; + int in_sched_functions(unsigned long addr) { return in_lock_functions(addr) || @@ -8204,11 +9403,15 @@ static void init_rt_rq(struct rt_rq *rt_rq, struct rq *rq) __set_bit(MAX_RT_PRIO, array->bitmap); #if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED - rt_rq->highest_prio = MAX_RT_PRIO; + rt_rq->highest_prio.curr = MAX_RT_PRIO; +#ifdef CONFIG_SMP + rt_rq->highest_prio.next = MAX_RT_PRIO; +#endif #endif #ifdef CONFIG_SMP rt_rq->rt_nr_migratory = 0; rt_rq->overloaded = 0; + plist_head_init(&rt_rq->pushable_tasks, &rq->lock); #endif rt_rq->rt_time = 0; @@ -8295,12 +9498,15 @@ void __init sched_init(void) #ifdef CONFIG_USER_SCHED alloc_size *= 2; #endif +#ifdef CONFIG_CPUMASK_OFFSTACK + alloc_size += num_possible_cpus() * cpumask_size(); +#endif /* * As sched_init() is called before page_alloc is setup, * we use alloc_bootmem(). */ if (alloc_size) { - ptr = (unsigned long)alloc_bootmem(alloc_size); + ptr = (unsigned long)kzalloc(alloc_size, GFP_NOWAIT); #ifdef CONFIG_FAIR_GROUP_SCHED init_task_group.se = (struct sched_entity **)ptr; @@ -8332,6 +9538,12 @@ void __init sched_init(void) ptr += nr_cpu_ids * sizeof(void **); #endif /* CONFIG_USER_SCHED */ #endif /* CONFIG_RT_GROUP_SCHED */ +#ifdef CONFIG_CPUMASK_OFFSTACK + for_each_possible_cpu(i) { + per_cpu(load_balance_tmpmask, i) = (void *)ptr; + ptr += cpumask_size(); + } +#endif /* CONFIG_CPUMASK_OFFSTACK */ } #ifdef CONFIG_SMP @@ -8367,6 +9579,8 @@ void __init sched_init(void) rq = cpu_rq(i); spin_lock_init(&rq->lock); rq->nr_running = 0; + rq->calc_load_active = 0; + rq->calc_load_update = jiffies + LOAD_FREQ; init_cfs_rq(&rq->cfs, rq); init_rt_rq(&rq->rt, rq); #ifdef CONFIG_FAIR_GROUP_SCHED @@ -8387,7 +9601,7 @@ void __init sched_init(void) * 1024) and two child groups A0 and A1 (of weight 1024 each), * then A0's share of the cpu resource is: * - * A0's bandwidth = 1024 / (10*1024 + 1024 + 1024) = 8.33% + * A0's bandwidth = 1024 / (10*1024 + 1024 + 1024) = 8.33% * * We achieve this by letting init_task_group's tasks sit * directly in rq->cfs (i.e init_task_group->se[] = NULL). @@ -8404,11 +9618,11 @@ void __init sched_init(void) * system cpu resource, based on the weight assigned to root * user's cpu share (INIT_TASK_GROUP_LOAD). This is accomplished * by letting tasks of init_task_group sit in a separate cfs_rq - * (init_cfs_rq) and having one entity represent this group of + * (init_tg_cfs_rq) and having one entity represent this group of * tasks in rq->cfs (i.e init_task_group->se[] != NULL). */ init_tg_cfs_entry(&init_task_group, - &per_cpu(init_cfs_rq, i), + &per_cpu(init_tg_cfs_rq, i), &per_cpu(init_sched_entity, i), i, 1, root_task_group.se[i]); @@ -8434,6 +9648,7 @@ void __init sched_init(void) #ifdef CONFIG_SMP rq->sd = NULL; rq->rd = NULL; + rq->post_schedule = 0; rq->active_balance = 0; rq->next_balance = jiffies; rq->push_cpu = 0; @@ -8474,31 +9689,44 @@ void __init sched_init(void) * when this runqueue becomes "idle". */ init_idle(current, smp_processor_id()); + + calc_load_update = jiffies + LOAD_FREQ; + /* * During early bootup we pretend to be a normal task: */ current->sched_class = &fair_sched_class; /* Allocate the nohz_cpu_mask if CONFIG_CPUMASK_OFFSTACK */ - alloc_bootmem_cpumask_var(&nohz_cpu_mask); + alloc_cpumask_var(&nohz_cpu_mask, GFP_NOWAIT); #ifdef CONFIG_SMP #ifdef CONFIG_NO_HZ - alloc_bootmem_cpumask_var(&nohz.cpu_mask); + alloc_cpumask_var(&nohz.cpu_mask, GFP_NOWAIT); + alloc_cpumask_var(&nohz.ilb_grp_nohz_mask, GFP_NOWAIT); #endif - alloc_bootmem_cpumask_var(&cpu_isolated_map); + alloc_cpumask_var(&cpu_isolated_map, GFP_NOWAIT); #endif /* SMP */ + perf_counter_init(); + scheduler_running = 1; } #ifdef CONFIG_DEBUG_SPINLOCK_SLEEP -void __might_sleep(char *file, int line) +static inline int preempt_count_equals(int preempt_offset) +{ + int nested = preempt_count() & ~PREEMPT_ACTIVE; + + return (nested == PREEMPT_INATOMIC_BASE + preempt_offset); +} + +void __might_sleep(char *file, int line, int preempt_offset) { #ifdef in_atomic static unsigned long prev_jiffy; /* ratelimiting */ - if ((!in_atomic() && !irqs_disabled()) || - system_state != SYSTEM_RUNNING || oops_in_progress) + if ((preempt_count_equals(preempt_offset) && !irqs_disabled()) || + system_state != SYSTEM_RUNNING || oops_in_progress) return; if (time_before(jiffies, prev_jiffy + HZ) && prev_jiffy) return; @@ -9210,6 +10438,16 @@ static int sched_rt_global_constraints(void) return ret; } + +int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk) +{ + /* Don't accept realtime tasks when there is no way for them to run */ + if (rt_task(tsk) && tg->rt_bandwidth.rt_runtime == 0) + return 0; + + return 1; +} + #else /* !CONFIG_RT_GROUP_SCHED */ static int sched_rt_global_constraints(void) { @@ -9219,6 +10457,13 @@ static int sched_rt_global_constraints(void) if (sysctl_sched_rt_period <= 0) return -EINVAL; + /* + * There's always some RT tasks in the root group + * -- migration, kstopmachine etc.. + */ + if (sysctl_sched_rt_runtime == 0) + return -EBUSY; + spin_lock_irqsave(&def_rt_bandwidth.rt_runtime_lock, flags); for_each_possible_cpu(i) { struct rt_rq *rt_rq = &cpu_rq(i)->rt; @@ -9303,8 +10548,7 @@ cpu_cgroup_can_attach(struct cgroup_subsys *ss, struct cgroup *cgrp, struct task_struct *tsk) { #ifdef CONFIG_RT_GROUP_SCHED - /* Don't accept realtime tasks when there is no way for them to run */ - if (rt_task(tsk) && cgroup_tg(cgrp)->rt_bandwidth.rt_runtime == 0) + if (!sched_rt_can_attach(cgroup_tg(cgrp), tsk)) return -EINVAL; #else /* We don't support RT-tasks being in separate groups */ @@ -9415,6 +10659,7 @@ struct cpuacct { struct cgroup_subsys_state css; /* cpuusage holds pointer to a u64-type object on every cpu */ u64 *cpuusage; + struct percpu_counter cpustat[CPUACCT_STAT_NSTATS]; struct cpuacct *parent; }; @@ -9439,20 +10684,32 @@ static struct cgroup_subsys_state *cpuacct_create( struct cgroup_subsys *ss, struct cgroup *cgrp) { struct cpuacct *ca = kzalloc(sizeof(*ca), GFP_KERNEL); + int i; if (!ca) - return ERR_PTR(-ENOMEM); + goto out; ca->cpuusage = alloc_percpu(u64); - if (!ca->cpuusage) { - kfree(ca); - return ERR_PTR(-ENOMEM); - } + if (!ca->cpuusage) + goto out_free_ca; + + for (i = 0; i < CPUACCT_STAT_NSTATS; i++) + if (percpu_counter_init(&ca->cpustat[i], 0)) + goto out_free_counters; if (cgrp->parent) ca->parent = cgroup_ca(cgrp->parent); return &ca->css; + +out_free_counters: + while (--i >= 0) + percpu_counter_destroy(&ca->cpustat[i]); + free_percpu(ca->cpuusage); +out_free_ca: + kfree(ca); +out: + return ERR_PTR(-ENOMEM); } /* destroy an existing cpu accounting group */ @@ -9460,14 +10717,17 @@ static void cpuacct_destroy(struct cgroup_subsys *ss, struct cgroup *cgrp) { struct cpuacct *ca = cgroup_ca(cgrp); + int i; + for (i = 0; i < CPUACCT_STAT_NSTATS; i++) + percpu_counter_destroy(&ca->cpustat[i]); free_percpu(ca->cpuusage); kfree(ca); } static u64 cpuacct_cpuusage_read(struct cpuacct *ca, int cpu) { - u64 *cpuusage = percpu_ptr(ca->cpuusage, cpu); + u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu); u64 data; #ifndef CONFIG_64BIT @@ -9486,7 +10746,7 @@ static u64 cpuacct_cpuusage_read(struct cpuacct *ca, int cpu) static void cpuacct_cpuusage_write(struct cpuacct *ca, int cpu, u64 val) { - u64 *cpuusage = percpu_ptr(ca->cpuusage, cpu); + u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu); #ifndef CONFIG_64BIT /* @@ -9547,6 +10807,25 @@ static int cpuacct_percpu_seq_read(struct cgroup *cgroup, struct cftype *cft, return 0; } +static const char *cpuacct_stat_desc[] = { + [CPUACCT_STAT_USER] = "user", + [CPUACCT_STAT_SYSTEM] = "system", +}; + +static int cpuacct_stats_show(struct cgroup *cgrp, struct cftype *cft, + struct cgroup_map_cb *cb) +{ + struct cpuacct *ca = cgroup_ca(cgrp); + int i; + + for (i = 0; i < CPUACCT_STAT_NSTATS; i++) { + s64 val = percpu_counter_read(&ca->cpustat[i]); + val = cputime64_to_clock_t(val); + cb->fill(cb, cpuacct_stat_desc[i], val); + } + return 0; +} + static struct cftype files[] = { { .name = "usage", @@ -9557,7 +10836,10 @@ static struct cftype files[] = { .name = "usage_percpu", .read_seq_string = cpuacct_percpu_seq_read, }, - + { + .name = "stat", + .read_map = cpuacct_stats_show, + }, }; static int cpuacct_populate(struct cgroup_subsys *ss, struct cgroup *cgrp) @@ -9575,16 +10857,42 @@ static void cpuacct_charge(struct task_struct *tsk, u64 cputime) struct cpuacct *ca; int cpu; - if (!cpuacct_subsys.active) + if (unlikely(!cpuacct_subsys.active)) return; cpu = task_cpu(tsk); + + rcu_read_lock(); + ca = task_ca(tsk); for (; ca; ca = ca->parent) { - u64 *cpuusage = percpu_ptr(ca->cpuusage, cpu); + u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu); *cpuusage += cputime; } + + rcu_read_unlock(); +} + +/* + * Charge the system/user time to the task's accounting group. + */ +static void cpuacct_update_stats(struct task_struct *tsk, + enum cpuacct_stat_index idx, cputime_t val) +{ + struct cpuacct *ca; + + if (unlikely(!cpuacct_subsys.active)) + return; + + rcu_read_lock(); + ca = task_ca(tsk); + + do { + percpu_counter_add(&ca->cpustat[idx], val); + ca = ca->parent; + } while (ca); + rcu_read_unlock(); } struct cgroup_subsys cpuacct_subsys = { @@ -9595,3 +10903,113 @@ struct cgroup_subsys cpuacct_subsys = { .subsys_id = cpuacct_subsys_id, }; #endif /* CONFIG_CGROUP_CPUACCT */ + +#ifndef CONFIG_SMP + +int rcu_expedited_torture_stats(char *page) +{ + return 0; +} +EXPORT_SYMBOL_GPL(rcu_expedited_torture_stats); + +void synchronize_sched_expedited(void) +{ +} +EXPORT_SYMBOL_GPL(synchronize_sched_expedited); + +#else /* #ifndef CONFIG_SMP */ + +static DEFINE_PER_CPU(struct migration_req, rcu_migration_req); +static DEFINE_MUTEX(rcu_sched_expedited_mutex); + +#define RCU_EXPEDITED_STATE_POST -2 +#define RCU_EXPEDITED_STATE_IDLE -1 + +static int rcu_expedited_state = RCU_EXPEDITED_STATE_IDLE; + +int rcu_expedited_torture_stats(char *page) +{ + int cnt = 0; + int cpu; + + cnt += sprintf(&page[cnt], "state: %d /", rcu_expedited_state); + for_each_online_cpu(cpu) { + cnt += sprintf(&page[cnt], " %d:%d", + cpu, per_cpu(rcu_migration_req, cpu).dest_cpu); + } + cnt += sprintf(&page[cnt], "\n"); + return cnt; +} +EXPORT_SYMBOL_GPL(rcu_expedited_torture_stats); + +static long synchronize_sched_expedited_count; + +/* + * Wait for an rcu-sched grace period to elapse, but use "big hammer" + * approach to force grace period to end quickly. This consumes + * significant time on all CPUs, and is thus not recommended for + * any sort of common-case code. + * + * Note that it is illegal to call this function while holding any + * lock that is acquired by a CPU-hotplug notifier. Failing to + * observe this restriction will result in deadlock. + */ +void synchronize_sched_expedited(void) +{ + int cpu; + unsigned long flags; + bool need_full_sync = 0; + struct rq *rq; + struct migration_req *req; + long snap; + int trycount = 0; + + smp_mb(); /* ensure prior mod happens before capturing snap. */ + snap = ACCESS_ONCE(synchronize_sched_expedited_count) + 1; + get_online_cpus(); + while (!mutex_trylock(&rcu_sched_expedited_mutex)) { + put_online_cpus(); + if (trycount++ < 10) + udelay(trycount * num_online_cpus()); + else { + synchronize_sched(); + return; + } + if (ACCESS_ONCE(synchronize_sched_expedited_count) - snap > 0) { + smp_mb(); /* ensure test happens before caller kfree */ + return; + } + get_online_cpus(); + } + rcu_expedited_state = RCU_EXPEDITED_STATE_POST; + for_each_online_cpu(cpu) { + rq = cpu_rq(cpu); + req = &per_cpu(rcu_migration_req, cpu); + init_completion(&req->done); + req->task = NULL; + req->dest_cpu = RCU_MIGRATION_NEED_QS; + spin_lock_irqsave(&rq->lock, flags); + list_add(&req->list, &rq->migration_queue); + spin_unlock_irqrestore(&rq->lock, flags); + wake_up_process(rq->migration_thread); + } + for_each_online_cpu(cpu) { + rcu_expedited_state = cpu; + req = &per_cpu(rcu_migration_req, cpu); + rq = cpu_rq(cpu); + wait_for_completion(&req->done); + spin_lock_irqsave(&rq->lock, flags); + if (unlikely(req->dest_cpu == RCU_MIGRATION_MUST_SYNC)) + need_full_sync = 1; + req->dest_cpu = RCU_MIGRATION_IDLE; + spin_unlock_irqrestore(&rq->lock, flags); + } + rcu_expedited_state = RCU_EXPEDITED_STATE_IDLE; + mutex_unlock(&rcu_sched_expedited_mutex); + put_online_cpus(); + if (need_full_sync) + synchronize_sched(); +} +EXPORT_SYMBOL_GPL(synchronize_sched_expedited); + +#endif /* #else #ifndef CONFIG_SMP */ diff --git a/kernel/sched_clock.c b/kernel/sched_clock.c index a0b0852414c..e1d16c9a768 100644 --- a/kernel/sched_clock.c +++ b/kernel/sched_clock.c @@ -24,11 +24,12 @@ * The clock: sched_clock_cpu() is monotonic per cpu, and should be somewhat * consistent between cpus (never more than 2 jiffies difference). */ -#include <linux/sched.h> -#include <linux/percpu.h> #include <linux/spinlock.h> -#include <linux/ktime.h> +#include <linux/hardirq.h> #include <linux/module.h> +#include <linux/percpu.h> +#include <linux/ktime.h> +#include <linux/sched.h> /* * Scheduler clock - returns current time in nanosec units. @@ -37,12 +38,14 @@ */ unsigned long long __attribute__((weak)) sched_clock(void) { - return (unsigned long long)jiffies * (NSEC_PER_SEC / HZ); + return (unsigned long long)(jiffies - INITIAL_JIFFIES) + * (NSEC_PER_SEC / HZ); } static __read_mostly int sched_clock_running; #ifdef CONFIG_HAVE_UNSTABLE_SCHED_CLOCK +__read_mostly int sched_clock_stable; struct sched_clock_data { /* @@ -87,7 +90,7 @@ void sched_clock_init(void) } /* - * min,max except they take wrapping into account + * min, max except they take wrapping into account */ static inline u64 wrap_min(u64 x, u64 y) @@ -111,15 +114,13 @@ static u64 __update_sched_clock(struct sched_clock_data *scd, u64 now) s64 delta = now - scd->tick_raw; u64 clock, min_clock, max_clock; - WARN_ON_ONCE(!irqs_disabled()); - if (unlikely(delta < 0)) delta = 0; /* * scd->clock = clamp(scd->tick_gtod + delta, - * max(scd->tick_gtod, scd->clock), - * scd->tick_gtod + TICK_NSEC); + * max(scd->tick_gtod, scd->clock), + * scd->tick_gtod + TICK_NSEC); */ clock = scd->tick_gtod + delta; @@ -148,8 +149,20 @@ static void lock_double_clock(struct sched_clock_data *data1, u64 sched_clock_cpu(int cpu) { - struct sched_clock_data *scd = cpu_sdc(cpu); u64 now, clock, this_clock, remote_clock; + struct sched_clock_data *scd; + + if (sched_clock_stable) + return sched_clock(); + + scd = cpu_sdc(cpu); + + /* + * Normally this is not called in NMI context - but if it is, + * trying to do any locking here is totally lethal. + */ + if (unlikely(in_nmi())) + return scd->clock; if (unlikely(!sched_clock_running)) return 0ull; @@ -195,14 +208,18 @@ u64 sched_clock_cpu(int cpu) void sched_clock_tick(void) { - struct sched_clock_data *scd = this_scd(); + struct sched_clock_data *scd; u64 now, now_gtod; + if (sched_clock_stable) + return; + if (unlikely(!sched_clock_running)) return; WARN_ON_ONCE(!irqs_disabled()); + scd = this_scd(); now_gtod = ktime_to_ns(ktime_get()); now = sched_clock(); @@ -250,7 +267,7 @@ u64 sched_clock_cpu(int cpu) return sched_clock(); } -#endif +#endif /* CONFIG_HAVE_UNSTABLE_SCHED_CLOCK */ unsigned long long cpu_clock(int cpu) { diff --git a/kernel/sched_cpupri.c b/kernel/sched_cpupri.c index 1e00bfacf9b..0f052fc674d 100644 --- a/kernel/sched_cpupri.c +++ b/kernel/sched_cpupri.c @@ -55,7 +55,7 @@ static int convert_prio(int prio) * cpupri_find - find the best (lowest-pri) CPU in the system * @cp: The cpupri context * @p: The task - * @lowest_mask: A mask to fill in with selected CPUs + * @lowest_mask: A mask to fill in with selected CPUs (or NULL) * * Note: This function returns the recommended CPUs as calculated during the * current invokation. By the time the call returns, the CPUs may have in @@ -81,7 +81,21 @@ int cpupri_find(struct cpupri *cp, struct task_struct *p, if (cpumask_any_and(&p->cpus_allowed, vec->mask) >= nr_cpu_ids) continue; - cpumask_and(lowest_mask, &p->cpus_allowed, vec->mask); + if (lowest_mask) { + cpumask_and(lowest_mask, &p->cpus_allowed, vec->mask); + + /* + * We have to ensure that we have at least one bit + * still set in the array, since the map could have + * been concurrently emptied between the first and + * second reads of vec->mask. If we hit this + * condition, simply act as though we never hit this + * priority level and continue on. + */ + if (cpumask_any(lowest_mask) >= nr_cpu_ids) + continue; + } + return 1; } @@ -113,21 +127,11 @@ void cpupri_set(struct cpupri *cp, int cpu, int newpri) /* * If the cpu was currently mapped to a different value, we - * first need to unmap the old value + * need to map it to the new value then remove the old value. + * Note, we must add the new value first, otherwise we risk the + * cpu being cleared from pri_active, and this cpu could be + * missed for a push or pull. */ - if (likely(oldpri != CPUPRI_INVALID)) { - struct cpupri_vec *vec = &cp->pri_to_cpu[oldpri]; - - spin_lock_irqsave(&vec->lock, flags); - - vec->count--; - if (!vec->count) - clear_bit(oldpri, cp->pri_active); - cpumask_clear_cpu(cpu, vec->mask); - - spin_unlock_irqrestore(&vec->lock, flags); - } - if (likely(newpri != CPUPRI_INVALID)) { struct cpupri_vec *vec = &cp->pri_to_cpu[newpri]; @@ -140,6 +144,18 @@ void cpupri_set(struct cpupri *cp, int cpu, int newpri) spin_unlock_irqrestore(&vec->lock, flags); } + if (likely(oldpri != CPUPRI_INVALID)) { + struct cpupri_vec *vec = &cp->pri_to_cpu[oldpri]; + + spin_lock_irqsave(&vec->lock, flags); + + vec->count--; + if (!vec->count) + clear_bit(oldpri, cp->pri_active); + cpumask_clear_cpu(cpu, vec->mask); + + spin_unlock_irqrestore(&vec->lock, flags); + } *currpri = newpri; } @@ -151,10 +167,14 @@ void cpupri_set(struct cpupri *cp, int cpu, int newpri) * * Returns: -ENOMEM if memory fails. */ -int __init_refok cpupri_init(struct cpupri *cp, bool bootmem) +int cpupri_init(struct cpupri *cp, bool bootmem) { + gfp_t gfp = GFP_KERNEL; int i; + if (bootmem) + gfp = GFP_NOWAIT; + memset(cp, 0, sizeof(*cp)); for (i = 0; i < CPUPRI_NR_PRIORITIES; i++) { @@ -162,9 +182,7 @@ int __init_refok cpupri_init(struct cpupri *cp, bool bootmem) spin_lock_init(&vec->lock); vec->count = 0; - if (bootmem) - alloc_bootmem_cpumask_var(&vec->mask); - else if (!alloc_cpumask_var(&vec->mask, GFP_KERNEL)) + if (!zalloc_cpumask_var(&vec->mask, gfp)) goto cleanup; } diff --git a/kernel/sched_cpupri.h b/kernel/sched_cpupri.h index 642a94ef8a0..9a7e859b8fb 100644 --- a/kernel/sched_cpupri.h +++ b/kernel/sched_cpupri.h @@ -25,7 +25,7 @@ struct cpupri { #ifdef CONFIG_SMP int cpupri_find(struct cpupri *cp, - struct task_struct *p, cpumask_t *lowest_mask); + struct task_struct *p, struct cpumask *lowest_mask); void cpupri_set(struct cpupri *cp, int cpu, int pri); int cpupri_init(struct cpupri *cp, bool bootmem); void cpupri_cleanup(struct cpupri *cp); diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c index 16eeba4e416..5ddbd089126 100644 --- a/kernel/sched_debug.c +++ b/kernel/sched_debug.c @@ -162,7 +162,7 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) { s64 MIN_vruntime = -1, min_vruntime, max_vruntime = -1, spread, rq0_min_vruntime, spread0; - struct rq *rq = &per_cpu(runqueues, cpu); + struct rq *rq = cpu_rq(cpu); struct sched_entity *last; unsigned long flags; @@ -191,7 +191,7 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq) if (last) max_vruntime = last->vruntime; min_vruntime = cfs_rq->min_vruntime; - rq0_min_vruntime = per_cpu(runqueues, 0).cfs.min_vruntime; + rq0_min_vruntime = cpu_rq(0)->cfs.min_vruntime; spin_unlock_irqrestore(&rq->lock, flags); SEQ_printf(m, " .%-30s: %Ld.%06ld\n", "MIN_vruntime", SPLIT_NS(MIN_vruntime)); @@ -248,7 +248,7 @@ void print_rt_rq(struct seq_file *m, int cpu, struct rt_rq *rt_rq) static void print_cpu(struct seq_file *m, int cpu) { - struct rq *rq = &per_cpu(runqueues, cpu); + struct rq *rq = cpu_rq(cpu); #ifdef CONFIG_X86 { @@ -272,7 +272,6 @@ static void print_cpu(struct seq_file *m, int cpu) P(nr_switches); P(nr_load_updates); P(nr_uninterruptible); - SEQ_printf(m, " .%-30s: %lu\n", "jiffies", jiffies); PN(next_balance); P(curr->pid); PN(clock); @@ -287,9 +286,6 @@ static void print_cpu(struct seq_file *m, int cpu) #ifdef CONFIG_SCHEDSTATS #define P(n) SEQ_printf(m, " .%-30s: %d\n", #n, rq->n); - P(yld_exp_empty); - P(yld_act_empty); - P(yld_both_empty); P(yld_count); P(sched_switch); @@ -314,7 +310,7 @@ static int sched_debug_show(struct seq_file *m, void *v) u64 now = ktime_to_ns(ktime_get()); int cpu; - SEQ_printf(m, "Sched Debug Version: v0.08, %s %.*s\n", + SEQ_printf(m, "Sched Debug Version: v0.09, %s %.*s\n", init_utsname()->release, (int)strcspn(init_utsname()->version, " "), init_utsname()->version); @@ -325,6 +321,7 @@ static int sched_debug_show(struct seq_file *m, void *v) SEQ_printf(m, " .%-40s: %Ld\n", #x, (long long)(x)) #define PN(x) \ SEQ_printf(m, " .%-40s: %Ld.%06ld\n", #x, SPLIT_NS(x)) + P(jiffies); PN(sysctl_sched_latency); PN(sysctl_sched_min_granularity); PN(sysctl_sched_wakeup_granularity); @@ -397,6 +394,7 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) PN(se.vruntime); PN(se.sum_exec_runtime); PN(se.avg_overlap); + PN(se.avg_wakeup); nr_switches = p->nvcsw + p->nivcsw; @@ -411,6 +409,8 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m) PN(se.wait_max); PN(se.wait_sum); P(se.wait_count); + PN(se.iowait_sum); + P(se.iowait_count); P(sched_info.bkl_count); P(se.nr_migrations); P(se.nr_migrations_cold); @@ -481,6 +481,8 @@ void proc_sched_set_task(struct task_struct *p) p->se.wait_max = 0; p->se.wait_sum = 0; p->se.wait_count = 0; + p->se.iowait_sum = 0; + p->se.iowait_count = 0; p->se.sleep_max = 0; p->se.sum_sleep_runtime = 0; p->se.block_max = 0; diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index 5cc1c162044..aa7f8412101 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -24,7 +24,7 @@ /* * Targeted preemption latency for CPU-bound tasks: - * (default: 20ms * (1 + ilog(ncpus)), units: nanoseconds) + * (default: 5ms * (1 + ilog(ncpus)), units: nanoseconds) * * NOTE: this latency value is not the same as the concept of * 'timeslice length' - timeslices in CFS are of variable length @@ -34,13 +34,13 @@ * (to see the precise effective timeslice length of your workload, * run vmstat and monitor the context-switches (cs) field) */ -unsigned int sysctl_sched_latency = 20000000ULL; +unsigned int sysctl_sched_latency = 5000000ULL; /* * Minimal preemption granularity for CPU-bound tasks: - * (default: 4 msec * (1 + ilog(ncpus)), units: nanoseconds) + * (default: 1 msec * (1 + ilog(ncpus)), units: nanoseconds) */ -unsigned int sysctl_sched_min_granularity = 4000000ULL; +unsigned int sysctl_sched_min_granularity = 1000000ULL; /* * is kept at sysctl_sched_latency / sysctl_sched_min_granularity @@ -48,10 +48,10 @@ unsigned int sysctl_sched_min_granularity = 4000000ULL; static unsigned int sched_nr_latency = 5; /* - * After fork, child runs first. (default) If set to 0 then + * After fork, child runs first. If set to 0 (default) then * parent will (try to) run first. */ -const_debug unsigned int sysctl_sched_child_runs_first = 1; +unsigned int sysctl_sched_child_runs_first __read_mostly; /* * sys_sched_yield() compat mode @@ -63,13 +63,13 @@ unsigned int __read_mostly sysctl_sched_compat_yield; /* * SCHED_OTHER wake-up granularity. - * (default: 5 msec * (1 + ilog(ncpus)), units: nanoseconds) + * (default: 1 msec * (1 + ilog(ncpus)), units: nanoseconds) * * This option delays the preemption effects of decoupled workloads * and reduces their over-scheduling. Synchronous workloads will still * have immediate wakeup/sleep latencies. */ -unsigned int sysctl_sched_wakeup_granularity = 5000000UL; +unsigned int sysctl_sched_wakeup_granularity = 1000000UL; const_debug unsigned int sysctl_sched_migration_cost = 500000UL; @@ -79,11 +79,6 @@ static const struct sched_class fair_sched_class; * CFS operations on generic schedulable entities: */ -static inline struct task_struct *task_of(struct sched_entity *se) -{ - return container_of(se, struct task_struct, se); -} - #ifdef CONFIG_FAIR_GROUP_SCHED /* cpu runqueue to which this cfs_rq is attached */ @@ -95,6 +90,14 @@ static inline struct rq *rq_of(struct cfs_rq *cfs_rq) /* An entity is a task if it doesn't "own" a runqueue */ #define entity_is_task(se) (!se->my_q) +static inline struct task_struct *task_of(struct sched_entity *se) +{ +#ifdef CONFIG_SCHED_DEBUG + WARN_ON_ONCE(!entity_is_task(se)); +#endif + return container_of(se, struct task_struct, se); +} + /* Walk up scheduling entities hierarchy */ #define for_each_sched_entity(se) \ for (; se; se = se->parent) @@ -186,7 +189,12 @@ find_matching_se(struct sched_entity **se, struct sched_entity **pse) } } -#else /* CONFIG_FAIR_GROUP_SCHED */ +#else /* !CONFIG_FAIR_GROUP_SCHED */ + +static inline struct task_struct *task_of(struct sched_entity *se) +{ + return container_of(se, struct task_struct, se); +} static inline struct rq *rq_of(struct cfs_rq *cfs_rq) { @@ -266,6 +274,12 @@ static inline u64 min_vruntime(u64 min_vruntime, u64 vruntime) return min_vruntime; } +static inline int entity_before(struct sched_entity *a, + struct sched_entity *b) +{ + return (s64)(a->vruntime - b->vruntime) < 0; +} + static inline s64 entity_key(struct cfs_rq *cfs_rq, struct sched_entity *se) { return se->vruntime - cfs_rq->min_vruntime; @@ -430,12 +444,13 @@ static u64 sched_slice(struct cfs_rq *cfs_rq, struct sched_entity *se) for_each_sched_entity(se) { struct load_weight *load; + struct load_weight lw; cfs_rq = cfs_rq_of(se); load = &cfs_rq->load; if (unlikely(!se->on_rq)) { - struct load_weight lw = cfs_rq->load; + lw = cfs_rq->load; update_load_add(&lw, se->load.weight); load = &lw; @@ -530,6 +545,12 @@ update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se) schedstat_set(se->wait_count, se->wait_count + 1); schedstat_set(se->wait_sum, se->wait_sum + rq_of(cfs_rq)->clock - se->wait_start); +#ifdef CONFIG_SCHEDSTATS + if (entity_is_task(se)) { + trace_sched_stat_wait(task_of(se), + rq_of(cfs_rq)->clock - se->wait_start); + } +#endif schedstat_set(se->wait_start, 0); } @@ -604,9 +625,13 @@ account_entity_dequeue(struct cfs_rq *cfs_rq, struct sched_entity *se) static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) { #ifdef CONFIG_SCHEDSTATS + struct task_struct *tsk = NULL; + + if (entity_is_task(se)) + tsk = task_of(se); + if (se->sleep_start) { u64 delta = rq_of(cfs_rq)->clock - se->sleep_start; - struct task_struct *tsk = task_of(se); if ((s64)delta < 0) delta = 0; @@ -617,11 +642,13 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) se->sleep_start = 0; se->sum_sleep_runtime += delta; - account_scheduler_latency(tsk, delta >> 10, 1); + if (tsk) { + account_scheduler_latency(tsk, delta >> 10, 1); + trace_sched_stat_sleep(tsk, delta); + } } if (se->block_start) { u64 delta = rq_of(cfs_rq)->clock - se->block_start; - struct task_struct *tsk = task_of(se); if ((s64)delta < 0) delta = 0; @@ -632,17 +659,25 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) se->block_start = 0; se->sum_sleep_runtime += delta; - /* - * Blocking time is in units of nanosecs, so shift by 20 to - * get a milliseconds-range estimation of the amount of - * time that the task spent sleeping: - */ - if (unlikely(prof_on == SLEEP_PROFILING)) { + if (tsk) { + if (tsk->in_iowait) { + se->iowait_sum += delta; + se->iowait_count++; + trace_sched_stat_iowait(tsk, delta); + } - profile_hits(SLEEP_PROFILING, (void *)get_wchan(tsk), - delta >> 20); + /* + * Blocking time is in units of nanosecs, so shift by + * 20 to get a milliseconds-range estimation of the + * amount of time that the task spent sleeping: + */ + if (unlikely(prof_on == SLEEP_PROFILING)) { + profile_hits(SLEEP_PROFILING, + (void *)get_wchan(tsk), + delta >> 20); + } + account_scheduler_latency(tsk, delta >> 10, 0); } - account_scheduler_latency(tsk, delta >> 10, 0); } #endif } @@ -686,16 +721,17 @@ place_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int initial) * all of which have the same weight. */ if (sched_feat(NORMALIZED_SLEEPER) && - task_of(se)->policy != SCHED_IDLE) + (!entity_is_task(se) || + task_of(se)->policy != SCHED_IDLE)) thresh = calc_delta_fair(thresh, se); vruntime -= thresh; } - - /* ensure we never gain time by being placed backwards. */ - vruntime = max_vruntime(se->vruntime, vruntime); } + /* ensure we never gain time by being placed backwards. */ + vruntime = max_vruntime(se->vruntime, vruntime); + se->vruntime = vruntime; } @@ -719,7 +755,7 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int wakeup) __enqueue_entity(cfs_rq, se); } -static void clear_buddies(struct cfs_rq *cfs_rq, struct sched_entity *se) +static void __clear_buddies(struct cfs_rq *cfs_rq, struct sched_entity *se) { if (cfs_rq->last == se) cfs_rq->last = NULL; @@ -728,6 +764,12 @@ static void clear_buddies(struct cfs_rq *cfs_rq, struct sched_entity *se) cfs_rq->next = NULL; } +static void clear_buddies(struct cfs_rq *cfs_rq, struct sched_entity *se) +{ + for_each_sched_entity(se) + __clear_buddies(cfs_rq_of(se), se); +} + static void dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int sleep) { @@ -768,8 +810,14 @@ check_preempt_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr) ideal_runtime = sched_slice(cfs_rq, curr); delta_exec = curr->sum_exec_runtime - curr->prev_sum_exec_runtime; - if (delta_exec > ideal_runtime) + if (delta_exec > ideal_runtime) { resched_task(rq_of(cfs_rq)->curr); + /* + * The current task ran long enough, ensure it doesn't get + * re-elected due to buddy favours. + */ + clear_buddies(cfs_rq, curr); + } } static void @@ -1003,7 +1051,7 @@ static void yield_task_fair(struct rq *rq) /* * Already in the rightmost position? */ - if (unlikely(!rightmost || rightmost->vruntime < se->vruntime)) + if (unlikely(!rightmost || entity_before(rightmost, se))) return; /* @@ -1020,17 +1068,21 @@ static void yield_task_fair(struct rq *rq) * search starts with cpus closest then further out as needed, * so we always favor a closer, idle cpu. * Domains may include CPUs that are not usable for migration, - * hence we need to mask them out (cpu_active_mask) + * hence we need to mask them out (rq->rd->online) * * Returns the CPU we should wake onto. */ #if defined(ARCH_HAS_SCHED_WAKE_IDLE) + +#define cpu_rd_active(cpu, rq) cpumask_test_cpu(cpu, rq->rd->online) + static int wake_idle(int cpu, struct task_struct *p) { struct sched_domain *sd; int i; unsigned int chosen_wakeup_cpu; int this_cpu; + struct rq *task_rq = task_rq(p); /* * At POWERSAVINGS_BALANCE_WAKEUP level, if both this_cpu and prev_cpu @@ -1063,10 +1115,10 @@ static int wake_idle(int cpu, struct task_struct *p) for_each_domain(cpu, sd) { if ((sd->flags & SD_WAKE_IDLE) || ((sd->flags & SD_WAKE_IDLE_FAR) - && !task_hot(p, task_rq(p)->clock, sd))) { + && !task_hot(p, task_rq->clock, sd))) { for_each_cpu_and(i, sched_domain_span(sd), &p->cpus_allowed) { - if (cpu_active(i) && idle_cpu(i)) { + if (cpu_rd_active(i, task_rq) && idle_cpu(i)) { if (i != task_cpu(p)) { schedstat_inc(p, se.nr_wakeups_idle); @@ -1209,7 +1261,17 @@ wake_affine(struct sched_domain *this_sd, struct rq *this_rq, tg = task_group(p); weight = p->se.load.weight; - balanced = 100*(tl + effective_load(tg, this_cpu, weight, weight)) <= + /* + * In low-load situations, where prev_cpu is idle and this_cpu is idle + * due to the sync cause above having dropped tl to 0, we'll always have + * an imbalance, but there's really nothing you can do about that, so + * that's good too. + * + * Otherwise check if either cpus are near enough in load to allow this + * task to be woken on this_cpu. + */ + balanced = !tl || + 100*(tl + effective_load(tg, this_cpu, weight, weight)) <= imbalance*(load + effective_load(tg, prev_cpu, 0, weight)); /* @@ -1252,8 +1314,6 @@ static int select_task_rq_fair(struct task_struct *p, int sync) this_rq = cpu_rq(this_cpu); new_cpu = prev_cpu; - if (prev_cpu == this_cpu) - goto out; /* * 'this_sd' is the first domain that both * this_cpu and prev_cpu are present in: @@ -1302,16 +1362,63 @@ out: } #endif /* CONFIG_SMP */ -static unsigned long wakeup_gran(struct sched_entity *se) +/* + * Adaptive granularity + * + * se->avg_wakeup gives the average time a task runs until it does a wakeup, + * with the limit of wakeup_gran -- when it never does a wakeup. + * + * So the smaller avg_wakeup is the faster we want this task to preempt, + * but we don't want to treat the preemptee unfairly and therefore allow it + * to run for at least the amount of time we'd like to run. + * + * NOTE: we use 2*avg_wakeup to increase the probability of actually doing one + * + * NOTE: we use *nr_running to scale with load, this nicely matches the + * degrading latency on load. + */ +static unsigned long +adaptive_gran(struct sched_entity *curr, struct sched_entity *se) +{ + u64 this_run = curr->sum_exec_runtime - curr->prev_sum_exec_runtime; + u64 expected_wakeup = 2*se->avg_wakeup * cfs_rq_of(se)->nr_running; + u64 gran = 0; + + if (this_run < expected_wakeup) + gran = expected_wakeup - this_run; + + return min_t(s64, gran, sysctl_sched_wakeup_granularity); +} + +static unsigned long +wakeup_gran(struct sched_entity *curr, struct sched_entity *se) { unsigned long gran = sysctl_sched_wakeup_granularity; + if (cfs_rq_of(curr)->curr && sched_feat(ADAPTIVE_GRAN)) + gran = adaptive_gran(curr, se); + /* - * More easily preempt - nice tasks, while not making it harder for - * + nice tasks. + * Since its curr running now, convert the gran from real-time + * to virtual-time in his units. */ - if (!sched_feat(ASYM_GRAN) || se->load.weight > NICE_0_LOAD) - gran = calc_delta_fair(sysctl_sched_wakeup_granularity, se); + if (sched_feat(ASYM_GRAN)) { + /* + * By using 'se' instead of 'curr' we penalize light tasks, so + * they get preempted easier. That is, if 'se' < 'curr' then + * the resulting gran will be larger, therefore penalizing the + * lighter, if otoh 'se' > 'curr' then the resulting gran will + * be smaller, again penalizing the lighter task. + * + * This is especially important for buddies when the leftmost + * task is higher priority than the buddy. + */ + if (unlikely(se->load.weight != NICE_0_LOAD)) + gran = calc_delta_fair(gran, se); + } else { + if (unlikely(curr->load.weight != NICE_0_LOAD)) + gran = calc_delta_fair(gran, curr); + } return gran; } @@ -1338,7 +1445,7 @@ wakeup_preempt_entity(struct sched_entity *curr, struct sched_entity *se) if (vdiff <= 0) return -1; - gran = wakeup_gran(curr); + gran = wakeup_gran(curr, se); if (vdiff > gran) return 1; @@ -1428,17 +1535,10 @@ static void check_preempt_wakeup(struct rq *rq, struct task_struct *p, int sync) find_matching_se(&se, &pse); - while (se) { - BUG_ON(!pse); - - if (wakeup_preempt_entity(se, pse) == 1) { - resched_task(curr); - break; - } + BUG_ON(!pse); - se = parent_entity(se); - pse = parent_entity(pse); - } + if (wakeup_preempt_entity(se, pse) == 1) + resched_task(curr); } static struct task_struct *pick_next_task_fair(struct rq *rq) @@ -1452,6 +1552,11 @@ static struct task_struct *pick_next_task_fair(struct rq *rq) do { se = pick_next_entity(cfs_rq); + /* + * If se was a buddy, clear it so that it will have to earn + * the favour again. + */ + __clear_buddies(cfs_rq, se); set_next_entity(cfs_rq, se); cfs_rq = group_cfs_rq(se); } while (cfs_rq); @@ -1650,11 +1755,13 @@ static void task_new_fair(struct rq *rq, struct task_struct *p) sched_info_queued(p); update_curr(cfs_rq); + if (curr) + se->vruntime = curr->vruntime; place_entity(cfs_rq, se, 1); /* 'curr' will be NULL if the child belongs to a different group */ if (sysctl_sched_child_runs_first && this_cpu == task_cpu(p) && - curr && curr->vruntime < se->vruntime) { + curr && entity_before(curr, se)) { /* * Upon rescheduling, sched_class::put_prev_task() will place * 'current' within the tree based on its new key value. diff --git a/kernel/sched_features.h b/kernel/sched_features.h index da5d93b5d2c..e2dc63a5815 100644 --- a/kernel/sched_features.h +++ b/kernel/sched_features.h @@ -1,5 +1,6 @@ -SCHED_FEAT(NEW_FAIR_SLEEPERS, 1) -SCHED_FEAT(NORMALIZED_SLEEPER, 1) +SCHED_FEAT(NEW_FAIR_SLEEPERS, 0) +SCHED_FEAT(NORMALIZED_SLEEPER, 0) +SCHED_FEAT(ADAPTIVE_GRAN, 1) SCHED_FEAT(WAKEUP_PREEMPT, 1) SCHED_FEAT(START_DEBIT, 1) SCHED_FEAT(AFFINE_WAKEUPS, 1) @@ -13,3 +14,4 @@ SCHED_FEAT(LB_WAKEUP_UPDATE, 1) SCHED_FEAT(ASYM_EFF_LOAD, 1) SCHED_FEAT(WAKEUP_OVERLAP, 0) SCHED_FEAT(LAST_BUDDY, 1) +SCHED_FEAT(OWNER_SPIN, 1) diff --git a/kernel/sched_idletask.c b/kernel/sched_idletask.c index 8a21a2e28c1..499672c10cb 100644 --- a/kernel/sched_idletask.c +++ b/kernel/sched_idletask.c @@ -22,7 +22,8 @@ static void check_preempt_curr_idle(struct rq *rq, struct task_struct *p, int sy static struct task_struct *pick_next_task_idle(struct rq *rq) { schedstat_inc(rq, sched_goidle); - + /* adjust the active tasks as we might go into a long sleep */ + calc_load_account_active(rq); return rq->idle; } diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c index 954e1a81b79..2eb4bd6a526 100644 --- a/kernel/sched_rt.c +++ b/kernel/sched_rt.c @@ -3,6 +3,52 @@ * policies) */ +#ifdef CONFIG_RT_GROUP_SCHED + +#define rt_entity_is_task(rt_se) (!(rt_se)->my_q) + +static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se) +{ +#ifdef CONFIG_SCHED_DEBUG + WARN_ON_ONCE(!rt_entity_is_task(rt_se)); +#endif + return container_of(rt_se, struct task_struct, rt); +} + +static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq) +{ + return rt_rq->rq; +} + +static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se) +{ + return rt_se->rt_rq; +} + +#else /* CONFIG_RT_GROUP_SCHED */ + +#define rt_entity_is_task(rt_se) (1) + +static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se) +{ + return container_of(rt_se, struct task_struct, rt); +} + +static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq) +{ + return container_of(rt_rq, struct rq, rt); +} + +static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se) +{ + struct task_struct *p = rt_task_of(rt_se); + struct rq *rq = task_rq(p); + + return &rq->rt; +} + +#endif /* CONFIG_RT_GROUP_SCHED */ + #ifdef CONFIG_SMP static inline int rt_overloaded(struct rq *rq) @@ -37,25 +83,86 @@ static inline void rt_clear_overload(struct rq *rq) cpumask_clear_cpu(rq->cpu, rq->rd->rto_mask); } -static void update_rt_migration(struct rq *rq) +static void update_rt_migration(struct rt_rq *rt_rq) { - if (rq->rt.rt_nr_migratory && (rq->rt.rt_nr_running > 1)) { - if (!rq->rt.overloaded) { - rt_set_overload(rq); - rq->rt.overloaded = 1; + if (rt_rq->rt_nr_migratory && rt_rq->rt_nr_total > 1) { + if (!rt_rq->overloaded) { + rt_set_overload(rq_of_rt_rq(rt_rq)); + rt_rq->overloaded = 1; } - } else if (rq->rt.overloaded) { - rt_clear_overload(rq); - rq->rt.overloaded = 0; + } else if (rt_rq->overloaded) { + rt_clear_overload(rq_of_rt_rq(rt_rq)); + rt_rq->overloaded = 0; } } -#endif /* CONFIG_SMP */ -static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se) +static void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +{ + if (!rt_entity_is_task(rt_se)) + return; + + rt_rq = &rq_of_rt_rq(rt_rq)->rt; + + rt_rq->rt_nr_total++; + if (rt_se->nr_cpus_allowed > 1) + rt_rq->rt_nr_migratory++; + + update_rt_migration(rt_rq); +} + +static void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +{ + if (!rt_entity_is_task(rt_se)) + return; + + rt_rq = &rq_of_rt_rq(rt_rq)->rt; + + rt_rq->rt_nr_total--; + if (rt_se->nr_cpus_allowed > 1) + rt_rq->rt_nr_migratory--; + + update_rt_migration(rt_rq); +} + +static void enqueue_pushable_task(struct rq *rq, struct task_struct *p) +{ + plist_del(&p->pushable_tasks, &rq->rt.pushable_tasks); + plist_node_init(&p->pushable_tasks, p->prio); + plist_add(&p->pushable_tasks, &rq->rt.pushable_tasks); +} + +static void dequeue_pushable_task(struct rq *rq, struct task_struct *p) +{ + plist_del(&p->pushable_tasks, &rq->rt.pushable_tasks); +} + +static inline int has_pushable_tasks(struct rq *rq) +{ + return !plist_head_empty(&rq->rt.pushable_tasks); +} + +#else + +static inline void enqueue_pushable_task(struct rq *rq, struct task_struct *p) +{ +} + +static inline void dequeue_pushable_task(struct rq *rq, struct task_struct *p) +{ +} + +static inline +void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +{ +} + +static inline +void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) { - return container_of(rt_se, struct task_struct, rt); } +#endif /* CONFIG_SMP */ + static inline int on_rt_rq(struct sched_rt_entity *rt_se) { return !list_empty(&rt_se->run_list); @@ -79,16 +186,6 @@ static inline u64 sched_rt_period(struct rt_rq *rt_rq) #define for_each_leaf_rt_rq(rt_rq, rq) \ list_for_each_entry_rcu(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list) -static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq) -{ - return rt_rq->rq; -} - -static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se) -{ - return rt_se->rt_rq; -} - #define for_each_sched_rt_entity(rt_se) \ for (; rt_se; rt_se = rt_se->parent) @@ -108,7 +205,7 @@ static void sched_rt_rq_enqueue(struct rt_rq *rt_rq) if (rt_rq->rt_nr_running) { if (rt_se && !on_rt_rq(rt_se)) enqueue_rt_entity(rt_se); - if (rt_rq->highest_prio < curr->prio) + if (rt_rq->highest_prio.curr < curr->prio) resched_task(curr); } } @@ -176,19 +273,6 @@ static inline u64 sched_rt_period(struct rt_rq *rt_rq) #define for_each_leaf_rt_rq(rt_rq, rq) \ for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL) -static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq) -{ - return container_of(rt_rq, struct rq, rt); -} - -static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se) -{ - struct task_struct *p = rt_task_of(rt_se); - struct rq *rq = task_rq(p); - - return &rq->rt; -} - #define for_each_sched_rt_entity(rt_se) \ for (; rt_se; rt_se = NULL) @@ -473,7 +557,7 @@ static inline int rt_se_prio(struct sched_rt_entity *rt_se) struct rt_rq *rt_rq = group_rt_rq(rt_se); if (rt_rq) - return rt_rq->highest_prio; + return rt_rq->highest_prio.curr; #endif return rt_task_of(rt_se)->prio; @@ -531,6 +615,8 @@ static void update_curr_rt(struct rq *rq) curr->se.exec_start = rq->clock; cpuacct_charge(curr, delta_exec); + sched_rt_avg_update(rq, delta_exec); + if (!rt_bandwidth_enabled()) return; @@ -547,91 +633,174 @@ static void update_curr_rt(struct rq *rq) } } -static inline -void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +#if defined CONFIG_SMP + +static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu); + +static inline int next_prio(struct rq *rq) { - WARN_ON(!rt_prio(rt_se_prio(rt_se))); - rt_rq->rt_nr_running++; -#if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED - if (rt_se_prio(rt_se) < rt_rq->highest_prio) { -#ifdef CONFIG_SMP - struct rq *rq = rq_of_rt_rq(rt_rq); -#endif + struct task_struct *next = pick_next_highest_task_rt(rq, rq->cpu); + + if (next && rt_prio(next->prio)) + return next->prio; + else + return MAX_RT_PRIO; +} + +static void +inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) +{ + struct rq *rq = rq_of_rt_rq(rt_rq); + + if (prio < prev_prio) { + + /* + * If the new task is higher in priority than anything on the + * run-queue, we know that the previous high becomes our + * next-highest. + */ + rt_rq->highest_prio.next = prev_prio; - rt_rq->highest_prio = rt_se_prio(rt_se); -#ifdef CONFIG_SMP if (rq->online) - cpupri_set(&rq->rd->cpupri, rq->cpu, - rt_se_prio(rt_se)); -#endif - } -#endif -#ifdef CONFIG_SMP - if (rt_se->nr_cpus_allowed > 1) { - struct rq *rq = rq_of_rt_rq(rt_rq); + cpupri_set(&rq->rd->cpupri, rq->cpu, prio); - rq->rt.rt_nr_migratory++; - } + } else if (prio == rt_rq->highest_prio.curr) + /* + * If the next task is equal in priority to the highest on + * the run-queue, then we implicitly know that the next highest + * task cannot be any lower than current + */ + rt_rq->highest_prio.next = prio; + else if (prio < rt_rq->highest_prio.next) + /* + * Otherwise, we need to recompute next-highest + */ + rt_rq->highest_prio.next = next_prio(rq); +} - update_rt_migration(rq_of_rt_rq(rt_rq)); -#endif -#ifdef CONFIG_RT_GROUP_SCHED - if (rt_se_boosted(rt_se)) - rt_rq->rt_nr_boosted++; +static void +dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) +{ + struct rq *rq = rq_of_rt_rq(rt_rq); - if (rt_rq->tg) - start_rt_bandwidth(&rt_rq->tg->rt_bandwidth); -#else - start_rt_bandwidth(&def_rt_bandwidth); -#endif + if (rt_rq->rt_nr_running && (prio <= rt_rq->highest_prio.next)) + rt_rq->highest_prio.next = next_prio(rq); + + if (rq->online && rt_rq->highest_prio.curr != prev_prio) + cpupri_set(&rq->rd->cpupri, rq->cpu, rt_rq->highest_prio.curr); } +#else /* CONFIG_SMP */ + static inline -void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) -{ -#ifdef CONFIG_SMP - int highest_prio = rt_rq->highest_prio; -#endif +void inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) {} +static inline +void dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) {} + +#endif /* CONFIG_SMP */ - WARN_ON(!rt_prio(rt_se_prio(rt_se))); - WARN_ON(!rt_rq->rt_nr_running); - rt_rq->rt_nr_running--; #if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED +static void +inc_rt_prio(struct rt_rq *rt_rq, int prio) +{ + int prev_prio = rt_rq->highest_prio.curr; + + if (prio < prev_prio) + rt_rq->highest_prio.curr = prio; + + inc_rt_prio_smp(rt_rq, prio, prev_prio); +} + +static void +dec_rt_prio(struct rt_rq *rt_rq, int prio) +{ + int prev_prio = rt_rq->highest_prio.curr; + if (rt_rq->rt_nr_running) { - struct rt_prio_array *array; - WARN_ON(rt_se_prio(rt_se) < rt_rq->highest_prio); - if (rt_se_prio(rt_se) == rt_rq->highest_prio) { - /* recalculate */ - array = &rt_rq->active; - rt_rq->highest_prio = + WARN_ON(prio < prev_prio); + + /* + * This may have been our highest task, and therefore + * we may have some recomputation to do + */ + if (prio == prev_prio) { + struct rt_prio_array *array = &rt_rq->active; + + rt_rq->highest_prio.curr = sched_find_first_bit(array->bitmap); - } /* otherwise leave rq->highest prio alone */ + } + } else - rt_rq->highest_prio = MAX_RT_PRIO; -#endif -#ifdef CONFIG_SMP - if (rt_se->nr_cpus_allowed > 1) { - struct rq *rq = rq_of_rt_rq(rt_rq); - rq->rt.rt_nr_migratory--; - } + rt_rq->highest_prio.curr = MAX_RT_PRIO; - if (rt_rq->highest_prio != highest_prio) { - struct rq *rq = rq_of_rt_rq(rt_rq); + dec_rt_prio_smp(rt_rq, prio, prev_prio); +} - if (rq->online) - cpupri_set(&rq->rd->cpupri, rq->cpu, - rt_rq->highest_prio); - } +#else + +static inline void inc_rt_prio(struct rt_rq *rt_rq, int prio) {} +static inline void dec_rt_prio(struct rt_rq *rt_rq, int prio) {} + +#endif /* CONFIG_SMP || CONFIG_RT_GROUP_SCHED */ - update_rt_migration(rq_of_rt_rq(rt_rq)); -#endif /* CONFIG_SMP */ #ifdef CONFIG_RT_GROUP_SCHED + +static void +inc_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +{ + if (rt_se_boosted(rt_se)) + rt_rq->rt_nr_boosted++; + + if (rt_rq->tg) + start_rt_bandwidth(&rt_rq->tg->rt_bandwidth); +} + +static void +dec_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +{ if (rt_se_boosted(rt_se)) rt_rq->rt_nr_boosted--; WARN_ON(!rt_rq->rt_nr_running && rt_rq->rt_nr_boosted); -#endif +} + +#else /* CONFIG_RT_GROUP_SCHED */ + +static void +inc_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +{ + start_rt_bandwidth(&def_rt_bandwidth); +} + +static inline +void dec_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) {} + +#endif /* CONFIG_RT_GROUP_SCHED */ + +static inline +void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +{ + int prio = rt_se_prio(rt_se); + + WARN_ON(!rt_prio(prio)); + rt_rq->rt_nr_running++; + + inc_rt_prio(rt_rq, prio); + inc_rt_migration(rt_se, rt_rq); + inc_rt_group(rt_se, rt_rq); +} + +static inline +void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) +{ + WARN_ON(!rt_prio(rt_se_prio(rt_se))); + WARN_ON(!rt_rq->rt_nr_running); + rt_rq->rt_nr_running--; + + dec_rt_prio(rt_rq, rt_se_prio(rt_se)); + dec_rt_migration(rt_se, rt_rq); + dec_rt_group(rt_se, rt_rq); } static void __enqueue_rt_entity(struct sched_rt_entity *rt_se) @@ -718,7 +887,8 @@ static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup) enqueue_rt_entity(rt_se); - inc_cpu_load(rq, p->se.load.weight); + if (!task_current(rq, p) && p->rt.nr_cpus_allowed > 1) + enqueue_pushable_task(rq, p); } static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int sleep) @@ -728,7 +898,7 @@ static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int sleep) update_curr_rt(rq); dequeue_rt_entity(rt_se); - dec_cpu_load(rq, p->se.load.weight); + dequeue_pushable_task(rq, p); } /* @@ -805,20 +975,15 @@ static int select_task_rq_rt(struct task_struct *p, int sync) static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p) { - cpumask_var_t mask; - if (rq->curr->rt.nr_cpus_allowed == 1) return; - if (!alloc_cpumask_var(&mask, GFP_ATOMIC)) - return; - if (p->rt.nr_cpus_allowed != 1 - && cpupri_find(&rq->rd->cpupri, p, mask)) - goto free; + && cpupri_find(&rq->rd->cpupri, p, NULL)) + return; - if (!cpupri_find(&rq->rd->cpupri, rq->curr, mask)) - goto free; + if (!cpupri_find(&rq->rd->cpupri, rq->curr, NULL)) + return; /* * There appears to be other cpus that can accept @@ -827,8 +992,6 @@ static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p) */ requeue_task_rt(rq, p, 1); resched_task(rq->curr); -free: - free_cpumask_var(mask); } #endif /* CONFIG_SMP */ @@ -878,7 +1041,7 @@ static struct sched_rt_entity *pick_next_rt_entity(struct rq *rq, return next; } -static struct task_struct *pick_next_task_rt(struct rq *rq) +static struct task_struct *_pick_next_task_rt(struct rq *rq) { struct sched_rt_entity *rt_se; struct task_struct *p; @@ -900,6 +1063,26 @@ static struct task_struct *pick_next_task_rt(struct rq *rq) p = rt_task_of(rt_se); p->se.exec_start = rq->clock; + + return p; +} + +static struct task_struct *pick_next_task_rt(struct rq *rq) +{ + struct task_struct *p = _pick_next_task_rt(rq); + + /* The running task is never eligible for pushing */ + if (p) + dequeue_pushable_task(rq, p); + +#ifdef CONFIG_SMP + /* + * We detect this state here so that we can avoid taking the RQ + * lock again later if there is no need to push + */ + rq->post_schedule = has_pushable_tasks(rq); +#endif + return p; } @@ -907,6 +1090,13 @@ static void put_prev_task_rt(struct rq *rq, struct task_struct *p) { update_curr_rt(rq); p->se.exec_start = 0; + + /* + * The previous task needs to be made eligible for pushing + * if it is still active + */ + if (p->se.on_rq && p->rt.nr_cpus_allowed > 1) + enqueue_pushable_task(rq, p); } #ifdef CONFIG_SMP @@ -960,16 +1150,17 @@ static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu) static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask); -static inline int pick_optimal_cpu(int this_cpu, cpumask_t *mask) +static inline int pick_optimal_cpu(int this_cpu, + const struct cpumask *mask) { int first; /* "this_cpu" is cheaper to preempt than a remote processor */ - if ((this_cpu != -1) && cpu_isset(this_cpu, *mask)) + if ((this_cpu != -1) && cpumask_test_cpu(this_cpu, mask)) return this_cpu; - first = first_cpu(*mask); - if (first != NR_CPUS) + first = cpumask_first(mask); + if (first < nr_cpu_ids) return first; return -1; @@ -981,6 +1172,7 @@ static int find_lowest_rq(struct task_struct *task) struct cpumask *lowest_mask = __get_cpu_var(local_cpu_mask); int this_cpu = smp_processor_id(); int cpu = task_cpu(task); + cpumask_var_t domain_mask; if (task->rt.nr_cpus_allowed == 1) return -1; /* No other targets possible */ @@ -989,13 +1181,6 @@ static int find_lowest_rq(struct task_struct *task) return -1; /* No targets found */ /* - * Only consider CPUs that are usable for migration. - * I guess we might want to change cpupri_find() to ignore those - * in the first place. - */ - cpumask_and(lowest_mask, lowest_mask, cpu_active_mask); - - /* * At this point we have built a mask of cpus representing the * lowest priority tasks in the system. Now we want to elect * the best one based on our affinity and topology. @@ -1013,19 +1198,25 @@ static int find_lowest_rq(struct task_struct *task) if (this_cpu == cpu) this_cpu = -1; /* Skip this_cpu opt if the same */ - for_each_domain(cpu, sd) { - if (sd->flags & SD_WAKE_AFFINE) { - cpumask_t domain_mask; - int best_cpu; + if (alloc_cpumask_var(&domain_mask, GFP_ATOMIC)) { + for_each_domain(cpu, sd) { + if (sd->flags & SD_WAKE_AFFINE) { + int best_cpu; - cpumask_and(&domain_mask, sched_domain_span(sd), - lowest_mask); + cpumask_and(domain_mask, + sched_domain_span(sd), + lowest_mask); - best_cpu = pick_optimal_cpu(this_cpu, - &domain_mask); - if (best_cpu != -1) - return best_cpu; + best_cpu = pick_optimal_cpu(this_cpu, + domain_mask); + + if (best_cpu != -1) { + free_cpumask_var(domain_mask); + return best_cpu; + } + } } + free_cpumask_var(domain_mask); } /* @@ -1072,7 +1263,7 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq) } /* If this rq is still suitable use it. */ - if (lowest_rq->rt.highest_prio > task->prio) + if (lowest_rq->rt.highest_prio.curr > task->prio) break; /* try again */ @@ -1083,6 +1274,26 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq) return lowest_rq; } +static struct task_struct *pick_next_pushable_task(struct rq *rq) +{ + struct task_struct *p; + + if (!has_pushable_tasks(rq)) + return NULL; + + p = plist_first_entry(&rq->rt.pushable_tasks, + struct task_struct, pushable_tasks); + + BUG_ON(rq->cpu != task_cpu(p)); + BUG_ON(task_current(rq, p)); + BUG_ON(p->rt.nr_cpus_allowed <= 1); + + BUG_ON(!p->se.on_rq); + BUG_ON(!rt_task(p)); + + return p; +} + /* * If the current CPU has more than one RT task, see if the non * running task can migrate over to a CPU that is running a task @@ -1092,13 +1303,11 @@ static int push_rt_task(struct rq *rq) { struct task_struct *next_task; struct rq *lowest_rq; - int ret = 0; - int paranoid = RT_MAX_TRIES; if (!rq->rt.overloaded) return 0; - next_task = pick_next_highest_task_rt(rq, -1); + next_task = pick_next_pushable_task(rq); if (!next_task) return 0; @@ -1127,16 +1336,34 @@ static int push_rt_task(struct rq *rq) struct task_struct *task; /* * find lock_lowest_rq releases rq->lock - * so it is possible that next_task has changed. - * If it has, then try again. + * so it is possible that next_task has migrated. + * + * We need to make sure that the task is still on the same + * run-queue and is also still the next task eligible for + * pushing. */ - task = pick_next_highest_task_rt(rq, -1); - if (unlikely(task != next_task) && task && paranoid--) { - put_task_struct(next_task); - next_task = task; - goto retry; + task = pick_next_pushable_task(rq); + if (task_cpu(next_task) == rq->cpu && task == next_task) { + /* + * If we get here, the task hasnt moved at all, but + * it has failed to push. We will not try again, + * since the other cpus will pull from us when they + * are ready. + */ + dequeue_pushable_task(rq, next_task); + goto out; } - goto out; + + if (!task) + /* No more tasks, just exit */ + goto out; + + /* + * Something has shifted, try again. + */ + put_task_struct(next_task); + next_task = task; + goto retry; } deactivate_task(rq, next_task, 0); @@ -1147,23 +1374,12 @@ static int push_rt_task(struct rq *rq) double_unlock_balance(rq, lowest_rq); - ret = 1; out: put_task_struct(next_task); - return ret; + return 1; } -/* - * TODO: Currently we just use the second highest prio task on - * the queue, and stop when it can't migrate (or there's - * no more RT tasks). There may be a case where a lower - * priority RT task has a different affinity than the - * higher RT task. In this case the lower RT task could - * possibly be able to migrate where as the higher priority - * RT task could not. We currently ignore this issue. - * Enhancements are welcome! - */ static void push_rt_tasks(struct rq *rq) { /* push_rt_task will return true if it moved an RT */ @@ -1174,33 +1390,35 @@ static void push_rt_tasks(struct rq *rq) static int pull_rt_task(struct rq *this_rq) { int this_cpu = this_rq->cpu, ret = 0, cpu; - struct task_struct *p, *next; + struct task_struct *p; struct rq *src_rq; if (likely(!rt_overloaded(this_rq))) return 0; - next = pick_next_task_rt(this_rq); - for_each_cpu(cpu, this_rq->rd->rto_mask) { if (this_cpu == cpu) continue; src_rq = cpu_rq(cpu); + + /* + * Don't bother taking the src_rq->lock if the next highest + * task is known to be lower-priority than our current task. + * This may look racy, but if this value is about to go + * logically higher, the src_rq will push this task away. + * And if its going logically lower, we do not care + */ + if (src_rq->rt.highest_prio.next >= + this_rq->rt.highest_prio.curr) + continue; + /* * We can potentially drop this_rq's lock in * double_lock_balance, and another CPU could - * steal our next task - hence we must cause - * the caller to recalculate the next task - * in that case: + * alter this_rq */ - if (double_lock_balance(this_rq, src_rq)) { - struct task_struct *old_next = next; - - next = pick_next_task_rt(this_rq); - if (next != old_next) - ret = 1; - } + double_lock_balance(this_rq, src_rq); /* * Are there still pullable RT tasks? @@ -1214,7 +1432,7 @@ static int pull_rt_task(struct rq *this_rq) * Do we have an RT task that preempts * the to-be-scheduled task? */ - if (p && (!next || (p->prio < next->prio))) { + if (p && (p->prio < this_rq->rt.highest_prio.curr)) { WARN_ON(p == src_rq->curr); WARN_ON(!p->se.on_rq); @@ -1224,12 +1442,9 @@ static int pull_rt_task(struct rq *this_rq) * This is just that p is wakeing up and hasn't * had a chance to schedule. We only pull * p if it is lower in priority than the - * current task on the run queue or - * this_rq next task is lower in prio than - * the current task on that rq. + * current task on the run queue */ - if (p->prio < src_rq->curr->prio || - (next && next->prio < src_rq->curr->prio)) + if (p->prio < src_rq->curr->prio) goto skip; ret = 1; @@ -1242,13 +1457,7 @@ static int pull_rt_task(struct rq *this_rq) * case there's an even higher prio task * in another runqueue. (low likelyhood * but possible) - * - * Update next so that we won't pick a task - * on another cpu with a priority lower (or equal) - * than the one we just picked. */ - next = p; - } skip: double_unlock_balance(this_rq, src_rq); @@ -1260,24 +1469,13 @@ static int pull_rt_task(struct rq *this_rq) static void pre_schedule_rt(struct rq *rq, struct task_struct *prev) { /* Try to pull RT tasks here if we lower this rq's prio */ - if (unlikely(rt_task(prev)) && rq->rt.highest_prio > prev->prio) + if (unlikely(rt_task(prev)) && rq->rt.highest_prio.curr > prev->prio) pull_rt_task(rq); } static void post_schedule_rt(struct rq *rq) { - /* - * If we have more than one rt_task queued, then - * see if we can push the other rt_tasks off to other CPUS. - * Note we may release the rq lock, and since - * the lock was owned by prev, we need to release it - * first via finish_lock_switch and then reaquire it here. - */ - if (unlikely(rq->rt.overloaded)) { - spin_lock_irq(&rq->lock); - push_rt_tasks(rq); - spin_unlock_irq(&rq->lock); - } + push_rt_tasks(rq); } /* @@ -1288,7 +1486,8 @@ static void task_wake_up_rt(struct rq *rq, struct task_struct *p) { if (!task_running(rq, p) && !test_tsk_need_resched(rq->curr) && - rq->rt.overloaded) + has_pushable_tasks(rq) && + p->rt.nr_cpus_allowed > 1) push_rt_tasks(rq); } @@ -1324,6 +1523,24 @@ static void set_cpus_allowed_rt(struct task_struct *p, if (p->se.on_rq && (weight != p->rt.nr_cpus_allowed)) { struct rq *rq = task_rq(p); + if (!task_current(rq, p)) { + /* + * Make sure we dequeue this task from the pushable list + * before going further. It will either remain off of + * the list because we are no longer pushable, or it + * will be requeued. + */ + if (p->rt.nr_cpus_allowed > 1) + dequeue_pushable_task(rq, p); + + /* + * Requeue if our weight is changing and still > 1 + */ + if (weight > 1) + enqueue_pushable_task(rq, p); + + } + if ((p->rt.nr_cpus_allowed <= 1) && (weight > 1)) { rq->rt.rt_nr_migratory++; } else if ((p->rt.nr_cpus_allowed > 1) && (weight <= 1)) { @@ -1331,7 +1548,7 @@ static void set_cpus_allowed_rt(struct task_struct *p, rq->rt.rt_nr_migratory--; } - update_rt_migration(rq); + update_rt_migration(&rq->rt); } cpumask_copy(&p->cpus_allowed, new_mask); @@ -1346,7 +1563,7 @@ static void rq_online_rt(struct rq *rq) __enable_runtime(rq); - cpupri_set(&rq->rd->cpupri, rq->cpu, rq->rt.highest_prio); + cpupri_set(&rq->rd->cpupri, rq->cpu, rq->rt.highest_prio.curr); } /* Assumes rq->lock is held */ @@ -1383,7 +1600,7 @@ static inline void init_sched_rt_class(void) unsigned int i; for_each_possible_cpu(i) - alloc_cpumask_var_node(&per_cpu(local_cpu_mask, i), + zalloc_cpumask_var_node(&per_cpu(local_cpu_mask, i), GFP_KERNEL, cpu_to_node(i)); } #endif /* CONFIG_SMP */ @@ -1438,7 +1655,7 @@ static void prio_changed_rt(struct rq *rq, struct task_struct *p, * can release the rq lock and p could migrate. * Only reschedule if p is still on the same runqueue. */ - if (p->prio > rq->rt.highest_prio && rq->curr == p) + if (p->prio > rq->rt.highest_prio.curr && rq->curr == p) resched_task(p); #else /* For UP simply resched on drop of prio */ @@ -1509,6 +1726,9 @@ static void set_curr_task_rt(struct rq *rq) struct task_struct *p = rq->curr; p->se.exec_start = rq->clock; + + /* The running task is never eligible for pushing */ + dequeue_pushable_task(rq, p); } static const struct sched_class rt_sched_class = { diff --git a/kernel/sched_stats.h b/kernel/sched_stats.h index f2773b5d122..32d2bd4061b 100644 --- a/kernel/sched_stats.h +++ b/kernel/sched_stats.h @@ -4,7 +4,7 @@ * bump this up when changing the output format or the meaning of an existing * format, so that tools can adapt (or abort) */ -#define SCHEDSTAT_VERSION 14 +#define SCHEDSTAT_VERSION 15 static int show_schedstat(struct seq_file *seq, void *v) { @@ -26,9 +26,8 @@ static int show_schedstat(struct seq_file *seq, void *v) /* runqueue-specific stats */ seq_printf(seq, - "cpu%d %u %u %u %u %u %u %u %u %u %llu %llu %lu", - cpu, rq->yld_both_empty, - rq->yld_act_empty, rq->yld_exp_empty, rq->yld_count, + "cpu%d %u %u %u %u %u %u %llu %llu %lu", + cpu, rq->yld_count, rq->sched_switch, rq->sched_count, rq->sched_goidle, rq->ttwu_count, rq->ttwu_local, rq->rq_cpu_time, @@ -296,20 +295,21 @@ sched_info_switch(struct task_struct *prev, struct task_struct *next) static inline void account_group_user_time(struct task_struct *tsk, cputime_t cputime) { - struct signal_struct *sig; + struct thread_group_cputimer *cputimer; /* tsk == current, ensure it is safe to use ->signal */ if (unlikely(tsk->exit_state)) return; - sig = tsk->signal; - if (sig->cputime.totals) { - struct task_cputime *times; + cputimer = &tsk->signal->cputimer; - times = per_cpu_ptr(sig->cputime.totals, get_cpu()); - times->utime = cputime_add(times->utime, cputime); - put_cpu_no_resched(); - } + if (!cputimer->running) + return; + + spin_lock(&cputimer->lock); + cputimer->cputime.utime = + cputime_add(cputimer->cputime.utime, cputime); + spin_unlock(&cputimer->lock); } /** @@ -325,20 +325,21 @@ static inline void account_group_user_time(struct task_struct *tsk, static inline void account_group_system_time(struct task_struct *tsk, cputime_t cputime) { - struct signal_struct *sig; + struct thread_group_cputimer *cputimer; /* tsk == current, ensure it is safe to use ->signal */ if (unlikely(tsk->exit_state)) return; - sig = tsk->signal; - if (sig->cputime.totals) { - struct task_cputime *times; + cputimer = &tsk->signal->cputimer; - times = per_cpu_ptr(sig->cputime.totals, get_cpu()); - times->stime = cputime_add(times->stime, cputime); - put_cpu_no_resched(); - } + if (!cputimer->running) + return; + + spin_lock(&cputimer->lock); + cputimer->cputime.stime = + cputime_add(cputimer->cputime.stime, cputime); + spin_unlock(&cputimer->lock); } /** @@ -354,6 +355,7 @@ static inline void account_group_system_time(struct task_struct *tsk, static inline void account_group_exec_runtime(struct task_struct *tsk, unsigned long long ns) { + struct thread_group_cputimer *cputimer; struct signal_struct *sig; sig = tsk->signal; @@ -362,11 +364,12 @@ static inline void account_group_exec_runtime(struct task_struct *tsk, if (unlikely(!sig)) return; - if (sig->cputime.totals) { - struct task_cputime *times; + cputimer = &sig->cputimer; - times = per_cpu_ptr(sig->cputime.totals, get_cpu()); - times->sum_exec_runtime += ns; - put_cpu_no_resched(); - } + if (!cputimer->running) + return; + + spin_lock(&cputimer->lock); + cputimer->cputime.sum_exec_runtime += ns; + spin_unlock(&cputimer->lock); } diff --git a/kernel/seccomp.c b/kernel/seccomp.c index ad64fcb731f..57d4b13b631 100644 --- a/kernel/seccomp.c +++ b/kernel/seccomp.c @@ -8,6 +8,7 @@ #include <linux/seccomp.h> #include <linux/sched.h> +#include <linux/compat.h> /* #define SECCOMP_DEBUG 1 */ #define NR_SECCOMP_MODES 1 @@ -22,7 +23,7 @@ static int mode1_syscalls[] = { 0, /* null terminated */ }; -#ifdef TIF_32BIT +#ifdef CONFIG_COMPAT static int mode1_syscalls_32[] = { __NR_seccomp_read_32, __NR_seccomp_write_32, __NR_seccomp_exit_32, __NR_seccomp_sigreturn_32, 0, /* null terminated */ @@ -37,8 +38,8 @@ void __secure_computing(int this_syscall) switch (mode) { case 1: syscall = mode1_syscalls; -#ifdef TIF_32BIT - if (test_thread_flag(TIF_32BIT)) +#ifdef CONFIG_COMPAT + if (is_compat_task()) syscall = mode1_syscalls_32; #endif do { diff --git a/kernel/signal.c b/kernel/signal.c index e73759783dc..64c5deeaca5 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -27,7 +27,7 @@ #include <linux/freezer.h> #include <linux/pid_namespace.h> #include <linux/nsproxy.h> -#include <trace/sched.h> +#include <trace/events/sched.h> #include <asm/param.h> #include <asm/uaccess.h> @@ -41,8 +41,6 @@ static struct kmem_cache *sigqueue_cachep; -DEFINE_TRACE(sched_signal_send); - static void __user *sig_handler(struct task_struct *t, int sig) { return t->sighand->action[sig - 1].sa.sa_handler; @@ -55,10 +53,22 @@ static int sig_handler_ignored(void __user *handler, int sig) (handler == SIG_DFL && sig_kernel_ignore(sig)); } -static int sig_ignored(struct task_struct *t, int sig) +static int sig_task_ignored(struct task_struct *t, int sig, + int from_ancestor_ns) { void __user *handler; + handler = sig_handler(t, sig); + + if (unlikely(t->signal->flags & SIGNAL_UNKILLABLE) && + handler == SIG_DFL && !from_ancestor_ns) + return 1; + + return sig_handler_ignored(handler, sig); +} + +static int sig_ignored(struct task_struct *t, int sig, int from_ancestor_ns) +{ /* * Blocked signals are never ignored, since the * signal handler may change by the time it is @@ -67,14 +77,13 @@ static int sig_ignored(struct task_struct *t, int sig) if (sigismember(&t->blocked, sig) || sigismember(&t->real_blocked, sig)) return 0; - handler = sig_handler(t, sig); - if (!sig_handler_ignored(handler, sig)) + if (!sig_task_ignored(t, sig, from_ancestor_ns)) return 0; /* * Tracers may want to know about even ignored signals. */ - return !tracehook_consider_ignored_signal(t, sig, handler); + return !tracehook_consider_ignored_signal(t, sig); } /* @@ -238,14 +247,19 @@ void flush_sigqueue(struct sigpending *queue) /* * Flush all pending signals for a task. */ +void __flush_signals(struct task_struct *t) +{ + clear_tsk_thread_flag(t, TIF_SIGPENDING); + flush_sigqueue(&t->pending); + flush_sigqueue(&t->signal->shared_pending); +} + void flush_signals(struct task_struct *t) { unsigned long flags; spin_lock_irqsave(&t->sighand->siglock, flags); - clear_tsk_thread_flag(t, TIF_SIGPENDING); - flush_sigqueue(&t->pending); - flush_sigqueue(&t->signal->shared_pending); + __flush_signals(t); spin_unlock_irqrestore(&t->sighand->siglock, flags); } @@ -318,7 +332,7 @@ int unhandled_signal(struct task_struct *tsk, int sig) return 1; if (handler != SIG_IGN && handler != SIG_DFL) return 0; - return !tracehook_consider_fatal_signal(tsk, sig, handler); + return !tracehook_consider_fatal_signal(tsk, sig); } @@ -624,7 +638,7 @@ static int check_kill_permission(int sig, struct siginfo *info, * Returns true if the signal should be actually delivered, otherwise * it should be dropped. */ -static int prepare_signal(int sig, struct task_struct *p) +static int prepare_signal(int sig, struct task_struct *p, int from_ancestor_ns) { struct signal_struct *signal = p->signal; struct task_struct *t; @@ -708,7 +722,7 @@ static int prepare_signal(int sig, struct task_struct *p) } } - return !sig_ignored(p, sig); + return !sig_ignored(p, sig, from_ancestor_ns); } /* @@ -777,7 +791,7 @@ static void complete_signal(int sig, struct task_struct *p, int group) !(signal->flags & (SIGNAL_UNKILLABLE | SIGNAL_GROUP_EXIT)) && !sigismember(&t->real_blocked, sig) && (sig == SIGKILL || - !tracehook_consider_fatal_signal(t, sig, SIG_DFL))) { + !tracehook_consider_fatal_signal(t, sig))) { /* * This signal will be fatal to the whole group. */ @@ -813,16 +827,18 @@ static inline int legacy_queue(struct sigpending *signals, int sig) return (sig < SIGRTMIN) && sigismember(&signals->signal, sig); } -static int send_signal(int sig, struct siginfo *info, struct task_struct *t, - int group) +static int __send_signal(int sig, struct siginfo *info, struct task_struct *t, + int group, int from_ancestor_ns) { struct sigpending *pending; struct sigqueue *q; + int override_rlimit; trace_sched_signal_send(sig, t); assert_spin_locked(&t->sighand->siglock); - if (!prepare_signal(sig, t)) + + if (!prepare_signal(sig, t, from_ancestor_ns)) return 0; pending = group ? &t->signal->shared_pending : &t->pending; @@ -848,9 +864,13 @@ static int send_signal(int sig, struct siginfo *info, struct task_struct *t, make sure at least one signal gets delivered and don't pass on the info struct. */ - q = __sigqueue_alloc(t, GFP_ATOMIC, (sig < SIGRTMIN && - (is_si_special(info) || - info->si_code >= 0))); + if (sig < SIGRTMIN) + override_rlimit = (is_si_special(info) || info->si_code >= 0); + else + override_rlimit = 0; + + q = __sigqueue_alloc(t, GFP_ATOMIC | __GFP_NOTRACK_FALSE_POSITIVE, + override_rlimit); if (q) { list_add_tail(&q->list, &pending->list); switch ((unsigned long) info) { @@ -871,6 +891,8 @@ static int send_signal(int sig, struct siginfo *info, struct task_struct *t, break; default: copy_siginfo(&q->info, info); + if (from_ancestor_ns) + q->info.si_pid = 0; break; } } else if (!is_si_special(info)) { @@ -889,6 +911,20 @@ out_set: return 0; } +static int send_signal(int sig, struct siginfo *info, struct task_struct *t, + int group) +{ + int from_ancestor_ns = 0; + +#ifdef CONFIG_PID_NS + if (!is_si_special(info) && SI_FROMUSER(info) && + task_pid_nr_ns(current, task_active_pid_ns(t)) <= 0) + from_ancestor_ns = 1; +#endif + + return __send_signal(sig, info, t, group, from_ancestor_ns); +} + int print_fatal_signals; static void print_fatal_signal(struct pt_regs *regs, int signr) @@ -909,7 +945,9 @@ static void print_fatal_signal(struct pt_regs *regs, int signr) } #endif printk("\n"); + preempt_disable(); show_regs(regs); + preempt_enable(); } static int __init setup_print_fatal_signals(char *str) @@ -1131,7 +1169,7 @@ int kill_pid_info_as_uid(int sig, struct siginfo *info, struct pid *pid, if (sig && p->sighand) { unsigned long flags; spin_lock_irqsave(&p->sighand->siglock, flags); - ret = __group_send_sig_info(sig, info, p); + ret = __send_signal(sig, info, p, 1, 0); spin_unlock_irqrestore(&p->sighand->siglock, flags); } out_unlock: @@ -1318,7 +1356,7 @@ int send_sigqueue(struct sigqueue *q, struct task_struct *t, int group) goto ret; ret = 1; /* the signal is ignored */ - if (!prepare_signal(sig, t)) + if (!prepare_signal(sig, t, 0)) goto out; ret = 0; @@ -1365,7 +1403,6 @@ int do_notify_parent(struct task_struct *tsk, int sig) struct siginfo info; unsigned long flags; struct sighand_struct *psig; - struct task_cputime cputime; int ret = sig; BUG_ON(sig == -1); @@ -1373,7 +1410,7 @@ int do_notify_parent(struct task_struct *tsk, int sig) /* do_notify_parent_cldstop should have been called instead. */ BUG_ON(task_is_stopped_or_traced(tsk)); - BUG_ON(!tsk->ptrace && + BUG_ON(!task_ptrace(tsk) && (tsk->group_leader != tsk || !thread_group_empty(tsk))); info.si_signo = sig; @@ -1395,9 +1432,10 @@ int do_notify_parent(struct task_struct *tsk, int sig) info.si_uid = __task_cred(tsk)->uid; rcu_read_unlock(); - thread_group_cputime(tsk, &cputime); - info.si_utime = cputime_to_jiffies(cputime.utime); - info.si_stime = cputime_to_jiffies(cputime.stime); + info.si_utime = cputime_to_clock_t(cputime_add(tsk->utime, + tsk->signal->utime)); + info.si_stime = cputime_to_clock_t(cputime_add(tsk->stime, + tsk->signal->stime)); info.si_status = tsk->exit_code & 0x7f; if (tsk->exit_code & 0x80) @@ -1411,7 +1449,7 @@ int do_notify_parent(struct task_struct *tsk, int sig) psig = tsk->parent->sighand; spin_lock_irqsave(&psig->siglock, flags); - if (!tsk->ptrace && sig == SIGCHLD && + if (!task_ptrace(tsk) && sig == SIGCHLD && (psig->action[SIGCHLD-1].sa.sa_handler == SIG_IGN || (psig->action[SIGCHLD-1].sa.sa_flags & SA_NOCLDWAIT))) { /* @@ -1448,7 +1486,7 @@ static void do_notify_parent_cldstop(struct task_struct *tsk, int why) struct task_struct *parent; struct sighand_struct *sighand; - if (tsk->ptrace & PT_PTRACED) + if (task_ptrace(tsk)) parent = tsk->parent; else { tsk = tsk->group_leader; @@ -1461,7 +1499,7 @@ static void do_notify_parent_cldstop(struct task_struct *tsk, int why) * see comment in do_notify_parent() abot the following 3 lines */ rcu_read_lock(); - info.si_pid = task_pid_nr_ns(tsk, tsk->parent->nsproxy->pid_ns); + info.si_pid = task_pid_nr_ns(tsk, parent->nsproxy->pid_ns); info.si_uid = __task_cred(tsk)->uid; rcu_read_unlock(); @@ -1497,7 +1535,7 @@ static void do_notify_parent_cldstop(struct task_struct *tsk, int why) static inline int may_ptrace_stop(void) { - if (!likely(current->ptrace & PT_PTRACED)) + if (!likely(task_ptrace(current))) return 0; /* * Are we in the middle of do_coredump? @@ -1573,7 +1611,15 @@ static void ptrace_stop(int exit_code, int clear_code, siginfo_t *info) read_lock(&tasklist_lock); if (may_ptrace_stop()) { do_notify_parent_cldstop(current, CLD_TRAPPED); + /* + * Don't want to allow preemption here, because + * sys_ptrace() needs this task to be inactive. + * + * XXX: implement read_unlock_no_resched(). + */ + preempt_disable(); read_unlock(&tasklist_lock); + preempt_enable_no_resched(); schedule(); } else { /* @@ -1707,7 +1753,7 @@ static int do_signal_stop(int signr) static int ptrace_signal(int signr, siginfo_t *info, struct pt_regs *regs, void *cookie) { - if (!(current->ptrace & PT_PTRACED)) + if (!task_ptrace(current)) return signr; ptrace_signal_deliver(regs, cookie); @@ -1834,9 +1880,16 @@ relock: /* * Global init gets no signals it doesn't want. + * Container-init gets no signals it doesn't want from same + * container. + * + * Note that if global/container-init sees a sig_kernel_only() + * signal here, the signal must have been generated internally + * or must have come from an ancestor namespace. In either + * case, the signal cannot be dropped. */ if (unlikely(signal->flags & SIGNAL_UNKILLABLE) && - !signal_group_exit(signal)) + !sig_kernel_only(signr)) continue; if (sig_kernel_stop(signr)) { @@ -2233,24 +2286,17 @@ SYSCALL_DEFINE2(kill, pid_t, pid, int, sig) return kill_something_info(sig, &info, pid); } -static int do_tkill(pid_t tgid, pid_t pid, int sig) +static int +do_send_specific(pid_t tgid, pid_t pid, int sig, struct siginfo *info) { - int error; - struct siginfo info; struct task_struct *p; unsigned long flags; - - error = -ESRCH; - info.si_signo = sig; - info.si_errno = 0; - info.si_code = SI_TKILL; - info.si_pid = task_tgid_vnr(current); - info.si_uid = current_uid(); + int error = -ESRCH; rcu_read_lock(); p = find_task_by_vpid(pid); if (p && (tgid <= 0 || task_tgid_vnr(p) == tgid)) { - error = check_kill_permission(sig, &info, p); + error = check_kill_permission(sig, info, p); /* * The null signal is a permissions and process existence * probe. No signal is actually delivered. @@ -2260,7 +2306,7 @@ static int do_tkill(pid_t tgid, pid_t pid, int sig) * signal is private anyway. */ if (!error && sig && lock_task_sighand(p, &flags)) { - error = specific_send_sig_info(sig, &info, p); + error = specific_send_sig_info(sig, info, p); unlock_task_sighand(p, &flags); } } @@ -2269,6 +2315,19 @@ static int do_tkill(pid_t tgid, pid_t pid, int sig) return error; } +static int do_tkill(pid_t tgid, pid_t pid, int sig) +{ + struct siginfo info; + + info.si_signo = sig; + info.si_errno = 0; + info.si_code = SI_TKILL; + info.si_pid = task_tgid_vnr(current); + info.si_uid = current_uid(); + + return do_send_specific(tgid, pid, sig, &info); +} + /** * sys_tgkill - send signal to one specific thread * @tgid: the thread group ID of the thread @@ -2318,6 +2377,32 @@ SYSCALL_DEFINE3(rt_sigqueueinfo, pid_t, pid, int, sig, return kill_proc_info(sig, &info, pid); } +long do_rt_tgsigqueueinfo(pid_t tgid, pid_t pid, int sig, siginfo_t *info) +{ + /* This is only valid for single tasks */ + if (pid <= 0 || tgid <= 0) + return -EINVAL; + + /* Not even root can pretend to send signals from the kernel. + Nor can they impersonate a kill(), which adds source info. */ + if (info->si_code >= 0) + return -EPERM; + info->si_signo = sig; + + return do_send_specific(tgid, pid, sig, info); +} + +SYSCALL_DEFINE4(rt_tgsigqueueinfo, pid_t, tgid, pid_t, pid, int, sig, + siginfo_t __user *, uinfo) +{ + siginfo_t info; + + if (copy_from_user(&info, uinfo, sizeof(siginfo_t))) + return -EFAULT; + + return do_rt_tgsigqueueinfo(tgid, pid, sig, &info); +} + int do_sigaction(int sig, struct k_sigaction *act, struct k_sigaction *oact) { struct task_struct *t = current; @@ -2369,11 +2454,9 @@ do_sigaltstack (const stack_t __user *uss, stack_t __user *uoss, unsigned long s stack_t oss; int error; - if (uoss) { - oss.ss_sp = (void __user *) current->sas_ss_sp; - oss.ss_size = current->sas_ss_size; - oss.ss_flags = sas_ss_flags(sp); - } + oss.ss_sp = (void __user *) current->sas_ss_sp; + oss.ss_size = current->sas_ss_size; + oss.ss_flags = sas_ss_flags(sp); if (uss) { void __user *ss_sp; @@ -2381,10 +2464,12 @@ do_sigaltstack (const stack_t __user *uss, stack_t __user *uoss, unsigned long s int ss_flags; error = -EFAULT; - if (!access_ok(VERIFY_READ, uss, sizeof(*uss)) - || __get_user(ss_sp, &uss->ss_sp) - || __get_user(ss_flags, &uss->ss_flags) - || __get_user(ss_size, &uss->ss_size)) + if (!access_ok(VERIFY_READ, uss, sizeof(*uss))) + goto out; + error = __get_user(ss_sp, &uss->ss_sp) | + __get_user(ss_flags, &uss->ss_flags) | + __get_user(ss_size, &uss->ss_size); + if (error) goto out; error = -EPERM; @@ -2416,13 +2501,16 @@ do_sigaltstack (const stack_t __user *uss, stack_t __user *uoss, unsigned long s current->sas_ss_size = ss_size; } + error = 0; if (uoss) { error = -EFAULT; - if (copy_to_user(uoss, &oss, sizeof(oss))) + if (!access_ok(VERIFY_WRITE, uoss, sizeof(*uoss))) goto out; + error = __put_user(oss.ss_sp, &uoss->ss_sp) | + __put_user(oss.ss_size, &uoss->ss_size) | + __put_user(oss.ss_flags, &uoss->ss_flags); } - error = 0; out: return error; } diff --git a/kernel/slow-work.c b/kernel/slow-work.c new file mode 100644 index 00000000000..09d7519557d --- /dev/null +++ b/kernel/slow-work.c @@ -0,0 +1,645 @@ +/* Worker thread pool for slow items, such as filesystem lookups or mkdirs + * + * Copyright (C) 2008 Red Hat, Inc. All Rights Reserved. + * Written by David Howells (dhowells@redhat.com) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public Licence + * as published by the Free Software Foundation; either version + * 2 of the Licence, or (at your option) any later version. + * + * See Documentation/slow-work.txt + */ + +#include <linux/module.h> +#include <linux/slow-work.h> +#include <linux/kthread.h> +#include <linux/freezer.h> +#include <linux/wait.h> + +#define SLOW_WORK_CULL_TIMEOUT (5 * HZ) /* cull threads 5s after running out of + * things to do */ +#define SLOW_WORK_OOM_TIMEOUT (5 * HZ) /* can't start new threads for 5s after + * OOM */ + +static void slow_work_cull_timeout(unsigned long); +static void slow_work_oom_timeout(unsigned long); + +#ifdef CONFIG_SYSCTL +static int slow_work_min_threads_sysctl(struct ctl_table *, int, struct file *, + void __user *, size_t *, loff_t *); + +static int slow_work_max_threads_sysctl(struct ctl_table *, int , struct file *, + void __user *, size_t *, loff_t *); +#endif + +/* + * The pool of threads has at least min threads in it as long as someone is + * using the facility, and may have as many as max. + * + * A portion of the pool may be processing very slow operations. + */ +static unsigned slow_work_min_threads = 2; +static unsigned slow_work_max_threads = 4; +static unsigned vslow_work_proportion = 50; /* % of threads that may process + * very slow work */ + +#ifdef CONFIG_SYSCTL +static const int slow_work_min_min_threads = 2; +static int slow_work_max_max_threads = 255; +static const int slow_work_min_vslow = 1; +static const int slow_work_max_vslow = 99; + +ctl_table slow_work_sysctls[] = { + { + .ctl_name = CTL_UNNUMBERED, + .procname = "min-threads", + .data = &slow_work_min_threads, + .maxlen = sizeof(unsigned), + .mode = 0644, + .proc_handler = slow_work_min_threads_sysctl, + .extra1 = (void *) &slow_work_min_min_threads, + .extra2 = &slow_work_max_threads, + }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "max-threads", + .data = &slow_work_max_threads, + .maxlen = sizeof(unsigned), + .mode = 0644, + .proc_handler = slow_work_max_threads_sysctl, + .extra1 = &slow_work_min_threads, + .extra2 = (void *) &slow_work_max_max_threads, + }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "vslow-percentage", + .data = &vslow_work_proportion, + .maxlen = sizeof(unsigned), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .extra1 = (void *) &slow_work_min_vslow, + .extra2 = (void *) &slow_work_max_vslow, + }, + { .ctl_name = 0 } +}; +#endif + +/* + * The active state of the thread pool + */ +static atomic_t slow_work_thread_count; +static atomic_t vslow_work_executing_count; + +static bool slow_work_may_not_start_new_thread; +static bool slow_work_cull; /* cull a thread due to lack of activity */ +static DEFINE_TIMER(slow_work_cull_timer, slow_work_cull_timeout, 0, 0); +static DEFINE_TIMER(slow_work_oom_timer, slow_work_oom_timeout, 0, 0); +static struct slow_work slow_work_new_thread; /* new thread starter */ + +/* + * The queues of work items and the lock governing access to them. These are + * shared between all the CPUs. It doesn't make sense to have per-CPU queues + * as the number of threads bears no relation to the number of CPUs. + * + * There are two queues of work items: one for slow work items, and one for + * very slow work items. + */ +static LIST_HEAD(slow_work_queue); +static LIST_HEAD(vslow_work_queue); +static DEFINE_SPINLOCK(slow_work_queue_lock); + +/* + * The thread controls. A variable used to signal to the threads that they + * should exit when the queue is empty, a waitqueue used by the threads to wait + * for signals, and a completion set by the last thread to exit. + */ +static bool slow_work_threads_should_exit; +static DECLARE_WAIT_QUEUE_HEAD(slow_work_thread_wq); +static DECLARE_COMPLETION(slow_work_last_thread_exited); + +/* + * The number of users of the thread pool and its lock. Whilst this is zero we + * have no threads hanging around, and when this reaches zero, we wait for all + * active or queued work items to complete and kill all the threads we do have. + */ +static int slow_work_user_count; +static DEFINE_MUTEX(slow_work_user_lock); + +/* + * Calculate the maximum number of active threads in the pool that are + * permitted to process very slow work items. + * + * The answer is rounded up to at least 1, but may not equal or exceed the + * maximum number of the threads in the pool. This means we always have at + * least one thread that can process slow work items, and we always have at + * least one thread that won't get tied up doing so. + */ +static unsigned slow_work_calc_vsmax(void) +{ + unsigned vsmax; + + vsmax = atomic_read(&slow_work_thread_count) * vslow_work_proportion; + vsmax /= 100; + vsmax = max(vsmax, 1U); + return min(vsmax, slow_work_max_threads - 1); +} + +/* + * Attempt to execute stuff queued on a slow thread. Return true if we managed + * it, false if there was nothing to do. + */ +static bool slow_work_execute(void) +{ + struct slow_work *work = NULL; + unsigned vsmax; + bool very_slow; + + vsmax = slow_work_calc_vsmax(); + + /* see if we can schedule a new thread to be started if we're not + * keeping up with the work */ + if (!waitqueue_active(&slow_work_thread_wq) && + (!list_empty(&slow_work_queue) || !list_empty(&vslow_work_queue)) && + atomic_read(&slow_work_thread_count) < slow_work_max_threads && + !slow_work_may_not_start_new_thread) + slow_work_enqueue(&slow_work_new_thread); + + /* find something to execute */ + spin_lock_irq(&slow_work_queue_lock); + if (!list_empty(&vslow_work_queue) && + atomic_read(&vslow_work_executing_count) < vsmax) { + work = list_entry(vslow_work_queue.next, + struct slow_work, link); + if (test_and_set_bit_lock(SLOW_WORK_EXECUTING, &work->flags)) + BUG(); + list_del_init(&work->link); + atomic_inc(&vslow_work_executing_count); + very_slow = true; + } else if (!list_empty(&slow_work_queue)) { + work = list_entry(slow_work_queue.next, + struct slow_work, link); + if (test_and_set_bit_lock(SLOW_WORK_EXECUTING, &work->flags)) + BUG(); + list_del_init(&work->link); + very_slow = false; + } else { + very_slow = false; /* avoid the compiler warning */ + } + spin_unlock_irq(&slow_work_queue_lock); + + if (!work) + return false; + + if (!test_and_clear_bit(SLOW_WORK_PENDING, &work->flags)) + BUG(); + + work->ops->execute(work); + + if (very_slow) + atomic_dec(&vslow_work_executing_count); + clear_bit_unlock(SLOW_WORK_EXECUTING, &work->flags); + + /* if someone tried to enqueue the item whilst we were executing it, + * then it'll be left unenqueued to avoid multiple threads trying to + * execute it simultaneously + * + * there is, however, a race between us testing the pending flag and + * getting the spinlock, and between the enqueuer setting the pending + * flag and getting the spinlock, so we use a deferral bit to tell us + * if the enqueuer got there first + */ + if (test_bit(SLOW_WORK_PENDING, &work->flags)) { + spin_lock_irq(&slow_work_queue_lock); + + if (!test_bit(SLOW_WORK_EXECUTING, &work->flags) && + test_and_clear_bit(SLOW_WORK_ENQ_DEFERRED, &work->flags)) + goto auto_requeue; + + spin_unlock_irq(&slow_work_queue_lock); + } + + work->ops->put_ref(work); + return true; + +auto_requeue: + /* we must complete the enqueue operation + * - we transfer our ref on the item back to the appropriate queue + * - don't wake another thread up as we're awake already + */ + if (test_bit(SLOW_WORK_VERY_SLOW, &work->flags)) + list_add_tail(&work->link, &vslow_work_queue); + else + list_add_tail(&work->link, &slow_work_queue); + spin_unlock_irq(&slow_work_queue_lock); + return true; +} + +/** + * slow_work_enqueue - Schedule a slow work item for processing + * @work: The work item to queue + * + * Schedule a slow work item for processing. If the item is already undergoing + * execution, this guarantees not to re-enter the execution routine until the + * first execution finishes. + * + * The item is pinned by this function as it retains a reference to it, managed + * through the item operations. The item is unpinned once it has been + * executed. + * + * An item may hog the thread that is running it for a relatively large amount + * of time, sufficient, for example, to perform several lookup, mkdir, create + * and setxattr operations. It may sleep on I/O and may sleep to obtain locks. + * + * Conversely, if a number of items are awaiting processing, it may take some + * time before any given item is given attention. The number of threads in the + * pool may be increased to deal with demand, but only up to a limit. + * + * If SLOW_WORK_VERY_SLOW is set on the work item, then it will be placed in + * the very slow queue, from which only a portion of the threads will be + * allowed to pick items to execute. This ensures that very slow items won't + * overly block ones that are just ordinarily slow. + * + * Returns 0 if successful, -EAGAIN if not. + */ +int slow_work_enqueue(struct slow_work *work) +{ + unsigned long flags; + + BUG_ON(slow_work_user_count <= 0); + BUG_ON(!work); + BUG_ON(!work->ops); + BUG_ON(!work->ops->get_ref); + + /* when honouring an enqueue request, we only promise that we will run + * the work function in the future; we do not promise to run it once + * per enqueue request + * + * we use the PENDING bit to merge together repeat requests without + * having to disable IRQs and take the spinlock, whilst still + * maintaining our promise + */ + if (!test_and_set_bit_lock(SLOW_WORK_PENDING, &work->flags)) { + spin_lock_irqsave(&slow_work_queue_lock, flags); + + /* we promise that we will not attempt to execute the work + * function in more than one thread simultaneously + * + * this, however, leaves us with a problem if we're asked to + * enqueue the work whilst someone is executing the work + * function as simply queueing the work immediately means that + * another thread may try executing it whilst it is already + * under execution + * + * to deal with this, we set the ENQ_DEFERRED bit instead of + * enqueueing, and the thread currently executing the work + * function will enqueue the work item when the work function + * returns and it has cleared the EXECUTING bit + */ + if (test_bit(SLOW_WORK_EXECUTING, &work->flags)) { + set_bit(SLOW_WORK_ENQ_DEFERRED, &work->flags); + } else { + if (work->ops->get_ref(work) < 0) + goto cant_get_ref; + if (test_bit(SLOW_WORK_VERY_SLOW, &work->flags)) + list_add_tail(&work->link, &vslow_work_queue); + else + list_add_tail(&work->link, &slow_work_queue); + wake_up(&slow_work_thread_wq); + } + + spin_unlock_irqrestore(&slow_work_queue_lock, flags); + } + return 0; + +cant_get_ref: + spin_unlock_irqrestore(&slow_work_queue_lock, flags); + return -EAGAIN; +} +EXPORT_SYMBOL(slow_work_enqueue); + +/* + * Schedule a cull of the thread pool at some time in the near future + */ +static void slow_work_schedule_cull(void) +{ + mod_timer(&slow_work_cull_timer, + round_jiffies(jiffies + SLOW_WORK_CULL_TIMEOUT)); +} + +/* + * Worker thread culling algorithm + */ +static bool slow_work_cull_thread(void) +{ + unsigned long flags; + bool do_cull = false; + + spin_lock_irqsave(&slow_work_queue_lock, flags); + + if (slow_work_cull) { + slow_work_cull = false; + + if (list_empty(&slow_work_queue) && + list_empty(&vslow_work_queue) && + atomic_read(&slow_work_thread_count) > + slow_work_min_threads) { + slow_work_schedule_cull(); + do_cull = true; + } + } + + spin_unlock_irqrestore(&slow_work_queue_lock, flags); + return do_cull; +} + +/* + * Determine if there is slow work available for dispatch + */ +static inline bool slow_work_available(int vsmax) +{ + return !list_empty(&slow_work_queue) || + (!list_empty(&vslow_work_queue) && + atomic_read(&vslow_work_executing_count) < vsmax); +} + +/* + * Worker thread dispatcher + */ +static int slow_work_thread(void *_data) +{ + int vsmax; + + DEFINE_WAIT(wait); + + set_freezable(); + set_user_nice(current, -5); + + for (;;) { + vsmax = vslow_work_proportion; + vsmax *= atomic_read(&slow_work_thread_count); + vsmax /= 100; + + prepare_to_wait_exclusive(&slow_work_thread_wq, &wait, + TASK_INTERRUPTIBLE); + if (!freezing(current) && + !slow_work_threads_should_exit && + !slow_work_available(vsmax) && + !slow_work_cull) + schedule(); + finish_wait(&slow_work_thread_wq, &wait); + + try_to_freeze(); + + vsmax = vslow_work_proportion; + vsmax *= atomic_read(&slow_work_thread_count); + vsmax /= 100; + + if (slow_work_available(vsmax) && slow_work_execute()) { + cond_resched(); + if (list_empty(&slow_work_queue) && + list_empty(&vslow_work_queue) && + atomic_read(&slow_work_thread_count) > + slow_work_min_threads) + slow_work_schedule_cull(); + continue; + } + + if (slow_work_threads_should_exit) + break; + + if (slow_work_cull && slow_work_cull_thread()) + break; + } + + if (atomic_dec_and_test(&slow_work_thread_count)) + complete_and_exit(&slow_work_last_thread_exited, 0); + return 0; +} + +/* + * Handle thread cull timer expiration + */ +static void slow_work_cull_timeout(unsigned long data) +{ + slow_work_cull = true; + wake_up(&slow_work_thread_wq); +} + +/* + * Get a reference on slow work thread starter + */ +static int slow_work_new_thread_get_ref(struct slow_work *work) +{ + return 0; +} + +/* + * Drop a reference on slow work thread starter + */ +static void slow_work_new_thread_put_ref(struct slow_work *work) +{ +} + +/* + * Start a new slow work thread + */ +static void slow_work_new_thread_execute(struct slow_work *work) +{ + struct task_struct *p; + + if (slow_work_threads_should_exit) + return; + + if (atomic_read(&slow_work_thread_count) >= slow_work_max_threads) + return; + + if (!mutex_trylock(&slow_work_user_lock)) + return; + + slow_work_may_not_start_new_thread = true; + atomic_inc(&slow_work_thread_count); + p = kthread_run(slow_work_thread, NULL, "kslowd"); + if (IS_ERR(p)) { + printk(KERN_DEBUG "Slow work thread pool: OOM\n"); + if (atomic_dec_and_test(&slow_work_thread_count)) + BUG(); /* we're running on a slow work thread... */ + mod_timer(&slow_work_oom_timer, + round_jiffies(jiffies + SLOW_WORK_OOM_TIMEOUT)); + } else { + /* ratelimit the starting of new threads */ + mod_timer(&slow_work_oom_timer, jiffies + 1); + } + + mutex_unlock(&slow_work_user_lock); +} + +static const struct slow_work_ops slow_work_new_thread_ops = { + .get_ref = slow_work_new_thread_get_ref, + .put_ref = slow_work_new_thread_put_ref, + .execute = slow_work_new_thread_execute, +}; + +/* + * post-OOM new thread start suppression expiration + */ +static void slow_work_oom_timeout(unsigned long data) +{ + slow_work_may_not_start_new_thread = false; +} + +#ifdef CONFIG_SYSCTL +/* + * Handle adjustment of the minimum number of threads + */ +static int slow_work_min_threads_sysctl(struct ctl_table *table, int write, + struct file *filp, void __user *buffer, + size_t *lenp, loff_t *ppos) +{ + int ret = proc_dointvec_minmax(table, write, filp, buffer, lenp, ppos); + int n; + + if (ret == 0) { + mutex_lock(&slow_work_user_lock); + if (slow_work_user_count > 0) { + /* see if we need to start or stop threads */ + n = atomic_read(&slow_work_thread_count) - + slow_work_min_threads; + + if (n < 0 && !slow_work_may_not_start_new_thread) + slow_work_enqueue(&slow_work_new_thread); + else if (n > 0) + slow_work_schedule_cull(); + } + mutex_unlock(&slow_work_user_lock); + } + + return ret; +} + +/* + * Handle adjustment of the maximum number of threads + */ +static int slow_work_max_threads_sysctl(struct ctl_table *table, int write, + struct file *filp, void __user *buffer, + size_t *lenp, loff_t *ppos) +{ + int ret = proc_dointvec_minmax(table, write, filp, buffer, lenp, ppos); + int n; + + if (ret == 0) { + mutex_lock(&slow_work_user_lock); + if (slow_work_user_count > 0) { + /* see if we need to stop threads */ + n = slow_work_max_threads - + atomic_read(&slow_work_thread_count); + + if (n < 0) + slow_work_schedule_cull(); + } + mutex_unlock(&slow_work_user_lock); + } + + return ret; +} +#endif /* CONFIG_SYSCTL */ + +/** + * slow_work_register_user - Register a user of the facility + * + * Register a user of the facility, starting up the initial threads if there + * aren't any other users at this point. This will return 0 if successful, or + * an error if not. + */ +int slow_work_register_user(void) +{ + struct task_struct *p; + int loop; + + mutex_lock(&slow_work_user_lock); + + if (slow_work_user_count == 0) { + printk(KERN_NOTICE "Slow work thread pool: Starting up\n"); + init_completion(&slow_work_last_thread_exited); + + slow_work_threads_should_exit = false; + slow_work_init(&slow_work_new_thread, + &slow_work_new_thread_ops); + slow_work_may_not_start_new_thread = false; + slow_work_cull = false; + + /* start the minimum number of threads */ + for (loop = 0; loop < slow_work_min_threads; loop++) { + atomic_inc(&slow_work_thread_count); + p = kthread_run(slow_work_thread, NULL, "kslowd"); + if (IS_ERR(p)) + goto error; + } + printk(KERN_NOTICE "Slow work thread pool: Ready\n"); + } + + slow_work_user_count++; + mutex_unlock(&slow_work_user_lock); + return 0; + +error: + if (atomic_dec_and_test(&slow_work_thread_count)) + complete(&slow_work_last_thread_exited); + if (loop > 0) { + printk(KERN_ERR "Slow work thread pool:" + " Aborting startup on ENOMEM\n"); + slow_work_threads_should_exit = true; + wake_up_all(&slow_work_thread_wq); + wait_for_completion(&slow_work_last_thread_exited); + printk(KERN_ERR "Slow work thread pool: Aborted\n"); + } + mutex_unlock(&slow_work_user_lock); + return PTR_ERR(p); +} +EXPORT_SYMBOL(slow_work_register_user); + +/** + * slow_work_unregister_user - Unregister a user of the facility + * + * Unregister a user of the facility, killing all the threads if this was the + * last one. + */ +void slow_work_unregister_user(void) +{ + mutex_lock(&slow_work_user_lock); + + BUG_ON(slow_work_user_count <= 0); + + slow_work_user_count--; + if (slow_work_user_count == 0) { + printk(KERN_NOTICE "Slow work thread pool: Shutting down\n"); + slow_work_threads_should_exit = true; + del_timer_sync(&slow_work_cull_timer); + del_timer_sync(&slow_work_oom_timer); + wake_up_all(&slow_work_thread_wq); + wait_for_completion(&slow_work_last_thread_exited); + printk(KERN_NOTICE "Slow work thread pool:" + " Shut down complete\n"); + } + + mutex_unlock(&slow_work_user_lock); +} +EXPORT_SYMBOL(slow_work_unregister_user); + +/* + * Initialise the slow work facility + */ +static int __init init_slow_work(void) +{ + unsigned nr_cpus = num_possible_cpus(); + + if (slow_work_max_threads < nr_cpus) + slow_work_max_threads = nr_cpus; +#ifdef CONFIG_SYSCTL + if (slow_work_max_max_threads < nr_cpus * 2) + slow_work_max_max_threads = nr_cpus * 2; +#endif + return 0; +} + +subsys_initcall(init_slow_work); diff --git a/kernel/smp.c b/kernel/smp.c index 5cfa0e5e3e8..94188b8ecc3 100644 --- a/kernel/smp.c +++ b/kernel/smp.c @@ -2,39 +2,82 @@ * Generic helpers for smp ipi calls * * (C) Jens Axboe <jens.axboe@oracle.com> 2008 - * */ -#include <linux/init.h> -#include <linux/module.h> -#include <linux/percpu.h> #include <linux/rcupdate.h> #include <linux/rculist.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/percpu.h> +#include <linux/init.h> #include <linux/smp.h> +#include <linux/cpu.h> static DEFINE_PER_CPU(struct call_single_queue, call_single_queue); -static LIST_HEAD(call_function_queue); -__cacheline_aligned_in_smp DEFINE_SPINLOCK(call_function_lock); + +static struct { + struct list_head queue; + spinlock_t lock; +} call_function __cacheline_aligned_in_smp = + { + .queue = LIST_HEAD_INIT(call_function.queue), + .lock = __SPIN_LOCK_UNLOCKED(call_function.lock), + }; enum { - CSD_FLAG_WAIT = 0x01, - CSD_FLAG_ALLOC = 0x02, + CSD_FLAG_LOCK = 0x01, }; struct call_function_data { - struct call_single_data csd; - spinlock_t lock; - unsigned int refs; - struct rcu_head rcu_head; - unsigned long cpumask_bits[]; + struct call_single_data csd; + spinlock_t lock; + unsigned int refs; + cpumask_var_t cpumask; }; struct call_single_queue { - struct list_head list; - spinlock_t lock; + struct list_head list; + spinlock_t lock; +}; + +static DEFINE_PER_CPU(struct call_function_data, cfd_data) = { + .lock = __SPIN_LOCK_UNLOCKED(cfd_data.lock), +}; + +static int +hotplug_cfd(struct notifier_block *nfb, unsigned long action, void *hcpu) +{ + long cpu = (long)hcpu; + struct call_function_data *cfd = &per_cpu(cfd_data, cpu); + + switch (action) { + case CPU_UP_PREPARE: + case CPU_UP_PREPARE_FROZEN: + if (!zalloc_cpumask_var_node(&cfd->cpumask, GFP_KERNEL, + cpu_to_node(cpu))) + return NOTIFY_BAD; + break; + +#ifdef CONFIG_HOTPLUG_CPU + case CPU_UP_CANCELED: + case CPU_UP_CANCELED_FROZEN: + + case CPU_DEAD: + case CPU_DEAD_FROZEN: + free_cpumask_var(cfd->cpumask); + break; +#endif + }; + + return NOTIFY_OK; +} + +static struct notifier_block __cpuinitdata hotplug_cfd_notifier = { + .notifier_call = hotplug_cfd, }; static int __cpuinit init_call_single_data(void) { + void *cpu = (void *)(long)smp_processor_id(); int i; for_each_possible_cpu(i) { @@ -43,29 +86,63 @@ static int __cpuinit init_call_single_data(void) spin_lock_init(&q->lock); INIT_LIST_HEAD(&q->list); } + + hotplug_cfd(&hotplug_cfd_notifier, CPU_UP_PREPARE, cpu); + register_cpu_notifier(&hotplug_cfd_notifier); + return 0; } early_initcall(init_call_single_data); -static void csd_flag_wait(struct call_single_data *data) +/* + * csd_lock/csd_unlock used to serialize access to per-cpu csd resources + * + * For non-synchronous ipi calls the csd can still be in use by the + * previous function call. For multi-cpu calls its even more interesting + * as we'll have to ensure no other cpu is observing our csd. + */ +static void csd_lock_wait(struct call_single_data *data) { - /* Wait for response */ - do { - if (!(data->flags & CSD_FLAG_WAIT)) - break; + while (data->flags & CSD_FLAG_LOCK) cpu_relax(); - } while (1); +} + +static void csd_lock(struct call_single_data *data) +{ + csd_lock_wait(data); + data->flags = CSD_FLAG_LOCK; + + /* + * prevent CPU from reordering the above assignment + * to ->flags with any subsequent assignments to other + * fields of the specified call_single_data structure: + */ + smp_mb(); +} + +static void csd_unlock(struct call_single_data *data) +{ + WARN_ON(!(data->flags & CSD_FLAG_LOCK)); + + /* + * ensure we're all done before releasing data: + */ + smp_mb(); + + data->flags &= ~CSD_FLAG_LOCK; } /* - * Insert a previously allocated call_single_data element for execution - * on the given CPU. data must already have ->func, ->info, and ->flags set. + * Insert a previously allocated call_single_data element + * for execution on the given CPU. data must already have + * ->func, ->info, and ->flags set. */ -static void generic_exec_single(int cpu, struct call_single_data *data) +static +void generic_exec_single(int cpu, struct call_single_data *data, int wait) { struct call_single_queue *dst = &per_cpu(call_single_queue, cpu); - int wait = data->flags & CSD_FLAG_WAIT, ipi; unsigned long flags; + int ipi; spin_lock_irqsave(&dst->lock, flags); ipi = list_empty(&dst->list); @@ -73,24 +150,21 @@ static void generic_exec_single(int cpu, struct call_single_data *data) spin_unlock_irqrestore(&dst->lock, flags); /* - * Make the list addition visible before sending the ipi. + * The list addition should be visible before sending the IPI + * handler locks the list to pull the entry off it because of + * normal cache coherency rules implied by spinlocks. + * + * If IPIs can go out of order to the cache coherency protocol + * in an architecture, sufficient synchronisation should be added + * to arch code to make it appear to obey cache coherency WRT + * locking and barrier primitives. Generic code isn't really + * equipped to do the right thing... */ - smp_mb(); - if (ipi) arch_send_call_function_single_ipi(cpu); if (wait) - csd_flag_wait(data); -} - -static void rcu_free_call_data(struct rcu_head *head) -{ - struct call_function_data *data; - - data = container_of(head, struct call_function_data, rcu_head); - - kfree(data); + csd_lock_wait(data); } /* @@ -103,99 +177,88 @@ void generic_smp_call_function_interrupt(void) int cpu = get_cpu(); /* - * It's ok to use list_for_each_rcu() here even though we may delete - * 'pos', since list_del_rcu() doesn't clear ->next + * Ensure entry is visible on call_function_queue after we have + * entered the IPI. See comment in smp_call_function_many. + * If we don't have this, then we may miss an entry on the list + * and never get another IPI to process it. + */ + smp_mb(); + + /* + * It's ok to use list_for_each_rcu() here even though we may + * delete 'pos', since list_del_rcu() doesn't clear ->next */ - rcu_read_lock(); - list_for_each_entry_rcu(data, &call_function_queue, csd.list) { + list_for_each_entry_rcu(data, &call_function.queue, csd.list) { int refs; - if (!cpumask_test_cpu(cpu, to_cpumask(data->cpumask_bits))) + spin_lock(&data->lock); + if (!cpumask_test_cpu(cpu, data->cpumask)) { + spin_unlock(&data->lock); continue; + } + cpumask_clear_cpu(cpu, data->cpumask); + spin_unlock(&data->lock); data->csd.func(data->csd.info); spin_lock(&data->lock); - cpumask_clear_cpu(cpu, to_cpumask(data->cpumask_bits)); WARN_ON(data->refs == 0); - data->refs--; - refs = data->refs; + refs = --data->refs; + if (!refs) { + spin_lock(&call_function.lock); + list_del_rcu(&data->csd.list); + spin_unlock(&call_function.lock); + } spin_unlock(&data->lock); if (refs) continue; - spin_lock(&call_function_lock); - list_del_rcu(&data->csd.list); - spin_unlock(&call_function_lock); - - if (data->csd.flags & CSD_FLAG_WAIT) { - /* - * serialize stores to data with the flag clear - * and wakeup - */ - smp_wmb(); - data->csd.flags &= ~CSD_FLAG_WAIT; - } - if (data->csd.flags & CSD_FLAG_ALLOC) - call_rcu(&data->rcu_head, rcu_free_call_data); + csd_unlock(&data->csd); } - rcu_read_unlock(); put_cpu(); } /* - * Invoked by arch to handle an IPI for call function single. Must be called - * from the arch with interrupts disabled. + * Invoked by arch to handle an IPI for call function single. Must be + * called from the arch with interrupts disabled. */ void generic_smp_call_function_single_interrupt(void) { struct call_single_queue *q = &__get_cpu_var(call_single_queue); + unsigned int data_flags; LIST_HEAD(list); - /* - * Need to see other stores to list head for checking whether - * list is empty without holding q->lock - */ - smp_read_barrier_depends(); - while (!list_empty(&q->list)) { - unsigned int data_flags; - - spin_lock(&q->lock); - list_replace_init(&q->list, &list); - spin_unlock(&q->lock); - - while (!list_empty(&list)) { - struct call_single_data *data; - - data = list_entry(list.next, struct call_single_data, - list); - list_del(&data->list); - - /* - * 'data' can be invalid after this call if - * flags == 0 (when called through - * generic_exec_single(), so save them away before - * making the call. - */ - data_flags = data->flags; - - data->func(data->info); - - if (data_flags & CSD_FLAG_WAIT) { - smp_wmb(); - data->flags &= ~CSD_FLAG_WAIT; - } else if (data_flags & CSD_FLAG_ALLOC) - kfree(data); - } + spin_lock(&q->lock); + list_replace_init(&q->list, &list); + spin_unlock(&q->lock); + + while (!list_empty(&list)) { + struct call_single_data *data; + + data = list_entry(list.next, struct call_single_data, list); + list_del(&data->list); + /* - * See comment on outer loop + * 'data' can be invalid after this call if flags == 0 + * (when called through generic_exec_single()), + * so save them away before making the call: */ - smp_read_barrier_depends(); + data_flags = data->flags; + + data->func(data->info); + + /* + * Unlocked CSDs are valid through generic_exec_single(): + */ + if (data_flags & CSD_FLAG_LOCK) + csd_unlock(data); } } +static DEFINE_PER_CPU(struct call_single_data, csd_data); + /* * smp_call_function_single - Run a function on a specific CPU * @func: The function to run. This must be fast and non-blocking. @@ -209,41 +272,45 @@ void generic_smp_call_function_single_interrupt(void) int smp_call_function_single(int cpu, void (*func) (void *info), void *info, int wait) { - struct call_single_data d; + struct call_single_data d = { + .flags = 0, + }; unsigned long flags; - /* prevent preemption and reschedule on another processor, - as well as CPU removal */ - int me = get_cpu(); + int this_cpu; int err = 0; + /* + * prevent preemption and reschedule on another processor, + * as well as CPU removal + */ + this_cpu = get_cpu(); + /* Can deadlock when called with interrupts disabled */ - WARN_ON(irqs_disabled()); + WARN_ON_ONCE(irqs_disabled() && !oops_in_progress); - if (cpu == me) { + if (cpu == this_cpu) { local_irq_save(flags); func(info); local_irq_restore(flags); - } else if ((unsigned)cpu < nr_cpu_ids && cpu_online(cpu)) { - struct call_single_data *data = NULL; + } else { + if ((unsigned)cpu < nr_cpu_ids && cpu_online(cpu)) { + struct call_single_data *data = &d; - if (!wait) { - data = kmalloc(sizeof(*data), GFP_ATOMIC); - if (data) - data->flags = CSD_FLAG_ALLOC; - } - if (!data) { - data = &d; - data->flags = CSD_FLAG_WAIT; - } + if (!wait) + data = &__get_cpu_var(csd_data); - data->func = func; - data->info = info; - generic_exec_single(cpu, data); - } else { - err = -ENXIO; /* CPU not online */ + csd_lock(data); + + data->func = func; + data->info = info; + generic_exec_single(cpu, data, wait); + } else { + err = -ENXIO; /* CPU not online */ + } } put_cpu(); + return err; } EXPORT_SYMBOL(smp_call_function_single); @@ -253,23 +320,26 @@ EXPORT_SYMBOL(smp_call_function_single); * @cpu: The CPU to run on. * @data: Pre-allocated and setup data structure * - * Like smp_call_function_single(), but allow caller to pass in a pre-allocated - * data structure. Useful for embedding @data inside other structures, for - * instance. - * + * Like smp_call_function_single(), but allow caller to pass in a + * pre-allocated data structure. Useful for embedding @data inside + * other structures, for instance. */ -void __smp_call_function_single(int cpu, struct call_single_data *data) +void __smp_call_function_single(int cpu, struct call_single_data *data, + int wait) { + csd_lock(data); + /* Can deadlock when called with interrupts disabled */ - WARN_ON((data->flags & CSD_FLAG_WAIT) && irqs_disabled()); + WARN_ON_ONCE(wait && irqs_disabled() && !oops_in_progress); - generic_exec_single(cpu, data); + generic_exec_single(cpu, data, wait); } -/* FIXME: Shim for archs using old arch_send_call_function_ipi API. */ +/* Deprecated: shim for archs using old arch_send_call_function_ipi API. */ + #ifndef arch_send_call_function_ipi_mask -#define arch_send_call_function_ipi_mask(maskp) \ - arch_send_call_function_ipi(*(maskp)) +# define arch_send_call_function_ipi_mask(maskp) \ + arch_send_call_function_ipi(*(maskp)) #endif /** @@ -277,7 +347,8 @@ void __smp_call_function_single(int cpu, struct call_single_data *data) * @mask: The set of cpus to run on (only runs on online subset). * @func: The function to run. This must be fast and non-blocking. * @info: An arbitrary pointer to pass to the function. - * @wait: If true, wait (atomically) until function has completed on other CPUs. + * @wait: If true, wait (atomically) until function has completed + * on other CPUs. * * If @wait is true, then returns once @func has returned. Note that @wait * will be implicitly turned on in case of allocation failures, since @@ -288,27 +359,27 @@ void __smp_call_function_single(int cpu, struct call_single_data *data) * must be disabled when calling this function. */ void smp_call_function_many(const struct cpumask *mask, - void (*func)(void *), void *info, - bool wait) + void (*func)(void *), void *info, bool wait) { struct call_function_data *data; unsigned long flags; - int cpu, next_cpu; + int cpu, next_cpu, this_cpu = smp_processor_id(); /* Can deadlock when called with interrupts disabled */ - WARN_ON(irqs_disabled()); + WARN_ON_ONCE(irqs_disabled() && !oops_in_progress); - /* So, what's a CPU they want? Ignoring this one. */ + /* So, what's a CPU they want? Ignoring this one. */ cpu = cpumask_first_and(mask, cpu_online_mask); - if (cpu == smp_processor_id()) + if (cpu == this_cpu) cpu = cpumask_next_and(cpu, mask, cpu_online_mask); + /* No online cpus? We're done. */ if (cpu >= nr_cpu_ids) return; /* Do we have another CPU which isn't us? */ next_cpu = cpumask_next_and(cpu, mask, cpu_online_mask); - if (next_cpu == smp_processor_id()) + if (next_cpu == this_cpu) next_cpu = cpumask_next_and(next_cpu, mask, cpu_online_mask); /* Fastpath: do that cpu by itself. */ @@ -317,43 +388,40 @@ void smp_call_function_many(const struct cpumask *mask, return; } - data = kmalloc(sizeof(*data) + cpumask_size(), GFP_ATOMIC); - if (unlikely(!data)) { - /* Slow path. */ - for_each_online_cpu(cpu) { - if (cpu == smp_processor_id()) - continue; - if (cpumask_test_cpu(cpu, mask)) - smp_call_function_single(cpu, func, info, wait); - } - return; - } + data = &__get_cpu_var(cfd_data); + csd_lock(&data->csd); - spin_lock_init(&data->lock); - data->csd.flags = CSD_FLAG_ALLOC; - if (wait) - data->csd.flags |= CSD_FLAG_WAIT; + spin_lock_irqsave(&data->lock, flags); data->csd.func = func; data->csd.info = info; - cpumask_and(to_cpumask(data->cpumask_bits), mask, cpu_online_mask); - cpumask_clear_cpu(smp_processor_id(), to_cpumask(data->cpumask_bits)); - data->refs = cpumask_weight(to_cpumask(data->cpumask_bits)); + cpumask_and(data->cpumask, mask, cpu_online_mask); + cpumask_clear_cpu(this_cpu, data->cpumask); + data->refs = cpumask_weight(data->cpumask); - spin_lock_irqsave(&call_function_lock, flags); - list_add_tail_rcu(&data->csd.list, &call_function_queue); - spin_unlock_irqrestore(&call_function_lock, flags); + spin_lock(&call_function.lock); + /* + * Place entry at the _HEAD_ of the list, so that any cpu still + * observing the entry in generic_smp_call_function_interrupt() + * will not miss any other list entries: + */ + list_add_rcu(&data->csd.list, &call_function.queue); + spin_unlock(&call_function.lock); + + spin_unlock_irqrestore(&data->lock, flags); /* * Make the list addition visible before sending the ipi. + * (IPIs must obey or appear to obey normal Linux cache + * coherency rules -- see comment in generic_exec_single). */ smp_mb(); /* Send a message to all CPUs in the map */ - arch_send_call_function_ipi_mask(to_cpumask(data->cpumask_bits)); + arch_send_call_function_ipi_mask(data->cpumask); - /* optionally wait for the CPUs to complete */ + /* Optionally wait for the CPUs to complete */ if (wait) - csd_flag_wait(&data->csd); + csd_lock_wait(&data->csd); } EXPORT_SYMBOL(smp_call_function_many); @@ -361,7 +429,8 @@ EXPORT_SYMBOL(smp_call_function_many); * smp_call_function(): Run a function on all other CPUs. * @func: The function to run. This must be fast and non-blocking. * @info: An arbitrary pointer to pass to the function. - * @wait: If true, wait (atomically) until function has completed on other CPUs. + * @wait: If true, wait (atomically) until function has completed + * on other CPUs. * * Returns 0. * @@ -377,26 +446,27 @@ int smp_call_function(void (*func)(void *), void *info, int wait) preempt_disable(); smp_call_function_many(cpu_online_mask, func, info, wait); preempt_enable(); + return 0; } EXPORT_SYMBOL(smp_call_function); void ipi_call_lock(void) { - spin_lock(&call_function_lock); + spin_lock(&call_function.lock); } void ipi_call_unlock(void) { - spin_unlock(&call_function_lock); + spin_unlock(&call_function.lock); } void ipi_call_lock_irq(void) { - spin_lock_irq(&call_function_lock); + spin_lock_irq(&call_function.lock); } void ipi_call_unlock_irq(void) { - spin_unlock_irq(&call_function_lock); + spin_unlock_irq(&call_function.lock); } diff --git a/kernel/softirq.c b/kernel/softirq.c index bdbe9de9cd8..7db25067cd2 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -21,9 +21,13 @@ #include <linux/freezer.h> #include <linux/kthread.h> #include <linux/rcupdate.h> +#include <linux/ftrace.h> #include <linux/smp.h> #include <linux/tick.h> +#define CREATE_TRACE_POINTS +#include <trace/events/irq.h> + #include <asm/irq.h> /* - No shared variables, all the data are CPU local. @@ -52,13 +56,18 @@ static struct softirq_action softirq_vec[NR_SOFTIRQS] __cacheline_aligned_in_smp static DEFINE_PER_CPU(struct task_struct *, ksoftirqd); +char *softirq_to_name[NR_SOFTIRQS] = { + "HI", "TIMER", "NET_TX", "NET_RX", "BLOCK", + "TASKLET", "SCHED", "HRTIMER", "RCU" +}; + /* * we cannot loop indefinitely here to avoid userspace starvation, * but we also don't want to introduce a worst case 1/HZ latency * to the pending events, so lets the scheduler to balance * the softirq load for us. */ -static inline void wakeup_softirqd(void) +void wakeup_softirqd(void) { /* Interrupts are disabled: no need to stop preemption */ struct task_struct *tsk = __get_cpu_var(ksoftirqd); @@ -79,13 +88,23 @@ static void __local_bh_disable(unsigned long ip) WARN_ON_ONCE(in_irq()); raw_local_irq_save(flags); - add_preempt_count(SOFTIRQ_OFFSET); + /* + * The preempt tracer hooks into add_preempt_count and will break + * lockdep because it calls back into lockdep after SOFTIRQ_OFFSET + * is set and before current->softirq_enabled is cleared. + * We must manually increment preempt_count here and manually + * call the trace_preempt_off later. + */ + preempt_count() += SOFTIRQ_OFFSET; /* * Were softirqs turned off above: */ if (softirq_count() == SOFTIRQ_OFFSET) trace_softirqs_off(ip); raw_local_irq_restore(flags); + + if (preempt_count() == SOFTIRQ_OFFSET) + trace_preempt_off(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1)); } #else /* !CONFIG_TRACE_IRQFLAGS */ static inline void __local_bh_disable(unsigned long ip) @@ -180,7 +199,7 @@ asmlinkage void __do_softirq(void) account_system_vtime(current); __local_bh_disable((unsigned long)__builtin_return_address(0)); - trace_softirq_enter(); + lockdep_softirq_enter(); cpu = smp_processor_id(); restart: @@ -194,18 +213,21 @@ restart: do { if (pending & 1) { int prev_count = preempt_count(); + kstat_incr_softirqs_this_cpu(h - softirq_vec); + trace_softirq_entry(h, softirq_vec); h->action(h); - + trace_softirq_exit(h, softirq_vec); if (unlikely(prev_count != preempt_count())) { - printk(KERN_ERR "huh, entered softirq %td %p" + printk(KERN_ERR "huh, entered softirq %td %s %p" "with preempt_count %08x," " exited with %08x?\n", h - softirq_vec, + softirq_to_name[h - softirq_vec], h->action, prev_count, preempt_count()); preempt_count() = prev_count; } - rcu_bh_qsctr_inc(cpu); + rcu_bh_qs(cpu); } h++; pending >>= 1; @@ -220,7 +242,7 @@ restart: if (pending) wakeup_softirqd(); - trace_softirq_exit(); + lockdep_softirq_exit(); account_system_vtime(current); _local_bh_enable(); @@ -323,7 +345,9 @@ void open_softirq(int nr, void (*action)(struct softirq_action *)) softirq_vec[nr].action = action; } -/* Tasklets */ +/* + * Tasklets + */ struct tasklet_head { struct tasklet_struct *head; @@ -361,6 +385,17 @@ void __tasklet_hi_schedule(struct tasklet_struct *t) EXPORT_SYMBOL(__tasklet_hi_schedule); +void __tasklet_hi_schedule_first(struct tasklet_struct *t) +{ + BUG_ON(!irqs_disabled()); + + t->next = __get_cpu_var(tasklet_hi_vec).head; + __get_cpu_var(tasklet_hi_vec).head = t; + __raise_softirq_irqoff(HI_SOFTIRQ); +} + +EXPORT_SYMBOL(__tasklet_hi_schedule_first); + static void tasklet_action(struct softirq_action *a) { struct tasklet_struct *list; @@ -450,9 +485,9 @@ void tasklet_kill(struct tasklet_struct *t) printk("Attempt to kill tasklet from interrupt\n"); while (test_and_set_bit(TASKLET_STATE_SCHED, &t->state)) { - do + do { yield(); - while (test_bit(TASKLET_STATE_SCHED, &t->state)); + } while (test_bit(TASKLET_STATE_SCHED, &t->state)); } tasklet_unlock_wait(t); clear_bit(TASKLET_STATE_SCHED, &t->state); @@ -460,6 +495,66 @@ void tasklet_kill(struct tasklet_struct *t) EXPORT_SYMBOL(tasklet_kill); +/* + * tasklet_hrtimer + */ + +/* + * The trampoline is called when the hrtimer expires. If this is + * called from the hrtimer interrupt then we schedule the tasklet as + * the timer callback function expects to run in softirq context. If + * it's called in softirq context anyway (i.e. high resolution timers + * disabled) then the hrtimer callback is called right away. + */ +static enum hrtimer_restart __hrtimer_tasklet_trampoline(struct hrtimer *timer) +{ + struct tasklet_hrtimer *ttimer = + container_of(timer, struct tasklet_hrtimer, timer); + + if (hrtimer_is_hres_active(timer)) { + tasklet_hi_schedule(&ttimer->tasklet); + return HRTIMER_NORESTART; + } + return ttimer->function(timer); +} + +/* + * Helper function which calls the hrtimer callback from + * tasklet/softirq context + */ +static void __tasklet_hrtimer_trampoline(unsigned long data) +{ + struct tasklet_hrtimer *ttimer = (void *)data; + enum hrtimer_restart restart; + + restart = ttimer->function(&ttimer->timer); + if (restart != HRTIMER_NORESTART) + hrtimer_restart(&ttimer->timer); +} + +/** + * tasklet_hrtimer_init - Init a tasklet/hrtimer combo for softirq callbacks + * @ttimer: tasklet_hrtimer which is initialized + * @function: hrtimer callback funtion which gets called from softirq context + * @which_clock: clock id (CLOCK_MONOTONIC/CLOCK_REALTIME) + * @mode: hrtimer mode (HRTIMER_MODE_ABS/HRTIMER_MODE_REL) + */ +void tasklet_hrtimer_init(struct tasklet_hrtimer *ttimer, + enum hrtimer_restart (*function)(struct hrtimer *), + clockid_t which_clock, enum hrtimer_mode mode) +{ + hrtimer_init(&ttimer->timer, which_clock, mode); + ttimer->timer.function = __hrtimer_tasklet_trampoline; + tasklet_init(&ttimer->tasklet, __tasklet_hrtimer_trampoline, + (unsigned long)ttimer); + ttimer->function = function; +} +EXPORT_SYMBOL_GPL(tasklet_hrtimer_init); + +/* + * Remote softirq bits + */ + DEFINE_PER_CPU(struct list_head [NR_SOFTIRQS], softirq_work_list); EXPORT_PER_CPU_SYMBOL(softirq_work_list); @@ -496,7 +591,7 @@ static int __try_remote_softirq(struct call_single_data *cp, int cpu, int softir cp->flags = 0; cp->priv = softirq; - __smp_call_function_single(cpu, cp); + __smp_call_function_single(cpu, cp, 0); return 0; } return 1; @@ -626,6 +721,7 @@ static int ksoftirqd(void * __bind_cpu) preempt_enable_no_resched(); cond_resched(); preempt_disable(); + rcu_sched_qs((long)__bind_cpu); } preempt_enable(); set_current_state(TASK_INTERRUPTIBLE); @@ -795,12 +891,17 @@ int __init __weak early_irq_init(void) return 0; } +int __init __weak arch_probe_nr_irqs(void) +{ + return 0; +} + int __init __weak arch_early_irq_init(void) { return 0; } -int __weak arch_init_chip_data(struct irq_desc *desc, int cpu) +int __weak arch_init_chip_data(struct irq_desc *desc, int node) { return 0; } diff --git a/kernel/softlockup.c b/kernel/softlockup.c index d9188c66278..88796c33083 100644 --- a/kernel/softlockup.c +++ b/kernel/softlockup.c @@ -16,6 +16,7 @@ #include <linux/lockdep.h> #include <linux/notifier.h> #include <linux/module.h> +#include <linux/sysctl.h> #include <asm/irq_regs.h> @@ -88,6 +89,14 @@ void touch_all_softlockup_watchdogs(void) } EXPORT_SYMBOL(touch_all_softlockup_watchdogs); +int proc_dosoftlockup_thresh(struct ctl_table *table, int write, + struct file *filp, void __user *buffer, + size_t *lenp, loff_t *ppos) +{ + touch_all_softlockup_watchdogs(); + return proc_dointvec_minmax(table, write, filp, buffer, lenp, ppos); +} + /* * This callback runs from the timer interrupt, and checks * whether the watchdog thread has hung or not: @@ -157,97 +166,11 @@ void softlockup_tick(void) } /* - * Have a reasonable limit on the number of tasks checked: - */ -unsigned long __read_mostly sysctl_hung_task_check_count = 1024; - -/* - * Zero means infinite timeout - no checking done: - */ -unsigned long __read_mostly sysctl_hung_task_timeout_secs = 480; - -unsigned long __read_mostly sysctl_hung_task_warnings = 10; - -/* - * Only do the hung-tasks check on one CPU: - */ -static int check_cpu __read_mostly = -1; - -static void check_hung_task(struct task_struct *t, unsigned long now) -{ - unsigned long switch_count = t->nvcsw + t->nivcsw; - - if (t->flags & PF_FROZEN) - return; - - if (switch_count != t->last_switch_count || !t->last_switch_timestamp) { - t->last_switch_count = switch_count; - t->last_switch_timestamp = now; - return; - } - if ((long)(now - t->last_switch_timestamp) < - sysctl_hung_task_timeout_secs) - return; - if (!sysctl_hung_task_warnings) - return; - sysctl_hung_task_warnings--; - - /* - * Ok, the task did not get scheduled for more than 2 minutes, - * complain: - */ - printk(KERN_ERR "INFO: task %s:%d blocked for more than " - "%ld seconds.\n", t->comm, t->pid, - sysctl_hung_task_timeout_secs); - printk(KERN_ERR "\"echo 0 > /proc/sys/kernel/hung_task_timeout_secs\"" - " disables this message.\n"); - sched_show_task(t); - __debug_show_held_locks(t); - - t->last_switch_timestamp = now; - touch_nmi_watchdog(); - - if (softlockup_panic) - panic("softlockup: blocked tasks"); -} - -/* - * Check whether a TASK_UNINTERRUPTIBLE does not get woken up for - * a really long time (120 seconds). If that happens, print out - * a warning. - */ -static void check_hung_uninterruptible_tasks(int this_cpu) -{ - int max_count = sysctl_hung_task_check_count; - unsigned long now = get_timestamp(this_cpu); - struct task_struct *g, *t; - - /* - * If the system crashed already then all bets are off, - * do not report extra hung tasks: - */ - if (test_taint(TAINT_DIE) || did_panic) - return; - - read_lock(&tasklist_lock); - do_each_thread(g, t) { - if (!--max_count) - goto unlock; - /* use "==" to skip the TASK_KILLABLE tasks waiting on NFS */ - if (t->state == TASK_UNINTERRUPTIBLE) - check_hung_task(t, now); - } while_each_thread(g, t); - unlock: - read_unlock(&tasklist_lock); -} - -/* * The watchdog thread - runs every second and touches the timestamp. */ static int watchdog(void *__bind_cpu) { struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 }; - int this_cpu = (long)__bind_cpu; sched_setscheduler(current, SCHED_FIFO, ¶m); @@ -267,11 +190,6 @@ static int watchdog(void *__bind_cpu) if (kthread_should_stop()) break; - if (this_cpu == check_cpu) { - if (sysctl_hung_task_timeout_secs) - check_hung_uninterruptible_tasks(this_cpu); - } - set_current_state(TASK_INTERRUPTIBLE); } __set_current_state(TASK_RUNNING); @@ -303,18 +221,9 @@ cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) break; case CPU_ONLINE: case CPU_ONLINE_FROZEN: - check_cpu = cpumask_any(cpu_online_mask); wake_up_process(per_cpu(watchdog_task, hotcpu)); break; #ifdef CONFIG_HOTPLUG_CPU - case CPU_DOWN_PREPARE: - case CPU_DOWN_PREPARE_FROZEN: - if (hotcpu == check_cpu) { - /* Pick any other online cpu. */ - check_cpu = cpumask_any_but(cpu_online_mask, hotcpu); - } - break; - case CPU_UP_CANCELED: case CPU_UP_CANCELED_FROZEN: if (!per_cpu(watchdog_task, hotcpu)) diff --git a/kernel/spinlock.c b/kernel/spinlock.c index 29ab20749dd..5ddab730cb2 100644 --- a/kernel/spinlock.c +++ b/kernel/spinlock.c @@ -21,44 +21,29 @@ #include <linux/debug_locks.h> #include <linux/module.h> +#ifndef _spin_trylock int __lockfunc _spin_trylock(spinlock_t *lock) { - preempt_disable(); - if (_raw_spin_trylock(lock)) { - spin_acquire(&lock->dep_map, 0, 1, _RET_IP_); - return 1; - } - - preempt_enable(); - return 0; + return __spin_trylock(lock); } EXPORT_SYMBOL(_spin_trylock); +#endif +#ifndef _read_trylock int __lockfunc _read_trylock(rwlock_t *lock) { - preempt_disable(); - if (_raw_read_trylock(lock)) { - rwlock_acquire_read(&lock->dep_map, 0, 1, _RET_IP_); - return 1; - } - - preempt_enable(); - return 0; + return __read_trylock(lock); } EXPORT_SYMBOL(_read_trylock); +#endif +#ifndef _write_trylock int __lockfunc _write_trylock(rwlock_t *lock) { - preempt_disable(); - if (_raw_write_trylock(lock)) { - rwlock_acquire(&lock->dep_map, 0, 1, _RET_IP_); - return 1; - } - - preempt_enable(); - return 0; + return __write_trylock(lock); } EXPORT_SYMBOL(_write_trylock); +#endif /* * If lockdep is enabled then we use the non-preemption spin-ops @@ -67,130 +52,101 @@ EXPORT_SYMBOL(_write_trylock); */ #if !defined(CONFIG_GENERIC_LOCKBREAK) || defined(CONFIG_DEBUG_LOCK_ALLOC) +#ifndef _read_lock void __lockfunc _read_lock(rwlock_t *lock) { - preempt_disable(); - rwlock_acquire_read(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_read_trylock, _raw_read_lock); + __read_lock(lock); } EXPORT_SYMBOL(_read_lock); +#endif +#ifndef _spin_lock_irqsave unsigned long __lockfunc _spin_lock_irqsave(spinlock_t *lock) { - unsigned long flags; - - local_irq_save(flags); - preempt_disable(); - spin_acquire(&lock->dep_map, 0, 0, _RET_IP_); - /* - * On lockdep we dont want the hand-coded irq-enable of - * _raw_spin_lock_flags() code, because lockdep assumes - * that interrupts are not re-enabled during lock-acquire: - */ -#ifdef CONFIG_LOCKDEP - LOCK_CONTENDED(lock, _raw_spin_trylock, _raw_spin_lock); -#else - _raw_spin_lock_flags(lock, &flags); -#endif - return flags; + return __spin_lock_irqsave(lock); } EXPORT_SYMBOL(_spin_lock_irqsave); +#endif +#ifndef _spin_lock_irq void __lockfunc _spin_lock_irq(spinlock_t *lock) { - local_irq_disable(); - preempt_disable(); - spin_acquire(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_spin_trylock, _raw_spin_lock); + __spin_lock_irq(lock); } EXPORT_SYMBOL(_spin_lock_irq); +#endif +#ifndef _spin_lock_bh void __lockfunc _spin_lock_bh(spinlock_t *lock) { - local_bh_disable(); - preempt_disable(); - spin_acquire(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_spin_trylock, _raw_spin_lock); + __spin_lock_bh(lock); } EXPORT_SYMBOL(_spin_lock_bh); +#endif +#ifndef _read_lock_irqsave unsigned long __lockfunc _read_lock_irqsave(rwlock_t *lock) { - unsigned long flags; - - local_irq_save(flags); - preempt_disable(); - rwlock_acquire_read(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_read_trylock, _raw_read_lock); - return flags; + return __read_lock_irqsave(lock); } EXPORT_SYMBOL(_read_lock_irqsave); +#endif +#ifndef _read_lock_irq void __lockfunc _read_lock_irq(rwlock_t *lock) { - local_irq_disable(); - preempt_disable(); - rwlock_acquire_read(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_read_trylock, _raw_read_lock); + __read_lock_irq(lock); } EXPORT_SYMBOL(_read_lock_irq); +#endif +#ifndef _read_lock_bh void __lockfunc _read_lock_bh(rwlock_t *lock) { - local_bh_disable(); - preempt_disable(); - rwlock_acquire_read(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_read_trylock, _raw_read_lock); + __read_lock_bh(lock); } EXPORT_SYMBOL(_read_lock_bh); +#endif +#ifndef _write_lock_irqsave unsigned long __lockfunc _write_lock_irqsave(rwlock_t *lock) { - unsigned long flags; - - local_irq_save(flags); - preempt_disable(); - rwlock_acquire(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_write_trylock, _raw_write_lock); - return flags; + return __write_lock_irqsave(lock); } EXPORT_SYMBOL(_write_lock_irqsave); +#endif +#ifndef _write_lock_irq void __lockfunc _write_lock_irq(rwlock_t *lock) { - local_irq_disable(); - preempt_disable(); - rwlock_acquire(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_write_trylock, _raw_write_lock); + __write_lock_irq(lock); } EXPORT_SYMBOL(_write_lock_irq); +#endif +#ifndef _write_lock_bh void __lockfunc _write_lock_bh(rwlock_t *lock) { - local_bh_disable(); - preempt_disable(); - rwlock_acquire(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_write_trylock, _raw_write_lock); + __write_lock_bh(lock); } EXPORT_SYMBOL(_write_lock_bh); +#endif +#ifndef _spin_lock void __lockfunc _spin_lock(spinlock_t *lock) { - preempt_disable(); - spin_acquire(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_spin_trylock, _raw_spin_lock); + __spin_lock(lock); } - EXPORT_SYMBOL(_spin_lock); +#endif +#ifndef _write_lock void __lockfunc _write_lock(rwlock_t *lock) { - preempt_disable(); - rwlock_acquire(&lock->dep_map, 0, 0, _RET_IP_); - LOCK_CONTENDED(lock, _raw_write_trylock, _raw_write_lock); + __write_lock(lock); } - EXPORT_SYMBOL(_write_lock); +#endif #else /* CONFIG_PREEMPT: */ @@ -299,16 +255,8 @@ unsigned long __lockfunc _spin_lock_irqsave_nested(spinlock_t *lock, int subclas local_irq_save(flags); preempt_disable(); spin_acquire(&lock->dep_map, subclass, 0, _RET_IP_); - /* - * On lockdep we dont want the hand-coded irq-enable of - * _raw_spin_lock_flags() code, because lockdep assumes - * that interrupts are not re-enabled during lock-acquire: - */ -#ifdef CONFIG_LOCKDEP - LOCK_CONTENDED(lock, _raw_spin_trylock, _raw_spin_lock); -#else - _raw_spin_lock_flags(lock, &flags); -#endif + LOCK_CONTENDED_FLAGS(lock, _raw_spin_trylock, _raw_spin_lock, + _raw_spin_lock_flags, &flags); return flags; } EXPORT_SYMBOL(_spin_lock_irqsave_nested); @@ -324,125 +272,109 @@ EXPORT_SYMBOL(_spin_lock_nest_lock); #endif +#ifndef _spin_unlock void __lockfunc _spin_unlock(spinlock_t *lock) { - spin_release(&lock->dep_map, 1, _RET_IP_); - _raw_spin_unlock(lock); - preempt_enable(); + __spin_unlock(lock); } EXPORT_SYMBOL(_spin_unlock); +#endif +#ifndef _write_unlock void __lockfunc _write_unlock(rwlock_t *lock) { - rwlock_release(&lock->dep_map, 1, _RET_IP_); - _raw_write_unlock(lock); - preempt_enable(); + __write_unlock(lock); } EXPORT_SYMBOL(_write_unlock); +#endif +#ifndef _read_unlock void __lockfunc _read_unlock(rwlock_t *lock) { - rwlock_release(&lock->dep_map, 1, _RET_IP_); - _raw_read_unlock(lock); - preempt_enable(); + __read_unlock(lock); } EXPORT_SYMBOL(_read_unlock); +#endif +#ifndef _spin_unlock_irqrestore void __lockfunc _spin_unlock_irqrestore(spinlock_t *lock, unsigned long flags) { - spin_release(&lock->dep_map, 1, _RET_IP_); - _raw_spin_unlock(lock); - local_irq_restore(flags); - preempt_enable(); + __spin_unlock_irqrestore(lock, flags); } EXPORT_SYMBOL(_spin_unlock_irqrestore); +#endif +#ifndef _spin_unlock_irq void __lockfunc _spin_unlock_irq(spinlock_t *lock) { - spin_release(&lock->dep_map, 1, _RET_IP_); - _raw_spin_unlock(lock); - local_irq_enable(); - preempt_enable(); + __spin_unlock_irq(lock); } EXPORT_SYMBOL(_spin_unlock_irq); +#endif +#ifndef _spin_unlock_bh void __lockfunc _spin_unlock_bh(spinlock_t *lock) { - spin_release(&lock->dep_map, 1, _RET_IP_); - _raw_spin_unlock(lock); - preempt_enable_no_resched(); - local_bh_enable_ip((unsigned long)__builtin_return_address(0)); + __spin_unlock_bh(lock); } EXPORT_SYMBOL(_spin_unlock_bh); +#endif +#ifndef _read_unlock_irqrestore void __lockfunc _read_unlock_irqrestore(rwlock_t *lock, unsigned long flags) { - rwlock_release(&lock->dep_map, 1, _RET_IP_); - _raw_read_unlock(lock); - local_irq_restore(flags); - preempt_enable(); + __read_unlock_irqrestore(lock, flags); } EXPORT_SYMBOL(_read_unlock_irqrestore); +#endif +#ifndef _read_unlock_irq void __lockfunc _read_unlock_irq(rwlock_t *lock) { - rwlock_release(&lock->dep_map, 1, _RET_IP_); - _raw_read_unlock(lock); - local_irq_enable(); - preempt_enable(); + __read_unlock_irq(lock); } EXPORT_SYMBOL(_read_unlock_irq); +#endif +#ifndef _read_unlock_bh void __lockfunc _read_unlock_bh(rwlock_t *lock) { - rwlock_release(&lock->dep_map, 1, _RET_IP_); - _raw_read_unlock(lock); - preempt_enable_no_resched(); - local_bh_enable_ip((unsigned long)__builtin_return_address(0)); + __read_unlock_bh(lock); } EXPORT_SYMBOL(_read_unlock_bh); +#endif +#ifndef _write_unlock_irqrestore void __lockfunc _write_unlock_irqrestore(rwlock_t *lock, unsigned long flags) { - rwlock_release(&lock->dep_map, 1, _RET_IP_); - _raw_write_unlock(lock); - local_irq_restore(flags); - preempt_enable(); + __write_unlock_irqrestore(lock, flags); } EXPORT_SYMBOL(_write_unlock_irqrestore); +#endif +#ifndef _write_unlock_irq void __lockfunc _write_unlock_irq(rwlock_t *lock) { - rwlock_release(&lock->dep_map, 1, _RET_IP_); - _raw_write_unlock(lock); - local_irq_enable(); - preempt_enable(); + __write_unlock_irq(lock); } EXPORT_SYMBOL(_write_unlock_irq); +#endif +#ifndef _write_unlock_bh void __lockfunc _write_unlock_bh(rwlock_t *lock) { - rwlock_release(&lock->dep_map, 1, _RET_IP_); - _raw_write_unlock(lock); - preempt_enable_no_resched(); - local_bh_enable_ip((unsigned long)__builtin_return_address(0)); + __write_unlock_bh(lock); } EXPORT_SYMBOL(_write_unlock_bh); +#endif +#ifndef _spin_trylock_bh int __lockfunc _spin_trylock_bh(spinlock_t *lock) { - local_bh_disable(); - preempt_disable(); - if (_raw_spin_trylock(lock)) { - spin_acquire(&lock->dep_map, 0, 1, _RET_IP_); - return 1; - } - - preempt_enable_no_resched(); - local_bh_enable_ip((unsigned long)__builtin_return_address(0)); - return 0; + return __spin_trylock_bh(lock); } EXPORT_SYMBOL(_spin_trylock_bh); +#endif notrace int in_lock_functions(unsigned long addr) { diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c index 0cd415ee62a..912823e2a11 100644 --- a/kernel/stop_machine.c +++ b/kernel/stop_machine.c @@ -44,7 +44,7 @@ static DEFINE_MUTEX(setup_lock); static int refcount; static struct workqueue_struct *stop_machine_wq; static struct stop_machine_data active, idle; -static const cpumask_t *active_cpus; +static const struct cpumask *active_cpus; static void *stop_machine_work; static void set_state(enum stopmachine_state newstate) @@ -170,7 +170,7 @@ int __stop_machine(int (*fn)(void *), void *data, const struct cpumask *cpus) * doesn't hit this CPU until we're ready. */ get_cpu(); for_each_online_cpu(i) { - sm_work = percpu_ptr(stop_machine_work, i); + sm_work = per_cpu_ptr(stop_machine_work, i); INIT_WORK(sm_work, stop_cpu); queue_work_on(i, stop_machine_wq, sm_work); } diff --git a/kernel/sys.c b/kernel/sys.c index e7dc0e10a48..b3f1097c76f 100644 --- a/kernel/sys.c +++ b/kernel/sys.c @@ -14,6 +14,7 @@ #include <linux/prctl.h> #include <linux/highuid.h> #include <linux/fs.h> +#include <linux/perf_counter.h> #include <linux/resource.h> #include <linux/kernel.h> #include <linux/kexec.h> @@ -34,6 +35,7 @@ #include <linux/seccomp.h> #include <linux/cpu.h> #include <linux/ptrace.h> +#include <linux/fs_struct.h> #include <linux/compat.h> #include <linux/syscalls.h> @@ -359,6 +361,7 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, void __user *, arg) { char buffer[256]; + int ret = 0; /* We only trust the superuser with rebooting the system. */ if (!capable(CAP_SYS_BOOT)) @@ -396,7 +399,7 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, kernel_halt(); unlock_kernel(); do_exit(0); - break; + panic("cannot halt"); case LINUX_REBOOT_CMD_POWER_OFF: kernel_power_off(); @@ -416,29 +419,22 @@ SYSCALL_DEFINE4(reboot, int, magic1, int, magic2, unsigned int, cmd, #ifdef CONFIG_KEXEC case LINUX_REBOOT_CMD_KEXEC: - { - int ret; - ret = kernel_kexec(); - unlock_kernel(); - return ret; - } + ret = kernel_kexec(); + break; #endif #ifdef CONFIG_HIBERNATION case LINUX_REBOOT_CMD_SW_SUSPEND: - { - int ret = hibernate(); - unlock_kernel(); - return ret; - } + ret = hibernate(); + break; #endif default: - unlock_kernel(); - return -EINVAL; + ret = -EINVAL; + break; } unlock_kernel(); - return 0; + return ret; } static void deferred_cad(struct work_struct *dummy) @@ -559,7 +555,7 @@ error: abort_creds(new); return retval; } - + /* * change the user struct in a credentials set to match the new UID */ @@ -571,6 +567,11 @@ static int set_user(struct cred *new) if (!new_user) return -EAGAIN; + if (!task_can_switch_user(new_user, current)) { + free_uid(new_user); + return -EINVAL; + } + if (atomic_read(&new_user->processes) >= current->signal->rlim[RLIMIT_NPROC].rlim_cur && new_user != INIT_USER) { @@ -631,10 +632,11 @@ SYSCALL_DEFINE2(setreuid, uid_t, ruid, uid_t, euid) goto error; } - retval = -EAGAIN; - if (new->uid != old->uid && set_user(new) < 0) - goto error; - + if (new->uid != old->uid) { + retval = set_user(new); + if (retval < 0) + goto error; + } if (ruid != (uid_t) -1 || (euid != (uid_t) -1 && euid != old->uid)) new->suid = new->euid; @@ -680,9 +682,10 @@ SYSCALL_DEFINE1(setuid, uid_t, uid) retval = -EPERM; if (capable(CAP_SETUID)) { new->suid = new->uid = uid; - if (uid != old->uid && set_user(new) < 0) { - retval = -EAGAIN; - goto error; + if (uid != old->uid) { + retval = set_user(new); + if (retval < 0) + goto error; } } else if (uid != old->uid && uid != new->suid) { goto error; @@ -734,11 +737,13 @@ SYSCALL_DEFINE3(setresuid, uid_t, ruid, uid_t, euid, uid_t, suid) goto error; } - retval = -EAGAIN; if (ruid != (uid_t) -1) { new->uid = ruid; - if (ruid != old->uid && set_user(new) < 0) - goto error; + if (ruid != old->uid) { + retval = set_user(new); + if (retval < 0) + goto error; + } } if (euid != (uid_t) -1) new->euid = euid; @@ -1004,10 +1009,8 @@ SYSCALL_DEFINE2(setpgid, pid_t, pid, pid_t, pgid) if (err) goto out; - if (task_pgrp(p) != pgrp) { + if (task_pgrp(p) != pgrp) change_pid(p, PIDTYPE_PGID, pgrp); - set_task_pgrp(p, pid_nr(pgrp)); - } err = 0; out: @@ -1110,289 +1113,6 @@ out: return err; } -/* - * Supplementary group IDs - */ - -/* init to 2 - one for init_task, one to ensure it is never freed */ -struct group_info init_groups = { .usage = ATOMIC_INIT(2) }; - -struct group_info *groups_alloc(int gidsetsize) -{ - struct group_info *group_info; - int nblocks; - int i; - - nblocks = (gidsetsize + NGROUPS_PER_BLOCK - 1) / NGROUPS_PER_BLOCK; - /* Make sure we always allocate at least one indirect block pointer */ - nblocks = nblocks ? : 1; - group_info = kmalloc(sizeof(*group_info) + nblocks*sizeof(gid_t *), GFP_USER); - if (!group_info) - return NULL; - group_info->ngroups = gidsetsize; - group_info->nblocks = nblocks; - atomic_set(&group_info->usage, 1); - - if (gidsetsize <= NGROUPS_SMALL) - group_info->blocks[0] = group_info->small_block; - else { - for (i = 0; i < nblocks; i++) { - gid_t *b; - b = (void *)__get_free_page(GFP_USER); - if (!b) - goto out_undo_partial_alloc; - group_info->blocks[i] = b; - } - } - return group_info; - -out_undo_partial_alloc: - while (--i >= 0) { - free_page((unsigned long)group_info->blocks[i]); - } - kfree(group_info); - return NULL; -} - -EXPORT_SYMBOL(groups_alloc); - -void groups_free(struct group_info *group_info) -{ - if (group_info->blocks[0] != group_info->small_block) { - int i; - for (i = 0; i < group_info->nblocks; i++) - free_page((unsigned long)group_info->blocks[i]); - } - kfree(group_info); -} - -EXPORT_SYMBOL(groups_free); - -/* export the group_info to a user-space array */ -static int groups_to_user(gid_t __user *grouplist, - const struct group_info *group_info) -{ - int i; - unsigned int count = group_info->ngroups; - - for (i = 0; i < group_info->nblocks; i++) { - unsigned int cp_count = min(NGROUPS_PER_BLOCK, count); - unsigned int len = cp_count * sizeof(*grouplist); - - if (copy_to_user(grouplist, group_info->blocks[i], len)) - return -EFAULT; - - grouplist += NGROUPS_PER_BLOCK; - count -= cp_count; - } - return 0; -} - -/* fill a group_info from a user-space array - it must be allocated already */ -static int groups_from_user(struct group_info *group_info, - gid_t __user *grouplist) -{ - int i; - unsigned int count = group_info->ngroups; - - for (i = 0; i < group_info->nblocks; i++) { - unsigned int cp_count = min(NGROUPS_PER_BLOCK, count); - unsigned int len = cp_count * sizeof(*grouplist); - - if (copy_from_user(group_info->blocks[i], grouplist, len)) - return -EFAULT; - - grouplist += NGROUPS_PER_BLOCK; - count -= cp_count; - } - return 0; -} - -/* a simple Shell sort */ -static void groups_sort(struct group_info *group_info) -{ - int base, max, stride; - int gidsetsize = group_info->ngroups; - - for (stride = 1; stride < gidsetsize; stride = 3 * stride + 1) - ; /* nothing */ - stride /= 3; - - while (stride) { - max = gidsetsize - stride; - for (base = 0; base < max; base++) { - int left = base; - int right = left + stride; - gid_t tmp = GROUP_AT(group_info, right); - - while (left >= 0 && GROUP_AT(group_info, left) > tmp) { - GROUP_AT(group_info, right) = - GROUP_AT(group_info, left); - right = left; - left -= stride; - } - GROUP_AT(group_info, right) = tmp; - } - stride /= 3; - } -} - -/* a simple bsearch */ -int groups_search(const struct group_info *group_info, gid_t grp) -{ - unsigned int left, right; - - if (!group_info) - return 0; - - left = 0; - right = group_info->ngroups; - while (left < right) { - unsigned int mid = (left+right)/2; - int cmp = grp - GROUP_AT(group_info, mid); - if (cmp > 0) - left = mid + 1; - else if (cmp < 0) - right = mid; - else - return 1; - } - return 0; -} - -/** - * set_groups - Change a group subscription in a set of credentials - * @new: The newly prepared set of credentials to alter - * @group_info: The group list to install - * - * Validate a group subscription and, if valid, insert it into a set - * of credentials. - */ -int set_groups(struct cred *new, struct group_info *group_info) -{ - int retval; - - retval = security_task_setgroups(group_info); - if (retval) - return retval; - - put_group_info(new->group_info); - groups_sort(group_info); - get_group_info(group_info); - new->group_info = group_info; - return 0; -} - -EXPORT_SYMBOL(set_groups); - -/** - * set_current_groups - Change current's group subscription - * @group_info: The group list to impose - * - * Validate a group subscription and, if valid, impose it upon current's task - * security record. - */ -int set_current_groups(struct group_info *group_info) -{ - struct cred *new; - int ret; - - new = prepare_creds(); - if (!new) - return -ENOMEM; - - ret = set_groups(new, group_info); - if (ret < 0) { - abort_creds(new); - return ret; - } - - return commit_creds(new); -} - -EXPORT_SYMBOL(set_current_groups); - -SYSCALL_DEFINE2(getgroups, int, gidsetsize, gid_t __user *, grouplist) -{ - const struct cred *cred = current_cred(); - int i; - - if (gidsetsize < 0) - return -EINVAL; - - /* no need to grab task_lock here; it cannot change */ - i = cred->group_info->ngroups; - if (gidsetsize) { - if (i > gidsetsize) { - i = -EINVAL; - goto out; - } - if (groups_to_user(grouplist, cred->group_info)) { - i = -EFAULT; - goto out; - } - } -out: - return i; -} - -/* - * SMP: Our groups are copy-on-write. We can set them safely - * without another task interfering. - */ - -SYSCALL_DEFINE2(setgroups, int, gidsetsize, gid_t __user *, grouplist) -{ - struct group_info *group_info; - int retval; - - if (!capable(CAP_SETGID)) - return -EPERM; - if ((unsigned)gidsetsize > NGROUPS_MAX) - return -EINVAL; - - group_info = groups_alloc(gidsetsize); - if (!group_info) - return -ENOMEM; - retval = groups_from_user(group_info, grouplist); - if (retval) { - put_group_info(group_info); - return retval; - } - - retval = set_current_groups(group_info); - put_group_info(group_info); - - return retval; -} - -/* - * Check whether we're fsgid/egid or in the supplemental group.. - */ -int in_group_p(gid_t grp) -{ - const struct cred *cred = current_cred(); - int retval = 1; - - if (grp != cred->fsgid) - retval = groups_search(cred->group_info, grp); - return retval; -} - -EXPORT_SYMBOL(in_group_p); - -int in_egroup_p(gid_t grp) -{ - const struct cred *cred = current_cred(); - int retval = 1; - - if (grp != cred->egid) - retval = groups_search(cred->group_info, grp); - return retval; -} - -EXPORT_SYMBOL(in_egroup_p); - DECLARE_RWSEM(uts_sem); SYSCALL_DEFINE1(newuname, struct new_utsname __user *, name) @@ -1525,22 +1245,14 @@ SYSCALL_DEFINE2(setrlimit, unsigned int, resource, struct rlimit __user *, rlim) return -EINVAL; if (copy_from_user(&new_rlim, rlim, sizeof(*rlim))) return -EFAULT; + if (new_rlim.rlim_cur > new_rlim.rlim_max) + return -EINVAL; old_rlim = current->signal->rlim + resource; if ((new_rlim.rlim_max > old_rlim->rlim_max) && !capable(CAP_SYS_RESOURCE)) return -EPERM; - - if (resource == RLIMIT_NOFILE) { - if (new_rlim.rlim_max == RLIM_INFINITY) - new_rlim.rlim_max = sysctl_nr_open; - if (new_rlim.rlim_cur == RLIM_INFINITY) - new_rlim.rlim_cur = sysctl_nr_open; - if (new_rlim.rlim_max > sysctl_nr_open) - return -EPERM; - } - - if (new_rlim.rlim_cur > new_rlim.rlim_max) - return -EINVAL; + if (resource == RLIMIT_NOFILE && new_rlim.rlim_max > sysctl_nr_open) + return -EPERM; retval = security_task_setrlimit(resource, &new_rlim); if (retval) @@ -1799,6 +1511,12 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, case PR_SET_TSC: error = SET_TSC_CTL(arg2); break; + case PR_TASK_PERF_COUNTERS_DISABLE: + error = perf_counter_task_disable(); + break; + case PR_TASK_PERF_COUNTERS_ENABLE: + error = perf_counter_task_enable(); + break; case PR_GET_TIMERSLACK: error = current->timer_slack_ns; break; diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 27dad296738..44e5936118d 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -49,6 +49,7 @@ cond_syscall(sys_sendmsg); cond_syscall(compat_sys_sendmsg); cond_syscall(sys_recvmsg); cond_syscall(compat_sys_recvmsg); +cond_syscall(compat_sys_recvfrom); cond_syscall(sys_socketcall); cond_syscall(sys_futex); cond_syscall(compat_sys_futex); @@ -175,3 +176,6 @@ cond_syscall(compat_sys_timerfd_settime); cond_syscall(compat_sys_timerfd_gettime); cond_syscall(sys_eventfd); cond_syscall(sys_eventfd2); + +/* performance counters: */ +cond_syscall(sys_perf_counter_open); diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 368d1638ee7..3125cff1c57 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -27,6 +27,7 @@ #include <linux/security.h> #include <linux/ctype.h> #include <linux/utsname.h> +#include <linux/kmemcheck.h> #include <linux/smp_lock.h> #include <linux/fs.h> #include <linux/init.h> @@ -48,6 +49,8 @@ #include <linux/acpi.h> #include <linux/reboot.h> #include <linux/ftrace.h> +#include <linux/slow-work.h> +#include <linux/perf_counter.h> #include <asm/uaccess.h> #include <asm/processor.h> @@ -95,14 +98,15 @@ static int sixty = 60; static int neg_one = -1; #endif -#if defined(CONFIG_MMU) && defined(CONFIG_FILE_LOCKING) -static int two = 2; -#endif - static int zero; -static int one = 1; +static int __maybe_unused one = 1; +static int __maybe_unused two = 2; +static unsigned long one_ul = 1; static int one_hundred = 100; +/* this is needed for the proc_doulongvec_minmax of vm_dirty_bytes */ +static unsigned long dirty_bytes_min = 2 * PAGE_SIZE; + /* this is needed for the proc_dointvec_minmax for [fs_]overflow UID and GID */ static int maxolduid = 65535; static int minolduid; @@ -112,6 +116,7 @@ static int ngroups_max = NGROUPS_MAX; #ifdef CONFIG_MODULES extern char modprobe_path[]; +extern int modules_disabled; #endif #ifdef CONFIG_CHR_DEV_SG extern int sg_big_buff; @@ -240,6 +245,14 @@ static int max_wakeup_granularity_ns = NSEC_PER_SEC; /* 1 second */ #endif static struct ctl_table kern_table[] = { + { + .ctl_name = CTL_UNNUMBERED, + .procname = "sched_child_runs_first", + .data = &sysctl_sched_child_runs_first, + .maxlen = sizeof(unsigned int), + .mode = 0644, + .proc_handler = &proc_dointvec, + }, #ifdef CONFIG_SCHED_DEBUG { .ctl_name = CTL_UNNUMBERED, @@ -294,14 +307,6 @@ static struct ctl_table kern_table[] = { }, { .ctl_name = CTL_UNNUMBERED, - .procname = "sched_child_runs_first", - .data = &sysctl_sched_child_runs_first, - .maxlen = sizeof(unsigned int), - .mode = 0644, - .proc_handler = &proc_dointvec, - }, - { - .ctl_name = CTL_UNNUMBERED, .procname = "sched_features", .data = &sysctl_sched_features, .maxlen = sizeof(unsigned int), @@ -324,6 +329,25 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = &proc_dointvec, }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "sched_time_avg", + .data = &sysctl_sched_time_avg, + .maxlen = sizeof(unsigned int), + .mode = 0644, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "timer_migration", + .data = &sysctl_timer_migration, + .maxlen = sizeof(unsigned int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .strategy = &sysctl_intvec, + .extra1 = &zero, + .extra2 = &one, + }, #endif { .ctl_name = CTL_UNNUMBERED, @@ -532,6 +556,17 @@ static struct ctl_table kern_table[] = { .proc_handler = &proc_dostring, .strategy = &sysctl_string, }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "modules_disabled", + .data = &modules_disabled, + .maxlen = sizeof(int), + .mode = 0644, + /* only handle a transition from default "0" to "1" */ + .proc_handler = &proc_dointvec_minmax, + .extra1 = &one, + .extra2 = &one, + }, #endif #if defined(CONFIG_HOTPLUG) && defined(CONFIG_NET) { @@ -720,6 +755,14 @@ static struct ctl_table kern_table[] = { .proc_handler = &proc_dointvec, }, { + .ctl_name = CTL_UNNUMBERED, + .procname = "panic_on_io_nmi", + .data = &panic_on_io_nmi, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec, + }, + { .ctl_name = KERN_BOOTLOADER_TYPE, .procname = "bootloader_type", .data = &bootloader_type, @@ -729,6 +772,14 @@ static struct ctl_table kern_table[] = { }, { .ctl_name = CTL_UNNUMBERED, + .procname = "bootloader_version", + .data = &bootloader_version, + .maxlen = sizeof (int), + .mode = 0444, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = CTL_UNNUMBERED, .procname = "kstack_depth_to_print", .data = &kstack_depth_to_print, .maxlen = sizeof(int), @@ -809,11 +860,24 @@ static struct ctl_table kern_table[] = { .data = &softlockup_thresh, .maxlen = sizeof(int), .mode = 0644, - .proc_handler = &proc_dointvec_minmax, + .proc_handler = &proc_dosoftlockup_thresh, .strategy = &sysctl_intvec, .extra1 = &neg_one, .extra2 = &sixty, }, +#endif +#ifdef CONFIG_DETECT_HUNG_TASK + { + .ctl_name = CTL_UNNUMBERED, + .procname = "hung_task_panic", + .data = &sysctl_hung_task_panic, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .strategy = &sysctl_intvec, + .extra1 = &zero, + .extra2 = &one, + }, { .ctl_name = CTL_UNNUMBERED, .procname = "hung_task_check_count", @@ -829,7 +893,7 @@ static struct ctl_table kern_table[] = { .data = &sysctl_hung_task_timeout_secs, .maxlen = sizeof(unsigned long), .mode = 0644, - .proc_handler = &proc_doulongvec_minmax, + .proc_handler = &proc_dohung_task_timeout_secs, .strategy = &sysctl_intvec, }, { @@ -889,16 +953,51 @@ static struct ctl_table kern_table[] = { .proc_handler = &proc_dointvec, }, #endif -#ifdef CONFIG_UNEVICTABLE_LRU +#ifdef CONFIG_SLOW_WORK { .ctl_name = CTL_UNNUMBERED, - .procname = "scan_unevictable_pages", - .data = &scan_unevictable_pages, - .maxlen = sizeof(scan_unevictable_pages), + .procname = "slow-work", + .mode = 0555, + .child = slow_work_sysctls, + }, +#endif +#ifdef CONFIG_PERF_COUNTERS + { + .ctl_name = CTL_UNNUMBERED, + .procname = "perf_counter_paranoid", + .data = &sysctl_perf_counter_paranoid, + .maxlen = sizeof(sysctl_perf_counter_paranoid), .mode = 0644, - .proc_handler = &scan_unevictable_handler, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "perf_counter_mlock_kb", + .data = &sysctl_perf_counter_mlock, + .maxlen = sizeof(sysctl_perf_counter_mlock), + .mode = 0644, + .proc_handler = &proc_dointvec, + }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "perf_counter_max_sample_rate", + .data = &sysctl_perf_counter_sample_rate, + .maxlen = sizeof(sysctl_perf_counter_sample_rate), + .mode = 0644, + .proc_handler = &proc_dointvec, + }, +#endif +#ifdef CONFIG_KMEMCHECK + { + .ctl_name = CTL_UNNUMBERED, + .procname = "kmemcheck", + .data = &kmemcheck_enabled, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec, }, #endif + /* * NOTE: do not add new entries to this table unless you have read * Documentation/sysctl/ctl_unnumbered.txt @@ -974,7 +1073,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = &dirty_background_bytes_handler, .strategy = &sysctl_intvec, - .extra1 = &one, + .extra1 = &one_ul, }, { .ctl_name = VM_DIRTY_RATIO, @@ -995,7 +1094,7 @@ static struct ctl_table vm_table[] = { .mode = 0644, .proc_handler = &dirty_bytes_handler, .strategy = &sysctl_intvec, - .extra1 = &one, + .extra1 = &dirty_bytes_min, }, { .procname = "dirty_writeback_centisecs", @@ -1009,7 +1108,7 @@ static struct ctl_table vm_table[] = { .data = &dirty_expire_interval, .maxlen = sizeof(dirty_expire_interval), .mode = 0644, - .proc_handler = &proc_dointvec_userhz_jiffies, + .proc_handler = &proc_dointvec, }, { .ctl_name = VM_NR_PDFLUSH_THREADS, @@ -1212,16 +1311,14 @@ static struct ctl_table vm_table[] = { .strategy = &sysctl_jiffies, }, #endif -#ifdef CONFIG_SECURITY { .ctl_name = CTL_UNNUMBERED, .procname = "mmap_min_addr", - .data = &mmap_min_addr, - .maxlen = sizeof(unsigned long), + .data = &dac_mmap_min_addr, + .maxlen = sizeof(unsigned long), .mode = 0644, - .proc_handler = &proc_doulongvec_minmax, + .proc_handler = &mmap_min_addr_handler, }, -#endif #ifdef CONFIG_NUMA { .ctl_name = CTL_UNNUMBERED, @@ -1259,6 +1356,14 @@ static struct ctl_table vm_table[] = { .extra2 = &one, }, #endif + { + .ctl_name = CTL_UNNUMBERED, + .procname = "scan_unevictable_pages", + .data = &scan_unevictable_pages, + .maxlen = sizeof(scan_unevictable_pages), + .mode = 0644, + .proc_handler = &scan_unevictable_handler, + }, /* * NOTE: do not add new entries to this table unless you have read * Documentation/sysctl/ctl_unnumbered.txt @@ -1372,10 +1477,7 @@ static struct ctl_table fs_table[] = { .data = &lease_break_time, .maxlen = sizeof(int), .mode = 0644, - .proc_handler = &proc_dointvec_minmax, - .strategy = &sysctl_intvec, - .extra1 = &zero, - .extra2 = &two, + .proc_handler = &proc_dointvec, }, #endif #ifdef CONFIG_AIO @@ -1416,7 +1518,10 @@ static struct ctl_table fs_table[] = { .data = &suid_dumpable, .maxlen = sizeof(int), .mode = 0644, - .proc_handler = &proc_dointvec, + .proc_handler = &proc_dointvec_minmax, + .strategy = &sysctl_intvec, + .extra1 = &zero, + .extra2 = &two, }, #if defined(CONFIG_BINFMT_MISC) || defined(CONFIG_BINFMT_MISC_MODULE) { @@ -2197,7 +2302,7 @@ static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table, void *data) { #define TMPBUFLEN 21 - int *i, vleft, first=1, neg, val; + int *i, vleft, first = 1, neg; unsigned long lval; size_t left, len; @@ -2250,8 +2355,6 @@ static int __do_proc_dointvec(void *tbl_data, struct ctl_table *table, len = p-buf; if ((len < left) && *p && !isspace(*p)) break; - if (neg) - val = -val; s += len; left -= len; diff --git a/kernel/sysctl_check.c b/kernel/sysctl_check.c index fafeb48f27c..b38423ca711 100644 --- a/kernel/sysctl_check.c +++ b/kernel/sysctl_check.c @@ -219,6 +219,7 @@ static const struct trans_ctl_table trans_net_ipv4_conf_vars_table[] = { { NET_IPV4_CONF_ARP_IGNORE, "arp_ignore" }, { NET_IPV4_CONF_PROMOTE_SECONDARIES, "promote_secondaries" }, { NET_IPV4_CONF_ARP_ACCEPT, "arp_accept" }, + { NET_IPV4_CONF_ARP_NOTIFY, "arp_notify" }, {} }; diff --git a/kernel/taskstats.c b/kernel/taskstats.c index 888adbcca30..ea8384d3caa 100644 --- a/kernel/taskstats.c +++ b/kernel/taskstats.c @@ -108,7 +108,7 @@ static int prepare_reply(struct genl_info *info, u8 cmd, struct sk_buff **skbp, /* * Send taskstats data in @skb to listener with nl_pid @pid */ -static int send_reply(struct sk_buff *skb, pid_t pid) +static int send_reply(struct sk_buff *skb, struct genl_info *info) { struct genlmsghdr *genlhdr = nlmsg_data(nlmsg_hdr(skb)); void *reply = genlmsg_data(genlhdr); @@ -120,7 +120,7 @@ static int send_reply(struct sk_buff *skb, pid_t pid) return rc; } - return genlmsg_unicast(skb, pid); + return genlmsg_reply(skb, info); } /* @@ -150,7 +150,7 @@ static void send_cpu_listeners(struct sk_buff *skb, if (!skb_next) break; } - rc = genlmsg_unicast(skb_cur, s->pid); + rc = genlmsg_unicast(&init_net, skb_cur, s->pid); if (rc == -ECONNREFUSED) { s->valid = 0; delcount++; @@ -418,7 +418,7 @@ static int cgroupstats_user_cmd(struct sk_buff *skb, struct genl_info *info) goto err; } - rc = send_reply(rep_skb, info->snd_pid); + rc = send_reply(rep_skb, info); err: fput_light(file, fput_needed); @@ -487,7 +487,7 @@ free_return_rc: } else goto err; - return send_reply(rep_skb, info->snd_pid); + return send_reply(rep_skb, info); err: nlmsg_free(rep_skb); return rc; diff --git a/kernel/time/Makefile b/kernel/time/Makefile index 905b0b50792..0b0a6366c9d 100644 --- a/kernel/time/Makefile +++ b/kernel/time/Makefile @@ -1,4 +1,4 @@ -obj-y += timekeeping.o ntp.o clocksource.o jiffies.o timer_list.o +obj-y += timekeeping.o ntp.o clocksource.o jiffies.o timer_list.o timecompare.o obj-$(CONFIG_GENERIC_CLOCKEVENTS_BUILD) += clockevents.o obj-$(CONFIG_GENERIC_CLOCKEVENTS) += tick-common.o diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index ea2f48af83c..620b58abdc3 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -18,6 +18,7 @@ #include <linux/notifier.h> #include <linux/smp.h> #include <linux/sysdev.h> +#include <linux/tick.h> /* The registered clock event devices */ static LIST_HEAD(clockevent_devices); @@ -54,6 +55,7 @@ unsigned long clockevent_delta2ns(unsigned long latch, return (unsigned long) clc; } +EXPORT_SYMBOL_GPL(clockevent_delta2ns); /** * clockevents_set_mode - set the operating mode of a clock event device @@ -68,6 +70,17 @@ void clockevents_set_mode(struct clock_event_device *dev, if (dev->mode != mode) { dev->set_mode(mode, dev); dev->mode = mode; + + /* + * A nsec2cyc multiplicator of 0 is invalid and we'd crash + * on it, so fix it up and emit a warning: + */ + if (mode == CLOCK_EVT_MODE_ONESHOT) { + if (unlikely(!dev->mult)) { + dev->mult = 1; + WARN_ON(1); + } + } } } @@ -124,11 +137,12 @@ int clockevents_program_event(struct clock_event_device *dev, ktime_t expires, */ int clockevents_register_notifier(struct notifier_block *nb) { + unsigned long flags; int ret; - spin_lock(&clockevents_lock); + spin_lock_irqsave(&clockevents_lock, flags); ret = raw_notifier_chain_register(&clockevents_chain, nb); - spin_unlock(&clockevents_lock); + spin_unlock_irqrestore(&clockevents_lock, flags); return ret; } @@ -165,26 +179,20 @@ static void clockevents_notify_released(void) */ void clockevents_register_device(struct clock_event_device *dev) { + unsigned long flags; + BUG_ON(dev->mode != CLOCK_EVT_MODE_UNUSED); BUG_ON(!dev->cpumask); - /* - * A nsec2cyc multiplicator of 0 is invalid and we'd crash - * on it, so fix it up and emit a warning: - */ - if (unlikely(!dev->mult)) { - dev->mult = 1; - WARN_ON(1); - } - - spin_lock(&clockevents_lock); + spin_lock_irqsave(&clockevents_lock, flags); list_add(&dev->list, &clockevent_devices); clockevents_do_notify(CLOCK_EVT_NOTIFY_ADD, dev); clockevents_notify_released(); - spin_unlock(&clockevents_lock); + spin_unlock_irqrestore(&clockevents_lock, flags); } +EXPORT_SYMBOL_GPL(clockevents_register_device); /* * Noop handler when we shut down an event device @@ -230,8 +238,9 @@ void clockevents_exchange_device(struct clock_event_device *old, void clockevents_notify(unsigned long reason, void *arg) { struct list_head *node, *tmp; + unsigned long flags; - spin_lock(&clockevents_lock); + spin_lock_irqsave(&clockevents_lock, flags); clockevents_do_notify(reason, arg); switch (reason) { @@ -246,7 +255,7 @@ void clockevents_notify(unsigned long reason, void *arg) default: break; } - spin_unlock(&clockevents_lock); + spin_unlock_irqrestore(&clockevents_lock, flags); } EXPORT_SYMBOL_GPL(clockevents_notify); #endif diff --git a/kernel/time/clocksource.c b/kernel/time/clocksource.c index ca89e1593f0..7466cb81125 100644 --- a/kernel/time/clocksource.c +++ b/kernel/time/clocksource.c @@ -31,6 +31,82 @@ #include <linux/sched.h> /* for spin_unlock_irq() using preempt_count() m68k */ #include <linux/tick.h> +void timecounter_init(struct timecounter *tc, + const struct cyclecounter *cc, + u64 start_tstamp) +{ + tc->cc = cc; + tc->cycle_last = cc->read(cc); + tc->nsec = start_tstamp; +} +EXPORT_SYMBOL(timecounter_init); + +/** + * timecounter_read_delta - get nanoseconds since last call of this function + * @tc: Pointer to time counter + * + * When the underlying cycle counter runs over, this will be handled + * correctly as long as it does not run over more than once between + * calls. + * + * The first call to this function for a new time counter initializes + * the time tracking and returns an undefined result. + */ +static u64 timecounter_read_delta(struct timecounter *tc) +{ + cycle_t cycle_now, cycle_delta; + u64 ns_offset; + + /* read cycle counter: */ + cycle_now = tc->cc->read(tc->cc); + + /* calculate the delta since the last timecounter_read_delta(): */ + cycle_delta = (cycle_now - tc->cycle_last) & tc->cc->mask; + + /* convert to nanoseconds: */ + ns_offset = cyclecounter_cyc2ns(tc->cc, cycle_delta); + + /* update time stamp of timecounter_read_delta() call: */ + tc->cycle_last = cycle_now; + + return ns_offset; +} + +u64 timecounter_read(struct timecounter *tc) +{ + u64 nsec; + + /* increment time by nanoseconds since last call */ + nsec = timecounter_read_delta(tc); + nsec += tc->nsec; + tc->nsec = nsec; + + return nsec; +} +EXPORT_SYMBOL(timecounter_read); + +u64 timecounter_cyc2time(struct timecounter *tc, + cycle_t cycle_tstamp) +{ + u64 cycle_delta = (cycle_tstamp - tc->cycle_last) & tc->cc->mask; + u64 nsec; + + /* + * Instead of always treating cycle_tstamp as more recent + * than tc->cycle_last, detect when it is too far in the + * future and treat it as old time stamp instead. + */ + if (cycle_delta > tc->cc->mask / 2) { + cycle_delta = (tc->cycle_last - cycle_tstamp) & tc->cc->mask; + nsec = tc->nsec - cyclecounter_cyc2ns(tc->cc, cycle_delta); + } else { + nsec = cyclecounter_cyc2ns(tc->cc, cycle_delta) + tc->nsec; + } + + return nsec; +} +EXPORT_SYMBOL(timecounter_cyc2time); + /* XXX - Would like a better way for initializing curr_clocksource */ extern struct clocksource clocksource_jiffies; @@ -105,12 +181,12 @@ static void clocksource_watchdog(unsigned long data) resumed = test_and_clear_bit(0, &watchdog_resumed); - wdnow = watchdog->read(); + wdnow = watchdog->read(watchdog); wd_nsec = cyc2ns(watchdog, (wdnow - watchdog_last) & watchdog->mask); watchdog_last = wdnow; list_for_each_entry_safe(cs, tmp, &watchdog_list, wd_list) { - csnow = cs->read(); + csnow = cs->read(cs); if (unlikely(resumed)) { cs->wd_last = csnow; @@ -171,7 +247,7 @@ static void clocksource_check_watchdog(struct clocksource *cs) list_add(&cs->wd_list, &watchdog_list); if (!started && watchdog) { - watchdog_last = watchdog->read(); + watchdog_last = watchdog->read(watchdog); watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL; add_timer_on(&watchdog_timer, cpumask_first(cpu_online_mask)); @@ -192,7 +268,7 @@ static void clocksource_check_watchdog(struct clocksource *cs) cse->flags &= ~CLOCK_SOURCE_WATCHDOG; /* Start if list is not empty */ if (!list_empty(&watchdog_list)) { - watchdog_last = watchdog->read(); + watchdog_last = watchdog->read(watchdog); watchdog_timer.expires = jiffies + WATCHDOG_INTERVAL; add_timer_on(&watchdog_timer, @@ -326,9 +402,6 @@ int clocksource_register(struct clocksource *c) unsigned long flags; int ret; - /* save mult_orig on registration */ - c->mult_orig = c->mult; - spin_lock_irqsave(&clocksource_lock, flags); ret = clocksource_enqueue(c); if (!ret) @@ -436,6 +509,18 @@ static ssize_t sysfs_override_clocksource(struct sys_device *dev, } } + /* + * Check to make sure we don't switch to a non-highres capable + * clocksource if the tick code is in oneshot mode (highres or nohz) + */ + if (tick_oneshot_mode_active() && ovr && + !(ovr->flags & CLOCK_SOURCE_VALID_FOR_HRES)) { + printk(KERN_WARNING "%s clocksource is not HRT compatible. " + "Cannot switch while in HRT/NOHZ mode\n", ovr->name); + ovr = NULL; + override_name[0] = 0; + } + /* Reselect, when the override name has changed */ if (ovr != clocksource_override) { clocksource_override = ovr; @@ -464,7 +549,13 @@ sysfs_show_available_clocksources(struct sys_device *dev, spin_lock_irq(&clocksource_lock); list_for_each_entry(src, &clocksource_list, list) { - count += snprintf(buf + count, + /* + * Don't show non-HRES clocksource if the tick code is + * in one shot mode (highres=on or nohz=on) + */ + if (!tick_oneshot_mode_active() || + (src->flags & CLOCK_SOURCE_VALID_FOR_HRES)) + count += snprintf(buf + count, max((ssize_t)PAGE_SIZE - count, (ssize_t)0), "%s ", src->name); } diff --git a/kernel/time/jiffies.c b/kernel/time/jiffies.c index 06f197560f3..c3f6c30816e 100644 --- a/kernel/time/jiffies.c +++ b/kernel/time/jiffies.c @@ -50,7 +50,7 @@ */ #define JIFFIES_SHIFT 8 -static cycle_t jiffies_read(void) +static cycle_t jiffies_read(struct clocksource *cs) { return (cycle_t) jiffies; } diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c index f5f793d9241..7fc64375ff4 100644 --- a/kernel/time/ntp.c +++ b/kernel/time/ntp.c @@ -1,71 +1,129 @@ /* - * linux/kernel/time/ntp.c - * * NTP state machine interfaces and logic. * * This code was mainly moved from kernel/timer.c and kernel/time.c * Please see those files for relevant copyright info and historical * changelogs. */ - -#include <linux/mm.h> -#include <linux/time.h> -#include <linux/timex.h> -#include <linux/jiffies.h> -#include <linux/hrtimer.h> #include <linux/capability.h> -#include <linux/math64.h> #include <linux/clocksource.h> #include <linux/workqueue.h> -#include <asm/timex.h> +#include <linux/hrtimer.h> +#include <linux/jiffies.h> +#include <linux/math64.h> +#include <linux/timex.h> +#include <linux/time.h> +#include <linux/mm.h> /* - * Timekeeping variables + * NTP timekeeping variables: */ -unsigned long tick_usec = TICK_USEC; /* USER_HZ period (usec) */ -unsigned long tick_nsec; /* ACTHZ period (nsec) */ -u64 tick_length; -static u64 tick_length_base; -static struct hrtimer leap_timer; +/* USER_HZ period (usecs): */ +unsigned long tick_usec = TICK_USEC; -#define MAX_TICKADJ 500 /* microsecs */ -#define MAX_TICKADJ_SCALED (((u64)(MAX_TICKADJ * NSEC_PER_USEC) << \ - NTP_SCALE_SHIFT) / NTP_INTERVAL_FREQ) +/* ACTHZ period (nsecs): */ +unsigned long tick_nsec; + +u64 tick_length; +static u64 tick_length_base; + +static struct hrtimer leap_timer; + +#define MAX_TICKADJ 500LL /* usecs */ +#define MAX_TICKADJ_SCALED \ + (((MAX_TICKADJ * NSEC_PER_USEC) << NTP_SCALE_SHIFT) / NTP_INTERVAL_FREQ) /* * phase-lock loop variables */ -/* TIME_ERROR prevents overwriting the CMOS clock */ -static int time_state = TIME_OK; /* clock synchronization status */ -int time_status = STA_UNSYNC; /* clock status bits */ -static long time_tai; /* TAI offset (s) */ -static s64 time_offset; /* time adjustment (ns) */ -static long time_constant = 2; /* pll time constant */ -long time_maxerror = NTP_PHASE_LIMIT; /* maximum error (us) */ -long time_esterror = NTP_PHASE_LIMIT; /* estimated error (us) */ -static s64 time_freq; /* frequency offset (scaled ns/s)*/ -static long time_reftime; /* time at last adjustment (s) */ -long time_adjust; -static long ntp_tick_adj; +/* + * clock synchronization status + * + * (TIME_ERROR prevents overwriting the CMOS clock) + */ +static int time_state = TIME_OK; + +/* clock status bits: */ +int time_status = STA_UNSYNC; + +/* TAI offset (secs): */ +static long time_tai; + +/* time adjustment (nsecs): */ +static s64 time_offset; + +/* pll time constant: */ +static long time_constant = 2; + +/* maximum error (usecs): */ +long time_maxerror = NTP_PHASE_LIMIT; + +/* estimated error (usecs): */ +long time_esterror = NTP_PHASE_LIMIT; + +/* frequency offset (scaled nsecs/secs): */ +static s64 time_freq; + +/* time at last adjustment (secs): */ +static long time_reftime; + +long time_adjust; + +/* constant (boot-param configurable) NTP tick adjustment (upscaled) */ +static s64 ntp_tick_adj; + +/* + * NTP methods: + */ + +/* + * Update (tick_length, tick_length_base, tick_nsec), based + * on (tick_usec, ntp_tick_adj, time_freq): + */ static void ntp_update_frequency(void) { - u64 second_length = (u64)(tick_usec * NSEC_PER_USEC * USER_HZ) - << NTP_SCALE_SHIFT; - second_length += (s64)ntp_tick_adj << NTP_SCALE_SHIFT; - second_length += time_freq; + u64 second_length; + u64 new_base; + + second_length = (u64)(tick_usec * NSEC_PER_USEC * USER_HZ) + << NTP_SCALE_SHIFT; + + second_length += ntp_tick_adj; + second_length += time_freq; - tick_length_base = second_length; + tick_nsec = div_u64(second_length, HZ) >> NTP_SCALE_SHIFT; + new_base = div_u64(second_length, NTP_INTERVAL_FREQ); - tick_nsec = div_u64(second_length, HZ) >> NTP_SCALE_SHIFT; - tick_length_base = div_u64(tick_length_base, NTP_INTERVAL_FREQ); + /* + * Don't wait for the next second_overflow, apply + * the change to the tick length immediately: + */ + tick_length += new_base - tick_length_base; + tick_length_base = new_base; +} + +static inline s64 ntp_update_offset_fll(s64 offset64, long secs) +{ + time_status &= ~STA_MODE; + + if (secs < MINSEC) + return 0; + + if (!(time_status & STA_FLL) && (secs <= MAXSEC)) + return 0; + + time_status |= STA_MODE; + + return div_s64(offset64 << (NTP_SCALE_SHIFT - SHIFT_FLL), secs); } static void ntp_update_offset(long offset) { - long mtemp; s64 freq_adj; + s64 offset64; + long secs; if (!(time_status & STA_PLL)) return; @@ -84,24 +142,23 @@ static void ntp_update_offset(long offset) * Select how the frequency is to be controlled * and in which mode (PLL or FLL). */ - if (time_status & STA_FREQHOLD || time_reftime == 0) - time_reftime = xtime.tv_sec; - mtemp = xtime.tv_sec - time_reftime; + secs = xtime.tv_sec - time_reftime; + if (unlikely(time_status & STA_FREQHOLD)) + secs = 0; + time_reftime = xtime.tv_sec; - freq_adj = (s64)offset * mtemp; - freq_adj <<= NTP_SCALE_SHIFT - 2 * (SHIFT_PLL + 2 + time_constant); - time_status &= ~STA_MODE; - if (mtemp >= MINSEC && (time_status & STA_FLL || mtemp > MAXSEC)) { - freq_adj += div_s64((s64)offset << (NTP_SCALE_SHIFT - SHIFT_FLL), - mtemp); - time_status |= STA_MODE; - } - freq_adj += time_freq; - freq_adj = min(freq_adj, MAXFREQ_SCALED); - time_freq = max(freq_adj, -MAXFREQ_SCALED); + offset64 = offset; + freq_adj = (offset64 * secs) << + (NTP_SCALE_SHIFT - 2 * (SHIFT_PLL + 2 + time_constant)); - time_offset = div_s64((s64)offset << NTP_SCALE_SHIFT, NTP_INTERVAL_FREQ); + freq_adj += ntp_update_offset_fll(offset64, secs); + + freq_adj = min(freq_adj + time_freq, MAXFREQ_SCALED); + + time_freq = max(freq_adj, -MAXFREQ_SCALED); + + time_offset = div_s64(offset64 << NTP_SCALE_SHIFT, NTP_INTERVAL_FREQ); } /** @@ -111,15 +168,15 @@ static void ntp_update_offset(long offset) */ void ntp_clear(void) { - time_adjust = 0; /* stop active adjtime() */ - time_status |= STA_UNSYNC; - time_maxerror = NTP_PHASE_LIMIT; - time_esterror = NTP_PHASE_LIMIT; + time_adjust = 0; /* stop active adjtime() */ + time_status |= STA_UNSYNC; + time_maxerror = NTP_PHASE_LIMIT; + time_esterror = NTP_PHASE_LIMIT; ntp_update_frequency(); - tick_length = tick_length_base; - time_offset = 0; + tick_length = tick_length_base; + time_offset = 0; } /* @@ -140,8 +197,8 @@ static enum hrtimer_restart ntp_leap_second(struct hrtimer *timer) xtime.tv_sec--; wall_to_monotonic.tv_sec++; time_state = TIME_OOP; - printk(KERN_NOTICE "Clock: " - "inserting leap second 23:59:60 UTC\n"); + printk(KERN_NOTICE + "Clock: inserting leap second 23:59:60 UTC\n"); hrtimer_add_expires_ns(&leap_timer, NSEC_PER_SEC); res = HRTIMER_RESTART; break; @@ -150,8 +207,8 @@ static enum hrtimer_restart ntp_leap_second(struct hrtimer *timer) time_tai--; wall_to_monotonic.tv_sec--; time_state = TIME_WAIT; - printk(KERN_NOTICE "Clock: " - "deleting leap second 23:59:59 UTC\n"); + printk(KERN_NOTICE + "Clock: deleting leap second 23:59:59 UTC\n"); break; case TIME_OOP: time_tai++; @@ -179,7 +236,7 @@ static enum hrtimer_restart ntp_leap_second(struct hrtimer *timer) */ void second_overflow(void) { - s64 time_adj; + s64 delta; /* Bump the maxerror field */ time_maxerror += MAXFREQ / NSEC_PER_USEC; @@ -192,24 +249,30 @@ void second_overflow(void) * Compute the phase adjustment for the next second. The offset is * reduced by a fixed factor times the time constant. */ - tick_length = tick_length_base; - time_adj = shift_right(time_offset, SHIFT_PLL + time_constant); - time_offset -= time_adj; - tick_length += time_adj; - - if (unlikely(time_adjust)) { - if (time_adjust > MAX_TICKADJ) { - time_adjust -= MAX_TICKADJ; - tick_length += MAX_TICKADJ_SCALED; - } else if (time_adjust < -MAX_TICKADJ) { - time_adjust += MAX_TICKADJ; - tick_length -= MAX_TICKADJ_SCALED; - } else { - tick_length += (s64)(time_adjust * NSEC_PER_USEC / - NTP_INTERVAL_FREQ) << NTP_SCALE_SHIFT; - time_adjust = 0; - } + tick_length = tick_length_base; + + delta = shift_right(time_offset, SHIFT_PLL + time_constant); + time_offset -= delta; + tick_length += delta; + + if (!time_adjust) + return; + + if (time_adjust > MAX_TICKADJ) { + time_adjust -= MAX_TICKADJ; + tick_length += MAX_TICKADJ_SCALED; + return; } + + if (time_adjust < -MAX_TICKADJ) { + time_adjust += MAX_TICKADJ; + tick_length -= MAX_TICKADJ_SCALED; + return; + } + + tick_length += (s64)(time_adjust * NSEC_PER_USEC / NTP_INTERVAL_FREQ) + << NTP_SCALE_SHIFT; + time_adjust = 0; } #ifdef CONFIG_GENERIC_CMOS_UPDATE @@ -233,12 +296,13 @@ static void sync_cmos_clock(struct work_struct *work) * This code is run on a timer. If the clock is set, that timer * may not expire at the correct time. Thus, we adjust... */ - if (!ntp_synced()) + if (!ntp_synced()) { /* * Not synced, exit, do not restart a timer (if one is * running, let it run out). */ return; + } getnstimeofday(&now); if (abs(now.tv_nsec - (NSEC_PER_SEC / 2)) <= tick_nsec / 2) @@ -270,7 +334,116 @@ static void notify_cmos_timer(void) static inline void notify_cmos_timer(void) { } #endif -/* adjtimex mainly allows reading (and writing, if superuser) of +/* + * Start the leap seconds timer: + */ +static inline void ntp_start_leap_timer(struct timespec *ts) +{ + long now = ts->tv_sec; + + if (time_status & STA_INS) { + time_state = TIME_INS; + now += 86400 - now % 86400; + hrtimer_start(&leap_timer, ktime_set(now, 0), HRTIMER_MODE_ABS); + + return; + } + + if (time_status & STA_DEL) { + time_state = TIME_DEL; + now += 86400 - (now + 1) % 86400; + hrtimer_start(&leap_timer, ktime_set(now, 0), HRTIMER_MODE_ABS); + } +} + +/* + * Propagate a new txc->status value into the NTP state: + */ +static inline void process_adj_status(struct timex *txc, struct timespec *ts) +{ + if ((time_status & STA_PLL) && !(txc->status & STA_PLL)) { + time_state = TIME_OK; + time_status = STA_UNSYNC; + } + + /* + * If we turn on PLL adjustments then reset the + * reference time to current time. + */ + if (!(time_status & STA_PLL) && (txc->status & STA_PLL)) + time_reftime = xtime.tv_sec; + + /* only set allowed bits */ + time_status &= STA_RONLY; + time_status |= txc->status & ~STA_RONLY; + + switch (time_state) { + case TIME_OK: + ntp_start_leap_timer(ts); + break; + case TIME_INS: + case TIME_DEL: + time_state = TIME_OK; + ntp_start_leap_timer(ts); + case TIME_WAIT: + if (!(time_status & (STA_INS | STA_DEL))) + time_state = TIME_OK; + break; + case TIME_OOP: + hrtimer_restart(&leap_timer); + break; + } +} +/* + * Called with the xtime lock held, so we can access and modify + * all the global NTP state: + */ +static inline void process_adjtimex_modes(struct timex *txc, struct timespec *ts) +{ + if (txc->modes & ADJ_STATUS) + process_adj_status(txc, ts); + + if (txc->modes & ADJ_NANO) + time_status |= STA_NANO; + + if (txc->modes & ADJ_MICRO) + time_status &= ~STA_NANO; + + if (txc->modes & ADJ_FREQUENCY) { + time_freq = txc->freq * PPM_SCALE; + time_freq = min(time_freq, MAXFREQ_SCALED); + time_freq = max(time_freq, -MAXFREQ_SCALED); + } + + if (txc->modes & ADJ_MAXERROR) + time_maxerror = txc->maxerror; + + if (txc->modes & ADJ_ESTERROR) + time_esterror = txc->esterror; + + if (txc->modes & ADJ_TIMECONST) { + time_constant = txc->constant; + if (!(time_status & STA_NANO)) + time_constant += 4; + time_constant = min(time_constant, (long)MAXTC); + time_constant = max(time_constant, 0l); + } + + if (txc->modes & ADJ_TAI && txc->constant > 0) + time_tai = txc->constant; + + if (txc->modes & ADJ_OFFSET) + ntp_update_offset(txc->offset); + + if (txc->modes & ADJ_TICK) + tick_usec = txc->tick; + + if (txc->modes & (ADJ_TICK|ADJ_FREQUENCY|ADJ_OFFSET)) + ntp_update_frequency(); +} + +/* + * adjtimex mainly allows reading (and writing, if superuser) of * kernel time-keeping variables. used by xntpd. */ int do_adjtimex(struct timex *txc) @@ -291,11 +464,14 @@ int do_adjtimex(struct timex *txc) if (txc->modes && !capable(CAP_SYS_TIME)) return -EPERM; - /* if the quartz is off by more than 10% something is VERY wrong! */ + /* + * if the quartz is off by more than 10% then + * something is VERY wrong! + */ if (txc->modes & ADJ_TICK && (txc->tick < 900000/USER_HZ || txc->tick > 1100000/USER_HZ)) - return -EINVAL; + return -EINVAL; if (txc->modes & ADJ_STATUS && time_state != TIME_OK) hrtimer_cancel(&leap_timer); @@ -305,7 +481,6 @@ int do_adjtimex(struct timex *txc) write_seqlock_irq(&xtime_lock); - /* If there are input parameters, then process them */ if (txc->modes & ADJ_ADJTIME) { long save_adjust = time_adjust; @@ -315,98 +490,24 @@ int do_adjtimex(struct timex *txc) ntp_update_frequency(); } txc->offset = save_adjust; - goto adj_done; - } - if (txc->modes) { - long sec; - - if (txc->modes & ADJ_STATUS) { - if ((time_status & STA_PLL) && - !(txc->status & STA_PLL)) { - time_state = TIME_OK; - time_status = STA_UNSYNC; - } - /* only set allowed bits */ - time_status &= STA_RONLY; - time_status |= txc->status & ~STA_RONLY; - - switch (time_state) { - case TIME_OK: - start_timer: - sec = ts.tv_sec; - if (time_status & STA_INS) { - time_state = TIME_INS; - sec += 86400 - sec % 86400; - hrtimer_start(&leap_timer, ktime_set(sec, 0), HRTIMER_MODE_ABS); - } else if (time_status & STA_DEL) { - time_state = TIME_DEL; - sec += 86400 - (sec + 1) % 86400; - hrtimer_start(&leap_timer, ktime_set(sec, 0), HRTIMER_MODE_ABS); - } - break; - case TIME_INS: - case TIME_DEL: - time_state = TIME_OK; - goto start_timer; - break; - case TIME_WAIT: - if (!(time_status & (STA_INS | STA_DEL))) - time_state = TIME_OK; - break; - case TIME_OOP: - hrtimer_restart(&leap_timer); - break; - } - } - - if (txc->modes & ADJ_NANO) - time_status |= STA_NANO; - if (txc->modes & ADJ_MICRO) - time_status &= ~STA_NANO; - - if (txc->modes & ADJ_FREQUENCY) { - time_freq = (s64)txc->freq * PPM_SCALE; - time_freq = min(time_freq, MAXFREQ_SCALED); - time_freq = max(time_freq, -MAXFREQ_SCALED); - } - - if (txc->modes & ADJ_MAXERROR) - time_maxerror = txc->maxerror; - if (txc->modes & ADJ_ESTERROR) - time_esterror = txc->esterror; - - if (txc->modes & ADJ_TIMECONST) { - time_constant = txc->constant; - if (!(time_status & STA_NANO)) - time_constant += 4; - time_constant = min(time_constant, (long)MAXTC); - time_constant = max(time_constant, 0l); - } - - if (txc->modes & ADJ_TAI && txc->constant > 0) - time_tai = txc->constant; - - if (txc->modes & ADJ_OFFSET) - ntp_update_offset(txc->offset); - if (txc->modes & ADJ_TICK) - tick_usec = txc->tick; + } else { - if (txc->modes & (ADJ_TICK|ADJ_FREQUENCY|ADJ_OFFSET)) - ntp_update_frequency(); - } + /* If there are input parameters, then process them: */ + if (txc->modes) + process_adjtimex_modes(txc, &ts); - txc->offset = shift_right(time_offset * NTP_INTERVAL_FREQ, + txc->offset = shift_right(time_offset * NTP_INTERVAL_FREQ, NTP_SCALE_SHIFT); - if (!(time_status & STA_NANO)) - txc->offset /= NSEC_PER_USEC; + if (!(time_status & STA_NANO)) + txc->offset /= NSEC_PER_USEC; + } -adj_done: result = time_state; /* mostly `TIME_OK' */ if (time_status & (STA_UNSYNC|STA_CLOCKERR)) result = TIME_ERROR; txc->freq = shift_right((time_freq >> PPM_SCALE_INV_SHIFT) * - (s64)PPM_SCALE_INV, NTP_SCALE_SHIFT); + PPM_SCALE_INV, NTP_SCALE_SHIFT); txc->maxerror = time_maxerror; txc->esterror = time_esterror; txc->status = time_status; @@ -425,6 +526,7 @@ adj_done: txc->calcnt = 0; txc->errcnt = 0; txc->stbcnt = 0; + write_sequnlock_irq(&xtime_lock); txc->time.tv_sec = ts.tv_sec; @@ -440,6 +542,8 @@ adj_done: static int __init ntp_tick_adj_setup(char *str) { ntp_tick_adj = simple_strtol(str, NULL, 0); + ntp_tick_adj <<= NTP_SCALE_SHIFT; + return 1; } diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c index 118a3b3b3f9..c2ec25087a3 100644 --- a/kernel/time/tick-broadcast.c +++ b/kernel/time/tick-broadcast.c @@ -27,7 +27,7 @@ * timer stops in C3 state. */ -struct tick_device tick_broadcast_device; +static struct tick_device tick_broadcast_device; /* FIXME: Use cpumask_var_t. */ static DECLARE_BITMAP(tick_broadcast_mask, NR_CPUS); static DECLARE_BITMAP(tmpmask, NR_CPUS); @@ -205,11 +205,11 @@ static void tick_handle_periodic_broadcast(struct clock_event_device *dev) * Powerstate information: The system enters/leaves a state, where * affected devices might stop */ -static void tick_do_broadcast_on_off(void *why) +static void tick_do_broadcast_on_off(unsigned long *reason) { struct clock_event_device *bc, *dev; struct tick_device *td; - unsigned long flags, *reason = why; + unsigned long flags; int cpu, bc_stopped; spin_lock_irqsave(&tick_broadcast_lock, flags); @@ -276,8 +276,7 @@ void tick_broadcast_on_off(unsigned long reason, int *oncpu) printk(KERN_ERR "tick-broadcast: ignoring broadcast for " "offline CPU #%d\n", *oncpu); else - smp_call_function_single(*oncpu, tick_do_broadcast_on_off, - &reason, 1); + tick_do_broadcast_on_off(&reason); } /* diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c index 63e05d423a0..83c4417b6a3 100644 --- a/kernel/time/tick-common.c +++ b/kernel/time/tick-common.c @@ -93,7 +93,17 @@ void tick_handle_periodic(struct clock_event_device *dev) for (;;) { if (!clockevents_program_event(dev, next, ktime_get())) return; - tick_periodic(cpu); + /* + * Have to be careful here. If we're in oneshot mode, + * before we call tick_periodic() in a loop, we need + * to be sure we're using a real hardware clocksource. + * Otherwise we could get trapped in an infinite + * loop, as the tick_periodic() increments jiffies, + * when then will increment time, posibly causing + * the loop to trigger again and again. + */ + if (timekeeping_valid_for_hres()) + tick_periodic(cpu); next = ktime_add(next, tick_period); } } @@ -274,6 +284,21 @@ out_bc: } /* + * Transfer the do_timer job away from a dying cpu. + * + * Called with interrupts disabled. + */ +static void tick_handover_do_timer(int *cpup) +{ + if (*cpup == tick_do_timer_cpu) { + int cpu = cpumask_first(cpu_online_mask); + + tick_do_timer_cpu = (cpu < nr_cpu_ids) ? cpu : + TICK_DO_TIMER_NONE; + } +} + +/* * Shutdown an event device on a given cpu: * * This is called on a life CPU, when a CPU is dead. So we cannot @@ -297,13 +322,6 @@ static void tick_shutdown(unsigned int *cpup) clockevents_exchange_device(dev, NULL); td->evtdev = NULL; } - /* Transfer the do_timer job away from this cpu */ - if (*cpup == tick_do_timer_cpu) { - int cpu = cpumask_first(cpu_online_mask); - - tick_do_timer_cpu = (cpu < nr_cpu_ids) ? cpu : - TICK_DO_TIMER_NONE; - } spin_unlock_irqrestore(&tick_device_lock, flags); } @@ -357,6 +375,10 @@ static int tick_notify(struct notifier_block *nb, unsigned long reason, tick_broadcast_oneshot_control(reason); break; + case CLOCK_EVT_NOTIFY_CPU_DYING: + tick_handover_do_timer(dev); + break; + case CLOCK_EVT_NOTIFY_CPU_DEAD: tick_shutdown_broadcast_oneshot(dev); tick_shutdown_broadcast(dev); diff --git a/kernel/time/tick-oneshot.c b/kernel/time/tick-oneshot.c index 2e8de678e76..a96c0e2b89c 100644 --- a/kernel/time/tick-oneshot.c +++ b/kernel/time/tick-oneshot.c @@ -128,6 +128,23 @@ int tick_switch_to_oneshot(void (*handler)(struct clock_event_device *)) return 0; } +/** + * tick_check_oneshot_mode - check whether the system is in oneshot mode + * + * returns 1 when either nohz or highres are enabled. otherwise 0. + */ +int tick_oneshot_mode_active(void) +{ + unsigned long flags; + int ret; + + local_irq_save(flags); + ret = __get_cpu_var(tick_cpu_device).mode == TICKDEV_MODE_ONESHOT; + local_irq_restore(flags); + + return ret; +} + #ifdef CONFIG_HIGH_RES_TIMERS /** * tick_init_highres - switch to high resolution mode diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c index 1b6c05bd0d0..e0f59a21c06 100644 --- a/kernel/time/tick-sched.c +++ b/kernel/time/tick-sched.c @@ -134,7 +134,7 @@ __setup("nohz=", setup_tick_nohz); * value. We do this unconditionally on any cpu, as we don't know whether the * cpu, which has the update task assigned is in a long sleep. */ -void tick_nohz_update_jiffies(void) +static void tick_nohz_update_jiffies(void) { int cpu = smp_processor_id(); struct tick_sched *ts = &per_cpu(tick_cpu_sched, cpu); @@ -222,6 +222,15 @@ void tick_nohz_stop_sched_tick(int inidle) cpu = smp_processor_id(); ts = &per_cpu(tick_cpu_sched, cpu); + + /* + * Call to tick_nohz_start_idle stops the last_update_time from being + * updated. Thus, it must not be called in the event we are called from + * irq_exit() with the prior state different than idle. + */ + if (!inidle && !ts->inidle) + goto end; + now = tick_nohz_start_idle(ts); /* @@ -239,9 +248,6 @@ void tick_nohz_stop_sched_tick(int inidle) if (unlikely(ts->nohz_mode == NOHZ_MODE_INACTIVE)) goto end; - if (!inidle && !ts->inidle) - goto end; - ts->inidle = 1; if (need_resched()) @@ -349,7 +355,7 @@ void tick_nohz_stop_sched_tick(int inidle) if (ts->nohz_mode == NOHZ_MODE_HIGHRES) { hrtimer_start(&ts->sched_timer, expires, - HRTIMER_MODE_ABS); + HRTIMER_MODE_ABS_PINNED); /* Check, if the timer was already in the past */ if (hrtimer_active(&ts->sched_timer)) goto out; @@ -395,7 +401,7 @@ static void tick_nohz_restart(struct tick_sched *ts, ktime_t now) if (ts->nohz_mode == NOHZ_MODE_HIGHRES) { hrtimer_start_expires(&ts->sched_timer, - HRTIMER_MODE_ABS); + HRTIMER_MODE_ABS_PINNED); /* Check, if the timer was already in the past */ if (hrtimer_active(&ts->sched_timer)) break; @@ -698,7 +704,8 @@ void tick_setup_sched_timer(void) for (;;) { hrtimer_forward(&ts->sched_timer, now, tick_period); - hrtimer_start_expires(&ts->sched_timer, HRTIMER_MODE_ABS); + hrtimer_start_expires(&ts->sched_timer, + HRTIMER_MODE_ABS_PINNED); /* Check, if the timer was already in the past */ if (hrtimer_active(&ts->sched_timer)) break; diff --git a/kernel/time/timecompare.c b/kernel/time/timecompare.c new file mode 100644 index 00000000000..71e7f1a1915 --- /dev/null +++ b/kernel/time/timecompare.c @@ -0,0 +1,191 @@ +/* + * Copyright (C) 2009 Intel Corporation. + * Author: Patrick Ohly <patrick.ohly@intel.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include <linux/timecompare.h> +#include <linux/module.h> +#include <linux/math64.h> + +/* + * fixed point arithmetic scale factor for skew + * + * Usually one would measure skew in ppb (parts per billion, 1e9), but + * using a factor of 2 simplifies the math. + */ +#define TIMECOMPARE_SKEW_RESOLUTION (((s64)1)<<30) + +ktime_t timecompare_transform(struct timecompare *sync, + u64 source_tstamp) +{ + u64 nsec; + + nsec = source_tstamp + sync->offset; + nsec += (s64)(source_tstamp - sync->last_update) * sync->skew / + TIMECOMPARE_SKEW_RESOLUTION; + + return ns_to_ktime(nsec); +} +EXPORT_SYMBOL(timecompare_transform); + +int timecompare_offset(struct timecompare *sync, + s64 *offset, + u64 *source_tstamp) +{ + u64 start_source = 0, end_source = 0; + struct { + s64 offset; + s64 duration_target; + } buffer[10], sample, *samples; + int counter = 0, i; + int used; + int index; + int num_samples = sync->num_samples; + + if (num_samples > sizeof(buffer)/sizeof(buffer[0])) { + samples = kmalloc(sizeof(*samples) * num_samples, GFP_ATOMIC); + if (!samples) { + samples = buffer; + num_samples = sizeof(buffer)/sizeof(buffer[0]); + } + } else { + samples = buffer; + } + + /* run until we have enough valid samples, but do not try forever */ + i = 0; + counter = 0; + while (1) { + u64 ts; + ktime_t start, end; + + start = sync->target(); + ts = timecounter_read(sync->source); + end = sync->target(); + + if (!i) + start_source = ts; + + /* ignore negative durations */ + sample.duration_target = ktime_to_ns(ktime_sub(end, start)); + if (sample.duration_target >= 0) { + /* + * assume symetric delay to and from source: + * average target time corresponds to measured + * source time + */ + sample.offset = + ktime_to_ns(ktime_add(end, start)) / 2 - + ts; + + /* simple insertion sort based on duration */ + index = counter - 1; + while (index >= 0) { + if (samples[index].duration_target < + sample.duration_target) + break; + samples[index + 1] = samples[index]; + index--; + } + samples[index + 1] = sample; + counter++; + } + + i++; + if (counter >= num_samples || i >= 100000) { + end_source = ts; + break; + } + } + + *source_tstamp = (end_source + start_source) / 2; + + /* remove outliers by only using 75% of the samples */ + used = counter * 3 / 4; + if (!used) + used = counter; + if (used) { + /* calculate average */ + s64 off = 0; + for (index = 0; index < used; index++) + off += samples[index].offset; + *offset = div_s64(off, used); + } + + if (samples && samples != buffer) + kfree(samples); + + return used; +} +EXPORT_SYMBOL(timecompare_offset); + +void __timecompare_update(struct timecompare *sync, + u64 source_tstamp) +{ + s64 offset; + u64 average_time; + + if (!timecompare_offset(sync, &offset, &average_time)) + return; + + if (!sync->last_update) { + sync->last_update = average_time; + sync->offset = offset; + sync->skew = 0; + } else { + s64 delta_nsec = average_time - sync->last_update; + + /* avoid division by negative or small deltas */ + if (delta_nsec >= 10000) { + s64 delta_offset_nsec = offset - sync->offset; + s64 skew; /* delta_offset_nsec * + TIMECOMPARE_SKEW_RESOLUTION / + delta_nsec */ + u64 divisor; + + /* div_s64() is limited to 32 bit divisor */ + skew = delta_offset_nsec * TIMECOMPARE_SKEW_RESOLUTION; + divisor = delta_nsec; + while (unlikely(divisor >= ((s64)1) << 32)) { + /* divide both by 2; beware, right shift + of negative value has undefined + behavior and can only be used for + the positive divisor */ + skew = div_s64(skew, 2); + divisor >>= 1; + } + skew = div_s64(skew, divisor); + + /* + * Calculate new overall skew as 4/16 the + * old value and 12/16 the new one. This is + * a rather arbitrary tradeoff between + * only using the latest measurement (0/16 and + * 16/16) and even more weight on past measurements. + */ +#define TIMECOMPARE_NEW_SKEW_PER_16 12 + sync->skew = + div_s64((16 - TIMECOMPARE_NEW_SKEW_PER_16) * + sync->skew + + TIMECOMPARE_NEW_SKEW_PER_16 * skew, + 16); + sync->last_update = average_time; + sync->offset = offset; + } + } +} +EXPORT_SYMBOL(__timecompare_update); diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 900f1b6598d..e8c77d9c633 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -22,7 +22,7 @@ /* * This read-write spinlock protects us from races in SMP while - * playing with xtime and avenrun. + * playing with xtime. */ __cacheline_aligned_in_smp DEFINE_SEQLOCK(xtime_lock); @@ -77,6 +77,10 @@ static void clocksource_forward_now(void) clock->cycle_last = cycle_now; nsec = cyc2ns(clock, cycle_delta); + + /* If arch requires, add in gettimeoffset() */ + nsec += arch_gettimeoffset(); + timespec_add_ns(&xtime, nsec); nsec = ((s64)cycle_delta * clock->mult_orig) >> clock->shift; @@ -111,6 +115,9 @@ void getnstimeofday(struct timespec *ts) /* convert to nanoseconds: */ nsecs = cyc2ns(clock, cycle_delta); + /* If arch requires, add in gettimeoffset() */ + nsecs += arch_gettimeoffset(); + } while (read_seqretry(&xtime_lock, seq)); timespec_add_ns(ts, nsecs); @@ -182,7 +189,7 @@ EXPORT_SYMBOL(do_settimeofday); */ static void change_clocksource(void) { - struct clocksource *new; + struct clocksource *new, *old; new = clocksource_get_next(); @@ -191,11 +198,16 @@ static void change_clocksource(void) clocksource_forward_now(); - new->raw_time = clock->raw_time; + if (clocksource_enable(new)) + return; + new->raw_time = clock->raw_time; + old = clock; clock = new; + clocksource_disable(old); + clock->cycle_last = 0; - clock->cycle_last = clocksource_read(new); + clock->cycle_last = clocksource_read(clock); clock->error = 0; clock->xtime_nsec = 0; clocksource_calculate_interval(clock, NTP_INTERVAL_LENGTH); @@ -292,6 +304,7 @@ void __init timekeeping_init(void) ntp_init(); clock = clocksource_get_next(); + clocksource_enable(clock); clocksource_calculate_interval(clock, NTP_INTERVAL_LENGTH); clock->cycle_last = clocksource_read(clock); diff --git a/kernel/time/timer_list.c b/kernel/time/timer_list.c index a999b92a127..fddd69d16e0 100644 --- a/kernel/time/timer_list.c +++ b/kernel/time/timer_list.c @@ -286,7 +286,7 @@ static int __init init_timer_list_procfs(void) { struct proc_dir_entry *pe; - pe = proc_create("timer_list", 0644, NULL, &timer_list_fops); + pe = proc_create("timer_list", 0444, NULL, &timer_list_fops); if (!pe) return -ENOMEM; return 0; diff --git a/kernel/time/timer_stats.c b/kernel/time/timer_stats.c index c994530d166..4cde8b9c716 100644 --- a/kernel/time/timer_stats.c +++ b/kernel/time/timer_stats.c @@ -96,7 +96,7 @@ static DEFINE_MUTEX(show_mutex); /* * Collection status, active/inactive: */ -static int __read_mostly active; +int __read_mostly timer_stats_active; /* * Beginning/end timestamps of measurement: @@ -242,7 +242,7 @@ void timer_stats_update_stats(void *timer, pid_t pid, void *startf, struct entry *entry, input; unsigned long flags; - if (likely(!active)) + if (likely(!timer_stats_active)) return; lock = &per_cpu(lookup_lock, raw_smp_processor_id()); @@ -254,7 +254,7 @@ void timer_stats_update_stats(void *timer, pid_t pid, void *startf, input.timer_flag = timer_flag; spin_lock_irqsave(lock, flags); - if (!active) + if (!timer_stats_active) goto out_unlock; entry = tstat_lookup(&input, comm); @@ -290,7 +290,7 @@ static int tstats_show(struct seq_file *m, void *v) /* * If still active then calculate up to now: */ - if (active) + if (timer_stats_active) time_stop = ktime_get(); time = ktime_sub(time_stop, time_start); @@ -368,18 +368,18 @@ static ssize_t tstats_write(struct file *file, const char __user *buf, mutex_lock(&show_mutex); switch (ctl[0]) { case '0': - if (active) { - active = 0; + if (timer_stats_active) { + timer_stats_active = 0; time_stop = ktime_get(); sync_access(); } break; case '1': - if (!active) { + if (!timer_stats_active) { reset_entries(); time_start = ktime_get(); smp_mb(); - active = 1; + timer_stats_active = 1; } break; default: diff --git a/kernel/timer.c b/kernel/timer.c index 13dd64fe143..a3d25f41501 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -37,6 +37,8 @@ #include <linux/delay.h> #include <linux/tick.h> #include <linux/kallsyms.h> +#include <linux/perf_counter.h> +#include <linux/sched.h> #include <asm/uaccess.h> #include <asm/unistd.h> @@ -378,6 +380,8 @@ static void timer_stats_account_timer(struct timer_list *timer) { unsigned int flag = 0; + if (likely(!timer->start_site)) + return; if (unlikely(tbase_get_deferrable(timer->base))) flag |= TIMER_STATS_FLAG_DEFERRABLE; @@ -491,14 +495,18 @@ static inline void debug_timer_free(struct timer_list *timer) debug_object_free(timer, &timer_debug_descr); } -static void __init_timer(struct timer_list *timer); +static void __init_timer(struct timer_list *timer, + const char *name, + struct lock_class_key *key); -void init_timer_on_stack(struct timer_list *timer) +void init_timer_on_stack_key(struct timer_list *timer, + const char *name, + struct lock_class_key *key) { debug_object_init_on_stack(timer, &timer_debug_descr); - __init_timer(timer); + __init_timer(timer, name, key); } -EXPORT_SYMBOL_GPL(init_timer_on_stack); +EXPORT_SYMBOL_GPL(init_timer_on_stack_key); void destroy_timer_on_stack(struct timer_list *timer) { @@ -512,7 +520,9 @@ static inline void debug_timer_activate(struct timer_list *timer) { } static inline void debug_timer_deactivate(struct timer_list *timer) { } #endif -static void __init_timer(struct timer_list *timer) +static void __init_timer(struct timer_list *timer, + const char *name, + struct lock_class_key *key) { timer->entry.next = NULL; timer->base = __raw_get_cpu_var(tvec_bases); @@ -521,28 +531,36 @@ static void __init_timer(struct timer_list *timer) timer->start_pid = -1; memset(timer->start_comm, 0, TASK_COMM_LEN); #endif + lockdep_init_map(&timer->lockdep_map, name, key, 0); } /** - * init_timer - initialize a timer. + * init_timer_key - initialize a timer * @timer: the timer to be initialized + * @name: name of the timer + * @key: lockdep class key of the fake lock used for tracking timer + * sync lock dependencies * - * init_timer() must be done to a timer prior calling *any* of the + * init_timer_key() must be done to a timer prior calling *any* of the * other timer functions. */ -void init_timer(struct timer_list *timer) +void init_timer_key(struct timer_list *timer, + const char *name, + struct lock_class_key *key) { debug_timer_init(timer); - __init_timer(timer); + __init_timer(timer, name, key); } -EXPORT_SYMBOL(init_timer); +EXPORT_SYMBOL(init_timer_key); -void init_timer_deferrable(struct timer_list *timer) +void init_timer_deferrable_key(struct timer_list *timer, + const char *name, + struct lock_class_key *key) { - init_timer(timer); + init_timer_key(timer, name, key); timer_set_deferrable(timer); } -EXPORT_SYMBOL(init_timer_deferrable); +EXPORT_SYMBOL(init_timer_deferrable_key); static inline void detach_timer(struct timer_list *timer, int clear_pending) @@ -589,11 +607,13 @@ static struct tvec_base *lock_timer_base(struct timer_list *timer, } } -int __mod_timer(struct timer_list *timer, unsigned long expires) +static inline int +__mod_timer(struct timer_list *timer, unsigned long expires, + bool pending_only, int pinned) { struct tvec_base *base, *new_base; unsigned long flags; - int ret = 0; + int ret = 0 , cpu; timer_stats_timer_set_start_info(timer); BUG_ON(!timer->function); @@ -603,12 +623,27 @@ int __mod_timer(struct timer_list *timer, unsigned long expires) if (timer_pending(timer)) { detach_timer(timer, 0); ret = 1; + } else { + if (pending_only) + goto out_unlock; } debug_timer_activate(timer); new_base = __get_cpu_var(tvec_bases); + cpu = smp_processor_id(); + +#if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP) + if (!pinned && get_sysctl_timer_migration() && idle_cpu(cpu)) { + int preferred_cpu = get_nohz_load_balancer(); + + if (preferred_cpu >= 0) + cpu = preferred_cpu; + } +#endif + new_base = per_cpu(tvec_bases, cpu); + if (base != new_base) { /* * We are trying to schedule the timer on the local CPU. @@ -629,42 +664,28 @@ int __mod_timer(struct timer_list *timer, unsigned long expires) timer->expires = expires; internal_add_timer(base, timer); + +out_unlock: spin_unlock_irqrestore(&base->lock, flags); return ret; } -EXPORT_SYMBOL(__mod_timer); - /** - * add_timer_on - start a timer on a particular CPU - * @timer: the timer to be added - * @cpu: the CPU to start it on + * mod_timer_pending - modify a pending timer's timeout + * @timer: the pending timer to be modified + * @expires: new timeout in jiffies * - * This is not very scalable on SMP. Double adds are not possible. + * mod_timer_pending() is the same for pending timers as mod_timer(), + * but will not re-activate and modify already deleted timers. + * + * It is useful for unserialized use of timers. */ -void add_timer_on(struct timer_list *timer, int cpu) +int mod_timer_pending(struct timer_list *timer, unsigned long expires) { - struct tvec_base *base = per_cpu(tvec_bases, cpu); - unsigned long flags; - - timer_stats_timer_set_start_info(timer); - BUG_ON(timer_pending(timer) || !timer->function); - spin_lock_irqsave(&base->lock, flags); - timer_set_base(timer, base); - debug_timer_activate(timer); - internal_add_timer(base, timer); - /* - * Check whether the other CPU is idle and needs to be - * triggered to reevaluate the timer wheel when nohz is - * active. We are protected against the other CPU fiddling - * with the timer by holding the timer base lock. This also - * makes sure that a CPU on the way to idle can not evaluate - * the timer wheel. - */ - wake_up_idle_cpu(cpu); - spin_unlock_irqrestore(&base->lock, flags); + return __mod_timer(timer, expires, true, TIMER_NOT_PINNED); } +EXPORT_SYMBOL(mod_timer_pending); /** * mod_timer - modify a timer's timeout @@ -688,21 +709,91 @@ void add_timer_on(struct timer_list *timer, int cpu) */ int mod_timer(struct timer_list *timer, unsigned long expires) { - BUG_ON(!timer->function); - - timer_stats_timer_set_start_info(timer); /* * This is a common optimization triggered by the * networking code - if the timer is re-modified * to be the same thing then just return: */ + if (timer_pending(timer) && timer->expires == expires) + return 1; + + return __mod_timer(timer, expires, false, TIMER_NOT_PINNED); +} +EXPORT_SYMBOL(mod_timer); + +/** + * mod_timer_pinned - modify a timer's timeout + * @timer: the timer to be modified + * @expires: new timeout in jiffies + * + * mod_timer_pinned() is a way to update the expire field of an + * active timer (if the timer is inactive it will be activated) + * and not allow the timer to be migrated to a different CPU. + * + * mod_timer_pinned(timer, expires) is equivalent to: + * + * del_timer(timer); timer->expires = expires; add_timer(timer); + */ +int mod_timer_pinned(struct timer_list *timer, unsigned long expires) +{ if (timer->expires == expires && timer_pending(timer)) return 1; - return __mod_timer(timer, expires); + return __mod_timer(timer, expires, false, TIMER_PINNED); } +EXPORT_SYMBOL(mod_timer_pinned); -EXPORT_SYMBOL(mod_timer); +/** + * add_timer - start a timer + * @timer: the timer to be added + * + * The kernel will do a ->function(->data) callback from the + * timer interrupt at the ->expires point in the future. The + * current time is 'jiffies'. + * + * The timer's ->expires, ->function (and if the handler uses it, ->data) + * fields must be set prior calling this function. + * + * Timers with an ->expires field in the past will be executed in the next + * timer tick. + */ +void add_timer(struct timer_list *timer) +{ + BUG_ON(timer_pending(timer)); + mod_timer(timer, timer->expires); +} +EXPORT_SYMBOL(add_timer); + +/** + * add_timer_on - start a timer on a particular CPU + * @timer: the timer to be added + * @cpu: the CPU to start it on + * + * This is not very scalable on SMP. Double adds are not possible. + */ +void add_timer_on(struct timer_list *timer, int cpu) +{ + struct tvec_base *base = per_cpu(tvec_bases, cpu); + unsigned long flags; + + timer_stats_timer_set_start_info(timer); + BUG_ON(timer_pending(timer) || !timer->function); + spin_lock_irqsave(&base->lock, flags); + timer_set_base(timer, base); + debug_timer_activate(timer); + internal_add_timer(base, timer); + /* + * Check whether the other CPU is idle and needs to be + * triggered to reevaluate the timer wheel when nohz is + * active. We are protected against the other CPU fiddling + * with the timer by holding the timer base lock. This also + * makes sure that a CPU on the way to idle can not evaluate + * the timer wheel. + */ + wake_up_idle_cpu(cpu); + spin_unlock_irqrestore(&base->lock, flags); +} +EXPORT_SYMBOL_GPL(add_timer_on); /** * del_timer - deactive a timer. @@ -733,7 +824,6 @@ int del_timer(struct timer_list *timer) return ret; } - EXPORT_SYMBOL(del_timer); #ifdef CONFIG_SMP @@ -767,7 +857,6 @@ out: return ret; } - EXPORT_SYMBOL(try_to_del_timer_sync); /** @@ -789,6 +878,15 @@ EXPORT_SYMBOL(try_to_del_timer_sync); */ int del_timer_sync(struct timer_list *timer) { +#ifdef CONFIG_LOCKDEP + unsigned long flags; + + local_irq_save(flags); + lock_map_acquire(&timer->lockdep_map); + lock_map_release(&timer->lockdep_map); + local_irq_restore(flags); +#endif + for (;;) { int ret = try_to_del_timer_sync(timer); if (ret >= 0) @@ -796,7 +894,6 @@ int del_timer_sync(struct timer_list *timer) cpu_relax(); } } - EXPORT_SYMBOL(del_timer_sync); #endif @@ -861,10 +958,36 @@ static inline void __run_timers(struct tvec_base *base) set_running_timer(base, timer); detach_timer(timer, 1); + spin_unlock_irq(&base->lock); { int preempt_count = preempt_count(); + +#ifdef CONFIG_LOCKDEP + /* + * It is permissible to free the timer from + * inside the function that is called from + * it, this we need to take into account for + * lockdep too. To avoid bogus "held lock + * freed" warnings as well as problems when + * looking into timer->lockdep_map, make a + * copy and use that here. + */ + struct lockdep_map lockdep_map = + timer->lockdep_map; +#endif + /* + * Couple the lock chain with the lock chain at + * del_timer_sync() by acquiring the lock_map + * around the fn() call here and in + * del_timer_sync(). + */ + lock_map_acquire(&lockdep_map); + fn(data); + + lock_map_release(&lockdep_map); + if (preempt_count != preempt_count()) { printk(KERN_ERR "huh, entered %p " "with preempt_count %08x, exited" @@ -930,6 +1053,9 @@ cascade: index = slot = timer_jiffies & TVN_MASK; do { list_for_each_entry(nte, varp->vec + slot, entry) { + if (tbase_get_deferrable(nte->base)) + continue; + found = 1; if (time_before(nte->expires, expires)) expires = nte->expires; @@ -1030,61 +1156,21 @@ void update_process_times(int user_tick) /* Note: this timer irq context must be accounted for as well. */ account_process_tick(p, user_tick); run_local_timers(); - if (rcu_pending(cpu)) - rcu_check_callbacks(cpu, user_tick); + rcu_check_callbacks(cpu, user_tick); printk_tick(); scheduler_tick(); run_posix_cpu_timers(p); } /* - * Nr of active tasks - counted in fixed-point numbers - */ -static unsigned long count_active_tasks(void) -{ - return nr_active() * FIXED_1; -} - -/* - * Hmm.. Changed this, as the GNU make sources (load.c) seems to - * imply that avenrun[] is the standard name for this kind of thing. - * Nothing else seems to be standardized: the fractional size etc - * all seem to differ on different machines. - * - * Requires xtime_lock to access. - */ -unsigned long avenrun[3]; - -EXPORT_SYMBOL(avenrun); - -/* - * calc_load - given tick count, update the avenrun load estimates. - * This is called while holding a write_lock on xtime_lock. - */ -static inline void calc_load(unsigned long ticks) -{ - unsigned long active_tasks; /* fixed-point */ - static int count = LOAD_FREQ; - - count -= ticks; - if (unlikely(count < 0)) { - active_tasks = count_active_tasks(); - do { - CALC_LOAD(avenrun[0], EXP_1, active_tasks); - CALC_LOAD(avenrun[1], EXP_5, active_tasks); - CALC_LOAD(avenrun[2], EXP_15, active_tasks); - count += LOAD_FREQ; - } while (count < 0); - } -} - -/* * This function runs timers and the timer-tq in bottom half context. */ static void run_timer_softirq(struct softirq_action *h) { struct tvec_base *base = __get_cpu_var(tvec_bases); + perf_counter_do_pending(); + hrtimer_run_pending(); if (time_after_eq(jiffies, base->timer_jiffies)) @@ -1102,16 +1188,6 @@ void run_local_timers(void) } /* - * Called by the timer interrupt. xtime_lock must already be taken - * by the timer IRQ! - */ -static inline void update_times(unsigned long ticks) -{ - update_wall_time(); - calc_load(ticks); -} - -/* * The 64-bit jiffies value is not atomic - you MUST NOT read it * without sampling the sequence number in xtime_lock. * jiffies is defined in the linker script... @@ -1120,7 +1196,8 @@ static inline void update_times(unsigned long ticks) void do_timer(unsigned long ticks) { jiffies_64 += ticks; - update_times(ticks); + update_wall_time(); + calc_global_load(); } #ifdef __ARCH_WANT_SYS_ALARM @@ -1268,7 +1345,7 @@ signed long __sched schedule_timeout(signed long timeout) expire = timeout + jiffies; setup_timer_on_stack(&timer, process_timeout, (unsigned long)current); - __mod_timer(&timer, expire); + __mod_timer(&timer, expire, false, TIMER_NOT_PINNED); schedule(); del_singleshot_timer_sync(&timer); @@ -1321,37 +1398,17 @@ int do_sysinfo(struct sysinfo *info) { unsigned long mem_total, sav_total; unsigned int mem_unit, bitcount; - unsigned long seq; + struct timespec tp; memset(info, 0, sizeof(struct sysinfo)); - do { - struct timespec tp; - seq = read_seqbegin(&xtime_lock); - - /* - * This is annoying. The below is the same thing - * posix_get_clock_monotonic() does, but it wants to - * take the lock which we want to cover the loads stuff - * too. - */ - - getnstimeofday(&tp); - tp.tv_sec += wall_to_monotonic.tv_sec; - tp.tv_nsec += wall_to_monotonic.tv_nsec; - monotonic_to_bootbased(&tp); - if (tp.tv_nsec - NSEC_PER_SEC >= 0) { - tp.tv_nsec = tp.tv_nsec - NSEC_PER_SEC; - tp.tv_sec++; - } - info->uptime = tp.tv_sec + (tp.tv_nsec ? 1 : 0); + ktime_get_ts(&tp); + monotonic_to_bootbased(&tp); + info->uptime = tp.tv_sec + (tp.tv_nsec ? 1 : 0); - info->loads[0] = avenrun[0] << (SI_LOAD_SHIFT - FSHIFT); - info->loads[1] = avenrun[1] << (SI_LOAD_SHIFT - FSHIFT); - info->loads[2] = avenrun[2] << (SI_LOAD_SHIFT - FSHIFT); + get_avenrun(info->loads, 0, SI_LOAD_SHIFT - FSHIFT); - info->procs = nr_threads; - } while (read_seqretry(&xtime_lock, seq)); + info->procs = nr_threads; si_meminfo(info); si_swapinfo(info); diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index e2a4ff6fc3a..1ea0d1234f4 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -9,12 +9,22 @@ config USER_STACKTRACE_SUPPORT config NOP_TRACER bool +config HAVE_FTRACE_NMI_ENTER + bool + config HAVE_FUNCTION_TRACER bool config HAVE_FUNCTION_GRAPH_TRACER bool +config HAVE_FUNCTION_GRAPH_FP_TEST + bool + help + An arch may pass in a unique value (frame pointer) to both the + entering and exiting of a function. On exit, the value is compared + and if it does not match, then it will panic the kernel. + config HAVE_FUNCTION_TRACE_MCOUNT_TEST bool help @@ -31,12 +41,40 @@ config HAVE_FTRACE_MCOUNT_RECORD config HAVE_HW_BRANCH_TRACER bool +config HAVE_SYSCALL_TRACEPOINTS + bool + config TRACER_MAX_TRACE bool config RING_BUFFER bool +config FTRACE_NMI_ENTER + bool + depends on HAVE_FTRACE_NMI_ENTER + default y + +config EVENT_TRACING + select CONTEXT_SWITCH_TRACER + bool + +config CONTEXT_SWITCH_TRACER + bool + +config RING_BUFFER_ALLOW_SWAP + bool + help + Allow the use of ring_buffer_swap_cpu. + Adds a very slight overhead to tracing when enabled. + +# All tracer options should select GENERIC_TRACER. For those options that are +# enabled by all tracers (context switch and event tracer) they select TRACING. +# This allows those options to appear when no other tracer is selected. But the +# options do not appear when something else selects it. We need the two options +# GENERIC_TRACER and TRACING to avoid circular dependencies to accomplish the +# hidding of the automatic options options. + config TRACING bool select DEBUG_FS @@ -44,15 +82,43 @@ config TRACING select STACKTRACE if STACKTRACE_SUPPORT select TRACEPOINTS select NOP_TRACER + select BINARY_PRINTF + select EVENT_TRACING -menu "Tracers" +config GENERIC_TRACER + bool + select TRACING + +# +# Minimum requirements an architecture has to meet for us to +# be able to offer generic tracing facilities: +# +config TRACING_SUPPORT + bool + # PPC32 has no irqflags tracing support, but it can use most of the + # tracers anyway, they were tested to build and work. Note that new + # exceptions to this list aren't welcomed, better implement the + # irqflags tracing for your architecture. + depends on TRACE_IRQFLAGS_SUPPORT || PPC32 + depends on STACKTRACE_SUPPORT + default y + +if TRACING_SUPPORT + +menuconfig FTRACE + bool "Tracers" + default y if DEBUG_KERNEL + help + Enable the kernel tracing infrastructure. + +if FTRACE config FUNCTION_TRACER bool "Kernel Function Tracer" depends on HAVE_FUNCTION_TRACER - depends on DEBUG_KERNEL select FRAME_POINTER - select TRACING + select KALLSYMS + select GENERIC_TRACER select CONTEXT_SWITCH_TRACER help Enable the kernel to trace every kernel function. This is done @@ -67,25 +133,26 @@ config FUNCTION_GRAPH_TRACER bool "Kernel Function Graph Tracer" depends on HAVE_FUNCTION_GRAPH_TRACER depends on FUNCTION_TRACER + depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE default y help Enable the kernel to trace a function at both its return and its entry. - It's first purpose is to trace the duration of functions and - draw a call graph for each thread with some informations like - the return value. - This is done by setting the current return address on the current - task structure into a stack of calls. + Its first purpose is to trace the duration of functions and + draw a call graph for each thread with some information like + the return value. This is done by setting the current return + address on the current task structure into a stack of calls. + config IRQSOFF_TRACER bool "Interrupts-off Latency Tracer" default n depends on TRACE_IRQFLAGS_SUPPORT depends on GENERIC_TIME - depends on DEBUG_KERNEL select TRACE_IRQFLAGS - select TRACING + select GENERIC_TRACER select TRACER_MAX_TRACE + select RING_BUFFER_ALLOW_SWAP help This option measures the time spent in irqs-off critical sections, with microsecond accuracy. @@ -94,7 +161,7 @@ config IRQSOFF_TRACER disabled by default and can be runtime (re-)started via: - echo 0 > /debugfs/tracing/tracing_max_latency + echo 0 > /sys/kernel/debug/tracing/tracing_max_latency (Note that kernel size and overhead increases with this option enabled. This option and the preempt-off timing option can be @@ -105,9 +172,9 @@ config PREEMPT_TRACER default n depends on GENERIC_TIME depends on PREEMPT - depends on DEBUG_KERNEL - select TRACING + select GENERIC_TRACER select TRACER_MAX_TRACE + select RING_BUFFER_ALLOW_SWAP help This option measures the time spent in preemption off critical sections, with microsecond accuracy. @@ -116,7 +183,7 @@ config PREEMPT_TRACER disabled by default and can be runtime (re-)started via: - echo 0 > /debugfs/tracing/tracing_max_latency + echo 0 > /sys/kernel/debug/tracing/tracing_max_latency (Note that kernel size and overhead increases with this option enabled. This option and the irqs-off timing option can be @@ -125,79 +192,111 @@ config PREEMPT_TRACER config SYSPROF_TRACER bool "Sysprof Tracer" depends on X86 - select TRACING + select GENERIC_TRACER + select CONTEXT_SWITCH_TRACER help This tracer provides the trace needed by the 'Sysprof' userspace tool. config SCHED_TRACER bool "Scheduling Latency Tracer" - depends on DEBUG_KERNEL - select TRACING + select GENERIC_TRACER select CONTEXT_SWITCH_TRACER select TRACER_MAX_TRACE help This tracer tracks the latency of the highest priority task to be scheduled in, starting from the point it has woken up. -config CONTEXT_SWITCH_TRACER - bool "Trace process context switches" - depends on DEBUG_KERNEL +config ENABLE_DEFAULT_TRACERS + bool "Trace process context switches and events" + depends on !GENERIC_TRACER select TRACING - select MARKERS help - This tracer gets called from the context switch and records - all switching of tasks. + This tracer hooks to various trace points in the kernel + allowing the user to pick and choose which trace point they + want to trace. It also includes the sched_switch tracer plugin. + +config FTRACE_SYSCALLS + bool "Trace syscalls" + depends on HAVE_SYSCALL_TRACEPOINTS + select GENERIC_TRACER + select KALLSYMS + help + Basic tracer to catch the syscall entry and exit events. config BOOT_TRACER bool "Trace boot initcalls" - depends on DEBUG_KERNEL - select TRACING + select GENERIC_TRACER select CONTEXT_SWITCH_TRACER help This tracer helps developers to optimize boot times: it records the timings of the initcalls and traces key events and the identity of tasks that can cause boot delays, such as context-switches. - Its aim is to be parsed by the /scripts/bootgraph.pl tool to + Its aim is to be parsed by the scripts/bootgraph.pl tool to produce pretty graphics about boot inefficiencies, giving a visual representation of the delays during initcalls - but the raw /debug/tracing/trace text output is readable too. - ( Note that tracing self tests can't be enabled if this tracer is - selected, because the self-tests are an initcall as well and that - would invalidate the boot trace. ) + You must pass in initcall_debug and ftrace=initcall to the kernel + command line to enable this on bootup. config TRACE_BRANCH_PROFILING + bool + select GENERIC_TRACER + +choice + prompt "Branch Profiling" + default BRANCH_PROFILE_NONE + help + The branch profiling is a software profiler. It will add hooks + into the C conditionals to test which path a branch takes. + + The likely/unlikely profiler only looks at the conditions that + are annotated with a likely or unlikely macro. + + The "all branch" profiler will profile every if statement in the + kernel. This profiler will also enable the likely/unlikely + profiler as well. + + Either of the above profilers add a bit of overhead to the system. + If unsure choose "No branch profiling". + +config BRANCH_PROFILE_NONE + bool "No branch profiling" + help + No branch profiling. Branch profiling adds a bit of overhead. + Only enable it if you want to analyse the branching behavior. + Otherwise keep it disabled. + +config PROFILE_ANNOTATED_BRANCHES bool "Trace likely/unlikely profiler" - depends on DEBUG_KERNEL - select TRACING + select TRACE_BRANCH_PROFILING help This tracer profiles all the the likely and unlikely macros in the kernel. It will display the results in: - /debugfs/tracing/profile_annotated_branch + /sys/kernel/debug/tracing/profile_annotated_branch Note: this will add a significant overhead, only turn this on if you need to profile the system's use of these macros. - Say N if unsure. - config PROFILE_ALL_BRANCHES bool "Profile all if conditionals" - depends on TRACE_BRANCH_PROFILING + select TRACE_BRANCH_PROFILING help This tracer profiles all branch conditions. Every if () taken in the kernel is recorded whether it hit or miss. The results will be displayed in: - /debugfs/tracing/profile_branch + /sys/kernel/debug/tracing/profile_branch + + This option also enables the likely/unlikely profiler. This configuration, when enabled, will impose a great overhead on the system. This should only be enabled when the system is to be analyzed - - Say N if unsure. +endchoice config TRACING_BRANCHES bool @@ -223,9 +322,8 @@ config BRANCH_TRACER config POWER_TRACER bool "Trace power consumption behavior" - depends on DEBUG_KERNEL depends on X86 - select TRACING + select GENERIC_TRACER help This tracer helps developers to analyze and optimize the kernels power management decisions, specifically the C-state and P-state @@ -235,12 +333,12 @@ config POWER_TRACER config STACK_TRACER bool "Trace max stack" depends on HAVE_FUNCTION_TRACER - depends on DEBUG_KERNEL select FUNCTION_TRACER select STACKTRACE + select KALLSYMS help This special tracer records the maximum stack footprint of the - kernel and displays it in debugfs/tracing/stack_trace. + kernel and displays it in /sys/kernel/debug/tracing/stack_trace. This tracer works by hooking into every function call that the kernel executes, and keeping a maximum stack depth value and @@ -259,16 +357,71 @@ config STACK_TRACER config HW_BRANCH_TRACER depends on HAVE_HW_BRANCH_TRACER bool "Trace hw branches" - select TRACING + select GENERIC_TRACER help This tracer records all branches on the system in a circular buffer giving access to the last N branches for each cpu. +config KMEMTRACE + bool "Trace SLAB allocations" + select GENERIC_TRACER + help + kmemtrace provides tracing for slab allocator functions, such as + kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected + data is then fed to the userspace application in order to analyse + allocation hotspots, internal fragmentation and so on, making it + possible to see how well an allocator performs, as well as debug + and profile kernel code. + + This requires an userspace application to use. See + Documentation/trace/kmemtrace.txt for more information. + + Saying Y will make the kernel somewhat larger and slower. However, + if you disable kmemtrace at run-time or boot-time, the performance + impact is minimal (depending on the arch the kernel is built for). + + If unsure, say N. + +config WORKQUEUE_TRACER + bool "Trace workqueues" + select GENERIC_TRACER + help + The workqueue tracer provides some statistical informations + about each cpu workqueue thread such as the number of the + works inserted and executed since their creation. It can help + to evaluate the amount of work each of them have to perform. + For example it can help a developer to decide whether he should + choose a per cpu workqueue instead of a singlethreaded one. + +config BLK_DEV_IO_TRACE + bool "Support for tracing block io actions" + depends on SYSFS + depends on BLOCK + select RELAY + select DEBUG_FS + select TRACEPOINTS + select GENERIC_TRACER + select STACKTRACE + help + Say Y here if you want to be able to trace the block layer actions + on a given queue. Tracing allows you to see any traffic happening + on a block device queue. For more information (and the userspace + support tools needed), fetch the blktrace tools from: + + git://git.kernel.dk/blktrace.git + + Tracing also is possible using the ftrace interface, e.g.: + + echo 1 > /sys/block/sda/sda1/trace/enable + echo blk > /sys/kernel/debug/tracing/current_tracer + cat /sys/kernel/debug/tracing/trace_pipe + + If unsure, say N. + config DYNAMIC_FTRACE bool "enable/disable ftrace tracepoints dynamically" depends on FUNCTION_TRACER depends on HAVE_DYNAMIC_FTRACE - depends on DEBUG_KERNEL default y help This option will modify all the calls to ftrace dynamically @@ -284,6 +437,20 @@ config DYNAMIC_FTRACE were made. If so, it runs stop_machine (stops all CPUS) and modifies the code to jump over the call to ftrace. +config FUNCTION_PROFILER + bool "Kernel function profiler" + depends on FUNCTION_TRACER + default n + help + This option enables the kernel function profiler. A file is created + in debugfs called function_profile_enabled which defaults to zero. + When a 1 is echoed into this file profiling begins, and when a + zero is entered, profiling stops. A file in the trace_stats + directory called functions, that show the list of functions that + have been hit and their counters. + + If in doubt, say N + config FTRACE_MCOUNT_RECORD def_bool y depends on DYNAMIC_FTRACE @@ -294,7 +461,7 @@ config FTRACE_SELFTEST config FTRACE_STARTUP_TEST bool "Perform a startup test on ftrace" - depends on TRACING && DEBUG_KERNEL && !BOOT_TRACER + depends on GENERIC_TRACER select FTRACE_SELFTEST help This option performs a series of startup tests on ftrace. On bootup @@ -302,4 +469,46 @@ config FTRACE_STARTUP_TEST functioning properly. It will do tests on all the configured tracers of ftrace. -endmenu +config MMIOTRACE + bool "Memory mapped IO tracing" + depends on HAVE_MMIOTRACE_SUPPORT && PCI + select GENERIC_TRACER + help + Mmiotrace traces Memory Mapped I/O access and is meant for + debugging and reverse engineering. It is called from the ioremap + implementation and works via page faults. Tracing is disabled by + default and can be enabled at run-time. + + See Documentation/trace/mmiotrace.txt. + If you are not helping to develop drivers, say N. + +config MMIOTRACE_TEST + tristate "Test module for mmiotrace" + depends on MMIOTRACE && m + help + This is a dumb module for testing mmiotrace. It is very dangerous + as it will write garbage to IO memory starting at a given address. + However, it should be safe to use on e.g. unused portion of VRAM. + + Say N, unless you absolutely know what you are doing. + +config RING_BUFFER_BENCHMARK + tristate "Ring buffer benchmark stress tester" + depends on RING_BUFFER + help + This option creates a test to stress the ring buffer and bench mark it. + It creates its own ring buffer such that it will not interfer with + any other users of the ring buffer (such as ftrace). It then creates + a producer and consumer that will run for 10 seconds and sleep for + 10 seconds. Each interval it will print out the number of events + it recorded and give a rough estimate of how long each iteration took. + + It does not disable interrupts or raise its priority, so it may be + affected by processes that are running. + + If unsure, say N + +endif # FTRACE + +endif # TRACING_SUPPORT + diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 349d5a93653..844164dca90 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -15,10 +15,20 @@ ifdef CONFIG_TRACING_BRANCHES KBUILD_CFLAGS += -DDISABLE_BRANCH_PROFILING endif +# +# Make the trace clocks available generally: it's infrastructure +# relied on by ptrace for example: +# +obj-y += trace_clock.o + obj-$(CONFIG_FUNCTION_TRACER) += libftrace.o obj-$(CONFIG_RING_BUFFER) += ring_buffer.o +obj-$(CONFIG_RING_BUFFER_BENCHMARK) += ring_buffer_benchmark.o obj-$(CONFIG_TRACING) += trace.o +obj-$(CONFIG_TRACING) += trace_output.o +obj-$(CONFIG_TRACING) += trace_stat.o +obj-$(CONFIG_TRACING) += trace_printk.o obj-$(CONFIG_CONTEXT_SWITCH_TRACER) += trace_sched_switch.o obj-$(CONFIG_SYSPROF_TRACER) += trace_sysprof.o obj-$(CONFIG_FUNCTION_TRACER) += trace_functions.o @@ -33,5 +43,16 @@ obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += trace_functions_graph.o obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_branch.o obj-$(CONFIG_HW_BRANCH_TRACER) += trace_hw_branches.o obj-$(CONFIG_POWER_TRACER) += trace_power.o +obj-$(CONFIG_KMEMTRACE) += kmemtrace.o +obj-$(CONFIG_WORKQUEUE_TRACER) += trace_workqueue.o +obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o +ifeq ($(CONFIG_BLOCK),y) +obj-$(CONFIG_EVENT_TRACING) += blktrace.o +endif +obj-$(CONFIG_EVENT_TRACING) += trace_events.o +obj-$(CONFIG_EVENT_TRACING) += trace_export.o +obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o +obj-$(CONFIG_EVENT_PROFILE) += trace_event_profile.o +obj-$(CONFIG_EVENT_TRACING) += trace_events_filter.o libftrace-y := ftrace.o diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c new file mode 100644 index 00000000000..3eb159c277c --- /dev/null +++ b/kernel/trace/blktrace.c @@ -0,0 +1,1728 @@ +/* + * Copyright (C) 2006 Jens Axboe <axboe@kernel.dk> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + * + */ +#include <linux/kernel.h> +#include <linux/blkdev.h> +#include <linux/blktrace_api.h> +#include <linux/percpu.h> +#include <linux/init.h> +#include <linux/mutex.h> +#include <linux/debugfs.h> +#include <linux/smp_lock.h> +#include <linux/time.h> +#include <linux/uaccess.h> + +#include <trace/events/block.h> + +#include "trace_output.h" + +#ifdef CONFIG_BLK_DEV_IO_TRACE + +static unsigned int blktrace_seq __read_mostly = 1; + +static struct trace_array *blk_tr; +static bool blk_tracer_enabled __read_mostly; + +/* Select an alternative, minimalistic output than the original one */ +#define TRACE_BLK_OPT_CLASSIC 0x1 + +static struct tracer_opt blk_tracer_opts[] = { + /* Default disable the minimalistic output */ + { TRACER_OPT(blk_classic, TRACE_BLK_OPT_CLASSIC) }, + { } +}; + +static struct tracer_flags blk_tracer_flags = { + .val = 0, + .opts = blk_tracer_opts, +}; + +/* Global reference count of probes */ +static atomic_t blk_probes_ref = ATOMIC_INIT(0); + +static void blk_register_tracepoints(void); +static void blk_unregister_tracepoints(void); + +/* + * Send out a notify message. + */ +static void trace_note(struct blk_trace *bt, pid_t pid, int action, + const void *data, size_t len) +{ + struct blk_io_trace *t; + struct ring_buffer_event *event = NULL; + struct ring_buffer *buffer = NULL; + int pc = 0; + int cpu = smp_processor_id(); + bool blk_tracer = blk_tracer_enabled; + + if (blk_tracer) { + buffer = blk_tr->buffer; + pc = preempt_count(); + event = trace_buffer_lock_reserve(buffer, TRACE_BLK, + sizeof(*t) + len, + 0, pc); + if (!event) + return; + t = ring_buffer_event_data(event); + goto record_it; + } + + if (!bt->rchan) + return; + + t = relay_reserve(bt->rchan, sizeof(*t) + len); + if (t) { + t->magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION; + t->time = ktime_to_ns(ktime_get()); +record_it: + t->device = bt->dev; + t->action = action; + t->pid = pid; + t->cpu = cpu; + t->pdu_len = len; + memcpy((void *) t + sizeof(*t), data, len); + + if (blk_tracer) + trace_buffer_unlock_commit(buffer, event, 0, pc); + } +} + +/* + * Send out a notify for this process, if we haven't done so since a trace + * started + */ +static void trace_note_tsk(struct blk_trace *bt, struct task_struct *tsk) +{ + tsk->btrace_seq = blktrace_seq; + trace_note(bt, tsk->pid, BLK_TN_PROCESS, tsk->comm, sizeof(tsk->comm)); +} + +static void trace_note_time(struct blk_trace *bt) +{ + struct timespec now; + unsigned long flags; + u32 words[2]; + + getnstimeofday(&now); + words[0] = now.tv_sec; + words[1] = now.tv_nsec; + + local_irq_save(flags); + trace_note(bt, 0, BLK_TN_TIMESTAMP, words, sizeof(words)); + local_irq_restore(flags); +} + +void __trace_note_message(struct blk_trace *bt, const char *fmt, ...) +{ + int n; + va_list args; + unsigned long flags; + char *buf; + + if (unlikely(bt->trace_state != Blktrace_running && + !blk_tracer_enabled)) + return; + + local_irq_save(flags); + buf = per_cpu_ptr(bt->msg_data, smp_processor_id()); + va_start(args, fmt); + n = vscnprintf(buf, BLK_TN_MAX_MSG, fmt, args); + va_end(args); + + trace_note(bt, 0, BLK_TN_MESSAGE, buf, n); + local_irq_restore(flags); +} +EXPORT_SYMBOL_GPL(__trace_note_message); + +static int act_log_check(struct blk_trace *bt, u32 what, sector_t sector, + pid_t pid) +{ + if (((bt->act_mask << BLK_TC_SHIFT) & what) == 0) + return 1; + if (sector && (sector < bt->start_lba || sector > bt->end_lba)) + return 1; + if (bt->pid && pid != bt->pid) + return 1; + + return 0; +} + +/* + * Data direction bit lookup + */ +static const u32 ddir_act[2] = { BLK_TC_ACT(BLK_TC_READ), + BLK_TC_ACT(BLK_TC_WRITE) }; + +/* The ilog2() calls fall out because they're constant */ +#define MASK_TC_BIT(rw, __name) ((rw & (1 << BIO_RW_ ## __name)) << \ + (ilog2(BLK_TC_ ## __name) + BLK_TC_SHIFT - BIO_RW_ ## __name)) + +/* + * The worker for the various blk_add_trace*() types. Fills out a + * blk_io_trace structure and places it in a per-cpu subbuffer. + */ +static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, + int rw, u32 what, int error, int pdu_len, void *pdu_data) +{ + struct task_struct *tsk = current; + struct ring_buffer_event *event = NULL; + struct ring_buffer *buffer = NULL; + struct blk_io_trace *t; + unsigned long flags = 0; + unsigned long *sequence; + pid_t pid; + int cpu, pc = 0; + bool blk_tracer = blk_tracer_enabled; + + if (unlikely(bt->trace_state != Blktrace_running && !blk_tracer)) + return; + + what |= ddir_act[rw & WRITE]; + what |= MASK_TC_BIT(rw, BARRIER); + what |= MASK_TC_BIT(rw, SYNCIO); + what |= MASK_TC_BIT(rw, AHEAD); + what |= MASK_TC_BIT(rw, META); + what |= MASK_TC_BIT(rw, DISCARD); + + pid = tsk->pid; + if (act_log_check(bt, what, sector, pid)) + return; + cpu = raw_smp_processor_id(); + + if (blk_tracer) { + tracing_record_cmdline(current); + + buffer = blk_tr->buffer; + pc = preempt_count(); + event = trace_buffer_lock_reserve(buffer, TRACE_BLK, + sizeof(*t) + pdu_len, + 0, pc); + if (!event) + return; + t = ring_buffer_event_data(event); + goto record_it; + } + + /* + * A word about the locking here - we disable interrupts to reserve + * some space in the relay per-cpu buffer, to prevent an irq + * from coming in and stepping on our toes. + */ + local_irq_save(flags); + + if (unlikely(tsk->btrace_seq != blktrace_seq)) + trace_note_tsk(bt, tsk); + + t = relay_reserve(bt->rchan, sizeof(*t) + pdu_len); + if (t) { + sequence = per_cpu_ptr(bt->sequence, cpu); + + t->magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION; + t->sequence = ++(*sequence); + t->time = ktime_to_ns(ktime_get()); +record_it: + /* + * These two are not needed in ftrace as they are in the + * generic trace_entry, filled by tracing_generic_entry_update, + * but for the trace_event->bin() synthesizer benefit we do it + * here too. + */ + t->cpu = cpu; + t->pid = pid; + + t->sector = sector; + t->bytes = bytes; + t->action = what; + t->device = bt->dev; + t->error = error; + t->pdu_len = pdu_len; + + if (pdu_len) + memcpy((void *) t + sizeof(*t), pdu_data, pdu_len); + + if (blk_tracer) { + trace_buffer_unlock_commit(buffer, event, 0, pc); + return; + } + } + + local_irq_restore(flags); +} + +static struct dentry *blk_tree_root; +static DEFINE_MUTEX(blk_tree_mutex); + +static void blk_trace_free(struct blk_trace *bt) +{ + debugfs_remove(bt->msg_file); + debugfs_remove(bt->dropped_file); + relay_close(bt->rchan); + debugfs_remove(bt->dir); + free_percpu(bt->sequence); + free_percpu(bt->msg_data); + kfree(bt); +} + +static void blk_trace_cleanup(struct blk_trace *bt) +{ + blk_trace_free(bt); + if (atomic_dec_and_test(&blk_probes_ref)) + blk_unregister_tracepoints(); +} + +int blk_trace_remove(struct request_queue *q) +{ + struct blk_trace *bt; + + bt = xchg(&q->blk_trace, NULL); + if (!bt) + return -EINVAL; + + if (bt->trace_state != Blktrace_running) + blk_trace_cleanup(bt); + + return 0; +} +EXPORT_SYMBOL_GPL(blk_trace_remove); + +static int blk_dropped_open(struct inode *inode, struct file *filp) +{ + filp->private_data = inode->i_private; + + return 0; +} + +static ssize_t blk_dropped_read(struct file *filp, char __user *buffer, + size_t count, loff_t *ppos) +{ + struct blk_trace *bt = filp->private_data; + char buf[16]; + + snprintf(buf, sizeof(buf), "%u\n", atomic_read(&bt->dropped)); + + return simple_read_from_buffer(buffer, count, ppos, buf, strlen(buf)); +} + +static const struct file_operations blk_dropped_fops = { + .owner = THIS_MODULE, + .open = blk_dropped_open, + .read = blk_dropped_read, +}; + +static int blk_msg_open(struct inode *inode, struct file *filp) +{ + filp->private_data = inode->i_private; + + return 0; +} + +static ssize_t blk_msg_write(struct file *filp, const char __user *buffer, + size_t count, loff_t *ppos) +{ + char *msg; + struct blk_trace *bt; + + if (count >= BLK_TN_MAX_MSG) + return -EINVAL; + + msg = kmalloc(count + 1, GFP_KERNEL); + if (msg == NULL) + return -ENOMEM; + + if (copy_from_user(msg, buffer, count)) { + kfree(msg); + return -EFAULT; + } + + msg[count] = '\0'; + bt = filp->private_data; + __trace_note_message(bt, "%s", msg); + kfree(msg); + + return count; +} + +static const struct file_operations blk_msg_fops = { + .owner = THIS_MODULE, + .open = blk_msg_open, + .write = blk_msg_write, +}; + +/* + * Keep track of how many times we encountered a full subbuffer, to aid + * the user space app in telling how many lost events there were. + */ +static int blk_subbuf_start_callback(struct rchan_buf *buf, void *subbuf, + void *prev_subbuf, size_t prev_padding) +{ + struct blk_trace *bt; + + if (!relay_buf_full(buf)) + return 1; + + bt = buf->chan->private_data; + atomic_inc(&bt->dropped); + return 0; +} + +static int blk_remove_buf_file_callback(struct dentry *dentry) +{ + debugfs_remove(dentry); + + return 0; +} + +static struct dentry *blk_create_buf_file_callback(const char *filename, + struct dentry *parent, + int mode, + struct rchan_buf *buf, + int *is_global) +{ + return debugfs_create_file(filename, mode, parent, buf, + &relay_file_operations); +} + +static struct rchan_callbacks blk_relay_callbacks = { + .subbuf_start = blk_subbuf_start_callback, + .create_buf_file = blk_create_buf_file_callback, + .remove_buf_file = blk_remove_buf_file_callback, +}; + +static void blk_trace_setup_lba(struct blk_trace *bt, + struct block_device *bdev) +{ + struct hd_struct *part = NULL; + + if (bdev) + part = bdev->bd_part; + + if (part) { + bt->start_lba = part->start_sect; + bt->end_lba = part->start_sect + part->nr_sects; + } else { + bt->start_lba = 0; + bt->end_lba = -1ULL; + } +} + +/* + * Setup everything required to start tracing + */ +int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, + struct block_device *bdev, + struct blk_user_trace_setup *buts) +{ + struct blk_trace *old_bt, *bt = NULL; + struct dentry *dir = NULL; + int ret, i; + + if (!buts->buf_size || !buts->buf_nr) + return -EINVAL; + + strncpy(buts->name, name, BLKTRACE_BDEV_SIZE); + buts->name[BLKTRACE_BDEV_SIZE - 1] = '\0'; + + /* + * some device names have larger paths - convert the slashes + * to underscores for this to work as expected + */ + for (i = 0; i < strlen(buts->name); i++) + if (buts->name[i] == '/') + buts->name[i] = '_'; + + bt = kzalloc(sizeof(*bt), GFP_KERNEL); + if (!bt) + return -ENOMEM; + + ret = -ENOMEM; + bt->sequence = alloc_percpu(unsigned long); + if (!bt->sequence) + goto err; + + bt->msg_data = __alloc_percpu(BLK_TN_MAX_MSG, __alignof__(char)); + if (!bt->msg_data) + goto err; + + ret = -ENOENT; + + mutex_lock(&blk_tree_mutex); + if (!blk_tree_root) { + blk_tree_root = debugfs_create_dir("block", NULL); + if (!blk_tree_root) { + mutex_unlock(&blk_tree_mutex); + goto err; + } + } + mutex_unlock(&blk_tree_mutex); + + dir = debugfs_create_dir(buts->name, blk_tree_root); + + if (!dir) + goto err; + + bt->dir = dir; + bt->dev = dev; + atomic_set(&bt->dropped, 0); + + ret = -EIO; + bt->dropped_file = debugfs_create_file("dropped", 0444, dir, bt, + &blk_dropped_fops); + if (!bt->dropped_file) + goto err; + + bt->msg_file = debugfs_create_file("msg", 0222, dir, bt, &blk_msg_fops); + if (!bt->msg_file) + goto err; + + bt->rchan = relay_open("trace", dir, buts->buf_size, + buts->buf_nr, &blk_relay_callbacks, bt); + if (!bt->rchan) + goto err; + + bt->act_mask = buts->act_mask; + if (!bt->act_mask) + bt->act_mask = (u16) -1; + + blk_trace_setup_lba(bt, bdev); + + /* overwrite with user settings */ + if (buts->start_lba) + bt->start_lba = buts->start_lba; + if (buts->end_lba) + bt->end_lba = buts->end_lba; + + bt->pid = buts->pid; + bt->trace_state = Blktrace_setup; + + ret = -EBUSY; + old_bt = xchg(&q->blk_trace, bt); + if (old_bt) { + (void) xchg(&q->blk_trace, old_bt); + goto err; + } + + if (atomic_inc_return(&blk_probes_ref) == 1) + blk_register_tracepoints(); + + return 0; +err: + blk_trace_free(bt); + return ret; +} + +int blk_trace_setup(struct request_queue *q, char *name, dev_t dev, + struct block_device *bdev, + char __user *arg) +{ + struct blk_user_trace_setup buts; + int ret; + + ret = copy_from_user(&buts, arg, sizeof(buts)); + if (ret) + return -EFAULT; + + ret = do_blk_trace_setup(q, name, dev, bdev, &buts); + if (ret) + return ret; + + if (copy_to_user(arg, &buts, sizeof(buts))) + return -EFAULT; + + return 0; +} +EXPORT_SYMBOL_GPL(blk_trace_setup); + +int blk_trace_startstop(struct request_queue *q, int start) +{ + int ret; + struct blk_trace *bt = q->blk_trace; + + if (bt == NULL) + return -EINVAL; + + /* + * For starting a trace, we can transition from a setup or stopped + * trace. For stopping a trace, the state must be running + */ + ret = -EINVAL; + if (start) { + if (bt->trace_state == Blktrace_setup || + bt->trace_state == Blktrace_stopped) { + blktrace_seq++; + smp_mb(); + bt->trace_state = Blktrace_running; + + trace_note_time(bt); + ret = 0; + } + } else { + if (bt->trace_state == Blktrace_running) { + bt->trace_state = Blktrace_stopped; + relay_flush(bt->rchan); + ret = 0; + } + } + + return ret; +} +EXPORT_SYMBOL_GPL(blk_trace_startstop); + +/** + * blk_trace_ioctl: - handle the ioctls associated with tracing + * @bdev: the block device + * @cmd: the ioctl cmd + * @arg: the argument data, if any + * + **/ +int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg) +{ + struct request_queue *q; + int ret, start = 0; + char b[BDEVNAME_SIZE]; + + q = bdev_get_queue(bdev); + if (!q) + return -ENXIO; + + mutex_lock(&bdev->bd_mutex); + + switch (cmd) { + case BLKTRACESETUP: + bdevname(bdev, b); + ret = blk_trace_setup(q, b, bdev->bd_dev, bdev, arg); + break; + case BLKTRACESTART: + start = 1; + case BLKTRACESTOP: + ret = blk_trace_startstop(q, start); + break; + case BLKTRACETEARDOWN: + ret = blk_trace_remove(q); + break; + default: + ret = -ENOTTY; + break; + } + + mutex_unlock(&bdev->bd_mutex); + return ret; +} + +/** + * blk_trace_shutdown: - stop and cleanup trace structures + * @q: the request queue associated with the device + * + **/ +void blk_trace_shutdown(struct request_queue *q) +{ + if (q->blk_trace) { + blk_trace_startstop(q, 0); + blk_trace_remove(q); + } +} + +/* + * blktrace probes + */ + +/** + * blk_add_trace_rq - Add a trace for a request oriented action + * @q: queue the io is for + * @rq: the source request + * @what: the action + * + * Description: + * Records an action against a request. Will log the bio offset + size. + * + **/ +static void blk_add_trace_rq(struct request_queue *q, struct request *rq, + u32 what) +{ + struct blk_trace *bt = q->blk_trace; + int rw = rq->cmd_flags & 0x03; + + if (likely(!bt)) + return; + + if (blk_discard_rq(rq)) + rw |= (1 << BIO_RW_DISCARD); + + if (blk_pc_request(rq)) { + what |= BLK_TC_ACT(BLK_TC_PC); + __blk_add_trace(bt, 0, blk_rq_bytes(rq), rw, + what, rq->errors, rq->cmd_len, rq->cmd); + } else { + what |= BLK_TC_ACT(BLK_TC_FS); + __blk_add_trace(bt, blk_rq_pos(rq), blk_rq_bytes(rq), rw, + what, rq->errors, 0, NULL); + } +} + +static void blk_add_trace_rq_abort(struct request_queue *q, struct request *rq) +{ + blk_add_trace_rq(q, rq, BLK_TA_ABORT); +} + +static void blk_add_trace_rq_insert(struct request_queue *q, struct request *rq) +{ + blk_add_trace_rq(q, rq, BLK_TA_INSERT); +} + +static void blk_add_trace_rq_issue(struct request_queue *q, struct request *rq) +{ + blk_add_trace_rq(q, rq, BLK_TA_ISSUE); +} + +static void blk_add_trace_rq_requeue(struct request_queue *q, + struct request *rq) +{ + blk_add_trace_rq(q, rq, BLK_TA_REQUEUE); +} + +static void blk_add_trace_rq_complete(struct request_queue *q, + struct request *rq) +{ + blk_add_trace_rq(q, rq, BLK_TA_COMPLETE); +} + +/** + * blk_add_trace_bio - Add a trace for a bio oriented action + * @q: queue the io is for + * @bio: the source bio + * @what: the action + * + * Description: + * Records an action against a bio. Will log the bio offset + size. + * + **/ +static void blk_add_trace_bio(struct request_queue *q, struct bio *bio, + u32 what) +{ + struct blk_trace *bt = q->blk_trace; + + if (likely(!bt)) + return; + + __blk_add_trace(bt, bio->bi_sector, bio->bi_size, bio->bi_rw, what, + !bio_flagged(bio, BIO_UPTODATE), 0, NULL); +} + +static void blk_add_trace_bio_bounce(struct request_queue *q, struct bio *bio) +{ + blk_add_trace_bio(q, bio, BLK_TA_BOUNCE); +} + +static void blk_add_trace_bio_complete(struct request_queue *q, struct bio *bio) +{ + blk_add_trace_bio(q, bio, BLK_TA_COMPLETE); +} + +static void blk_add_trace_bio_backmerge(struct request_queue *q, + struct bio *bio) +{ + blk_add_trace_bio(q, bio, BLK_TA_BACKMERGE); +} + +static void blk_add_trace_bio_frontmerge(struct request_queue *q, + struct bio *bio) +{ + blk_add_trace_bio(q, bio, BLK_TA_FRONTMERGE); +} + +static void blk_add_trace_bio_queue(struct request_queue *q, struct bio *bio) +{ + blk_add_trace_bio(q, bio, BLK_TA_QUEUE); +} + +static void blk_add_trace_getrq(struct request_queue *q, + struct bio *bio, int rw) +{ + if (bio) + blk_add_trace_bio(q, bio, BLK_TA_GETRQ); + else { + struct blk_trace *bt = q->blk_trace; + + if (bt) + __blk_add_trace(bt, 0, 0, rw, BLK_TA_GETRQ, 0, 0, NULL); + } +} + + +static void blk_add_trace_sleeprq(struct request_queue *q, + struct bio *bio, int rw) +{ + if (bio) + blk_add_trace_bio(q, bio, BLK_TA_SLEEPRQ); + else { + struct blk_trace *bt = q->blk_trace; + + if (bt) + __blk_add_trace(bt, 0, 0, rw, BLK_TA_SLEEPRQ, + 0, 0, NULL); + } +} + +static void blk_add_trace_plug(struct request_queue *q) +{ + struct blk_trace *bt = q->blk_trace; + + if (bt) + __blk_add_trace(bt, 0, 0, 0, BLK_TA_PLUG, 0, 0, NULL); +} + +static void blk_add_trace_unplug_io(struct request_queue *q) +{ + struct blk_trace *bt = q->blk_trace; + + if (bt) { + unsigned int pdu = q->rq.count[READ] + q->rq.count[WRITE]; + __be64 rpdu = cpu_to_be64(pdu); + + __blk_add_trace(bt, 0, 0, 0, BLK_TA_UNPLUG_IO, 0, + sizeof(rpdu), &rpdu); + } +} + +static void blk_add_trace_unplug_timer(struct request_queue *q) +{ + struct blk_trace *bt = q->blk_trace; + + if (bt) { + unsigned int pdu = q->rq.count[READ] + q->rq.count[WRITE]; + __be64 rpdu = cpu_to_be64(pdu); + + __blk_add_trace(bt, 0, 0, 0, BLK_TA_UNPLUG_TIMER, 0, + sizeof(rpdu), &rpdu); + } +} + +static void blk_add_trace_split(struct request_queue *q, struct bio *bio, + unsigned int pdu) +{ + struct blk_trace *bt = q->blk_trace; + + if (bt) { + __be64 rpdu = cpu_to_be64(pdu); + + __blk_add_trace(bt, bio->bi_sector, bio->bi_size, bio->bi_rw, + BLK_TA_SPLIT, !bio_flagged(bio, BIO_UPTODATE), + sizeof(rpdu), &rpdu); + } +} + +/** + * blk_add_trace_remap - Add a trace for a remap operation + * @q: queue the io is for + * @bio: the source bio + * @dev: target device + * @from: source sector + * + * Description: + * Device mapper or raid target sometimes need to split a bio because + * it spans a stripe (or similar). Add a trace for that action. + * + **/ +static void blk_add_trace_remap(struct request_queue *q, struct bio *bio, + dev_t dev, sector_t from) +{ + struct blk_trace *bt = q->blk_trace; + struct blk_io_trace_remap r; + + if (likely(!bt)) + return; + + r.device_from = cpu_to_be32(dev); + r.device_to = cpu_to_be32(bio->bi_bdev->bd_dev); + r.sector_from = cpu_to_be64(from); + + __blk_add_trace(bt, bio->bi_sector, bio->bi_size, bio->bi_rw, + BLK_TA_REMAP, !bio_flagged(bio, BIO_UPTODATE), + sizeof(r), &r); +} + +/** + * blk_add_driver_data - Add binary message with driver-specific data + * @q: queue the io is for + * @rq: io request + * @data: driver-specific data + * @len: length of driver-specific data + * + * Description: + * Some drivers might want to write driver-specific data per request. + * + **/ +void blk_add_driver_data(struct request_queue *q, + struct request *rq, + void *data, size_t len) +{ + struct blk_trace *bt = q->blk_trace; + + if (likely(!bt)) + return; + + if (blk_pc_request(rq)) + __blk_add_trace(bt, 0, blk_rq_bytes(rq), 0, + BLK_TA_DRV_DATA, rq->errors, len, data); + else + __blk_add_trace(bt, blk_rq_pos(rq), blk_rq_bytes(rq), 0, + BLK_TA_DRV_DATA, rq->errors, len, data); +} +EXPORT_SYMBOL_GPL(blk_add_driver_data); + +static void blk_register_tracepoints(void) +{ + int ret; + + ret = register_trace_block_rq_abort(blk_add_trace_rq_abort); + WARN_ON(ret); + ret = register_trace_block_rq_insert(blk_add_trace_rq_insert); + WARN_ON(ret); + ret = register_trace_block_rq_issue(blk_add_trace_rq_issue); + WARN_ON(ret); + ret = register_trace_block_rq_requeue(blk_add_trace_rq_requeue); + WARN_ON(ret); + ret = register_trace_block_rq_complete(blk_add_trace_rq_complete); + WARN_ON(ret); + ret = register_trace_block_bio_bounce(blk_add_trace_bio_bounce); + WARN_ON(ret); + ret = register_trace_block_bio_complete(blk_add_trace_bio_complete); + WARN_ON(ret); + ret = register_trace_block_bio_backmerge(blk_add_trace_bio_backmerge); + WARN_ON(ret); + ret = register_trace_block_bio_frontmerge(blk_add_trace_bio_frontmerge); + WARN_ON(ret); + ret = register_trace_block_bio_queue(blk_add_trace_bio_queue); + WARN_ON(ret); + ret = register_trace_block_getrq(blk_add_trace_getrq); + WARN_ON(ret); + ret = register_trace_block_sleeprq(blk_add_trace_sleeprq); + WARN_ON(ret); + ret = register_trace_block_plug(blk_add_trace_plug); + WARN_ON(ret); + ret = register_trace_block_unplug_timer(blk_add_trace_unplug_timer); + WARN_ON(ret); + ret = register_trace_block_unplug_io(blk_add_trace_unplug_io); + WARN_ON(ret); + ret = register_trace_block_split(blk_add_trace_split); + WARN_ON(ret); + ret = register_trace_block_remap(blk_add_trace_remap); + WARN_ON(ret); +} + +static void blk_unregister_tracepoints(void) +{ + unregister_trace_block_remap(blk_add_trace_remap); + unregister_trace_block_split(blk_add_trace_split); + unregister_trace_block_unplug_io(blk_add_trace_unplug_io); + unregister_trace_block_unplug_timer(blk_add_trace_unplug_timer); + unregister_trace_block_plug(blk_add_trace_plug); + unregister_trace_block_sleeprq(blk_add_trace_sleeprq); + unregister_trace_block_getrq(blk_add_trace_getrq); + unregister_trace_block_bio_queue(blk_add_trace_bio_queue); + unregister_trace_block_bio_frontmerge(blk_add_trace_bio_frontmerge); + unregister_trace_block_bio_backmerge(blk_add_trace_bio_backmerge); + unregister_trace_block_bio_complete(blk_add_trace_bio_complete); + unregister_trace_block_bio_bounce(blk_add_trace_bio_bounce); + unregister_trace_block_rq_complete(blk_add_trace_rq_complete); + unregister_trace_block_rq_requeue(blk_add_trace_rq_requeue); + unregister_trace_block_rq_issue(blk_add_trace_rq_issue); + unregister_trace_block_rq_insert(blk_add_trace_rq_insert); + unregister_trace_block_rq_abort(blk_add_trace_rq_abort); + + tracepoint_synchronize_unregister(); +} + +/* + * struct blk_io_tracer formatting routines + */ + +static void fill_rwbs(char *rwbs, const struct blk_io_trace *t) +{ + int i = 0; + int tc = t->action >> BLK_TC_SHIFT; + + if (t->action == BLK_TN_MESSAGE) { + rwbs[i++] = 'N'; + goto out; + } + + if (tc & BLK_TC_DISCARD) + rwbs[i++] = 'D'; + else if (tc & BLK_TC_WRITE) + rwbs[i++] = 'W'; + else if (t->bytes) + rwbs[i++] = 'R'; + else + rwbs[i++] = 'N'; + + if (tc & BLK_TC_AHEAD) + rwbs[i++] = 'A'; + if (tc & BLK_TC_BARRIER) + rwbs[i++] = 'B'; + if (tc & BLK_TC_SYNC) + rwbs[i++] = 'S'; + if (tc & BLK_TC_META) + rwbs[i++] = 'M'; +out: + rwbs[i] = '\0'; +} + +static inline +const struct blk_io_trace *te_blk_io_trace(const struct trace_entry *ent) +{ + return (const struct blk_io_trace *)ent; +} + +static inline const void *pdu_start(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent) + 1; +} + +static inline u32 t_action(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent)->action; +} + +static inline u32 t_bytes(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent)->bytes; +} + +static inline u32 t_sec(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent)->bytes >> 9; +} + +static inline unsigned long long t_sector(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent)->sector; +} + +static inline __u16 t_error(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent)->error; +} + +static __u64 get_pdu_int(const struct trace_entry *ent) +{ + const __u64 *val = pdu_start(ent); + return be64_to_cpu(*val); +} + +static void get_pdu_remap(const struct trace_entry *ent, + struct blk_io_trace_remap *r) +{ + const struct blk_io_trace_remap *__r = pdu_start(ent); + __u64 sector_from = __r->sector_from; + + r->device_from = be32_to_cpu(__r->device_from); + r->device_to = be32_to_cpu(__r->device_to); + r->sector_from = be64_to_cpu(sector_from); +} + +typedef int (blk_log_action_t) (struct trace_iterator *iter, const char *act); + +static int blk_log_action_classic(struct trace_iterator *iter, const char *act) +{ + char rwbs[6]; + unsigned long long ts = iter->ts; + unsigned long nsec_rem = do_div(ts, NSEC_PER_SEC); + unsigned secs = (unsigned long)ts; + const struct blk_io_trace *t = te_blk_io_trace(iter->ent); + + fill_rwbs(rwbs, t); + + return trace_seq_printf(&iter->seq, + "%3d,%-3d %2d %5d.%09lu %5u %2s %3s ", + MAJOR(t->device), MINOR(t->device), iter->cpu, + secs, nsec_rem, iter->ent->pid, act, rwbs); +} + +static int blk_log_action(struct trace_iterator *iter, const char *act) +{ + char rwbs[6]; + const struct blk_io_trace *t = te_blk_io_trace(iter->ent); + + fill_rwbs(rwbs, t); + return trace_seq_printf(&iter->seq, "%3d,%-3d %2s %3s ", + MAJOR(t->device), MINOR(t->device), act, rwbs); +} + +static int blk_log_dump_pdu(struct trace_seq *s, const struct trace_entry *ent) +{ + const unsigned char *pdu_buf; + int pdu_len; + int i, end, ret; + + pdu_buf = pdu_start(ent); + pdu_len = te_blk_io_trace(ent)->pdu_len; + + if (!pdu_len) + return 1; + + /* find the last zero that needs to be printed */ + for (end = pdu_len - 1; end >= 0; end--) + if (pdu_buf[end]) + break; + end++; + + if (!trace_seq_putc(s, '(')) + return 0; + + for (i = 0; i < pdu_len; i++) { + + ret = trace_seq_printf(s, "%s%02x", + i == 0 ? "" : " ", pdu_buf[i]); + if (!ret) + return ret; + + /* + * stop when the rest is just zeroes and indicate so + * with a ".." appended + */ + if (i == end && end != pdu_len - 1) + return trace_seq_puts(s, " ..) "); + } + + return trace_seq_puts(s, ") "); +} + +static int blk_log_generic(struct trace_seq *s, const struct trace_entry *ent) +{ + char cmd[TASK_COMM_LEN]; + + trace_find_cmdline(ent->pid, cmd); + + if (t_action(ent) & BLK_TC_ACT(BLK_TC_PC)) { + int ret; + + ret = trace_seq_printf(s, "%u ", t_bytes(ent)); + if (!ret) + return 0; + ret = blk_log_dump_pdu(s, ent); + if (!ret) + return 0; + return trace_seq_printf(s, "[%s]\n", cmd); + } else { + if (t_sec(ent)) + return trace_seq_printf(s, "%llu + %u [%s]\n", + t_sector(ent), t_sec(ent), cmd); + return trace_seq_printf(s, "[%s]\n", cmd); + } +} + +static int blk_log_with_error(struct trace_seq *s, + const struct trace_entry *ent) +{ + if (t_action(ent) & BLK_TC_ACT(BLK_TC_PC)) { + int ret; + + ret = blk_log_dump_pdu(s, ent); + if (ret) + return trace_seq_printf(s, "[%d]\n", t_error(ent)); + return 0; + } else { + if (t_sec(ent)) + return trace_seq_printf(s, "%llu + %u [%d]\n", + t_sector(ent), + t_sec(ent), t_error(ent)); + return trace_seq_printf(s, "%llu [%d]\n", + t_sector(ent), t_error(ent)); + } +} + +static int blk_log_remap(struct trace_seq *s, const struct trace_entry *ent) +{ + struct blk_io_trace_remap r = { .device_from = 0, }; + + get_pdu_remap(ent, &r); + return trace_seq_printf(s, "%llu + %u <- (%d,%d) %llu\n", + t_sector(ent), t_sec(ent), + MAJOR(r.device_from), MINOR(r.device_from), + (unsigned long long)r.sector_from); +} + +static int blk_log_plug(struct trace_seq *s, const struct trace_entry *ent) +{ + char cmd[TASK_COMM_LEN]; + + trace_find_cmdline(ent->pid, cmd); + + return trace_seq_printf(s, "[%s]\n", cmd); +} + +static int blk_log_unplug(struct trace_seq *s, const struct trace_entry *ent) +{ + char cmd[TASK_COMM_LEN]; + + trace_find_cmdline(ent->pid, cmd); + + return trace_seq_printf(s, "[%s] %llu\n", cmd, get_pdu_int(ent)); +} + +static int blk_log_split(struct trace_seq *s, const struct trace_entry *ent) +{ + char cmd[TASK_COMM_LEN]; + + trace_find_cmdline(ent->pid, cmd); + + return trace_seq_printf(s, "%llu / %llu [%s]\n", t_sector(ent), + get_pdu_int(ent), cmd); +} + +static int blk_log_msg(struct trace_seq *s, const struct trace_entry *ent) +{ + int ret; + const struct blk_io_trace *t = te_blk_io_trace(ent); + + ret = trace_seq_putmem(s, t + 1, t->pdu_len); + if (ret) + return trace_seq_putc(s, '\n'); + return ret; +} + +/* + * struct tracer operations + */ + +static void blk_tracer_print_header(struct seq_file *m) +{ + if (!(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC)) + return; + seq_puts(m, "# DEV CPU TIMESTAMP PID ACT FLG\n" + "# | | | | | |\n"); +} + +static void blk_tracer_start(struct trace_array *tr) +{ + blk_tracer_enabled = true; +} + +static int blk_tracer_init(struct trace_array *tr) +{ + blk_tr = tr; + blk_tracer_start(tr); + return 0; +} + +static void blk_tracer_stop(struct trace_array *tr) +{ + blk_tracer_enabled = false; +} + +static void blk_tracer_reset(struct trace_array *tr) +{ + blk_tracer_stop(tr); +} + +static const struct { + const char *act[2]; + int (*print)(struct trace_seq *s, const struct trace_entry *ent); +} what2act[] = { + [__BLK_TA_QUEUE] = {{ "Q", "queue" }, blk_log_generic }, + [__BLK_TA_BACKMERGE] = {{ "M", "backmerge" }, blk_log_generic }, + [__BLK_TA_FRONTMERGE] = {{ "F", "frontmerge" }, blk_log_generic }, + [__BLK_TA_GETRQ] = {{ "G", "getrq" }, blk_log_generic }, + [__BLK_TA_SLEEPRQ] = {{ "S", "sleeprq" }, blk_log_generic }, + [__BLK_TA_REQUEUE] = {{ "R", "requeue" }, blk_log_with_error }, + [__BLK_TA_ISSUE] = {{ "D", "issue" }, blk_log_generic }, + [__BLK_TA_COMPLETE] = {{ "C", "complete" }, blk_log_with_error }, + [__BLK_TA_PLUG] = {{ "P", "plug" }, blk_log_plug }, + [__BLK_TA_UNPLUG_IO] = {{ "U", "unplug_io" }, blk_log_unplug }, + [__BLK_TA_UNPLUG_TIMER] = {{ "UT", "unplug_timer" }, blk_log_unplug }, + [__BLK_TA_INSERT] = {{ "I", "insert" }, blk_log_generic }, + [__BLK_TA_SPLIT] = {{ "X", "split" }, blk_log_split }, + [__BLK_TA_BOUNCE] = {{ "B", "bounce" }, blk_log_generic }, + [__BLK_TA_REMAP] = {{ "A", "remap" }, blk_log_remap }, +}; + +static enum print_line_t print_one_line(struct trace_iterator *iter, + bool classic) +{ + struct trace_seq *s = &iter->seq; + const struct blk_io_trace *t; + u16 what; + int ret; + bool long_act; + blk_log_action_t *log_action; + + t = te_blk_io_trace(iter->ent); + what = t->action & ((1 << BLK_TC_SHIFT) - 1); + long_act = !!(trace_flags & TRACE_ITER_VERBOSE); + log_action = classic ? &blk_log_action_classic : &blk_log_action; + + if (t->action == BLK_TN_MESSAGE) { + ret = log_action(iter, long_act ? "message" : "m"); + if (ret) + ret = blk_log_msg(s, iter->ent); + goto out; + } + + if (unlikely(what == 0 || what >= ARRAY_SIZE(what2act))) + ret = trace_seq_printf(s, "Unknown action %x\n", what); + else { + ret = log_action(iter, what2act[what].act[long_act]); + if (ret) + ret = what2act[what].print(s, iter->ent); + } +out: + return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; +} + +static enum print_line_t blk_trace_event_print(struct trace_iterator *iter, + int flags) +{ + return print_one_line(iter, false); +} + +static int blk_trace_synthesize_old_trace(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct blk_io_trace *t = (struct blk_io_trace *)iter->ent; + const int offset = offsetof(struct blk_io_trace, sector); + struct blk_io_trace old = { + .magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION, + .time = iter->ts, + }; + + if (!trace_seq_putmem(s, &old, offset)) + return 0; + return trace_seq_putmem(s, &t->sector, + sizeof(old) - offset + t->pdu_len); +} + +static enum print_line_t +blk_trace_event_print_binary(struct trace_iterator *iter, int flags) +{ + return blk_trace_synthesize_old_trace(iter) ? + TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; +} + +static enum print_line_t blk_tracer_print_line(struct trace_iterator *iter) +{ + if (!(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC)) + return TRACE_TYPE_UNHANDLED; + + return print_one_line(iter, true); +} + +static int blk_tracer_set_flag(u32 old_flags, u32 bit, int set) +{ + /* don't output context-info for blk_classic output */ + if (bit == TRACE_BLK_OPT_CLASSIC) { + if (set) + trace_flags &= ~TRACE_ITER_CONTEXT_INFO; + else + trace_flags |= TRACE_ITER_CONTEXT_INFO; + } + return 0; +} + +static struct tracer blk_tracer __read_mostly = { + .name = "blk", + .init = blk_tracer_init, + .reset = blk_tracer_reset, + .start = blk_tracer_start, + .stop = blk_tracer_stop, + .print_header = blk_tracer_print_header, + .print_line = blk_tracer_print_line, + .flags = &blk_tracer_flags, + .set_flag = blk_tracer_set_flag, +}; + +static struct trace_event trace_blk_event = { + .type = TRACE_BLK, + .trace = blk_trace_event_print, + .binary = blk_trace_event_print_binary, +}; + +static int __init init_blk_tracer(void) +{ + if (!register_ftrace_event(&trace_blk_event)) { + pr_warning("Warning: could not register block events\n"); + return 1; + } + + if (register_tracer(&blk_tracer) != 0) { + pr_warning("Warning: could not register the block tracer\n"); + unregister_ftrace_event(&trace_blk_event); + return 1; + } + + return 0; +} + +device_initcall(init_blk_tracer); + +static int blk_trace_remove_queue(struct request_queue *q) +{ + struct blk_trace *bt; + + bt = xchg(&q->blk_trace, NULL); + if (bt == NULL) + return -EINVAL; + + if (atomic_dec_and_test(&blk_probes_ref)) + blk_unregister_tracepoints(); + + blk_trace_free(bt); + return 0; +} + +/* + * Setup everything required to start tracing + */ +static int blk_trace_setup_queue(struct request_queue *q, + struct block_device *bdev) +{ + struct blk_trace *old_bt, *bt = NULL; + int ret = -ENOMEM; + + bt = kzalloc(sizeof(*bt), GFP_KERNEL); + if (!bt) + return -ENOMEM; + + bt->msg_data = __alloc_percpu(BLK_TN_MAX_MSG, __alignof__(char)); + if (!bt->msg_data) + goto free_bt; + + bt->dev = bdev->bd_dev; + bt->act_mask = (u16)-1; + + blk_trace_setup_lba(bt, bdev); + + old_bt = xchg(&q->blk_trace, bt); + if (old_bt != NULL) { + (void)xchg(&q->blk_trace, old_bt); + ret = -EBUSY; + goto free_bt; + } + + if (atomic_inc_return(&blk_probes_ref) == 1) + blk_register_tracepoints(); + return 0; + +free_bt: + blk_trace_free(bt); + return ret; +} + +/* + * sysfs interface to enable and configure tracing + */ + +static ssize_t sysfs_blk_trace_attr_show(struct device *dev, + struct device_attribute *attr, + char *buf); +static ssize_t sysfs_blk_trace_attr_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count); +#define BLK_TRACE_DEVICE_ATTR(_name) \ + DEVICE_ATTR(_name, S_IRUGO | S_IWUSR, \ + sysfs_blk_trace_attr_show, \ + sysfs_blk_trace_attr_store) + +static BLK_TRACE_DEVICE_ATTR(enable); +static BLK_TRACE_DEVICE_ATTR(act_mask); +static BLK_TRACE_DEVICE_ATTR(pid); +static BLK_TRACE_DEVICE_ATTR(start_lba); +static BLK_TRACE_DEVICE_ATTR(end_lba); + +static struct attribute *blk_trace_attrs[] = { + &dev_attr_enable.attr, + &dev_attr_act_mask.attr, + &dev_attr_pid.attr, + &dev_attr_start_lba.attr, + &dev_attr_end_lba.attr, + NULL +}; + +struct attribute_group blk_trace_attr_group = { + .name = "trace", + .attrs = blk_trace_attrs, +}; + +static const struct { + int mask; + const char *str; +} mask_maps[] = { + { BLK_TC_READ, "read" }, + { BLK_TC_WRITE, "write" }, + { BLK_TC_BARRIER, "barrier" }, + { BLK_TC_SYNC, "sync" }, + { BLK_TC_QUEUE, "queue" }, + { BLK_TC_REQUEUE, "requeue" }, + { BLK_TC_ISSUE, "issue" }, + { BLK_TC_COMPLETE, "complete" }, + { BLK_TC_FS, "fs" }, + { BLK_TC_PC, "pc" }, + { BLK_TC_AHEAD, "ahead" }, + { BLK_TC_META, "meta" }, + { BLK_TC_DISCARD, "discard" }, + { BLK_TC_DRV_DATA, "drv_data" }, +}; + +static int blk_trace_str2mask(const char *str) +{ + int i; + int mask = 0; + char *buf, *s, *token; + + buf = kstrdup(str, GFP_KERNEL); + if (buf == NULL) + return -ENOMEM; + s = strstrip(buf); + + while (1) { + token = strsep(&s, ","); + if (token == NULL) + break; + + if (*token == '\0') + continue; + + for (i = 0; i < ARRAY_SIZE(mask_maps); i++) { + if (strcasecmp(token, mask_maps[i].str) == 0) { + mask |= mask_maps[i].mask; + break; + } + } + if (i == ARRAY_SIZE(mask_maps)) { + mask = -EINVAL; + break; + } + } + kfree(buf); + + return mask; +} + +static ssize_t blk_trace_mask2str(char *buf, int mask) +{ + int i; + char *p = buf; + + for (i = 0; i < ARRAY_SIZE(mask_maps); i++) { + if (mask & mask_maps[i].mask) { + p += sprintf(p, "%s%s", + (p == buf) ? "" : ",", mask_maps[i].str); + } + } + *p++ = '\n'; + + return p - buf; +} + +static struct request_queue *blk_trace_get_queue(struct block_device *bdev) +{ + if (bdev->bd_disk == NULL) + return NULL; + + return bdev_get_queue(bdev); +} + +static ssize_t sysfs_blk_trace_attr_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct hd_struct *p = dev_to_part(dev); + struct request_queue *q; + struct block_device *bdev; + ssize_t ret = -ENXIO; + + lock_kernel(); + bdev = bdget(part_devt(p)); + if (bdev == NULL) + goto out_unlock_kernel; + + q = blk_trace_get_queue(bdev); + if (q == NULL) + goto out_bdput; + + mutex_lock(&bdev->bd_mutex); + + if (attr == &dev_attr_enable) { + ret = sprintf(buf, "%u\n", !!q->blk_trace); + goto out_unlock_bdev; + } + + if (q->blk_trace == NULL) + ret = sprintf(buf, "disabled\n"); + else if (attr == &dev_attr_act_mask) + ret = blk_trace_mask2str(buf, q->blk_trace->act_mask); + else if (attr == &dev_attr_pid) + ret = sprintf(buf, "%u\n", q->blk_trace->pid); + else if (attr == &dev_attr_start_lba) + ret = sprintf(buf, "%llu\n", q->blk_trace->start_lba); + else if (attr == &dev_attr_end_lba) + ret = sprintf(buf, "%llu\n", q->blk_trace->end_lba); + +out_unlock_bdev: + mutex_unlock(&bdev->bd_mutex); +out_bdput: + bdput(bdev); +out_unlock_kernel: + unlock_kernel(); + return ret; +} + +static ssize_t sysfs_blk_trace_attr_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct block_device *bdev; + struct request_queue *q; + struct hd_struct *p; + u64 value; + ssize_t ret = -EINVAL; + + if (count == 0) + goto out; + + if (attr == &dev_attr_act_mask) { + if (sscanf(buf, "%llx", &value) != 1) { + /* Assume it is a list of trace category names */ + ret = blk_trace_str2mask(buf); + if (ret < 0) + goto out; + value = ret; + } + } else if (sscanf(buf, "%llu", &value) != 1) + goto out; + + ret = -ENXIO; + + lock_kernel(); + p = dev_to_part(dev); + bdev = bdget(part_devt(p)); + if (bdev == NULL) + goto out_unlock_kernel; + + q = blk_trace_get_queue(bdev); + if (q == NULL) + goto out_bdput; + + mutex_lock(&bdev->bd_mutex); + + if (attr == &dev_attr_enable) { + if (value) + ret = blk_trace_setup_queue(q, bdev); + else + ret = blk_trace_remove_queue(q); + goto out_unlock_bdev; + } + + ret = 0; + if (q->blk_trace == NULL) + ret = blk_trace_setup_queue(q, bdev); + + if (ret == 0) { + if (attr == &dev_attr_act_mask) + q->blk_trace->act_mask = value; + else if (attr == &dev_attr_pid) + q->blk_trace->pid = value; + else if (attr == &dev_attr_start_lba) + q->blk_trace->start_lba = value; + else if (attr == &dev_attr_end_lba) + q->blk_trace->end_lba = value; + } + +out_unlock_bdev: + mutex_unlock(&bdev->bd_mutex); +out_bdput: + bdput(bdev); +out_unlock_kernel: + unlock_kernel(); +out: + return ret ? ret : count; +} + +int blk_trace_init_sysfs(struct device *dev) +{ + return sysfs_create_group(&dev->kobj, &blk_trace_attr_group); +} + +#endif /* CONFIG_BLK_DEV_IO_TRACE */ + +#ifdef CONFIG_EVENT_TRACING + +void blk_dump_cmd(char *buf, struct request *rq) +{ + int i, end; + int len = rq->cmd_len; + unsigned char *cmd = rq->cmd; + + if (!blk_pc_request(rq)) { + buf[0] = '\0'; + return; + } + + for (end = len - 1; end >= 0; end--) + if (cmd[end]) + break; + end++; + + for (i = 0; i < len; i++) { + buf += sprintf(buf, "%s%02x", i == 0 ? "" : " ", cmd[i]); + if (i == end && end != len - 1) { + sprintf(buf, " .."); + break; + } + } +} + +void blk_fill_rwbs(char *rwbs, u32 rw, int bytes) +{ + int i = 0; + + if (rw & WRITE) + rwbs[i++] = 'W'; + else if (rw & 1 << BIO_RW_DISCARD) + rwbs[i++] = 'D'; + else if (bytes) + rwbs[i++] = 'R'; + else + rwbs[i++] = 'N'; + + if (rw & 1 << BIO_RW_AHEAD) + rwbs[i++] = 'A'; + if (rw & 1 << BIO_RW_BARRIER) + rwbs[i++] = 'B'; + if (rw & 1 << BIO_RW_SYNCIO) + rwbs[i++] = 'S'; + if (rw & 1 << BIO_RW_META) + rwbs[i++] = 'M'; + + rwbs[i] = '\0'; +} + +void blk_fill_rwbs_rq(char *rwbs, struct request *rq) +{ + int rw = rq->cmd_flags & 0x03; + int bytes; + + if (blk_discard_rq(rq)) + rw |= (1 << BIO_RW_DISCARD); + + bytes = blk_rq_bytes(rq); + + blk_fill_rwbs(rwbs, rw, bytes); +} + +#endif /* CONFIG_EVENT_TRACING */ + diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 2f32969c09d..8c804e24f96 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -17,6 +17,7 @@ #include <linux/clocksource.h> #include <linux/kallsyms.h> #include <linux/seq_file.h> +#include <linux/suspend.h> #include <linux/debugfs.h> #include <linux/hardirq.h> #include <linux/kthread.h> @@ -26,10 +27,15 @@ #include <linux/sysctl.h> #include <linux/ctype.h> #include <linux/list.h> +#include <linux/hash.h> + +#include <trace/events/sched.h> #include <asm/ftrace.h> +#include <asm/setup.h> -#include "trace.h" +#include "trace_output.h" +#include "trace_stat.h" #define FTRACE_WARN_ON(cond) \ do { \ @@ -43,14 +49,14 @@ ftrace_kill(); \ } while (0) +/* hash bits for specific function selection */ +#define FTRACE_HASH_BITS 7 +#define FTRACE_FUNC_HASHSIZE (1 << FTRACE_HASH_BITS) + /* ftrace_enabled is a method to turn ftrace on or off */ int ftrace_enabled __read_mostly; static int last_ftrace_enabled; -/* set when tracing only a pid */ -struct pid *ftrace_pid_trace; -static struct pid * const ftrace_swapper_pid = &init_struct_pid; - /* Quick disabling of function tracer. */ int function_trace_stop; @@ -60,13 +66,11 @@ int function_trace_stop; */ static int ftrace_disabled __read_mostly; -static DEFINE_SPINLOCK(ftrace_lock); -static DEFINE_MUTEX(ftrace_sysctl_lock); -static DEFINE_MUTEX(ftrace_start_lock); +static DEFINE_MUTEX(ftrace_lock); static struct ftrace_ops ftrace_list_end __read_mostly = { - .func = ftrace_stub, + .func = ftrace_stub, }; static struct ftrace_ops *ftrace_list __read_mostly = &ftrace_list_end; @@ -133,9 +137,6 @@ static void ftrace_test_stop_func(unsigned long ip, unsigned long parent_ip) static int __register_ftrace_function(struct ftrace_ops *ops) { - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); - ops->next = ftrace_list; /* * We are entering ops into the ftrace_list but another @@ -171,18 +172,12 @@ static int __register_ftrace_function(struct ftrace_ops *ops) #endif } - spin_unlock(&ftrace_lock); - return 0; } static int __unregister_ftrace_function(struct ftrace_ops *ops) { struct ftrace_ops **p; - int ret = 0; - - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); /* * If we are removing the last function, then simply point @@ -191,17 +186,15 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops) if (ftrace_list == ops && ops->next == &ftrace_list_end) { ftrace_trace_function = ftrace_stub; ftrace_list = &ftrace_list_end; - goto out; + return 0; } for (p = &ftrace_list; *p != &ftrace_list_end; p = &(*p)->next) if (*p == ops) break; - if (*p != ops) { - ret = -1; - goto out; - } + if (*p != ops) + return -1; *p = (*p)->next; @@ -222,21 +215,15 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops) } } - out: - spin_unlock(&ftrace_lock); - - return ret; + return 0; } static void ftrace_update_pid_func(void) { ftrace_func_t func; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); - if (ftrace_trace_function == ftrace_stub) - goto out; + return; func = ftrace_trace_function; @@ -253,23 +240,603 @@ static void ftrace_update_pid_func(void) #else __ftrace_trace_function = func; #endif +} + +#ifdef CONFIG_FUNCTION_PROFILER +struct ftrace_profile { + struct hlist_node node; + unsigned long ip; + unsigned long counter; +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + unsigned long long time; +#endif +}; + +struct ftrace_profile_page { + struct ftrace_profile_page *next; + unsigned long index; + struct ftrace_profile records[]; +}; + +struct ftrace_profile_stat { + atomic_t disabled; + struct hlist_head *hash; + struct ftrace_profile_page *pages; + struct ftrace_profile_page *start; + struct tracer_stat stat; +}; + +#define PROFILE_RECORDS_SIZE \ + (PAGE_SIZE - offsetof(struct ftrace_profile_page, records)) + +#define PROFILES_PER_PAGE \ + (PROFILE_RECORDS_SIZE / sizeof(struct ftrace_profile)) + +static int ftrace_profile_bits __read_mostly; +static int ftrace_profile_enabled __read_mostly; + +/* ftrace_profile_lock - synchronize the enable and disable of the profiler */ +static DEFINE_MUTEX(ftrace_profile_lock); + +static DEFINE_PER_CPU(struct ftrace_profile_stat, ftrace_profile_stats); + +#define FTRACE_PROFILE_HASH_SIZE 1024 /* must be power of 2 */ + +static void * +function_stat_next(void *v, int idx) +{ + struct ftrace_profile *rec = v; + struct ftrace_profile_page *pg; + + pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK); + + again: + if (idx != 0) + rec++; + + if ((void *)rec >= (void *)&pg->records[pg->index]) { + pg = pg->next; + if (!pg) + return NULL; + rec = &pg->records[0]; + if (!rec->counter) + goto again; + } + + return rec; +} + +static void *function_stat_start(struct tracer_stat *trace) +{ + struct ftrace_profile_stat *stat = + container_of(trace, struct ftrace_profile_stat, stat); + + if (!stat || !stat->start) + return NULL; + + return function_stat_next(&stat->start->records[0], 0); +} + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +/* function graph compares on total time */ +static int function_stat_cmp(void *p1, void *p2) +{ + struct ftrace_profile *a = p1; + struct ftrace_profile *b = p2; + + if (a->time < b->time) + return -1; + if (a->time > b->time) + return 1; + else + return 0; +} +#else +/* not function graph compares against hits */ +static int function_stat_cmp(void *p1, void *p2) +{ + struct ftrace_profile *a = p1; + struct ftrace_profile *b = p2; + + if (a->counter < b->counter) + return -1; + if (a->counter > b->counter) + return 1; + else + return 0; +} +#endif + +static int function_stat_headers(struct seq_file *m) +{ +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + seq_printf(m, " Function " + "Hit Time Avg\n" + " -------- " + "--- ---- ---\n"); +#else + seq_printf(m, " Function Hit\n" + " -------- ---\n"); +#endif + return 0; +} + +static int function_stat_show(struct seq_file *m, void *v) +{ + struct ftrace_profile *rec = v; + char str[KSYM_SYMBOL_LEN]; +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + static DEFINE_MUTEX(mutex); + static struct trace_seq s; + unsigned long long avg; +#endif + + kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); + seq_printf(m, " %-30.30s %10lu", str, rec->counter); + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + seq_printf(m, " "); + avg = rec->time; + do_div(avg, rec->counter); + + mutex_lock(&mutex); + trace_seq_init(&s); + trace_print_graph_duration(rec->time, &s); + trace_seq_puts(&s, " "); + trace_print_graph_duration(avg, &s); + trace_print_seq(m, &s); + mutex_unlock(&mutex); +#endif + seq_putc(m, '\n'); + + return 0; +} + +static void ftrace_profile_reset(struct ftrace_profile_stat *stat) +{ + struct ftrace_profile_page *pg; + + pg = stat->pages = stat->start; + + while (pg) { + memset(pg->records, 0, PROFILE_RECORDS_SIZE); + pg->index = 0; + pg = pg->next; + } + + memset(stat->hash, 0, + FTRACE_PROFILE_HASH_SIZE * sizeof(struct hlist_head)); +} + +int ftrace_profile_pages_init(struct ftrace_profile_stat *stat) +{ + struct ftrace_profile_page *pg; + int functions; + int pages; + int i; + + /* If we already allocated, do nothing */ + if (stat->pages) + return 0; + + stat->pages = (void *)get_zeroed_page(GFP_KERNEL); + if (!stat->pages) + return -ENOMEM; + +#ifdef CONFIG_DYNAMIC_FTRACE + functions = ftrace_update_tot_cnt; +#else + /* + * We do not know the number of functions that exist because + * dynamic tracing is what counts them. With past experience + * we have around 20K functions. That should be more than enough. + * It is highly unlikely we will execute every function in + * the kernel. + */ + functions = 20000; +#endif + + pg = stat->start = stat->pages; + + pages = DIV_ROUND_UP(functions, PROFILES_PER_PAGE); + + for (i = 0; i < pages; i++) { + pg->next = (void *)get_zeroed_page(GFP_KERNEL); + if (!pg->next) + goto out_free; + pg = pg->next; + } + + return 0; + + out_free: + pg = stat->start; + while (pg) { + unsigned long tmp = (unsigned long)pg; + + pg = pg->next; + free_page(tmp); + } + + free_page((unsigned long)stat->pages); + stat->pages = NULL; + stat->start = NULL; + + return -ENOMEM; +} + +static int ftrace_profile_init_cpu(int cpu) +{ + struct ftrace_profile_stat *stat; + int size; + + stat = &per_cpu(ftrace_profile_stats, cpu); + + if (stat->hash) { + /* If the profile is already created, simply reset it */ + ftrace_profile_reset(stat); + return 0; + } + + /* + * We are profiling all functions, but usually only a few thousand + * functions are hit. We'll make a hash of 1024 items. + */ + size = FTRACE_PROFILE_HASH_SIZE; + + stat->hash = kzalloc(sizeof(struct hlist_head) * size, GFP_KERNEL); + + if (!stat->hash) + return -ENOMEM; + + if (!ftrace_profile_bits) { + size--; + + for (; size; size >>= 1) + ftrace_profile_bits++; + } + + /* Preallocate the function profiling pages */ + if (ftrace_profile_pages_init(stat) < 0) { + kfree(stat->hash); + stat->hash = NULL; + return -ENOMEM; + } + + return 0; +} + +static int ftrace_profile_init(void) +{ + int cpu; + int ret = 0; + + for_each_online_cpu(cpu) { + ret = ftrace_profile_init_cpu(cpu); + if (ret) + break; + } + + return ret; +} + +/* interrupts must be disabled */ +static struct ftrace_profile * +ftrace_find_profiled_func(struct ftrace_profile_stat *stat, unsigned long ip) +{ + struct ftrace_profile *rec; + struct hlist_head *hhd; + struct hlist_node *n; + unsigned long key; + + key = hash_long(ip, ftrace_profile_bits); + hhd = &stat->hash[key]; + + if (hlist_empty(hhd)) + return NULL; + hlist_for_each_entry_rcu(rec, n, hhd, node) { + if (rec->ip == ip) + return rec; + } + + return NULL; +} + +static void ftrace_add_profile(struct ftrace_profile_stat *stat, + struct ftrace_profile *rec) +{ + unsigned long key; + + key = hash_long(rec->ip, ftrace_profile_bits); + hlist_add_head_rcu(&rec->node, &stat->hash[key]); +} + +/* + * The memory is already allocated, this simply finds a new record to use. + */ +static struct ftrace_profile * +ftrace_profile_alloc(struct ftrace_profile_stat *stat, unsigned long ip) +{ + struct ftrace_profile *rec = NULL; + + /* prevent recursion (from NMIs) */ + if (atomic_inc_return(&stat->disabled) != 1) + goto out; + + /* + * Try to find the function again since an NMI + * could have added it + */ + rec = ftrace_find_profiled_func(stat, ip); + if (rec) + goto out; + + if (stat->pages->index == PROFILES_PER_PAGE) { + if (!stat->pages->next) + goto out; + stat->pages = stat->pages->next; + } + + rec = &stat->pages->records[stat->pages->index++]; + rec->ip = ip; + ftrace_add_profile(stat, rec); + + out: + atomic_dec(&stat->disabled); + + return rec; +} + +static void +function_profile_call(unsigned long ip, unsigned long parent_ip) +{ + struct ftrace_profile_stat *stat; + struct ftrace_profile *rec; + unsigned long flags; + + if (!ftrace_profile_enabled) + return; + + local_irq_save(flags); + + stat = &__get_cpu_var(ftrace_profile_stats); + if (!stat->hash || !ftrace_profile_enabled) + goto out; + + rec = ftrace_find_profiled_func(stat, ip); + if (!rec) { + rec = ftrace_profile_alloc(stat, ip); + if (!rec) + goto out; + } + + rec->counter++; out: - spin_unlock(&ftrace_lock); + local_irq_restore(flags); } +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +static int profile_graph_entry(struct ftrace_graph_ent *trace) +{ + function_profile_call(trace->func, 0); + return 1; +} + +static void profile_graph_return(struct ftrace_graph_ret *trace) +{ + struct ftrace_profile_stat *stat; + unsigned long long calltime; + struct ftrace_profile *rec; + unsigned long flags; + + local_irq_save(flags); + stat = &__get_cpu_var(ftrace_profile_stats); + if (!stat->hash || !ftrace_profile_enabled) + goto out; + + calltime = trace->rettime - trace->calltime; + + if (!(trace_flags & TRACE_ITER_GRAPH_TIME)) { + int index; + + index = trace->depth; + + /* Append this call time to the parent time to subtract */ + if (index) + current->ret_stack[index - 1].subtime += calltime; + + if (current->ret_stack[index].subtime < calltime) + calltime -= current->ret_stack[index].subtime; + else + calltime = 0; + } + + rec = ftrace_find_profiled_func(stat, trace->func); + if (rec) + rec->time += calltime; + + out: + local_irq_restore(flags); +} + +static int register_ftrace_profiler(void) +{ + return register_ftrace_graph(&profile_graph_return, + &profile_graph_entry); +} + +static void unregister_ftrace_profiler(void) +{ + unregister_ftrace_graph(); +} +#else +static struct ftrace_ops ftrace_profile_ops __read_mostly = +{ + .func = function_profile_call, +}; + +static int register_ftrace_profiler(void) +{ + return register_ftrace_function(&ftrace_profile_ops); +} + +static void unregister_ftrace_profiler(void) +{ + unregister_ftrace_function(&ftrace_profile_ops); +} +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ + +static ssize_t +ftrace_profile_write(struct file *filp, const char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + unsigned long val; + char buf[64]; /* big enough to hold a number */ + int ret; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + ret = strict_strtoul(buf, 10, &val); + if (ret < 0) + return ret; + + val = !!val; + + mutex_lock(&ftrace_profile_lock); + if (ftrace_profile_enabled ^ val) { + if (val) { + ret = ftrace_profile_init(); + if (ret < 0) { + cnt = ret; + goto out; + } + + ret = register_ftrace_profiler(); + if (ret < 0) { + cnt = ret; + goto out; + } + ftrace_profile_enabled = 1; + } else { + ftrace_profile_enabled = 0; + /* + * unregister_ftrace_profiler calls stop_machine + * so this acts like an synchronize_sched. + */ + unregister_ftrace_profiler(); + } + } + out: + mutex_unlock(&ftrace_profile_lock); + + filp->f_pos += cnt; + + return cnt; +} + +static ssize_t +ftrace_profile_read(struct file *filp, char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + char buf[64]; /* big enough to hold a number */ + int r; + + r = sprintf(buf, "%u\n", ftrace_profile_enabled); + return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); +} + +static const struct file_operations ftrace_profile_fops = { + .open = tracing_open_generic, + .read = ftrace_profile_read, + .write = ftrace_profile_write, +}; + +/* used to initialize the real stat files */ +static struct tracer_stat function_stats __initdata = { + .name = "functions", + .stat_start = function_stat_start, + .stat_next = function_stat_next, + .stat_cmp = function_stat_cmp, + .stat_headers = function_stat_headers, + .stat_show = function_stat_show +}; + +static __init void ftrace_profile_debugfs(struct dentry *d_tracer) +{ + struct ftrace_profile_stat *stat; + struct dentry *entry; + char *name; + int ret; + int cpu; + + for_each_possible_cpu(cpu) { + stat = &per_cpu(ftrace_profile_stats, cpu); + + /* allocate enough for function name + cpu number */ + name = kmalloc(32, GFP_KERNEL); + if (!name) { + /* + * The files created are permanent, if something happens + * we still do not free memory. + */ + WARN(1, + "Could not allocate stat file for cpu %d\n", + cpu); + return; + } + stat->stat = function_stats; + snprintf(name, 32, "function%d", cpu); + stat->stat.name = name; + ret = register_stat_tracer(&stat->stat); + if (ret) { + WARN(1, + "Could not register function stat for cpu %d\n", + cpu); + kfree(name); + return; + } + } + + entry = debugfs_create_file("function_profile_enabled", 0644, + d_tracer, NULL, &ftrace_profile_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'function_profile_enabled' entry\n"); +} + +#else /* CONFIG_FUNCTION_PROFILER */ +static __init void ftrace_profile_debugfs(struct dentry *d_tracer) +{ +} +#endif /* CONFIG_FUNCTION_PROFILER */ + +/* set when tracing only a pid */ +struct pid *ftrace_pid_trace; +static struct pid * const ftrace_swapper_pid = &init_struct_pid; + #ifdef CONFIG_DYNAMIC_FTRACE + #ifndef CONFIG_FTRACE_MCOUNT_RECORD # error Dynamic ftrace depends on MCOUNT_RECORD #endif -/* - * Since MCOUNT_ADDR may point to mcount itself, we do not want - * to get it confused by reading a reference in the code as we - * are parsing on objcopy output of text. Use a variable for - * it instead. - */ -static unsigned long mcount_addr = MCOUNT_ADDR; +static struct hlist_head ftrace_func_hash[FTRACE_FUNC_HASHSIZE] __read_mostly; + +struct ftrace_func_probe { + struct hlist_node node; + struct ftrace_probe_ops *ops; + unsigned long flags; + unsigned long ip; + void *data; + struct rcu_head rcu; +}; enum { FTRACE_ENABLE_CALLS = (1 << 0), @@ -283,13 +850,13 @@ enum { static int ftrace_filtered; -static LIST_HEAD(ftrace_new_addrs); +static struct dyn_ftrace *ftrace_new_addrs; static DEFINE_MUTEX(ftrace_regex_lock); struct ftrace_page { struct ftrace_page *next; - unsigned long index; + int index; struct dyn_ftrace records[]; }; @@ -304,6 +871,19 @@ static struct ftrace_page *ftrace_pages; static struct dyn_ftrace *ftrace_free_records; +/* + * This is a double for. Do not use 'break' to break out of the loop, + * you must use a goto. + */ +#define do_for_each_ftrace_rec(pg, rec) \ + for (pg = ftrace_pages_start; pg; pg = pg->next) { \ + int _____i; \ + for (_____i = 0; _____i < pg->index; _____i++) { \ + rec = &pg->records[_____i]; + +#define while_for_each_ftrace_rec() \ + } \ + } #ifdef CONFIG_KPROBES @@ -337,36 +917,11 @@ static inline int record_frozen(struct dyn_ftrace *rec) static void ftrace_free_rec(struct dyn_ftrace *rec) { - rec->ip = (unsigned long)ftrace_free_records; + rec->freelist = ftrace_free_records; ftrace_free_records = rec; rec->flags |= FTRACE_FL_FREE; } -void ftrace_release(void *start, unsigned long size) -{ - struct dyn_ftrace *rec; - struct ftrace_page *pg; - unsigned long s = (unsigned long)start; - unsigned long e = s + size; - int i; - - if (ftrace_disabled || !start) - return; - - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); - - for (pg = ftrace_pages_start; pg; pg = pg->next) { - for (i = 0; i < pg->index; i++) { - rec = &pg->records[i]; - - if ((rec->ip >= s) && (rec->ip < e)) - ftrace_free_rec(rec); - } - } - spin_unlock(&ftrace_lock); -} - static struct dyn_ftrace *ftrace_alloc_dyn_node(unsigned long ip) { struct dyn_ftrace *rec; @@ -381,7 +936,7 @@ static struct dyn_ftrace *ftrace_alloc_dyn_node(unsigned long ip) return NULL; } - ftrace_free_records = (void *)rec->ip; + ftrace_free_records = rec->freelist; memset(rec, 0, sizeof(*rec)); return rec; } @@ -413,8 +968,8 @@ ftrace_record_ip(unsigned long ip) return NULL; rec->ip = ip; - - list_add(&rec->list, &ftrace_new_addrs); + rec->newlist = ftrace_new_addrs; + ftrace_new_addrs = rec; return rec; } @@ -460,111 +1015,75 @@ static void ftrace_bug(int failed, unsigned long ip) static int __ftrace_replace_code(struct dyn_ftrace *rec, int enable) { - unsigned long ip, fl; unsigned long ftrace_addr; + unsigned long flag = 0UL; - ftrace_addr = (unsigned long)ftrace_caller; - - ip = rec->ip; + ftrace_addr = (unsigned long)FTRACE_ADDR; /* - * If this record is not to be traced and - * it is not enabled then do nothing. + * If this record is not to be traced or we want to disable it, + * then disable it. * - * If this record is not to be traced and - * it is enabled then disabled it. + * If we want to enable it and filtering is off, then enable it. * + * If we want to enable it and filtering is on, enable it only if + * it's filtered */ - if (rec->flags & FTRACE_FL_NOTRACE) { - if (rec->flags & FTRACE_FL_ENABLED) - rec->flags &= ~FTRACE_FL_ENABLED; - else - return 0; - - } else if (ftrace_filtered && enable) { - /* - * Filtering is on: - */ - - fl = rec->flags & (FTRACE_FL_FILTER | FTRACE_FL_ENABLED); - - /* Record is filtered and enabled, do nothing */ - if (fl == (FTRACE_FL_FILTER | FTRACE_FL_ENABLED)) - return 0; - - /* Record is not filtered and is not enabled do nothing */ - if (!fl) - return 0; - - /* Record is not filtered but enabled, disable it */ - if (fl == FTRACE_FL_ENABLED) - rec->flags &= ~FTRACE_FL_ENABLED; - else - /* Otherwise record is filtered but not enabled, enable it */ - rec->flags |= FTRACE_FL_ENABLED; - } else { - /* Disable or not filtered */ - - if (enable) { - /* if record is enabled, do nothing */ - if (rec->flags & FTRACE_FL_ENABLED) - return 0; - - rec->flags |= FTRACE_FL_ENABLED; - - } else { + if (enable && !(rec->flags & FTRACE_FL_NOTRACE)) { + if (!ftrace_filtered || (rec->flags & FTRACE_FL_FILTER)) + flag = FTRACE_FL_ENABLED; + } - /* if record is not enabled do nothing */ - if (!(rec->flags & FTRACE_FL_ENABLED)) - return 0; + /* If the state of this record hasn't changed, then do nothing */ + if ((rec->flags & FTRACE_FL_ENABLED) == flag) + return 0; - rec->flags &= ~FTRACE_FL_ENABLED; - } + if (flag) { + rec->flags |= FTRACE_FL_ENABLED; + return ftrace_make_call(rec, ftrace_addr); } - if (rec->flags & FTRACE_FL_ENABLED) - return ftrace_make_call(rec, ftrace_addr); - else - return ftrace_make_nop(NULL, rec, ftrace_addr); + rec->flags &= ~FTRACE_FL_ENABLED; + return ftrace_make_nop(NULL, rec, ftrace_addr); } static void ftrace_replace_code(int enable) { - int i, failed; struct dyn_ftrace *rec; struct ftrace_page *pg; + int failed; - for (pg = ftrace_pages_start; pg; pg = pg->next) { - for (i = 0; i < pg->index; i++) { - rec = &pg->records[i]; - - /* - * Skip over free records and records that have - * failed. - */ - if (rec->flags & FTRACE_FL_FREE || - rec->flags & FTRACE_FL_FAILED) - continue; + do_for_each_ftrace_rec(pg, rec) { + /* + * Skip over free records, records that have + * failed and not converted. + */ + if (rec->flags & FTRACE_FL_FREE || + rec->flags & FTRACE_FL_FAILED || + !(rec->flags & FTRACE_FL_CONVERTED)) + continue; - /* ignore updates to this record's mcount site */ - if (get_kprobe((void *)rec->ip)) { - freeze_record(rec); - continue; - } else { - unfreeze_record(rec); - } + /* ignore updates to this record's mcount site */ + if (get_kprobe((void *)rec->ip)) { + freeze_record(rec); + continue; + } else { + unfreeze_record(rec); + } - failed = __ftrace_replace_code(rec, enable); - if (failed && (rec->flags & FTRACE_FL_CONVERTED)) { - rec->flags |= FTRACE_FL_FAILED; - if ((system_state == SYSTEM_BOOTING) || - !core_kernel_text(rec->ip)) { - ftrace_free_rec(rec); - } else - ftrace_bug(failed, rec->ip); - } + failed = __ftrace_replace_code(rec, enable); + if (failed) { + rec->flags |= FTRACE_FL_FAILED; + if ((system_state == SYSTEM_BOOTING) || + !core_kernel_text(rec->ip)) { + ftrace_free_rec(rec); + } else { + ftrace_bug(failed, rec->ip); + /* Stop processing */ + return; + } } - } + } while_for_each_ftrace_rec(); } static int @@ -575,7 +1094,7 @@ ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec) ip = rec->ip; - ret = ftrace_make_nop(mod, rec, mcount_addr); + ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR); if (ret) { ftrace_bug(ret, ip); rec->flags |= FTRACE_FL_FAILED; @@ -584,6 +1103,24 @@ ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec) return 1; } +/* + * archs can override this function if they must do something + * before the modifying code is performed. + */ +int __weak ftrace_arch_code_modify_prepare(void) +{ + return 0; +} + +/* + * archs can override this function if they must do something + * after the modifying code is performed. + */ +int __weak ftrace_arch_code_modify_post_process(void) +{ + return 0; +} + static int __ftrace_modify_code(void *data) { int *command = data; @@ -606,7 +1143,17 @@ static int __ftrace_modify_code(void *data) static void ftrace_run_update_code(int command) { + int ret; + + ret = ftrace_arch_code_modify_prepare(); + FTRACE_WARN_ON(ret); + if (ret) + return; + stop_machine(__ftrace_modify_code, &command, NULL); + + ret = ftrace_arch_code_modify_post_process(); + FTRACE_WARN_ON(ret); } static ftrace_func_t saved_ftrace_func; @@ -630,13 +1177,10 @@ static void ftrace_startup(int command) if (unlikely(ftrace_disabled)) return; - mutex_lock(&ftrace_start_lock); ftrace_start_up++; command |= FTRACE_ENABLE_CALLS; ftrace_startup_enable(command); - - mutex_unlock(&ftrace_start_lock); } static void ftrace_shutdown(int command) @@ -644,8 +1188,14 @@ static void ftrace_shutdown(int command) if (unlikely(ftrace_disabled)) return; - mutex_lock(&ftrace_start_lock); ftrace_start_up--; + /* + * Just warn in case of unbalance, no need to kill ftrace, it's not + * critical but the ftrace_call callers may be never nopped again after + * further ftrace uses. + */ + WARN_ON_ONCE(ftrace_start_up < 0); + if (!ftrace_start_up) command |= FTRACE_DISABLE_CALLS; @@ -655,11 +1205,9 @@ static void ftrace_shutdown(int command) } if (!command || !ftrace_enabled) - goto out; + return; ftrace_run_update_code(command); - out: - mutex_unlock(&ftrace_start_lock); } static void ftrace_startup_sysctl(void) @@ -669,7 +1217,6 @@ static void ftrace_startup_sysctl(void) if (unlikely(ftrace_disabled)) return; - mutex_lock(&ftrace_start_lock); /* Force update next time */ saved_ftrace_func = NULL; /* ftrace_start_up is true if we want ftrace running */ @@ -677,7 +1224,6 @@ static void ftrace_startup_sysctl(void) command |= FTRACE_ENABLE_CALLS; ftrace_run_update_code(command); - mutex_unlock(&ftrace_start_lock); } static void ftrace_shutdown_sysctl(void) @@ -687,13 +1233,11 @@ static void ftrace_shutdown_sysctl(void) if (unlikely(ftrace_disabled)) return; - mutex_lock(&ftrace_start_lock); /* ftrace_start_up is true if ftrace is running */ if (ftrace_start_up) command |= FTRACE_DISABLE_CALLS; ftrace_run_update_code(command); - mutex_unlock(&ftrace_start_lock); } static cycle_t ftrace_update_time; @@ -702,19 +1246,21 @@ unsigned long ftrace_update_tot_cnt; static int ftrace_update_code(struct module *mod) { - struct dyn_ftrace *p, *t; + struct dyn_ftrace *p; cycle_t start, stop; start = ftrace_now(raw_smp_processor_id()); ftrace_update_cnt = 0; - list_for_each_entry_safe(p, t, &ftrace_new_addrs, list) { + while (ftrace_new_addrs) { /* If something went wrong, bail without enabling anything */ if (unlikely(ftrace_disabled)) return -1; - list_del_init(&p->list); + p = ftrace_new_addrs; + ftrace_new_addrs = p->newlist; + p->flags = 0L; /* convert record (i.e, patch mcount-call with NOP) */ if (ftrace_code_disable(mod, p)) { @@ -780,37 +1326,116 @@ enum { FTRACE_ITER_CONT = (1 << 1), FTRACE_ITER_NOTRACE = (1 << 2), FTRACE_ITER_FAILURES = (1 << 3), + FTRACE_ITER_PRINTALL = (1 << 4), + FTRACE_ITER_HASH = (1 << 5), }; #define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */ struct ftrace_iterator { struct ftrace_page *pg; - unsigned idx; + int hidx; + int idx; unsigned flags; unsigned char buffer[FTRACE_BUFF_MAX+1]; unsigned buffer_idx; - unsigned filtered; }; static void * +t_hash_next(struct seq_file *m, void *v, loff_t *pos) +{ + struct ftrace_iterator *iter = m->private; + struct hlist_node *hnd = v; + struct hlist_head *hhd; + + WARN_ON(!(iter->flags & FTRACE_ITER_HASH)); + + (*pos)++; + + retry: + if (iter->hidx >= FTRACE_FUNC_HASHSIZE) + return NULL; + + hhd = &ftrace_func_hash[iter->hidx]; + + if (hlist_empty(hhd)) { + iter->hidx++; + hnd = NULL; + goto retry; + } + + if (!hnd) + hnd = hhd->first; + else { + hnd = hnd->next; + if (!hnd) { + iter->hidx++; + goto retry; + } + } + + return hnd; +} + +static void *t_hash_start(struct seq_file *m, loff_t *pos) +{ + struct ftrace_iterator *iter = m->private; + void *p = NULL; + loff_t l; + + if (!(iter->flags & FTRACE_ITER_HASH)) + *pos = 0; + + iter->flags |= FTRACE_ITER_HASH; + + iter->hidx = 0; + for (l = 0; l <= *pos; ) { + p = t_hash_next(m, p, &l); + if (!p) + break; + } + return p; +} + +static int t_hash_show(struct seq_file *m, void *v) +{ + struct ftrace_func_probe *rec; + struct hlist_node *hnd = v; + + rec = hlist_entry(hnd, struct ftrace_func_probe, node); + + if (rec->ops->print) + return rec->ops->print(m, rec->ip, rec->ops, rec->data); + + seq_printf(m, "%pf:%pf", (void *)rec->ip, (void *)rec->ops->func); + + if (rec->data) + seq_printf(m, ":%p", rec->data); + seq_putc(m, '\n'); + + return 0; +} + +static void * t_next(struct seq_file *m, void *v, loff_t *pos) { struct ftrace_iterator *iter = m->private; struct dyn_ftrace *rec = NULL; + if (iter->flags & FTRACE_ITER_HASH) + return t_hash_next(m, v, pos); + (*pos)++; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + if (iter->flags & FTRACE_ITER_PRINTALL) + return NULL; + retry: if (iter->idx >= iter->pg->index) { if (iter->pg->next) { iter->pg = iter->pg->next; iter->idx = 0; goto retry; - } else { - iter->idx = -1; } } else { rec = &iter->pg->records[iter->idx++]; @@ -831,7 +1456,6 @@ t_next(struct seq_file *m, void *v, loff_t *pos) goto retry; } } - spin_unlock(&ftrace_lock); return rec; } @@ -840,34 +1464,60 @@ static void *t_start(struct seq_file *m, loff_t *pos) { struct ftrace_iterator *iter = m->private; void *p = NULL; + loff_t l; + + mutex_lock(&ftrace_lock); + /* + * For set_ftrace_filter reading, if we have the filter + * off, we can short cut and just print out that all + * functions are enabled. + */ + if (iter->flags & FTRACE_ITER_FILTER && !ftrace_filtered) { + if (*pos > 0) + return t_hash_start(m, pos); + iter->flags |= FTRACE_ITER_PRINTALL; + return iter; + } + + if (iter->flags & FTRACE_ITER_HASH) + return t_hash_start(m, pos); - if (*pos > 0) { - if (iter->idx < 0) - return p; - (*pos)--; - iter->idx--; + iter->pg = ftrace_pages_start; + iter->idx = 0; + for (l = 0; l <= *pos; ) { + p = t_next(m, p, &l); + if (!p) + break; } - p = t_next(m, p, pos); + if (!p && iter->flags & FTRACE_ITER_FILTER) + return t_hash_start(m, pos); return p; } static void t_stop(struct seq_file *m, void *p) { + mutex_unlock(&ftrace_lock); } static int t_show(struct seq_file *m, void *v) { + struct ftrace_iterator *iter = m->private; struct dyn_ftrace *rec = v; - char str[KSYM_SYMBOL_LEN]; - if (!rec) + if (iter->flags & FTRACE_ITER_HASH) + return t_hash_show(m, v); + + if (iter->flags & FTRACE_ITER_PRINTALL) { + seq_printf(m, "#### all functions enabled ####\n"); return 0; + } - kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); + if (!rec) + return 0; - seq_printf(m, "%s\n", str); + seq_printf(m, "%pf\n", (void *)rec->ip); return 0; } @@ -906,17 +1556,6 @@ ftrace_avail_open(struct inode *inode, struct file *file) return ret; } -int ftrace_avail_release(struct inode *inode, struct file *file) -{ - struct seq_file *m = (struct seq_file *)file->private_data; - struct ftrace_iterator *iter = m->private; - - seq_release(inode, file); - kfree(iter); - - return 0; -} - static int ftrace_failures_open(struct inode *inode, struct file *file) { @@ -940,23 +1579,16 @@ static void ftrace_filter_reset(int enable) struct ftrace_page *pg; struct dyn_ftrace *rec; unsigned long type = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; - unsigned i; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + mutex_lock(&ftrace_lock); if (enable) ftrace_filtered = 0; - pg = ftrace_pages_start; - while (pg) { - for (i = 0; i < pg->index; i++) { - rec = &pg->records[i]; - if (rec->flags & FTRACE_FL_FAILED) - continue; - rec->flags &= ~type; - } - pg = pg->next; - } - spin_unlock(&ftrace_lock); + do_for_each_ftrace_rec(pg, rec) { + if (rec->flags & FTRACE_FL_FAILED) + continue; + rec->flags &= ~type; + } while_for_each_ftrace_rec(); + mutex_unlock(&ftrace_lock); } static int @@ -974,7 +1606,7 @@ ftrace_regex_open(struct inode *inode, struct file *file, int enable) mutex_lock(&ftrace_regex_lock); if ((file->f_mode & FMODE_WRITE) && - !(file->f_flags & O_APPEND)) + (file->f_flags & O_TRUNC)) ftrace_filter_reset(enable); if (file->f_mode & FMODE_READ) { @@ -1007,16 +1639,6 @@ ftrace_notrace_open(struct inode *inode, struct file *file) return ftrace_regex_open(inode, file, 0); } -static ssize_t -ftrace_regex_read(struct file *file, char __user *ubuf, - size_t cnt, loff_t *ppos) -{ - if (file->f_mode & FMODE_READ) - return seq_read(file, ubuf, cnt, ppos); - else - return -EPERM; -} - static loff_t ftrace_regex_lseek(struct file *file, loff_t offset, int origin) { @@ -1037,86 +1659,536 @@ enum { MATCH_END_ONLY, }; -static void -ftrace_match(unsigned char *buff, int len, int enable) +/* + * (static function - no need for kernel doc) + * + * Pass in a buffer containing a glob and this function will + * set search to point to the search part of the buffer and + * return the type of search it is (see enum above). + * This does modify buff. + * + * Returns enum type. + * search returns the pointer to use for comparison. + * not returns 1 if buff started with a '!' + * 0 otherwise. + */ +static int +ftrace_setup_glob(char *buff, int len, char **search, int *not) { - char str[KSYM_SYMBOL_LEN]; - char *search = NULL; - struct ftrace_page *pg; - struct dyn_ftrace *rec; int type = MATCH_FULL; - unsigned long flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; - unsigned i, match = 0, search_len = 0; - int not = 0; + int i; if (buff[0] == '!') { - not = 1; + *not = 1; buff++; len--; - } + } else + *not = 0; + + *search = buff; for (i = 0; i < len; i++) { if (buff[i] == '*') { if (!i) { - search = buff + i + 1; + *search = buff + 1; type = MATCH_END_ONLY; - search_len = len - (i + 1); } else { - if (type == MATCH_END_ONLY) { + if (type == MATCH_END_ONLY) type = MATCH_MIDDLE_ONLY; - } else { - match = i; + else type = MATCH_FRONT_ONLY; - } buff[i] = 0; break; } } } - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); - if (enable) - ftrace_filtered = 1; - pg = ftrace_pages_start; - while (pg) { - for (i = 0; i < pg->index; i++) { - int matched = 0; - char *ptr; + return type; +} + +static int ftrace_match(char *str, char *regex, int len, int type) +{ + int matched = 0; + char *ptr; + + switch (type) { + case MATCH_FULL: + if (strcmp(str, regex) == 0) + matched = 1; + break; + case MATCH_FRONT_ONLY: + if (strncmp(str, regex, len) == 0) + matched = 1; + break; + case MATCH_MIDDLE_ONLY: + if (strstr(str, regex)) + matched = 1; + break; + case MATCH_END_ONLY: + ptr = strstr(str, regex); + if (ptr && (ptr[len] == 0)) + matched = 1; + break; + } + + return matched; +} + +static int +ftrace_match_record(struct dyn_ftrace *rec, char *regex, int len, int type) +{ + char str[KSYM_SYMBOL_LEN]; + + kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); + return ftrace_match(str, regex, len, type); +} + +static void ftrace_match_records(char *buff, int len, int enable) +{ + unsigned int search_len; + struct ftrace_page *pg; + struct dyn_ftrace *rec; + unsigned long flag; + char *search; + int type; + int not; + + flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; + type = ftrace_setup_glob(buff, len, &search, ¬); + + search_len = strlen(search); + + mutex_lock(&ftrace_lock); + do_for_each_ftrace_rec(pg, rec) { + + if (rec->flags & FTRACE_FL_FAILED) + continue; + + if (ftrace_match_record(rec, search, search_len, type)) { + if (not) + rec->flags &= ~flag; + else + rec->flags |= flag; + } + /* + * Only enable filtering if we have a function that + * is filtered on. + */ + if (enable && (rec->flags & FTRACE_FL_FILTER)) + ftrace_filtered = 1; + } while_for_each_ftrace_rec(); + mutex_unlock(&ftrace_lock); +} + +static int +ftrace_match_module_record(struct dyn_ftrace *rec, char *mod, + char *regex, int len, int type) +{ + char str[KSYM_SYMBOL_LEN]; + char *modname; + + kallsyms_lookup(rec->ip, NULL, NULL, &modname, str); + + if (!modname || strcmp(modname, mod)) + return 0; + + /* blank search means to match all funcs in the mod */ + if (len) + return ftrace_match(str, regex, len, type); + else + return 1; +} + +static void ftrace_match_module_records(char *buff, char *mod, int enable) +{ + unsigned search_len = 0; + struct ftrace_page *pg; + struct dyn_ftrace *rec; + int type = MATCH_FULL; + char *search = buff; + unsigned long flag; + int not = 0; + + flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; + + /* blank or '*' mean the same */ + if (strcmp(buff, "*") == 0) + buff[0] = 0; - rec = &pg->records[i]; - if (rec->flags & FTRACE_FL_FAILED) + /* handle the case of 'dont filter this module' */ + if (strcmp(buff, "!") == 0 || strcmp(buff, "!*") == 0) { + buff[0] = 0; + not = 1; + } + + if (strlen(buff)) { + type = ftrace_setup_glob(buff, strlen(buff), &search, ¬); + search_len = strlen(search); + } + + mutex_lock(&ftrace_lock); + do_for_each_ftrace_rec(pg, rec) { + + if (rec->flags & FTRACE_FL_FAILED) + continue; + + if (ftrace_match_module_record(rec, mod, + search, search_len, type)) { + if (not) + rec->flags &= ~flag; + else + rec->flags |= flag; + } + if (enable && (rec->flags & FTRACE_FL_FILTER)) + ftrace_filtered = 1; + + } while_for_each_ftrace_rec(); + mutex_unlock(&ftrace_lock); +} + +/* + * We register the module command as a template to show others how + * to register the a command as well. + */ + +static int +ftrace_mod_callback(char *func, char *cmd, char *param, int enable) +{ + char *mod; + + /* + * cmd == 'mod' because we only registered this func + * for the 'mod' ftrace_func_command. + * But if you register one func with multiple commands, + * you can tell which command was used by the cmd + * parameter. + */ + + /* we must have a module name */ + if (!param) + return -EINVAL; + + mod = strsep(¶m, ":"); + if (!strlen(mod)) + return -EINVAL; + + ftrace_match_module_records(func, mod, enable); + return 0; +} + +static struct ftrace_func_command ftrace_mod_cmd = { + .name = "mod", + .func = ftrace_mod_callback, +}; + +static int __init ftrace_mod_cmd_init(void) +{ + return register_ftrace_command(&ftrace_mod_cmd); +} +device_initcall(ftrace_mod_cmd_init); + +static void +function_trace_probe_call(unsigned long ip, unsigned long parent_ip) +{ + struct ftrace_func_probe *entry; + struct hlist_head *hhd; + struct hlist_node *n; + unsigned long key; + int resched; + + key = hash_long(ip, FTRACE_HASH_BITS); + + hhd = &ftrace_func_hash[key]; + + if (hlist_empty(hhd)) + return; + + /* + * Disable preemption for these calls to prevent a RCU grace + * period. This syncs the hash iteration and freeing of items + * on the hash. rcu_read_lock is too dangerous here. + */ + resched = ftrace_preempt_disable(); + hlist_for_each_entry_rcu(entry, n, hhd, node) { + if (entry->ip == ip) + entry->ops->func(ip, parent_ip, &entry->data); + } + ftrace_preempt_enable(resched); +} + +static struct ftrace_ops trace_probe_ops __read_mostly = +{ + .func = function_trace_probe_call, +}; + +static int ftrace_probe_registered; + +static void __enable_ftrace_function_probe(void) +{ + int i; + + if (ftrace_probe_registered) + return; + + for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) { + struct hlist_head *hhd = &ftrace_func_hash[i]; + if (hhd->first) + break; + } + /* Nothing registered? */ + if (i == FTRACE_FUNC_HASHSIZE) + return; + + __register_ftrace_function(&trace_probe_ops); + ftrace_startup(0); + ftrace_probe_registered = 1; +} + +static void __disable_ftrace_function_probe(void) +{ + int i; + + if (!ftrace_probe_registered) + return; + + for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) { + struct hlist_head *hhd = &ftrace_func_hash[i]; + if (hhd->first) + return; + } + + /* no more funcs left */ + __unregister_ftrace_function(&trace_probe_ops); + ftrace_shutdown(0); + ftrace_probe_registered = 0; +} + + +static void ftrace_free_entry_rcu(struct rcu_head *rhp) +{ + struct ftrace_func_probe *entry = + container_of(rhp, struct ftrace_func_probe, rcu); + + if (entry->ops->free) + entry->ops->free(&entry->data); + kfree(entry); +} + + +int +register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops, + void *data) +{ + struct ftrace_func_probe *entry; + struct ftrace_page *pg; + struct dyn_ftrace *rec; + int type, len, not; + unsigned long key; + int count = 0; + char *search; + + type = ftrace_setup_glob(glob, strlen(glob), &search, ¬); + len = strlen(search); + + /* we do not support '!' for function probes */ + if (WARN_ON(not)) + return -EINVAL; + + mutex_lock(&ftrace_lock); + do_for_each_ftrace_rec(pg, rec) { + + if (rec->flags & FTRACE_FL_FAILED) + continue; + + if (!ftrace_match_record(rec, search, len, type)) + continue; + + entry = kmalloc(sizeof(*entry), GFP_KERNEL); + if (!entry) { + /* If we did not process any, then return error */ + if (!count) + count = -ENOMEM; + goto out_unlock; + } + + count++; + + entry->data = data; + + /* + * The caller might want to do something special + * for each function we find. We call the callback + * to give the caller an opportunity to do so. + */ + if (ops->callback) { + if (ops->callback(rec->ip, &entry->data) < 0) { + /* caller does not like this func */ + kfree(entry); continue; - kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); - switch (type) { - case MATCH_FULL: - if (strcmp(str, buff) == 0) - matched = 1; - break; - case MATCH_FRONT_ONLY: - if (memcmp(str, buff, match) == 0) - matched = 1; - break; - case MATCH_MIDDLE_ONLY: - if (strstr(str, search)) - matched = 1; - break; - case MATCH_END_ONLY: - ptr = strstr(str, search); - if (ptr && (ptr[search_len] == 0)) - matched = 1; - break; } - if (matched) { - if (not) - rec->flags &= ~flag; - else - rec->flags |= flag; + } + + entry->ops = ops; + entry->ip = rec->ip; + + key = hash_long(entry->ip, FTRACE_HASH_BITS); + hlist_add_head_rcu(&entry->node, &ftrace_func_hash[key]); + + } while_for_each_ftrace_rec(); + __enable_ftrace_function_probe(); + + out_unlock: + mutex_unlock(&ftrace_lock); + + return count; +} + +enum { + PROBE_TEST_FUNC = 1, + PROBE_TEST_DATA = 2 +}; + +static void +__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops, + void *data, int flags) +{ + struct ftrace_func_probe *entry; + struct hlist_node *n, *tmp; + char str[KSYM_SYMBOL_LEN]; + int type = MATCH_FULL; + int i, len = 0; + char *search; + + if (glob && (strcmp(glob, "*") || !strlen(glob))) + glob = NULL; + else { + int not; + + type = ftrace_setup_glob(glob, strlen(glob), &search, ¬); + len = strlen(search); + + /* we do not support '!' for function probes */ + if (WARN_ON(not)) + return; + } + + mutex_lock(&ftrace_lock); + for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) { + struct hlist_head *hhd = &ftrace_func_hash[i]; + + hlist_for_each_entry_safe(entry, n, tmp, hhd, node) { + + /* break up if statements for readability */ + if ((flags & PROBE_TEST_FUNC) && entry->ops != ops) + continue; + + if ((flags & PROBE_TEST_DATA) && entry->data != data) + continue; + + /* do this last, since it is the most expensive */ + if (glob) { + kallsyms_lookup(entry->ip, NULL, NULL, + NULL, str); + if (!ftrace_match(str, glob, len, type)) + continue; } + + hlist_del(&entry->node); + call_rcu(&entry->rcu, ftrace_free_entry_rcu); + } + } + __disable_ftrace_function_probe(); + mutex_unlock(&ftrace_lock); +} + +void +unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops, + void *data) +{ + __unregister_ftrace_function_probe(glob, ops, data, + PROBE_TEST_FUNC | PROBE_TEST_DATA); +} + +void +unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops) +{ + __unregister_ftrace_function_probe(glob, ops, NULL, PROBE_TEST_FUNC); +} + +void unregister_ftrace_function_probe_all(char *glob) +{ + __unregister_ftrace_function_probe(glob, NULL, NULL, 0); +} + +static LIST_HEAD(ftrace_commands); +static DEFINE_MUTEX(ftrace_cmd_mutex); + +int register_ftrace_command(struct ftrace_func_command *cmd) +{ + struct ftrace_func_command *p; + int ret = 0; + + mutex_lock(&ftrace_cmd_mutex); + list_for_each_entry(p, &ftrace_commands, list) { + if (strcmp(cmd->name, p->name) == 0) { + ret = -EBUSY; + goto out_unlock; + } + } + list_add(&cmd->list, &ftrace_commands); + out_unlock: + mutex_unlock(&ftrace_cmd_mutex); + + return ret; +} + +int unregister_ftrace_command(struct ftrace_func_command *cmd) +{ + struct ftrace_func_command *p, *n; + int ret = -ENODEV; + + mutex_lock(&ftrace_cmd_mutex); + list_for_each_entry_safe(p, n, &ftrace_commands, list) { + if (strcmp(cmd->name, p->name) == 0) { + ret = 0; + list_del_init(&p->list); + goto out_unlock; + } + } + out_unlock: + mutex_unlock(&ftrace_cmd_mutex); + + return ret; +} + +static int ftrace_process_regex(char *buff, int len, int enable) +{ + char *func, *command, *next = buff; + struct ftrace_func_command *p; + int ret = -EINVAL; + + func = strsep(&next, ":"); + + if (!next) { + ftrace_match_records(func, len, enable); + return 0; + } + + /* command found */ + + command = strsep(&next, ":"); + + mutex_lock(&ftrace_cmd_mutex); + list_for_each_entry(p, &ftrace_commands, list) { + if (strcmp(p->name, command) == 0) { + ret = p->func(func, command, next, enable); + goto out_unlock; } - pg = pg->next; } - spin_unlock(&ftrace_lock); + out_unlock: + mutex_unlock(&ftrace_cmd_mutex); + + return ret; } static ssize_t @@ -1150,7 +2222,11 @@ ftrace_regex_write(struct file *file, const char __user *ubuf, read++; cnt--; - if (!(iter->flags & ~FTRACE_ITER_CONT)) { + /* + * If the parser haven't finished with the last write, + * continue reading the user input without skipping spaces. + */ + if (!(iter->flags & FTRACE_ITER_CONT)) { /* skip white space */ while (cnt && isspace(ch)) { ret = get_user(ch, ubuf++); @@ -1160,8 +2236,9 @@ ftrace_regex_write(struct file *file, const char __user *ubuf, cnt--; } + /* only spaces were written */ if (isspace(ch)) { - file->f_pos += read; + *ppos += read; ret = read; goto out; } @@ -1184,16 +2261,18 @@ ftrace_regex_write(struct file *file, const char __user *ubuf, } if (isspace(ch)) { - iter->filtered++; iter->buffer[iter->buffer_idx] = 0; - ftrace_match(iter->buffer, iter->buffer_idx, enable); + ret = ftrace_process_regex(iter->buffer, + iter->buffer_idx, enable); + if (ret) + goto out; iter->buffer_idx = 0; - } else + } else { iter->flags |= FTRACE_ITER_CONT; + iter->buffer[iter->buffer_idx++] = ch; + } - - file->f_pos += read; - + *ppos += read; ret = read; out: mutex_unlock(&ftrace_regex_lock); @@ -1225,7 +2304,7 @@ ftrace_set_regex(unsigned char *buf, int len, int reset, int enable) if (reset) ftrace_filter_reset(enable); if (buf) - ftrace_match(buf, len, enable); + ftrace_match_records(buf, len, enable); mutex_unlock(&ftrace_regex_lock); } @@ -1258,6 +2337,45 @@ void ftrace_set_notrace(unsigned char *buf, int len, int reset) ftrace_set_regex(buf, len, reset, 0); } +/* + * command line interface to allow users to set filters on boot up. + */ +#define FTRACE_FILTER_SIZE COMMAND_LINE_SIZE +static char ftrace_notrace_buf[FTRACE_FILTER_SIZE] __initdata; +static char ftrace_filter_buf[FTRACE_FILTER_SIZE] __initdata; + +static int __init set_ftrace_notrace(char *str) +{ + strncpy(ftrace_notrace_buf, str, FTRACE_FILTER_SIZE); + return 1; +} +__setup("ftrace_notrace=", set_ftrace_notrace); + +static int __init set_ftrace_filter(char *str) +{ + strncpy(ftrace_filter_buf, str, FTRACE_FILTER_SIZE); + return 1; +} +__setup("ftrace_filter=", set_ftrace_filter); + +static void __init set_ftrace_early_filter(char *buf, int enable) +{ + char *func; + + while (buf) { + func = strsep(&buf, ","); + ftrace_set_regex(func, strlen(func), 0, enable); + } +} + +static void __init set_ftrace_early_filters(void) +{ + if (ftrace_filter_buf[0]) + set_ftrace_early_filter(ftrace_filter_buf, 1); + if (ftrace_notrace_buf[0]) + set_ftrace_early_filter(ftrace_notrace_buf, 0); +} + static int ftrace_regex_release(struct inode *inode, struct file *file, int enable) { @@ -1273,17 +2391,14 @@ ftrace_regex_release(struct inode *inode, struct file *file, int enable) iter = file->private_data; if (iter->buffer_idx) { - iter->filtered++; iter->buffer[iter->buffer_idx] = 0; - ftrace_match(iter->buffer, iter->buffer_idx, enable); + ftrace_match_records(iter->buffer, iter->buffer_idx, enable); } - mutex_lock(&ftrace_sysctl_lock); - mutex_lock(&ftrace_start_lock); + mutex_lock(&ftrace_lock); if (ftrace_start_up && ftrace_enabled) ftrace_run_update_code(FTRACE_ENABLE_CALLS); - mutex_unlock(&ftrace_start_lock); - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); kfree(iter); mutex_unlock(&ftrace_regex_lock); @@ -1302,31 +2417,31 @@ ftrace_notrace_release(struct inode *inode, struct file *file) return ftrace_regex_release(inode, file, 0); } -static struct file_operations ftrace_avail_fops = { +static const struct file_operations ftrace_avail_fops = { .open = ftrace_avail_open, .read = seq_read, .llseek = seq_lseek, - .release = ftrace_avail_release, + .release = seq_release_private, }; -static struct file_operations ftrace_failures_fops = { +static const struct file_operations ftrace_failures_fops = { .open = ftrace_failures_open, .read = seq_read, .llseek = seq_lseek, - .release = ftrace_avail_release, + .release = seq_release_private, }; -static struct file_operations ftrace_filter_fops = { +static const struct file_operations ftrace_filter_fops = { .open = ftrace_filter_open, - .read = ftrace_regex_read, + .read = seq_read, .write = ftrace_filter_write, .llseek = ftrace_regex_lseek, .release = ftrace_filter_release, }; -static struct file_operations ftrace_notrace_fops = { +static const struct file_operations ftrace_notrace_fops = { .open = ftrace_notrace_open, - .read = ftrace_regex_read, + .read = seq_read, .write = ftrace_notrace_write, .llseek = ftrace_regex_lseek, .release = ftrace_notrace_release, @@ -1340,28 +2455,31 @@ int ftrace_graph_count; unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly; static void * -g_next(struct seq_file *m, void *v, loff_t *pos) +__g_next(struct seq_file *m, loff_t *pos) { unsigned long *array = m->private; - int index = *pos; - - (*pos)++; - if (index >= ftrace_graph_count) + if (*pos >= ftrace_graph_count) return NULL; + return &array[*pos]; +} - return &array[index]; +static void * +g_next(struct seq_file *m, void *v, loff_t *pos) +{ + (*pos)++; + return __g_next(m, pos); } static void *g_start(struct seq_file *m, loff_t *pos) { - void *p = NULL; - mutex_lock(&graph_lock); - p = g_next(m, p, pos); + /* Nothing, tell g_show to print all functions are enabled */ + if (!ftrace_graph_count && !*pos) + return (void *)1; - return p; + return __g_next(m, pos); } static void g_stop(struct seq_file *m, void *p) @@ -1372,14 +2490,16 @@ static void g_stop(struct seq_file *m, void *p) static int g_show(struct seq_file *m, void *v) { unsigned long *ptr = v; - char str[KSYM_SYMBOL_LEN]; if (!ptr) return 0; - kallsyms_lookup(*ptr, NULL, NULL, NULL, str); + if (ptr == (unsigned long *)1) { + seq_printf(m, "#### all functions enabled ####\n"); + return 0; + } - seq_printf(m, "%s\n", str); + seq_printf(m, "%pf\n", v); return 0; } @@ -1401,7 +2521,7 @@ ftrace_graph_open(struct inode *inode, struct file *file) mutex_lock(&graph_lock); if ((file->f_mode & FMODE_WRITE) && - !(file->f_flags & O_APPEND)) { + (file->f_flags & O_TRUNC)) { ftrace_graph_count = 0; memset(ftrace_graph_funcs, 0, sizeof(ftrace_graph_funcs)); } @@ -1419,53 +2539,61 @@ ftrace_graph_open(struct inode *inode, struct file *file) return ret; } -static ssize_t -ftrace_graph_read(struct file *file, char __user *ubuf, - size_t cnt, loff_t *ppos) +static int +ftrace_graph_release(struct inode *inode, struct file *file) { if (file->f_mode & FMODE_READ) - return seq_read(file, ubuf, cnt, ppos); - else - return -EPERM; + seq_release(inode, file); + return 0; } static int -ftrace_set_func(unsigned long *array, int idx, char *buffer) +ftrace_set_func(unsigned long *array, int *idx, char *buffer) { - char str[KSYM_SYMBOL_LEN]; struct dyn_ftrace *rec; struct ftrace_page *pg; + int search_len; int found = 0; - int i, j; + int type, not; + char *search; + bool exists; + int i; if (ftrace_disabled) return -ENODEV; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + /* decode regex */ + type = ftrace_setup_glob(buffer, strlen(buffer), &search, ¬); + if (not) + return -EINVAL; - for (pg = ftrace_pages_start; pg; pg = pg->next) { - for (i = 0; i < pg->index; i++) { - rec = &pg->records[i]; + search_len = strlen(search); - if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) - continue; + mutex_lock(&ftrace_lock); + do_for_each_ftrace_rec(pg, rec) { + + if (*idx >= FTRACE_GRAPH_MAX_FUNCS) + break; + + if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) + continue; - kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); - if (strcmp(str, buffer) == 0) { + if (ftrace_match_record(rec, search, search_len, type)) { + /* ensure it is not already in the array */ + exists = false; + for (i = 0; i < *idx; i++) + if (array[i] == rec->ip) { + exists = true; + break; + } + if (!exists) { + array[(*idx)++] = rec->ip; found = 1; - for (j = 0; j < idx; j++) - if (array[j] == rec->ip) { - found = 0; - break; - } - if (found) - array[idx] = rec->ip; - break; } } - } - spin_unlock(&ftrace_lock); + } while_for_each_ftrace_rec(); + + mutex_unlock(&ftrace_lock); return found ? 0 : -EINVAL; } @@ -1533,13 +2661,11 @@ ftrace_graph_write(struct file *file, const char __user *ubuf, } buffer[index] = 0; - /* we allow only one at a time */ - ret = ftrace_set_func(array, ftrace_graph_count, buffer); + /* we allow only one expression at a time */ + ret = ftrace_set_func(array, &ftrace_graph_count, buffer); if (ret) goto out; - ftrace_graph_count++; - file->f_pos += read; ret = read; @@ -1550,46 +2676,32 @@ ftrace_graph_write(struct file *file, const char __user *ubuf, } static const struct file_operations ftrace_graph_fops = { - .open = ftrace_graph_open, - .read = ftrace_graph_read, - .write = ftrace_graph_write, + .open = ftrace_graph_open, + .read = seq_read, + .write = ftrace_graph_write, + .release = ftrace_graph_release, }; #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ static __init int ftrace_init_dyn_debugfs(struct dentry *d_tracer) { - struct dentry *entry; - entry = debugfs_create_file("available_filter_functions", 0444, - d_tracer, NULL, &ftrace_avail_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'available_filter_functions' entry\n"); + trace_create_file("available_filter_functions", 0444, + d_tracer, NULL, &ftrace_avail_fops); - entry = debugfs_create_file("failures", 0444, - d_tracer, NULL, &ftrace_failures_fops); - if (!entry) - pr_warning("Could not create debugfs 'failures' entry\n"); + trace_create_file("failures", 0444, + d_tracer, NULL, &ftrace_failures_fops); - entry = debugfs_create_file("set_ftrace_filter", 0644, d_tracer, - NULL, &ftrace_filter_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'set_ftrace_filter' entry\n"); + trace_create_file("set_ftrace_filter", 0644, d_tracer, + NULL, &ftrace_filter_fops); - entry = debugfs_create_file("set_ftrace_notrace", 0644, d_tracer, + trace_create_file("set_ftrace_notrace", 0644, d_tracer, NULL, &ftrace_notrace_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'set_ftrace_notrace' entry\n"); #ifdef CONFIG_FUNCTION_GRAPH_TRACER - entry = debugfs_create_file("set_graph_function", 0444, d_tracer, + trace_create_file("set_graph_function", 0444, d_tracer, NULL, &ftrace_graph_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'set_graph_function' entry\n"); #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ return 0; @@ -1603,7 +2715,7 @@ static int ftrace_convert_nops(struct module *mod, unsigned long addr; unsigned long flags; - mutex_lock(&ftrace_start_lock); + mutex_lock(&ftrace_lock); p = start; while (p < end) { addr = ftrace_call_adjust(*p++); @@ -1622,19 +2734,77 @@ static int ftrace_convert_nops(struct module *mod, local_irq_save(flags); ftrace_update_code(mod); local_irq_restore(flags); - mutex_unlock(&ftrace_start_lock); + mutex_unlock(&ftrace_lock); return 0; } -void ftrace_init_module(struct module *mod, - unsigned long *start, unsigned long *end) +#ifdef CONFIG_MODULES +void ftrace_release(void *start, void *end) +{ + struct dyn_ftrace *rec; + struct ftrace_page *pg; + unsigned long s = (unsigned long)start; + unsigned long e = (unsigned long)end; + + if (ftrace_disabled || !start || start == end) + return; + + mutex_lock(&ftrace_lock); + do_for_each_ftrace_rec(pg, rec) { + if ((rec->ip >= s) && (rec->ip < e)) { + /* + * rec->ip is changed in ftrace_free_rec() + * It should not between s and e if record was freed. + */ + FTRACE_WARN_ON(rec->flags & FTRACE_FL_FREE); + ftrace_free_rec(rec); + } + } while_for_each_ftrace_rec(); + mutex_unlock(&ftrace_lock); +} + +static void ftrace_init_module(struct module *mod, + unsigned long *start, unsigned long *end) { if (ftrace_disabled || start == end) return; ftrace_convert_nops(mod, start, end); } +static int ftrace_module_notify(struct notifier_block *self, + unsigned long val, void *data) +{ + struct module *mod = data; + + switch (val) { + case MODULE_STATE_COMING: + ftrace_init_module(mod, mod->ftrace_callsites, + mod->ftrace_callsites + + mod->num_ftrace_callsites); + break; + case MODULE_STATE_GOING: + ftrace_release(mod->ftrace_callsites, + mod->ftrace_callsites + + mod->num_ftrace_callsites); + break; + } + + return 0; +} +#else +static int ftrace_module_notify(struct notifier_block *self, + unsigned long val, void *data) +{ + return 0; +} +#endif /* CONFIG_MODULES */ + +struct notifier_block ftrace_module_nb = { + .notifier_call = ftrace_module_notify, + .priority = 0, +}; + extern unsigned long __start_mcount_loc[]; extern unsigned long __stop_mcount_loc[]; @@ -1666,6 +2836,12 @@ void __init ftrace_init(void) __start_mcount_loc, __stop_mcount_loc); + ret = register_module_notifier(&ftrace_module_nb); + if (ret) + pr_warning("Failed to register trace ftrace module notifier\n"); + + set_ftrace_early_filters(); + return; failed: ftrace_disabled = 1; @@ -1699,7 +2875,7 @@ ftrace_pid_read(struct file *file, char __user *ubuf, if (ftrace_pid_trace == ftrace_swapper_pid) r = sprintf(buf, "swapper tasks\n"); else if (ftrace_pid_trace) - r = sprintf(buf, "%u\n", pid_nr(ftrace_pid_trace)); + r = sprintf(buf, "%u\n", pid_vnr(ftrace_pid_trace)); else r = sprintf(buf, "no pid\n"); @@ -1736,9 +2912,12 @@ static void clear_ftrace_pid(struct pid *pid) { struct task_struct *p; + rcu_read_lock(); do_each_pid_task(pid, PIDTYPE_PID, p) { clear_tsk_trace_trace(p); } while_each_pid_task(pid, PIDTYPE_PID, p); + rcu_read_unlock(); + put_pid(pid); } @@ -1746,9 +2925,11 @@ static void set_ftrace_pid(struct pid *pid) { struct task_struct *p; + rcu_read_lock(); do_each_pid_task(pid, PIDTYPE_PID, p) { set_tsk_trace_trace(p); } while_each_pid_task(pid, PIDTYPE_PID, p); + rcu_read_unlock(); } static void clear_ftrace_pid_task(struct pid **pid) @@ -1790,7 +2971,7 @@ ftrace_pid_write(struct file *filp, const char __user *ubuf, if (ret < 0) return ret; - mutex_lock(&ftrace_start_lock); + mutex_lock(&ftrace_lock); if (val < 0) { /* disable pid tracing */ if (!ftrace_pid_trace) @@ -1829,12 +3010,12 @@ ftrace_pid_write(struct file *filp, const char __user *ubuf, ftrace_startup_enable(0); out: - mutex_unlock(&ftrace_start_lock); + mutex_unlock(&ftrace_lock); return cnt; } -static struct file_operations ftrace_pid_fops = { +static const struct file_operations ftrace_pid_fops = { .read = ftrace_pid_read, .write = ftrace_pid_write, }; @@ -1842,7 +3023,6 @@ static struct file_operations ftrace_pid_fops = { static __init int ftrace_init_debugfs(void) { struct dentry *d_tracer; - struct dentry *entry; d_tracer = tracing_init_dentry(); if (!d_tracer) @@ -1850,14 +3030,13 @@ static __init int ftrace_init_debugfs(void) ftrace_init_dyn_debugfs(d_tracer); - entry = debugfs_create_file("set_ftrace_pid", 0644, d_tracer, - NULL, &ftrace_pid_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'set_ftrace_pid' entry\n"); + trace_create_file("set_ftrace_pid", 0644, d_tracer, + NULL, &ftrace_pid_fops); + + ftrace_profile_debugfs(d_tracer); + return 0; } - fs_initcall(ftrace_init_debugfs); /** @@ -1892,17 +3071,17 @@ int register_ftrace_function(struct ftrace_ops *ops) if (unlikely(ftrace_disabled)) return -1; - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); ret = __register_ftrace_function(ops); ftrace_startup(0); - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); return ret; } /** - * unregister_ftrace_function - unresgister a function for profiling. + * unregister_ftrace_function - unregister a function for profiling. * @ops - ops structure that holds the function to unregister * * Unregister a function that was added to be called by ftrace profiling. @@ -1911,10 +3090,10 @@ int unregister_ftrace_function(struct ftrace_ops *ops) { int ret; - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); ret = __unregister_ftrace_function(ops); ftrace_shutdown(0); - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); return ret; } @@ -1929,14 +3108,14 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, if (unlikely(ftrace_disabled)) return -ENODEV; - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); ret = proc_dointvec(table, write, file, buffer, lenp, ppos); - if (ret || !write || (last_ftrace_enabled == ftrace_enabled)) + if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled)) goto out; - last_ftrace_enabled = ftrace_enabled; + last_ftrace_enabled = !!ftrace_enabled; if (ftrace_enabled) { @@ -1958,13 +3137,14 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, } out: - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); return ret; } #ifdef CONFIG_FUNCTION_GRAPH_TRACER -static atomic_t ftrace_graph_active; +static int ftrace_graph_active; +static struct notifier_block ftrace_suspend_notifier; int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace) { @@ -2005,12 +3185,12 @@ static int alloc_retstack_tasklist(struct ftrace_ret_stack **ret_stack_list) } if (t->ret_stack == NULL) { - t->curr_ret_stack = -1; - /* Make sure IRQs see the -1 first: */ - barrier(); - t->ret_stack = ret_stack_list[start++]; atomic_set(&t->tracing_graph_pause, 0); atomic_set(&t->trace_overrun, 0); + t->curr_ret_stack = -1; + /* Make sure the tasks see the -1 first: */ + smp_wmb(); + t->ret_stack = ret_stack_list[start++]; } } while_each_thread(g, t); @@ -2022,11 +3202,43 @@ free: return ret; } +static void +ftrace_graph_probe_sched_switch(struct rq *__rq, struct task_struct *prev, + struct task_struct *next) +{ + unsigned long long timestamp; + int index; + + /* + * Does the user want to count the time a function was asleep. + * If so, do not update the time stamps. + */ + if (trace_flags & TRACE_ITER_SLEEP_TIME) + return; + + timestamp = trace_clock_local(); + + prev->ftrace_timestamp = timestamp; + + /* only process tasks that we timestamped */ + if (!next->ftrace_timestamp) + return; + + /* + * Update all the counters in next to make up for the + * time next was sleeping. + */ + timestamp -= next->ftrace_timestamp; + + for (index = next->curr_ret_stack; index >= 0; index--) + next->ret_stack[index].calltime += timestamp; +} + /* Allocate a return stack for each task */ static int start_graph_tracing(void) { struct ftrace_ret_stack **ret_stack_list; - int ret; + int ret, cpu; ret_stack_list = kmalloc(FTRACE_RETSTACK_ALLOC_SIZE * sizeof(struct ftrace_ret_stack *), @@ -2035,25 +3247,68 @@ static int start_graph_tracing(void) if (!ret_stack_list) return -ENOMEM; + /* The cpu_boot init_task->ret_stack will never be freed */ + for_each_online_cpu(cpu) { + if (!idle_task(cpu)->ret_stack) + ftrace_graph_init_task(idle_task(cpu)); + } + do { ret = alloc_retstack_tasklist(ret_stack_list); } while (ret == -EAGAIN); + if (!ret) { + ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch); + if (ret) + pr_info("ftrace_graph: Couldn't activate tracepoint" + " probe to kernel_sched_switch\n"); + } + kfree(ret_stack_list); return ret; } +/* + * Hibernation protection. + * The state of the current task is too much unstable during + * suspend/restore to disk. We want to protect against that. + */ +static int +ftrace_suspend_notifier_call(struct notifier_block *bl, unsigned long state, + void *unused) +{ + switch (state) { + case PM_HIBERNATION_PREPARE: + pause_graph_tracing(); + break; + + case PM_POST_HIBERNATION: + unpause_graph_tracing(); + break; + } + return NOTIFY_DONE; +} + int register_ftrace_graph(trace_func_graph_ret_t retfunc, trace_func_graph_ent_t entryfunc) { int ret = 0; - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); - atomic_inc(&ftrace_graph_active); + /* we currently allow only one tracer registered at a time */ + if (ftrace_graph_active) { + ret = -EBUSY; + goto out; + } + + ftrace_suspend_notifier.notifier_call = ftrace_suspend_notifier_call; + register_pm_notifier(&ftrace_suspend_notifier); + + ftrace_graph_active++; ret = start_graph_tracing(); if (ret) { - atomic_dec(&ftrace_graph_active); + ftrace_graph_active--; goto out; } @@ -2063,36 +3318,50 @@ int register_ftrace_graph(trace_func_graph_ret_t retfunc, ftrace_startup(FTRACE_START_FUNC_RET); out: - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); return ret; } void unregister_ftrace_graph(void) { - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); + + if (unlikely(!ftrace_graph_active)) + goto out; - atomic_dec(&ftrace_graph_active); + ftrace_graph_active--; + unregister_trace_sched_switch(ftrace_graph_probe_sched_switch); ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; ftrace_graph_entry = ftrace_graph_entry_stub; ftrace_shutdown(FTRACE_STOP_FUNC_RET); + unregister_pm_notifier(&ftrace_suspend_notifier); - mutex_unlock(&ftrace_sysctl_lock); + out: + mutex_unlock(&ftrace_lock); } /* Allocate a return stack for newly created task */ void ftrace_graph_init_task(struct task_struct *t) { - if (atomic_read(&ftrace_graph_active)) { - t->ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH + /* Make sure we do not use the parent ret_stack */ + t->ret_stack = NULL; + + if (ftrace_graph_active) { + struct ftrace_ret_stack *ret_stack; + + ret_stack = kmalloc(FTRACE_RETFUNC_DEPTH * sizeof(struct ftrace_ret_stack), GFP_KERNEL); - if (!t->ret_stack) + if (!ret_stack) return; t->curr_ret_stack = -1; atomic_set(&t->tracing_graph_pause, 0); atomic_set(&t->trace_overrun, 0); - } else - t->ret_stack = NULL; + t->ftrace_timestamp = 0; + /* make curr_ret_stack visable before we add the ret_stack */ + smp_wmb(); + t->ret_stack = ret_stack; + } } void ftrace_graph_exit_task(struct task_struct *t) diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c new file mode 100644 index 00000000000..81b1645c854 --- /dev/null +++ b/kernel/trace/kmemtrace.c @@ -0,0 +1,511 @@ +/* + * Memory allocator tracing + * + * Copyright (C) 2008 Eduard - Gabriel Munteanu + * Copyright (C) 2008 Pekka Enberg <penberg@cs.helsinki.fi> + * Copyright (C) 2008 Frederic Weisbecker <fweisbec@gmail.com> + */ + +#include <linux/tracepoint.h> +#include <linux/seq_file.h> +#include <linux/debugfs.h> +#include <linux/dcache.h> +#include <linux/fs.h> + +#include <linux/kmemtrace.h> + +#include "trace_output.h" +#include "trace.h" + +/* Select an alternative, minimalistic output than the original one */ +#define TRACE_KMEM_OPT_MINIMAL 0x1 + +static struct tracer_opt kmem_opts[] = { + /* Default disable the minimalistic output */ + { TRACER_OPT(kmem_minimalistic, TRACE_KMEM_OPT_MINIMAL) }, + { } +}; + +static struct tracer_flags kmem_tracer_flags = { + .val = 0, + .opts = kmem_opts +}; + +static struct trace_array *kmemtrace_array; + +/* Trace allocations */ +static inline void kmemtrace_alloc(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags, + int node) +{ + struct ftrace_event_call *call = &event_kmem_alloc; + struct trace_array *tr = kmemtrace_array; + struct kmemtrace_alloc_entry *entry; + struct ring_buffer_event *event; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + if (!event) + return; + + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, 0, 0); + + entry->ent.type = TRACE_KMEM_ALLOC; + entry->type_id = type_id; + entry->call_site = call_site; + entry->ptr = ptr; + entry->bytes_req = bytes_req; + entry->bytes_alloc = bytes_alloc; + entry->gfp_flags = gfp_flags; + entry->node = node; + + if (!filter_check_discard(call, entry, tr->buffer, event)) + ring_buffer_unlock_commit(tr->buffer, event); + + trace_wake_up(); +} + +static inline void kmemtrace_free(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr) +{ + struct ftrace_event_call *call = &event_kmem_free; + struct trace_array *tr = kmemtrace_array; + struct kmemtrace_free_entry *entry; + struct ring_buffer_event *event; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + if (!event) + return; + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, 0, 0); + + entry->ent.type = TRACE_KMEM_FREE; + entry->type_id = type_id; + entry->call_site = call_site; + entry->ptr = ptr; + + if (!filter_check_discard(call, entry, tr->buffer, event)) + ring_buffer_unlock_commit(tr->buffer, event); + + trace_wake_up(); +} + +static void kmemtrace_kmalloc(unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags) +{ + kmemtrace_alloc(KMEMTRACE_TYPE_KMALLOC, call_site, ptr, + bytes_req, bytes_alloc, gfp_flags, -1); +} + +static void kmemtrace_kmem_cache_alloc(unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags) +{ + kmemtrace_alloc(KMEMTRACE_TYPE_CACHE, call_site, ptr, + bytes_req, bytes_alloc, gfp_flags, -1); +} + +static void kmemtrace_kmalloc_node(unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags, + int node) +{ + kmemtrace_alloc(KMEMTRACE_TYPE_KMALLOC, call_site, ptr, + bytes_req, bytes_alloc, gfp_flags, node); +} + +static void kmemtrace_kmem_cache_alloc_node(unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags, + int node) +{ + kmemtrace_alloc(KMEMTRACE_TYPE_CACHE, call_site, ptr, + bytes_req, bytes_alloc, gfp_flags, node); +} + +static void kmemtrace_kfree(unsigned long call_site, const void *ptr) +{ + kmemtrace_free(KMEMTRACE_TYPE_KMALLOC, call_site, ptr); +} + +static void kmemtrace_kmem_cache_free(unsigned long call_site, const void *ptr) +{ + kmemtrace_free(KMEMTRACE_TYPE_CACHE, call_site, ptr); +} + +static int kmemtrace_start_probes(void) +{ + int err; + + err = register_trace_kmalloc(kmemtrace_kmalloc); + if (err) + return err; + err = register_trace_kmem_cache_alloc(kmemtrace_kmem_cache_alloc); + if (err) + return err; + err = register_trace_kmalloc_node(kmemtrace_kmalloc_node); + if (err) + return err; + err = register_trace_kmem_cache_alloc_node(kmemtrace_kmem_cache_alloc_node); + if (err) + return err; + err = register_trace_kfree(kmemtrace_kfree); + if (err) + return err; + err = register_trace_kmem_cache_free(kmemtrace_kmem_cache_free); + + return err; +} + +static void kmemtrace_stop_probes(void) +{ + unregister_trace_kmalloc(kmemtrace_kmalloc); + unregister_trace_kmem_cache_alloc(kmemtrace_kmem_cache_alloc); + unregister_trace_kmalloc_node(kmemtrace_kmalloc_node); + unregister_trace_kmem_cache_alloc_node(kmemtrace_kmem_cache_alloc_node); + unregister_trace_kfree(kmemtrace_kfree); + unregister_trace_kmem_cache_free(kmemtrace_kmem_cache_free); +} + +static int kmem_trace_init(struct trace_array *tr) +{ + kmemtrace_array = tr; + + tracing_reset_online_cpus(tr); + + kmemtrace_start_probes(); + + return 0; +} + +static void kmem_trace_reset(struct trace_array *tr) +{ + kmemtrace_stop_probes(); +} + +static void kmemtrace_headers(struct seq_file *s) +{ + /* Don't need headers for the original kmemtrace output */ + if (!(kmem_tracer_flags.val & TRACE_KMEM_OPT_MINIMAL)) + return; + + seq_printf(s, "#\n"); + seq_printf(s, "# ALLOC TYPE REQ GIVEN FLAGS " + " POINTER NODE CALLER\n"); + seq_printf(s, "# FREE | | | | " + " | | | |\n"); + seq_printf(s, "# |\n\n"); +} + +/* + * The following functions give the original output from kmemtrace, + * plus the origin CPU, since reordering occurs in-kernel now. + */ + +#define KMEMTRACE_USER_ALLOC 0 +#define KMEMTRACE_USER_FREE 1 + +struct kmemtrace_user_event { + u8 event_id; + u8 type_id; + u16 event_size; + u32 cpu; + u64 timestamp; + unsigned long call_site; + unsigned long ptr; +}; + +struct kmemtrace_user_event_alloc { + size_t bytes_req; + size_t bytes_alloc; + unsigned gfp_flags; + int node; +}; + +static enum print_line_t +kmemtrace_print_alloc(struct trace_iterator *iter, int flags) +{ + struct trace_seq *s = &iter->seq; + struct kmemtrace_alloc_entry *entry; + int ret; + + trace_assign_type(entry, iter->ent); + + ret = trace_seq_printf(s, "type_id %d call_site %pF ptr %lu " + "bytes_req %lu bytes_alloc %lu gfp_flags %lu node %d\n", + entry->type_id, (void *)entry->call_site, (unsigned long)entry->ptr, + (unsigned long)entry->bytes_req, (unsigned long)entry->bytes_alloc, + (unsigned long)entry->gfp_flags, entry->node); + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t +kmemtrace_print_free(struct trace_iterator *iter, int flags) +{ + struct trace_seq *s = &iter->seq; + struct kmemtrace_free_entry *entry; + int ret; + + trace_assign_type(entry, iter->ent); + + ret = trace_seq_printf(s, "type_id %d call_site %pF ptr %lu\n", + entry->type_id, (void *)entry->call_site, + (unsigned long)entry->ptr); + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t +kmemtrace_print_alloc_user(struct trace_iterator *iter, int flags) +{ + struct trace_seq *s = &iter->seq; + struct kmemtrace_alloc_entry *entry; + struct kmemtrace_user_event *ev; + struct kmemtrace_user_event_alloc *ev_alloc; + + trace_assign_type(entry, iter->ent); + + ev = trace_seq_reserve(s, sizeof(*ev)); + if (!ev) + return TRACE_TYPE_PARTIAL_LINE; + + ev->event_id = KMEMTRACE_USER_ALLOC; + ev->type_id = entry->type_id; + ev->event_size = sizeof(*ev) + sizeof(*ev_alloc); + ev->cpu = iter->cpu; + ev->timestamp = iter->ts; + ev->call_site = entry->call_site; + ev->ptr = (unsigned long)entry->ptr; + + ev_alloc = trace_seq_reserve(s, sizeof(*ev_alloc)); + if (!ev_alloc) + return TRACE_TYPE_PARTIAL_LINE; + + ev_alloc->bytes_req = entry->bytes_req; + ev_alloc->bytes_alloc = entry->bytes_alloc; + ev_alloc->gfp_flags = entry->gfp_flags; + ev_alloc->node = entry->node; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t +kmemtrace_print_free_user(struct trace_iterator *iter, int flags) +{ + struct trace_seq *s = &iter->seq; + struct kmemtrace_free_entry *entry; + struct kmemtrace_user_event *ev; + + trace_assign_type(entry, iter->ent); + + ev = trace_seq_reserve(s, sizeof(*ev)); + if (!ev) + return TRACE_TYPE_PARTIAL_LINE; + + ev->event_id = KMEMTRACE_USER_FREE; + ev->type_id = entry->type_id; + ev->event_size = sizeof(*ev); + ev->cpu = iter->cpu; + ev->timestamp = iter->ts; + ev->call_site = entry->call_site; + ev->ptr = (unsigned long)entry->ptr; + + return TRACE_TYPE_HANDLED; +} + +/* The two other following provide a more minimalistic output */ +static enum print_line_t +kmemtrace_print_alloc_compress(struct trace_iterator *iter) +{ + struct kmemtrace_alloc_entry *entry; + struct trace_seq *s = &iter->seq; + int ret; + + trace_assign_type(entry, iter->ent); + + /* Alloc entry */ + ret = trace_seq_printf(s, " + "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Type */ + switch (entry->type_id) { + case KMEMTRACE_TYPE_KMALLOC: + ret = trace_seq_printf(s, "K "); + break; + case KMEMTRACE_TYPE_CACHE: + ret = trace_seq_printf(s, "C "); + break; + case KMEMTRACE_TYPE_PAGES: + ret = trace_seq_printf(s, "P "); + break; + default: + ret = trace_seq_printf(s, "? "); + } + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Requested */ + ret = trace_seq_printf(s, "%4zu ", entry->bytes_req); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Allocated */ + ret = trace_seq_printf(s, "%4zu ", entry->bytes_alloc); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Flags + * TODO: would be better to see the name of the GFP flag names + */ + ret = trace_seq_printf(s, "%08x ", entry->gfp_flags); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Pointer to allocated */ + ret = trace_seq_printf(s, "0x%tx ", (ptrdiff_t)entry->ptr); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Node and call site*/ + ret = trace_seq_printf(s, "%4d %pf\n", entry->node, + (void *)entry->call_site); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t +kmemtrace_print_free_compress(struct trace_iterator *iter) +{ + struct kmemtrace_free_entry *entry; + struct trace_seq *s = &iter->seq; + int ret; + + trace_assign_type(entry, iter->ent); + + /* Free entry */ + ret = trace_seq_printf(s, " - "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Type */ + switch (entry->type_id) { + case KMEMTRACE_TYPE_KMALLOC: + ret = trace_seq_printf(s, "K "); + break; + case KMEMTRACE_TYPE_CACHE: + ret = trace_seq_printf(s, "C "); + break; + case KMEMTRACE_TYPE_PAGES: + ret = trace_seq_printf(s, "P "); + break; + default: + ret = trace_seq_printf(s, "? "); + } + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Skip requested/allocated/flags */ + ret = trace_seq_printf(s, " "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Pointer to allocated */ + ret = trace_seq_printf(s, "0x%tx ", (ptrdiff_t)entry->ptr); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Skip node and print call site*/ + ret = trace_seq_printf(s, " %pf\n", (void *)entry->call_site); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t kmemtrace_print_line(struct trace_iterator *iter) +{ + struct trace_entry *entry = iter->ent; + + if (!(kmem_tracer_flags.val & TRACE_KMEM_OPT_MINIMAL)) + return TRACE_TYPE_UNHANDLED; + + switch (entry->type) { + case TRACE_KMEM_ALLOC: + return kmemtrace_print_alloc_compress(iter); + case TRACE_KMEM_FREE: + return kmemtrace_print_free_compress(iter); + default: + return TRACE_TYPE_UNHANDLED; + } +} + +static struct trace_event kmem_trace_alloc = { + .type = TRACE_KMEM_ALLOC, + .trace = kmemtrace_print_alloc, + .binary = kmemtrace_print_alloc_user, +}; + +static struct trace_event kmem_trace_free = { + .type = TRACE_KMEM_FREE, + .trace = kmemtrace_print_free, + .binary = kmemtrace_print_free_user, +}; + +static struct tracer kmem_tracer __read_mostly = { + .name = "kmemtrace", + .init = kmem_trace_init, + .reset = kmem_trace_reset, + .print_line = kmemtrace_print_line, + .print_header = kmemtrace_headers, + .flags = &kmem_tracer_flags +}; + +void kmemtrace_init(void) +{ + /* earliest opportunity to start kmem tracing */ +} + +static int __init init_kmem_tracer(void) +{ + if (!register_ftrace_event(&kmem_trace_alloc)) { + pr_warning("Warning: could not register kmem events\n"); + return 1; + } + + if (!register_ftrace_event(&kmem_trace_free)) { + pr_warning("Warning: could not register kmem events\n"); + return 1; + } + + if (!register_tracer(&kmem_tracer)) { + pr_warning("Warning: could not register the kmem tracer\n"); + return 1; + } + + return 0; +} +device_initcall(init_kmem_tracer); diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 8b0daf0662e..454e74e718c 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -4,21 +4,115 @@ * Copyright (C) 2008 Steven Rostedt <srostedt@redhat.com> */ #include <linux/ring_buffer.h> +#include <linux/trace_clock.h> +#include <linux/ftrace_irq.h> #include <linux/spinlock.h> #include <linux/debugfs.h> #include <linux/uaccess.h> +#include <linux/hardirq.h> +#include <linux/kmemcheck.h> #include <linux/module.h> #include <linux/percpu.h> #include <linux/mutex.h> -#include <linux/sched.h> /* used for sched_clock() (for now) */ #include <linux/init.h> #include <linux/hash.h> #include <linux/list.h> +#include <linux/cpu.h> #include <linux/fs.h> #include "trace.h" /* + * The ring buffer header is special. We must manually up keep it. + */ +int ring_buffer_print_entry_header(struct trace_seq *s) +{ + int ret; + + ret = trace_seq_printf(s, "# compressed entry header\n"); + ret = trace_seq_printf(s, "\ttype_len : 5 bits\n"); + ret = trace_seq_printf(s, "\ttime_delta : 27 bits\n"); + ret = trace_seq_printf(s, "\tarray : 32 bits\n"); + ret = trace_seq_printf(s, "\n"); + ret = trace_seq_printf(s, "\tpadding : type == %d\n", + RINGBUF_TYPE_PADDING); + ret = trace_seq_printf(s, "\ttime_extend : type == %d\n", + RINGBUF_TYPE_TIME_EXTEND); + ret = trace_seq_printf(s, "\tdata max type_len == %d\n", + RINGBUF_TYPE_DATA_TYPE_LEN_MAX); + + return ret; +} + +/* + * The ring buffer is made up of a list of pages. A separate list of pages is + * allocated for each CPU. A writer may only write to a buffer that is + * associated with the CPU it is currently executing on. A reader may read + * from any per cpu buffer. + * + * The reader is special. For each per cpu buffer, the reader has its own + * reader page. When a reader has read the entire reader page, this reader + * page is swapped with another page in the ring buffer. + * + * Now, as long as the writer is off the reader page, the reader can do what + * ever it wants with that page. The writer will never write to that page + * again (as long as it is out of the ring buffer). + * + * Here's some silly ASCII art. + * + * +------+ + * |reader| RING BUFFER + * |page | + * +------+ +---+ +---+ +---+ + * | |-->| |-->| | + * +---+ +---+ +---+ + * ^ | + * | | + * +---------------+ + * + * + * +------+ + * |reader| RING BUFFER + * |page |------------------v + * +------+ +---+ +---+ +---+ + * | |-->| |-->| | + * +---+ +---+ +---+ + * ^ | + * | | + * +---------------+ + * + * + * +------+ + * |reader| RING BUFFER + * |page |------------------v + * +------+ +---+ +---+ +---+ + * ^ | |-->| |-->| | + * | +---+ +---+ +---+ + * | | + * | | + * +------------------------------+ + * + * + * +------+ + * |buffer| RING BUFFER + * |page |------------------v + * +------+ +---+ +---+ +---+ + * ^ | | | |-->| | + * | New +---+ +---+ +---+ + * | Reader------^ | + * | page | + * +------------------------------+ + * + * + * After we make this swap, the reader can hand this page off to the splice + * code and be done with it. It can even allocate a new page if it needs to + * and swap that into the ring buffer. + * + * We will be using cmpxchg soon to make all this lockless. + * + */ + +/* * A fast way to enable or disable all ring buffers is to * call tracing_on or tracing_off. Turning off the ring buffers * prevents all ring buffers from being recorded to. @@ -57,7 +151,9 @@ enum { RB_BUFFERS_DISABLED = 1 << RB_BUFFERS_DISABLED_BIT, }; -static long ring_buffer_flags __read_mostly = RB_BUFFERS_ON; +static unsigned long ring_buffer_flags __read_mostly = RB_BUFFERS_ON; + +#define BUF_PAGE_HDR_SIZE offsetof(struct buffer_data_page, data) /** * tracing_on - enable all tracing buffers @@ -89,59 +185,71 @@ EXPORT_SYMBOL_GPL(tracing_off); * tracing_off_permanent - permanently disable ring buffers * * This function, once called, will disable all ring buffers - * permanenty. + * permanently. */ void tracing_off_permanent(void) { set_bit(RB_BUFFERS_DISABLED_BIT, &ring_buffer_flags); } +/** + * tracing_is_on - show state of ring buffers enabled + */ +int tracing_is_on(void) +{ + return ring_buffer_flags == RB_BUFFERS_ON; +} +EXPORT_SYMBOL_GPL(tracing_is_on); + #include "trace.h" -/* Up this if you want to test the TIME_EXTENTS and normalization */ -#define DEBUG_SHIFT 0 +#define RB_EVNT_HDR_SIZE (offsetof(struct ring_buffer_event, array)) +#define RB_ALIGNMENT 4U +#define RB_MAX_SMALL_DATA (RB_ALIGNMENT * RINGBUF_TYPE_DATA_TYPE_LEN_MAX) +#define RB_EVNT_MIN_SIZE 8U /* two 32bit words */ -/* FIXME!!! */ -u64 ring_buffer_time_stamp(int cpu) -{ - u64 time; +/* define RINGBUF_TYPE_DATA for 'case RINGBUF_TYPE_DATA:' */ +#define RINGBUF_TYPE_DATA 0 ... RINGBUF_TYPE_DATA_TYPE_LEN_MAX - preempt_disable_notrace(); - /* shift to debug/test normalization and TIME_EXTENTS */ - time = sched_clock() << DEBUG_SHIFT; - preempt_enable_no_resched_notrace(); +enum { + RB_LEN_TIME_EXTEND = 8, + RB_LEN_TIME_STAMP = 16, +}; - return time; +static inline int rb_null_event(struct ring_buffer_event *event) +{ + return event->type_len == RINGBUF_TYPE_PADDING && !event->time_delta; } -EXPORT_SYMBOL_GPL(ring_buffer_time_stamp); -void ring_buffer_normalize_time_stamp(int cpu, u64 *ts) +static void rb_event_set_padding(struct ring_buffer_event *event) { - /* Just stupid testing the normalize function and deltas */ - *ts >>= DEBUG_SHIFT; + /* padding has a NULL time_delta */ + event->type_len = RINGBUF_TYPE_PADDING; + event->time_delta = 0; } -EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp); -#define RB_EVNT_HDR_SIZE (sizeof(struct ring_buffer_event)) -#define RB_ALIGNMENT_SHIFT 2 -#define RB_ALIGNMENT (1 << RB_ALIGNMENT_SHIFT) -#define RB_MAX_SMALL_DATA 28 +static unsigned +rb_event_data_length(struct ring_buffer_event *event) +{ + unsigned length; -enum { - RB_LEN_TIME_EXTEND = 8, - RB_LEN_TIME_STAMP = 16, -}; + if (event->type_len) + length = event->type_len * RB_ALIGNMENT; + else + length = event->array[0]; + return length + RB_EVNT_HDR_SIZE; +} /* inline for ring buffer fast paths */ -static inline unsigned +static unsigned rb_event_length(struct ring_buffer_event *event) { - unsigned length; - - switch (event->type) { + switch (event->type_len) { case RINGBUF_TYPE_PADDING: - /* undefined */ - return -1; + if (rb_null_event(event)) + /* undefined */ + return -1; + return event->array[0] + RB_EVNT_HDR_SIZE; case RINGBUF_TYPE_TIME_EXTEND: return RB_LEN_TIME_EXTEND; @@ -150,11 +258,7 @@ rb_event_length(struct ring_buffer_event *event) return RB_LEN_TIME_STAMP; case RINGBUF_TYPE_DATA: - if (event->len) - length = event->len << RB_ALIGNMENT_SHIFT; - else - length = event->array[0]; - return length + RB_EVNT_HDR_SIZE; + return rb_event_data_length(event); default: BUG(); } @@ -169,7 +273,7 @@ rb_event_length(struct ring_buffer_event *event) unsigned ring_buffer_event_length(struct ring_buffer_event *event) { unsigned length = rb_event_length(event); - if (event->type != RINGBUF_TYPE_DATA) + if (event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX) return length; length -= RB_EVNT_HDR_SIZE; if (length > RB_MAX_SMALL_DATA + sizeof(event->array[0])) @@ -179,12 +283,12 @@ unsigned ring_buffer_event_length(struct ring_buffer_event *event) EXPORT_SYMBOL_GPL(ring_buffer_event_length); /* inline for ring buffer fast paths */ -static inline void * +static void * rb_event_data(struct ring_buffer_event *event) { - BUG_ON(event->type != RINGBUF_TYPE_DATA); + BUG_ON(event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX); /* If length is in len field, then array[0] has the data */ - if (event->len) + if (event->type_len) return (void *)&event->array[0]; /* Otherwise length is in array[0] and array[1] has the data */ return (void *)&event->array[1]; @@ -209,30 +313,65 @@ EXPORT_SYMBOL_GPL(ring_buffer_event_data); struct buffer_data_page { u64 time_stamp; /* page time stamp */ - local_t commit; /* write commited index */ + local_t commit; /* write committed index */ unsigned char data[]; /* data of buffer page */ }; +/* + * Note, the buffer_page list must be first. The buffer pages + * are allocated in cache lines, which means that each buffer + * page will be at the beginning of a cache line, and thus + * the least significant bits will be zero. We use this to + * add flags in the list struct pointers, to make the ring buffer + * lockless. + */ struct buffer_page { + struct list_head list; /* list of buffer pages */ local_t write; /* index for next write */ unsigned read; /* index for next read */ - struct list_head list; /* list of free pages */ + local_t entries; /* entries on this page */ struct buffer_data_page *page; /* Actual data page */ }; +/* + * The buffer page counters, write and entries, must be reset + * atomically when crossing page boundaries. To synchronize this + * update, two counters are inserted into the number. One is + * the actual counter for the write position or count on the page. + * + * The other is a counter of updaters. Before an update happens + * the update partition of the counter is incremented. This will + * allow the updater to update the counter atomically. + * + * The counter is 20 bits, and the state data is 12. + */ +#define RB_WRITE_MASK 0xfffff +#define RB_WRITE_INTCNT (1 << 20) + static void rb_init_page(struct buffer_data_page *bpage) { local_set(&bpage->commit, 0); } +/** + * ring_buffer_page_len - the size of data on the page. + * @page: The page to read + * + * Returns the amount of data on the page, including buffer page header. + */ +size_t ring_buffer_page_len(void *page) +{ + return local_read(&((struct buffer_data_page *)page)->commit) + + BUF_PAGE_HDR_SIZE; +} + /* * Also stolen from mm/slob.c. Thanks to Mathieu Desnoyers for pointing * this issue out. */ -static inline void free_buffer_page(struct buffer_page *bpage) +static void free_buffer_page(struct buffer_page *bpage) { - if (bpage->page) - free_page((unsigned long)bpage->page); + free_page((unsigned long)bpage->page); kfree(bpage); } @@ -246,7 +385,35 @@ static inline int test_time_stamp(u64 delta) return 0; } -#define BUF_PAGE_SIZE (PAGE_SIZE - sizeof(struct buffer_data_page)) +#define BUF_PAGE_SIZE (PAGE_SIZE - BUF_PAGE_HDR_SIZE) + +/* Max payload is BUF_PAGE_SIZE - header (8bytes) */ +#define BUF_MAX_DATA_SIZE (BUF_PAGE_SIZE - (sizeof(u32) * 2)) + +/* Max number of timestamps that can fit on a page */ +#define RB_TIMESTAMPS_PER_PAGE (BUF_PAGE_SIZE / RB_LEN_TIME_STAMP) + +int ring_buffer_print_page_header(struct trace_seq *s) +{ + struct buffer_data_page field; + int ret; + + ret = trace_seq_printf(s, "\tfield: u64 timestamp;\t" + "offset:0;\tsize:%u;\n", + (unsigned int)sizeof(field.time_stamp)); + + ret = trace_seq_printf(s, "\tfield: local_t commit;\t" + "offset:%u;\tsize:%u;\n", + (unsigned int)offsetof(typeof(field), commit), + (unsigned int)sizeof(field.commit)); + + ret = trace_seq_printf(s, "\tfield: char data;\t" + "offset:%u;\tsize:%u;\n", + (unsigned int)offsetof(typeof(field), data), + (unsigned int)BUF_PAGE_SIZE); + + return ret; +} /* * head_page == tail_page && head == tail then buffer is empty. @@ -254,16 +421,20 @@ static inline int test_time_stamp(u64 delta) struct ring_buffer_per_cpu { int cpu; struct ring_buffer *buffer; - spinlock_t reader_lock; /* serialize readers */ + spinlock_t reader_lock; /* serialize readers */ raw_spinlock_t lock; struct lock_class_key lock_key; - struct list_head pages; + struct list_head *pages; struct buffer_page *head_page; /* read from head */ struct buffer_page *tail_page; /* write to tail */ - struct buffer_page *commit_page; /* commited pages */ + struct buffer_page *commit_page; /* committed pages */ struct buffer_page *reader_page; - unsigned long overrun; - unsigned long entries; + local_t commit_overrun; + local_t overrun; + local_t entries; + local_t committing; + local_t commits; + unsigned long read; u64 write_stamp; u64 read_stamp; atomic_t record_disabled; @@ -273,12 +444,19 @@ struct ring_buffer { unsigned pages; unsigned flags; int cpus; - cpumask_var_t cpumask; atomic_t record_disabled; + cpumask_var_t cpumask; + + struct lock_class_key *reader_lock_key; struct mutex mutex; struct ring_buffer_per_cpu **buffers; + +#ifdef CONFIG_HOTPLUG_CPU + struct notifier_block cpu_notify; +#endif + u64 (*clock)(void); }; struct ring_buffer_iter { @@ -289,33 +467,456 @@ struct ring_buffer_iter { }; /* buffer may be either ring_buffer or ring_buffer_per_cpu */ -#define RB_WARN_ON(buffer, cond) \ - ({ \ - int _____ret = unlikely(cond); \ - if (_____ret) { \ - atomic_inc(&buffer->record_disabled); \ - WARN_ON(1); \ - } \ - _____ret; \ +#define RB_WARN_ON(b, cond) \ + ({ \ + int _____ret = unlikely(cond); \ + if (_____ret) { \ + if (__same_type(*(b), struct ring_buffer_per_cpu)) { \ + struct ring_buffer_per_cpu *__b = \ + (void *)b; \ + atomic_inc(&__b->buffer->record_disabled); \ + } else \ + atomic_inc(&b->record_disabled); \ + WARN_ON(1); \ + } \ + _____ret; \ }) +/* Up this if you want to test the TIME_EXTENTS and normalization */ +#define DEBUG_SHIFT 0 + +static inline u64 rb_time_stamp(struct ring_buffer *buffer, int cpu) +{ + /* shift to debug/test normalization and TIME_EXTENTS */ + return buffer->clock() << DEBUG_SHIFT; +} + +u64 ring_buffer_time_stamp(struct ring_buffer *buffer, int cpu) +{ + u64 time; + + preempt_disable_notrace(); + time = rb_time_stamp(buffer, cpu); + preempt_enable_no_resched_notrace(); + + return time; +} +EXPORT_SYMBOL_GPL(ring_buffer_time_stamp); + +void ring_buffer_normalize_time_stamp(struct ring_buffer *buffer, + int cpu, u64 *ts) +{ + /* Just stupid testing the normalize function and deltas */ + *ts >>= DEBUG_SHIFT; +} +EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp); + +/* + * Making the ring buffer lockless makes things tricky. + * Although writes only happen on the CPU that they are on, + * and they only need to worry about interrupts. Reads can + * happen on any CPU. + * + * The reader page is always off the ring buffer, but when the + * reader finishes with a page, it needs to swap its page with + * a new one from the buffer. The reader needs to take from + * the head (writes go to the tail). But if a writer is in overwrite + * mode and wraps, it must push the head page forward. + * + * Here lies the problem. + * + * The reader must be careful to replace only the head page, and + * not another one. As described at the top of the file in the + * ASCII art, the reader sets its old page to point to the next + * page after head. It then sets the page after head to point to + * the old reader page. But if the writer moves the head page + * during this operation, the reader could end up with the tail. + * + * We use cmpxchg to help prevent this race. We also do something + * special with the page before head. We set the LSB to 1. + * + * When the writer must push the page forward, it will clear the + * bit that points to the head page, move the head, and then set + * the bit that points to the new head page. + * + * We also don't want an interrupt coming in and moving the head + * page on another writer. Thus we use the second LSB to catch + * that too. Thus: + * + * head->list->prev->next bit 1 bit 0 + * ------- ------- + * Normal page 0 0 + * Points to head page 0 1 + * New head page 1 0 + * + * Note we can not trust the prev pointer of the head page, because: + * + * +----+ +-----+ +-----+ + * | |------>| T |---X--->| N | + * | |<------| | | | + * +----+ +-----+ +-----+ + * ^ ^ | + * | +-----+ | | + * +----------| R |----------+ | + * | |<-----------+ + * +-----+ + * + * Key: ---X--> HEAD flag set in pointer + * T Tail page + * R Reader page + * N Next page + * + * (see __rb_reserve_next() to see where this happens) + * + * What the above shows is that the reader just swapped out + * the reader page with a page in the buffer, but before it + * could make the new header point back to the new page added + * it was preempted by a writer. The writer moved forward onto + * the new page added by the reader and is about to move forward + * again. + * + * You can see, it is legitimate for the previous pointer of + * the head (or any page) not to point back to itself. But only + * temporarially. + */ + +#define RB_PAGE_NORMAL 0UL +#define RB_PAGE_HEAD 1UL +#define RB_PAGE_UPDATE 2UL + + +#define RB_FLAG_MASK 3UL + +/* PAGE_MOVED is not part of the mask */ +#define RB_PAGE_MOVED 4UL + +/* + * rb_list_head - remove any bit + */ +static struct list_head *rb_list_head(struct list_head *list) +{ + unsigned long val = (unsigned long)list; + + return (struct list_head *)(val & ~RB_FLAG_MASK); +} + +/* + * rb_is_head_page - test if the give page is the head page + * + * Because the reader may move the head_page pointer, we can + * not trust what the head page is (it may be pointing to + * the reader page). But if the next page is a header page, + * its flags will be non zero. + */ +static int inline +rb_is_head_page(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *page, struct list_head *list) +{ + unsigned long val; + + val = (unsigned long)list->next; + + if ((val & ~RB_FLAG_MASK) != (unsigned long)&page->list) + return RB_PAGE_MOVED; + + return val & RB_FLAG_MASK; +} + +/* + * rb_is_reader_page + * + * The unique thing about the reader page, is that, if the + * writer is ever on it, the previous pointer never points + * back to the reader page. + */ +static int rb_is_reader_page(struct buffer_page *page) +{ + struct list_head *list = page->list.prev; + + return rb_list_head(list->next) != &page->list; +} + +/* + * rb_set_list_to_head - set a list_head to be pointing to head. + */ +static void rb_set_list_to_head(struct ring_buffer_per_cpu *cpu_buffer, + struct list_head *list) +{ + unsigned long *ptr; + + ptr = (unsigned long *)&list->next; + *ptr |= RB_PAGE_HEAD; + *ptr &= ~RB_PAGE_UPDATE; +} + +/* + * rb_head_page_activate - sets up head page + */ +static void rb_head_page_activate(struct ring_buffer_per_cpu *cpu_buffer) +{ + struct buffer_page *head; + + head = cpu_buffer->head_page; + if (!head) + return; + + /* + * Set the previous list pointer to have the HEAD flag. + */ + rb_set_list_to_head(cpu_buffer, head->list.prev); +} + +static void rb_list_head_clear(struct list_head *list) +{ + unsigned long *ptr = (unsigned long *)&list->next; + + *ptr &= ~RB_FLAG_MASK; +} + +/* + * rb_head_page_dactivate - clears head page ptr (for free list) + */ +static void +rb_head_page_deactivate(struct ring_buffer_per_cpu *cpu_buffer) +{ + struct list_head *hd; + + /* Go through the whole list and clear any pointers found. */ + rb_list_head_clear(cpu_buffer->pages); + + list_for_each(hd, cpu_buffer->pages) + rb_list_head_clear(hd); +} + +static int rb_head_page_set(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *head, + struct buffer_page *prev, + int old_flag, int new_flag) +{ + struct list_head *list; + unsigned long val = (unsigned long)&head->list; + unsigned long ret; + + list = &prev->list; + + val &= ~RB_FLAG_MASK; + + ret = (unsigned long)cmpxchg(&list->next, + val | old_flag, val | new_flag); + + /* check if the reader took the page */ + if ((ret & ~RB_FLAG_MASK) != val) + return RB_PAGE_MOVED; + + return ret & RB_FLAG_MASK; +} + +static int rb_head_page_set_update(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *head, + struct buffer_page *prev, + int old_flag) +{ + return rb_head_page_set(cpu_buffer, head, prev, + old_flag, RB_PAGE_UPDATE); +} + +static int rb_head_page_set_head(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *head, + struct buffer_page *prev, + int old_flag) +{ + return rb_head_page_set(cpu_buffer, head, prev, + old_flag, RB_PAGE_HEAD); +} + +static int rb_head_page_set_normal(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *head, + struct buffer_page *prev, + int old_flag) +{ + return rb_head_page_set(cpu_buffer, head, prev, + old_flag, RB_PAGE_NORMAL); +} + +static inline void rb_inc_page(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page **bpage) +{ + struct list_head *p = rb_list_head((*bpage)->list.next); + + *bpage = list_entry(p, struct buffer_page, list); +} + +static struct buffer_page * +rb_set_head_page(struct ring_buffer_per_cpu *cpu_buffer) +{ + struct buffer_page *head; + struct buffer_page *page; + struct list_head *list; + int i; + + if (RB_WARN_ON(cpu_buffer, !cpu_buffer->head_page)) + return NULL; + + /* sanity check */ + list = cpu_buffer->pages; + if (RB_WARN_ON(cpu_buffer, rb_list_head(list->prev->next) != list)) + return NULL; + + page = head = cpu_buffer->head_page; + /* + * It is possible that the writer moves the header behind + * where we started, and we miss in one loop. + * A second loop should grab the header, but we'll do + * three loops just because I'm paranoid. + */ + for (i = 0; i < 3; i++) { + do { + if (rb_is_head_page(cpu_buffer, page, page->list.prev)) { + cpu_buffer->head_page = page; + return page; + } + rb_inc_page(cpu_buffer, &page); + } while (page != head); + } + + RB_WARN_ON(cpu_buffer, 1); + + return NULL; +} + +static int rb_head_page_replace(struct buffer_page *old, + struct buffer_page *new) +{ + unsigned long *ptr = (unsigned long *)&old->list.prev->next; + unsigned long val; + unsigned long ret; + + val = *ptr & ~RB_FLAG_MASK; + val |= RB_PAGE_HEAD; + + ret = cmpxchg(ptr, val, &new->list); + + return ret == val; +} + +/* + * rb_tail_page_update - move the tail page forward + * + * Returns 1 if moved tail page, 0 if someone else did. + */ +static int rb_tail_page_update(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *tail_page, + struct buffer_page *next_page) +{ + struct buffer_page *old_tail; + unsigned long old_entries; + unsigned long old_write; + int ret = 0; + + /* + * The tail page now needs to be moved forward. + * + * We need to reset the tail page, but without messing + * with possible erasing of data brought in by interrupts + * that have moved the tail page and are currently on it. + * + * We add a counter to the write field to denote this. + */ + old_write = local_add_return(RB_WRITE_INTCNT, &next_page->write); + old_entries = local_add_return(RB_WRITE_INTCNT, &next_page->entries); + + /* + * Just make sure we have seen our old_write and synchronize + * with any interrupts that come in. + */ + barrier(); + + /* + * If the tail page is still the same as what we think + * it is, then it is up to us to update the tail + * pointer. + */ + if (tail_page == cpu_buffer->tail_page) { + /* Zero the write counter */ + unsigned long val = old_write & ~RB_WRITE_MASK; + unsigned long eval = old_entries & ~RB_WRITE_MASK; + + /* + * This will only succeed if an interrupt did + * not come in and change it. In which case, we + * do not want to modify it. + * + * We add (void) to let the compiler know that we do not care + * about the return value of these functions. We use the + * cmpxchg to only update if an interrupt did not already + * do it for us. If the cmpxchg fails, we don't care. + */ + (void)local_cmpxchg(&next_page->write, old_write, val); + (void)local_cmpxchg(&next_page->entries, old_entries, eval); + + /* + * No need to worry about races with clearing out the commit. + * it only can increment when a commit takes place. But that + * only happens in the outer most nested commit. + */ + local_set(&next_page->page->commit, 0); + + old_tail = cmpxchg(&cpu_buffer->tail_page, + tail_page, next_page); + + if (old_tail == tail_page) + ret = 1; + } + + return ret; +} + +static int rb_check_bpage(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *bpage) +{ + unsigned long val = (unsigned long)bpage; + + if (RB_WARN_ON(cpu_buffer, val & RB_FLAG_MASK)) + return 1; + + return 0; +} + +/** + * rb_check_list - make sure a pointer to a list has the last bits zero + */ +static int rb_check_list(struct ring_buffer_per_cpu *cpu_buffer, + struct list_head *list) +{ + if (RB_WARN_ON(cpu_buffer, rb_list_head(list->prev) != list->prev)) + return 1; + if (RB_WARN_ON(cpu_buffer, rb_list_head(list->next) != list->next)) + return 1; + return 0; +} + /** * check_pages - integrity check of buffer pages * @cpu_buffer: CPU buffer with pages to test * - * As a safty measure we check to make sure the data pages have not + * As a safety measure we check to make sure the data pages have not * been corrupted. */ static int rb_check_pages(struct ring_buffer_per_cpu *cpu_buffer) { - struct list_head *head = &cpu_buffer->pages; + struct list_head *head = cpu_buffer->pages; struct buffer_page *bpage, *tmp; + rb_head_page_deactivate(cpu_buffer); + if (RB_WARN_ON(cpu_buffer, head->next->prev != head)) return -1; if (RB_WARN_ON(cpu_buffer, head->prev->next != head)) return -1; + if (rb_check_list(cpu_buffer, head)) + return -1; + list_for_each_entry_safe(bpage, tmp, head, list) { if (RB_WARN_ON(cpu_buffer, bpage->list.next->prev != &bpage->list)) @@ -323,25 +924,33 @@ static int rb_check_pages(struct ring_buffer_per_cpu *cpu_buffer) if (RB_WARN_ON(cpu_buffer, bpage->list.prev->next != &bpage->list)) return -1; + if (rb_check_list(cpu_buffer, &bpage->list)) + return -1; } + rb_head_page_activate(cpu_buffer); + return 0; } static int rb_allocate_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned nr_pages) { - struct list_head *head = &cpu_buffer->pages; struct buffer_page *bpage, *tmp; unsigned long addr; LIST_HEAD(pages); unsigned i; + WARN_ON(!nr_pages); + for (i = 0; i < nr_pages; i++) { bpage = kzalloc_node(ALIGN(sizeof(*bpage), cache_line_size()), GFP_KERNEL, cpu_to_node(cpu_buffer->cpu)); if (!bpage) goto free_pages; + + rb_check_bpage(cpu_buffer, bpage); + list_add(&bpage->list, &pages); addr = __get_free_page(GFP_KERNEL); @@ -351,7 +960,13 @@ static int rb_allocate_pages(struct ring_buffer_per_cpu *cpu_buffer, rb_init_page(bpage->page); } - list_splice(&pages, head); + /* + * The ring buffer page list is a circular list that does not + * start and end with a list head. All page list items point to + * other pages. + */ + cpu_buffer->pages = pages.next; + list_del(&pages); rb_check_pages(cpu_buffer); @@ -381,14 +996,16 @@ rb_allocate_cpu_buffer(struct ring_buffer *buffer, int cpu) cpu_buffer->cpu = cpu; cpu_buffer->buffer = buffer; spin_lock_init(&cpu_buffer->reader_lock); + lockdep_set_class(&cpu_buffer->reader_lock, buffer->reader_lock_key); cpu_buffer->lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; - INIT_LIST_HEAD(&cpu_buffer->pages); bpage = kzalloc_node(ALIGN(sizeof(*bpage), cache_line_size()), GFP_KERNEL, cpu_to_node(cpu)); if (!bpage) goto fail_free_buffer; + rb_check_bpage(cpu_buffer, bpage); + cpu_buffer->reader_page = bpage; addr = __get_free_page(GFP_KERNEL); if (!addr) @@ -403,9 +1020,11 @@ rb_allocate_cpu_buffer(struct ring_buffer *buffer, int cpu) goto fail_free_reader; cpu_buffer->head_page - = list_entry(cpu_buffer->pages.next, struct buffer_page, list); + = list_entry(cpu_buffer->pages, struct buffer_page, list); cpu_buffer->tail_page = cpu_buffer->commit_page = cpu_buffer->head_page; + rb_head_page_activate(cpu_buffer); + return cpu_buffer; fail_free_reader: @@ -418,24 +1037,29 @@ rb_allocate_cpu_buffer(struct ring_buffer *buffer, int cpu) static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer) { - struct list_head *head = &cpu_buffer->pages; + struct list_head *head = cpu_buffer->pages; struct buffer_page *bpage, *tmp; - list_del_init(&cpu_buffer->reader_page->list); free_buffer_page(cpu_buffer->reader_page); - list_for_each_entry_safe(bpage, tmp, head, list) { - list_del_init(&bpage->list); + rb_head_page_deactivate(cpu_buffer); + + if (head) { + list_for_each_entry_safe(bpage, tmp, head, list) { + list_del_init(&bpage->list); + free_buffer_page(bpage); + } + bpage = list_entry(head, struct buffer_page, list); free_buffer_page(bpage); } + kfree(cpu_buffer); } -/* - * Causes compile errors if the struct buffer_page gets bigger - * than the struct page. - */ -extern int ring_buffer_page_too_big(void); +#ifdef CONFIG_HOTPLUG_CPU +static int rb_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu); +#endif /** * ring_buffer_alloc - allocate a new ring_buffer @@ -447,17 +1071,13 @@ extern int ring_buffer_page_too_big(void); * when the buffer wraps. If this flag is not set, the buffer will * drop data when the tail hits the head. */ -struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) +struct ring_buffer *__ring_buffer_alloc(unsigned long size, unsigned flags, + struct lock_class_key *key) { struct ring_buffer *buffer; int bsize; int cpu; - /* Paranoid! Optimizes out when all is well */ - if (sizeof(struct buffer_page) > sizeof(struct page)) - ring_buffer_page_too_big(); - - /* keep it in its own cache line */ buffer = kzalloc(ALIGN(sizeof(*buffer), cache_line_size()), GFP_KERNEL); @@ -469,12 +1089,24 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) buffer->pages = DIV_ROUND_UP(size, BUF_PAGE_SIZE); buffer->flags = flags; + buffer->clock = trace_clock_local; + buffer->reader_lock_key = key; /* need at least two pages */ - if (buffer->pages == 1) - buffer->pages++; + if (buffer->pages < 2) + buffer->pages = 2; + /* + * In case of non-hotplug cpu, if the ring-buffer is allocated + * in early initcall, it will not be notified of secondary cpus. + * In that off case, we need to allocate for all possible cpus. + */ +#ifdef CONFIG_HOTPLUG_CPU + get_online_cpus(); + cpumask_copy(buffer->cpumask, cpu_online_mask); +#else cpumask_copy(buffer->cpumask, cpu_possible_mask); +#endif buffer->cpus = nr_cpu_ids; bsize = sizeof(void *) * nr_cpu_ids; @@ -490,6 +1122,13 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) goto fail_free_buffers; } +#ifdef CONFIG_HOTPLUG_CPU + buffer->cpu_notify.notifier_call = rb_cpu_notify; + buffer->cpu_notify.priority = 0; + register_cpu_notifier(&buffer->cpu_notify); +#endif + + put_online_cpus(); mutex_init(&buffer->mutex); return buffer; @@ -503,12 +1142,13 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) fail_free_cpumask: free_cpumask_var(buffer->cpumask); + put_online_cpus(); fail_free_buffer: kfree(buffer); return NULL; } -EXPORT_SYMBOL_GPL(ring_buffer_alloc); +EXPORT_SYMBOL_GPL(__ring_buffer_alloc); /** * ring_buffer_free - free a ring buffer. @@ -519,15 +1159,30 @@ ring_buffer_free(struct ring_buffer *buffer) { int cpu; + get_online_cpus(); + +#ifdef CONFIG_HOTPLUG_CPU + unregister_cpu_notifier(&buffer->cpu_notify); +#endif + for_each_buffer_cpu(buffer, cpu) rb_free_cpu_buffer(buffer->buffers[cpu]); + put_online_cpus(); + + kfree(buffer->buffers); free_cpumask_var(buffer->cpumask); kfree(buffer); } EXPORT_SYMBOL_GPL(ring_buffer_free); +void ring_buffer_set_clock(struct ring_buffer *buffer, + u64 (*clock)(void)) +{ + buffer->clock = clock; +} + static void rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer); static void @@ -540,15 +1195,17 @@ rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned nr_pages) atomic_inc(&cpu_buffer->record_disabled); synchronize_sched(); + rb_head_page_deactivate(cpu_buffer); + for (i = 0; i < nr_pages; i++) { - if (RB_WARN_ON(cpu_buffer, list_empty(&cpu_buffer->pages))) + if (RB_WARN_ON(cpu_buffer, list_empty(cpu_buffer->pages))) return; - p = cpu_buffer->pages.next; + p = cpu_buffer->pages->next; bpage = list_entry(p, struct buffer_page, list); list_del_init(&bpage->list); free_buffer_page(bpage); } - if (RB_WARN_ON(cpu_buffer, list_empty(&cpu_buffer->pages))) + if (RB_WARN_ON(cpu_buffer, list_empty(cpu_buffer->pages))) return; rb_reset_cpu(cpu_buffer); @@ -570,15 +1227,19 @@ rb_insert_pages(struct ring_buffer_per_cpu *cpu_buffer, atomic_inc(&cpu_buffer->record_disabled); synchronize_sched(); + spin_lock_irq(&cpu_buffer->reader_lock); + rb_head_page_deactivate(cpu_buffer); + for (i = 0; i < nr_pages; i++) { if (RB_WARN_ON(cpu_buffer, list_empty(pages))) return; p = pages->next; bpage = list_entry(p, struct buffer_page, list); list_del_init(&bpage->list); - list_add_tail(&bpage->list, &cpu_buffer->pages); + list_add_tail(&bpage->list, cpu_buffer->pages); } rb_reset_cpu(cpu_buffer); + spin_unlock_irq(&cpu_buffer->reader_lock); rb_check_pages(cpu_buffer); @@ -627,16 +1288,15 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) return size; mutex_lock(&buffer->mutex); + get_online_cpus(); nr_pages = DIV_ROUND_UP(size, BUF_PAGE_SIZE); if (size < buffer_size) { /* easy case, just free pages */ - if (RB_WARN_ON(buffer, nr_pages >= buffer->pages)) { - mutex_unlock(&buffer->mutex); - return -1; - } + if (RB_WARN_ON(buffer, nr_pages >= buffer->pages)) + goto out_fail; rm_pages = buffer->pages - nr_pages; @@ -655,10 +1315,8 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) * add these pages to the cpu_buffers. Otherwise we just free * them all and return -ENOMEM; */ - if (RB_WARN_ON(buffer, nr_pages <= buffer->pages)) { - mutex_unlock(&buffer->mutex); - return -1; - } + if (RB_WARN_ON(buffer, nr_pages <= buffer->pages)) + goto out_fail; new_pages = nr_pages - buffer->pages; @@ -683,13 +1341,12 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) rb_insert_pages(cpu_buffer, &pages, new_pages); } - if (RB_WARN_ON(buffer, !list_empty(&pages))) { - mutex_unlock(&buffer->mutex); - return -1; - } + if (RB_WARN_ON(buffer, !list_empty(&pages))) + goto out_fail; out: buffer->pages = nr_pages; + put_online_cpus(); mutex_unlock(&buffer->mutex); return size; @@ -699,15 +1356,20 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) list_del_init(&bpage->list); free_buffer_page(bpage); } + put_online_cpus(); mutex_unlock(&buffer->mutex); return -ENOMEM; -} -EXPORT_SYMBOL_GPL(ring_buffer_resize); -static inline int rb_null_event(struct ring_buffer_event *event) -{ - return event->type == RINGBUF_TYPE_PADDING; + /* + * Something went totally wrong, and we are too paranoid + * to even clean up the mess. + */ + out_fail: + put_online_cpus(); + mutex_unlock(&buffer->mutex); + return -1; } +EXPORT_SYMBOL_GPL(ring_buffer_resize); static inline void * __rb_data_page_index(struct buffer_data_page *bpage, unsigned index) @@ -728,21 +1390,14 @@ rb_reader_event(struct ring_buffer_per_cpu *cpu_buffer) } static inline struct ring_buffer_event * -rb_head_event(struct ring_buffer_per_cpu *cpu_buffer) -{ - return __rb_page_index(cpu_buffer->head_page, - cpu_buffer->head_page->read); -} - -static inline struct ring_buffer_event * rb_iter_head_event(struct ring_buffer_iter *iter) { return __rb_page_index(iter->head_page, iter->head); } -static inline unsigned rb_page_write(struct buffer_page *bpage) +static inline unsigned long rb_page_write(struct buffer_page *bpage) { - return local_read(&bpage->write); + return local_read(&bpage->write) & RB_WRITE_MASK; } static inline unsigned rb_page_commit(struct buffer_page *bpage) @@ -750,6 +1405,11 @@ static inline unsigned rb_page_commit(struct buffer_page *bpage) return local_read(&bpage->page->commit); } +static inline unsigned long rb_page_entries(struct buffer_page *bpage) +{ + return local_read(&bpage->entries) & RB_WRITE_MASK; +} + /* Size is determined by what has been commited */ static inline unsigned rb_page_size(struct buffer_page *bpage) { @@ -762,58 +1422,17 @@ rb_commit_index(struct ring_buffer_per_cpu *cpu_buffer) return rb_page_commit(cpu_buffer->commit_page); } -static inline unsigned rb_head_size(struct ring_buffer_per_cpu *cpu_buffer) -{ - return rb_page_commit(cpu_buffer->head_page); -} - -/* - * When the tail hits the head and the buffer is in overwrite mode, - * the head jumps to the next page and all content on the previous - * page is discarded. But before doing so, we update the overrun - * variable of the buffer. - */ -static void rb_update_overflow(struct ring_buffer_per_cpu *cpu_buffer) -{ - struct ring_buffer_event *event; - unsigned long head; - - for (head = 0; head < rb_head_size(cpu_buffer); - head += rb_event_length(event)) { - - event = __rb_page_index(cpu_buffer->head_page, head); - if (RB_WARN_ON(cpu_buffer, rb_null_event(event))) - return; - /* Only count data entries */ - if (event->type != RINGBUF_TYPE_DATA) - continue; - cpu_buffer->overrun++; - cpu_buffer->entries--; - } -} - -static inline void rb_inc_page(struct ring_buffer_per_cpu *cpu_buffer, - struct buffer_page **bpage) -{ - struct list_head *p = (*bpage)->list.next; - - if (p == &cpu_buffer->pages) - p = p->next; - - *bpage = list_entry(p, struct buffer_page, list); -} - static inline unsigned rb_event_index(struct ring_buffer_event *event) { unsigned long addr = (unsigned long)event; - return (addr & ~PAGE_MASK) - (PAGE_SIZE - BUF_PAGE_SIZE); + return (addr & ~PAGE_MASK) - BUF_PAGE_HDR_SIZE; } static inline int -rb_is_commit(struct ring_buffer_per_cpu *cpu_buffer, - struct ring_buffer_event *event) +rb_event_is_commit(struct ring_buffer_per_cpu *cpu_buffer, + struct ring_buffer_event *event) { unsigned long addr = (unsigned long)event; unsigned long index; @@ -825,34 +1444,11 @@ rb_is_commit(struct ring_buffer_per_cpu *cpu_buffer, rb_commit_index(cpu_buffer) == index; } -static inline void -rb_set_commit_event(struct ring_buffer_per_cpu *cpu_buffer, - struct ring_buffer_event *event) -{ - unsigned long addr = (unsigned long)event; - unsigned long index; - - index = rb_event_index(event); - addr &= PAGE_MASK; - - while (cpu_buffer->commit_page->page != (void *)addr) { - if (RB_WARN_ON(cpu_buffer, - cpu_buffer->commit_page == cpu_buffer->tail_page)) - return; - cpu_buffer->commit_page->page->commit = - cpu_buffer->commit_page->write; - rb_inc_page(cpu_buffer, &cpu_buffer->commit_page); - cpu_buffer->write_stamp = - cpu_buffer->commit_page->page->time_stamp; - } - - /* Now set the commit to the event's index */ - local_set(&cpu_buffer->commit_page->page->commit, index); -} - -static inline void +static void rb_set_commit_to_write(struct ring_buffer_per_cpu *cpu_buffer) { + unsigned long max_count; + /* * We only race with interrupts and NMIs on this CPU. * If we own the commit event, then we can commit @@ -862,9 +1458,16 @@ rb_set_commit_to_write(struct ring_buffer_per_cpu *cpu_buffer) * assign the commit to the tail. */ again: + max_count = cpu_buffer->buffer->pages * 100; + while (cpu_buffer->commit_page != cpu_buffer->tail_page) { - cpu_buffer->commit_page->page->commit = - cpu_buffer->commit_page->write; + if (RB_WARN_ON(cpu_buffer, !(--max_count))) + return; + if (RB_WARN_ON(cpu_buffer, + rb_is_reader_page(cpu_buffer->tail_page))) + return; + local_set(&cpu_buffer->commit_page->page->commit, + rb_page_write(cpu_buffer->commit_page)); rb_inc_page(cpu_buffer, &cpu_buffer->commit_page); cpu_buffer->write_stamp = cpu_buffer->commit_page->page->time_stamp; @@ -873,8 +1476,12 @@ rb_set_commit_to_write(struct ring_buffer_per_cpu *cpu_buffer) } while (rb_commit_index(cpu_buffer) != rb_page_write(cpu_buffer->commit_page)) { - cpu_buffer->commit_page->page->commit = - cpu_buffer->commit_page->write; + + local_set(&cpu_buffer->commit_page->page->commit, + rb_page_write(cpu_buffer->commit_page)); + RB_WARN_ON(cpu_buffer, + local_read(&cpu_buffer->commit_page->page->commit) & + ~RB_WRITE_MASK); barrier(); } @@ -896,7 +1503,7 @@ static void rb_reset_reader_page(struct ring_buffer_per_cpu *cpu_buffer) cpu_buffer->reader_page->read = 0; } -static inline void rb_inc_iter(struct ring_buffer_iter *iter) +static void rb_inc_iter(struct ring_buffer_iter *iter) { struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer; @@ -907,7 +1514,7 @@ static inline void rb_inc_iter(struct ring_buffer_iter *iter) * to the head page instead of next. */ if (iter->head_page == cpu_buffer->reader_page) - iter->head_page = cpu_buffer->head_page; + iter->head_page = rb_set_head_page(cpu_buffer); else rb_inc_page(cpu_buffer, &iter->head_page); @@ -926,45 +1533,189 @@ static inline void rb_inc_iter(struct ring_buffer_iter *iter) * and with this, we can determine what to place into the * data field. */ -static inline void +static void rb_update_event(struct ring_buffer_event *event, unsigned type, unsigned length) { - event->type = type; + event->type_len = type; switch (type) { case RINGBUF_TYPE_PADDING: - break; - case RINGBUF_TYPE_TIME_EXTEND: - event->len = - (RB_LEN_TIME_EXTEND + (RB_ALIGNMENT-1)) - >> RB_ALIGNMENT_SHIFT; - break; - case RINGBUF_TYPE_TIME_STAMP: - event->len = - (RB_LEN_TIME_STAMP + (RB_ALIGNMENT-1)) - >> RB_ALIGNMENT_SHIFT; break; - case RINGBUF_TYPE_DATA: + case 0: length -= RB_EVNT_HDR_SIZE; - if (length > RB_MAX_SMALL_DATA) { - event->len = 0; + if (length > RB_MAX_SMALL_DATA) event->array[0] = length; - } else - event->len = - (length + (RB_ALIGNMENT-1)) - >> RB_ALIGNMENT_SHIFT; + else + event->type_len = DIV_ROUND_UP(length, RB_ALIGNMENT); break; default: BUG(); } } -static inline unsigned rb_calculate_event_length(unsigned length) +/* + * rb_handle_head_page - writer hit the head page + * + * Returns: +1 to retry page + * 0 to continue + * -1 on error + */ +static int +rb_handle_head_page(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *tail_page, + struct buffer_page *next_page) +{ + struct buffer_page *new_head; + int entries; + int type; + int ret; + + entries = rb_page_entries(next_page); + + /* + * The hard part is here. We need to move the head + * forward, and protect against both readers on + * other CPUs and writers coming in via interrupts. + */ + type = rb_head_page_set_update(cpu_buffer, next_page, tail_page, + RB_PAGE_HEAD); + + /* + * type can be one of four: + * NORMAL - an interrupt already moved it for us + * HEAD - we are the first to get here. + * UPDATE - we are the interrupt interrupting + * a current move. + * MOVED - a reader on another CPU moved the next + * pointer to its reader page. Give up + * and try again. + */ + + switch (type) { + case RB_PAGE_HEAD: + /* + * We changed the head to UPDATE, thus + * it is our responsibility to update + * the counters. + */ + local_add(entries, &cpu_buffer->overrun); + + /* + * The entries will be zeroed out when we move the + * tail page. + */ + + /* still more to do */ + break; + + case RB_PAGE_UPDATE: + /* + * This is an interrupt that interrupt the + * previous update. Still more to do. + */ + break; + case RB_PAGE_NORMAL: + /* + * An interrupt came in before the update + * and processed this for us. + * Nothing left to do. + */ + return 1; + case RB_PAGE_MOVED: + /* + * The reader is on another CPU and just did + * a swap with our next_page. + * Try again. + */ + return 1; + default: + RB_WARN_ON(cpu_buffer, 1); /* WTF??? */ + return -1; + } + + /* + * Now that we are here, the old head pointer is + * set to UPDATE. This will keep the reader from + * swapping the head page with the reader page. + * The reader (on another CPU) will spin till + * we are finished. + * + * We just need to protect against interrupts + * doing the job. We will set the next pointer + * to HEAD. After that, we set the old pointer + * to NORMAL, but only if it was HEAD before. + * otherwise we are an interrupt, and only + * want the outer most commit to reset it. + */ + new_head = next_page; + rb_inc_page(cpu_buffer, &new_head); + + ret = rb_head_page_set_head(cpu_buffer, new_head, next_page, + RB_PAGE_NORMAL); + + /* + * Valid returns are: + * HEAD - an interrupt came in and already set it. + * NORMAL - One of two things: + * 1) We really set it. + * 2) A bunch of interrupts came in and moved + * the page forward again. + */ + switch (ret) { + case RB_PAGE_HEAD: + case RB_PAGE_NORMAL: + /* OK */ + break; + default: + RB_WARN_ON(cpu_buffer, 1); + return -1; + } + + /* + * It is possible that an interrupt came in, + * set the head up, then more interrupts came in + * and moved it again. When we get back here, + * the page would have been set to NORMAL but we + * just set it back to HEAD. + * + * How do you detect this? Well, if that happened + * the tail page would have moved. + */ + if (ret == RB_PAGE_NORMAL) { + /* + * If the tail had moved passed next, then we need + * to reset the pointer. + */ + if (cpu_buffer->tail_page != tail_page && + cpu_buffer->tail_page != next_page) + rb_head_page_set_normal(cpu_buffer, new_head, + next_page, + RB_PAGE_HEAD); + } + + /* + * If this was the outer most commit (the one that + * changed the original pointer from HEAD to UPDATE), + * then it is up to us to reset it to NORMAL. + */ + if (type == RB_PAGE_HEAD) { + ret = rb_head_page_set_normal(cpu_buffer, next_page, + tail_page, + RB_PAGE_UPDATE); + if (RB_WARN_ON(cpu_buffer, + ret != RB_PAGE_UPDATE)) + return -1; + } + + return 0; +} + +static unsigned rb_calculate_event_length(unsigned length) { struct ring_buffer_event event; /* Used only for sizeof array */ @@ -981,133 +1732,239 @@ static inline unsigned rb_calculate_event_length(unsigned length) return length; } -static struct ring_buffer_event * -__rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, - unsigned type, unsigned long length, u64 *ts) +static inline void +rb_reset_tail(struct ring_buffer_per_cpu *cpu_buffer, + struct buffer_page *tail_page, + unsigned long tail, unsigned long length) { - struct buffer_page *tail_page, *head_page, *reader_page, *commit_page; - unsigned long tail, write; - struct ring_buffer *buffer = cpu_buffer->buffer; struct ring_buffer_event *event; - unsigned long flags; - commit_page = cpu_buffer->commit_page; - /* we just need to protect against interrupts */ - barrier(); - tail_page = cpu_buffer->tail_page; - write = local_add_return(length, &tail_page->write); - tail = write - length; + /* + * Only the event that crossed the page boundary + * must fill the old tail_page with padding. + */ + if (tail >= BUF_PAGE_SIZE) { + local_sub(length, &tail_page->write); + return; + } - /* See if we shot pass the end of this buffer page */ - if (write > BUF_PAGE_SIZE) { - struct buffer_page *next_page = tail_page; + event = __rb_page_index(tail_page, tail); + kmemcheck_annotate_bitfield(event, bitfield); - local_irq_save(flags); - __raw_spin_lock(&cpu_buffer->lock); + /* + * If this event is bigger than the minimum size, then + * we need to be careful that we don't subtract the + * write counter enough to allow another writer to slip + * in on this page. + * We put in a discarded commit instead, to make sure + * that this space is not used again. + * + * If we are less than the minimum size, we don't need to + * worry about it. + */ + if (tail > (BUF_PAGE_SIZE - RB_EVNT_MIN_SIZE)) { + /* No room for any events */ - rb_inc_page(cpu_buffer, &next_page); + /* Mark the rest of the page with padding */ + rb_event_set_padding(event); - head_page = cpu_buffer->head_page; - reader_page = cpu_buffer->reader_page; + /* Set the write back to the previous setting */ + local_sub(length, &tail_page->write); + return; + } - /* we grabbed the lock before incrementing */ - if (RB_WARN_ON(cpu_buffer, next_page == reader_page)) - goto out_unlock; + /* Put in a discarded event */ + event->array[0] = (BUF_PAGE_SIZE - tail) - RB_EVNT_HDR_SIZE; + event->type_len = RINGBUF_TYPE_PADDING; + /* time delta must be non zero */ + event->time_delta = 1; - /* - * If for some reason, we had an interrupt storm that made - * it all the way around the buffer, bail, and warn - * about it. - */ - if (unlikely(next_page == commit_page)) { - WARN_ON_ONCE(1); - goto out_unlock; - } + /* Set write to end of buffer */ + length = (tail + length) - BUF_PAGE_SIZE; + local_sub(length, &tail_page->write); +} - if (next_page == head_page) { - if (!(buffer->flags & RB_FL_OVERWRITE)) { - /* reset write */ - if (tail <= BUF_PAGE_SIZE) - local_set(&tail_page->write, tail); - goto out_unlock; - } +static struct ring_buffer_event * +rb_move_tail(struct ring_buffer_per_cpu *cpu_buffer, + unsigned long length, unsigned long tail, + struct buffer_page *commit_page, + struct buffer_page *tail_page, u64 *ts) +{ + struct ring_buffer *buffer = cpu_buffer->buffer; + struct buffer_page *next_page; + int ret; - /* tail_page has not moved yet? */ - if (tail_page == cpu_buffer->tail_page) { - /* count overflows */ - rb_update_overflow(cpu_buffer); + next_page = tail_page; - rb_inc_page(cpu_buffer, &head_page); - cpu_buffer->head_page = head_page; - cpu_buffer->head_page->read = 0; - } - } + rb_inc_page(cpu_buffer, &next_page); + + /* + * If for some reason, we had an interrupt storm that made + * it all the way around the buffer, bail, and warn + * about it. + */ + if (unlikely(next_page == commit_page)) { + local_inc(&cpu_buffer->commit_overrun); + goto out_reset; + } + + /* + * This is where the fun begins! + * + * We are fighting against races between a reader that + * could be on another CPU trying to swap its reader + * page with the buffer head. + * + * We are also fighting against interrupts coming in and + * moving the head or tail on us as well. + * + * If the next page is the head page then we have filled + * the buffer, unless the commit page is still on the + * reader page. + */ + if (rb_is_head_page(cpu_buffer, next_page, &tail_page->list)) { /* - * If the tail page is still the same as what we think - * it is, then it is up to us to update the tail - * pointer. + * If the commit is not on the reader page, then + * move the header page. */ - if (tail_page == cpu_buffer->tail_page) { - local_set(&next_page->write, 0); - local_set(&next_page->page->commit, 0); - cpu_buffer->tail_page = next_page; - - /* reread the time stamp */ - *ts = ring_buffer_time_stamp(cpu_buffer->cpu); - cpu_buffer->tail_page->page->time_stamp = *ts; + if (!rb_is_reader_page(cpu_buffer->commit_page)) { + /* + * If we are not in overwrite mode, + * this is easy, just stop here. + */ + if (!(buffer->flags & RB_FL_OVERWRITE)) + goto out_reset; + + ret = rb_handle_head_page(cpu_buffer, + tail_page, + next_page); + if (ret < 0) + goto out_reset; + if (ret) + goto out_again; + } else { + /* + * We need to be careful here too. The + * commit page could still be on the reader + * page. We could have a small buffer, and + * have filled up the buffer with events + * from interrupts and such, and wrapped. + * + * Note, if the tail page is also the on the + * reader_page, we let it move out. + */ + if (unlikely((cpu_buffer->commit_page != + cpu_buffer->tail_page) && + (cpu_buffer->commit_page == + cpu_buffer->reader_page))) { + local_inc(&cpu_buffer->commit_overrun); + goto out_reset; + } } + } + ret = rb_tail_page_update(cpu_buffer, tail_page, next_page); + if (ret) { /* - * The actual tail page has moved forward. + * Nested commits always have zero deltas, so + * just reread the time stamp */ - if (tail < BUF_PAGE_SIZE) { - /* Mark the rest of the page with padding */ - event = __rb_page_index(tail_page, tail); - event->type = RINGBUF_TYPE_PADDING; - } + *ts = rb_time_stamp(buffer, cpu_buffer->cpu); + next_page->page->time_stamp = *ts; + } - if (tail <= BUF_PAGE_SIZE) - /* Set the write back to the previous setting */ - local_set(&tail_page->write, tail); + out_again: - /* - * If this was a commit entry that failed, - * increment that too - */ - if (tail_page == cpu_buffer->commit_page && - tail == rb_commit_index(cpu_buffer)) { - rb_set_commit_to_write(cpu_buffer); - } + rb_reset_tail(cpu_buffer, tail_page, tail, length); - __raw_spin_unlock(&cpu_buffer->lock); - local_irq_restore(flags); + /* fail and let the caller try again */ + return ERR_PTR(-EAGAIN); - /* fail and let the caller try again */ - return ERR_PTR(-EAGAIN); - } + out_reset: + /* reset write */ + rb_reset_tail(cpu_buffer, tail_page, tail, length); - /* We reserved something on the buffer */ + return NULL; +} - if (RB_WARN_ON(cpu_buffer, write > BUF_PAGE_SIZE)) - return NULL; +static struct ring_buffer_event * +__rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, + unsigned type, unsigned long length, u64 *ts) +{ + struct buffer_page *tail_page, *commit_page; + struct ring_buffer_event *event; + unsigned long tail, write; + + commit_page = cpu_buffer->commit_page; + /* we just need to protect against interrupts */ + barrier(); + tail_page = cpu_buffer->tail_page; + write = local_add_return(length, &tail_page->write); + + /* set write to only the index of the write */ + write &= RB_WRITE_MASK; + tail = write - length; + + /* See if we shot pass the end of this buffer page */ + if (write > BUF_PAGE_SIZE) + return rb_move_tail(cpu_buffer, length, tail, + commit_page, tail_page, ts); + + /* We reserved something on the buffer */ event = __rb_page_index(tail_page, tail); + kmemcheck_annotate_bitfield(event, bitfield); rb_update_event(event, type, length); + /* The passed in type is zero for DATA */ + if (likely(!type)) + local_inc(&tail_page->entries); + /* - * If this is a commit and the tail is zero, then update - * this page's time stamp. + * If this is the first commit on the page, then update + * its timestamp. */ - if (!tail && rb_is_commit(cpu_buffer, event)) - cpu_buffer->commit_page->page->time_stamp = *ts; + if (!tail) + tail_page->page->time_stamp = *ts; return event; +} - out_unlock: - __raw_spin_unlock(&cpu_buffer->lock); - local_irq_restore(flags); - return NULL; +static inline int +rb_try_to_discard(struct ring_buffer_per_cpu *cpu_buffer, + struct ring_buffer_event *event) +{ + unsigned long new_index, old_index; + struct buffer_page *bpage; + unsigned long index; + unsigned long addr; + + new_index = rb_event_index(event); + old_index = new_index + rb_event_length(event); + addr = (unsigned long)event; + addr &= PAGE_MASK; + + bpage = cpu_buffer->tail_page; + + if (bpage->page == (void *)addr && rb_page_write(bpage) == old_index) { + unsigned long write_mask = + local_read(&bpage->write) & ~RB_WRITE_MASK; + /* + * This is on the tail page. It is possible that + * a write could come in and move the tail page + * and write to the next page. That is fine + * because we just shorten what is on this page. + */ + old_index += write_mask; + new_index += write_mask; + index = local_cmpxchg(&bpage->write, old_index, new_index); + if (index == old_index) + return 1; + } + + /* could not discard */ + return 0; } static int @@ -1142,26 +1999,33 @@ rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, return -EAGAIN; /* Only a commited time event can update the write stamp */ - if (rb_is_commit(cpu_buffer, event)) { + if (rb_event_is_commit(cpu_buffer, event)) { /* - * If this is the first on the page, then we need to - * update the page itself, and just put in a zero. + * If this is the first on the page, then it was + * updated with the page itself. Try to discard it + * and if we can't just make it zero. */ if (rb_event_index(event)) { event->time_delta = *delta & TS_MASK; event->array[0] = *delta >> TS_SHIFT; } else { - cpu_buffer->commit_page->page->time_stamp = *ts; - event->time_delta = 0; - event->array[0] = 0; + /* try to discard, since we do not need this */ + if (!rb_try_to_discard(cpu_buffer, event)) { + /* nope, just zero it */ + event->time_delta = 0; + event->array[0] = 0; + } } cpu_buffer->write_stamp = *ts; /* let the caller know this was the commit */ ret = 1; } else { - /* Darn, this is just wasted space */ - event->time_delta = 0; - event->array[0] = 0; + /* Try to discard the event */ + if (!rb_try_to_discard(cpu_buffer, event)) { + /* Darn, this is just wasted space */ + event->time_delta = 0; + event->array[0] = 0; + } ret = 0; } @@ -1170,15 +2034,72 @@ rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, return ret; } +static void rb_start_commit(struct ring_buffer_per_cpu *cpu_buffer) +{ + local_inc(&cpu_buffer->committing); + local_inc(&cpu_buffer->commits); +} + +static void rb_end_commit(struct ring_buffer_per_cpu *cpu_buffer) +{ + unsigned long commits; + + if (RB_WARN_ON(cpu_buffer, + !local_read(&cpu_buffer->committing))) + return; + + again: + commits = local_read(&cpu_buffer->commits); + /* synchronize with interrupts */ + barrier(); + if (local_read(&cpu_buffer->committing) == 1) + rb_set_commit_to_write(cpu_buffer); + + local_dec(&cpu_buffer->committing); + + /* synchronize with interrupts */ + barrier(); + + /* + * Need to account for interrupts coming in between the + * updating of the commit page and the clearing of the + * committing counter. + */ + if (unlikely(local_read(&cpu_buffer->commits) != commits) && + !local_read(&cpu_buffer->committing)) { + local_inc(&cpu_buffer->committing); + goto again; + } +} + static struct ring_buffer_event * -rb_reserve_next_event(struct ring_buffer_per_cpu *cpu_buffer, - unsigned type, unsigned long length) +rb_reserve_next_event(struct ring_buffer *buffer, + struct ring_buffer_per_cpu *cpu_buffer, + unsigned long length) { struct ring_buffer_event *event; - u64 ts, delta; + u64 ts, delta = 0; int commit = 0; int nr_loops = 0; + rb_start_commit(cpu_buffer); + +#ifdef CONFIG_RING_BUFFER_ALLOW_SWAP + /* + * Due to the ability to swap a cpu buffer from a buffer + * it is possible it was swapped before we committed. + * (committing stops a swap). We check for it here and + * if it happened, we have to fail the write. + */ + barrier(); + if (unlikely(ACCESS_ONCE(cpu_buffer->buffer) != buffer)) { + local_dec(&cpu_buffer->committing); + local_dec(&cpu_buffer->commits); + return NULL; + } +#endif + + length = rb_calculate_event_length(length); again: /* * We allow for interrupts to reenter here and do a trace. @@ -1190,9 +2111,9 @@ rb_reserve_next_event(struct ring_buffer_per_cpu *cpu_buffer, * Bail! */ if (RB_WARN_ON(cpu_buffer, ++nr_loops > 1000)) - return NULL; + goto out_fail; - ts = ring_buffer_time_stamp(cpu_buffer->cpu); + ts = rb_time_stamp(cpu_buffer->buffer, cpu_buffer->cpu); /* * Only the first commit can update the timestamp. @@ -1202,70 +2123,99 @@ rb_reserve_next_event(struct ring_buffer_per_cpu *cpu_buffer, * also be made. But only the entry that did the actual * commit will be something other than zero. */ - if (cpu_buffer->tail_page == cpu_buffer->commit_page && - rb_page_write(cpu_buffer->tail_page) == - rb_commit_index(cpu_buffer)) { + if (likely(cpu_buffer->tail_page == cpu_buffer->commit_page && + rb_page_write(cpu_buffer->tail_page) == + rb_commit_index(cpu_buffer))) { + u64 diff; - delta = ts - cpu_buffer->write_stamp; + diff = ts - cpu_buffer->write_stamp; - /* make sure this delta is calculated here */ + /* make sure this diff is calculated here */ barrier(); /* Did the write stamp get updated already? */ if (unlikely(ts < cpu_buffer->write_stamp)) - delta = 0; + goto get_event; - if (test_time_stamp(delta)) { + delta = diff; + if (unlikely(test_time_stamp(delta))) { commit = rb_add_time_stamp(cpu_buffer, &ts, &delta); - if (commit == -EBUSY) - return NULL; + goto out_fail; if (commit == -EAGAIN) goto again; RB_WARN_ON(cpu_buffer, commit < 0); } - } else - /* Non commits have zero deltas */ - delta = 0; + } - event = __rb_reserve_next(cpu_buffer, type, length, &ts); - if (PTR_ERR(event) == -EAGAIN) + get_event: + event = __rb_reserve_next(cpu_buffer, 0, length, &ts); + if (unlikely(PTR_ERR(event) == -EAGAIN)) goto again; - if (!event) { - if (unlikely(commit)) - /* - * Ouch! We needed a timestamp and it was commited. But - * we didn't get our event reserved. - */ - rb_set_commit_to_write(cpu_buffer); - return NULL; - } + if (!event) + goto out_fail; - /* - * If the timestamp was commited, make the commit our entry - * now so that we will update it when needed. - */ - if (commit) - rb_set_commit_event(cpu_buffer, event); - else if (!rb_is_commit(cpu_buffer, event)) + if (!rb_event_is_commit(cpu_buffer, event)) delta = 0; event->time_delta = delta; return event; + + out_fail: + rb_end_commit(cpu_buffer); + return NULL; +} + +#ifdef CONFIG_TRACING + +#define TRACE_RECURSIVE_DEPTH 16 + +static int trace_recursive_lock(void) +{ + current->trace_recursion++; + + if (likely(current->trace_recursion < TRACE_RECURSIVE_DEPTH)) + return 0; + + /* Disable all tracing before we do anything else */ + tracing_off_permanent(); + + printk_once(KERN_WARNING "Tracing recursion: depth[%ld]:" + "HC[%lu]:SC[%lu]:NMI[%lu]\n", + current->trace_recursion, + hardirq_count() >> HARDIRQ_SHIFT, + softirq_count() >> SOFTIRQ_SHIFT, + in_nmi()); + + WARN_ON_ONCE(1); + return -1; +} + +static void trace_recursive_unlock(void) +{ + WARN_ON_ONCE(!current->trace_recursion); + + current->trace_recursion--; } +#else + +#define trace_recursive_lock() (0) +#define trace_recursive_unlock() do { } while (0) + +#endif + static DEFINE_PER_CPU(int, rb_need_resched); /** * ring_buffer_lock_reserve - reserve a part of the buffer * @buffer: the ring buffer to reserve from * @length: the length of the data to reserve (excluding event header) - * @flags: a pointer to save the interrupt flags * * Returns a reseverd event on the ring buffer to copy directly to. * The user of this interface will need to get the body to write into @@ -1278,9 +2228,7 @@ static DEFINE_PER_CPU(int, rb_need_resched); * If NULL is returned, then nothing has been allocated or locked. */ struct ring_buffer_event * -ring_buffer_lock_reserve(struct ring_buffer *buffer, - unsigned long length, - unsigned long *flags) +ring_buffer_lock_reserve(struct ring_buffer *buffer, unsigned long length) { struct ring_buffer_per_cpu *cpu_buffer; struct ring_buffer_event *event; @@ -1295,6 +2243,9 @@ ring_buffer_lock_reserve(struct ring_buffer *buffer, /* If we are tracing schedule, we don't want to recurse */ resched = ftrace_preempt_disable(); + if (trace_recursive_lock()) + goto out_nocheck; + cpu = raw_smp_processor_id(); if (!cpumask_test_cpu(cpu, buffer->cpumask)) @@ -1305,11 +2256,10 @@ ring_buffer_lock_reserve(struct ring_buffer *buffer, if (atomic_read(&cpu_buffer->record_disabled)) goto out; - length = rb_calculate_event_length(length); - if (length > BUF_PAGE_SIZE) + if (length > BUF_MAX_DATA_SIZE) goto out; - event = rb_reserve_next_event(cpu_buffer, RINGBUF_TYPE_DATA, length); + event = rb_reserve_next_event(buffer, cpu_buffer, length); if (!event) goto out; @@ -1324,38 +2274,45 @@ ring_buffer_lock_reserve(struct ring_buffer *buffer, return event; out: + trace_recursive_unlock(); + + out_nocheck: ftrace_preempt_enable(resched); return NULL; } EXPORT_SYMBOL_GPL(ring_buffer_lock_reserve); -static void rb_commit(struct ring_buffer_per_cpu *cpu_buffer, +static void +rb_update_write_stamp(struct ring_buffer_per_cpu *cpu_buffer, struct ring_buffer_event *event) { - cpu_buffer->entries++; - - /* Only process further if we own the commit */ - if (!rb_is_commit(cpu_buffer, event)) - return; - - cpu_buffer->write_stamp += event->time_delta; + /* + * The event first in the commit queue updates the + * time stamp. + */ + if (rb_event_is_commit(cpu_buffer, event)) + cpu_buffer->write_stamp += event->time_delta; +} - rb_set_commit_to_write(cpu_buffer); +static void rb_commit(struct ring_buffer_per_cpu *cpu_buffer, + struct ring_buffer_event *event) +{ + local_inc(&cpu_buffer->entries); + rb_update_write_stamp(cpu_buffer, event); + rb_end_commit(cpu_buffer); } /** * ring_buffer_unlock_commit - commit a reserved * @buffer: The buffer to commit to * @event: The event pointer to commit. - * @flags: the interrupt flags received from ring_buffer_lock_reserve. * * This commits the data to the ring buffer, and releases any locks held. * * Must be paired with ring_buffer_lock_reserve. */ int ring_buffer_unlock_commit(struct ring_buffer *buffer, - struct ring_buffer_event *event, - unsigned long flags) + struct ring_buffer_event *event) { struct ring_buffer_per_cpu *cpu_buffer; int cpu = raw_smp_processor_id(); @@ -1364,6 +2321,8 @@ int ring_buffer_unlock_commit(struct ring_buffer *buffer, rb_commit(cpu_buffer, event); + trace_recursive_unlock(); + /* * Only the last preempt count needs to restore preemption. */ @@ -1376,6 +2335,119 @@ int ring_buffer_unlock_commit(struct ring_buffer *buffer, } EXPORT_SYMBOL_GPL(ring_buffer_unlock_commit); +static inline void rb_event_discard(struct ring_buffer_event *event) +{ + /* array[0] holds the actual length for the discarded event */ + event->array[0] = rb_event_data_length(event) - RB_EVNT_HDR_SIZE; + event->type_len = RINGBUF_TYPE_PADDING; + /* time delta must be non zero */ + if (!event->time_delta) + event->time_delta = 1; +} + +/* + * Decrement the entries to the page that an event is on. + * The event does not even need to exist, only the pointer + * to the page it is on. This may only be called before the commit + * takes place. + */ +static inline void +rb_decrement_entry(struct ring_buffer_per_cpu *cpu_buffer, + struct ring_buffer_event *event) +{ + unsigned long addr = (unsigned long)event; + struct buffer_page *bpage = cpu_buffer->commit_page; + struct buffer_page *start; + + addr &= PAGE_MASK; + + /* Do the likely case first */ + if (likely(bpage->page == (void *)addr)) { + local_dec(&bpage->entries); + return; + } + + /* + * Because the commit page may be on the reader page we + * start with the next page and check the end loop there. + */ + rb_inc_page(cpu_buffer, &bpage); + start = bpage; + do { + if (bpage->page == (void *)addr) { + local_dec(&bpage->entries); + return; + } + rb_inc_page(cpu_buffer, &bpage); + } while (bpage != start); + + /* commit not part of this buffer?? */ + RB_WARN_ON(cpu_buffer, 1); +} + +/** + * ring_buffer_commit_discard - discard an event that has not been committed + * @buffer: the ring buffer + * @event: non committed event to discard + * + * Sometimes an event that is in the ring buffer needs to be ignored. + * This function lets the user discard an event in the ring buffer + * and then that event will not be read later. + * + * This function only works if it is called before the the item has been + * committed. It will try to free the event from the ring buffer + * if another event has not been added behind it. + * + * If another event has been added behind it, it will set the event + * up as discarded, and perform the commit. + * + * If this function is called, do not call ring_buffer_unlock_commit on + * the event. + */ +void ring_buffer_discard_commit(struct ring_buffer *buffer, + struct ring_buffer_event *event) +{ + struct ring_buffer_per_cpu *cpu_buffer; + int cpu; + + /* The event is discarded regardless */ + rb_event_discard(event); + + cpu = smp_processor_id(); + cpu_buffer = buffer->buffers[cpu]; + + /* + * This must only be called if the event has not been + * committed yet. Thus we can assume that preemption + * is still disabled. + */ + RB_WARN_ON(buffer, !local_read(&cpu_buffer->committing)); + + rb_decrement_entry(cpu_buffer, event); + if (rb_try_to_discard(cpu_buffer, event)) + goto out; + + /* + * The commit is still visible by the reader, so we + * must still update the timestamp. + */ + rb_update_write_stamp(cpu_buffer, event); + out: + rb_end_commit(cpu_buffer); + + trace_recursive_unlock(); + + /* + * Only the last preempt count needs to restore preemption. + */ + if (preempt_count() == 1) + ftrace_preempt_enable(per_cpu(rb_need_resched, cpu)); + else + preempt_enable_no_resched_notrace(); + +} +EXPORT_SYMBOL_GPL(ring_buffer_discard_commit); + /** * ring_buffer_write - write data to the buffer without reserving * @buffer: The ring buffer to write to. @@ -1395,7 +2467,6 @@ int ring_buffer_write(struct ring_buffer *buffer, { struct ring_buffer_per_cpu *cpu_buffer; struct ring_buffer_event *event; - unsigned long event_length; void *body; int ret = -EBUSY; int cpu, resched; @@ -1418,9 +2489,10 @@ int ring_buffer_write(struct ring_buffer *buffer, if (atomic_read(&cpu_buffer->record_disabled)) goto out; - event_length = rb_calculate_event_length(length); - event = rb_reserve_next_event(cpu_buffer, - RINGBUF_TYPE_DATA, event_length); + if (length > BUF_MAX_DATA_SIZE) + goto out; + + event = rb_reserve_next_event(buffer, cpu_buffer, length); if (!event) goto out; @@ -1438,12 +2510,16 @@ int ring_buffer_write(struct ring_buffer *buffer, } EXPORT_SYMBOL_GPL(ring_buffer_write); -static inline int rb_per_cpu_empty(struct ring_buffer_per_cpu *cpu_buffer) +static int rb_per_cpu_empty(struct ring_buffer_per_cpu *cpu_buffer) { struct buffer_page *reader = cpu_buffer->reader_page; - struct buffer_page *head = cpu_buffer->head_page; + struct buffer_page *head = rb_set_head_page(cpu_buffer); struct buffer_page *commit = cpu_buffer->commit_page; + /* In case of error, head will be NULL */ + if (unlikely(!head)) + return 1; + return reader->read == rb_page_commit(reader) && (commit == reader || (commit == head && @@ -1528,12 +2604,16 @@ EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu); unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; + unsigned long ret; if (!cpumask_test_cpu(cpu, buffer->cpumask)) return 0; cpu_buffer = buffer->buffers[cpu]; - return cpu_buffer->entries; + ret = (local_read(&cpu_buffer->entries) - local_read(&cpu_buffer->overrun)) + - cpu_buffer->read; + + return ret; } EXPORT_SYMBOL_GPL(ring_buffer_entries_cpu); @@ -1545,16 +2625,40 @@ EXPORT_SYMBOL_GPL(ring_buffer_entries_cpu); unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; + unsigned long ret; if (!cpumask_test_cpu(cpu, buffer->cpumask)) return 0; cpu_buffer = buffer->buffers[cpu]; - return cpu_buffer->overrun; + ret = local_read(&cpu_buffer->overrun); + + return ret; } EXPORT_SYMBOL_GPL(ring_buffer_overrun_cpu); /** + * ring_buffer_commit_overrun_cpu - get the number of overruns caused by commits + * @buffer: The ring buffer + * @cpu: The per CPU buffer to get the number of overruns from + */ +unsigned long +ring_buffer_commit_overrun_cpu(struct ring_buffer *buffer, int cpu) +{ + struct ring_buffer_per_cpu *cpu_buffer; + unsigned long ret; + + if (!cpumask_test_cpu(cpu, buffer->cpumask)) + return 0; + + cpu_buffer = buffer->buffers[cpu]; + ret = local_read(&cpu_buffer->commit_overrun); + + return ret; +} +EXPORT_SYMBOL_GPL(ring_buffer_commit_overrun_cpu); + +/** * ring_buffer_entries - get the number of entries in a buffer * @buffer: The ring buffer * @@ -1570,7 +2674,8 @@ unsigned long ring_buffer_entries(struct ring_buffer *buffer) /* if you care about this being correct, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; - entries += cpu_buffer->entries; + entries += (local_read(&cpu_buffer->entries) - + local_read(&cpu_buffer->overrun)) - cpu_buffer->read; } return entries; @@ -1593,7 +2698,7 @@ unsigned long ring_buffer_overruns(struct ring_buffer *buffer) /* if you care about this being correct, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; - overruns += cpu_buffer->overrun; + overruns += local_read(&cpu_buffer->overrun); } return overruns; @@ -1606,8 +2711,10 @@ static void rb_iter_reset(struct ring_buffer_iter *iter) /* Iterator usage is expected to have record disabled */ if (list_empty(&cpu_buffer->reader_page->list)) { - iter->head_page = cpu_buffer->head_page; - iter->head = cpu_buffer->head_page->read; + iter->head_page = rb_set_head_page(cpu_buffer); + if (unlikely(!iter->head_page)) + return; + iter->head = iter->head_page->read; } else { iter->head_page = cpu_buffer->reader_page; iter->head = cpu_buffer->reader_page->read; @@ -1627,9 +2734,14 @@ static void rb_iter_reset(struct ring_buffer_iter *iter) */ void ring_buffer_iter_reset(struct ring_buffer_iter *iter) { - struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer; + struct ring_buffer_per_cpu *cpu_buffer; unsigned long flags; + if (!iter) + return; + + cpu_buffer = iter->cpu_buffer; + spin_lock_irqsave(&cpu_buffer->reader_lock, flags); rb_iter_reset(iter); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); @@ -1657,7 +2769,7 @@ rb_update_read_stamp(struct ring_buffer_per_cpu *cpu_buffer, { u64 delta; - switch (event->type) { + switch (event->type_len) { case RINGBUF_TYPE_PADDING: return; @@ -1688,7 +2800,7 @@ rb_update_iter_read_stamp(struct ring_buffer_iter *iter, { u64 delta; - switch (event->type) { + switch (event->type_len) { case RINGBUF_TYPE_PADDING: return; @@ -1719,6 +2831,7 @@ rb_get_reader_page(struct ring_buffer_per_cpu *cpu_buffer) struct buffer_page *reader = NULL; unsigned long flags; int nr_loops = 0; + int ret; local_irq_save(flags); __raw_spin_lock(&cpu_buffer->lock); @@ -1752,29 +2865,56 @@ rb_get_reader_page(struct ring_buffer_per_cpu *cpu_buffer) goto out; /* - * Splice the empty reader page into the list around the head. * Reset the reader page to size zero. */ + local_set(&cpu_buffer->reader_page->write, 0); + local_set(&cpu_buffer->reader_page->entries, 0); + local_set(&cpu_buffer->reader_page->page->commit, 0); - reader = cpu_buffer->head_page; + spin: + /* + * Splice the empty reader page into the list around the head. + */ + reader = rb_set_head_page(cpu_buffer); cpu_buffer->reader_page->list.next = reader->list.next; cpu_buffer->reader_page->list.prev = reader->list.prev; - local_set(&cpu_buffer->reader_page->write, 0); - local_set(&cpu_buffer->reader_page->page->commit, 0); + /* + * cpu_buffer->pages just needs to point to the buffer, it + * has no specific buffer page to point to. Lets move it out + * of our way so we don't accidently swap it. + */ + cpu_buffer->pages = reader->list.prev; - /* Make the reader page now replace the head */ - reader->list.prev->next = &cpu_buffer->reader_page->list; - reader->list.next->prev = &cpu_buffer->reader_page->list; + /* The reader page will be pointing to the new head */ + rb_set_list_to_head(cpu_buffer, &cpu_buffer->reader_page->list); + + /* + * Here's the tricky part. + * + * We need to move the pointer past the header page. + * But we can only do that if a writer is not currently + * moving it. The page before the header page has the + * flag bit '1' set if it is pointing to the page we want. + * but if the writer is in the process of moving it + * than it will be '2' or already moved '0'. + */ + + ret = rb_head_page_replace(reader, cpu_buffer->reader_page); /* - * If the tail is on the reader, then we must set the head - * to the inserted page, otherwise we set it one before. + * If we did not convert it, then we must try again. */ - cpu_buffer->head_page = cpu_buffer->reader_page; + if (!ret) + goto spin; - if (cpu_buffer->commit_page != reader) - rb_inc_page(cpu_buffer, &cpu_buffer->head_page); + /* + * Yeah! We succeeded in replacing the page. + * + * Now make the new head point back to the reader page. + */ + reader->list.next->prev = &cpu_buffer->reader_page->list; + rb_inc_page(cpu_buffer, &cpu_buffer->head_page); /* Finally update the reader page to the new head */ cpu_buffer->reader_page = reader; @@ -1803,8 +2943,8 @@ static void rb_advance_reader(struct ring_buffer_per_cpu *cpu_buffer) event = rb_reader_event(cpu_buffer); - if (event->type == RINGBUF_TYPE_DATA) - cpu_buffer->entries--; + if (event->type_len <= RINGBUF_TYPE_DATA_TYPE_LEN_MAX) + cpu_buffer->read++; rb_update_read_stamp(cpu_buffer, event); @@ -1826,8 +2966,8 @@ static void rb_advance_iter(struct ring_buffer_iter *iter) * Check if we are at the end of the buffer. */ if (iter->head >= rb_page_size(iter->head_page)) { - if (RB_WARN_ON(buffer, - iter->head_page == cpu_buffer->commit_page)) + /* discarded commits can make the page empty */ + if (iter->head_page == cpu_buffer->commit_page) return; rb_inc_iter(iter); return; @@ -1864,21 +3004,16 @@ rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) struct buffer_page *reader; int nr_loops = 0; - if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return NULL; - cpu_buffer = buffer->buffers[cpu]; again: /* * We repeat when a timestamp is encountered. It is possible * to get multiple timestamps from an interrupt entering just - * as one timestamp is about to be written. The max times - * that this can happen is the number of nested interrupts we - * can have. Nesting 10 deep of interrupts is clearly - * an anomaly. + * as one timestamp is about to be written, or from discarded + * commits. The most that we can have is the number on a single page. */ - if (RB_WARN_ON(cpu_buffer, ++nr_loops > 10)) + if (RB_WARN_ON(cpu_buffer, ++nr_loops > RB_TIMESTAMPS_PER_PAGE)) return NULL; reader = rb_get_reader_page(cpu_buffer); @@ -1887,11 +3022,19 @@ rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) event = rb_reader_event(cpu_buffer); - switch (event->type) { + switch (event->type_len) { case RINGBUF_TYPE_PADDING: - RB_WARN_ON(cpu_buffer, 1); - rb_advance_reader(cpu_buffer); - return NULL; + if (rb_null_event(event)) + RB_WARN_ON(cpu_buffer, 1); + /* + * Because the writer could be discarding every + * event it creates (which would probably be bad) + * if we were to go back to "again" then we may never + * catch up, and will trigger the warn on, or lock + * the box. Return the padding, and we will release + * the current locks, and try again. + */ + return event; case RINGBUF_TYPE_TIME_EXTEND: /* Internal data, OK to advance */ @@ -1906,7 +3049,8 @@ rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) case RINGBUF_TYPE_DATA: if (ts) { *ts = cpu_buffer->read_stamp + event->time_delta; - ring_buffer_normalize_time_stamp(cpu_buffer->cpu, ts); + ring_buffer_normalize_time_stamp(buffer, + cpu_buffer->cpu, ts); } return event; @@ -1934,14 +3078,14 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) again: /* - * We repeat when a timestamp is encountered. It is possible - * to get multiple timestamps from an interrupt entering just - * as one timestamp is about to be written. The max times - * that this can happen is the number of nested interrupts we - * can have. Nesting 10 deep of interrupts is clearly - * an anomaly. + * We repeat when a timestamp is encountered. + * We can get multiple timestamps by nested interrupts or also + * if filtering is on (discarding commits). Since discarding + * commits can be frequent we can get a lot of timestamps. + * But we limit them by not adding timestamps if they begin + * at the start of a page. */ - if (RB_WARN_ON(cpu_buffer, ++nr_loops > 10)) + if (RB_WARN_ON(cpu_buffer, ++nr_loops > RB_TIMESTAMPS_PER_PAGE)) return NULL; if (rb_per_cpu_empty(cpu_buffer)) @@ -1949,10 +3093,14 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) event = rb_iter_head_event(iter); - switch (event->type) { + switch (event->type_len) { case RINGBUF_TYPE_PADDING: - rb_inc_iter(iter); - goto again; + if (rb_null_event(event)) { + rb_inc_iter(iter); + goto again; + } + rb_advance_iter(iter); + return event; case RINGBUF_TYPE_TIME_EXTEND: /* Internal data, OK to advance */ @@ -1967,7 +3115,8 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) case RINGBUF_TYPE_DATA: if (ts) { *ts = iter->read_stamp + event->time_delta; - ring_buffer_normalize_time_stamp(cpu_buffer->cpu, ts); + ring_buffer_normalize_time_stamp(buffer, + cpu_buffer->cpu, ts); } return event; @@ -1979,6 +3128,21 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) } EXPORT_SYMBOL_GPL(ring_buffer_iter_peek); +static inline int rb_ok_to_lock(void) +{ + /* + * If an NMI die dumps out the content of the ring buffer + * do not grab locks. We also permanently disable the ring + * buffer too. A one time deal is all you get from reading + * the ring buffer from an NMI. + */ + if (likely(!in_nmi())) + return 1; + + tracing_off_permanent(); + return 0; +} + /** * ring_buffer_peek - peek at the next event to be read * @buffer: The ring buffer to read @@ -1994,10 +3158,25 @@ ring_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; struct ring_buffer_event *event; unsigned long flags; + int dolock; - spin_lock_irqsave(&cpu_buffer->reader_lock, flags); + if (!cpumask_test_cpu(cpu, buffer->cpumask)) + return NULL; + + dolock = rb_ok_to_lock(); + again: + local_irq_save(flags); + if (dolock) + spin_lock(&cpu_buffer->reader_lock); event = rb_buffer_peek(buffer, cpu, ts); - spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + if (event && event->type_len == RINGBUF_TYPE_PADDING) + rb_advance_reader(cpu_buffer); + if (dolock) + spin_unlock(&cpu_buffer->reader_lock); + local_irq_restore(flags); + + if (event && event->type_len == RINGBUF_TYPE_PADDING) + goto again; return event; } @@ -2017,10 +3196,14 @@ ring_buffer_iter_peek(struct ring_buffer_iter *iter, u64 *ts) struct ring_buffer_event *event; unsigned long flags; + again: spin_lock_irqsave(&cpu_buffer->reader_lock, flags); event = rb_iter_peek(iter, ts); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + if (event && event->type_len == RINGBUF_TYPE_PADDING) + goto again; + return event; } @@ -2035,23 +3218,38 @@ ring_buffer_iter_peek(struct ring_buffer_iter *iter, u64 *ts) struct ring_buffer_event * ring_buffer_consume(struct ring_buffer *buffer, int cpu, u64 *ts) { - struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; - struct ring_buffer_event *event; + struct ring_buffer_per_cpu *cpu_buffer; + struct ring_buffer_event *event = NULL; unsigned long flags; + int dolock; + + dolock = rb_ok_to_lock(); + + again: + /* might be called in atomic */ + preempt_disable(); if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return NULL; + goto out; - spin_lock_irqsave(&cpu_buffer->reader_lock, flags); + cpu_buffer = buffer->buffers[cpu]; + local_irq_save(flags); + if (dolock) + spin_lock(&cpu_buffer->reader_lock); event = rb_buffer_peek(buffer, cpu, ts); - if (!event) - goto out; + if (event) + rb_advance_reader(cpu_buffer); - rb_advance_reader(cpu_buffer); + if (dolock) + spin_unlock(&cpu_buffer->reader_lock); + local_irq_restore(flags); out: - spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + preempt_enable(); + + if (event && event->type_len == RINGBUF_TYPE_PADDING) + goto again; return event; } @@ -2132,10 +3330,14 @@ ring_buffer_read(struct ring_buffer_iter *iter, u64 *ts) unsigned long flags; spin_lock_irqsave(&cpu_buffer->reader_lock, flags); + again: event = rb_iter_peek(iter, ts); if (!event) goto out; + if (event->type_len == RINGBUF_TYPE_PADDING) + goto again; + rb_advance_iter(iter); out: spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); @@ -2157,9 +3359,12 @@ EXPORT_SYMBOL_GPL(ring_buffer_size); static void rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer) { + rb_head_page_deactivate(cpu_buffer); + cpu_buffer->head_page - = list_entry(cpu_buffer->pages.next, struct buffer_page, list); + = list_entry(cpu_buffer->pages, struct buffer_page, list); local_set(&cpu_buffer->head_page->write, 0); + local_set(&cpu_buffer->head_page->entries, 0); local_set(&cpu_buffer->head_page->page->commit, 0); cpu_buffer->head_page->read = 0; @@ -2169,11 +3374,21 @@ rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer) INIT_LIST_HEAD(&cpu_buffer->reader_page->list); local_set(&cpu_buffer->reader_page->write, 0); + local_set(&cpu_buffer->reader_page->entries, 0); local_set(&cpu_buffer->reader_page->page->commit, 0); cpu_buffer->reader_page->read = 0; - cpu_buffer->overrun = 0; - cpu_buffer->entries = 0; + local_set(&cpu_buffer->commit_overrun, 0); + local_set(&cpu_buffer->overrun, 0); + local_set(&cpu_buffer->entries, 0); + local_set(&cpu_buffer->committing, 0); + local_set(&cpu_buffer->commits, 0); + cpu_buffer->read = 0; + + cpu_buffer->write_stamp = 0; + cpu_buffer->read_stamp = 0; + + rb_head_page_activate(cpu_buffer); } /** @@ -2189,15 +3404,23 @@ void ring_buffer_reset_cpu(struct ring_buffer *buffer, int cpu) if (!cpumask_test_cpu(cpu, buffer->cpumask)) return; + atomic_inc(&cpu_buffer->record_disabled); + spin_lock_irqsave(&cpu_buffer->reader_lock, flags); + if (RB_WARN_ON(cpu_buffer, local_read(&cpu_buffer->committing))) + goto out; + __raw_spin_lock(&cpu_buffer->lock); rb_reset_cpu(cpu_buffer); __raw_spin_unlock(&cpu_buffer->lock); + out: spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + + atomic_dec(&cpu_buffer->record_disabled); } EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu); @@ -2221,14 +3444,28 @@ EXPORT_SYMBOL_GPL(ring_buffer_reset); int ring_buffer_empty(struct ring_buffer *buffer) { struct ring_buffer_per_cpu *cpu_buffer; + unsigned long flags; + int dolock; int cpu; + int ret; + + dolock = rb_ok_to_lock(); /* yes this is racy, but if you don't like the race, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; - if (!rb_per_cpu_empty(cpu_buffer)) + local_irq_save(flags); + if (dolock) + spin_lock(&cpu_buffer->reader_lock); + ret = rb_per_cpu_empty(cpu_buffer); + if (dolock) + spin_unlock(&cpu_buffer->reader_lock); + local_irq_restore(flags); + + if (!ret) return 0; } + return 1; } EXPORT_SYMBOL_GPL(ring_buffer_empty); @@ -2241,15 +3478,29 @@ EXPORT_SYMBOL_GPL(ring_buffer_empty); int ring_buffer_empty_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; + unsigned long flags; + int dolock; + int ret; if (!cpumask_test_cpu(cpu, buffer->cpumask)) return 1; + dolock = rb_ok_to_lock(); + cpu_buffer = buffer->buffers[cpu]; - return rb_per_cpu_empty(cpu_buffer); + local_irq_save(flags); + if (dolock) + spin_lock(&cpu_buffer->reader_lock); + ret = rb_per_cpu_empty(cpu_buffer); + if (dolock) + spin_unlock(&cpu_buffer->reader_lock); + local_irq_restore(flags); + + return ret; } EXPORT_SYMBOL_GPL(ring_buffer_empty_cpu); +#ifdef CONFIG_RING_BUFFER_ALLOW_SWAP /** * ring_buffer_swap_cpu - swap a CPU buffer between two ring buffers * @buffer_a: One buffer to swap with @@ -2265,18 +3516,36 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a, { struct ring_buffer_per_cpu *cpu_buffer_a; struct ring_buffer_per_cpu *cpu_buffer_b; + int ret = -EINVAL; if (!cpumask_test_cpu(cpu, buffer_a->cpumask) || !cpumask_test_cpu(cpu, buffer_b->cpumask)) - return -EINVAL; + goto out; /* At least make sure the two buffers are somewhat the same */ if (buffer_a->pages != buffer_b->pages) - return -EINVAL; + goto out; + + ret = -EAGAIN; + + if (ring_buffer_flags != RB_BUFFERS_ON) + goto out; + + if (atomic_read(&buffer_a->record_disabled)) + goto out; + + if (atomic_read(&buffer_b->record_disabled)) + goto out; cpu_buffer_a = buffer_a->buffers[cpu]; cpu_buffer_b = buffer_b->buffers[cpu]; + if (atomic_read(&cpu_buffer_a->record_disabled)) + goto out; + + if (atomic_read(&cpu_buffer_b->record_disabled)) + goto out; + /* * We can't do a synchronize_sched here because this * function can be called in atomic context. @@ -2286,39 +3555,28 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a, atomic_inc(&cpu_buffer_a->record_disabled); atomic_inc(&cpu_buffer_b->record_disabled); + ret = -EBUSY; + if (local_read(&cpu_buffer_a->committing)) + goto out_dec; + if (local_read(&cpu_buffer_b->committing)) + goto out_dec; + buffer_a->buffers[cpu] = cpu_buffer_b; buffer_b->buffers[cpu] = cpu_buffer_a; cpu_buffer_b->buffer = buffer_a; cpu_buffer_a->buffer = buffer_b; + ret = 0; + +out_dec: atomic_dec(&cpu_buffer_a->record_disabled); atomic_dec(&cpu_buffer_b->record_disabled); - - return 0; +out: + return ret; } EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu); - -static void rb_remove_entries(struct ring_buffer_per_cpu *cpu_buffer, - struct buffer_data_page *bpage) -{ - struct ring_buffer_event *event; - unsigned long head; - - __raw_spin_lock(&cpu_buffer->lock); - for (head = 0; head < local_read(&bpage->commit); - head += rb_event_length(event)) { - - event = __rb_data_page_index(bpage, head); - if (RB_WARN_ON(cpu_buffer, rb_null_event(event))) - return; - /* Only count data entries */ - if (event->type != RINGBUF_TYPE_DATA) - continue; - cpu_buffer->entries--; - } - __raw_spin_unlock(&cpu_buffer->lock); -} +#endif /* CONFIG_RING_BUFFER_ALLOW_SWAP */ /** * ring_buffer_alloc_read_page - allocate a page to read from buffer @@ -2337,8 +3595,8 @@ static void rb_remove_entries(struct ring_buffer_per_cpu *cpu_buffer, */ void *ring_buffer_alloc_read_page(struct ring_buffer *buffer) { - unsigned long addr; struct buffer_data_page *bpage; + unsigned long addr; addr = __get_free_page(GFP_KERNEL); if (!addr) @@ -2346,8 +3604,11 @@ void *ring_buffer_alloc_read_page(struct ring_buffer *buffer) bpage = (void *)addr; + rb_init_page(bpage); + return bpage; } +EXPORT_SYMBOL_GPL(ring_buffer_alloc_read_page); /** * ring_buffer_free_read_page - free an allocated read page @@ -2360,11 +3621,13 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) { free_page((unsigned long)data); } +EXPORT_SYMBOL_GPL(ring_buffer_free_read_page); /** * ring_buffer_read_page - extract a page from the ring buffer * @buffer: buffer to extract from * @data_page: the page to use allocated from ring_buffer_alloc_read_page + * @len: amount to extract * @cpu: the cpu of the buffer to extract * @full: should the extraction only happen when the page is full. * @@ -2374,12 +3637,12 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) * to swap with a page in the ring buffer. * * for example: - * rpage = ring_buffer_alloc_page(buffer); + * rpage = ring_buffer_alloc_read_page(buffer); * if (!rpage) * return error; - * ret = ring_buffer_read_page(buffer, &rpage, cpu, 0); - * if (ret) - * process_page(rpage); + * ret = ring_buffer_read_page(buffer, &rpage, len, cpu, 0); + * if (ret >= 0) + * process_page(rpage, ret); * * When @full is set, the function will not return true unless * the writer is off the reader page. @@ -2390,80 +3653,129 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) * responsible for that. * * Returns: - * 1 if data has been transferred - * 0 if no data has been transferred. + * >=0 if data has been transferred, returns the offset of consumed data. + * <0 if no data has been transferred. */ int ring_buffer_read_page(struct ring_buffer *buffer, - void **data_page, int cpu, int full) + void **data_page, size_t len, int cpu, int full) { struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; struct ring_buffer_event *event; struct buffer_data_page *bpage; + struct buffer_page *reader; unsigned long flags; - int ret = 0; + unsigned int commit; + unsigned int read; + u64 save_timestamp; + int ret = -1; + + if (!cpumask_test_cpu(cpu, buffer->cpumask)) + goto out; + + /* + * If len is not big enough to hold the page header, then + * we can not copy anything. + */ + if (len <= BUF_PAGE_HDR_SIZE) + goto out; + + len -= BUF_PAGE_HDR_SIZE; if (!data_page) - return 0; + goto out; bpage = *data_page; if (!bpage) - return 0; + goto out; spin_lock_irqsave(&cpu_buffer->reader_lock, flags); - /* - * rb_buffer_peek will get the next ring buffer if - * the current reader page is empty. - */ - event = rb_buffer_peek(buffer, cpu, NULL); - if (!event) - goto out; + reader = rb_get_reader_page(cpu_buffer); + if (!reader) + goto out_unlock; + + event = rb_reader_event(cpu_buffer); + + read = reader->read; + commit = rb_page_commit(reader); - /* check for data */ - if (!local_read(&cpu_buffer->reader_page->page->commit)) - goto out; /* - * If the writer is already off of the read page, then simply - * switch the read page with the given page. Otherwise - * we need to copy the data from the reader to the writer. + * If this page has been partially read or + * if len is not big enough to read the rest of the page or + * a writer is still on the page, then + * we must copy the data from the page to the buffer. + * Otherwise, we can simply swap the page with the one passed in. */ - if (cpu_buffer->reader_page == cpu_buffer->commit_page) { - unsigned int read = cpu_buffer->reader_page->read; + if (read || (len < (commit - read)) || + cpu_buffer->reader_page == cpu_buffer->commit_page) { + struct buffer_data_page *rpage = cpu_buffer->reader_page->page; + unsigned int rpos = read; + unsigned int pos = 0; + unsigned int size; if (full) - goto out; - /* The writer is still on the reader page, we must copy */ - bpage = cpu_buffer->reader_page->page; - memcpy(bpage->data, - cpu_buffer->reader_page->page->data + read, - local_read(&bpage->commit) - read); + goto out_unlock; - /* consume what was read */ - cpu_buffer->reader_page += read; + if (len > (commit - read)) + len = (commit - read); + size = rb_event_length(event); + + if (len < size) + goto out_unlock; + + /* save the current timestamp, since the user will need it */ + save_timestamp = cpu_buffer->read_stamp; + + /* Need to copy one event at a time */ + do { + memcpy(bpage->data + pos, rpage->data + rpos, size); + + len -= size; + + rb_advance_reader(cpu_buffer); + rpos = reader->read; + pos += size; + + event = rb_reader_event(cpu_buffer); + size = rb_event_length(event); + } while (len > size); + + /* update bpage */ + local_set(&bpage->commit, pos); + bpage->time_stamp = save_timestamp; + + /* we copied everything to the beginning */ + read = 0; } else { + /* update the entry counter */ + cpu_buffer->read += rb_page_entries(reader); + /* swap the pages */ rb_init_page(bpage); - bpage = cpu_buffer->reader_page->page; - cpu_buffer->reader_page->page = *data_page; - cpu_buffer->reader_page->read = 0; + bpage = reader->page; + reader->page = *data_page; + local_set(&reader->write, 0); + local_set(&reader->entries, 0); + reader->read = 0; *data_page = bpage; } - ret = 1; + ret = read; - /* update the entry counter */ - rb_remove_entries(cpu_buffer, bpage); - out: + out_unlock: spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + out: return ret; } +EXPORT_SYMBOL_GPL(ring_buffer_read_page); +#ifdef CONFIG_TRACING static ssize_t rb_simple_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { - long *p = filp->private_data; + unsigned long *p = filp->private_data; char buf[64]; int r; @@ -2479,9 +3791,9 @@ static ssize_t rb_simple_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) { - long *p = filp->private_data; + unsigned long *p = filp->private_data; char buf[64]; - long val; + unsigned long val; int ret; if (cnt >= sizeof(buf)) @@ -2506,7 +3818,7 @@ rb_simple_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations rb_simple_fops = { +static const struct file_operations rb_simple_fops = { .open = tracing_open_generic, .read = rb_simple_read, .write = rb_simple_write, @@ -2516,16 +3828,53 @@ static struct file_operations rb_simple_fops = { static __init int rb_init_debugfs(void) { struct dentry *d_tracer; - struct dentry *entry; d_tracer = tracing_init_dentry(); - entry = debugfs_create_file("tracing_on", 0644, d_tracer, - &ring_buffer_flags, &rb_simple_fops); - if (!entry) - pr_warning("Could not create debugfs 'tracing_on' entry\n"); + trace_create_file("tracing_on", 0644, d_tracer, + &ring_buffer_flags, &rb_simple_fops); return 0; } fs_initcall(rb_init_debugfs); +#endif + +#ifdef CONFIG_HOTPLUG_CPU +static int rb_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + struct ring_buffer *buffer = + container_of(self, struct ring_buffer, cpu_notify); + long cpu = (long)hcpu; + + switch (action) { + case CPU_UP_PREPARE: + case CPU_UP_PREPARE_FROZEN: + if (cpumask_test_cpu(cpu, buffer->cpumask)) + return NOTIFY_OK; + + buffer->buffers[cpu] = + rb_allocate_cpu_buffer(buffer, cpu); + if (!buffer->buffers[cpu]) { + WARN(1, "failed to allocate ring buffer on CPU %ld\n", + cpu); + return NOTIFY_OK; + } + smp_wmb(); + cpumask_set_cpu(cpu, buffer->cpumask); + break; + case CPU_DOWN_PREPARE: + case CPU_DOWN_PREPARE_FROZEN: + /* + * Do nothing. + * If we were to free the buffer, then the user would + * lose any trace that was in the buffer. + */ + break; + default: + break; + } + return NOTIFY_OK; +} +#endif diff --git a/kernel/trace/ring_buffer_benchmark.c b/kernel/trace/ring_buffer_benchmark.c new file mode 100644 index 00000000000..573d3cc762c --- /dev/null +++ b/kernel/trace/ring_buffer_benchmark.c @@ -0,0 +1,419 @@ +/* + * ring buffer tester and benchmark + * + * Copyright (C) 2009 Steven Rostedt <srostedt@redhat.com> + */ +#include <linux/ring_buffer.h> +#include <linux/completion.h> +#include <linux/kthread.h> +#include <linux/module.h> +#include <linux/time.h> + +struct rb_page { + u64 ts; + local_t commit; + char data[4080]; +}; + +/* run time and sleep time in seconds */ +#define RUN_TIME 10 +#define SLEEP_TIME 10 + +/* number of events for writer to wake up the reader */ +static int wakeup_interval = 100; + +static int reader_finish; +static struct completion read_start; +static struct completion read_done; + +static struct ring_buffer *buffer; +static struct task_struct *producer; +static struct task_struct *consumer; +static unsigned long read; + +static int disable_reader; +module_param(disable_reader, uint, 0644); +MODULE_PARM_DESC(disable_reader, "only run producer"); + +static int read_events; + +static int kill_test; + +#define KILL_TEST() \ + do { \ + if (!kill_test) { \ + kill_test = 1; \ + WARN_ON(1); \ + } \ + } while (0) + +enum event_status { + EVENT_FOUND, + EVENT_DROPPED, +}; + +static enum event_status read_event(int cpu) +{ + struct ring_buffer_event *event; + int *entry; + u64 ts; + + event = ring_buffer_consume(buffer, cpu, &ts); + if (!event) + return EVENT_DROPPED; + + entry = ring_buffer_event_data(event); + if (*entry != cpu) { + KILL_TEST(); + return EVENT_DROPPED; + } + + read++; + return EVENT_FOUND; +} + +static enum event_status read_page(int cpu) +{ + struct ring_buffer_event *event; + struct rb_page *rpage; + unsigned long commit; + void *bpage; + int *entry; + int ret; + int inc; + int i; + + bpage = ring_buffer_alloc_read_page(buffer); + if (!bpage) + return EVENT_DROPPED; + + ret = ring_buffer_read_page(buffer, &bpage, PAGE_SIZE, cpu, 1); + if (ret >= 0) { + rpage = bpage; + commit = local_read(&rpage->commit); + for (i = 0; i < commit && !kill_test; i += inc) { + + if (i >= (PAGE_SIZE - offsetof(struct rb_page, data))) { + KILL_TEST(); + break; + } + + inc = -1; + event = (void *)&rpage->data[i]; + switch (event->type_len) { + case RINGBUF_TYPE_PADDING: + /* failed writes may be discarded events */ + if (!event->time_delta) + KILL_TEST(); + inc = event->array[0] + 4; + break; + case RINGBUF_TYPE_TIME_EXTEND: + inc = 8; + break; + case 0: + entry = ring_buffer_event_data(event); + if (*entry != cpu) { + KILL_TEST(); + break; + } + read++; + if (!event->array[0]) { + KILL_TEST(); + break; + } + inc = event->array[0] + 4; + break; + default: + entry = ring_buffer_event_data(event); + if (*entry != cpu) { + KILL_TEST(); + break; + } + read++; + inc = ((event->type_len + 1) * 4); + } + if (kill_test) + break; + + if (inc <= 0) { + KILL_TEST(); + break; + } + } + } + ring_buffer_free_read_page(buffer, bpage); + + if (ret < 0) + return EVENT_DROPPED; + return EVENT_FOUND; +} + +static void ring_buffer_consumer(void) +{ + /* toggle between reading pages and events */ + read_events ^= 1; + + read = 0; + while (!reader_finish && !kill_test) { + int found; + + do { + int cpu; + + found = 0; + for_each_online_cpu(cpu) { + enum event_status stat; + + if (read_events) + stat = read_event(cpu); + else + stat = read_page(cpu); + + if (kill_test) + break; + if (stat == EVENT_FOUND) + found = 1; + } + } while (found && !kill_test); + + set_current_state(TASK_INTERRUPTIBLE); + if (reader_finish) + break; + + schedule(); + __set_current_state(TASK_RUNNING); + } + reader_finish = 0; + complete(&read_done); +} + +static void ring_buffer_producer(void) +{ + struct timeval start_tv; + struct timeval end_tv; + unsigned long long time; + unsigned long long entries; + unsigned long long overruns; + unsigned long missed = 0; + unsigned long hit = 0; + unsigned long avg; + int cnt = 0; + + /* + * Hammer the buffer for 10 secs (this may + * make the system stall) + */ + trace_printk("Starting ring buffer hammer\n"); + do_gettimeofday(&start_tv); + do { + struct ring_buffer_event *event; + int *entry; + + event = ring_buffer_lock_reserve(buffer, 10); + if (!event) { + missed++; + } else { + hit++; + entry = ring_buffer_event_data(event); + *entry = smp_processor_id(); + ring_buffer_unlock_commit(buffer, event); + } + do_gettimeofday(&end_tv); + + cnt++; + if (consumer && !(cnt % wakeup_interval)) + wake_up_process(consumer); + +#ifndef CONFIG_PREEMPT + /* + * If we are a non preempt kernel, the 10 second run will + * stop everything while it runs. Instead, we will call + * cond_resched and also add any time that was lost by a + * rescedule. + * + * Do a cond resched at the same frequency we would wake up + * the reader. + */ + if (cnt % wakeup_interval) + cond_resched(); +#endif + + } while (end_tv.tv_sec < (start_tv.tv_sec + RUN_TIME) && !kill_test); + trace_printk("End ring buffer hammer\n"); + + if (consumer) { + /* Init both completions here to avoid races */ + init_completion(&read_start); + init_completion(&read_done); + /* the completions must be visible before the finish var */ + smp_wmb(); + reader_finish = 1; + /* finish var visible before waking up the consumer */ + smp_wmb(); + wake_up_process(consumer); + wait_for_completion(&read_done); + } + + time = end_tv.tv_sec - start_tv.tv_sec; + time *= USEC_PER_SEC; + time += (long long)((long)end_tv.tv_usec - (long)start_tv.tv_usec); + + entries = ring_buffer_entries(buffer); + overruns = ring_buffer_overruns(buffer); + + if (kill_test) + trace_printk("ERROR!\n"); + trace_printk("Time: %lld (usecs)\n", time); + trace_printk("Overruns: %lld\n", overruns); + if (disable_reader) + trace_printk("Read: (reader disabled)\n"); + else + trace_printk("Read: %ld (by %s)\n", read, + read_events ? "events" : "pages"); + trace_printk("Entries: %lld\n", entries); + trace_printk("Total: %lld\n", entries + overruns + read); + trace_printk("Missed: %ld\n", missed); + trace_printk("Hit: %ld\n", hit); + + /* Convert time from usecs to millisecs */ + do_div(time, USEC_PER_MSEC); + if (time) + hit /= (long)time; + else + trace_printk("TIME IS ZERO??\n"); + + trace_printk("Entries per millisec: %ld\n", hit); + + if (hit) { + /* Calculate the average time in nanosecs */ + avg = NSEC_PER_MSEC / hit; + trace_printk("%ld ns per entry\n", avg); + } + + if (missed) { + if (time) + missed /= (long)time; + + trace_printk("Total iterations per millisec: %ld\n", + hit + missed); + + /* it is possible that hit + missed will overflow and be zero */ + if (!(hit + missed)) { + trace_printk("hit + missed overflowed and totalled zero!\n"); + hit--; /* make it non zero */ + } + + /* Caculate the average time in nanosecs */ + avg = NSEC_PER_MSEC / (hit + missed); + trace_printk("%ld ns per entry\n", avg); + } +} + +static void wait_to_die(void) +{ + set_current_state(TASK_INTERRUPTIBLE); + while (!kthread_should_stop()) { + schedule(); + set_current_state(TASK_INTERRUPTIBLE); + } + __set_current_state(TASK_RUNNING); +} + +static int ring_buffer_consumer_thread(void *arg) +{ + while (!kthread_should_stop() && !kill_test) { + complete(&read_start); + + ring_buffer_consumer(); + + set_current_state(TASK_INTERRUPTIBLE); + if (kthread_should_stop() || kill_test) + break; + + schedule(); + __set_current_state(TASK_RUNNING); + } + __set_current_state(TASK_RUNNING); + + if (kill_test) + wait_to_die(); + + return 0; +} + +static int ring_buffer_producer_thread(void *arg) +{ + init_completion(&read_start); + + while (!kthread_should_stop() && !kill_test) { + ring_buffer_reset(buffer); + + if (consumer) { + smp_wmb(); + wake_up_process(consumer); + wait_for_completion(&read_start); + } + + ring_buffer_producer(); + + trace_printk("Sleeping for 10 secs\n"); + set_current_state(TASK_INTERRUPTIBLE); + schedule_timeout(HZ * SLEEP_TIME); + __set_current_state(TASK_RUNNING); + } + + if (kill_test) + wait_to_die(); + + return 0; +} + +static int __init ring_buffer_benchmark_init(void) +{ + int ret; + + /* make a one meg buffer in overwite mode */ + buffer = ring_buffer_alloc(1000000, RB_FL_OVERWRITE); + if (!buffer) + return -ENOMEM; + + if (!disable_reader) { + consumer = kthread_create(ring_buffer_consumer_thread, + NULL, "rb_consumer"); + ret = PTR_ERR(consumer); + if (IS_ERR(consumer)) + goto out_fail; + } + + producer = kthread_run(ring_buffer_producer_thread, + NULL, "rb_producer"); + ret = PTR_ERR(producer); + + if (IS_ERR(producer)) + goto out_kill; + + return 0; + + out_kill: + if (consumer) + kthread_stop(consumer); + + out_fail: + ring_buffer_free(buffer); + return ret; +} + +static void __exit ring_buffer_benchmark_exit(void) +{ + kthread_stop(producer); + if (consumer) + kthread_stop(consumer); + ring_buffer_free(buffer); +} + +module_init(ring_buffer_benchmark_init); +module_exit(ring_buffer_benchmark_exit); + +MODULE_AUTHOR("Steven Rostedt"); +MODULE_DESCRIPTION("ring_buffer_benchmark"); +MODULE_LICENSE("GPL"); diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c580233add9..5c75deeefe3 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -11,47 +11,58 @@ * Copyright (C) 2004-2006 Ingo Molnar * Copyright (C) 2004 William Lee Irwin III */ +#include <linux/ring_buffer.h> #include <linux/utsrelease.h> +#include <linux/stacktrace.h> +#include <linux/writeback.h> #include <linux/kallsyms.h> #include <linux/seq_file.h> +#include <linux/smp_lock.h> #include <linux/notifier.h> +#include <linux/irqflags.h> #include <linux/debugfs.h> #include <linux/pagemap.h> #include <linux/hardirq.h> #include <linux/linkage.h> #include <linux/uaccess.h> +#include <linux/kprobes.h> #include <linux/ftrace.h> #include <linux/module.h> #include <linux/percpu.h> +#include <linux/splice.h> #include <linux/kdebug.h> +#include <linux/string.h> #include <linux/ctype.h> #include <linux/init.h> #include <linux/poll.h> #include <linux/gfp.h> #include <linux/fs.h> -#include <linux/kprobes.h> -#include <linux/writeback.h> - -#include <linux/stacktrace.h> -#include <linux/ring_buffer.h> -#include <linux/irqflags.h> #include "trace.h" +#include "trace_output.h" #define TRACE_BUFFER_FLAGS (RB_FL_OVERWRITE) -unsigned long __read_mostly tracing_max_latency = (cycle_t)ULONG_MAX; -unsigned long __read_mostly tracing_thresh; +/* + * On boot up, the ring buffer is set to the minimum size, so that + * we do not waste memory on systems that are not using tracing. + */ +int ring_buffer_expanded; /* * We need to change this state when a selftest is running. * A selftest will lurk into the ring-buffer to count the * entries inserted during the selftest although some concurrent - * insertions into the ring-buffer such as ftrace_printk could occurred + * insertions into the ring-buffer such as trace_printk could occurred * at the same time, giving false positive or negative results. */ static bool __read_mostly tracing_selftest_running; +/* + * If a tracer is running, we do not want to run SELFTEST. + */ +bool __read_mostly tracing_selftest_disabled; + /* For tracers that don't implement custom flags */ static struct tracer_opt dummy_tracer_opt[] = { { } @@ -73,9 +84,9 @@ static int dummy_set_flag(u32 old_flags, u32 bit, int set) * of the tracer is successful. But that is the only place that sets * this back to zero. */ -int tracing_disabled = 1; +static int tracing_disabled = 1; -static DEFINE_PER_CPU(local_t, ftrace_cpu_disabled); +DEFINE_PER_CPU(local_t, ftrace_cpu_disabled); static inline void ftrace_disable_cpu(void) { @@ -91,6 +102,9 @@ static inline void ftrace_enable_cpu(void) static cpumask_var_t __read_mostly tracing_buffer_mask; +/* Define which cpu buffers are currently read in trace_pipe */ +static cpumask_var_t tracing_reader_cpumask; + #define for_each_tracing_cpu(cpu) \ for_each_cpu(cpu, tracing_buffer_mask) @@ -109,14 +123,21 @@ static cpumask_var_t __read_mostly tracing_buffer_mask; */ int ftrace_dump_on_oops; -static int tracing_set_tracer(char *buf); +static int tracing_set_tracer(const char *buf); + +#define BOOTUP_TRACER_SIZE 100 +static char bootup_tracer_buf[BOOTUP_TRACER_SIZE] __initdata; +static char *default_bootup_tracer; static int __init set_ftrace(char *str) { - tracing_set_tracer(str); + strncpy(bootup_tracer_buf, str, BOOTUP_TRACER_SIZE); + default_bootup_tracer = bootup_tracer_buf; + /* We are using ftrace early, expand it */ + ring_buffer_expanded = 1; return 1; } -__setup("ftrace", set_ftrace); +__setup("ftrace=", set_ftrace); static int __init set_ftrace_dump_on_oops(char *str) { @@ -125,21 +146,13 @@ static int __init set_ftrace_dump_on_oops(char *str) } __setup("ftrace_dump_on_oops", set_ftrace_dump_on_oops); -long -ns2usecs(cycle_t nsec) +unsigned long long ns2usecs(cycle_t nsec) { nsec += 500; do_div(nsec, 1000); return nsec; } -cycle_t ftrace_now(int cpu) -{ - u64 ts = ring_buffer_time_stamp(cpu); - ring_buffer_normalize_time_stamp(cpu, &ts); - return ts; -} - /* * The global_trace is the descriptor that holds the tracing * buffers for the live tracing. For each CPU, it contains @@ -156,6 +169,28 @@ static struct trace_array global_trace; static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu); +int filter_current_check_discard(struct ring_buffer *buffer, + struct ftrace_event_call *call, void *rec, + struct ring_buffer_event *event) +{ + return filter_check_discard(call, rec, buffer, event); +} +EXPORT_SYMBOL_GPL(filter_current_check_discard); + +cycle_t ftrace_now(int cpu) +{ + u64 ts; + + /* Early boot up does not have a buffer yet */ + if (!global_trace.buffer) + return trace_clock_local(); + + ts = ring_buffer_time_stamp(global_trace.buffer, cpu); + ring_buffer_normalize_time_stamp(global_trace.buffer, cpu, &ts); + + return ts; +} + /* * The max_tr is used to snapshot the global_trace when a maximum * latency is reached. Some tracers will use this to store a maximum @@ -186,9 +221,6 @@ int tracing_is_enabled(void) return tracer_enabled; } -/* function tracing enabled */ -int ftrace_function_enabled; - /* * trace_buf_size is the size in bytes that is allocated * for a buffer. Note, the number of bytes is always rounded @@ -229,7 +261,11 @@ static DECLARE_WAIT_QUEUE_HEAD(trace_wait); /* trace_flags holds trace_options default values */ unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK | - TRACE_ITER_ANNOTATE; + TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO | TRACE_ITER_SLEEP_TIME | + TRACE_ITER_GRAPH_TIME; + +static int trace_stop_count; +static DEFINE_SPINLOCK(tracing_start_lock); /** * trace_wake_up - wake up tasks waiting for trace input @@ -250,13 +286,12 @@ void trace_wake_up(void) static int __init set_buf_size(char *str) { unsigned long buf_size; - int ret; if (!str) return 0; - ret = strict_strtoul(str, 0, &buf_size); + buf_size = memparse(str, &str); /* nr_entries can not be zero */ - if (ret < 0 || buf_size == 0) + if (buf_size == 0) return 0; trace_buf_size = buf_size; return 1; @@ -280,192 +315,58 @@ static const char *trace_options[] = { "block", "stacktrace", "sched-tree", - "ftrace_printk", + "trace_printk", "ftrace_preempt", "branch", "annotate", "userstacktrace", "sym-userobj", "printk-msg-only", + "context-info", + "latency-format", + "sleep-time", + "graph-time", NULL }; -/* - * ftrace_max_lock is used to protect the swapping of buffers - * when taking a max snapshot. The buffers themselves are - * protected by per_cpu spinlocks. But the action of the swap - * needs its own lock. - * - * This is defined as a raw_spinlock_t in order to help - * with performance when lockdep debugging is enabled. - */ -static raw_spinlock_t ftrace_max_lock = - (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; - -/* - * Copy the new maximum trace into the separate maximum-trace - * structure. (this way the maximum trace is permanently saved, - * for later retrieval via /debugfs/tracing/latency_trace) - */ -static void -__update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) -{ - struct trace_array_cpu *data = tr->data[cpu]; - - max_tr.cpu = cpu; - max_tr.time_start = data->preempt_timestamp; - - data = max_tr.data[cpu]; - data->saved_latency = tracing_max_latency; - - memcpy(data->comm, tsk->comm, TASK_COMM_LEN); - data->pid = tsk->pid; - data->uid = task_uid(tsk); - data->nice = tsk->static_prio - 20 - MAX_RT_PRIO; - data->policy = tsk->policy; - data->rt_priority = tsk->rt_priority; +static struct { + u64 (*func)(void); + const char *name; +} trace_clocks[] = { + { trace_clock_local, "local" }, + { trace_clock_global, "global" }, +}; - /* record this tasks comm */ - tracing_record_cmdline(current); -} +int trace_clock_id; -/** - * trace_seq_printf - sequence printing of trace information - * @s: trace sequence descriptor - * @fmt: printf format string - * - * The tracer may use either sequence operations or its own - * copy to user routines. To simplify formating of a trace - * trace_seq_printf is used to store strings into a special - * buffer (@s). Then the output may be either used by - * the sequencer or pulled into another buffer. - */ -int -trace_seq_printf(struct trace_seq *s, const char *fmt, ...) +ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) { - int len = (PAGE_SIZE - 1) - s->len; - va_list ap; + int len; int ret; - if (!len) - return 0; - - va_start(ap, fmt); - ret = vsnprintf(s->buffer + s->len, len, fmt, ap); - va_end(ap); - - /* If we can't write it all, don't bother writing anything */ - if (ret >= len) - return 0; - - s->len += ret; - - return len; -} - -/** - * trace_seq_puts - trace sequence printing of simple string - * @s: trace sequence descriptor - * @str: simple string to record - * - * The tracer may use either the sequence operations or its own - * copy to user routines. This function records a simple string - * into a special buffer (@s) for later retrieval by a sequencer - * or other mechanism. - */ -static int -trace_seq_puts(struct trace_seq *s, const char *str) -{ - int len = strlen(str); - - if (len > ((PAGE_SIZE - 1) - s->len)) - return 0; - - memcpy(s->buffer + s->len, str, len); - s->len += len; - - return len; -} - -static int -trace_seq_putc(struct trace_seq *s, unsigned char c) -{ - if (s->len >= (PAGE_SIZE - 1)) - return 0; - - s->buffer[s->len++] = c; - - return 1; -} - -static int -trace_seq_putmem(struct trace_seq *s, void *mem, size_t len) -{ - if (len > ((PAGE_SIZE - 1) - s->len)) + if (!cnt) return 0; - memcpy(s->buffer + s->len, mem, len); - s->len += len; - - return len; -} - -#define MAX_MEMHEX_BYTES 8 -#define HEX_CHARS (MAX_MEMHEX_BYTES*2 + 1) - -static int -trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len) -{ - unsigned char hex[HEX_CHARS]; - unsigned char *data = mem; - int i, j; - -#ifdef __BIG_ENDIAN - for (i = 0, j = 0; i < len; i++) { -#else - for (i = len-1, j = 0; i >= 0; i--) { -#endif - hex[j++] = hex_asc_hi(data[i]); - hex[j++] = hex_asc_lo(data[i]); - } - hex[j++] = ' '; - - return trace_seq_putmem(s, hex, j); -} - -static int -trace_seq_path(struct trace_seq *s, struct path *path) -{ - unsigned char *p; + if (s->len <= s->readpos) + return -EBUSY; - if (s->len >= (PAGE_SIZE - 1)) - return 0; - p = d_path(path, s->buffer + s->len, PAGE_SIZE - s->len); - if (!IS_ERR(p)) { - p = mangle_path(s->buffer + s->len, p, "\n"); - if (p) { - s->len = p - s->buffer; - return 1; - } - } else { - s->buffer[s->len++] = '?'; - return 1; - } + len = s->len - s->readpos; + if (cnt > len) + cnt = len; + ret = copy_to_user(ubuf, s->buffer + s->readpos, cnt); + if (ret == cnt) + return -EFAULT; - return 0; -} + cnt -= ret; -static void -trace_seq_reset(struct trace_seq *s) -{ - s->len = 0; - s->readpos = 0; + s->readpos += cnt; + return cnt; } -ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) +static ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt) { int len; - int ret; + void *ret; if (s->len <= s->readpos) return -EBUSY; @@ -473,23 +374,62 @@ ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) len = s->len - s->readpos; if (cnt > len) cnt = len; - ret = copy_to_user(ubuf, s->buffer + s->readpos, cnt); - if (ret) + ret = memcpy(buf, s->buffer + s->readpos, cnt); + if (!ret) return -EFAULT; - s->readpos += len; + s->readpos += cnt; return cnt; } +/* + * ftrace_max_lock is used to protect the swapping of buffers + * when taking a max snapshot. The buffers themselves are + * protected by per_cpu spinlocks. But the action of the swap + * needs its own lock. + * + * This is defined as a raw_spinlock_t in order to help + * with performance when lockdep debugging is enabled. + * + * It is also used in other places outside the update_max_tr + * so it needs to be defined outside of the + * CONFIG_TRACER_MAX_TRACE. + */ +static raw_spinlock_t ftrace_max_lock = + (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; + +#ifdef CONFIG_TRACER_MAX_TRACE +unsigned long __read_mostly tracing_max_latency; +unsigned long __read_mostly tracing_thresh; + +/* + * Copy the new maximum trace into the separate maximum-trace + * structure. (this way the maximum trace is permanently saved, + * for later retrieval via /sys/kernel/debug/tracing/latency_trace) + */ static void -trace_print_seq(struct seq_file *m, struct trace_seq *s) +__update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) { - int len = s->len >= PAGE_SIZE ? PAGE_SIZE - 1 : s->len; + struct trace_array_cpu *data = tr->data[cpu]; + struct trace_array_cpu *max_data = tr->data[cpu]; - s->buffer[len] = 0; - seq_puts(m, s->buffer); + max_tr.cpu = cpu; + max_tr.time_start = data->preempt_timestamp; + + max_data = max_tr.data[cpu]; + max_data->saved_latency = tracing_max_latency; + max_data->critical_start = data->critical_start; + max_data->critical_end = data->critical_end; - trace_seq_reset(s); + memcpy(data->comm, tsk->comm, TASK_COMM_LEN); + max_data->pid = tsk->pid; + max_data->uid = task_uid(tsk); + max_data->nice = tsk->static_prio - 20 - MAX_RT_PRIO; + max_data->policy = tsk->policy; + max_data->rt_priority = tsk->rt_priority; + + /* record this tasks comm */ + tracing_record_cmdline(tsk); } /** @@ -506,16 +446,15 @@ update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) { struct ring_buffer *buf = tr->buffer; + if (trace_stop_count) + return; + WARN_ON_ONCE(!irqs_disabled()); __raw_spin_lock(&ftrace_max_lock); tr->buffer = max_tr.buffer; max_tr.buffer = buf; - ftrace_disable_cpu(); - ring_buffer_reset(tr->buffer); - ftrace_enable_cpu(); - __update_max_tr(tr, tsk, cpu); __raw_spin_unlock(&ftrace_max_lock); } @@ -533,21 +472,35 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu) { int ret; + if (trace_stop_count) + return; + WARN_ON_ONCE(!irqs_disabled()); __raw_spin_lock(&ftrace_max_lock); ftrace_disable_cpu(); - ring_buffer_reset(max_tr.buffer); ret = ring_buffer_swap_cpu(max_tr.buffer, tr->buffer, cpu); + if (ret == -EBUSY) { + /* + * We failed to swap the buffer due to a commit taking + * place on this CPU. We fail to record, but we reset + * the max trace buffer (no one writes directly to it) + * and flag that it failed. + */ + trace_array_printk(&max_tr, _THIS_IP_, + "Failed to swap buffers due to commit in progress\n"); + } + ftrace_enable_cpu(); - WARN_ON_ONCE(ret); + WARN_ON_ONCE(ret && ret != -EAGAIN && ret != -EBUSY); __update_max_tr(tr, tsk, cpu); __raw_spin_unlock(&ftrace_max_lock); } +#endif /* CONFIG_TRACER_MAX_TRACE */ /** * register_tracer - register a tracer with the ftrace system. @@ -556,6 +509,8 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu) * Register a new plugin tracer. */ int register_tracer(struct tracer *type) +__releases(kernel_lock) +__acquires(kernel_lock) { struct tracer *t; int len; @@ -594,12 +549,14 @@ int register_tracer(struct tracer *type) else if (!type->flags->opts) type->flags->opts = dummy_tracer_opt; + if (!type->wait_pipe) + type->wait_pipe = default_wait_pipe; + #ifdef CONFIG_FTRACE_STARTUP_TEST - if (type->selftest) { + if (type->selftest && !tracing_selftest_disabled) { struct tracer *saved_tracer = current_trace; struct trace_array *tr = &global_trace; - int i; /* * Run a selftest on this tracer. @@ -608,8 +565,7 @@ int register_tracer(struct tracer *type) * internal tracing to verify that everything is in order. * If we fail, we do not register this tracer. */ - for_each_tracing_cpu(i) - tracing_reset(tr, i); + tracing_reset_online_cpus(tr); current_trace = type; /* the test is responsible for initializing and enabling */ @@ -622,8 +578,7 @@ int register_tracer(struct tracer *type) goto out; } /* Only reset on passing, to avoid touching corrupted buffers */ - for_each_tracing_cpu(i) - tracing_reset(tr, i); + tracing_reset_online_cpus(tr); printk(KERN_CONT "PASSED\n"); } @@ -638,8 +593,26 @@ int register_tracer(struct tracer *type) out: tracing_selftest_running = false; mutex_unlock(&trace_types_lock); - lock_kernel(); + if (ret || !default_bootup_tracer) + goto out_unlock; + + if (strncmp(default_bootup_tracer, type->name, BOOTUP_TRACER_SIZE)) + goto out_unlock; + + printk(KERN_INFO "Starting tracer '%s'\n", type->name); + /* Do we want this tracer to start on bootup? */ + tracing_set_tracer(type->name); + default_bootup_tracer = NULL; + /* disable other selftests, since this will break it. */ + tracing_selftest_disabled = 1; +#ifdef CONFIG_FTRACE_STARTUP_TEST + printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n", + type->name); +#endif + + out_unlock: + lock_kernel(); return ret; } @@ -658,6 +631,15 @@ void unregister_tracer(struct tracer *type) found: *t = (*t)->next; + + if (type == current_trace && tracer_enabled) { + tracer_enabled = 0; + tracing_stop(); + if (current_trace->stop) + current_trace->stop(&global_trace); + current_trace = &nop_trace; + } + if (strlen(type->name) != max_tracer_type_len) goto out; @@ -671,43 +653,72 @@ void unregister_tracer(struct tracer *type) mutex_unlock(&trace_types_lock); } -void tracing_reset(struct trace_array *tr, int cpu) +static void __tracing_reset(struct trace_array *tr, int cpu) { ftrace_disable_cpu(); ring_buffer_reset_cpu(tr->buffer, cpu); ftrace_enable_cpu(); } +void tracing_reset(struct trace_array *tr, int cpu) +{ + struct ring_buffer *buffer = tr->buffer; + + ring_buffer_record_disable(buffer); + + /* Make sure all commits have finished */ + synchronize_sched(); + __tracing_reset(tr, cpu); + + ring_buffer_record_enable(buffer); +} + void tracing_reset_online_cpus(struct trace_array *tr) { + struct ring_buffer *buffer = tr->buffer; int cpu; + ring_buffer_record_disable(buffer); + + /* Make sure all commits have finished */ + synchronize_sched(); + tr->time_start = ftrace_now(tr->cpu); for_each_online_cpu(cpu) - tracing_reset(tr, cpu); + __tracing_reset(tr, cpu); + + ring_buffer_record_enable(buffer); +} + +void tracing_reset_current(int cpu) +{ + tracing_reset(&global_trace, cpu); +} + +void tracing_reset_current_online_cpus(void) +{ + tracing_reset_online_cpus(&global_trace); } #define SAVED_CMDLINES 128 +#define NO_CMDLINE_MAP UINT_MAX static unsigned map_pid_to_cmdline[PID_MAX_DEFAULT+1]; static unsigned map_cmdline_to_pid[SAVED_CMDLINES]; static char saved_cmdlines[SAVED_CMDLINES][TASK_COMM_LEN]; static int cmdline_idx; -static DEFINE_SPINLOCK(trace_cmdline_lock); +static raw_spinlock_t trace_cmdline_lock = __RAW_SPIN_LOCK_UNLOCKED; /* temporary disable recording */ -atomic_t trace_record_cmdline_disabled __read_mostly; +static atomic_t trace_record_cmdline_disabled __read_mostly; static void trace_init_cmdlines(void) { - memset(&map_pid_to_cmdline, -1, sizeof(map_pid_to_cmdline)); - memset(&map_cmdline_to_pid, -1, sizeof(map_cmdline_to_pid)); + memset(&map_pid_to_cmdline, NO_CMDLINE_MAP, sizeof(map_pid_to_cmdline)); + memset(&map_cmdline_to_pid, NO_CMDLINE_MAP, sizeof(map_cmdline_to_pid)); cmdline_idx = 0; } -static int trace_stop_count; -static DEFINE_SPINLOCK(tracing_start_lock); - /** * ftrace_off_permanent - disable all ftrace code permanently * @@ -738,13 +749,12 @@ void tracing_start(void) return; spin_lock_irqsave(&tracing_start_lock, flags); - if (--trace_stop_count) - goto out; - - if (trace_stop_count < 0) { - /* Someone screwed up their debugging */ - WARN_ON_ONCE(1); - trace_stop_count = 0; + if (--trace_stop_count) { + if (trace_stop_count < 0) { + /* Someone screwed up their debugging */ + WARN_ON_ONCE(1); + trace_stop_count = 0; + } goto out; } @@ -794,8 +804,7 @@ void trace_stop_cmdline_recording(void); static void trace_save_cmdline(struct task_struct *tsk) { - unsigned map; - unsigned idx; + unsigned pid, idx; if (!tsk->pid || unlikely(tsk->pid > PID_MAX_DEFAULT)) return; @@ -806,17 +815,24 @@ static void trace_save_cmdline(struct task_struct *tsk) * nor do we want to disable interrupts, * so if we miss here, then better luck next time. */ - if (!spin_trylock(&trace_cmdline_lock)) + if (!__raw_spin_trylock(&trace_cmdline_lock)) return; idx = map_pid_to_cmdline[tsk->pid]; - if (idx >= SAVED_CMDLINES) { + if (idx == NO_CMDLINE_MAP) { idx = (cmdline_idx + 1) % SAVED_CMDLINES; - map = map_cmdline_to_pid[idx]; - if (map <= PID_MAX_DEFAULT) - map_pid_to_cmdline[map] = (unsigned)-1; + /* + * Check whether the cmdline buffer at idx has a pid + * mapped. We are going to overwrite that entry so we + * need to clear the map_pid_to_cmdline. Otherwise we + * would read the new comm for the old pid. + */ + pid = map_cmdline_to_pid[idx]; + if (pid != NO_CMDLINE_MAP) + map_pid_to_cmdline[pid] = NO_CMDLINE_MAP; + map_cmdline_to_pid[idx] = tsk->pid; map_pid_to_cmdline[tsk->pid] = idx; cmdline_idx = idx; @@ -824,33 +840,39 @@ static void trace_save_cmdline(struct task_struct *tsk) memcpy(&saved_cmdlines[idx], tsk->comm, TASK_COMM_LEN); - spin_unlock(&trace_cmdline_lock); + __raw_spin_unlock(&trace_cmdline_lock); } -char *trace_find_cmdline(int pid) +void trace_find_cmdline(int pid, char comm[]) { - char *cmdline = "<...>"; unsigned map; - if (!pid) - return "<idle>"; + if (!pid) { + strcpy(comm, "<idle>"); + return; + } - if (pid > PID_MAX_DEFAULT) - goto out; + if (pid > PID_MAX_DEFAULT) { + strcpy(comm, "<...>"); + return; + } + preempt_disable(); + __raw_spin_lock(&trace_cmdline_lock); map = map_pid_to_cmdline[pid]; - if (map >= SAVED_CMDLINES) - goto out; - - cmdline = saved_cmdlines[map]; + if (map != NO_CMDLINE_MAP) + strcpy(comm, saved_cmdlines[map]); + else + strcpy(comm, "<...>"); - out: - return cmdline; + __raw_spin_unlock(&trace_cmdline_lock); + preempt_enable(); } void tracing_record_cmdline(struct task_struct *tsk) { - if (atomic_read(&trace_record_cmdline_disabled)) + if (atomic_read(&trace_record_cmdline_disabled) || !tracer_enabled || + !tracing_is_on()) return; trace_save_cmdline(tsk); @@ -864,7 +886,7 @@ tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags, entry->preempt_count = pc & 0xff; entry->pid = (tsk) ? tsk->pid : 0; - entry->tgid = (tsk) ? tsk->tgid : 0; + entry->tgid = (tsk) ? tsk->tgid : 0; entry->flags = #ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT (irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) | @@ -875,81 +897,108 @@ tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags, ((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) | (need_resched() ? TRACE_FLAG_NEED_RESCHED : 0); } +EXPORT_SYMBOL_GPL(tracing_generic_entry_update); -void -trace_function(struct trace_array *tr, struct trace_array_cpu *data, - unsigned long ip, unsigned long parent_ip, unsigned long flags, - int pc) +struct ring_buffer_event * +trace_buffer_lock_reserve(struct ring_buffer *buffer, + int type, + unsigned long len, + unsigned long flags, int pc) { struct ring_buffer_event *event; - struct ftrace_entry *entry; - unsigned long irq_flags; - /* If we are reading the ring buffer, don't trace */ - if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) - return; + event = ring_buffer_lock_reserve(buffer, len); + if (event != NULL) { + struct trace_entry *ent = ring_buffer_event_data(event); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); - if (!event) - return; - entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_FN; - entry->ip = ip; - entry->parent_ip = parent_ip; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + tracing_generic_entry_update(ent, flags, pc); + ent->type = type; + } + + return event; } -#ifdef CONFIG_FUNCTION_GRAPH_TRACER -static void __trace_graph_entry(struct trace_array *tr, - struct trace_array_cpu *data, - struct ftrace_graph_ent *trace, - unsigned long flags, - int pc) +static inline void +__trace_buffer_unlock_commit(struct ring_buffer *buffer, + struct ring_buffer_event *event, + unsigned long flags, int pc, + int wake) { - struct ring_buffer_event *event; - struct ftrace_graph_ent_entry *entry; - unsigned long irq_flags; + ring_buffer_unlock_commit(buffer, event); - if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) - return; + ftrace_trace_stack(buffer, flags, 6, pc); + ftrace_trace_userstack(buffer, flags, pc); - event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry), - &irq_flags); - if (!event) - return; - entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_GRAPH_ENT; - entry->graph_ent = *trace; - ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags); + if (wake) + trace_wake_up(); +} + +void trace_buffer_unlock_commit(struct ring_buffer *buffer, + struct ring_buffer_event *event, + unsigned long flags, int pc) +{ + __trace_buffer_unlock_commit(buffer, event, flags, pc, 1); } -static void __trace_graph_return(struct trace_array *tr, - struct trace_array_cpu *data, - struct ftrace_graph_ret *trace, - unsigned long flags, - int pc) +struct ring_buffer_event * +trace_current_buffer_lock_reserve(struct ring_buffer **current_rb, + int type, unsigned long len, + unsigned long flags, int pc) { + *current_rb = global_trace.buffer; + return trace_buffer_lock_reserve(*current_rb, + type, len, flags, pc); +} +EXPORT_SYMBOL_GPL(trace_current_buffer_lock_reserve); + +void trace_current_buffer_unlock_commit(struct ring_buffer *buffer, + struct ring_buffer_event *event, + unsigned long flags, int pc) +{ + __trace_buffer_unlock_commit(buffer, event, flags, pc, 1); +} +EXPORT_SYMBOL_GPL(trace_current_buffer_unlock_commit); + +void trace_nowake_buffer_unlock_commit(struct ring_buffer *buffer, + struct ring_buffer_event *event, + unsigned long flags, int pc) +{ + __trace_buffer_unlock_commit(buffer, event, flags, pc, 0); +} +EXPORT_SYMBOL_GPL(trace_nowake_buffer_unlock_commit); + +void trace_current_buffer_discard_commit(struct ring_buffer *buffer, + struct ring_buffer_event *event) +{ + ring_buffer_discard_commit(buffer, event); +} +EXPORT_SYMBOL_GPL(trace_current_buffer_discard_commit); + +void +trace_function(struct trace_array *tr, + unsigned long ip, unsigned long parent_ip, unsigned long flags, + int pc) +{ + struct ftrace_event_call *call = &event_function; + struct ring_buffer *buffer = tr->buffer; struct ring_buffer_event *event; - struct ftrace_graph_ret_entry *entry; - unsigned long irq_flags; + struct ftrace_entry *entry; + /* If we are reading the ring buffer, don't trace */ if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) return; - event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry), - &irq_flags); + event = trace_buffer_lock_reserve(buffer, TRACE_FN, sizeof(*entry), + flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_GRAPH_RET; - entry->ret = *trace; - ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags); + entry->ip = ip; + entry->parent_ip = parent_ip; + + if (!filter_check_discard(call, entry, buffer, event)) + ring_buffer_unlock_commit(buffer, event); } -#endif void ftrace(struct trace_array *tr, struct trace_array_cpu *data, @@ -957,31 +1006,24 @@ ftrace(struct trace_array *tr, struct trace_array_cpu *data, int pc) { if (likely(!atomic_read(&data->disabled))) - trace_function(tr, data, ip, parent_ip, flags, pc); + trace_function(tr, ip, parent_ip, flags, pc); } -static void ftrace_trace_stack(struct trace_array *tr, - struct trace_array_cpu *data, - unsigned long flags, - int skip, int pc) -{ #ifdef CONFIG_STACKTRACE +static void __ftrace_trace_stack(struct ring_buffer *buffer, + unsigned long flags, + int skip, int pc) +{ + struct ftrace_event_call *call = &event_kernel_stack; struct ring_buffer_event *event; struct stack_entry *entry; struct stack_trace trace; - unsigned long irq_flags; - - if (!(trace_flags & TRACE_ITER_STACKTRACE)) - return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = trace_buffer_lock_reserve(buffer, TRACE_STACK, + sizeof(*entry), flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_STACK; - memset(&entry->caller, 0, sizeof(entry->caller)); trace.nr_entries = 0; @@ -990,38 +1032,41 @@ static void ftrace_trace_stack(struct trace_array *tr, trace.entries = entry->caller; save_stack_trace(&trace); - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); -#endif + if (!filter_check_discard(call, entry, buffer, event)) + ring_buffer_unlock_commit(buffer, event); } -void __trace_stack(struct trace_array *tr, - struct trace_array_cpu *data, - unsigned long flags, - int skip) +void ftrace_trace_stack(struct ring_buffer *buffer, unsigned long flags, + int skip, int pc) { - ftrace_trace_stack(tr, data, flags, skip, preempt_count()); + if (!(trace_flags & TRACE_ITER_STACKTRACE)) + return; + + __ftrace_trace_stack(buffer, flags, skip, pc); } -static void ftrace_trace_userstack(struct trace_array *tr, - struct trace_array_cpu *data, - unsigned long flags, int pc) +void __trace_stack(struct trace_array *tr, unsigned long flags, int skip, + int pc) { -#ifdef CONFIG_STACKTRACE + __ftrace_trace_stack(tr->buffer, flags, skip, pc); +} + +void +ftrace_trace_userstack(struct ring_buffer *buffer, unsigned long flags, int pc) +{ + struct ftrace_event_call *call = &event_user_stack; struct ring_buffer_event *event; struct userstack_entry *entry; struct stack_trace trace; - unsigned long irq_flags; if (!(trace_flags & TRACE_ITER_USERSTACKTRACE)) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = trace_buffer_lock_reserve(buffer, TRACE_USER_STACK, + sizeof(*entry), flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_USER_STACK; memset(&entry->caller, 0, sizeof(entry->caller)); @@ -1031,112 +1076,45 @@ static void ftrace_trace_userstack(struct trace_array *tr, trace.entries = entry->caller; save_stack_trace_user(&trace); - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); -#endif + if (!filter_check_discard(call, entry, buffer, event)) + ring_buffer_unlock_commit(buffer, event); } -void __trace_userstack(struct trace_array *tr, - struct trace_array_cpu *data, - unsigned long flags) +#ifdef UNUSED +static void __trace_userstack(struct trace_array *tr, unsigned long flags) { - ftrace_trace_userstack(tr, data, flags, preempt_count()); + ftrace_trace_userstack(tr, flags, preempt_count()); } +#endif /* UNUSED */ + +#endif /* CONFIG_STACKTRACE */ static void -ftrace_trace_special(void *__tr, void *__data, +ftrace_trace_special(void *__tr, unsigned long arg1, unsigned long arg2, unsigned long arg3, int pc) { struct ring_buffer_event *event; - struct trace_array_cpu *data = __data; struct trace_array *tr = __tr; + struct ring_buffer *buffer = tr->buffer; struct special_entry *entry; - unsigned long irq_flags; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = trace_buffer_lock_reserve(buffer, TRACE_SPECIAL, + sizeof(*entry), 0, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, pc); - entry->ent.type = TRACE_SPECIAL; entry->arg1 = arg1; entry->arg2 = arg2; entry->arg3 = arg3; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - ftrace_trace_stack(tr, data, irq_flags, 4, pc); - ftrace_trace_userstack(tr, data, irq_flags, pc); - - trace_wake_up(); + trace_buffer_unlock_commit(buffer, event, 0, pc); } void __trace_special(void *__tr, void *__data, unsigned long arg1, unsigned long arg2, unsigned long arg3) { - ftrace_trace_special(__tr, __data, arg1, arg2, arg3, preempt_count()); -} - -void -tracing_sched_switch_trace(struct trace_array *tr, - struct trace_array_cpu *data, - struct task_struct *prev, - struct task_struct *next, - unsigned long flags, int pc) -{ - struct ring_buffer_event *event; - struct ctx_switch_entry *entry; - unsigned long irq_flags; - - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); - if (!event) - return; - entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_CTX; - entry->prev_pid = prev->pid; - entry->prev_prio = prev->prio; - entry->prev_state = prev->state; - entry->next_pid = next->pid; - entry->next_prio = next->prio; - entry->next_state = next->state; - entry->next_cpu = task_cpu(next); - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - ftrace_trace_stack(tr, data, flags, 5, pc); - ftrace_trace_userstack(tr, data, flags, pc); -} - -void -tracing_sched_wakeup_trace(struct trace_array *tr, - struct trace_array_cpu *data, - struct task_struct *wakee, - struct task_struct *curr, - unsigned long flags, int pc) -{ - struct ring_buffer_event *event; - struct ctx_switch_entry *entry; - unsigned long irq_flags; - - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); - if (!event) - return; - entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_WAKE; - entry->prev_pid = curr->pid; - entry->prev_prio = curr->prio; - entry->prev_state = curr->state; - entry->next_pid = wakee->pid; - entry->next_prio = wakee->prio; - entry->next_state = wakee->state; - entry->next_cpu = task_cpu(wakee); - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - ftrace_trace_stack(tr, data, flags, 6, pc); - ftrace_trace_userstack(tr, data, flags, pc); - - trace_wake_up(); + ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count()); } void @@ -1157,152 +1135,163 @@ ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3) data = tr->data[cpu]; if (likely(atomic_inc_return(&data->disabled) == 1)) - ftrace_trace_special(tr, data, arg1, arg2, arg3, pc); + ftrace_trace_special(tr, arg1, arg2, arg3, pc); atomic_dec(&data->disabled); local_irq_restore(flags); } -#ifdef CONFIG_FUNCTION_TRACER -static void -function_trace_call_preempt_only(unsigned long ip, unsigned long parent_ip) +/** + * trace_vbprintk - write binary msg to tracing buffer + * + */ +int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) { + static raw_spinlock_t trace_buf_lock = + (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; + static u32 trace_buf[TRACE_BUF_SIZE]; + + struct ftrace_event_call *call = &event_bprint; + struct ring_buffer_event *event; + struct ring_buffer *buffer; struct trace_array *tr = &global_trace; struct trace_array_cpu *data; + struct bprint_entry *entry; unsigned long flags; - long disabled; - int cpu, resched; - int pc; + int disable; + int resched; + int cpu, len = 0, size, pc; - if (unlikely(!ftrace_function_enabled)) - return; + if (unlikely(tracing_selftest_running || tracing_disabled)) + return 0; + + /* Don't pollute graph traces with trace_vprintk internals */ + pause_graph_tracing(); pc = preempt_count(); resched = ftrace_preempt_disable(); - local_save_flags(flags); cpu = raw_smp_processor_id(); data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - - if (likely(disabled == 1)) - trace_function(tr, data, ip, parent_ip, flags, pc); - atomic_dec(&data->disabled); - ftrace_preempt_enable(resched); -} + disable = atomic_inc_return(&data->disabled); + if (unlikely(disable != 1)) + goto out; -static void -function_trace_call(unsigned long ip, unsigned long parent_ip) -{ - struct trace_array *tr = &global_trace; - struct trace_array_cpu *data; - unsigned long flags; - long disabled; - int cpu; - int pc; + /* Lockdep uses trace_printk for lock tracing */ + local_irq_save(flags); + __raw_spin_lock(&trace_buf_lock); + len = vbin_printf(trace_buf, TRACE_BUF_SIZE, fmt, args); - if (unlikely(!ftrace_function_enabled)) - return; + if (len > TRACE_BUF_SIZE || len < 0) + goto out_unlock; - /* - * Need to use raw, since this must be called before the - * recursive protection is performed. - */ - local_irq_save(flags); - cpu = raw_smp_processor_id(); - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); + size = sizeof(*entry) + sizeof(u32) * len; + buffer = tr->buffer; + event = trace_buffer_lock_reserve(buffer, TRACE_BPRINT, size, + flags, pc); + if (!event) + goto out_unlock; + entry = ring_buffer_event_data(event); + entry->ip = ip; + entry->fmt = fmt; - if (likely(disabled == 1)) { - pc = preempt_count(); - trace_function(tr, data, ip, parent_ip, flags, pc); - } + memcpy(entry->buf, trace_buf, sizeof(u32) * len); + if (!filter_check_discard(call, entry, buffer, event)) + ring_buffer_unlock_commit(buffer, event); - atomic_dec(&data->disabled); +out_unlock: + __raw_spin_unlock(&trace_buf_lock); local_irq_restore(flags); + +out: + atomic_dec_return(&data->disabled); + ftrace_preempt_enable(resched); + unpause_graph_tracing(); + + return len; } +EXPORT_SYMBOL_GPL(trace_vbprintk); -#ifdef CONFIG_FUNCTION_GRAPH_TRACER -int trace_graph_entry(struct ftrace_graph_ent *trace) +int trace_array_printk(struct trace_array *tr, + unsigned long ip, const char *fmt, ...) { - struct trace_array *tr = &global_trace; - struct trace_array_cpu *data; - unsigned long flags; - long disabled; - int cpu; - int pc; - - if (!ftrace_trace_task(current)) - return 0; + int ret; + va_list ap; - if (!ftrace_graph_addr(trace->func)) + if (!(trace_flags & TRACE_ITER_PRINTK)) return 0; - local_irq_save(flags); - cpu = raw_smp_processor_id(); - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - if (likely(disabled == 1)) { - pc = preempt_count(); - __trace_graph_entry(tr, data, trace, flags, pc); - } - /* Only do the atomic if it is not already set */ - if (!test_tsk_trace_graph(current)) - set_tsk_trace_graph(current); - atomic_dec(&data->disabled); - local_irq_restore(flags); - - return 1; + va_start(ap, fmt); + ret = trace_array_vprintk(tr, ip, fmt, ap); + va_end(ap); + return ret; } -void trace_graph_return(struct ftrace_graph_ret *trace) +int trace_array_vprintk(struct trace_array *tr, + unsigned long ip, const char *fmt, va_list args) { - struct trace_array *tr = &global_trace; + static raw_spinlock_t trace_buf_lock = __RAW_SPIN_LOCK_UNLOCKED; + static char trace_buf[TRACE_BUF_SIZE]; + + struct ftrace_event_call *call = &event_print; + struct ring_buffer_event *event; + struct ring_buffer *buffer; struct trace_array_cpu *data; - unsigned long flags; - long disabled; - int cpu; - int pc; + int cpu, len = 0, size, pc; + struct print_entry *entry; + unsigned long irq_flags; + int disable; - local_irq_save(flags); + if (tracing_disabled || tracing_selftest_running) + return 0; + + pc = preempt_count(); + preempt_disable_notrace(); cpu = raw_smp_processor_id(); data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - if (likely(disabled == 1)) { - pc = preempt_count(); - __trace_graph_return(tr, data, trace, flags, pc); - } - if (!trace->depth) - clear_tsk_trace_graph(current); - atomic_dec(&data->disabled); - local_irq_restore(flags); -} -#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ -static struct ftrace_ops trace_ops __read_mostly = -{ - .func = function_trace_call, -}; + disable = atomic_inc_return(&data->disabled); + if (unlikely(disable != 1)) + goto out; -void tracing_start_function_trace(void) -{ - ftrace_function_enabled = 0; + pause_graph_tracing(); + raw_local_irq_save(irq_flags); + __raw_spin_lock(&trace_buf_lock); + len = vsnprintf(trace_buf, TRACE_BUF_SIZE, fmt, args); - if (trace_flags & TRACE_ITER_PREEMPTONLY) - trace_ops.func = function_trace_call_preempt_only; - else - trace_ops.func = function_trace_call; + len = min(len, TRACE_BUF_SIZE-1); + trace_buf[len] = 0; + + size = sizeof(*entry) + len + 1; + buffer = tr->buffer; + event = trace_buffer_lock_reserve(buffer, TRACE_PRINT, size, + irq_flags, pc); + if (!event) + goto out_unlock; + entry = ring_buffer_event_data(event); + entry->ip = ip; - register_ftrace_function(&trace_ops); - ftrace_function_enabled = 1; + memcpy(&entry->buf, trace_buf, len); + entry->buf[len] = 0; + if (!filter_check_discard(call, entry, buffer, event)) + ring_buffer_unlock_commit(buffer, event); + + out_unlock: + __raw_spin_unlock(&trace_buf_lock); + raw_local_irq_restore(irq_flags); + unpause_graph_tracing(); + out: + atomic_dec_return(&data->disabled); + preempt_enable_notrace(); + + return len; } -void tracing_stop_function_trace(void) +int trace_vprintk(unsigned long ip, const char *fmt, va_list args) { - ftrace_function_enabled = 0; - unregister_ftrace_function(&trace_ops); + return trace_array_printk(&global_trace, ip, fmt, args); } -#endif +EXPORT_SYMBOL_GPL(trace_vprintk); enum trace_file_type { TRACE_FILE_LAT_FMT = 1, @@ -1345,10 +1334,25 @@ __find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts) { struct ring_buffer *buffer = iter->tr->buffer; struct trace_entry *ent, *next = NULL; + int cpu_file = iter->cpu_file; u64 next_ts = 0, ts; int next_cpu = -1; int cpu; + /* + * If we are in a per_cpu trace file, don't bother by iterating over + * all cpu and peek directly. + */ + if (cpu_file > TRACE_PIPE_ALL_CPU) { + if (ring_buffer_empty_cpu(buffer, cpu_file)) + return NULL; + ent = peek_next_entry(iter, cpu_file, ent_ts); + if (ent_cpu) + *ent_cpu = cpu_file; + + return ent; + } + for_each_tracing_cpu(cpu) { if (ring_buffer_empty_cpu(buffer, cpu)) @@ -1376,8 +1380,8 @@ __find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts) } /* Find the next real entry, without updating the iterator itself */ -static struct trace_entry * -find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts) +struct trace_entry *trace_find_next_entry(struct trace_iterator *iter, + int *ent_cpu, u64 *ent_ts) { return __find_next_entry(iter, ent_cpu, ent_ts); } @@ -1426,19 +1430,63 @@ static void *s_next(struct seq_file *m, void *v, loff_t *pos) return ent; } +static void tracing_iter_reset(struct trace_iterator *iter, int cpu) +{ + struct trace_array *tr = iter->tr; + struct ring_buffer_event *event; + struct ring_buffer_iter *buf_iter; + unsigned long entries = 0; + u64 ts; + + tr->data[cpu]->skipped_entries = 0; + + if (!iter->buffer_iter[cpu]) + return; + + buf_iter = iter->buffer_iter[cpu]; + ring_buffer_iter_reset(buf_iter); + + /* + * We could have the case with the max latency tracers + * that a reset never took place on a cpu. This is evident + * by the timestamp being before the start of the buffer. + */ + while ((event = ring_buffer_iter_peek(buf_iter, &ts))) { + if (ts >= iter->tr->time_start) + break; + entries++; + ring_buffer_read(buf_iter, NULL); + } + + tr->data[cpu]->skipped_entries = entries; +} + +/* + * No necessary locking here. The worst thing which can + * happen is loosing events consumed at the same time + * by a trace_pipe reader. + * Other than that, we don't risk to crash the ring buffer + * because it serializes the readers. + * + * The current tracer is copied to avoid a global locking + * all around. + */ static void *s_start(struct seq_file *m, loff_t *pos) { struct trace_iterator *iter = m->private; + static struct tracer *old_tracer; + int cpu_file = iter->cpu_file; void *p = NULL; loff_t l = 0; int cpu; + /* copy the tracer to avoid using a global lock all around */ mutex_lock(&trace_types_lock); - - if (!current_trace || current_trace != iter->trace) { - mutex_unlock(&trace_types_lock); - return NULL; + if (unlikely(old_tracer != current_trace && current_trace)) { + old_tracer = current_trace; + *iter->trace = *current_trace; } + mutex_unlock(&trace_types_lock); atomic_inc(&trace_record_cmdline_disabled); @@ -1449,9 +1497,11 @@ static void *s_start(struct seq_file *m, loff_t *pos) ftrace_disable_cpu(); - for_each_tracing_cpu(cpu) { - ring_buffer_iter_reset(iter->buffer_iter[cpu]); - } + if (cpu_file == TRACE_PIPE_ALL_CPU) { + for_each_tracing_cpu(cpu) + tracing_iter_reset(iter, cpu); + } else + tracing_iter_reset(iter, cpu_file); ftrace_enable_cpu(); @@ -1463,161 +1513,14 @@ static void *s_start(struct seq_file *m, loff_t *pos) p = s_next(m, p, &l); } + trace_event_read_lock(); return p; } static void s_stop(struct seq_file *m, void *p) { atomic_dec(&trace_record_cmdline_disabled); - mutex_unlock(&trace_types_lock); -} - -#ifdef CONFIG_KRETPROBES -static inline const char *kretprobed(const char *name) -{ - static const char tramp_name[] = "kretprobe_trampoline"; - int size = sizeof(tramp_name); - - if (strncmp(tramp_name, name, size) == 0) - return "[unknown/kretprobe'd]"; - return name; -} -#else -static inline const char *kretprobed(const char *name) -{ - return name; -} -#endif /* CONFIG_KRETPROBES */ - -static int -seq_print_sym_short(struct trace_seq *s, const char *fmt, unsigned long address) -{ -#ifdef CONFIG_KALLSYMS - char str[KSYM_SYMBOL_LEN]; - const char *name; - - kallsyms_lookup(address, NULL, NULL, NULL, str); - - name = kretprobed(str); - - return trace_seq_printf(s, fmt, name); -#endif - return 1; -} - -static int -seq_print_sym_offset(struct trace_seq *s, const char *fmt, - unsigned long address) -{ -#ifdef CONFIG_KALLSYMS - char str[KSYM_SYMBOL_LEN]; - const char *name; - - sprint_symbol(str, address); - name = kretprobed(str); - - return trace_seq_printf(s, fmt, name); -#endif - return 1; -} - -#ifndef CONFIG_64BIT -# define IP_FMT "%08lx" -#else -# define IP_FMT "%016lx" -#endif - -int -seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) -{ - int ret; - - if (!ip) - return trace_seq_printf(s, "0"); - - if (sym_flags & TRACE_ITER_SYM_OFFSET) - ret = seq_print_sym_offset(s, "%s", ip); - else - ret = seq_print_sym_short(s, "%s", ip); - - if (!ret) - return 0; - - if (sym_flags & TRACE_ITER_SYM_ADDR) - ret = trace_seq_printf(s, " <" IP_FMT ">", ip); - return ret; -} - -static inline int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, - unsigned long ip, unsigned long sym_flags) -{ - struct file *file = NULL; - unsigned long vmstart = 0; - int ret = 1; - - if (mm) { - const struct vm_area_struct *vma; - - down_read(&mm->mmap_sem); - vma = find_vma(mm, ip); - if (vma) { - file = vma->vm_file; - vmstart = vma->vm_start; - } - if (file) { - ret = trace_seq_path(s, &file->f_path); - if (ret) - ret = trace_seq_printf(s, "[+0x%lx]", ip - vmstart); - } - up_read(&mm->mmap_sem); - } - if (ret && ((sym_flags & TRACE_ITER_SYM_ADDR) || !file)) - ret = trace_seq_printf(s, " <" IP_FMT ">", ip); - return ret; -} - -static int -seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s, - unsigned long sym_flags) -{ - struct mm_struct *mm = NULL; - int ret = 1; - unsigned int i; - - if (trace_flags & TRACE_ITER_SYM_USEROBJ) { - struct task_struct *task; - /* - * we do the lookup on the thread group leader, - * since individual threads might have already quit! - */ - rcu_read_lock(); - task = find_task_by_vpid(entry->ent.tgid); - if (task) - mm = get_task_mm(task); - rcu_read_unlock(); - } - - for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { - unsigned long ip = entry->caller[i]; - - if (ip == ULONG_MAX || !ret) - break; - if (i && ret) - ret = trace_seq_puts(s, " <- "); - if (!ip) { - if (ret) - ret = trace_seq_puts(s, "??"); - continue; - } - if (!ret) - break; - if (ret) - ret = seq_print_user_ip(s, mm, ip, sym_flags); - } - - if (mm) - mmput(mm); - return ret; + trace_event_read_unlock(); } static void print_lat_help_header(struct seq_file *m) @@ -1647,22 +1550,38 @@ print_trace_header(struct seq_file *m, struct trace_iterator *iter) struct trace_array *tr = iter->tr; struct trace_array_cpu *data = tr->data[tr->cpu]; struct tracer *type = current_trace; - unsigned long total; - unsigned long entries; + unsigned long entries = 0; + unsigned long total = 0; + unsigned long count; const char *name = "preemption"; + int cpu; if (type) name = type->name; - entries = ring_buffer_entries(iter->tr->buffer); - total = entries + - ring_buffer_overruns(iter->tr->buffer); - seq_printf(m, "%s latency trace v1.1.5 on %s\n", + for_each_tracing_cpu(cpu) { + count = ring_buffer_entries_cpu(tr->buffer, cpu); + /* + * If this buffer has skipped entries, then we hold all + * entries for the trace and we need to ignore the + * ones before the time stamp. + */ + if (tr->data[cpu]->skipped_entries) { + count -= tr->data[cpu]->skipped_entries; + /* total is the same as the entries */ + total += count; + } else + total += count + + ring_buffer_overrun_cpu(tr->buffer, cpu); + entries += count; + } + + seq_printf(m, "# %s latency trace v1.1.5 on %s\n", name, UTS_RELEASE); - seq_puts(m, "-----------------------------------" + seq_puts(m, "# -----------------------------------" "---------------------------------\n"); - seq_printf(m, " latency: %lu us, #%lu/%lu, CPU#%d |" + seq_printf(m, "# latency: %lu us, #%lu/%lu, CPU#%d |" " (M:%s VP:%d, KP:%d, SP:%d HP:%d", nsecs_to_usecs(data->saved_latency), entries, @@ -1684,121 +1603,24 @@ print_trace_header(struct seq_file *m, struct trace_iterator *iter) #else seq_puts(m, ")\n"); #endif - seq_puts(m, " -----------------\n"); - seq_printf(m, " | task: %.16s-%d " + seq_puts(m, "# -----------------\n"); + seq_printf(m, "# | task: %.16s-%d " "(uid:%d nice:%ld policy:%ld rt_prio:%ld)\n", data->comm, data->pid, data->uid, data->nice, data->policy, data->rt_priority); - seq_puts(m, " -----------------\n"); + seq_puts(m, "# -----------------\n"); if (data->critical_start) { - seq_puts(m, " => started at: "); + seq_puts(m, "# => started at: "); seq_print_ip_sym(&iter->seq, data->critical_start, sym_flags); trace_print_seq(m, &iter->seq); - seq_puts(m, "\n => ended at: "); + seq_puts(m, "\n# => ended at: "); seq_print_ip_sym(&iter->seq, data->critical_end, sym_flags); trace_print_seq(m, &iter->seq); - seq_puts(m, "\n"); + seq_puts(m, "\n#\n"); } - seq_puts(m, "\n"); -} - -static void -lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu) -{ - int hardirq, softirq; - char *comm; - - comm = trace_find_cmdline(entry->pid); - - trace_seq_printf(s, "%8.8s-%-5d ", comm, entry->pid); - trace_seq_printf(s, "%3d", cpu); - trace_seq_printf(s, "%c%c", - (entry->flags & TRACE_FLAG_IRQS_OFF) ? 'd' : - (entry->flags & TRACE_FLAG_IRQS_NOSUPPORT) ? 'X' : '.', - ((entry->flags & TRACE_FLAG_NEED_RESCHED) ? 'N' : '.')); - - hardirq = entry->flags & TRACE_FLAG_HARDIRQ; - softirq = entry->flags & TRACE_FLAG_SOFTIRQ; - if (hardirq && softirq) { - trace_seq_putc(s, 'H'); - } else { - if (hardirq) { - trace_seq_putc(s, 'h'); - } else { - if (softirq) - trace_seq_putc(s, 's'); - else - trace_seq_putc(s, '.'); - } - } - - if (entry->preempt_count) - trace_seq_printf(s, "%x", entry->preempt_count); - else - trace_seq_puts(s, "."); -} - -unsigned long preempt_mark_thresh = 100; - -static void -lat_print_timestamp(struct trace_seq *s, u64 abs_usecs, - unsigned long rel_usecs) -{ - trace_seq_printf(s, " %4lldus", abs_usecs); - if (rel_usecs > preempt_mark_thresh) - trace_seq_puts(s, "!: "); - else if (rel_usecs > 1) - trace_seq_puts(s, "+: "); - else - trace_seq_puts(s, " : "); -} - -static const char state_to_char[] = TASK_STATE_TO_CHAR_STR; - -static int task_state_char(unsigned long state) -{ - int bit = state ? __ffs(state) + 1 : 0; - - return bit < sizeof(state_to_char) - 1 ? state_to_char[bit] : '?'; -} - -/* - * The message is supposed to contain an ending newline. - * If the printing stops prematurely, try to add a newline of our own. - */ -void trace_seq_print_cont(struct trace_seq *s, struct trace_iterator *iter) -{ - struct trace_entry *ent; - struct trace_field_cont *cont; - bool ok = true; - - ent = peek_next_entry(iter, iter->cpu, NULL); - if (!ent || ent->type != TRACE_CONT) { - trace_seq_putc(s, '\n'); - return; - } - - do { - cont = (struct trace_field_cont *)ent; - if (ok) - ok = (trace_seq_printf(s, "%s", cont->buf) > 0); - - ftrace_disable_cpu(); - - if (iter->buffer_iter[iter->cpu]) - ring_buffer_read(iter->buffer_iter[iter->cpu], NULL); - else - ring_buffer_consume(iter->tr->buffer, iter->cpu, NULL); - - ftrace_enable_cpu(); - - ent = peek_next_entry(iter, iter->cpu, NULL); - } while (ent && ent->type == TRACE_CONT); - - if (!ok) - trace_seq_putc(s, '\n'); + seq_puts(m, "#\n"); } static void test_cpu_buff_start(struct trace_iterator *iter) @@ -1814,142 +1636,15 @@ static void test_cpu_buff_start(struct trace_iterator *iter) if (cpumask_test_cpu(iter->cpu, iter->started)) return; - cpumask_set_cpu(iter->cpu, iter->started); - trace_seq_printf(s, "##### CPU %u buffer started ####\n", iter->cpu); -} - -static enum print_line_t -print_lat_fmt(struct trace_iterator *iter, unsigned int trace_idx, int cpu) -{ - struct trace_seq *s = &iter->seq; - unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); - struct trace_entry *next_entry; - unsigned long verbose = (trace_flags & TRACE_ITER_VERBOSE); - struct trace_entry *entry = iter->ent; - unsigned long abs_usecs; - unsigned long rel_usecs; - u64 next_ts; - char *comm; - int S, T; - int i; - - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - - test_cpu_buff_start(iter); - - next_entry = find_next_entry(iter, NULL, &next_ts); - if (!next_entry) - next_ts = iter->ts; - rel_usecs = ns2usecs(next_ts - iter->ts); - abs_usecs = ns2usecs(iter->ts - iter->tr->time_start); - - if (verbose) { - comm = trace_find_cmdline(entry->pid); - trace_seq_printf(s, "%16s %5d %3d %d %08x %08x [%08lx]" - " %ld.%03ldms (+%ld.%03ldms): ", - comm, - entry->pid, cpu, entry->flags, - entry->preempt_count, trace_idx, - ns2usecs(iter->ts), - abs_usecs/1000, - abs_usecs % 1000, rel_usecs/1000, - rel_usecs % 1000); - } else { - lat_print_generic(s, entry, cpu); - lat_print_timestamp(s, abs_usecs, rel_usecs); - } - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; - - trace_assign_type(field, entry); - - seq_print_ip_sym(s, field->ip, sym_flags); - trace_seq_puts(s, " ("); - seq_print_ip_sym(s, field->parent_ip, sym_flags); - trace_seq_puts(s, ")\n"); - break; - } - case TRACE_CTX: - case TRACE_WAKE: { - struct ctx_switch_entry *field; - - trace_assign_type(field, entry); - - T = task_state_char(field->next_state); - S = task_state_char(field->prev_state); - comm = trace_find_cmdline(field->next_pid); - trace_seq_printf(s, " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", - field->prev_pid, - field->prev_prio, - S, entry->type == TRACE_CTX ? "==>" : " +", - field->next_cpu, - field->next_pid, - field->next_prio, - T, comm); - break; - } - case TRACE_SPECIAL: { - struct special_entry *field; - - trace_assign_type(field, entry); - - trace_seq_printf(s, "# %ld %ld %ld\n", - field->arg1, - field->arg2, - field->arg3); - break; - } - case TRACE_STACK: { - struct stack_entry *field; - - trace_assign_type(field, entry); - - for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { - if (i) - trace_seq_puts(s, " <= "); - seq_print_ip_sym(s, field->caller[i], sym_flags); - } - trace_seq_puts(s, "\n"); - break; - } - case TRACE_PRINT: { - struct print_entry *field; - - trace_assign_type(field, entry); - - seq_print_ip_sym(s, field->ip, sym_flags); - trace_seq_printf(s, ": %s", field->buf); - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); - break; - } - case TRACE_BRANCH: { - struct trace_branch *field; - - trace_assign_type(field, entry); - - trace_seq_printf(s, "[%s] %s:%s:%d\n", - field->correct ? " ok " : " MISS ", - field->func, - field->file, - field->line); - break; - } - case TRACE_USER_STACK: { - struct userstack_entry *field; + if (iter->tr->data[iter->cpu]->skipped_entries) + return; - trace_assign_type(field, entry); + cpumask_set_cpu(iter->cpu, iter->started); - seq_print_userip_objs(field, s, sym_flags); - trace_seq_putc(s, '\n'); - break; - } - default: - trace_seq_printf(s, "Unknown type %d\n", entry->type); - } - return TRACE_TYPE_HANDLED; + /* Don't print started cpu buffer for the first entry of the trace */ + if (iter->idx > 1) + trace_seq_printf(s, "##### CPU %u buffer started ####\n", + iter->cpu); } static enum print_line_t print_trace_fmt(struct trace_iterator *iter) @@ -1957,333 +1652,84 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) struct trace_seq *s = &iter->seq; unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); struct trace_entry *entry; - unsigned long usec_rem; - unsigned long long t; - unsigned long secs; - char *comm; - int ret; - int S, T; - int i; + struct trace_event *event; entry = iter->ent; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - test_cpu_buff_start(iter); - comm = trace_find_cmdline(iter->ent->pid); - - t = ns2usecs(iter->ts); - usec_rem = do_div(t, 1000000ULL); - secs = (unsigned long)t; + event = ftrace_find_event(entry->type); - ret = trace_seq_printf(s, "%16s-%-5d ", comm, entry->pid); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, "[%03d] ", iter->cpu); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, "%5lu.%06lu: ", secs, usec_rem); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; - - trace_assign_type(field, entry); - - ret = seq_print_ip_sym(s, field->ip, sym_flags); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - if ((sym_flags & TRACE_ITER_PRINT_PARENT) && - field->parent_ip) { - ret = trace_seq_printf(s, " <-"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - ret = seq_print_ip_sym(s, - field->parent_ip, - sym_flags); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - ret = trace_seq_printf(s, "\n"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_CTX: - case TRACE_WAKE: { - struct ctx_switch_entry *field; - - trace_assign_type(field, entry); - - T = task_state_char(field->next_state); - S = task_state_char(field->prev_state); - ret = trace_seq_printf(s, " %5d:%3d:%c %s [%03d] %5d:%3d:%c\n", - field->prev_pid, - field->prev_prio, - S, - entry->type == TRACE_CTX ? "==>" : " +", - field->next_cpu, - field->next_pid, - field->next_prio, - T); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_SPECIAL: { - struct special_entry *field; - - trace_assign_type(field, entry); - - ret = trace_seq_printf(s, "# %ld %ld %ld\n", - field->arg1, - field->arg2, - field->arg3); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_STACK: { - struct stack_entry *field; - - trace_assign_type(field, entry); - - for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { - if (i) { - ret = trace_seq_puts(s, " <= "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - ret = seq_print_ip_sym(s, field->caller[i], - sym_flags); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + if (iter->iter_flags & TRACE_FILE_LAT_FMT) { + if (!trace_print_lat_context(iter)) + goto partial; + } else { + if (!trace_print_context(iter)) + goto partial; } - ret = trace_seq_puts(s, "\n"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_PRINT: { - struct print_entry *field; - - trace_assign_type(field, entry); - - seq_print_ip_sym(s, field->ip, sym_flags); - trace_seq_printf(s, ": %s", field->buf); - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); - break; } - case TRACE_GRAPH_RET: { - return print_graph_function(iter); - } - case TRACE_GRAPH_ENT: { - return print_graph_function(iter); - } - case TRACE_BRANCH: { - struct trace_branch *field; - - trace_assign_type(field, entry); - trace_seq_printf(s, "[%s] %s:%s:%d\n", - field->correct ? " ok " : " MISS ", - field->func, - field->file, - field->line); - break; - } - case TRACE_USER_STACK: { - struct userstack_entry *field; + if (event) + return event->trace(iter, sym_flags); - trace_assign_type(field, entry); + if (!trace_seq_printf(s, "Unknown type %d\n", entry->type)) + goto partial; - ret = seq_print_userip_objs(field, s, sym_flags); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_putc(s, '\n'); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - } return TRACE_TYPE_HANDLED; +partial: + return TRACE_TYPE_PARTIAL_LINE; } static enum print_line_t print_raw_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; struct trace_entry *entry; - int ret; - int S, T; + struct trace_event *event; entry = iter->ent; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - - ret = trace_seq_printf(s, "%d %d %llu ", - entry->pid, iter->cpu, iter->ts); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; - - trace_assign_type(field, entry); - - ret = trace_seq_printf(s, "%x %x\n", - field->ip, - field->parent_ip); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_CTX: - case TRACE_WAKE: { - struct ctx_switch_entry *field; - - trace_assign_type(field, entry); - - T = task_state_char(field->next_state); - S = entry->type == TRACE_WAKE ? '+' : - task_state_char(field->prev_state); - ret = trace_seq_printf(s, "%d %d %c %d %d %d %c\n", - field->prev_pid, - field->prev_prio, - S, - field->next_cpu, - field->next_pid, - field->next_prio, - T); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + if (!trace_seq_printf(s, "%d %d %llu ", + entry->pid, iter->cpu, iter->ts)) + goto partial; } - case TRACE_SPECIAL: - case TRACE_USER_STACK: - case TRACE_STACK: { - struct special_entry *field; - trace_assign_type(field, entry); + event = ftrace_find_event(entry->type); + if (event) + return event->raw(iter, 0); - ret = trace_seq_printf(s, "# %ld %ld %ld\n", - field->arg1, - field->arg2, - field->arg3); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_PRINT: { - struct print_entry *field; - - trace_assign_type(field, entry); + if (!trace_seq_printf(s, "%d ?\n", entry->type)) + goto partial; - trace_seq_printf(s, "# %lx %s", field->ip, field->buf); - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); - break; - } - } return TRACE_TYPE_HANDLED; +partial: + return TRACE_TYPE_PARTIAL_LINE; } -#define SEQ_PUT_FIELD_RET(s, x) \ -do { \ - if (!trace_seq_putmem(s, &(x), sizeof(x))) \ - return 0; \ -} while (0) - -#define SEQ_PUT_HEX_FIELD_RET(s, x) \ -do { \ - BUILD_BUG_ON(sizeof(x) > MAX_MEMHEX_BYTES); \ - if (!trace_seq_putmem_hex(s, &(x), sizeof(x))) \ - return 0; \ -} while (0) - static enum print_line_t print_hex_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; unsigned char newline = '\n'; struct trace_entry *entry; - int S, T; + struct trace_event *event; entry = iter->ent; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - - SEQ_PUT_HEX_FIELD_RET(s, entry->pid); - SEQ_PUT_HEX_FIELD_RET(s, iter->cpu); - SEQ_PUT_HEX_FIELD_RET(s, iter->ts); - - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; - - trace_assign_type(field, entry); - - SEQ_PUT_HEX_FIELD_RET(s, field->ip); - SEQ_PUT_HEX_FIELD_RET(s, field->parent_ip); - break; - } - case TRACE_CTX: - case TRACE_WAKE: { - struct ctx_switch_entry *field; - - trace_assign_type(field, entry); - - T = task_state_char(field->next_state); - S = entry->type == TRACE_WAKE ? '+' : - task_state_char(field->prev_state); - SEQ_PUT_HEX_FIELD_RET(s, field->prev_pid); - SEQ_PUT_HEX_FIELD_RET(s, field->prev_prio); - SEQ_PUT_HEX_FIELD_RET(s, S); - SEQ_PUT_HEX_FIELD_RET(s, field->next_cpu); - SEQ_PUT_HEX_FIELD_RET(s, field->next_pid); - SEQ_PUT_HEX_FIELD_RET(s, field->next_prio); - SEQ_PUT_HEX_FIELD_RET(s, T); - break; + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + SEQ_PUT_HEX_FIELD_RET(s, entry->pid); + SEQ_PUT_HEX_FIELD_RET(s, iter->cpu); + SEQ_PUT_HEX_FIELD_RET(s, iter->ts); } - case TRACE_SPECIAL: - case TRACE_USER_STACK: - case TRACE_STACK: { - struct special_entry *field; - trace_assign_type(field, entry); - - SEQ_PUT_HEX_FIELD_RET(s, field->arg1); - SEQ_PUT_HEX_FIELD_RET(s, field->arg2); - SEQ_PUT_HEX_FIELD_RET(s, field->arg3); - break; - } + event = ftrace_find_event(entry->type); + if (event) { + enum print_line_t ret = event->hex(iter, 0); + if (ret != TRACE_TYPE_HANDLED) + return ret; } - SEQ_PUT_FIELD_RET(s, newline); - - return TRACE_TYPE_HANDLED; -} -static enum print_line_t print_printk_msg_only(struct trace_iterator *iter) -{ - struct trace_seq *s = &iter->seq; - struct trace_entry *entry = iter->ent; - struct print_entry *field; - int ret; - - trace_assign_type(field, entry); - - ret = trace_seq_printf(s, field->buf); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); + SEQ_PUT_FIELD_RET(s, newline); return TRACE_TYPE_HANDLED; } @@ -2292,59 +1738,37 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; struct trace_entry *entry; + struct trace_event *event; entry = iter->ent; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - - SEQ_PUT_FIELD_RET(s, entry->pid); - SEQ_PUT_FIELD_RET(s, entry->cpu); - SEQ_PUT_FIELD_RET(s, iter->ts); - - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; - - trace_assign_type(field, entry); - - SEQ_PUT_FIELD_RET(s, field->ip); - SEQ_PUT_FIELD_RET(s, field->parent_ip); - break; + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + SEQ_PUT_FIELD_RET(s, entry->pid); + SEQ_PUT_FIELD_RET(s, iter->cpu); + SEQ_PUT_FIELD_RET(s, iter->ts); } - case TRACE_CTX: { - struct ctx_switch_entry *field; - trace_assign_type(field, entry); - - SEQ_PUT_FIELD_RET(s, field->prev_pid); - SEQ_PUT_FIELD_RET(s, field->prev_prio); - SEQ_PUT_FIELD_RET(s, field->prev_state); - SEQ_PUT_FIELD_RET(s, field->next_pid); - SEQ_PUT_FIELD_RET(s, field->next_prio); - SEQ_PUT_FIELD_RET(s, field->next_state); - break; - } - case TRACE_SPECIAL: - case TRACE_USER_STACK: - case TRACE_STACK: { - struct special_entry *field; - - trace_assign_type(field, entry); - - SEQ_PUT_FIELD_RET(s, field->arg1); - SEQ_PUT_FIELD_RET(s, field->arg2); - SEQ_PUT_FIELD_RET(s, field->arg3); - break; - } - } - return 1; + event = ftrace_find_event(entry->type); + return event ? event->binary(iter, 0) : TRACE_TYPE_HANDLED; } static int trace_empty(struct trace_iterator *iter) { int cpu; + /* If we are looking at one CPU buffer, only check that one */ + if (iter->cpu_file != TRACE_PIPE_ALL_CPU) { + cpu = iter->cpu_file; + if (iter->buffer_iter[cpu]) { + if (!ring_buffer_iter_empty(iter->buffer_iter[cpu])) + return 0; + } else { + if (!ring_buffer_empty_cpu(iter->tr->buffer, cpu)) + return 0; + } + return 1; + } + for_each_tracing_cpu(cpu) { if (iter->buffer_iter[cpu]) { if (!ring_buffer_iter_empty(iter->buffer_iter[cpu])) @@ -2358,6 +1782,7 @@ static int trace_empty(struct trace_iterator *iter) return 1; } +/* Called with trace_event_read_lock() held. */ static enum print_line_t print_trace_line(struct trace_iterator *iter) { enum print_line_t ret; @@ -2368,10 +1793,15 @@ static enum print_line_t print_trace_line(struct trace_iterator *iter) return ret; } + if (iter->ent->type == TRACE_BPRINT && + trace_flags & TRACE_ITER_PRINTK && + trace_flags & TRACE_ITER_PRINTK_MSGONLY) + return trace_print_bprintk_msg_only(iter); + if (iter->ent->type == TRACE_PRINT && trace_flags & TRACE_ITER_PRINTK && trace_flags & TRACE_ITER_PRINTK_MSGONLY) - return print_printk_msg_only(iter); + return trace_print_printk_msg_only(iter); if (trace_flags & TRACE_ITER_BIN) return print_bin_fmt(iter); @@ -2382,9 +1812,6 @@ static enum print_line_t print_trace_line(struct trace_iterator *iter) if (trace_flags & TRACE_ITER_RAW) return print_raw_fmt(iter); - if (iter->iter_flags & TRACE_FILE_LAT_FMT) - return print_lat_fmt(iter, iter->idx, iter->cpu); - return print_trace_fmt(iter); } @@ -2426,30 +1853,45 @@ static struct seq_operations tracer_seq_ops = { }; static struct trace_iterator * -__tracing_open(struct inode *inode, struct file *file, int *ret) +__tracing_open(struct inode *inode, struct file *file) { + long cpu_file = (long) inode->i_private; + void *fail_ret = ERR_PTR(-ENOMEM); struct trace_iterator *iter; struct seq_file *m; - int cpu; + int cpu, ret; - if (tracing_disabled) { - *ret = -ENODEV; - return NULL; - } + if (tracing_disabled) + return ERR_PTR(-ENODEV); iter = kzalloc(sizeof(*iter), GFP_KERNEL); - if (!iter) { - *ret = -ENOMEM; - goto out; - } + if (!iter) + return ERR_PTR(-ENOMEM); + /* + * We make a copy of the current tracer to avoid concurrent + * changes on it while we are reading. + */ mutex_lock(&trace_types_lock); + iter->trace = kzalloc(sizeof(*iter->trace), GFP_KERNEL); + if (!iter->trace) + goto fail; + + if (current_trace) + *iter->trace = *current_trace; + + if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) + goto fail; + + cpumask_clear(iter->started); + if (current_trace && current_trace->print_max) iter->tr = &max_tr; else - iter->tr = inode->i_private; - iter->trace = current_trace; + iter->tr = &global_trace; iter->pos = -1; + mutex_init(&iter->mutex); + iter->cpu_file = cpu_file; /* Notify the tracer early; before we stop tracing. */ if (iter->trace && iter->trace->open) @@ -2459,30 +1901,34 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) if (ring_buffer_overruns(iter->tr->buffer)) iter->iter_flags |= TRACE_FILE_ANNOTATE; + /* stop the trace while dumping */ + tracing_stop(); - for_each_tracing_cpu(cpu) { + if (iter->cpu_file == TRACE_PIPE_ALL_CPU) { + for_each_tracing_cpu(cpu) { + iter->buffer_iter[cpu] = + ring_buffer_read_start(iter->tr->buffer, cpu); + tracing_iter_reset(iter, cpu); + } + } else { + cpu = iter->cpu_file; iter->buffer_iter[cpu] = - ring_buffer_read_start(iter->tr->buffer, cpu); - - if (!iter->buffer_iter[cpu]) - goto fail_buffer; + ring_buffer_read_start(iter->tr->buffer, cpu); + tracing_iter_reset(iter, cpu); } - /* TODO stop tracer */ - *ret = seq_open(file, &tracer_seq_ops); - if (*ret) + ret = seq_open(file, &tracer_seq_ops); + if (ret < 0) { + fail_ret = ERR_PTR(ret); goto fail_buffer; + } m = file->private_data; m->private = iter; - /* stop the trace while dumping */ - tracing_stop(); - mutex_unlock(&trace_types_lock); - out: return iter; fail_buffer: @@ -2490,10 +1936,14 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) if (iter->buffer_iter[cpu]) ring_buffer_read_finish(iter->buffer_iter[cpu]); } + free_cpumask_var(iter->started); + tracing_start(); + fail: mutex_unlock(&trace_types_lock); + kfree(iter->trace); kfree(iter); - return ERR_PTR(-ENOMEM); + return fail_ret; } int tracing_open_generic(struct inode *inode, struct file *filp) @@ -2505,12 +1955,17 @@ int tracing_open_generic(struct inode *inode, struct file *filp) return 0; } -int tracing_release(struct inode *inode, struct file *file) +static int tracing_release(struct inode *inode, struct file *file) { struct seq_file *m = (struct seq_file *)file->private_data; - struct trace_iterator *iter = m->private; + struct trace_iterator *iter; int cpu; + if (!(file->f_mode & FMODE_READ)) + return 0; + + iter = m->private; + mutex_lock(&trace_types_lock); for_each_tracing_cpu(cpu) { if (iter->buffer_iter[cpu]) @@ -2525,55 +1980,59 @@ int tracing_release(struct inode *inode, struct file *file) mutex_unlock(&trace_types_lock); seq_release(inode, file); + mutex_destroy(&iter->mutex); + free_cpumask_var(iter->started); + kfree(iter->trace); kfree(iter); return 0; } static int tracing_open(struct inode *inode, struct file *file) { - int ret; - - __tracing_open(inode, file, &ret); - - return ret; -} - -static int tracing_lt_open(struct inode *inode, struct file *file) -{ struct trace_iterator *iter; - int ret; + int ret = 0; - iter = __tracing_open(inode, file, &ret); + /* If this file was open for write, then erase contents */ + if ((file->f_mode & FMODE_WRITE) && + (file->f_flags & O_TRUNC)) { + long cpu = (long) inode->i_private; - if (!ret) - iter->iter_flags |= TRACE_FILE_LAT_FMT; + if (cpu == TRACE_PIPE_ALL_CPU) + tracing_reset_online_cpus(&global_trace); + else + tracing_reset(&global_trace, cpu); + } + if (file->f_mode & FMODE_READ) { + iter = __tracing_open(inode, file); + if (IS_ERR(iter)) + ret = PTR_ERR(iter); + else if (trace_flags & TRACE_ITER_LATENCY_FMT) + iter->iter_flags |= TRACE_FILE_LAT_FMT; + } return ret; } - static void * t_next(struct seq_file *m, void *v, loff_t *pos) { - struct tracer *t = m->private; + struct tracer *t = v; (*pos)++; if (t) t = t->next; - m->private = t; - return t; } static void *t_start(struct seq_file *m, loff_t *pos) { - struct tracer *t = m->private; + struct tracer *t; loff_t l = 0; mutex_lock(&trace_types_lock); - for (; t && l < *pos; t = t_next(m, t, &l)) + for (t = trace_types; t && l < *pos; t = t_next(m, t, &l)) ; return t; @@ -2609,35 +2068,28 @@ static struct seq_operations show_traces_seq_ops = { static int show_traces_open(struct inode *inode, struct file *file) { - int ret; - if (tracing_disabled) return -ENODEV; - ret = seq_open(file, &show_traces_seq_ops); - if (!ret) { - struct seq_file *m = file->private_data; - m->private = trace_types; - } + return seq_open(file, &show_traces_seq_ops); +} - return ret; +static ssize_t +tracing_write_stub(struct file *filp, const char __user *ubuf, + size_t count, loff_t *ppos) +{ + return count; } -static struct file_operations tracing_fops = { +static const struct file_operations tracing_fops = { .open = tracing_open, .read = seq_read, + .write = tracing_write_stub, .llseek = seq_lseek, .release = tracing_release, }; -static struct file_operations tracing_lt_fops = { - .open = tracing_lt_open, - .read = seq_read, - .llseek = seq_lseek, - .release = tracing_release, -}; - -static struct file_operations show_traces_fops = { +static const struct file_operations show_traces_fops = { .open = show_traces_open, .read = seq_read, .release = seq_release, @@ -2692,11 +2144,12 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf, if (!alloc_cpumask_var(&tracing_cpumask_new, GFP_KERNEL)) return -ENOMEM; - mutex_lock(&tracing_cpumask_update_lock); err = cpumask_parse_user(ubuf, count, tracing_cpumask_new); if (err) goto err_unlock; + mutex_lock(&tracing_cpumask_update_lock); + local_irq_disable(); __raw_spin_lock(&ftrace_max_lock); for_each_tracing_cpu(cpu) { @@ -2724,13 +2177,12 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf, return count; err_unlock: - mutex_unlock(&tracing_cpumask_update_lock); - free_cpumask_var(tracing_cpumask); + free_cpumask_var(tracing_cpumask_new); return err; } -static struct file_operations tracing_cpumask_fops = { +static const struct file_operations tracing_cpumask_fops = { .open = tracing_open_generic, .read = tracing_cpumask_read, .write = tracing_cpumask_write, @@ -2740,80 +2192,85 @@ static ssize_t tracing_trace_options_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { - int i; + struct tracer_opt *trace_opts; + u32 tracer_flags; + int len = 0; char *buf; int r = 0; - int len = 0; - u32 tracer_flags = current_trace->flags->val; - struct tracer_opt *trace_opts = current_trace->flags->opts; + int i; - /* calulate max size */ + /* calculate max size */ for (i = 0; trace_options[i]; i++) { len += strlen(trace_options[i]); - len += 3; /* "no" and space */ + len += 3; /* "no" and newline */ } + mutex_lock(&trace_types_lock); + tracer_flags = current_trace->flags->val; + trace_opts = current_trace->flags->opts; + /* * Increase the size with names of options specific * of the current tracer. */ for (i = 0; trace_opts[i].name; i++) { len += strlen(trace_opts[i].name); - len += 3; /* "no" and space */ + len += 3; /* "no" and newline */ } - /* +2 for \n and \0 */ - buf = kmalloc(len + 2, GFP_KERNEL); - if (!buf) + /* +1 for \0 */ + buf = kmalloc(len + 1, GFP_KERNEL); + if (!buf) { + mutex_unlock(&trace_types_lock); return -ENOMEM; + } for (i = 0; trace_options[i]; i++) { if (trace_flags & (1 << i)) - r += sprintf(buf + r, "%s ", trace_options[i]); + r += sprintf(buf + r, "%s\n", trace_options[i]); else - r += sprintf(buf + r, "no%s ", trace_options[i]); + r += sprintf(buf + r, "no%s\n", trace_options[i]); } for (i = 0; trace_opts[i].name; i++) { if (tracer_flags & trace_opts[i].bit) - r += sprintf(buf + r, "%s ", + r += sprintf(buf + r, "%s\n", trace_opts[i].name); else - r += sprintf(buf + r, "no%s ", + r += sprintf(buf + r, "no%s\n", trace_opts[i].name); } + mutex_unlock(&trace_types_lock); - r += sprintf(buf + r, "\n"); - WARN_ON(r >= len + 2); + WARN_ON(r >= len + 1); r = simple_read_from_buffer(ubuf, cnt, ppos, buf, r); kfree(buf); - return r; } /* Try to assign a tracer specific option */ static int set_tracer_option(struct tracer *trace, char *cmp, int neg) { - struct tracer_flags *trace_flags = trace->flags; + struct tracer_flags *tracer_flags = trace->flags; struct tracer_opt *opts = NULL; int ret = 0, i = 0; int len; - for (i = 0; trace_flags->opts[i].name; i++) { - opts = &trace_flags->opts[i]; + for (i = 0; tracer_flags->opts[i].name; i++) { + opts = &tracer_flags->opts[i]; len = strlen(opts->name); if (strncmp(cmp, opts->name, len) == 0) { - ret = trace->set_flag(trace_flags->val, + ret = trace->set_flag(tracer_flags->val, opts->bit, !neg); break; } } /* Not found */ - if (!trace_flags->opts[i].name) + if (!tracer_flags->opts[i].name) return -EINVAL; /* Refused to handle */ @@ -2821,13 +2278,25 @@ static int set_tracer_option(struct tracer *trace, char *cmp, int neg) return ret; if (neg) - trace_flags->val &= ~opts->bit; + tracer_flags->val &= ~opts->bit; else - trace_flags->val |= opts->bit; + tracer_flags->val |= opts->bit; return 0; } +static void set_tracer_flags(unsigned int mask, int enabled) +{ + /* do nothing if flag is already set */ + if (!!(trace_flags & mask) == !!enabled) + return; + + if (enabled) + trace_flags |= mask; + else + trace_flags &= ~mask; +} + static ssize_t tracing_trace_options_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) @@ -2855,17 +2324,16 @@ tracing_trace_options_write(struct file *filp, const char __user *ubuf, int len = strlen(trace_options[i]); if (strncmp(cmp, trace_options[i], len) == 0) { - if (neg) - trace_flags &= ~(1 << i); - else - trace_flags |= (1 << i); + set_tracer_flags(1 << i, !neg); break; } } /* If no option could be set, test the specific tracer options */ if (!trace_options[i]) { + mutex_lock(&trace_types_lock); ret = set_tracer_option(current_trace, cmp, neg); + mutex_unlock(&trace_types_lock); if (ret) return ret; } @@ -2875,7 +2343,7 @@ tracing_trace_options_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations tracing_iter_fops = { +static const struct file_operations tracing_iter_fops = { .open = tracing_open_generic, .read = tracing_trace_options_read, .write = tracing_trace_options_write, @@ -2883,21 +2351,20 @@ static struct file_operations tracing_iter_fops = { static const char readme_msg[] = "tracing mini-HOWTO:\n\n" - "# mkdir /debug\n" - "# mount -t debugfs nodev /debug\n\n" - "# cat /debug/tracing/available_tracers\n" - "wakeup preemptirqsoff preemptoff irqsoff ftrace sched_switch none\n\n" - "# cat /debug/tracing/current_tracer\n" - "none\n" - "# echo sched_switch > /debug/tracing/current_tracer\n" - "# cat /debug/tracing/current_tracer\n" + "# mount -t debugfs nodev /sys/kernel/debug\n\n" + "# cat /sys/kernel/debug/tracing/available_tracers\n" + "wakeup preemptirqsoff preemptoff irqsoff function sched_switch nop\n\n" + "# cat /sys/kernel/debug/tracing/current_tracer\n" + "nop\n" + "# echo sched_switch > /sys/kernel/debug/tracing/current_tracer\n" + "# cat /sys/kernel/debug/tracing/current_tracer\n" "sched_switch\n" - "# cat /debug/tracing/trace_options\n" + "# cat /sys/kernel/debug/tracing/trace_options\n" "noprint-parent nosym-offset nosym-addr noverbose\n" - "# echo print-parent > /debug/tracing/trace_options\n" - "# echo 1 > /debug/tracing/tracing_enabled\n" - "# cat /debug/tracing/trace > /tmp/trace.txt\n" - "echo 0 > /debug/tracing/tracing_enabled\n" + "# echo print-parent > /sys/kernel/debug/tracing/trace_options\n" + "# echo 1 > /sys/kernel/debug/tracing/tracing_enabled\n" + "# cat /sys/kernel/debug/tracing/trace > /tmp/trace.txt\n" + "# echo 0 > /sys/kernel/debug/tracing/tracing_enabled\n" ; static ssize_t @@ -2908,12 +2375,62 @@ tracing_readme_read(struct file *filp, char __user *ubuf, readme_msg, strlen(readme_msg)); } -static struct file_operations tracing_readme_fops = { +static const struct file_operations tracing_readme_fops = { .open = tracing_open_generic, .read = tracing_readme_read, }; static ssize_t +tracing_saved_cmdlines_read(struct file *file, char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + char *buf_comm; + char *file_buf; + char *buf; + int len = 0; + int pid; + int i; + + file_buf = kmalloc(SAVED_CMDLINES*(16+TASK_COMM_LEN), GFP_KERNEL); + if (!file_buf) + return -ENOMEM; + + buf_comm = kmalloc(TASK_COMM_LEN, GFP_KERNEL); + if (!buf_comm) { + kfree(file_buf); + return -ENOMEM; + } + + buf = file_buf; + + for (i = 0; i < SAVED_CMDLINES; i++) { + int r; + + pid = map_cmdline_to_pid[i]; + if (pid == -1 || pid == NO_CMDLINE_MAP) + continue; + + trace_find_cmdline(pid, buf_comm); + r = sprintf(buf, "%d %s\n", pid, buf_comm); + buf += r; + len += r; + } + + len = simple_read_from_buffer(ubuf, cnt, ppos, + file_buf, len); + + kfree(file_buf); + kfree(buf_comm); + + return len; +} + +static const struct file_operations tracing_saved_cmdlines_fops = { + .open = tracing_open_generic, + .read = tracing_saved_cmdlines_read, +}; + +static ssize_t tracing_ctrl_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { @@ -2930,7 +2447,7 @@ tracing_ctrl_write(struct file *filp, const char __user *ubuf, { struct trace_array *tr = filp->private_data; char buf[64]; - long val; + unsigned long val; int ret; if (cnt >= sizeof(buf)) @@ -2985,13 +2502,105 @@ tracing_set_trace_read(struct file *filp, char __user *ubuf, return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); } -static int tracing_set_tracer(char *buf) +int tracer_init(struct tracer *t, struct trace_array *tr) +{ + tracing_reset_online_cpus(tr); + return t->init(tr); +} + +static int tracing_resize_ring_buffer(unsigned long size) +{ + int ret; + + /* + * If kernel or user changes the size of the ring buffer + * we use the size that was given, and we can forget about + * expanding it later. + */ + ring_buffer_expanded = 1; + + ret = ring_buffer_resize(global_trace.buffer, size); + if (ret < 0) + return ret; + + ret = ring_buffer_resize(max_tr.buffer, size); + if (ret < 0) { + int r; + + r = ring_buffer_resize(global_trace.buffer, + global_trace.entries); + if (r < 0) { + /* + * AARGH! We are left with different + * size max buffer!!!! + * The max buffer is our "snapshot" buffer. + * When a tracer needs a snapshot (one of the + * latency tracers), it swaps the max buffer + * with the saved snap shot. We succeeded to + * update the size of the main buffer, but failed to + * update the size of the max buffer. But when we tried + * to reset the main buffer to the original size, we + * failed there too. This is very unlikely to + * happen, but if it does, warn and kill all + * tracing. + */ + WARN_ON(1); + tracing_disabled = 1; + } + return ret; + } + + global_trace.entries = size; + + return ret; +} + +/** + * tracing_update_buffers - used by tracing facility to expand ring buffers + * + * To save on memory when the tracing is never used on a system with it + * configured in. The ring buffers are set to a minimum size. But once + * a user starts to use the tracing facility, then they need to grow + * to their default size. + * + * This function is to be called when a tracer is about to be used. + */ +int tracing_update_buffers(void) +{ + int ret = 0; + + mutex_lock(&trace_types_lock); + if (!ring_buffer_expanded) + ret = tracing_resize_ring_buffer(trace_buf_size); + mutex_unlock(&trace_types_lock); + + return ret; +} + +struct trace_option_dentry; + +static struct trace_option_dentry * +create_trace_option_files(struct tracer *tracer); + +static void +destroy_trace_option_files(struct trace_option_dentry *topts); + +static int tracing_set_tracer(const char *buf) { + static struct trace_option_dentry *topts; struct trace_array *tr = &global_trace; struct tracer *t; int ret = 0; mutex_lock(&trace_types_lock); + + if (!ring_buffer_expanded) { + ret = tracing_resize_ring_buffer(trace_buf_size); + if (ret < 0) + goto out; + ret = 0; + } + for (t = trace_types; t; t = t->next) { if (strcmp(t->name, buf) == 0) break; @@ -3007,9 +2616,14 @@ static int tracing_set_tracer(char *buf) if (current_trace && current_trace->reset) current_trace->reset(tr); + destroy_trace_option_files(topts); + current_trace = t; + + topts = create_trace_option_files(current_trace); + if (t->init) { - ret = t->init(tr); + ret = tracer_init(t, tr); if (ret) goto out; } @@ -3072,9 +2686,9 @@ static ssize_t tracing_max_lat_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) { - long *ptr = filp->private_data; + unsigned long *ptr = filp->private_data; char buf[64]; - long val; + unsigned long val; int ret; if (cnt >= sizeof(buf)) @@ -3094,54 +2708,99 @@ tracing_max_lat_write(struct file *filp, const char __user *ubuf, return cnt; } -static atomic_t tracing_reader; - static int tracing_open_pipe(struct inode *inode, struct file *filp) { + long cpu_file = (long) inode->i_private; struct trace_iterator *iter; + int ret = 0; if (tracing_disabled) return -ENODEV; - /* We only allow for reader of the pipe */ - if (atomic_inc_return(&tracing_reader) != 1) { - atomic_dec(&tracing_reader); - return -EBUSY; + mutex_lock(&trace_types_lock); + + /* We only allow one reader per cpu */ + if (cpu_file == TRACE_PIPE_ALL_CPU) { + if (!cpumask_empty(tracing_reader_cpumask)) { + ret = -EBUSY; + goto out; + } + cpumask_setall(tracing_reader_cpumask); + } else { + if (!cpumask_test_cpu(cpu_file, tracing_reader_cpumask)) + cpumask_set_cpu(cpu_file, tracing_reader_cpumask); + else { + ret = -EBUSY; + goto out; + } } /* create a buffer to store the information to pass to userspace */ iter = kzalloc(sizeof(*iter), GFP_KERNEL); - if (!iter) - return -ENOMEM; + if (!iter) { + ret = -ENOMEM; + goto out; + } - if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) { - kfree(iter); - return -ENOMEM; + /* + * We make a copy of the current tracer to avoid concurrent + * changes on it while we are reading. + */ + iter->trace = kmalloc(sizeof(*iter->trace), GFP_KERNEL); + if (!iter->trace) { + ret = -ENOMEM; + goto fail; } + if (current_trace) + *iter->trace = *current_trace; - mutex_lock(&trace_types_lock); + if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) { + ret = -ENOMEM; + goto fail; + } /* trace pipe does not show start of buffer */ cpumask_setall(iter->started); + if (trace_flags & TRACE_ITER_LATENCY_FMT) + iter->iter_flags |= TRACE_FILE_LAT_FMT; + + iter->cpu_file = cpu_file; iter->tr = &global_trace; - iter->trace = current_trace; + mutex_init(&iter->mutex); filp->private_data = iter; if (iter->trace->pipe_open) iter->trace->pipe_open(iter); + +out: mutex_unlock(&trace_types_lock); + return ret; - return 0; +fail: + kfree(iter->trace); + kfree(iter); + mutex_unlock(&trace_types_lock); + return ret; } static int tracing_release_pipe(struct inode *inode, struct file *file) { struct trace_iterator *iter = file->private_data; + mutex_lock(&trace_types_lock); + + if (iter->cpu_file == TRACE_PIPE_ALL_CPU) + cpumask_clear(tracing_reader_cpumask); + else + cpumask_clear_cpu(iter->cpu_file, tracing_reader_cpumask); + + mutex_unlock(&trace_types_lock); + free_cpumask_var(iter->started); + mutex_destroy(&iter->mutex); + kfree(iter->trace); kfree(iter); - atomic_dec(&tracing_reader); return 0; } @@ -3167,67 +2826,57 @@ tracing_poll_pipe(struct file *filp, poll_table *poll_table) } } -/* - * Consumer reader. - */ -static ssize_t -tracing_read_pipe(struct file *filp, char __user *ubuf, - size_t cnt, loff_t *ppos) + +void default_wait_pipe(struct trace_iterator *iter) { - struct trace_iterator *iter = filp->private_data; - ssize_t sret; + DEFINE_WAIT(wait); - /* return any leftover data */ - sret = trace_seq_to_user(&iter->seq, ubuf, cnt); - if (sret != -EBUSY) - return sret; + prepare_to_wait(&trace_wait, &wait, TASK_INTERRUPTIBLE); - trace_seq_reset(&iter->seq); + if (trace_empty(iter)) + schedule(); - mutex_lock(&trace_types_lock); - if (iter->trace->read) { - sret = iter->trace->read(iter, filp, ubuf, cnt, ppos); - if (sret) - goto out; - } + finish_wait(&trace_wait, &wait); +} + +/* + * This is a make-shift waitqueue. + * A tracer might use this callback on some rare cases: + * + * 1) the current tracer might hold the runqueue lock when it wakes up + * a reader, hence a deadlock (sched, function, and function graph tracers) + * 2) the function tracers, trace all functions, we don't want + * the overhead of calling wake_up and friends + * (and tracing them too) + * + * Anyway, this is really very primitive wakeup. + */ +void poll_wait_pipe(struct trace_iterator *iter) +{ + set_current_state(TASK_INTERRUPTIBLE); + /* sleep for 100 msecs, and try again. */ + schedule_timeout(HZ / 10); +} + +/* Must be called with trace_types_lock mutex held. */ +static int tracing_wait_pipe(struct file *filp) +{ + struct trace_iterator *iter = filp->private_data; -waitagain: - sret = 0; while (trace_empty(iter)) { if ((filp->f_flags & O_NONBLOCK)) { - sret = -EAGAIN; - goto out; + return -EAGAIN; } - /* - * This is a make-shift waitqueue. The reason we don't use - * an actual wait queue is because: - * 1) we only ever have one waiter - * 2) the tracing, traces all functions, we don't want - * the overhead of calling wake_up and friends - * (and tracing them too) - * Anyway, this is really very primitive wakeup. - */ - set_current_state(TASK_INTERRUPTIBLE); - iter->tr->waiter = current; - - mutex_unlock(&trace_types_lock); - - /* sleep for 100 msecs, and try again. */ - schedule_timeout(HZ/10); + mutex_unlock(&iter->mutex); - mutex_lock(&trace_types_lock); - - iter->tr->waiter = NULL; + iter->trace->wait_pipe(iter); - if (signal_pending(current)) { - sret = -EINTR; - goto out; - } + mutex_lock(&iter->mutex); - if (iter->trace != current_trace) - goto out; + if (signal_pending(current)) + return -EINTR; /* * We block until we read something and tracing is disabled. @@ -3240,13 +2889,59 @@ waitagain: */ if (!tracer_enabled && iter->pos) break; + } + + return 1; +} - continue; +/* + * Consumer reader. + */ +static ssize_t +tracing_read_pipe(struct file *filp, char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + struct trace_iterator *iter = filp->private_data; + static struct tracer *old_tracer; + ssize_t sret; + + /* return any leftover data */ + sret = trace_seq_to_user(&iter->seq, ubuf, cnt); + if (sret != -EBUSY) + return sret; + + trace_seq_init(&iter->seq); + + /* copy the tracer to avoid using a global lock all around */ + mutex_lock(&trace_types_lock); + if (unlikely(old_tracer != current_trace && current_trace)) { + old_tracer = current_trace; + *iter->trace = *current_trace; + } + mutex_unlock(&trace_types_lock); + + /* + * Avoid more than one consumer on a single file descriptor + * This is just a matter of traces coherency, the ring buffer itself + * is protected. + */ + mutex_lock(&iter->mutex); + if (iter->trace->read) { + sret = iter->trace->read(iter, filp, ubuf, cnt, ppos); + if (sret) + goto out; } +waitagain: + sret = tracing_wait_pipe(filp); + if (sret <= 0) + goto out; + /* stop when tracing is finished */ - if (trace_empty(iter)) + if (trace_empty(iter)) { + sret = 0; goto out; + } if (cnt >= PAGE_SIZE) cnt = PAGE_SIZE - 1; @@ -3257,6 +2952,7 @@ waitagain: offsetof(struct trace_iterator, seq)); iter->pos = -1; + trace_event_read_lock(); while (find_next_entry_inc(iter) != NULL) { enum print_line_t ret; int len = iter->seq.len; @@ -3267,17 +2963,18 @@ waitagain: iter->seq.len = len; break; } - - trace_consume(iter); + if (ret != TRACE_TYPE_NO_CONSUME) + trace_consume(iter); if (iter->seq.len >= cnt) break; } + trace_event_read_unlock(); /* Now copy what we have to the user */ sret = trace_seq_to_user(&iter->seq, ubuf, cnt); if (iter->seq.readpos >= iter->seq.len) - trace_seq_reset(&iter->seq); + trace_seq_init(&iter->seq); /* * If there was nothing to send to user, inspite of consuming trace @@ -3287,20 +2984,169 @@ waitagain: goto waitagain; out: - mutex_unlock(&trace_types_lock); + mutex_unlock(&iter->mutex); return sret; } +static void tracing_pipe_buf_release(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + __free_page(buf->page); +} + +static void tracing_spd_release_pipe(struct splice_pipe_desc *spd, + unsigned int idx) +{ + __free_page(spd->pages[idx]); +} + +static struct pipe_buf_operations tracing_pipe_buf_ops = { + .can_merge = 0, + .map = generic_pipe_buf_map, + .unmap = generic_pipe_buf_unmap, + .confirm = generic_pipe_buf_confirm, + .release = tracing_pipe_buf_release, + .steal = generic_pipe_buf_steal, + .get = generic_pipe_buf_get, +}; + +static size_t +tracing_fill_pipe_page(size_t rem, struct trace_iterator *iter) +{ + size_t count; + int ret; + + /* Seq buffer is page-sized, exactly what we need. */ + for (;;) { + count = iter->seq.len; + ret = print_trace_line(iter); + count = iter->seq.len - count; + if (rem < count) { + rem = 0; + iter->seq.len -= count; + break; + } + if (ret == TRACE_TYPE_PARTIAL_LINE) { + iter->seq.len -= count; + break; + } + + if (ret != TRACE_TYPE_NO_CONSUME) + trace_consume(iter); + rem -= count; + if (!find_next_entry_inc(iter)) { + rem = 0; + iter->ent = NULL; + break; + } + } + + return rem; +} + +static ssize_t tracing_splice_read_pipe(struct file *filp, + loff_t *ppos, + struct pipe_inode_info *pipe, + size_t len, + unsigned int flags) +{ + struct page *pages[PIPE_BUFFERS]; + struct partial_page partial[PIPE_BUFFERS]; + struct trace_iterator *iter = filp->private_data; + struct splice_pipe_desc spd = { + .pages = pages, + .partial = partial, + .nr_pages = 0, /* This gets updated below. */ + .flags = flags, + .ops = &tracing_pipe_buf_ops, + .spd_release = tracing_spd_release_pipe, + }; + static struct tracer *old_tracer; + ssize_t ret; + size_t rem; + unsigned int i; + + /* copy the tracer to avoid using a global lock all around */ + mutex_lock(&trace_types_lock); + if (unlikely(old_tracer != current_trace && current_trace)) { + old_tracer = current_trace; + *iter->trace = *current_trace; + } + mutex_unlock(&trace_types_lock); + + mutex_lock(&iter->mutex); + + if (iter->trace->splice_read) { + ret = iter->trace->splice_read(iter, filp, + ppos, pipe, len, flags); + if (ret) + goto out_err; + } + + ret = tracing_wait_pipe(filp); + if (ret <= 0) + goto out_err; + + if (!iter->ent && !find_next_entry_inc(iter)) { + ret = -EFAULT; + goto out_err; + } + + trace_event_read_lock(); + + /* Fill as many pages as possible. */ + for (i = 0, rem = len; i < PIPE_BUFFERS && rem; i++) { + pages[i] = alloc_page(GFP_KERNEL); + if (!pages[i]) + break; + + rem = tracing_fill_pipe_page(rem, iter); + + /* Copy the data into the page, so we can start over. */ + ret = trace_seq_to_buffer(&iter->seq, + page_address(pages[i]), + iter->seq.len); + if (ret < 0) { + __free_page(pages[i]); + break; + } + partial[i].offset = 0; + partial[i].len = iter->seq.len; + + trace_seq_init(&iter->seq); + } + + trace_event_read_unlock(); + mutex_unlock(&iter->mutex); + + spd.nr_pages = i; + + return splice_to_pipe(pipe, &spd); + +out_err: + mutex_unlock(&iter->mutex); + + return ret; +} + static ssize_t tracing_entries_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct trace_array *tr = filp->private_data; - char buf[64]; + char buf[96]; int r; - r = sprintf(buf, "%lu\n", tr->entries >> 10); + mutex_lock(&trace_types_lock); + if (!ring_buffer_expanded) + r = sprintf(buf, "%lu (expanded: %lu)\n", + tr->entries >> 10, + trace_buf_size >> 10); + else + r = sprintf(buf, "%lu\n", tr->entries >> 10); + mutex_unlock(&trace_types_lock); + return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); } @@ -3344,28 +3190,11 @@ tracing_entries_write(struct file *filp, const char __user *ubuf, val <<= 10; if (val != global_trace.entries) { - ret = ring_buffer_resize(global_trace.buffer, val); - if (ret < 0) { - cnt = ret; - goto out; - } - - ret = ring_buffer_resize(max_tr.buffer, val); + ret = tracing_resize_ring_buffer(val); if (ret < 0) { - int r; cnt = ret; - r = ring_buffer_resize(global_trace.buffer, - global_trace.entries); - if (r < 0) { - /* AARGH! We are left with different - * size max buffer!!!! */ - WARN_ON(1); - tracing_disabled = 1; - } goto out; } - - global_trace.entries = val; } filp->f_pos += cnt; @@ -3393,7 +3222,7 @@ static int mark_printk(const char *fmt, ...) int ret; va_list args; va_start(args, fmt); - ret = trace_vprintk(0, -1, fmt, args); + ret = trace_vprintk(0, fmt, args); va_end(args); return ret; } @@ -3433,42 +3262,397 @@ tracing_mark_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations tracing_max_lat_fops = { +static ssize_t tracing_clock_read(struct file *filp, char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + char buf[64]; + int bufiter = 0; + int i; + + for (i = 0; i < ARRAY_SIZE(trace_clocks); i++) + bufiter += snprintf(buf + bufiter, sizeof(buf) - bufiter, + "%s%s%s%s", i ? " " : "", + i == trace_clock_id ? "[" : "", trace_clocks[i].name, + i == trace_clock_id ? "]" : ""); + bufiter += snprintf(buf + bufiter, sizeof(buf) - bufiter, "\n"); + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, bufiter); +} + +static ssize_t tracing_clock_write(struct file *filp, const char __user *ubuf, + size_t cnt, loff_t *fpos) +{ + char buf[64]; + const char *clockstr; + int i; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + clockstr = strstrip(buf); + + for (i = 0; i < ARRAY_SIZE(trace_clocks); i++) { + if (strcmp(trace_clocks[i].name, clockstr) == 0) + break; + } + if (i == ARRAY_SIZE(trace_clocks)) + return -EINVAL; + + trace_clock_id = i; + + mutex_lock(&trace_types_lock); + + ring_buffer_set_clock(global_trace.buffer, trace_clocks[i].func); + if (max_tr.buffer) + ring_buffer_set_clock(max_tr.buffer, trace_clocks[i].func); + + mutex_unlock(&trace_types_lock); + + *fpos += cnt; + + return cnt; +} + +static const struct file_operations tracing_max_lat_fops = { .open = tracing_open_generic, .read = tracing_max_lat_read, .write = tracing_max_lat_write, }; -static struct file_operations tracing_ctrl_fops = { +static const struct file_operations tracing_ctrl_fops = { .open = tracing_open_generic, .read = tracing_ctrl_read, .write = tracing_ctrl_write, }; -static struct file_operations set_tracer_fops = { +static const struct file_operations set_tracer_fops = { .open = tracing_open_generic, .read = tracing_set_trace_read, .write = tracing_set_trace_write, }; -static struct file_operations tracing_pipe_fops = { +static const struct file_operations tracing_pipe_fops = { .open = tracing_open_pipe, .poll = tracing_poll_pipe, .read = tracing_read_pipe, + .splice_read = tracing_splice_read_pipe, .release = tracing_release_pipe, }; -static struct file_operations tracing_entries_fops = { +static const struct file_operations tracing_entries_fops = { .open = tracing_open_generic, .read = tracing_entries_read, .write = tracing_entries_write, }; -static struct file_operations tracing_mark_fops = { +static const struct file_operations tracing_mark_fops = { .open = tracing_open_generic, .write = tracing_mark_write, }; +static const struct file_operations trace_clock_fops = { + .open = tracing_open_generic, + .read = tracing_clock_read, + .write = tracing_clock_write, +}; + +struct ftrace_buffer_info { + struct trace_array *tr; + void *spare; + int cpu; + unsigned int read; +}; + +static int tracing_buffers_open(struct inode *inode, struct file *filp) +{ + int cpu = (int)(long)inode->i_private; + struct ftrace_buffer_info *info; + + if (tracing_disabled) + return -ENODEV; + + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) + return -ENOMEM; + + info->tr = &global_trace; + info->cpu = cpu; + info->spare = NULL; + /* Force reading ring buffer for first read */ + info->read = (unsigned int)-1; + + filp->private_data = info; + + return nonseekable_open(inode, filp); +} + +static ssize_t +tracing_buffers_read(struct file *filp, char __user *ubuf, + size_t count, loff_t *ppos) +{ + struct ftrace_buffer_info *info = filp->private_data; + unsigned int pos; + ssize_t ret; + size_t size; + + if (!count) + return 0; + + if (!info->spare) + info->spare = ring_buffer_alloc_read_page(info->tr->buffer); + if (!info->spare) + return -ENOMEM; + + /* Do we have previous read data to read? */ + if (info->read < PAGE_SIZE) + goto read; + + info->read = 0; + + ret = ring_buffer_read_page(info->tr->buffer, + &info->spare, + count, + info->cpu, 0); + if (ret < 0) + return 0; + + pos = ring_buffer_page_len(info->spare); + + if (pos < PAGE_SIZE) + memset(info->spare + pos, 0, PAGE_SIZE - pos); + +read: + size = PAGE_SIZE - info->read; + if (size > count) + size = count; + + ret = copy_to_user(ubuf, info->spare + info->read, size); + if (ret == size) + return -EFAULT; + size -= ret; + + *ppos += size; + info->read += size; + + return size; +} + +static int tracing_buffers_release(struct inode *inode, struct file *file) +{ + struct ftrace_buffer_info *info = file->private_data; + + if (info->spare) + ring_buffer_free_read_page(info->tr->buffer, info->spare); + kfree(info); + + return 0; +} + +struct buffer_ref { + struct ring_buffer *buffer; + void *page; + int ref; +}; + +static void buffer_pipe_buf_release(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + struct buffer_ref *ref = (struct buffer_ref *)buf->private; + + if (--ref->ref) + return; + + ring_buffer_free_read_page(ref->buffer, ref->page); + kfree(ref); + buf->private = 0; +} + +static int buffer_pipe_buf_steal(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + return 1; +} + +static void buffer_pipe_buf_get(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + struct buffer_ref *ref = (struct buffer_ref *)buf->private; + + ref->ref++; +} + +/* Pipe buffer operations for a buffer. */ +static struct pipe_buf_operations buffer_pipe_buf_ops = { + .can_merge = 0, + .map = generic_pipe_buf_map, + .unmap = generic_pipe_buf_unmap, + .confirm = generic_pipe_buf_confirm, + .release = buffer_pipe_buf_release, + .steal = buffer_pipe_buf_steal, + .get = buffer_pipe_buf_get, +}; + +/* + * Callback from splice_to_pipe(), if we need to release some pages + * at the end of the spd in case we error'ed out in filling the pipe. + */ +static void buffer_spd_release(struct splice_pipe_desc *spd, unsigned int i) +{ + struct buffer_ref *ref = + (struct buffer_ref *)spd->partial[i].private; + + if (--ref->ref) + return; + + ring_buffer_free_read_page(ref->buffer, ref->page); + kfree(ref); + spd->partial[i].private = 0; +} + +static ssize_t +tracing_buffers_splice_read(struct file *file, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags) +{ + struct ftrace_buffer_info *info = file->private_data; + struct partial_page partial[PIPE_BUFFERS]; + struct page *pages[PIPE_BUFFERS]; + struct splice_pipe_desc spd = { + .pages = pages, + .partial = partial, + .flags = flags, + .ops = &buffer_pipe_buf_ops, + .spd_release = buffer_spd_release, + }; + struct buffer_ref *ref; + int entries, size, i; + size_t ret; + + if (*ppos & (PAGE_SIZE - 1)) { + WARN_ONCE(1, "Ftrace: previous read must page-align\n"); + return -EINVAL; + } + + if (len & (PAGE_SIZE - 1)) { + WARN_ONCE(1, "Ftrace: splice_read should page-align\n"); + if (len < PAGE_SIZE) + return -EINVAL; + len &= PAGE_MASK; + } + + entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); + + for (i = 0; i < PIPE_BUFFERS && len && entries; i++, len -= PAGE_SIZE) { + struct page *page; + int r; + + ref = kzalloc(sizeof(*ref), GFP_KERNEL); + if (!ref) + break; + + ref->ref = 1; + ref->buffer = info->tr->buffer; + ref->page = ring_buffer_alloc_read_page(ref->buffer); + if (!ref->page) { + kfree(ref); + break; + } + + r = ring_buffer_read_page(ref->buffer, &ref->page, + len, info->cpu, 1); + if (r < 0) { + ring_buffer_free_read_page(ref->buffer, + ref->page); + kfree(ref); + break; + } + + /* + * zero out any left over data, this is going to + * user land. + */ + size = ring_buffer_page_len(ref->page); + if (size < PAGE_SIZE) + memset(ref->page + size, 0, PAGE_SIZE - size); + + page = virt_to_page(ref->page); + + spd.pages[i] = page; + spd.partial[i].len = PAGE_SIZE; + spd.partial[i].offset = 0; + spd.partial[i].private = (unsigned long)ref; + spd.nr_pages++; + *ppos += PAGE_SIZE; + + entries = ring_buffer_entries_cpu(info->tr->buffer, info->cpu); + } + + spd.nr_pages = i; + + /* did we read anything? */ + if (!spd.nr_pages) { + if (flags & SPLICE_F_NONBLOCK) + ret = -EAGAIN; + else + ret = 0; + /* TODO: block */ + return ret; + } + + ret = splice_to_pipe(pipe, &spd); + + return ret; +} + +static const struct file_operations tracing_buffers_fops = { + .open = tracing_buffers_open, + .read = tracing_buffers_read, + .release = tracing_buffers_release, + .splice_read = tracing_buffers_splice_read, + .llseek = no_llseek, +}; + +static ssize_t +tracing_stats_read(struct file *filp, char __user *ubuf, + size_t count, loff_t *ppos) +{ + unsigned long cpu = (unsigned long)filp->private_data; + struct trace_array *tr = &global_trace; + struct trace_seq *s; + unsigned long cnt; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return ENOMEM; + + trace_seq_init(s); + + cnt = ring_buffer_entries_cpu(tr->buffer, cpu); + trace_seq_printf(s, "entries: %ld\n", cnt); + + cnt = ring_buffer_overrun_cpu(tr->buffer, cpu); + trace_seq_printf(s, "overrun: %ld\n", cnt); + + cnt = ring_buffer_commit_overrun_cpu(tr->buffer, cpu); + trace_seq_printf(s, "commit overrun: %ld\n", cnt); + + count = simple_read_from_buffer(ubuf, count, ppos, s->buffer, s->len); + + kfree(s); + + return count; +} + +static const struct file_operations tracing_stats_fops = { + .open = tracing_open_generic, + .read = tracing_stats_read, +}; + #ifdef CONFIG_DYNAMIC_FTRACE int __weak ftrace_arch_read_dyn_info(char *buf, int size) @@ -3500,7 +3684,7 @@ tracing_read_dyn_info(struct file *filp, char __user *ubuf, return r; } -static struct file_operations tracing_dyn_info_fops = { +static const struct file_operations tracing_dyn_info_fops = { .open = tracing_open_generic, .read = tracing_read_dyn_info, }; @@ -3515,6 +3699,9 @@ struct dentry *tracing_init_dentry(void) if (d_tracer) return d_tracer; + if (!debugfs_initialized()) + return NULL; + d_tracer = debugfs_create_dir("tracing", NULL); if (!d_tracer && !once) { @@ -3526,170 +3713,405 @@ struct dentry *tracing_init_dentry(void) return d_tracer; } +static struct dentry *d_percpu; + +struct dentry *tracing_dentry_percpu(void) +{ + static int once; + struct dentry *d_tracer; + + if (d_percpu) + return d_percpu; + + d_tracer = tracing_init_dentry(); + + if (!d_tracer) + return NULL; + + d_percpu = debugfs_create_dir("per_cpu", d_tracer); + + if (!d_percpu && !once) { + once = 1; + pr_warning("Could not create debugfs directory 'per_cpu'\n"); + return NULL; + } + + return d_percpu; +} + +static void tracing_init_debugfs_percpu(long cpu) +{ + struct dentry *d_percpu = tracing_dentry_percpu(); + struct dentry *d_cpu; + /* strlen(cpu) + MAX(log10(cpu)) + '\0' */ + char cpu_dir[7]; + + if (cpu > 999 || cpu < 0) + return; + + sprintf(cpu_dir, "cpu%ld", cpu); + d_cpu = debugfs_create_dir(cpu_dir, d_percpu); + if (!d_cpu) { + pr_warning("Could not create debugfs '%s' entry\n", cpu_dir); + return; + } + + /* per cpu trace_pipe */ + trace_create_file("trace_pipe", 0444, d_cpu, + (void *) cpu, &tracing_pipe_fops); + + /* per cpu trace */ + trace_create_file("trace", 0644, d_cpu, + (void *) cpu, &tracing_fops); + + trace_create_file("trace_pipe_raw", 0444, d_cpu, + (void *) cpu, &tracing_buffers_fops); + + trace_create_file("stats", 0444, d_cpu, + (void *) cpu, &tracing_stats_fops); +} + #ifdef CONFIG_FTRACE_SELFTEST /* Let selftest have access to static functions in this file */ #include "trace_selftest.c" #endif -static __init int tracer_init_debugfs(void) +struct trace_option_dentry { + struct tracer_opt *opt; + struct tracer_flags *flags; + struct dentry *entry; +}; + +static ssize_t +trace_options_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct trace_option_dentry *topt = filp->private_data; + char *buf; + + if (topt->flags->val & topt->opt->bit) + buf = "1\n"; + else + buf = "0\n"; + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, 2); +} + +static ssize_t +trace_options_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct trace_option_dentry *topt = filp->private_data; + unsigned long val; + char buf[64]; + int ret; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + ret = strict_strtoul(buf, 10, &val); + if (ret < 0) + return ret; + + ret = 0; + switch (val) { + case 0: + /* do nothing if already cleared */ + if (!(topt->flags->val & topt->opt->bit)) + break; + + mutex_lock(&trace_types_lock); + if (current_trace->set_flag) + ret = current_trace->set_flag(topt->flags->val, + topt->opt->bit, 0); + mutex_unlock(&trace_types_lock); + if (ret) + return ret; + topt->flags->val &= ~topt->opt->bit; + break; + case 1: + /* do nothing if already set */ + if (topt->flags->val & topt->opt->bit) + break; + + mutex_lock(&trace_types_lock); + if (current_trace->set_flag) + ret = current_trace->set_flag(topt->flags->val, + topt->opt->bit, 1); + mutex_unlock(&trace_types_lock); + if (ret) + return ret; + topt->flags->val |= topt->opt->bit; + break; + + default: + return -EINVAL; + } + + *ppos += cnt; + + return cnt; +} + + +static const struct file_operations trace_options_fops = { + .open = tracing_open_generic, + .read = trace_options_read, + .write = trace_options_write, +}; + +static ssize_t +trace_options_core_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + long index = (long)filp->private_data; + char *buf; + + if (trace_flags & (1 << index)) + buf = "1\n"; + else + buf = "0\n"; + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, 2); +} + +static ssize_t +trace_options_core_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + long index = (long)filp->private_data; + char buf[64]; + unsigned long val; + int ret; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + ret = strict_strtoul(buf, 10, &val); + if (ret < 0) + return ret; + + if (val != 0 && val != 1) + return -EINVAL; + set_tracer_flags(1 << index, val); + + *ppos += cnt; + + return cnt; +} + +static const struct file_operations trace_options_core_fops = { + .open = tracing_open_generic, + .read = trace_options_core_read, + .write = trace_options_core_write, +}; + +struct dentry *trace_create_file(const char *name, + mode_t mode, + struct dentry *parent, + void *data, + const struct file_operations *fops) +{ + struct dentry *ret; + + ret = debugfs_create_file(name, mode, parent, data, fops); + if (!ret) + pr_warning("Could not create debugfs '%s' entry\n", name); + + return ret; +} + + +static struct dentry *trace_options_init_dentry(void) { struct dentry *d_tracer; - struct dentry *entry; + static struct dentry *t_options; + + if (t_options) + return t_options; d_tracer = tracing_init_dentry(); + if (!d_tracer) + return NULL; - entry = debugfs_create_file("tracing_enabled", 0644, d_tracer, - &global_trace, &tracing_ctrl_fops); - if (!entry) - pr_warning("Could not create debugfs 'tracing_enabled' entry\n"); - - entry = debugfs_create_file("trace_options", 0644, d_tracer, - NULL, &tracing_iter_fops); - if (!entry) - pr_warning("Could not create debugfs 'trace_options' entry\n"); - - entry = debugfs_create_file("tracing_cpumask", 0644, d_tracer, - NULL, &tracing_cpumask_fops); - if (!entry) - pr_warning("Could not create debugfs 'tracing_cpumask' entry\n"); - - entry = debugfs_create_file("latency_trace", 0444, d_tracer, - &global_trace, &tracing_lt_fops); - if (!entry) - pr_warning("Could not create debugfs 'latency_trace' entry\n"); - - entry = debugfs_create_file("trace", 0444, d_tracer, - &global_trace, &tracing_fops); - if (!entry) - pr_warning("Could not create debugfs 'trace' entry\n"); - - entry = debugfs_create_file("available_tracers", 0444, d_tracer, - &global_trace, &show_traces_fops); - if (!entry) - pr_warning("Could not create debugfs 'available_tracers' entry\n"); - - entry = debugfs_create_file("current_tracer", 0444, d_tracer, - &global_trace, &set_tracer_fops); - if (!entry) - pr_warning("Could not create debugfs 'current_tracer' entry\n"); - - entry = debugfs_create_file("tracing_max_latency", 0644, d_tracer, - &tracing_max_latency, - &tracing_max_lat_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'tracing_max_latency' entry\n"); - - entry = debugfs_create_file("tracing_thresh", 0644, d_tracer, - &tracing_thresh, &tracing_max_lat_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'tracing_thresh' entry\n"); - entry = debugfs_create_file("README", 0644, d_tracer, - NULL, &tracing_readme_fops); - if (!entry) - pr_warning("Could not create debugfs 'README' entry\n"); - - entry = debugfs_create_file("trace_pipe", 0644, d_tracer, - NULL, &tracing_pipe_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'trace_pipe' entry\n"); - - entry = debugfs_create_file("buffer_size_kb", 0644, d_tracer, - &global_trace, &tracing_entries_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'buffer_size_kb' entry\n"); - - entry = debugfs_create_file("trace_marker", 0220, d_tracer, - NULL, &tracing_mark_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'trace_marker' entry\n"); + t_options = debugfs_create_dir("options", d_tracer); + if (!t_options) { + pr_warning("Could not create debugfs directory 'options'\n"); + return NULL; + } -#ifdef CONFIG_DYNAMIC_FTRACE - entry = debugfs_create_file("dyn_ftrace_total_info", 0444, d_tracer, - &ftrace_update_tot_cnt, - &tracing_dyn_info_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'dyn_ftrace_total_info' entry\n"); -#endif -#ifdef CONFIG_SYSPROF_TRACER - init_tracer_sysprof_debugfs(d_tracer); -#endif - return 0; + return t_options; } -int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) +static void +create_trace_option_file(struct trace_option_dentry *topt, + struct tracer_flags *flags, + struct tracer_opt *opt) { - static DEFINE_SPINLOCK(trace_buf_lock); - static char trace_buf[TRACE_BUF_SIZE]; + struct dentry *t_options; - struct ring_buffer_event *event; - struct trace_array *tr = &global_trace; - struct trace_array_cpu *data; - int cpu, len = 0, size, pc; - struct print_entry *entry; - unsigned long irq_flags; + t_options = trace_options_init_dentry(); + if (!t_options) + return; - if (tracing_disabled || tracing_selftest_running) - return 0; + topt->flags = flags; + topt->opt = opt; - pc = preempt_count(); - preempt_disable_notrace(); - cpu = raw_smp_processor_id(); - data = tr->data[cpu]; + topt->entry = trace_create_file(opt->name, 0644, t_options, topt, + &trace_options_fops); - if (unlikely(atomic_read(&data->disabled))) - goto out; +} - pause_graph_tracing(); - spin_lock_irqsave(&trace_buf_lock, irq_flags); - len = vsnprintf(trace_buf, TRACE_BUF_SIZE, fmt, args); +static struct trace_option_dentry * +create_trace_option_files(struct tracer *tracer) +{ + struct trace_option_dentry *topts; + struct tracer_flags *flags; + struct tracer_opt *opts; + int cnt; - len = min(len, TRACE_BUF_SIZE-1); - trace_buf[len] = 0; + if (!tracer) + return NULL; - size = sizeof(*entry) + len + 1; - event = ring_buffer_lock_reserve(tr->buffer, size, &irq_flags); - if (!event) - goto out_unlock; - entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, irq_flags, pc); - entry->ent.type = TRACE_PRINT; - entry->ip = ip; - entry->depth = depth; + flags = tracer->flags; - memcpy(&entry->buf, trace_buf, len); - entry->buf[len] = 0; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + if (!flags || !flags->opts) + return NULL; - out_unlock: - spin_unlock_irqrestore(&trace_buf_lock, irq_flags); - unpause_graph_tracing(); - out: - preempt_enable_notrace(); + opts = flags->opts; - return len; + for (cnt = 0; opts[cnt].name; cnt++) + ; + + topts = kcalloc(cnt + 1, sizeof(*topts), GFP_KERNEL); + if (!topts) + return NULL; + + for (cnt = 0; opts[cnt].name; cnt++) + create_trace_option_file(&topts[cnt], flags, + &opts[cnt]); + + return topts; } -EXPORT_SYMBOL_GPL(trace_vprintk); -int __ftrace_printk(unsigned long ip, const char *fmt, ...) +static void +destroy_trace_option_files(struct trace_option_dentry *topts) { - int ret; - va_list ap; + int cnt; - if (!(trace_flags & TRACE_ITER_PRINTK)) - return 0; + if (!topts) + return; - va_start(ap, fmt); - ret = trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); - va_end(ap); - return ret; + for (cnt = 0; topts[cnt].opt; cnt++) { + if (topts[cnt].entry) + debugfs_remove(topts[cnt].entry); + } + + kfree(topts); +} + +static struct dentry * +create_trace_option_core_file(const char *option, long index) +{ + struct dentry *t_options; + + t_options = trace_options_init_dentry(); + if (!t_options) + return NULL; + + return trace_create_file(option, 0644, t_options, (void *)index, + &trace_options_core_fops); +} + +static __init void create_trace_options_dir(void) +{ + struct dentry *t_options; + int i; + + t_options = trace_options_init_dentry(); + if (!t_options) + return; + + for (i = 0; trace_options[i]; i++) + create_trace_option_core_file(trace_options[i], i); +} + +static __init int tracer_init_debugfs(void) +{ + struct dentry *d_tracer; + int cpu; + + d_tracer = tracing_init_dentry(); + + trace_create_file("tracing_enabled", 0644, d_tracer, + &global_trace, &tracing_ctrl_fops); + + trace_create_file("trace_options", 0644, d_tracer, + NULL, &tracing_iter_fops); + + trace_create_file("tracing_cpumask", 0644, d_tracer, + NULL, &tracing_cpumask_fops); + + trace_create_file("trace", 0644, d_tracer, + (void *) TRACE_PIPE_ALL_CPU, &tracing_fops); + + trace_create_file("available_tracers", 0444, d_tracer, + &global_trace, &show_traces_fops); + + trace_create_file("current_tracer", 0644, d_tracer, + &global_trace, &set_tracer_fops); + +#ifdef CONFIG_TRACER_MAX_TRACE + trace_create_file("tracing_max_latency", 0644, d_tracer, + &tracing_max_latency, &tracing_max_lat_fops); + + trace_create_file("tracing_thresh", 0644, d_tracer, + &tracing_thresh, &tracing_max_lat_fops); +#endif + + trace_create_file("README", 0444, d_tracer, + NULL, &tracing_readme_fops); + + trace_create_file("trace_pipe", 0444, d_tracer, + (void *) TRACE_PIPE_ALL_CPU, &tracing_pipe_fops); + + trace_create_file("buffer_size_kb", 0644, d_tracer, + &global_trace, &tracing_entries_fops); + + trace_create_file("trace_marker", 0220, d_tracer, + NULL, &tracing_mark_fops); + + trace_create_file("saved_cmdlines", 0444, d_tracer, + NULL, &tracing_saved_cmdlines_fops); + + trace_create_file("trace_clock", 0644, d_tracer, NULL, + &trace_clock_fops); + +#ifdef CONFIG_DYNAMIC_FTRACE + trace_create_file("dyn_ftrace_total_info", 0444, d_tracer, + &ftrace_update_tot_cnt, &tracing_dyn_info_fops); +#endif +#ifdef CONFIG_SYSPROF_TRACER + init_tracer_sysprof_debugfs(d_tracer); +#endif + + create_trace_options_dir(); + + for_each_tracing_cpu(cpu) + tracing_init_debugfs_percpu(cpu); + + return 0; } -EXPORT_SYMBOL_GPL(__ftrace_printk); static int trace_panic_handler(struct notifier_block *this, unsigned long event, void *unused) @@ -3736,7 +4158,7 @@ static struct notifier_block trace_die_notifier = { * it if we decide to change what log level the ftrace dump * should be at. */ -#define KERN_TRACE KERN_INFO +#define KERN_TRACE KERN_EMERG static void trace_printk_seq(struct trace_seq *s) @@ -3750,39 +4172,48 @@ trace_printk_seq(struct trace_seq *s) printk(KERN_TRACE "%s", s->buffer); - trace_seq_reset(s); + trace_seq_init(s); } -void ftrace_dump(void) +static void __ftrace_dump(bool disable_tracing) { - static DEFINE_SPINLOCK(ftrace_dump_lock); + static raw_spinlock_t ftrace_dump_lock = + (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; /* use static because iter can be a bit big for the stack */ static struct trace_iterator iter; + unsigned int old_userobj; static int dump_ran; unsigned long flags; int cnt = 0, cpu; /* only one dump */ - spin_lock_irqsave(&ftrace_dump_lock, flags); + local_irq_save(flags); + __raw_spin_lock(&ftrace_dump_lock); if (dump_ran) goto out; dump_ran = 1; - /* No turning back! */ - ftrace_kill(); + tracing_off(); + + if (disable_tracing) + ftrace_kill(); for_each_tracing_cpu(cpu) { atomic_inc(&global_trace.data[cpu]->disabled); } + old_userobj = trace_flags & TRACE_ITER_SYM_USEROBJ; + /* don't look at user memory in panic mode */ trace_flags &= ~TRACE_ITER_SYM_USEROBJ; printk(KERN_TRACE "Dumping ftrace buffer:\n"); + /* Simulate the iterator */ iter.tr = &global_trace; iter.trace = current_trace; + iter.cpu_file = TRACE_PIPE_ALL_CPU; /* * We need to stop all tracing on all CPUS to read the @@ -3806,8 +4237,11 @@ void ftrace_dump(void) iter.pos = -1; if (find_next_entry_inc(&iter) != NULL) { - print_trace_line(&iter); - trace_consume(&iter); + int ret; + + ret = print_trace_line(&iter); + if (ret != TRACE_TYPE_NO_CONSUME) + trace_consume(&iter); } trace_printk_seq(&iter.seq); @@ -3818,13 +4252,30 @@ void ftrace_dump(void) else printk(KERN_TRACE "---------------------------------\n"); + /* Re-enable tracing if requested */ + if (!disable_tracing) { + trace_flags |= old_userobj; + + for_each_tracing_cpu(cpu) { + atomic_dec(&global_trace.data[cpu]->disabled); + } + tracing_on(); + } + out: - spin_unlock_irqrestore(&ftrace_dump_lock, flags); + __raw_spin_unlock(&ftrace_dump_lock); + local_irq_restore(flags); +} + +/* By default: disable tracing after the dump */ +void ftrace_dump(void) +{ + __ftrace_dump(true); } __init static int tracer_alloc_buffers(void) { - struct trace_array_cpu *data; + int ring_buf_size; int i; int ret = -ENOMEM; @@ -3834,11 +4285,21 @@ __init static int tracer_alloc_buffers(void) if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL)) goto out_free_buffer_mask; + if (!alloc_cpumask_var(&tracing_reader_cpumask, GFP_KERNEL)) + goto out_free_tracing_cpumask; + + /* To save memory, keep the ring buffer size to its minimum */ + if (ring_buffer_expanded) + ring_buf_size = trace_buf_size; + else + ring_buf_size = 1; + cpumask_copy(tracing_buffer_mask, cpu_possible_mask); cpumask_copy(tracing_cpumask, cpu_all_mask); + cpumask_clear(tracing_reader_cpumask); /* TODO: make the number of buffers hot pluggable with CPUS */ - global_trace.buffer = ring_buffer_alloc(trace_buf_size, + global_trace.buffer = ring_buffer_alloc(ring_buf_size, TRACE_BUFFER_FLAGS); if (!global_trace.buffer) { printk(KERN_ERR "tracer: failed to allocate ring buffer!\n"); @@ -3849,7 +4310,7 @@ __init static int tracer_alloc_buffers(void) #ifdef CONFIG_TRACER_MAX_TRACE - max_tr.buffer = ring_buffer_alloc(trace_buf_size, + max_tr.buffer = ring_buffer_alloc(ring_buf_size, TRACE_BUFFER_FLAGS); if (!max_tr.buffer) { printk(KERN_ERR "tracer: failed to allocate max ring buffer!\n"); @@ -3863,21 +4324,17 @@ __init static int tracer_alloc_buffers(void) /* Allocate the first page for all buffers */ for_each_tracing_cpu(i) { - data = global_trace.data[i] = &per_cpu(global_trace_cpu, i); + global_trace.data[i] = &per_cpu(global_trace_cpu, i); max_tr.data[i] = &per_cpu(max_data, i); } trace_init_cmdlines(); register_tracer(&nop_trace); + current_trace = &nop_trace; #ifdef CONFIG_BOOT_TRACER register_tracer(&boot_tracer); - current_trace = &boot_tracer; - current_trace->init(&global_trace); -#else - current_trace = &nop_trace; #endif - /* All seems OK, enable tracing */ tracing_disabled = 0; @@ -3885,14 +4342,38 @@ __init static int tracer_alloc_buffers(void) &trace_panic_notifier); register_die_notifier(&trace_die_notifier); - ret = 0; + + return 0; out_free_cpumask: + free_cpumask_var(tracing_reader_cpumask); +out_free_tracing_cpumask: free_cpumask_var(tracing_cpumask); out_free_buffer_mask: free_cpumask_var(tracing_buffer_mask); out: return ret; } + +__init static int clear_boot_tracer(void) +{ + /* + * The default tracer at boot buffer is an init section. + * This function is called in lateinit. If we did not + * find the boot tracer, then clear it out, to prevent + * later registration from accessing the buffer that is + * about to be freed. + */ + if (!default_bootup_tracer) + return 0; + + printk(KERN_INFO "ftrace bootup tracer '%s' not registered.\n", + default_bootup_tracer); + default_bootup_tracer = NULL; + + return 0; +} + early_initcall(tracer_alloc_buffers); fs_initcall(tracer_init_debugfs); +late_initcall(clear_boot_tracer); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 4d3d381bfd9..fa1dccb579d 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -9,6 +9,11 @@ #include <linux/mmiotrace.h> #include <linux/ftrace.h> #include <trace/boot.h> +#include <linux/kmemtrace.h> +#include <trace/power.h> + +#include <linux/trace_seq.h> +#include <linux/ftrace_event.h> enum trace_type { __TRACE_FIRST_TYPE = 0, @@ -16,9 +21,9 @@ enum trace_type { TRACE_FN, TRACE_CTX, TRACE_WAKE, - TRACE_CONT, TRACE_STACK, TRACE_PRINT, + TRACE_BPRINT, TRACE_SPECIAL, TRACE_MMIO_RW, TRACE_MMIO_MAP, @@ -29,24 +34,12 @@ enum trace_type { TRACE_GRAPH_ENT, TRACE_USER_STACK, TRACE_HW_BRANCHES, + TRACE_KMEM_ALLOC, + TRACE_KMEM_FREE, TRACE_POWER, + TRACE_BLK, - __TRACE_LAST_TYPE -}; - -/* - * The trace entry - the most basic unit of tracing. This is what - * is printed in the end as a single line in the trace output, such as: - * - * bash-15816 [01] 235.197585: idle_cpu <- irq_enter - */ -struct trace_entry { - unsigned char type; - unsigned char cpu; - unsigned char flags; - unsigned char preempt_count; - int pid; - int tgid; + __TRACE_LAST_TYPE, }; /* @@ -60,13 +53,13 @@ struct ftrace_entry { /* Function call entry */ struct ftrace_graph_ent_entry { - struct trace_entry ent; + struct trace_entry ent; struct ftrace_graph_ent graph_ent; }; /* Function return entry */ struct ftrace_graph_ret_entry { - struct trace_entry ent; + struct trace_entry ent; struct ftrace_graph_ret ret; }; extern struct tracer boot_tracer; @@ -112,12 +105,18 @@ struct userstack_entry { }; /* - * ftrace_printk entry: + * trace_printk entry: */ +struct bprint_entry { + struct trace_entry ent; + unsigned long ip; + const char *fmt; + u32 buf[]; +}; + struct print_entry { struct trace_entry ent; unsigned long ip; - int depth; char buf[]; }; @@ -170,15 +169,51 @@ struct trace_power { struct power_trace state_data; }; +enum kmemtrace_type_id { + KMEMTRACE_TYPE_KMALLOC = 0, /* kmalloc() or kfree(). */ + KMEMTRACE_TYPE_CACHE, /* kmem_cache_*(). */ + KMEMTRACE_TYPE_PAGES, /* __get_free_pages() and friends. */ +}; + +struct kmemtrace_alloc_entry { + struct trace_entry ent; + enum kmemtrace_type_id type_id; + unsigned long call_site; + const void *ptr; + size_t bytes_req; + size_t bytes_alloc; + gfp_t gfp_flags; + int node; +}; + +struct kmemtrace_free_entry { + struct trace_entry ent; + enum kmemtrace_type_id type_id; + unsigned long call_site; + const void *ptr; +}; + +struct syscall_trace_enter { + struct trace_entry ent; + int nr; + unsigned long args[]; +}; + +struct syscall_trace_exit { + struct trace_entry ent; + int nr; + unsigned long ret; +}; + + /* * trace_flag_type is an enumeration that holds different * states when a trace occurs. These are: * IRQS_OFF - interrupts were disabled - * IRQS_NOSUPPORT - arch does not support irqs_disabled_flags + * IRQS_NOSUPPORT - arch does not support irqs_disabled_flags * NEED_RESCED - reschedule is requested * HARDIRQ - inside an interrupt handler * SOFTIRQ - inside a softirq handler - * CONT - multiple entries hold the trace item */ enum trace_flag_type { TRACE_FLAG_IRQS_OFF = 0x01, @@ -186,7 +221,6 @@ enum trace_flag_type { TRACE_FLAG_NEED_RESCHED = 0x04, TRACE_FLAG_HARDIRQ = 0x08, TRACE_FLAG_SOFTIRQ = 0x10, - TRACE_FLAG_CONT = 0x20, }; #define TRACE_BUF_SIZE 1024 @@ -198,10 +232,8 @@ enum trace_flag_type { */ struct trace_array_cpu { atomic_t disabled; + void *buffer_page; /* ring buffer spare */ - /* these fields get copied into max-trace: */ - unsigned long trace_idx; - unsigned long overrun; unsigned long saved_latency; unsigned long critical_start; unsigned long critical_end; @@ -209,14 +241,13 @@ struct trace_array_cpu { unsigned long nice; unsigned long policy; unsigned long rt_priority; + unsigned long skipped_entries; cycle_t preempt_timestamp; pid_t pid; uid_t uid; char comm[TASK_COMM_LEN]; }; -struct trace_iterator; - /* * The trace array - an array of per-CPU trace arrays. This is the * highest level data structure that individual tracers deal with. @@ -262,10 +293,10 @@ extern void __ftrace_bad_type(void); do { \ IF_ASSIGN(var, ent, struct ftrace_entry, TRACE_FN); \ IF_ASSIGN(var, ent, struct ctx_switch_entry, 0); \ - IF_ASSIGN(var, ent, struct trace_field_cont, TRACE_CONT); \ IF_ASSIGN(var, ent, struct stack_entry, TRACE_STACK); \ IF_ASSIGN(var, ent, struct userstack_entry, TRACE_USER_STACK);\ IF_ASSIGN(var, ent, struct print_entry, TRACE_PRINT); \ + IF_ASSIGN(var, ent, struct bprint_entry, TRACE_BPRINT); \ IF_ASSIGN(var, ent, struct special_entry, 0); \ IF_ASSIGN(var, ent, struct trace_mmiotrace_rw, \ TRACE_MMIO_RW); \ @@ -279,26 +310,22 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct ftrace_graph_ret_entry, \ TRACE_GRAPH_RET); \ IF_ASSIGN(var, ent, struct hw_branch_entry, TRACE_HW_BRANCHES);\ - IF_ASSIGN(var, ent, struct trace_power, TRACE_POWER); \ + IF_ASSIGN(var, ent, struct trace_power, TRACE_POWER); \ + IF_ASSIGN(var, ent, struct kmemtrace_alloc_entry, \ + TRACE_KMEM_ALLOC); \ + IF_ASSIGN(var, ent, struct kmemtrace_free_entry, \ + TRACE_KMEM_FREE); \ __ftrace_bad_type(); \ } while (0) -/* Return values for print_line callback */ -enum print_line_t { - TRACE_TYPE_PARTIAL_LINE = 0, /* Retry after flushing the seq */ - TRACE_TYPE_HANDLED = 1, - TRACE_TYPE_UNHANDLED = 2 /* Relay to other output functions */ -}; - - /* * An option specific to a tracer. This is a boolean value. * The bit is the bit index that sets its value on the * flags value in struct tracer_flags. */ struct tracer_opt { - const char *name; /* Will appear on the trace_options file */ - u32 bit; /* Mask assigned in val field in tracer_flags */ + const char *name; /* Will appear on the trace_options file */ + u32 bit; /* Mask assigned in val field in tracer_flags */ }; /* @@ -307,28 +334,51 @@ struct tracer_opt { */ struct tracer_flags { u32 val; - struct tracer_opt *opts; + struct tracer_opt *opts; }; /* Makes more easy to define a tracer opt */ #define TRACER_OPT(s, b) .name = #s, .bit = b -/* - * A specific tracer, represented by methods that operate on a trace array: + +/** + * struct tracer - a specific tracer and its callbacks to interact with debugfs + * @name: the name chosen to select it on the available_tracers file + * @init: called when one switches to this tracer (echo name > current_tracer) + * @reset: called when one switches to another tracer + * @start: called when tracing is unpaused (echo 1 > tracing_enabled) + * @stop: called when tracing is paused (echo 0 > tracing_enabled) + * @open: called when the trace file is opened + * @pipe_open: called when the trace_pipe file is opened + * @wait_pipe: override how the user waits for traces on trace_pipe + * @close: called when the trace file is released + * @read: override the default read callback on trace_pipe + * @splice_read: override the default splice_read callback on trace_pipe + * @selftest: selftest to run on boot (see trace_selftest.c) + * @print_headers: override the first lines that describe your columns + * @print_line: callback that prints a trace + * @set_flag: signals one of your private flags changed (trace_options file) + * @flags: your private flags */ struct tracer { const char *name; - /* Your tracer should raise a warning if init fails */ int (*init)(struct trace_array *tr); void (*reset)(struct trace_array *tr); void (*start)(struct trace_array *tr); void (*stop)(struct trace_array *tr); void (*open)(struct trace_iterator *iter); void (*pipe_open)(struct trace_iterator *iter); + void (*wait_pipe)(struct trace_iterator *iter); void (*close)(struct trace_iterator *iter); ssize_t (*read)(struct trace_iterator *iter, struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos); + ssize_t (*splice_read)(struct trace_iterator *iter, + struct file *filp, + loff_t *ppos, + struct pipe_inode_info *pipe, + size_t len, + unsigned int flags); #ifdef CONFIG_FTRACE_STARTUP_TEST int (*selftest)(struct tracer *trace, struct trace_array *tr); @@ -339,51 +389,50 @@ struct tracer { int (*set_flag)(u32 old_flags, u32 bit, int set); struct tracer *next; int print_max; - struct tracer_flags *flags; + struct tracer_flags *flags; + struct tracer_stat *stats; }; -struct trace_seq { - unsigned char buffer[PAGE_SIZE]; - unsigned int len; - unsigned int readpos; -}; - -/* - * Trace iterator - used by printout routines who present trace - * results to users and which routines might sleep, etc: - */ -struct trace_iterator { - struct trace_array *tr; - struct tracer *trace; - void *private; - struct ring_buffer_iter *buffer_iter[NR_CPUS]; - - /* The below is zeroed out in pipe_read */ - struct trace_seq seq; - struct trace_entry *ent; - int cpu; - u64 ts; - unsigned long iter_flags; - loff_t pos; - long idx; - - cpumask_var_t started; -}; +#define TRACE_PIPE_ALL_CPU -1 +int tracer_init(struct tracer *t, struct trace_array *tr); int tracing_is_enabled(void); void trace_wake_up(void); void tracing_reset(struct trace_array *tr, int cpu); void tracing_reset_online_cpus(struct trace_array *tr); +void tracing_reset_current(int cpu); +void tracing_reset_current_online_cpus(void); int tracing_open_generic(struct inode *inode, struct file *filp); +struct dentry *trace_create_file(const char *name, + mode_t mode, + struct dentry *parent, + void *data, + const struct file_operations *fops); + struct dentry *tracing_init_dentry(void); void init_tracer_sysprof_debugfs(struct dentry *d_tracer); +struct ring_buffer_event; + +struct ring_buffer_event * +trace_buffer_lock_reserve(struct ring_buffer *buffer, + int type, + unsigned long len, + unsigned long flags, + int pc); +void trace_buffer_unlock_commit(struct ring_buffer *buffer, + struct ring_buffer_event *event, + unsigned long flags, int pc); + struct trace_entry *tracing_get_trace_entry(struct trace_array *tr, struct trace_array_cpu *data); -void tracing_generic_entry_update(struct trace_entry *entry, - unsigned long flags, - int pc); + +struct trace_entry *trace_find_next_entry(struct trace_iterator *iter, + int *ent_cpu, u64 *ent_ts); + +void default_wait_pipe(struct trace_iterator *iter); +void poll_wait_pipe(struct trace_iterator *iter); void ftrace(struct trace_array *tr, struct trace_array_cpu *data, @@ -391,14 +440,11 @@ void ftrace(struct trace_array *tr, unsigned long parent_ip, unsigned long flags, int pc); void tracing_sched_switch_trace(struct trace_array *tr, - struct trace_array_cpu *data, struct task_struct *prev, struct task_struct *next, unsigned long flags, int pc); -void tracing_record_cmdline(struct task_struct *tsk); void tracing_sched_wakeup_trace(struct trace_array *tr, - struct trace_array_cpu *data, struct task_struct *wakee, struct task_struct *cur, unsigned long flags, int pc); @@ -408,14 +454,13 @@ void trace_special(struct trace_array *tr, unsigned long arg2, unsigned long arg3, int pc); void trace_function(struct trace_array *tr, - struct trace_array_cpu *data, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc); void trace_graph_return(struct ftrace_graph_ret *trace); int trace_graph_entry(struct ftrace_graph_ent *trace); -void trace_hw_branch(struct trace_array *tr, u64 from, u64 to); +void set_graph_array(struct trace_array *tr); void tracing_start_cmdline_record(void); void tracing_stop_cmdline_record(void); @@ -427,22 +472,42 @@ void unregister_tracer(struct tracer *type); extern unsigned long nsecs_to_usecs(unsigned long nsecs); +#ifdef CONFIG_TRACER_MAX_TRACE extern unsigned long tracing_max_latency; extern unsigned long tracing_thresh; void update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu); void update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu); +#endif /* CONFIG_TRACER_MAX_TRACE */ -extern cycle_t ftrace_now(int cpu); +#ifdef CONFIG_STACKTRACE +void ftrace_trace_stack(struct ring_buffer *buffer, unsigned long flags, + int skip, int pc); -#ifdef CONFIG_FUNCTION_TRACER -void tracing_start_function_trace(void); -void tracing_stop_function_trace(void); +void ftrace_trace_userstack(struct ring_buffer *buffer, unsigned long flags, + int pc); + +void __trace_stack(struct trace_array *tr, unsigned long flags, int skip, + int pc); #else -# define tracing_start_function_trace() do { } while (0) -# define tracing_stop_function_trace() do { } while (0) -#endif +static inline void ftrace_trace_stack(struct trace_array *tr, + unsigned long flags, int skip, int pc) +{ +} + +static inline void ftrace_trace_userstack(struct trace_array *tr, + unsigned long flags, int pc) +{ +} + +static inline void __trace_stack(struct trace_array *tr, unsigned long flags, + int skip, int pc) +{ +} +#endif /* CONFIG_STACKTRACE */ + +extern cycle_t ftrace_now(int cpu); #ifdef CONFIG_CONTEXT_SWITCH_TRACER typedef void @@ -456,19 +521,25 @@ struct tracer_switch_ops { void *private; struct tracer_switch_ops *next; }; - -char *trace_find_cmdline(int pid); #endif /* CONFIG_CONTEXT_SWITCH_TRACER */ +extern void trace_find_cmdline(int pid, char comm[]); + #ifdef CONFIG_DYNAMIC_FTRACE extern unsigned long ftrace_update_tot_cnt; #define DYN_FTRACE_TEST_NAME trace_selftest_dynamic_test_func extern int DYN_FTRACE_TEST_NAME(void); #endif +extern int ring_buffer_expanded; +extern bool tracing_selftest_disabled; +DECLARE_PER_CPU(local_t, ftrace_cpu_disabled); + #ifdef CONFIG_FTRACE_STARTUP_TEST extern int trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr); +extern int trace_selftest_startup_function_graph(struct tracer *trace, + struct trace_array *tr); extern int trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr); extern int trace_selftest_startup_preemptoff(struct tracer *trace, @@ -485,27 +556,31 @@ extern int trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr); extern int trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr); +extern int trace_selftest_startup_hw_branches(struct tracer *trace, + struct trace_array *tr); #endif /* CONFIG_FTRACE_STARTUP_TEST */ extern void *head_page(struct trace_array_cpu *data); -extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...); -extern void trace_seq_print_cont(struct trace_seq *s, - struct trace_iterator *iter); - +extern unsigned long long ns2usecs(cycle_t nsec); +extern int +trace_vbprintk(unsigned long ip, const char *fmt, va_list args); extern int -seq_print_ip_sym(struct trace_seq *s, unsigned long ip, - unsigned long sym_flags); -extern ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, - size_t cnt); -extern long ns2usecs(cycle_t nsec); +trace_vprintk(unsigned long ip, const char *fmt, va_list args); extern int -trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args); +trace_array_vprintk(struct trace_array *tr, + unsigned long ip, const char *fmt, va_list args); +int trace_array_printk(struct trace_array *tr, + unsigned long ip, const char *fmt, ...); extern unsigned long trace_flags; +extern int trace_clock_id; + /* Standard output formatting function used for function return traces */ #ifdef CONFIG_FUNCTION_GRAPH_TRACER extern enum print_line_t print_graph_function(struct trace_iterator *iter); +extern enum print_line_t +trace_print_graph_duration(unsigned long long duration, struct trace_seq *s); #ifdef CONFIG_DYNAMIC_FTRACE /* TODO: make this variable */ @@ -537,7 +612,6 @@ static inline int ftrace_graph_addr(unsigned long addr) return 1; } #endif /* CONFIG_DYNAMIC_FTRACE */ - #else /* CONFIG_FUNCTION_GRAPH_TRACER */ static inline enum print_line_t print_graph_function(struct trace_iterator *iter) @@ -548,6 +622,7 @@ print_graph_function(struct trace_iterator *iter) extern struct pid *ftrace_pid_trace; +#ifdef CONFIG_FUNCTION_TRACER static inline int ftrace_trace_task(struct task_struct *task) { if (!ftrace_pid_trace) @@ -555,6 +630,12 @@ static inline int ftrace_trace_task(struct task_struct *task) return test_tsk_trace_trace(task); } +#else +static inline int ftrace_trace_task(struct task_struct *task) +{ + return 1; +} +#endif /* * trace_iterator_flags is an enumeration that defines bit @@ -580,7 +661,11 @@ enum trace_iterator_flags { TRACE_ITER_ANNOTATE = 0x2000, TRACE_ITER_USERSTACKTRACE = 0x4000, TRACE_ITER_SYM_USEROBJ = 0x8000, - TRACE_ITER_PRINTK_MSGONLY = 0x10000 + TRACE_ITER_PRINTK_MSGONLY = 0x10000, + TRACE_ITER_CONTEXT_INFO = 0x20000, /* Print pid/cpu/time */ + TRACE_ITER_LATENCY_FMT = 0x40000, + TRACE_ITER_SLEEP_TIME = 0x80000, + TRACE_ITER_GRAPH_TIME = 0x100000, }; /* @@ -601,12 +686,12 @@ extern struct tracer nop_trace; * preempt_enable (after a disable), a schedule might take place * causing an infinite recursion. * - * To prevent this, we read the need_recshed flag before + * To prevent this, we read the need_resched flag before * disabling preemption. When we want to enable preemption we * check the flag, if it is set, then we call preempt_enable_no_resched. * Otherwise, we call preempt_enable. * - * The rational for doing the above is that if need resched is set + * The rational for doing the above is that if need_resched is set * and we have yet to reschedule, we are either in an atomic location * (where we do not need to check for scheduling) or we are inside * the scheduler and do not want to resched. @@ -627,7 +712,7 @@ static inline int ftrace_preempt_disable(void) * * This is a scheduler safe way to enable preemption and not miss * any preemption checks. The disabled saved the state of preemption. - * If resched is set, then we were either inside an atomic or + * If resched is set, then we are either inside an atomic or * are inside the scheduler (we would have already scheduled * otherwise). In this case, we do not want to call normal * preempt_enable, but preempt_enable_no_resched instead. @@ -664,4 +749,132 @@ static inline void trace_branch_disable(void) } #endif /* CONFIG_BRANCH_TRACER */ +/* set ring buffers to default size if not already done so */ +int tracing_update_buffers(void); + +/* trace event type bit fields, not numeric */ +enum { + TRACE_EVENT_TYPE_PRINTF = 1, + TRACE_EVENT_TYPE_RAW = 2, +}; + +struct ftrace_event_field { + struct list_head link; + char *name; + char *type; + int filter_type; + int offset; + int size; + int is_signed; +}; + +struct event_filter { + int n_preds; + struct filter_pred **preds; + char *filter_string; + bool no_reset; +}; + +struct event_subsystem { + struct list_head list; + const char *name; + struct dentry *entry; + struct event_filter *filter; + int nr_events; +}; + +struct filter_pred; + +typedef int (*filter_pred_fn_t) (struct filter_pred *pred, void *event, + int val1, int val2); + +struct filter_pred { + filter_pred_fn_t fn; + u64 val; + char str_val[MAX_FILTER_STR_VAL]; + int str_len; + char *field_name; + int offset; + int not; + int op; + int pop_n; +}; + +extern void print_event_filter(struct ftrace_event_call *call, + struct trace_seq *s); +extern int apply_event_filter(struct ftrace_event_call *call, + char *filter_string); +extern int apply_subsystem_event_filter(struct event_subsystem *system, + char *filter_string); +extern void print_subsystem_event_filter(struct event_subsystem *system, + struct trace_seq *s); +extern int filter_assign_type(const char *type); + +static inline int +filter_check_discard(struct ftrace_event_call *call, void *rec, + struct ring_buffer *buffer, + struct ring_buffer_event *event) +{ + if (unlikely(call->filter_active) && !filter_match_preds(call, rec)) { + ring_buffer_discard_commit(buffer, event); + return 1; + } + + return 0; +} + +#define DEFINE_COMPARISON_PRED(type) \ +static int filter_pred_##type(struct filter_pred *pred, void *event, \ + int val1, int val2) \ +{ \ + type *addr = (type *)(event + pred->offset); \ + type val = (type)pred->val; \ + int match = 0; \ + \ + switch (pred->op) { \ + case OP_LT: \ + match = (*addr < val); \ + break; \ + case OP_LE: \ + match = (*addr <= val); \ + break; \ + case OP_GT: \ + match = (*addr > val); \ + break; \ + case OP_GE: \ + match = (*addr >= val); \ + break; \ + default: \ + break; \ + } \ + \ + return match; \ +} + +#define DEFINE_EQUALITY_PRED(size) \ +static int filter_pred_##size(struct filter_pred *pred, void *event, \ + int val1, int val2) \ +{ \ + u##size *addr = (u##size *)(event + pred->offset); \ + u##size val = (u##size)pred->val; \ + int match; \ + \ + match = (val == *addr) ^ pred->not; \ + \ + return match; \ +} + +extern struct mutex event_mutex; +extern struct list_head ftrace_events; + +extern const char *__start___trace_bprintk_fmt[]; +extern const char *__stop___trace_bprintk_fmt[]; + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ + extern struct ftrace_event_call event_##call; +#undef TRACE_EVENT_FORMAT_NOFILTER +#define TRACE_EVENT_FORMAT_NOFILTER(call, proto, args, fmt, tstruct, tpfmt) +#include "trace_event_types.h" + #endif /* _LINUX_KERNEL_TRACE_H */ diff --git a/kernel/trace/trace_boot.c b/kernel/trace/trace_boot.c index 366c8c333e1..19bfc75d467 100644 --- a/kernel/trace/trace_boot.c +++ b/kernel/trace/trace_boot.c @@ -9,8 +9,10 @@ #include <linux/debugfs.h> #include <linux/ftrace.h> #include <linux/kallsyms.h> +#include <linux/time.h> #include "trace.h" +#include "trace_output.h" static struct trace_array *boot_trace; static bool pre_initcalls_finished; @@ -27,23 +29,24 @@ void start_boot_trace(void) void enable_boot_trace(void) { - if (pre_initcalls_finished) + if (boot_trace && pre_initcalls_finished) tracing_start_sched_switch_record(); } void disable_boot_trace(void) { - if (pre_initcalls_finished) + if (boot_trace && pre_initcalls_finished) tracing_stop_sched_switch_record(); } static int boot_trace_init(struct trace_array *tr) { - int cpu; boot_trace = tr; - for_each_cpu(cpu, cpu_possible_mask) - tracing_reset(tr, cpu); + if (!tr) + return 0; + + tracing_reset_online_cpus(tr); tracing_sched_switch_assign_trace(tr); return 0; @@ -63,7 +66,7 @@ initcall_call_print_line(struct trace_iterator *iter) trace_assign_type(field, entry); call = &field->boot_call; ts = iter->ts; - nsec_rem = do_div(ts, 1000000000); + nsec_rem = do_div(ts, NSEC_PER_SEC); ret = trace_seq_printf(s, "[%5ld.%09ld] calling %s @ %i\n", (unsigned long)ts, nsec_rem, call->func, call->caller); @@ -88,7 +91,7 @@ initcall_ret_print_line(struct trace_iterator *iter) trace_assign_type(field, entry); init_ret = &field->boot_ret; ts = iter->ts; - nsec_rem = do_div(ts, 1000000000); + nsec_rem = do_div(ts, NSEC_PER_SEC); ret = trace_seq_printf(s, "[%5ld.%09ld] initcall %s " "returned %d after %llu msecs\n", @@ -127,11 +130,11 @@ struct tracer boot_tracer __read_mostly = void trace_boot_call(struct boot_trace_call *bt, initcall_t fn) { struct ring_buffer_event *event; + struct ring_buffer *buffer; struct trace_boot_call *entry; - unsigned long irq_flags; struct trace_array *tr = boot_trace; - if (!pre_initcalls_finished) + if (!tr || !pre_initcalls_finished) return; /* Get its name now since this function could @@ -140,18 +143,14 @@ void trace_boot_call(struct boot_trace_call *bt, initcall_t fn) sprint_symbol(bt->func, (unsigned long)fn); preempt_disable(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + buffer = tr->buffer; + event = trace_buffer_lock_reserve(buffer, TRACE_BOOT_CALL, + sizeof(*entry), 0, 0); if (!event) goto out; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_BOOT_CALL; entry->boot_call = *bt; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - - trace_wake_up(); - + trace_buffer_unlock_commit(buffer, event, 0, 0); out: preempt_enable(); } @@ -159,28 +158,24 @@ void trace_boot_call(struct boot_trace_call *bt, initcall_t fn) void trace_boot_ret(struct boot_trace_ret *bt, initcall_t fn) { struct ring_buffer_event *event; + struct ring_buffer *buffer; struct trace_boot_ret *entry; - unsigned long irq_flags; struct trace_array *tr = boot_trace; - if (!pre_initcalls_finished) + if (!tr || !pre_initcalls_finished) return; sprint_symbol(bt->func, (unsigned long)fn); preempt_disable(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + buffer = tr->buffer; + event = trace_buffer_lock_reserve(buffer, TRACE_BOOT_RET, + sizeof(*entry), 0, 0); if (!event) goto out; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_BOOT_RET; entry->boot_ret = *bt; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - - trace_wake_up(); - + trace_buffer_unlock_commit(buffer, event, 0, 0); out: preempt_enable(); } diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 6c00feb3bac..7a7a9fd249a 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -14,21 +14,27 @@ #include <linux/hash.h> #include <linux/fs.h> #include <asm/local.h> + #include "trace.h" +#include "trace_stat.h" +#include "trace_output.h" #ifdef CONFIG_BRANCH_TRACER +static struct tracer branch_trace; static int branch_tracing_enabled __read_mostly; static DEFINE_MUTEX(branch_tracing_mutex); + static struct trace_array *branch_tracer; static void probe_likely_condition(struct ftrace_branch_data *f, int val, int expect) { + struct ftrace_event_call *call = &event_branch; struct trace_array *tr = branch_tracer; struct ring_buffer_event *event; struct trace_branch *entry; - unsigned long flags, irq_flags; + unsigned long flags; int cpu, pc; const char *p; @@ -47,15 +53,13 @@ probe_likely_condition(struct ftrace_branch_data *f, int val, int expect) if (atomic_inc_return(&tr->data[cpu]->disabled) != 1) goto out; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + pc = preempt_count(); + event = trace_buffer_lock_reserve(tr, TRACE_BRANCH, + sizeof(*entry), flags, pc); if (!event) goto out; - pc = preempt_count(); entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_BRANCH; /* Strip off the path, only save the file */ p = f->file + strlen(f->file); @@ -70,7 +74,8 @@ probe_likely_condition(struct ftrace_branch_data *f, int val, int expect) entry->line = f->line; entry->correct = val == expect; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + if (!filter_check_discard(call, entry, tr->buffer, event)) + ring_buffer_unlock_commit(tr->buffer, event); out: atomic_dec(&tr->data[cpu]->disabled); @@ -88,8 +93,6 @@ void trace_likely_condition(struct ftrace_branch_data *f, int val, int expect) int enable_branch_tracing(struct trace_array *tr) { - int ret = 0; - mutex_lock(&branch_tracing_mutex); branch_tracer = tr; /* @@ -100,7 +103,7 @@ int enable_branch_tracing(struct trace_array *tr) branch_tracing_enabled++; mutex_unlock(&branch_tracing_mutex); - return ret; + return 0; } void disable_branch_tracing(void) @@ -128,11 +131,6 @@ static void stop_branch_trace(struct trace_array *tr) static int branch_trace_init(struct trace_array *tr) { - int cpu; - - for_each_online_cpu(cpu) - tracing_reset(tr, cpu); - start_branch_trace(tr); return 0; } @@ -142,22 +140,61 @@ static void branch_trace_reset(struct trace_array *tr) stop_branch_trace(tr); } -struct tracer branch_trace __read_mostly = +static enum print_line_t trace_branch_print(struct trace_iterator *iter, + int flags) +{ + struct trace_branch *field; + + trace_assign_type(field, iter->ent); + + if (trace_seq_printf(&iter->seq, "[%s] %s:%s:%d\n", + field->correct ? " ok " : " MISS ", + field->func, + field->file, + field->line)) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static void branch_print_header(struct seq_file *s) +{ + seq_puts(s, "# TASK-PID CPU# TIMESTAMP CORRECT" + " FUNC:FILE:LINE\n"); + seq_puts(s, "# | | | | | " + " |\n"); +} + +static struct trace_event trace_branch_event = { + .type = TRACE_BRANCH, + .trace = trace_branch_print, +}; + +static struct tracer branch_trace __read_mostly = { .name = "branch", .init = branch_trace_init, .reset = branch_trace_reset, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_branch, -#endif +#endif /* CONFIG_FTRACE_SELFTEST */ + .print_header = branch_print_header, }; -__init static int init_branch_trace(void) +__init static int init_branch_tracer(void) { + int ret; + + ret = register_ftrace_event(&trace_branch_event); + if (!ret) { + printk(KERN_WARNING "Warning: could not register " + "branch events\n"); + return 1; + } return register_tracer(&branch_trace); } +device_initcall(init_branch_tracer); -device_initcall(init_branch_trace); #else static inline void trace_likely_condition(struct ftrace_branch_data *f, int val, int expect) @@ -183,66 +220,39 @@ void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect) } EXPORT_SYMBOL(ftrace_likely_update); -struct ftrace_pointer { - void *start; - void *stop; - int hit; -}; +extern unsigned long __start_annotated_branch_profile[]; +extern unsigned long __stop_annotated_branch_profile[]; -static void * -t_next(struct seq_file *m, void *v, loff_t *pos) +static int annotated_branch_stat_headers(struct seq_file *m) { - const struct ftrace_pointer *f = m->private; - struct ftrace_branch_data *p = v; - - (*pos)++; - - if (v == (void *)1) - return f->start; - - ++p; - - if ((void *)p >= (void *)f->stop) - return NULL; - - return p; + seq_printf(m, " correct incorrect %% "); + seq_printf(m, " Function " + " File Line\n" + " ------- --------- - " + " -------- " + " ---- ----\n"); + return 0; } -static void *t_start(struct seq_file *m, loff_t *pos) +static inline long get_incorrect_percent(struct ftrace_branch_data *p) { - void *t = (void *)1; - loff_t l = 0; - - for (; t && l < *pos; t = t_next(m, t, &l)) - ; + long percent; - return t; -} + if (p->correct) { + percent = p->incorrect * 100; + percent /= p->correct + p->incorrect; + } else + percent = p->incorrect ? 100 : -1; -static void t_stop(struct seq_file *m, void *p) -{ + return percent; } -static int t_show(struct seq_file *m, void *v) +static int branch_stat_show(struct seq_file *m, void *v) { - const struct ftrace_pointer *fp = m->private; struct ftrace_branch_data *p = v; const char *f; long percent; - if (v == (void *)1) { - if (fp->hit) - seq_printf(m, " miss hit %% "); - else - seq_printf(m, " correct incorrect %% "); - seq_printf(m, " Function " - " File Line\n" - " ------- --------- - " - " -------- " - " ---- ----\n"); - return 0; - } - /* Only print the file, not the path */ f = p->file + strlen(p->file); while (f >= p->file && *f != '/') @@ -252,11 +262,7 @@ static int t_show(struct seq_file *m, void *v) /* * The miss is overlayed on correct, and hit on incorrect. */ - if (p->correct) { - percent = p->incorrect * 100; - percent /= p->correct + p->incorrect; - } else - percent = p->incorrect ? 100 : -1; + percent = get_incorrect_percent(p); seq_printf(m, "%8lu %8lu ", p->correct, p->incorrect); if (percent < 0) @@ -267,76 +273,118 @@ static int t_show(struct seq_file *m, void *v) return 0; } -static struct seq_operations tracing_likely_seq_ops = { - .start = t_start, - .next = t_next, - .stop = t_stop, - .show = t_show, +static void *annotated_branch_stat_start(struct tracer_stat *trace) +{ + return __start_annotated_branch_profile; +} + +static void * +annotated_branch_stat_next(void *v, int idx) +{ + struct ftrace_branch_data *p = v; + + ++p; + + if ((void *)p >= (void *)__stop_annotated_branch_profile) + return NULL; + + return p; +} + +static int annotated_branch_stat_cmp(void *p1, void *p2) +{ + struct ftrace_branch_data *a = p1; + struct ftrace_branch_data *b = p2; + + long percent_a, percent_b; + + percent_a = get_incorrect_percent(a); + percent_b = get_incorrect_percent(b); + + if (percent_a < percent_b) + return -1; + if (percent_a > percent_b) + return 1; + else + return 0; +} + +static struct tracer_stat annotated_branch_stats = { + .name = "branch_annotated", + .stat_start = annotated_branch_stat_start, + .stat_next = annotated_branch_stat_next, + .stat_cmp = annotated_branch_stat_cmp, + .stat_headers = annotated_branch_stat_headers, + .stat_show = branch_stat_show }; -static int tracing_branch_open(struct inode *inode, struct file *file) +__init static int init_annotated_branch_stats(void) { int ret; - ret = seq_open(file, &tracing_likely_seq_ops); + ret = register_stat_tracer(&annotated_branch_stats); if (!ret) { - struct seq_file *m = file->private_data; - m->private = (void *)inode->i_private; + printk(KERN_WARNING "Warning: could not register " + "annotated branches stats\n"); + return 1; } - - return ret; + return 0; } - -static const struct file_operations tracing_branch_fops = { - .open = tracing_branch_open, - .read = seq_read, - .llseek = seq_lseek, -}; +fs_initcall(init_annotated_branch_stats); #ifdef CONFIG_PROFILE_ALL_BRANCHES + extern unsigned long __start_branch_profile[]; extern unsigned long __stop_branch_profile[]; -static const struct ftrace_pointer ftrace_branch_pos = { - .start = __start_branch_profile, - .stop = __stop_branch_profile, - .hit = 1, -}; +static int all_branch_stat_headers(struct seq_file *m) +{ + seq_printf(m, " miss hit %% "); + seq_printf(m, " Function " + " File Line\n" + " ------- --------- - " + " -------- " + " ---- ----\n"); + return 0; +} -#endif /* CONFIG_PROFILE_ALL_BRANCHES */ +static void *all_branch_stat_start(struct tracer_stat *trace) +{ + return __start_branch_profile; +} -extern unsigned long __start_annotated_branch_profile[]; -extern unsigned long __stop_annotated_branch_profile[]; +static void * +all_branch_stat_next(void *v, int idx) +{ + struct ftrace_branch_data *p = v; -static const struct ftrace_pointer ftrace_annotated_branch_pos = { - .start = __start_annotated_branch_profile, - .stop = __stop_annotated_branch_profile, -}; + ++p; -static __init int ftrace_branch_init(void) -{ - struct dentry *d_tracer; - struct dentry *entry; + if ((void *)p >= (void *)__stop_branch_profile) + return NULL; - d_tracer = tracing_init_dentry(); + return p; +} - entry = debugfs_create_file("profile_annotated_branch", 0444, d_tracer, - (void *)&ftrace_annotated_branch_pos, - &tracing_branch_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'profile_annotatet_branch' entry\n"); +static struct tracer_stat all_branch_stats = { + .name = "branch_all", + .stat_start = all_branch_stat_start, + .stat_next = all_branch_stat_next, + .stat_headers = all_branch_stat_headers, + .stat_show = branch_stat_show +}; -#ifdef CONFIG_PROFILE_ALL_BRANCHES - entry = debugfs_create_file("profile_branch", 0444, d_tracer, - (void *)&ftrace_branch_pos, - &tracing_branch_fops); - if (!entry) - pr_warning("Could not create debugfs" - " 'profile_branch' entry\n"); -#endif +__init static int all_annotated_branch_stats(void) +{ + int ret; + ret = register_stat_tracer(&all_branch_stats); + if (!ret) { + printk(KERN_WARNING "Warning: could not register " + "all branches stats\n"); + return 1; + } return 0; } - -device_initcall(ftrace_branch_init); +fs_initcall(all_annotated_branch_stats); +#endif /* CONFIG_PROFILE_ALL_BRANCHES */ diff --git a/kernel/trace/trace_clock.c b/kernel/trace/trace_clock.c new file mode 100644 index 00000000000..b588fd81f7f --- /dev/null +++ b/kernel/trace/trace_clock.c @@ -0,0 +1,109 @@ +/* + * tracing clocks + * + * Copyright (C) 2009 Red Hat, Inc., Ingo Molnar <mingo@redhat.com> + * + * Implements 3 trace clock variants, with differing scalability/precision + * tradeoffs: + * + * - local: CPU-local trace clock + * - medium: scalable global clock with some jitter + * - global: globally monotonic, serialized clock + * + * Tracer plugins will chose a default from these clocks. + */ +#include <linux/spinlock.h> +#include <linux/hardirq.h> +#include <linux/module.h> +#include <linux/percpu.h> +#include <linux/sched.h> +#include <linux/ktime.h> +#include <linux/trace_clock.h> + +/* + * trace_clock_local(): the simplest and least coherent tracing clock. + * + * Useful for tracing that does not cross to other CPUs nor + * does it go through idle events. + */ +u64 notrace trace_clock_local(void) +{ + unsigned long flags; + u64 clock; + + /* + * sched_clock() is an architecture implemented, fast, scalable, + * lockless clock. It is not guaranteed to be coherent across + * CPUs, nor across CPU idle events. + */ + raw_local_irq_save(flags); + clock = sched_clock(); + raw_local_irq_restore(flags); + + return clock; +} + +/* + * trace_clock(): 'inbetween' trace clock. Not completely serialized, + * but not completely incorrect when crossing CPUs either. + * + * This is based on cpu_clock(), which will allow at most ~1 jiffy of + * jitter between CPUs. So it's a pretty scalable clock, but there + * can be offsets in the trace data. + */ +u64 notrace trace_clock(void) +{ + return cpu_clock(raw_smp_processor_id()); +} + + +/* + * trace_clock_global(): special globally coherent trace clock + * + * It has higher overhead than the other trace clocks but is still + * an order of magnitude faster than GTOD derived hardware clocks. + * + * Used by plugins that need globally coherent timestamps. + */ + +static u64 prev_trace_clock_time; + +static raw_spinlock_t trace_clock_lock ____cacheline_aligned_in_smp = + (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; + +u64 notrace trace_clock_global(void) +{ + unsigned long flags; + int this_cpu; + u64 now; + + raw_local_irq_save(flags); + + this_cpu = raw_smp_processor_id(); + now = cpu_clock(this_cpu); + /* + * If in an NMI context then dont risk lockups and return the + * cpu_clock() time: + */ + if (unlikely(in_nmi())) + goto out; + + __raw_spin_lock(&trace_clock_lock); + + /* + * TODO: if this happens often then maybe we should reset + * my_scd->clock to prev_trace_clock_time+1, to make sure + * we start ticking with the local clock from now on? + */ + if ((s64)(now - prev_trace_clock_time) < 0) + now = prev_trace_clock_time + 1; + + prev_trace_clock_time = now; + + __raw_spin_unlock(&trace_clock_lock); + + out: + raw_local_irq_restore(flags); + + return now; +} diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c new file mode 100644 index 00000000000..11ba5bb4ed0 --- /dev/null +++ b/kernel/trace/trace_event_profile.c @@ -0,0 +1,39 @@ +/* + * trace event based perf counter profiling + * + * Copyright (C) 2009 Red Hat Inc, Peter Zijlstra <pzijlstr@redhat.com> + * + */ + +#include "trace.h" + +int ftrace_profile_enable(int event_id) +{ + struct ftrace_event_call *event; + int ret = -EINVAL; + + mutex_lock(&event_mutex); + list_for_each_entry(event, &ftrace_events, list) { + if (event->id == event_id && event->profile_enable) { + ret = event->profile_enable(event); + break; + } + } + mutex_unlock(&event_mutex); + + return ret; +} + +void ftrace_profile_disable(int event_id) +{ + struct ftrace_event_call *event; + + mutex_lock(&event_mutex); + list_for_each_entry(event, &ftrace_events, list) { + if (event->id == event_id) { + event->profile_disable(event); + break; + } + } + mutex_unlock(&event_mutex); +} diff --git a/kernel/trace/trace_event_types.h b/kernel/trace/trace_event_types.h new file mode 100644 index 00000000000..6db005e1248 --- /dev/null +++ b/kernel/trace/trace_event_types.h @@ -0,0 +1,178 @@ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM ftrace + +/* + * We cheat and use the proto type field as the ID + * and args as the entry type (minus 'struct') + */ +TRACE_EVENT_FORMAT(function, TRACE_FN, ftrace_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, ip, ip) + TRACE_FIELD(unsigned long, parent_ip, parent_ip) + ), + TP_RAW_FMT(" %lx <-- %lx") +); + +TRACE_EVENT_FORMAT(funcgraph_entry, TRACE_GRAPH_ENT, + ftrace_graph_ent_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, graph_ent.func, func) + TRACE_FIELD(int, graph_ent.depth, depth) + ), + TP_RAW_FMT("--> %lx (%d)") +); + +TRACE_EVENT_FORMAT(funcgraph_exit, TRACE_GRAPH_RET, + ftrace_graph_ret_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, ret.func, func) + TRACE_FIELD(unsigned long long, ret.calltime, calltime) + TRACE_FIELD(unsigned long long, ret.rettime, rettime) + TRACE_FIELD(unsigned long, ret.overrun, overrun) + TRACE_FIELD(int, ret.depth, depth) + ), + TP_RAW_FMT("<-- %lx (%d)") +); + +TRACE_EVENT_FORMAT(wakeup, TRACE_WAKE, ctx_switch_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned int, prev_pid, prev_pid) + TRACE_FIELD(unsigned char, prev_prio, prev_prio) + TRACE_FIELD(unsigned char, prev_state, prev_state) + TRACE_FIELD(unsigned int, next_pid, next_pid) + TRACE_FIELD(unsigned char, next_prio, next_prio) + TRACE_FIELD(unsigned char, next_state, next_state) + TRACE_FIELD(unsigned int, next_cpu, next_cpu) + ), + TP_RAW_FMT("%u:%u:%u ==+ %u:%u:%u [%03u]") +); + +TRACE_EVENT_FORMAT(context_switch, TRACE_CTX, ctx_switch_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned int, prev_pid, prev_pid) + TRACE_FIELD(unsigned char, prev_prio, prev_prio) + TRACE_FIELD(unsigned char, prev_state, prev_state) + TRACE_FIELD(unsigned int, next_pid, next_pid) + TRACE_FIELD(unsigned char, next_prio, next_prio) + TRACE_FIELD(unsigned char, next_state, next_state) + TRACE_FIELD(unsigned int, next_cpu, next_cpu) + ), + TP_RAW_FMT("%u:%u:%u ==+ %u:%u:%u [%03u]") +); + +TRACE_EVENT_FORMAT_NOFILTER(special, TRACE_SPECIAL, special_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, arg1, arg1) + TRACE_FIELD(unsigned long, arg2, arg2) + TRACE_FIELD(unsigned long, arg3, arg3) + ), + TP_RAW_FMT("(%08lx) (%08lx) (%08lx)") +); + +/* + * Stack-trace entry: + */ + +/* #define FTRACE_STACK_ENTRIES 8 */ + +TRACE_EVENT_FORMAT(kernel_stack, TRACE_STACK, stack_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, caller[0], stack0) + TRACE_FIELD(unsigned long, caller[1], stack1) + TRACE_FIELD(unsigned long, caller[2], stack2) + TRACE_FIELD(unsigned long, caller[3], stack3) + TRACE_FIELD(unsigned long, caller[4], stack4) + TRACE_FIELD(unsigned long, caller[5], stack5) + TRACE_FIELD(unsigned long, caller[6], stack6) + TRACE_FIELD(unsigned long, caller[7], stack7) + ), + TP_RAW_FMT("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n" + "\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n") +); + +TRACE_EVENT_FORMAT(user_stack, TRACE_USER_STACK, userstack_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, caller[0], stack0) + TRACE_FIELD(unsigned long, caller[1], stack1) + TRACE_FIELD(unsigned long, caller[2], stack2) + TRACE_FIELD(unsigned long, caller[3], stack3) + TRACE_FIELD(unsigned long, caller[4], stack4) + TRACE_FIELD(unsigned long, caller[5], stack5) + TRACE_FIELD(unsigned long, caller[6], stack6) + TRACE_FIELD(unsigned long, caller[7], stack7) + ), + TP_RAW_FMT("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n" + "\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n") +); + +TRACE_EVENT_FORMAT(bprint, TRACE_BPRINT, bprint_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, ip, ip) + TRACE_FIELD(char *, fmt, fmt) + TRACE_FIELD_ZERO_CHAR(buf) + ), + TP_RAW_FMT("%08lx (%d) fmt:%p %s") +); + +TRACE_EVENT_FORMAT(print, TRACE_PRINT, print_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, ip, ip) + TRACE_FIELD_ZERO_CHAR(buf) + ), + TP_RAW_FMT("%08lx (%d) fmt:%p %s") +); + +TRACE_EVENT_FORMAT(branch, TRACE_BRANCH, trace_branch, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned int, line, line) + TRACE_FIELD_SPECIAL(char func[TRACE_FUNC_SIZE+1], func, + TRACE_FUNC_SIZE+1, func) + TRACE_FIELD_SPECIAL(char file[TRACE_FUNC_SIZE+1], file, + TRACE_FUNC_SIZE+1, file) + TRACE_FIELD(char, correct, correct) + ), + TP_RAW_FMT("%u:%s:%s (%u)") +); + +TRACE_EVENT_FORMAT(hw_branch, TRACE_HW_BRANCHES, hw_branch_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(u64, from, from) + TRACE_FIELD(u64, to, to) + ), + TP_RAW_FMT("from: %llx to: %llx") +); + +TRACE_EVENT_FORMAT(power, TRACE_POWER, trace_power, ignore, + TRACE_STRUCT( + TRACE_FIELD_SIGN(ktime_t, state_data.stamp, stamp, 1) + TRACE_FIELD_SIGN(ktime_t, state_data.end, end, 1) + TRACE_FIELD(int, state_data.type, type) + TRACE_FIELD(int, state_data.state, state) + ), + TP_RAW_FMT("%llx->%llx type:%u state:%u") +); + +TRACE_EVENT_FORMAT(kmem_alloc, TRACE_KMEM_ALLOC, kmemtrace_alloc_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(enum kmemtrace_type_id, type_id, type_id) + TRACE_FIELD(unsigned long, call_site, call_site) + TRACE_FIELD(const void *, ptr, ptr) + TRACE_FIELD(size_t, bytes_req, bytes_req) + TRACE_FIELD(size_t, bytes_alloc, bytes_alloc) + TRACE_FIELD(gfp_t, gfp_flags, gfp_flags) + TRACE_FIELD(int, node, node) + ), + TP_RAW_FMT("type:%u call_site:%lx ptr:%p req:%lu alloc:%lu" + " flags:%x node:%d") +); + +TRACE_EVENT_FORMAT(kmem_free, TRACE_KMEM_FREE, kmemtrace_free_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(enum kmemtrace_type_id, type_id, type_id) + TRACE_FIELD(unsigned long, call_site, call_site) + TRACE_FIELD(const void *, ptr, ptr) + ), + TP_RAW_FMT("type:%u call_site:%lx ptr:%p") +); + +#undef TRACE_SYSTEM diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c new file mode 100644 index 00000000000..78b1ed23017 --- /dev/null +++ b/kernel/trace/trace_events.c @@ -0,0 +1,1505 @@ +/* + * event tracer + * + * Copyright (C) 2008 Red Hat Inc, Steven Rostedt <srostedt@redhat.com> + * + * - Added format output of fields of the trace point. + * This was based off of work by Tom Zanussi <tzanussi@gmail.com>. + * + */ + +#include <linux/workqueue.h> +#include <linux/spinlock.h> +#include <linux/kthread.h> +#include <linux/debugfs.h> +#include <linux/uaccess.h> +#include <linux/module.h> +#include <linux/ctype.h> +#include <linux/delay.h> + +#include <asm/setup.h> + +#include "trace_output.h" + +#define TRACE_SYSTEM "TRACE_SYSTEM" + +DEFINE_MUTEX(event_mutex); + +LIST_HEAD(ftrace_events); + +int trace_define_field(struct ftrace_event_call *call, const char *type, + const char *name, int offset, int size, int is_signed, + int filter_type) +{ + struct ftrace_event_field *field; + + field = kzalloc(sizeof(*field), GFP_KERNEL); + if (!field) + goto err; + + field->name = kstrdup(name, GFP_KERNEL); + if (!field->name) + goto err; + + field->type = kstrdup(type, GFP_KERNEL); + if (!field->type) + goto err; + + if (filter_type == FILTER_OTHER) + field->filter_type = filter_assign_type(type); + else + field->filter_type = filter_type; + + field->offset = offset; + field->size = size; + field->is_signed = is_signed; + + list_add(&field->link, &call->fields); + + return 0; + +err: + if (field) { + kfree(field->name); + kfree(field->type); + } + kfree(field); + + return -ENOMEM; +} +EXPORT_SYMBOL_GPL(trace_define_field); + +#define __common_field(type, item) \ + ret = trace_define_field(call, #type, "common_" #item, \ + offsetof(typeof(ent), item), \ + sizeof(ent.item), \ + is_signed_type(type), FILTER_OTHER); \ + if (ret) \ + return ret; + +int trace_define_common_fields(struct ftrace_event_call *call) +{ + int ret; + struct trace_entry ent; + + __common_field(unsigned short, type); + __common_field(unsigned char, flags); + __common_field(unsigned char, preempt_count); + __common_field(int, pid); + __common_field(int, tgid); + + return ret; +} +EXPORT_SYMBOL_GPL(trace_define_common_fields); + +#ifdef CONFIG_MODULES + +static void trace_destroy_fields(struct ftrace_event_call *call) +{ + struct ftrace_event_field *field, *next; + + list_for_each_entry_safe(field, next, &call->fields, link) { + list_del(&field->link); + kfree(field->type); + kfree(field->name); + kfree(field); + } +} + +#endif /* CONFIG_MODULES */ + +static void ftrace_event_enable_disable(struct ftrace_event_call *call, + int enable) +{ + switch (enable) { + case 0: + if (call->enabled) { + call->enabled = 0; + tracing_stop_cmdline_record(); + call->unregfunc(call->data); + } + break; + case 1: + if (!call->enabled) { + call->enabled = 1; + tracing_start_cmdline_record(); + call->regfunc(call->data); + } + break; + } +} + +static void ftrace_clear_events(void) +{ + struct ftrace_event_call *call; + + mutex_lock(&event_mutex); + list_for_each_entry(call, &ftrace_events, list) { + ftrace_event_enable_disable(call, 0); + } + mutex_unlock(&event_mutex); +} + +/* + * __ftrace_set_clr_event(NULL, NULL, NULL, set) will set/unset all events. + */ +static int __ftrace_set_clr_event(const char *match, const char *sub, + const char *event, int set) +{ + struct ftrace_event_call *call; + int ret = -EINVAL; + + mutex_lock(&event_mutex); + list_for_each_entry(call, &ftrace_events, list) { + + if (!call->name || !call->regfunc) + continue; + + if (match && + strcmp(match, call->name) != 0 && + strcmp(match, call->system) != 0) + continue; + + if (sub && strcmp(sub, call->system) != 0) + continue; + + if (event && strcmp(event, call->name) != 0) + continue; + + ftrace_event_enable_disable(call, set); + + ret = 0; + } + mutex_unlock(&event_mutex); + + return ret; +} + +static int ftrace_set_clr_event(char *buf, int set) +{ + char *event = NULL, *sub = NULL, *match; + + /* + * The buf format can be <subsystem>:<event-name> + * *:<event-name> means any event by that name. + * :<event-name> is the same. + * + * <subsystem>:* means all events in that subsystem + * <subsystem>: means the same. + * + * <name> (no ':') means all events in a subsystem with + * the name <name> or any event that matches <name> + */ + + match = strsep(&buf, ":"); + if (buf) { + sub = match; + event = buf; + match = NULL; + + if (!strlen(sub) || strcmp(sub, "*") == 0) + sub = NULL; + if (!strlen(event) || strcmp(event, "*") == 0) + event = NULL; + } + + return __ftrace_set_clr_event(match, sub, event, set); +} + +/** + * trace_set_clr_event - enable or disable an event + * @system: system name to match (NULL for any system) + * @event: event name to match (NULL for all events, within system) + * @set: 1 to enable, 0 to disable + * + * This is a way for other parts of the kernel to enable or disable + * event recording. + * + * Returns 0 on success, -EINVAL if the parameters do not match any + * registered events. + */ +int trace_set_clr_event(const char *system, const char *event, int set) +{ + return __ftrace_set_clr_event(NULL, system, event, set); +} + +/* 128 should be much more than enough */ +#define EVENT_BUF_SIZE 127 + +static ssize_t +ftrace_event_write(struct file *file, const char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + size_t read = 0; + int i, set = 1; + ssize_t ret; + char *buf; + char ch; + + if (!cnt || cnt < 0) + return 0; + + ret = tracing_update_buffers(); + if (ret < 0) + return ret; + + ret = get_user(ch, ubuf++); + if (ret) + return ret; + read++; + cnt--; + + /* skip white space */ + while (cnt && isspace(ch)) { + ret = get_user(ch, ubuf++); + if (ret) + return ret; + read++; + cnt--; + } + + /* Only white space found? */ + if (isspace(ch)) { + file->f_pos += read; + ret = read; + return ret; + } + + buf = kmalloc(EVENT_BUF_SIZE+1, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + if (cnt > EVENT_BUF_SIZE) + cnt = EVENT_BUF_SIZE; + + i = 0; + while (cnt && !isspace(ch)) { + if (!i && ch == '!') + set = 0; + else + buf[i++] = ch; + + ret = get_user(ch, ubuf++); + if (ret) + goto out_free; + read++; + cnt--; + } + buf[i] = 0; + + file->f_pos += read; + + ret = ftrace_set_clr_event(buf, set); + if (ret) + goto out_free; + + ret = read; + + out_free: + kfree(buf); + + return ret; +} + +static void * +t_next(struct seq_file *m, void *v, loff_t *pos) +{ + struct list_head *list = m->private; + struct ftrace_event_call *call; + + (*pos)++; + + for (;;) { + if (list == &ftrace_events) + return NULL; + + call = list_entry(list, struct ftrace_event_call, list); + + /* + * The ftrace subsystem is for showing formats only. + * They can not be enabled or disabled via the event files. + */ + if (call->regfunc) + break; + + list = list->next; + } + + m->private = list->next; + + return call; +} + +static void *t_start(struct seq_file *m, loff_t *pos) +{ + struct ftrace_event_call *call = NULL; + loff_t l; + + mutex_lock(&event_mutex); + + m->private = ftrace_events.next; + for (l = 0; l <= *pos; ) { + call = t_next(m, NULL, &l); + if (!call) + break; + } + return call; +} + +static void * +s_next(struct seq_file *m, void *v, loff_t *pos) +{ + struct list_head *list = m->private; + struct ftrace_event_call *call; + + (*pos)++; + + retry: + if (list == &ftrace_events) + return NULL; + + call = list_entry(list, struct ftrace_event_call, list); + + if (!call->enabled) { + list = list->next; + goto retry; + } + + m->private = list->next; + + return call; +} + +static void *s_start(struct seq_file *m, loff_t *pos) +{ + struct ftrace_event_call *call = NULL; + loff_t l; + + mutex_lock(&event_mutex); + + m->private = ftrace_events.next; + for (l = 0; l <= *pos; ) { + call = s_next(m, NULL, &l); + if (!call) + break; + } + return call; +} + +static int t_show(struct seq_file *m, void *v) +{ + struct ftrace_event_call *call = v; + + if (strcmp(call->system, TRACE_SYSTEM) != 0) + seq_printf(m, "%s:", call->system); + seq_printf(m, "%s\n", call->name); + + return 0; +} + +static void t_stop(struct seq_file *m, void *p) +{ + mutex_unlock(&event_mutex); +} + +static int +ftrace_event_seq_open(struct inode *inode, struct file *file) +{ + const struct seq_operations *seq_ops; + + if ((file->f_mode & FMODE_WRITE) && + (file->f_flags & O_TRUNC)) + ftrace_clear_events(); + + seq_ops = inode->i_private; + return seq_open(file, seq_ops); +} + +static ssize_t +event_enable_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char *buf; + + if (call->enabled) + buf = "1\n"; + else + buf = "0\n"; + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, 2); +} + +static ssize_t +event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char buf[64]; + unsigned long val; + int ret; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + ret = strict_strtoul(buf, 10, &val); + if (ret < 0) + return ret; + + ret = tracing_update_buffers(); + if (ret < 0) + return ret; + + switch (val) { + case 0: + case 1: + mutex_lock(&event_mutex); + ftrace_event_enable_disable(call, val); + mutex_unlock(&event_mutex); + break; + + default: + return -EINVAL; + } + + *ppos += cnt; + + return cnt; +} + +static ssize_t +system_enable_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + const char set_to_char[4] = { '?', '0', '1', 'X' }; + const char *system = filp->private_data; + struct ftrace_event_call *call; + char buf[2]; + int set = 0; + int ret; + + mutex_lock(&event_mutex); + list_for_each_entry(call, &ftrace_events, list) { + if (!call->name || !call->regfunc) + continue; + + if (system && strcmp(call->system, system) != 0) + continue; + + /* + * We need to find out if all the events are set + * or if all events or cleared, or if we have + * a mixture. + */ + set |= (1 << !!call->enabled); + + /* + * If we have a mixture, no need to look further. + */ + if (set == 3) + break; + } + mutex_unlock(&event_mutex); + + buf[0] = set_to_char[set]; + buf[1] = '\n'; + + ret = simple_read_from_buffer(ubuf, cnt, ppos, buf, 2); + + return ret; +} + +static ssize_t +system_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + const char *system = filp->private_data; + unsigned long val; + char buf[64]; + ssize_t ret; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + ret = strict_strtoul(buf, 10, &val); + if (ret < 0) + return ret; + + ret = tracing_update_buffers(); + if (ret < 0) + return ret; + + if (val != 0 && val != 1) + return -EINVAL; + + ret = __ftrace_set_clr_event(NULL, system, NULL, val); + if (ret) + goto out; + + ret = cnt; + +out: + *ppos += cnt; + + return ret; +} + +extern char *__bad_type_size(void); + +#undef FIELD +#define FIELD(type, name) \ + sizeof(type) != sizeof(field.name) ? __bad_type_size() : \ + #type, "common_" #name, offsetof(typeof(field), name), \ + sizeof(field.name) + +static int trace_write_header(struct trace_seq *s) +{ + struct trace_entry field; + + /* struct trace_entry */ + return trace_seq_printf(s, + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\n", + FIELD(unsigned short, type), + FIELD(unsigned char, flags), + FIELD(unsigned char, preempt_count), + FIELD(int, pid), + FIELD(int, tgid)); +} + +static ssize_t +event_format_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + struct trace_seq *s; + char *buf; + int r; + + if (*ppos) + return 0; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + + /* If any of the first writes fail, so will the show_format. */ + + trace_seq_printf(s, "name: %s\n", call->name); + trace_seq_printf(s, "ID: %d\n", call->id); + trace_seq_printf(s, "format:\n"); + trace_write_header(s); + + r = call->show_format(call, s); + if (!r) { + /* + * ug! The format output is bigger than a PAGE!! + */ + buf = "FORMAT TOO BIG\n"; + r = simple_read_from_buffer(ubuf, cnt, ppos, + buf, strlen(buf)); + goto out; + } + + r = simple_read_from_buffer(ubuf, cnt, ppos, + s->buffer, s->len); + out: + kfree(s); + return r; +} + +static ssize_t +event_id_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + struct trace_seq *s; + int r; + + if (*ppos) + return 0; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + trace_seq_printf(s, "%d\n", call->id); + + r = simple_read_from_buffer(ubuf, cnt, ppos, + s->buffer, s->len); + kfree(s); + return r; +} + +static ssize_t +event_filter_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + struct trace_seq *s; + int r; + + if (*ppos) + return 0; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + + print_event_filter(call, s); + r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len); + + kfree(s); + + return r; +} + +static ssize_t +event_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char *buf; + int err; + + if (cnt >= PAGE_SIZE) + return -EINVAL; + + buf = (char *)__get_free_page(GFP_TEMPORARY); + if (!buf) + return -ENOMEM; + + if (copy_from_user(buf, ubuf, cnt)) { + free_page((unsigned long) buf); + return -EFAULT; + } + buf[cnt] = '\0'; + + err = apply_event_filter(call, buf); + free_page((unsigned long) buf); + if (err < 0) + return err; + + *ppos += cnt; + + return cnt; +} + +static ssize_t +subsystem_filter_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct event_subsystem *system = filp->private_data; + struct trace_seq *s; + int r; + + if (*ppos) + return 0; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + + print_subsystem_event_filter(system, s); + r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len); + + kfree(s); + + return r; +} + +static ssize_t +subsystem_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct event_subsystem *system = filp->private_data; + char *buf; + int err; + + if (cnt >= PAGE_SIZE) + return -EINVAL; + + buf = (char *)__get_free_page(GFP_TEMPORARY); + if (!buf) + return -ENOMEM; + + if (copy_from_user(buf, ubuf, cnt)) { + free_page((unsigned long) buf); + return -EFAULT; + } + buf[cnt] = '\0'; + + err = apply_subsystem_event_filter(system, buf); + free_page((unsigned long) buf); + if (err < 0) + return err; + + *ppos += cnt; + + return cnt; +} + +static ssize_t +show_header(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) +{ + int (*func)(struct trace_seq *s) = filp->private_data; + struct trace_seq *s; + int r; + + if (*ppos) + return 0; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + + func(s); + r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len); + + kfree(s); + + return r; +} + +static const struct seq_operations show_event_seq_ops = { + .start = t_start, + .next = t_next, + .show = t_show, + .stop = t_stop, +}; + +static const struct seq_operations show_set_event_seq_ops = { + .start = s_start, + .next = s_next, + .show = t_show, + .stop = t_stop, +}; + +static const struct file_operations ftrace_avail_fops = { + .open = ftrace_event_seq_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + +static const struct file_operations ftrace_set_event_fops = { + .open = ftrace_event_seq_open, + .read = seq_read, + .write = ftrace_event_write, + .llseek = seq_lseek, + .release = seq_release, +}; + +static const struct file_operations ftrace_enable_fops = { + .open = tracing_open_generic, + .read = event_enable_read, + .write = event_enable_write, +}; + +static const struct file_operations ftrace_event_format_fops = { + .open = tracing_open_generic, + .read = event_format_read, +}; + +static const struct file_operations ftrace_event_id_fops = { + .open = tracing_open_generic, + .read = event_id_read, +}; + +static const struct file_operations ftrace_event_filter_fops = { + .open = tracing_open_generic, + .read = event_filter_read, + .write = event_filter_write, +}; + +static const struct file_operations ftrace_subsystem_filter_fops = { + .open = tracing_open_generic, + .read = subsystem_filter_read, + .write = subsystem_filter_write, +}; + +static const struct file_operations ftrace_system_enable_fops = { + .open = tracing_open_generic, + .read = system_enable_read, + .write = system_enable_write, +}; + +static const struct file_operations ftrace_show_header_fops = { + .open = tracing_open_generic, + .read = show_header, +}; + +static struct dentry *event_trace_events_dir(void) +{ + static struct dentry *d_tracer; + static struct dentry *d_events; + + if (d_events) + return d_events; + + d_tracer = tracing_init_dentry(); + if (!d_tracer) + return NULL; + + d_events = debugfs_create_dir("events", d_tracer); + if (!d_events) + pr_warning("Could not create debugfs " + "'events' directory\n"); + + return d_events; +} + +static LIST_HEAD(event_subsystems); + +static struct dentry * +event_subsystem_dir(const char *name, struct dentry *d_events) +{ + struct event_subsystem *system; + struct dentry *entry; + + /* First see if we did not already create this dir */ + list_for_each_entry(system, &event_subsystems, list) { + if (strcmp(system->name, name) == 0) { + system->nr_events++; + return system->entry; + } + } + + /* need to create new entry */ + system = kmalloc(sizeof(*system), GFP_KERNEL); + if (!system) { + pr_warning("No memory to create event subsystem %s\n", + name); + return d_events; + } + + system->entry = debugfs_create_dir(name, d_events); + if (!system->entry) { + pr_warning("Could not create event subsystem %s\n", + name); + kfree(system); + return d_events; + } + + system->nr_events = 1; + system->name = kstrdup(name, GFP_KERNEL); + if (!system->name) { + debugfs_remove(system->entry); + kfree(system); + return d_events; + } + + list_add(&system->list, &event_subsystems); + + system->filter = NULL; + + system->filter = kzalloc(sizeof(struct event_filter), GFP_KERNEL); + if (!system->filter) { + pr_warning("Could not allocate filter for subsystem " + "'%s'\n", name); + return system->entry; + } + + entry = debugfs_create_file("filter", 0644, system->entry, system, + &ftrace_subsystem_filter_fops); + if (!entry) { + kfree(system->filter); + system->filter = NULL; + pr_warning("Could not create debugfs " + "'%s/filter' entry\n", name); + } + + entry = trace_create_file("enable", 0644, system->entry, + (void *)system->name, + &ftrace_system_enable_fops); + + return system->entry; +} + +static int +event_create_dir(struct ftrace_event_call *call, struct dentry *d_events, + const struct file_operations *id, + const struct file_operations *enable, + const struct file_operations *filter, + const struct file_operations *format) +{ + struct dentry *entry; + int ret; + + /* + * If the trace point header did not define TRACE_SYSTEM + * then the system would be called "TRACE_SYSTEM". + */ + if (strcmp(call->system, TRACE_SYSTEM) != 0) + d_events = event_subsystem_dir(call->system, d_events); + + call->dir = debugfs_create_dir(call->name, d_events); + if (!call->dir) { + pr_warning("Could not create debugfs " + "'%s' directory\n", call->name); + return -1; + } + + if (call->regfunc) + entry = trace_create_file("enable", 0644, call->dir, call, + enable); + + if (call->id && call->profile_enable) + entry = trace_create_file("id", 0444, call->dir, call, + id); + + if (call->define_fields) { + ret = call->define_fields(call); + if (ret < 0) { + pr_warning("Could not initialize trace point" + " events/%s\n", call->name); + return ret; + } + entry = trace_create_file("filter", 0644, call->dir, call, + filter); + } + + /* A trace may not want to export its format */ + if (!call->show_format) + return 0; + + entry = trace_create_file("format", 0444, call->dir, call, + format); + + return 0; +} + +#define for_each_event(event, start, end) \ + for (event = start; \ + (unsigned long)event < (unsigned long)end; \ + event++) + +#ifdef CONFIG_MODULES + +static LIST_HEAD(ftrace_module_file_list); + +/* + * Modules must own their file_operations to keep up with + * reference counting. + */ +struct ftrace_module_file_ops { + struct list_head list; + struct module *mod; + struct file_operations id; + struct file_operations enable; + struct file_operations format; + struct file_operations filter; +}; + +static void remove_subsystem_dir(const char *name) +{ + struct event_subsystem *system; + + if (strcmp(name, TRACE_SYSTEM) == 0) + return; + + list_for_each_entry(system, &event_subsystems, list) { + if (strcmp(system->name, name) == 0) { + if (!--system->nr_events) { + struct event_filter *filter = system->filter; + + debugfs_remove_recursive(system->entry); + list_del(&system->list); + if (filter) { + kfree(filter->filter_string); + kfree(filter); + } + kfree(system->name); + kfree(system); + } + break; + } + } +} + +static struct ftrace_module_file_ops * +trace_create_file_ops(struct module *mod) +{ + struct ftrace_module_file_ops *file_ops; + + /* + * This is a bit of a PITA. To allow for correct reference + * counting, modules must "own" their file_operations. + * To do this, we allocate the file operations that will be + * used in the event directory. + */ + + file_ops = kmalloc(sizeof(*file_ops), GFP_KERNEL); + if (!file_ops) + return NULL; + + file_ops->mod = mod; + + file_ops->id = ftrace_event_id_fops; + file_ops->id.owner = mod; + + file_ops->enable = ftrace_enable_fops; + file_ops->enable.owner = mod; + + file_ops->filter = ftrace_event_filter_fops; + file_ops->filter.owner = mod; + + file_ops->format = ftrace_event_format_fops; + file_ops->format.owner = mod; + + list_add(&file_ops->list, &ftrace_module_file_list); + + return file_ops; +} + +static void trace_module_add_events(struct module *mod) +{ + struct ftrace_module_file_ops *file_ops = NULL; + struct ftrace_event_call *call, *start, *end; + struct dentry *d_events; + int ret; + + start = mod->trace_events; + end = mod->trace_events + mod->num_trace_events; + + if (start == end) + return; + + d_events = event_trace_events_dir(); + if (!d_events) + return; + + for_each_event(call, start, end) { + /* The linker may leave blanks */ + if (!call->name) + continue; + if (call->raw_init) { + ret = call->raw_init(); + if (ret < 0) { + if (ret != -ENOSYS) + pr_warning("Could not initialize trace " + "point events/%s\n", call->name); + continue; + } + } + /* + * This module has events, create file ops for this module + * if not already done. + */ + if (!file_ops) { + file_ops = trace_create_file_ops(mod); + if (!file_ops) + return; + } + call->mod = mod; + list_add(&call->list, &ftrace_events); + event_create_dir(call, d_events, + &file_ops->id, &file_ops->enable, + &file_ops->filter, &file_ops->format); + } +} + +static void trace_module_remove_events(struct module *mod) +{ + struct ftrace_module_file_ops *file_ops; + struct ftrace_event_call *call, *p; + bool found = false; + + down_write(&trace_event_mutex); + list_for_each_entry_safe(call, p, &ftrace_events, list) { + if (call->mod == mod) { + found = true; + ftrace_event_enable_disable(call, 0); + if (call->event) + __unregister_ftrace_event(call->event); + debugfs_remove_recursive(call->dir); + list_del(&call->list); + trace_destroy_fields(call); + destroy_preds(call); + remove_subsystem_dir(call->system); + } + } + + /* Now free the file_operations */ + list_for_each_entry(file_ops, &ftrace_module_file_list, list) { + if (file_ops->mod == mod) + break; + } + if (&file_ops->list != &ftrace_module_file_list) { + list_del(&file_ops->list); + kfree(file_ops); + } + + /* + * It is safest to reset the ring buffer if the module being unloaded + * registered any events. + */ + if (found) + tracing_reset_current_online_cpus(); + up_write(&trace_event_mutex); +} + +static int trace_module_notify(struct notifier_block *self, + unsigned long val, void *data) +{ + struct module *mod = data; + + mutex_lock(&event_mutex); + switch (val) { + case MODULE_STATE_COMING: + trace_module_add_events(mod); + break; + case MODULE_STATE_GOING: + trace_module_remove_events(mod); + break; + } + mutex_unlock(&event_mutex); + + return 0; +} +#else +static int trace_module_notify(struct notifier_block *self, + unsigned long val, void *data) +{ + return 0; +} +#endif /* CONFIG_MODULES */ + +struct notifier_block trace_module_nb = { + .notifier_call = trace_module_notify, + .priority = 0, +}; + +extern struct ftrace_event_call __start_ftrace_events[]; +extern struct ftrace_event_call __stop_ftrace_events[]; + +static char bootup_event_buf[COMMAND_LINE_SIZE] __initdata; + +static __init int setup_trace_event(char *str) +{ + strlcpy(bootup_event_buf, str, COMMAND_LINE_SIZE); + ring_buffer_expanded = 1; + tracing_selftest_disabled = 1; + + return 1; +} +__setup("trace_event=", setup_trace_event); + +static __init int event_trace_init(void) +{ + struct ftrace_event_call *call; + struct dentry *d_tracer; + struct dentry *entry; + struct dentry *d_events; + int ret; + char *buf = bootup_event_buf; + char *token; + + d_tracer = tracing_init_dentry(); + if (!d_tracer) + return 0; + + entry = debugfs_create_file("available_events", 0444, d_tracer, + (void *)&show_event_seq_ops, + &ftrace_avail_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'available_events' entry\n"); + + entry = debugfs_create_file("set_event", 0644, d_tracer, + (void *)&show_set_event_seq_ops, + &ftrace_set_event_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'set_event' entry\n"); + + d_events = event_trace_events_dir(); + if (!d_events) + return 0; + + /* ring buffer internal formats */ + trace_create_file("header_page", 0444, d_events, + ring_buffer_print_page_header, + &ftrace_show_header_fops); + + trace_create_file("header_event", 0444, d_events, + ring_buffer_print_entry_header, + &ftrace_show_header_fops); + + trace_create_file("enable", 0644, d_events, + NULL, &ftrace_system_enable_fops); + + for_each_event(call, __start_ftrace_events, __stop_ftrace_events) { + /* The linker may leave blanks */ + if (!call->name) + continue; + if (call->raw_init) { + ret = call->raw_init(); + if (ret < 0) { + if (ret != -ENOSYS) + pr_warning("Could not initialize trace " + "point events/%s\n", call->name); + continue; + } + } + list_add(&call->list, &ftrace_events); + event_create_dir(call, d_events, &ftrace_event_id_fops, + &ftrace_enable_fops, &ftrace_event_filter_fops, + &ftrace_event_format_fops); + } + + while (true) { + token = strsep(&buf, ","); + + if (!token) + break; + if (!*token) + continue; + + ret = ftrace_set_clr_event(token, 1); + if (ret) + pr_warning("Failed to enable trace event: %s\n", token); + } + + ret = register_module_notifier(&trace_module_nb); + if (ret) + pr_warning("Failed to register trace events module notifier\n"); + + return 0; +} +fs_initcall(event_trace_init); + +#ifdef CONFIG_FTRACE_STARTUP_TEST + +static DEFINE_SPINLOCK(test_spinlock); +static DEFINE_SPINLOCK(test_spinlock_irq); +static DEFINE_MUTEX(test_mutex); + +static __init void test_work(struct work_struct *dummy) +{ + spin_lock(&test_spinlock); + spin_lock_irq(&test_spinlock_irq); + udelay(1); + spin_unlock_irq(&test_spinlock_irq); + spin_unlock(&test_spinlock); + + mutex_lock(&test_mutex); + msleep(1); + mutex_unlock(&test_mutex); +} + +static __init int event_test_thread(void *unused) +{ + void *test_malloc; + + test_malloc = kmalloc(1234, GFP_KERNEL); + if (!test_malloc) + pr_info("failed to kmalloc\n"); + + schedule_on_each_cpu(test_work); + + kfree(test_malloc); + + set_current_state(TASK_INTERRUPTIBLE); + while (!kthread_should_stop()) + schedule(); + + return 0; +} + +/* + * Do various things that may trigger events. + */ +static __init void event_test_stuff(void) +{ + struct task_struct *test_thread; + + test_thread = kthread_run(event_test_thread, NULL, "test-events"); + msleep(1); + kthread_stop(test_thread); +} + +/* + * For every trace event defined, we will test each trace point separately, + * and then by groups, and finally all trace points. + */ +static __init void event_trace_self_tests(void) +{ + struct ftrace_event_call *call; + struct event_subsystem *system; + int ret; + + pr_info("Running tests on trace events:\n"); + + list_for_each_entry(call, &ftrace_events, list) { + + /* Only test those that have a regfunc */ + if (!call->regfunc) + continue; + + pr_info("Testing event %s: ", call->name); + + /* + * If an event is already enabled, someone is using + * it and the self test should not be on. + */ + if (call->enabled) { + pr_warning("Enabled event during self test!\n"); + WARN_ON_ONCE(1); + continue; + } + + ftrace_event_enable_disable(call, 1); + event_test_stuff(); + ftrace_event_enable_disable(call, 0); + + pr_cont("OK\n"); + } + + /* Now test at the sub system level */ + + pr_info("Running tests on trace event systems:\n"); + + list_for_each_entry(system, &event_subsystems, list) { + + /* the ftrace system is special, skip it */ + if (strcmp(system->name, "ftrace") == 0) + continue; + + pr_info("Testing event system %s: ", system->name); + + ret = __ftrace_set_clr_event(NULL, system->name, NULL, 1); + if (WARN_ON_ONCE(ret)) { + pr_warning("error enabling system %s\n", + system->name); + continue; + } + + event_test_stuff(); + + ret = __ftrace_set_clr_event(NULL, system->name, NULL, 0); + if (WARN_ON_ONCE(ret)) + pr_warning("error disabling system %s\n", + system->name); + + pr_cont("OK\n"); + } + + /* Test with all events enabled */ + + pr_info("Running tests on all trace events:\n"); + pr_info("Testing all events: "); + + ret = __ftrace_set_clr_event(NULL, NULL, NULL, 1); + if (WARN_ON_ONCE(ret)) { + pr_warning("error enabling all events\n"); + return; + } + + event_test_stuff(); + + /* reset sysname */ + ret = __ftrace_set_clr_event(NULL, NULL, NULL, 0); + if (WARN_ON_ONCE(ret)) { + pr_warning("error disabling all events\n"); + return; + } + + pr_cont("OK\n"); +} + +#ifdef CONFIG_FUNCTION_TRACER + +static DEFINE_PER_CPU(atomic_t, test_event_disable); + +static void +function_test_events_call(unsigned long ip, unsigned long parent_ip) +{ + struct ring_buffer_event *event; + struct ring_buffer *buffer; + struct ftrace_entry *entry; + unsigned long flags; + long disabled; + int resched; + int cpu; + int pc; + + pc = preempt_count(); + resched = ftrace_preempt_disable(); + cpu = raw_smp_processor_id(); + disabled = atomic_inc_return(&per_cpu(test_event_disable, cpu)); + + if (disabled != 1) + goto out; + + local_save_flags(flags); + + event = trace_current_buffer_lock_reserve(&buffer, + TRACE_FN, sizeof(*entry), + flags, pc); + if (!event) + goto out; + entry = ring_buffer_event_data(event); + entry->ip = ip; + entry->parent_ip = parent_ip; + + trace_nowake_buffer_unlock_commit(buffer, event, flags, pc); + + out: + atomic_dec(&per_cpu(test_event_disable, cpu)); + ftrace_preempt_enable(resched); +} + +static struct ftrace_ops trace_ops __initdata = +{ + .func = function_test_events_call, +}; + +static __init void event_trace_self_test_with_function(void) +{ + register_ftrace_function(&trace_ops); + pr_info("Running tests again, along with the function tracer\n"); + event_trace_self_tests(); + unregister_ftrace_function(&trace_ops); +} +#else +static __init void event_trace_self_test_with_function(void) +{ +} +#endif + +static __init int event_trace_self_tests_init(void) +{ + if (!tracing_selftest_disabled) { + event_trace_self_tests(); + event_trace_self_test_with_function(); + } + + return 0; +} + +late_initcall(event_trace_self_tests_init); + +#endif diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c new file mode 100644 index 00000000000..93660fbbf62 --- /dev/null +++ b/kernel/trace/trace_events_filter.c @@ -0,0 +1,1223 @@ +/* + * trace_events_filter - generic event filtering + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (C) 2009 Tom Zanussi <tzanussi@gmail.com> + */ + +#include <linux/debugfs.h> +#include <linux/uaccess.h> +#include <linux/module.h> +#include <linux/ctype.h> +#include <linux/mutex.h> + +#include "trace.h" +#include "trace_output.h" + +enum filter_op_ids +{ + OP_OR, + OP_AND, + OP_NE, + OP_EQ, + OP_LT, + OP_LE, + OP_GT, + OP_GE, + OP_NONE, + OP_OPEN_PAREN, +}; + +struct filter_op { + int id; + char *string; + int precedence; +}; + +static struct filter_op filter_ops[] = { + { OP_OR, "||", 1 }, + { OP_AND, "&&", 2 }, + { OP_NE, "!=", 4 }, + { OP_EQ, "==", 4 }, + { OP_LT, "<", 5 }, + { OP_LE, "<=", 5 }, + { OP_GT, ">", 5 }, + { OP_GE, ">=", 5 }, + { OP_NONE, "OP_NONE", 0 }, + { OP_OPEN_PAREN, "(", 0 }, +}; + +enum { + FILT_ERR_NONE, + FILT_ERR_INVALID_OP, + FILT_ERR_UNBALANCED_PAREN, + FILT_ERR_TOO_MANY_OPERANDS, + FILT_ERR_OPERAND_TOO_LONG, + FILT_ERR_FIELD_NOT_FOUND, + FILT_ERR_ILLEGAL_FIELD_OP, + FILT_ERR_ILLEGAL_INTVAL, + FILT_ERR_BAD_SUBSYS_FILTER, + FILT_ERR_TOO_MANY_PREDS, + FILT_ERR_MISSING_FIELD, + FILT_ERR_INVALID_FILTER, +}; + +static char *err_text[] = { + "No error", + "Invalid operator", + "Unbalanced parens", + "Too many operands", + "Operand too long", + "Field not found", + "Illegal operation for field type", + "Illegal integer value", + "Couldn't find or set field in one of a subsystem's events", + "Too many terms in predicate expression", + "Missing field name and/or value", + "Meaningless filter expression", +}; + +struct opstack_op { + int op; + struct list_head list; +}; + +struct postfix_elt { + int op; + char *operand; + struct list_head list; +}; + +struct filter_parse_state { + struct filter_op *ops; + struct list_head opstack; + struct list_head postfix; + int lasterr; + int lasterr_pos; + + struct { + char *string; + unsigned int cnt; + unsigned int tail; + } infix; + + struct { + char string[MAX_FILTER_STR_VAL]; + int pos; + unsigned int tail; + } operand; +}; + +DEFINE_COMPARISON_PRED(s64); +DEFINE_COMPARISON_PRED(u64); +DEFINE_COMPARISON_PRED(s32); +DEFINE_COMPARISON_PRED(u32); +DEFINE_COMPARISON_PRED(s16); +DEFINE_COMPARISON_PRED(u16); +DEFINE_COMPARISON_PRED(s8); +DEFINE_COMPARISON_PRED(u8); + +DEFINE_EQUALITY_PRED(64); +DEFINE_EQUALITY_PRED(32); +DEFINE_EQUALITY_PRED(16); +DEFINE_EQUALITY_PRED(8); + +static int filter_pred_and(struct filter_pred *pred __attribute((unused)), + void *event __attribute((unused)), + int val1, int val2) +{ + return val1 && val2; +} + +static int filter_pred_or(struct filter_pred *pred __attribute((unused)), + void *event __attribute((unused)), + int val1, int val2) +{ + return val1 || val2; +} + +/* Filter predicate for fixed sized arrays of characters */ +static int filter_pred_string(struct filter_pred *pred, void *event, + int val1, int val2) +{ + char *addr = (char *)(event + pred->offset); + int cmp, match; + + cmp = strncmp(addr, pred->str_val, pred->str_len); + + match = (!cmp) ^ pred->not; + + return match; +} + +/* Filter predicate for char * pointers */ +static int filter_pred_pchar(struct filter_pred *pred, void *event, + int val1, int val2) +{ + char **addr = (char **)(event + pred->offset); + int cmp, match; + + cmp = strncmp(*addr, pred->str_val, pred->str_len); + + match = (!cmp) ^ pred->not; + + return match; +} + +/* + * Filter predicate for dynamic sized arrays of characters. + * These are implemented through a list of strings at the end + * of the entry. + * Also each of these strings have a field in the entry which + * contains its offset from the beginning of the entry. + * We have then first to get this field, dereference it + * and add it to the address of the entry, and at last we have + * the address of the string. + */ +static int filter_pred_strloc(struct filter_pred *pred, void *event, + int val1, int val2) +{ + u32 str_item = *(u32 *)(event + pred->offset); + int str_loc = str_item & 0xffff; + int str_len = str_item >> 16; + char *addr = (char *)(event + str_loc); + int cmp, match; + + cmp = strncmp(addr, pred->str_val, str_len); + + match = (!cmp) ^ pred->not; + + return match; +} + +static int filter_pred_none(struct filter_pred *pred, void *event, + int val1, int val2) +{ + return 0; +} + +/* return 1 if event matches, 0 otherwise (discard) */ +int filter_match_preds(struct ftrace_event_call *call, void *rec) +{ + struct event_filter *filter = call->filter; + int match, top = 0, val1 = 0, val2 = 0; + int stack[MAX_FILTER_PRED]; + struct filter_pred *pred; + int i; + + for (i = 0; i < filter->n_preds; i++) { + pred = filter->preds[i]; + if (!pred->pop_n) { + match = pred->fn(pred, rec, val1, val2); + stack[top++] = match; + continue; + } + if (pred->pop_n > top) { + WARN_ON_ONCE(1); + return 0; + } + val1 = stack[--top]; + val2 = stack[--top]; + match = pred->fn(pred, rec, val1, val2); + stack[top++] = match; + } + + return stack[--top]; +} +EXPORT_SYMBOL_GPL(filter_match_preds); + +static void parse_error(struct filter_parse_state *ps, int err, int pos) +{ + ps->lasterr = err; + ps->lasterr_pos = pos; +} + +static void remove_filter_string(struct event_filter *filter) +{ + kfree(filter->filter_string); + filter->filter_string = NULL; +} + +static int replace_filter_string(struct event_filter *filter, + char *filter_string) +{ + kfree(filter->filter_string); + filter->filter_string = kstrdup(filter_string, GFP_KERNEL); + if (!filter->filter_string) + return -ENOMEM; + + return 0; +} + +static int append_filter_string(struct event_filter *filter, + char *string) +{ + int newlen; + char *new_filter_string; + + BUG_ON(!filter->filter_string); + newlen = strlen(filter->filter_string) + strlen(string) + 1; + new_filter_string = kmalloc(newlen, GFP_KERNEL); + if (!new_filter_string) + return -ENOMEM; + + strcpy(new_filter_string, filter->filter_string); + strcat(new_filter_string, string); + kfree(filter->filter_string); + filter->filter_string = new_filter_string; + + return 0; +} + +static void append_filter_err(struct filter_parse_state *ps, + struct event_filter *filter) +{ + int pos = ps->lasterr_pos; + char *buf, *pbuf; + + buf = (char *)__get_free_page(GFP_TEMPORARY); + if (!buf) + return; + + append_filter_string(filter, "\n"); + memset(buf, ' ', PAGE_SIZE); + if (pos > PAGE_SIZE - 128) + pos = 0; + buf[pos] = '^'; + pbuf = &buf[pos] + 1; + + sprintf(pbuf, "\nparse_error: %s\n", err_text[ps->lasterr]); + append_filter_string(filter, buf); + free_page((unsigned long) buf); +} + +void print_event_filter(struct ftrace_event_call *call, struct trace_seq *s) +{ + struct event_filter *filter = call->filter; + + mutex_lock(&event_mutex); + if (filter && filter->filter_string) + trace_seq_printf(s, "%s\n", filter->filter_string); + else + trace_seq_printf(s, "none\n"); + mutex_unlock(&event_mutex); +} + +void print_subsystem_event_filter(struct event_subsystem *system, + struct trace_seq *s) +{ + struct event_filter *filter = system->filter; + + mutex_lock(&event_mutex); + if (filter && filter->filter_string) + trace_seq_printf(s, "%s\n", filter->filter_string); + else + trace_seq_printf(s, "none\n"); + mutex_unlock(&event_mutex); +} + +static struct ftrace_event_field * +find_event_field(struct ftrace_event_call *call, char *name) +{ + struct ftrace_event_field *field; + + list_for_each_entry(field, &call->fields, link) { + if (!strcmp(field->name, name)) + return field; + } + + return NULL; +} + +static void filter_free_pred(struct filter_pred *pred) +{ + if (!pred) + return; + + kfree(pred->field_name); + kfree(pred); +} + +static void filter_clear_pred(struct filter_pred *pred) +{ + kfree(pred->field_name); + pred->field_name = NULL; + pred->str_len = 0; +} + +static int filter_set_pred(struct filter_pred *dest, + struct filter_pred *src, + filter_pred_fn_t fn) +{ + *dest = *src; + if (src->field_name) { + dest->field_name = kstrdup(src->field_name, GFP_KERNEL); + if (!dest->field_name) + return -ENOMEM; + } + dest->fn = fn; + + return 0; +} + +static void filter_disable_preds(struct ftrace_event_call *call) +{ + struct event_filter *filter = call->filter; + int i; + + call->filter_active = 0; + filter->n_preds = 0; + + for (i = 0; i < MAX_FILTER_PRED; i++) + filter->preds[i]->fn = filter_pred_none; +} + +void destroy_preds(struct ftrace_event_call *call) +{ + struct event_filter *filter = call->filter; + int i; + + if (!filter) + return; + + for (i = 0; i < MAX_FILTER_PRED; i++) { + if (filter->preds[i]) + filter_free_pred(filter->preds[i]); + } + kfree(filter->preds); + kfree(filter->filter_string); + kfree(filter); + call->filter = NULL; +} + +static int init_preds(struct ftrace_event_call *call) +{ + struct event_filter *filter; + struct filter_pred *pred; + int i; + + if (call->filter) + return 0; + + filter = call->filter = kzalloc(sizeof(*filter), GFP_KERNEL); + if (!call->filter) + return -ENOMEM; + + filter->n_preds = 0; + + filter->preds = kzalloc(MAX_FILTER_PRED * sizeof(pred), GFP_KERNEL); + if (!filter->preds) + goto oom; + + for (i = 0; i < MAX_FILTER_PRED; i++) { + pred = kzalloc(sizeof(*pred), GFP_KERNEL); + if (!pred) + goto oom; + pred->fn = filter_pred_none; + filter->preds[i] = pred; + } + + return 0; + +oom: + destroy_preds(call); + + return -ENOMEM; +} + +static int init_subsystem_preds(struct event_subsystem *system) +{ + struct ftrace_event_call *call; + int err; + + list_for_each_entry(call, &ftrace_events, list) { + if (!call->define_fields) + continue; + + if (strcmp(call->system, system->name) != 0) + continue; + + err = init_preds(call); + if (err) + return err; + } + + return 0; +} + +enum { + FILTER_DISABLE_ALL, + FILTER_INIT_NO_RESET, + FILTER_SKIP_NO_RESET, +}; + +static void filter_free_subsystem_preds(struct event_subsystem *system, + int flag) +{ + struct ftrace_event_call *call; + + list_for_each_entry(call, &ftrace_events, list) { + if (!call->define_fields) + continue; + + if (strcmp(call->system, system->name) != 0) + continue; + + if (flag == FILTER_INIT_NO_RESET) { + call->filter->no_reset = false; + continue; + } + + if (flag == FILTER_SKIP_NO_RESET && call->filter->no_reset) + continue; + + filter_disable_preds(call); + remove_filter_string(call->filter); + } +} + +static int filter_add_pred_fn(struct filter_parse_state *ps, + struct ftrace_event_call *call, + struct filter_pred *pred, + filter_pred_fn_t fn) +{ + struct event_filter *filter = call->filter; + int idx, err; + + if (filter->n_preds == MAX_FILTER_PRED) { + parse_error(ps, FILT_ERR_TOO_MANY_PREDS, 0); + return -ENOSPC; + } + + idx = filter->n_preds; + filter_clear_pred(filter->preds[idx]); + err = filter_set_pred(filter->preds[idx], pred, fn); + if (err) + return err; + + filter->n_preds++; + call->filter_active = 1; + + return 0; +} + +int filter_assign_type(const char *type) +{ + if (strstr(type, "__data_loc") && strstr(type, "char")) + return FILTER_DYN_STRING; + + if (strchr(type, '[') && strstr(type, "char")) + return FILTER_STATIC_STRING; + + return FILTER_OTHER; +} + +static bool is_string_field(struct ftrace_event_field *field) +{ + return field->filter_type == FILTER_DYN_STRING || + field->filter_type == FILTER_STATIC_STRING || + field->filter_type == FILTER_PTR_STRING; +} + +static int is_legal_op(struct ftrace_event_field *field, int op) +{ + if (is_string_field(field) && (op != OP_EQ && op != OP_NE)) + return 0; + + return 1; +} + +static filter_pred_fn_t select_comparison_fn(int op, int field_size, + int field_is_signed) +{ + filter_pred_fn_t fn = NULL; + + switch (field_size) { + case 8: + if (op == OP_EQ || op == OP_NE) + fn = filter_pred_64; + else if (field_is_signed) + fn = filter_pred_s64; + else + fn = filter_pred_u64; + break; + case 4: + if (op == OP_EQ || op == OP_NE) + fn = filter_pred_32; + else if (field_is_signed) + fn = filter_pred_s32; + else + fn = filter_pred_u32; + break; + case 2: + if (op == OP_EQ || op == OP_NE) + fn = filter_pred_16; + else if (field_is_signed) + fn = filter_pred_s16; + else + fn = filter_pred_u16; + break; + case 1: + if (op == OP_EQ || op == OP_NE) + fn = filter_pred_8; + else if (field_is_signed) + fn = filter_pred_s8; + else + fn = filter_pred_u8; + break; + } + + return fn; +} + +static int filter_add_pred(struct filter_parse_state *ps, + struct ftrace_event_call *call, + struct filter_pred *pred, + bool dry_run) +{ + struct ftrace_event_field *field; + filter_pred_fn_t fn; + unsigned long long val; + int ret; + + pred->fn = filter_pred_none; + + if (pred->op == OP_AND) { + pred->pop_n = 2; + fn = filter_pred_and; + goto add_pred_fn; + } else if (pred->op == OP_OR) { + pred->pop_n = 2; + fn = filter_pred_or; + goto add_pred_fn; + } + + field = find_event_field(call, pred->field_name); + if (!field) { + parse_error(ps, FILT_ERR_FIELD_NOT_FOUND, 0); + return -EINVAL; + } + + pred->offset = field->offset; + + if (!is_legal_op(field, pred->op)) { + parse_error(ps, FILT_ERR_ILLEGAL_FIELD_OP, 0); + return -EINVAL; + } + + if (is_string_field(field)) { + pred->str_len = field->size; + + if (field->filter_type == FILTER_STATIC_STRING) + fn = filter_pred_string; + else if (field->filter_type == FILTER_DYN_STRING) + fn = filter_pred_strloc; + else { + fn = filter_pred_pchar; + pred->str_len = strlen(pred->str_val); + } + } else { + if (field->is_signed) + ret = strict_strtoll(pred->str_val, 0, &val); + else + ret = strict_strtoull(pred->str_val, 0, &val); + if (ret) { + parse_error(ps, FILT_ERR_ILLEGAL_INTVAL, 0); + return -EINVAL; + } + pred->val = val; + + fn = select_comparison_fn(pred->op, field->size, + field->is_signed); + if (!fn) { + parse_error(ps, FILT_ERR_INVALID_OP, 0); + return -EINVAL; + } + } + + if (pred->op == OP_NE) + pred->not = 1; + +add_pred_fn: + if (!dry_run) + return filter_add_pred_fn(ps, call, pred, fn); + return 0; +} + +static int filter_add_subsystem_pred(struct filter_parse_state *ps, + struct event_subsystem *system, + struct filter_pred *pred, + char *filter_string, + bool dry_run) +{ + struct ftrace_event_call *call; + int err = 0; + bool fail = true; + + list_for_each_entry(call, &ftrace_events, list) { + + if (!call->define_fields) + continue; + + if (strcmp(call->system, system->name)) + continue; + + if (call->filter->no_reset) + continue; + + err = filter_add_pred(ps, call, pred, dry_run); + if (err) + call->filter->no_reset = true; + else + fail = false; + + if (!dry_run) + replace_filter_string(call->filter, filter_string); + } + + if (fail) { + parse_error(ps, FILT_ERR_BAD_SUBSYS_FILTER, 0); + return err; + } + return 0; +} + +static void parse_init(struct filter_parse_state *ps, + struct filter_op *ops, + char *infix_string) +{ + memset(ps, '\0', sizeof(*ps)); + + ps->infix.string = infix_string; + ps->infix.cnt = strlen(infix_string); + ps->ops = ops; + + INIT_LIST_HEAD(&ps->opstack); + INIT_LIST_HEAD(&ps->postfix); +} + +static char infix_next(struct filter_parse_state *ps) +{ + ps->infix.cnt--; + + return ps->infix.string[ps->infix.tail++]; +} + +static char infix_peek(struct filter_parse_state *ps) +{ + if (ps->infix.tail == strlen(ps->infix.string)) + return 0; + + return ps->infix.string[ps->infix.tail]; +} + +static void infix_advance(struct filter_parse_state *ps) +{ + ps->infix.cnt--; + ps->infix.tail++; +} + +static inline int is_precedence_lower(struct filter_parse_state *ps, + int a, int b) +{ + return ps->ops[a].precedence < ps->ops[b].precedence; +} + +static inline int is_op_char(struct filter_parse_state *ps, char c) +{ + int i; + + for (i = 0; strcmp(ps->ops[i].string, "OP_NONE"); i++) { + if (ps->ops[i].string[0] == c) + return 1; + } + + return 0; +} + +static int infix_get_op(struct filter_parse_state *ps, char firstc) +{ + char nextc = infix_peek(ps); + char opstr[3]; + int i; + + opstr[0] = firstc; + opstr[1] = nextc; + opstr[2] = '\0'; + + for (i = 0; strcmp(ps->ops[i].string, "OP_NONE"); i++) { + if (!strcmp(opstr, ps->ops[i].string)) { + infix_advance(ps); + return ps->ops[i].id; + } + } + + opstr[1] = '\0'; + + for (i = 0; strcmp(ps->ops[i].string, "OP_NONE"); i++) { + if (!strcmp(opstr, ps->ops[i].string)) + return ps->ops[i].id; + } + + return OP_NONE; +} + +static inline void clear_operand_string(struct filter_parse_state *ps) +{ + memset(ps->operand.string, '\0', MAX_FILTER_STR_VAL); + ps->operand.tail = 0; +} + +static inline int append_operand_char(struct filter_parse_state *ps, char c) +{ + if (ps->operand.tail == MAX_FILTER_STR_VAL - 1) + return -EINVAL; + + ps->operand.string[ps->operand.tail++] = c; + + return 0; +} + +static int filter_opstack_push(struct filter_parse_state *ps, int op) +{ + struct opstack_op *opstack_op; + + opstack_op = kmalloc(sizeof(*opstack_op), GFP_KERNEL); + if (!opstack_op) + return -ENOMEM; + + opstack_op->op = op; + list_add(&opstack_op->list, &ps->opstack); + + return 0; +} + +static int filter_opstack_empty(struct filter_parse_state *ps) +{ + return list_empty(&ps->opstack); +} + +static int filter_opstack_top(struct filter_parse_state *ps) +{ + struct opstack_op *opstack_op; + + if (filter_opstack_empty(ps)) + return OP_NONE; + + opstack_op = list_first_entry(&ps->opstack, struct opstack_op, list); + + return opstack_op->op; +} + +static int filter_opstack_pop(struct filter_parse_state *ps) +{ + struct opstack_op *opstack_op; + int op; + + if (filter_opstack_empty(ps)) + return OP_NONE; + + opstack_op = list_first_entry(&ps->opstack, struct opstack_op, list); + op = opstack_op->op; + list_del(&opstack_op->list); + + kfree(opstack_op); + + return op; +} + +static void filter_opstack_clear(struct filter_parse_state *ps) +{ + while (!filter_opstack_empty(ps)) + filter_opstack_pop(ps); +} + +static char *curr_operand(struct filter_parse_state *ps) +{ + return ps->operand.string; +} + +static int postfix_append_operand(struct filter_parse_state *ps, char *operand) +{ + struct postfix_elt *elt; + + elt = kmalloc(sizeof(*elt), GFP_KERNEL); + if (!elt) + return -ENOMEM; + + elt->op = OP_NONE; + elt->operand = kstrdup(operand, GFP_KERNEL); + if (!elt->operand) { + kfree(elt); + return -ENOMEM; + } + + list_add_tail(&elt->list, &ps->postfix); + + return 0; +} + +static int postfix_append_op(struct filter_parse_state *ps, int op) +{ + struct postfix_elt *elt; + + elt = kmalloc(sizeof(*elt), GFP_KERNEL); + if (!elt) + return -ENOMEM; + + elt->op = op; + elt->operand = NULL; + + list_add_tail(&elt->list, &ps->postfix); + + return 0; +} + +static void postfix_clear(struct filter_parse_state *ps) +{ + struct postfix_elt *elt; + + while (!list_empty(&ps->postfix)) { + elt = list_first_entry(&ps->postfix, struct postfix_elt, list); + kfree(elt->operand); + list_del(&elt->list); + } +} + +static int filter_parse(struct filter_parse_state *ps) +{ + int in_string = 0; + int op, top_op; + char ch; + + while ((ch = infix_next(ps))) { + if (ch == '"') { + in_string ^= 1; + continue; + } + + if (in_string) + goto parse_operand; + + if (isspace(ch)) + continue; + + if (is_op_char(ps, ch)) { + op = infix_get_op(ps, ch); + if (op == OP_NONE) { + parse_error(ps, FILT_ERR_INVALID_OP, 0); + return -EINVAL; + } + + if (strlen(curr_operand(ps))) { + postfix_append_operand(ps, curr_operand(ps)); + clear_operand_string(ps); + } + + while (!filter_opstack_empty(ps)) { + top_op = filter_opstack_top(ps); + if (!is_precedence_lower(ps, top_op, op)) { + top_op = filter_opstack_pop(ps); + postfix_append_op(ps, top_op); + continue; + } + break; + } + + filter_opstack_push(ps, op); + continue; + } + + if (ch == '(') { + filter_opstack_push(ps, OP_OPEN_PAREN); + continue; + } + + if (ch == ')') { + if (strlen(curr_operand(ps))) { + postfix_append_operand(ps, curr_operand(ps)); + clear_operand_string(ps); + } + + top_op = filter_opstack_pop(ps); + while (top_op != OP_NONE) { + if (top_op == OP_OPEN_PAREN) + break; + postfix_append_op(ps, top_op); + top_op = filter_opstack_pop(ps); + } + if (top_op == OP_NONE) { + parse_error(ps, FILT_ERR_UNBALANCED_PAREN, 0); + return -EINVAL; + } + continue; + } +parse_operand: + if (append_operand_char(ps, ch)) { + parse_error(ps, FILT_ERR_OPERAND_TOO_LONG, 0); + return -EINVAL; + } + } + + if (strlen(curr_operand(ps))) + postfix_append_operand(ps, curr_operand(ps)); + + while (!filter_opstack_empty(ps)) { + top_op = filter_opstack_pop(ps); + if (top_op == OP_NONE) + break; + if (top_op == OP_OPEN_PAREN) { + parse_error(ps, FILT_ERR_UNBALANCED_PAREN, 0); + return -EINVAL; + } + postfix_append_op(ps, top_op); + } + + return 0; +} + +static struct filter_pred *create_pred(int op, char *operand1, char *operand2) +{ + struct filter_pred *pred; + + pred = kzalloc(sizeof(*pred), GFP_KERNEL); + if (!pred) + return NULL; + + pred->field_name = kstrdup(operand1, GFP_KERNEL); + if (!pred->field_name) { + kfree(pred); + return NULL; + } + + strcpy(pred->str_val, operand2); + pred->str_len = strlen(operand2); + + pred->op = op; + + return pred; +} + +static struct filter_pred *create_logical_pred(int op) +{ + struct filter_pred *pred; + + pred = kzalloc(sizeof(*pred), GFP_KERNEL); + if (!pred) + return NULL; + + pred->op = op; + + return pred; +} + +static int check_preds(struct filter_parse_state *ps) +{ + int n_normal_preds = 0, n_logical_preds = 0; + struct postfix_elt *elt; + + list_for_each_entry(elt, &ps->postfix, list) { + if (elt->op == OP_NONE) + continue; + + if (elt->op == OP_AND || elt->op == OP_OR) { + n_logical_preds++; + continue; + } + n_normal_preds++; + } + + if (!n_normal_preds || n_logical_preds >= n_normal_preds) { + parse_error(ps, FILT_ERR_INVALID_FILTER, 0); + return -EINVAL; + } + + return 0; +} + +static int replace_preds(struct event_subsystem *system, + struct ftrace_event_call *call, + struct filter_parse_state *ps, + char *filter_string, + bool dry_run) +{ + char *operand1 = NULL, *operand2 = NULL; + struct filter_pred *pred; + struct postfix_elt *elt; + int err; + int n_preds = 0; + + err = check_preds(ps); + if (err) + return err; + + list_for_each_entry(elt, &ps->postfix, list) { + if (elt->op == OP_NONE) { + if (!operand1) + operand1 = elt->operand; + else if (!operand2) + operand2 = elt->operand; + else { + parse_error(ps, FILT_ERR_TOO_MANY_OPERANDS, 0); + return -EINVAL; + } + continue; + } + + if (n_preds++ == MAX_FILTER_PRED) { + parse_error(ps, FILT_ERR_TOO_MANY_PREDS, 0); + return -ENOSPC; + } + + if (elt->op == OP_AND || elt->op == OP_OR) { + pred = create_logical_pred(elt->op); + goto add_pred; + } + + if (!operand1 || !operand2) { + parse_error(ps, FILT_ERR_MISSING_FIELD, 0); + return -EINVAL; + } + + pred = create_pred(elt->op, operand1, operand2); +add_pred: + if (!pred) + return -ENOMEM; + if (call) + err = filter_add_pred(ps, call, pred, false); + else + err = filter_add_subsystem_pred(ps, system, pred, + filter_string, dry_run); + filter_free_pred(pred); + if (err) + return err; + + operand1 = operand2 = NULL; + } + + return 0; +} + +int apply_event_filter(struct ftrace_event_call *call, char *filter_string) +{ + int err; + + struct filter_parse_state *ps; + + mutex_lock(&event_mutex); + + err = init_preds(call); + if (err) + goto out_unlock; + + if (!strcmp(strstrip(filter_string), "0")) { + filter_disable_preds(call); + remove_filter_string(call->filter); + mutex_unlock(&event_mutex); + return 0; + } + + err = -ENOMEM; + ps = kzalloc(sizeof(*ps), GFP_KERNEL); + if (!ps) + goto out_unlock; + + filter_disable_preds(call); + replace_filter_string(call->filter, filter_string); + + parse_init(ps, filter_ops, filter_string); + err = filter_parse(ps); + if (err) { + append_filter_err(ps, call->filter); + goto out; + } + + err = replace_preds(NULL, call, ps, filter_string, false); + if (err) + append_filter_err(ps, call->filter); + +out: + filter_opstack_clear(ps); + postfix_clear(ps); + kfree(ps); +out_unlock: + mutex_unlock(&event_mutex); + + return err; +} + +int apply_subsystem_event_filter(struct event_subsystem *system, + char *filter_string) +{ + int err; + + struct filter_parse_state *ps; + + mutex_lock(&event_mutex); + + err = init_subsystem_preds(system); + if (err) + goto out_unlock; + + if (!strcmp(strstrip(filter_string), "0")) { + filter_free_subsystem_preds(system, FILTER_DISABLE_ALL); + remove_filter_string(system->filter); + mutex_unlock(&event_mutex); + return 0; + } + + err = -ENOMEM; + ps = kzalloc(sizeof(*ps), GFP_KERNEL); + if (!ps) + goto out_unlock; + + replace_filter_string(system->filter, filter_string); + + parse_init(ps, filter_ops, filter_string); + err = filter_parse(ps); + if (err) { + append_filter_err(ps, system->filter); + goto out; + } + + filter_free_subsystem_preds(system, FILTER_INIT_NO_RESET); + + /* try to see the filter can be applied to which events */ + err = replace_preds(system, NULL, ps, filter_string, true); + if (err) { + append_filter_err(ps, system->filter); + goto out; + } + + filter_free_subsystem_preds(system, FILTER_SKIP_NO_RESET); + + /* really apply the filter to the events */ + err = replace_preds(system, NULL, ps, filter_string, false); + if (err) { + append_filter_err(ps, system->filter); + filter_free_subsystem_preds(system, 2); + } + +out: + filter_opstack_clear(ps); + postfix_clear(ps); + kfree(ps); +out_unlock: + mutex_unlock(&event_mutex); + + return err; +} + diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c new file mode 100644 index 00000000000..df1bf6e48bb --- /dev/null +++ b/kernel/trace/trace_export.c @@ -0,0 +1,206 @@ +/* + * trace_export.c - export basic ftrace utilities to user space + * + * Copyright (C) 2009 Steven Rostedt <srostedt@redhat.com> + */ +#include <linux/stringify.h> +#include <linux/kallsyms.h> +#include <linux/seq_file.h> +#include <linux/debugfs.h> +#include <linux/uaccess.h> +#include <linux/ftrace.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/fs.h> + +#include "trace_output.h" + + +#undef TRACE_STRUCT +#define TRACE_STRUCT(args...) args + +extern void __bad_type_size(void); + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign) \ + if (sizeof(type) != sizeof(field.item)) \ + __bad_type_size(); \ + ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ + "offset:%u;\tsize:%u;\n", \ + (unsigned int)offsetof(typeof(field), item), \ + (unsigned int)sizeof(field.item)); \ + if (!ret) \ + return 0; + + +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type_item, item, len, cmd) \ + ret = trace_seq_printf(s, "\tfield special:" #type_item ";\t" \ + "offset:%u;\tsize:%u;\n", \ + (unsigned int)offsetof(typeof(field), item), \ + (unsigned int)sizeof(field.item)); \ + if (!ret) \ + return 0; + +#undef TRACE_FIELD_ZERO_CHAR +#define TRACE_FIELD_ZERO_CHAR(item) \ + ret = trace_seq_printf(s, "\tfield:char " #item ";\t" \ + "offset:%u;\tsize:0;\n", \ + (unsigned int)offsetof(typeof(field), item)); \ + if (!ret) \ + return 0; + +#undef TRACE_FIELD_SIGN +#define TRACE_FIELD_SIGN(type, item, assign, is_signed) \ + TRACE_FIELD(type, item, assign) + +#undef TP_RAW_FMT +#define TP_RAW_FMT(args...) args + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +static int \ +ftrace_format_##call(struct ftrace_event_call *unused, \ + struct trace_seq *s) \ +{ \ + struct args field; \ + int ret; \ + \ + tstruct; \ + \ + trace_seq_printf(s, "\nprint fmt: \"%s\"\n", tpfmt); \ + \ + return ret; \ +} + +#undef TRACE_EVENT_FORMAT_NOFILTER +#define TRACE_EVENT_FORMAT_NOFILTER(call, proto, args, fmt, tstruct, \ + tpfmt) \ +static int \ +ftrace_format_##call(struct ftrace_event_call *unused, \ + struct trace_seq *s) \ +{ \ + struct args field; \ + int ret; \ + \ + tstruct; \ + \ + trace_seq_printf(s, "\nprint fmt: \"%s\"\n", tpfmt); \ + \ + return ret; \ +} + +#include "trace_event_types.h" + +#undef TRACE_ZERO_CHAR +#define TRACE_ZERO_CHAR(arg) + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign)\ + entry->item = assign; + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign)\ + entry->item = assign; + +#undef TRACE_FIELD_SIGN +#define TRACE_FIELD_SIGN(type, item, assign, is_signed) \ + TRACE_FIELD(type, item, assign) + +#undef TP_CMD +#define TP_CMD(cmd...) cmd + +#undef TRACE_ENTRY +#define TRACE_ENTRY entry + +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type_item, item, len, cmd) \ + cmd; + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +int ftrace_define_fields_##call(struct ftrace_event_call *event_call); \ +static int ftrace_raw_init_event_##call(void); \ + \ +struct ftrace_event_call __used \ +__attribute__((__aligned__(4))) \ +__attribute__((section("_ftrace_events"))) event_##call = { \ + .name = #call, \ + .id = proto, \ + .system = __stringify(TRACE_SYSTEM), \ + .raw_init = ftrace_raw_init_event_##call, \ + .show_format = ftrace_format_##call, \ + .define_fields = ftrace_define_fields_##call, \ +}; \ +static int ftrace_raw_init_event_##call(void) \ +{ \ + INIT_LIST_HEAD(&event_##call.fields); \ + return 0; \ +} \ + +#undef TRACE_EVENT_FORMAT_NOFILTER +#define TRACE_EVENT_FORMAT_NOFILTER(call, proto, args, fmt, tstruct, \ + tpfmt) \ + \ +struct ftrace_event_call __used \ +__attribute__((__aligned__(4))) \ +__attribute__((section("_ftrace_events"))) event_##call = { \ + .name = #call, \ + .id = proto, \ + .system = __stringify(TRACE_SYSTEM), \ + .show_format = ftrace_format_##call, \ +}; + +#include "trace_event_types.h" + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign) \ + ret = trace_define_field(event_call, #type, #item, \ + offsetof(typeof(field), item), \ + sizeof(field.item), \ + is_signed_type(type), FILTER_OTHER); \ + if (ret) \ + return ret; + +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type, item, len, cmd) \ + ret = trace_define_field(event_call, #type "[" #len "]", #item, \ + offsetof(typeof(field), item), \ + sizeof(field.item), 0, FILTER_OTHER); \ + if (ret) \ + return ret; + +#undef TRACE_FIELD_SIGN +#define TRACE_FIELD_SIGN(type, item, assign, is_signed) \ + ret = trace_define_field(event_call, #type, #item, \ + offsetof(typeof(field), item), \ + sizeof(field.item), is_signed, \ + FILTER_OTHER); \ + if (ret) \ + return ret; + +#undef TRACE_FIELD_ZERO_CHAR +#define TRACE_FIELD_ZERO_CHAR(item) + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +int \ +ftrace_define_fields_##call(struct ftrace_event_call *event_call) \ +{ \ + struct args field; \ + int ret; \ + \ + ret = trace_define_common_fields(event_call); \ + if (ret) \ + return ret; \ + \ + tstruct; \ + \ + return ret; \ +} + +#undef TRACE_EVENT_FORMAT_NOFILTER +#define TRACE_EVENT_FORMAT_NOFILTER(call, proto, args, fmt, tstruct, \ + tpfmt) + +#include "trace_event_types.h" diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 9236d7e25a1..5b01b94518f 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -9,6 +9,7 @@ * Copyright (C) 2004-2006 Ingo Molnar * Copyright (C) 2004 William Lee Irwin III */ +#include <linux/ring_buffer.h> #include <linux/debugfs.h> #include <linux/uaccess.h> #include <linux/ftrace.h> @@ -16,52 +17,387 @@ #include "trace.h" -static void start_function_trace(struct trace_array *tr) +/* function tracing enabled */ +static int ftrace_function_enabled; + +static struct trace_array *func_trace; + +static void tracing_start_function_trace(void); +static void tracing_stop_function_trace(void); + +static int function_trace_init(struct trace_array *tr) { + func_trace = tr; tr->cpu = get_cpu(); - tracing_reset_online_cpus(tr); put_cpu(); tracing_start_cmdline_record(); tracing_start_function_trace(); + return 0; } -static void stop_function_trace(struct trace_array *tr) +static void function_trace_reset(struct trace_array *tr) { tracing_stop_function_trace(); tracing_stop_cmdline_record(); } -static int function_trace_init(struct trace_array *tr) +static void function_trace_start(struct trace_array *tr) { - start_function_trace(tr); - return 0; + tracing_reset_online_cpus(tr); } -static void function_trace_reset(struct trace_array *tr) +static void +function_trace_call_preempt_only(unsigned long ip, unsigned long parent_ip) +{ + struct trace_array *tr = func_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu, resched; + int pc; + + if (unlikely(!ftrace_function_enabled)) + return; + + pc = preempt_count(); + resched = ftrace_preempt_disable(); + local_save_flags(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + + if (likely(disabled == 1)) + trace_function(tr, ip, parent_ip, flags, pc); + + atomic_dec(&data->disabled); + ftrace_preempt_enable(resched); +} + +static void +function_trace_call(unsigned long ip, unsigned long parent_ip) { - stop_function_trace(tr); + struct trace_array *tr = func_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu; + int pc; + + if (unlikely(!ftrace_function_enabled)) + return; + + /* + * Need to use raw, since this must be called before the + * recursive protection is performed. + */ + local_irq_save(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + + if (likely(disabled == 1)) { + pc = preempt_count(); + trace_function(tr, ip, parent_ip, flags, pc); + } + + atomic_dec(&data->disabled); + local_irq_restore(flags); } -static void function_trace_start(struct trace_array *tr) +static void +function_stack_trace_call(unsigned long ip, unsigned long parent_ip) { - tracing_reset_online_cpus(tr); + struct trace_array *tr = func_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu; + int pc; + + if (unlikely(!ftrace_function_enabled)) + return; + + /* + * Need to use raw, since this must be called before the + * recursive protection is performed. + */ + local_irq_save(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + + if (likely(disabled == 1)) { + pc = preempt_count(); + trace_function(tr, ip, parent_ip, flags, pc); + /* + * skip over 5 funcs: + * __ftrace_trace_stack, + * __trace_stack, + * function_stack_trace_call + * ftrace_list_func + * ftrace_call + */ + __trace_stack(tr, flags, 5, pc); + } + + atomic_dec(&data->disabled); + local_irq_restore(flags); +} + + +static struct ftrace_ops trace_ops __read_mostly = +{ + .func = function_trace_call, +}; + +static struct ftrace_ops trace_stack_ops __read_mostly = +{ + .func = function_stack_trace_call, +}; + +/* Our two options */ +enum { + TRACE_FUNC_OPT_STACK = 0x1, +}; + +static struct tracer_opt func_opts[] = { +#ifdef CONFIG_STACKTRACE + { TRACER_OPT(func_stack_trace, TRACE_FUNC_OPT_STACK) }, +#endif + { } /* Always set a last empty entry */ +}; + +static struct tracer_flags func_flags = { + .val = 0, /* By default: all flags disabled */ + .opts = func_opts +}; + +static void tracing_start_function_trace(void) +{ + ftrace_function_enabled = 0; + + if (trace_flags & TRACE_ITER_PREEMPTONLY) + trace_ops.func = function_trace_call_preempt_only; + else + trace_ops.func = function_trace_call; + + if (func_flags.val & TRACE_FUNC_OPT_STACK) + register_ftrace_function(&trace_stack_ops); + else + register_ftrace_function(&trace_ops); + + ftrace_function_enabled = 1; +} + +static void tracing_stop_function_trace(void) +{ + ftrace_function_enabled = 0; + + if (func_flags.val & TRACE_FUNC_OPT_STACK) + unregister_ftrace_function(&trace_stack_ops); + else + unregister_ftrace_function(&trace_ops); +} + +static int func_set_flag(u32 old_flags, u32 bit, int set) +{ + if (bit == TRACE_FUNC_OPT_STACK) { + /* do nothing if already set */ + if (!!set == !!(func_flags.val & TRACE_FUNC_OPT_STACK)) + return 0; + + if (set) { + unregister_ftrace_function(&trace_ops); + register_ftrace_function(&trace_stack_ops); + } else { + unregister_ftrace_function(&trace_stack_ops); + register_ftrace_function(&trace_ops); + } + + return 0; + } + + return -EINVAL; } static struct tracer function_trace __read_mostly = { - .name = "function", - .init = function_trace_init, - .reset = function_trace_reset, - .start = function_trace_start, + .name = "function", + .init = function_trace_init, + .reset = function_trace_reset, + .start = function_trace_start, + .wait_pipe = poll_wait_pipe, + .flags = &func_flags, + .set_flag = func_set_flag, #ifdef CONFIG_FTRACE_SELFTEST - .selftest = trace_selftest_startup_function, + .selftest = trace_selftest_startup_function, #endif }; +#ifdef CONFIG_DYNAMIC_FTRACE +static void +ftrace_traceon(unsigned long ip, unsigned long parent_ip, void **data) +{ + long *count = (long *)data; + + if (tracing_is_on()) + return; + + if (!*count) + return; + + if (*count != -1) + (*count)--; + + tracing_on(); +} + +static void +ftrace_traceoff(unsigned long ip, unsigned long parent_ip, void **data) +{ + long *count = (long *)data; + + if (!tracing_is_on()) + return; + + if (!*count) + return; + + if (*count != -1) + (*count)--; + + tracing_off(); +} + +static int +ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, + struct ftrace_probe_ops *ops, void *data); + +static struct ftrace_probe_ops traceon_probe_ops = { + .func = ftrace_traceon, + .print = ftrace_trace_onoff_print, +}; + +static struct ftrace_probe_ops traceoff_probe_ops = { + .func = ftrace_traceoff, + .print = ftrace_trace_onoff_print, +}; + +static int +ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, + struct ftrace_probe_ops *ops, void *data) +{ + long count = (long)data; + + seq_printf(m, "%pf:", (void *)ip); + + if (ops == &traceon_probe_ops) + seq_printf(m, "traceon"); + else + seq_printf(m, "traceoff"); + + if (count == -1) + seq_printf(m, ":unlimited\n"); + else + seq_printf(m, ":count=%ld\n", count); + + return 0; +} + +static int +ftrace_trace_onoff_unreg(char *glob, char *cmd, char *param) +{ + struct ftrace_probe_ops *ops; + + /* we register both traceon and traceoff to this callback */ + if (strcmp(cmd, "traceon") == 0) + ops = &traceon_probe_ops; + else + ops = &traceoff_probe_ops; + + unregister_ftrace_function_probe_func(glob, ops); + + return 0; +} + +static int +ftrace_trace_onoff_callback(char *glob, char *cmd, char *param, int enable) +{ + struct ftrace_probe_ops *ops; + void *count = (void *)-1; + char *number; + int ret; + + /* hash funcs only work with set_ftrace_filter */ + if (!enable) + return -EINVAL; + + if (glob[0] == '!') + return ftrace_trace_onoff_unreg(glob+1, cmd, param); + + /* we register both traceon and traceoff to this callback */ + if (strcmp(cmd, "traceon") == 0) + ops = &traceon_probe_ops; + else + ops = &traceoff_probe_ops; + + if (!param) + goto out_reg; + + number = strsep(¶m, ":"); + + if (!strlen(number)) + goto out_reg; + + /* + * We use the callback data field (which is a pointer) + * as our counter. + */ + ret = strict_strtoul(number, 0, (unsigned long *)&count); + if (ret) + return ret; + + out_reg: + ret = register_ftrace_function_probe(glob, ops, count); + + return ret < 0 ? ret : 0; +} + +static struct ftrace_func_command ftrace_traceon_cmd = { + .name = "traceon", + .func = ftrace_trace_onoff_callback, +}; + +static struct ftrace_func_command ftrace_traceoff_cmd = { + .name = "traceoff", + .func = ftrace_trace_onoff_callback, +}; + +static int __init init_func_cmd_traceon(void) +{ + int ret; + + ret = register_ftrace_command(&ftrace_traceoff_cmd); + if (ret) + return ret; + + ret = register_ftrace_command(&ftrace_traceon_cmd); + if (ret) + unregister_ftrace_command(&ftrace_traceoff_cmd); + return ret; +} +#else +static inline int init_func_cmd_traceon(void) +{ + return 0; +} +#endif /* CONFIG_DYNAMIC_FTRACE */ + static __init int init_function_trace(void) { + init_func_cmd_traceon(); return register_tracer(&function_trace); } - device_initcall(init_function_trace); + diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 930c08e5b38..b3749a2c313 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -1,7 +1,7 @@ /* * * Function graph tracer. - * Copyright (c) 2008 Frederic Weisbecker <fweisbec@gmail.com> + * Copyright (c) 2008-2009 Frederic Weisbecker <fweisbec@gmail.com> * Mostly borrowed from function tracer which * is Copyright (c) Steven Rostedt <srostedt@redhat.com> * @@ -12,6 +12,12 @@ #include <linux/fs.h> #include "trace.h" +#include "trace_output.h" + +struct fgraph_data { + pid_t last_pid; + int depth; +}; #define TRACE_GRAPH_INDENT 2 @@ -20,9 +26,11 @@ #define TRACE_GRAPH_PRINT_CPU 0x2 #define TRACE_GRAPH_PRINT_OVERHEAD 0x4 #define TRACE_GRAPH_PRINT_PROC 0x8 +#define TRACE_GRAPH_PRINT_DURATION 0x10 +#define TRACE_GRAPH_PRINT_ABS_TIME 0X20 static struct tracer_opt trace_opts[] = { - /* Display overruns ? */ + /* Display overruns? (for self-debug purpose) */ { TRACER_OPT(funcgraph-overrun, TRACE_GRAPH_PRINT_OVERRUN) }, /* Display CPU ? */ { TRACER_OPT(funcgraph-cpu, TRACE_GRAPH_PRINT_CPU) }, @@ -30,27 +38,251 @@ static struct tracer_opt trace_opts[] = { { TRACER_OPT(funcgraph-overhead, TRACE_GRAPH_PRINT_OVERHEAD) }, /* Display proc name/pid */ { TRACER_OPT(funcgraph-proc, TRACE_GRAPH_PRINT_PROC) }, + /* Display duration of execution */ + { TRACER_OPT(funcgraph-duration, TRACE_GRAPH_PRINT_DURATION) }, + /* Display absolute time of an entry */ + { TRACER_OPT(funcgraph-abstime, TRACE_GRAPH_PRINT_ABS_TIME) }, { } /* Empty entry */ }; static struct tracer_flags tracer_flags = { /* Don't display overruns and proc by default */ - .val = TRACE_GRAPH_PRINT_CPU | TRACE_GRAPH_PRINT_OVERHEAD, + .val = TRACE_GRAPH_PRINT_CPU | TRACE_GRAPH_PRINT_OVERHEAD | + TRACE_GRAPH_PRINT_DURATION, .opts = trace_opts }; -/* pid on the last trace processed */ -static pid_t last_pid[NR_CPUS] = { [0 ... NR_CPUS-1] = -1 }; +static struct trace_array *graph_array; -static int graph_trace_init(struct trace_array *tr) + +/* Add a function return address to the trace stack on thread info.*/ +int +ftrace_push_return_trace(unsigned long ret, unsigned long func, int *depth, + unsigned long frame_pointer) +{ + unsigned long long calltime; + int index; + + if (!current->ret_stack) + return -EBUSY; + + /* + * We must make sure the ret_stack is tested before we read + * anything else. + */ + smp_rmb(); + + /* The return trace stack is full */ + if (current->curr_ret_stack == FTRACE_RETFUNC_DEPTH - 1) { + atomic_inc(¤t->trace_overrun); + return -EBUSY; + } + + calltime = trace_clock_local(); + + index = ++current->curr_ret_stack; + barrier(); + current->ret_stack[index].ret = ret; + current->ret_stack[index].func = func; + current->ret_stack[index].calltime = calltime; + current->ret_stack[index].subtime = 0; + current->ret_stack[index].fp = frame_pointer; + *depth = index; + + return 0; +} + +/* Retrieve a function return address to the trace stack on thread info.*/ +static void +ftrace_pop_return_trace(struct ftrace_graph_ret *trace, unsigned long *ret, + unsigned long frame_pointer) { - int cpu, ret; + int index; - for_each_online_cpu(cpu) - tracing_reset(tr, cpu); + index = current->curr_ret_stack; + if (unlikely(index < 0)) { + ftrace_graph_stop(); + WARN_ON(1); + /* Might as well panic, otherwise we have no where to go */ + *ret = (unsigned long)panic; + return; + } + +#ifdef CONFIG_HAVE_FUNCTION_GRAPH_FP_TEST + /* + * The arch may choose to record the frame pointer used + * and check it here to make sure that it is what we expect it + * to be. If gcc does not set the place holder of the return + * address in the frame pointer, and does a copy instead, then + * the function graph trace will fail. This test detects this + * case. + * + * Currently, x86_32 with optimize for size (-Os) makes the latest + * gcc do the above. + */ + if (unlikely(current->ret_stack[index].fp != frame_pointer)) { + ftrace_graph_stop(); + WARN(1, "Bad frame pointer: expected %lx, received %lx\n" + " from func %pF return to %lx\n", + current->ret_stack[index].fp, + frame_pointer, + (void *)current->ret_stack[index].func, + current->ret_stack[index].ret); + *ret = (unsigned long)panic; + return; + } +#endif + + *ret = current->ret_stack[index].ret; + trace->func = current->ret_stack[index].func; + trace->calltime = current->ret_stack[index].calltime; + trace->overrun = atomic_read(¤t->trace_overrun); + trace->depth = index; +} + +/* + * Send the trace to the ring-buffer. + * @return the original return address. + */ +unsigned long ftrace_return_to_handler(unsigned long frame_pointer) +{ + struct ftrace_graph_ret trace; + unsigned long ret; + + ftrace_pop_return_trace(&trace, &ret, frame_pointer); + trace.rettime = trace_clock_local(); + ftrace_graph_return(&trace); + barrier(); + current->curr_ret_stack--; + + if (unlikely(!ret)) { + ftrace_graph_stop(); + WARN_ON(1); + /* Might as well panic. What else to do? */ + ret = (unsigned long)panic; + } + + return ret; +} + +static int __trace_graph_entry(struct trace_array *tr, + struct ftrace_graph_ent *trace, + unsigned long flags, + int pc) +{ + struct ftrace_event_call *call = &event_funcgraph_entry; + struct ring_buffer_event *event; + struct ring_buffer *buffer = tr->buffer; + struct ftrace_graph_ent_entry *entry; + + if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) + return 0; + + event = trace_buffer_lock_reserve(buffer, TRACE_GRAPH_ENT, + sizeof(*entry), flags, pc); + if (!event) + return 0; + entry = ring_buffer_event_data(event); + entry->graph_ent = *trace; + if (!filter_current_check_discard(buffer, call, entry, event)) + ring_buffer_unlock_commit(buffer, event); + + return 1; +} + +int trace_graph_entry(struct ftrace_graph_ent *trace) +{ + struct trace_array *tr = graph_array; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int ret; + int cpu; + int pc; + + if (unlikely(!tr)) + return 0; + + if (!ftrace_trace_task(current)) + return 0; + + if (!ftrace_graph_addr(trace->func)) + return 0; + + local_irq_save(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + if (likely(disabled == 1)) { + pc = preempt_count(); + ret = __trace_graph_entry(tr, trace, flags, pc); + } else { + ret = 0; + } + /* Only do the atomic if it is not already set */ + if (!test_tsk_trace_graph(current)) + set_tsk_trace_graph(current); + + atomic_dec(&data->disabled); + local_irq_restore(flags); + + return ret; +} + +static void __trace_graph_return(struct trace_array *tr, + struct ftrace_graph_ret *trace, + unsigned long flags, + int pc) +{ + struct ftrace_event_call *call = &event_funcgraph_exit; + struct ring_buffer_event *event; + struct ring_buffer *buffer = tr->buffer; + struct ftrace_graph_ret_entry *entry; + + if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) + return; + + event = trace_buffer_lock_reserve(buffer, TRACE_GRAPH_RET, + sizeof(*entry), flags, pc); + if (!event) + return; + entry = ring_buffer_event_data(event); + entry->ret = *trace; + if (!filter_current_check_discard(buffer, call, entry, event)) + ring_buffer_unlock_commit(buffer, event); +} + +void trace_graph_return(struct ftrace_graph_ret *trace) +{ + struct trace_array *tr = graph_array; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu; + int pc; + + local_irq_save(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + if (likely(disabled == 1)) { + pc = preempt_count(); + __trace_graph_return(tr, trace, flags, pc); + } + if (!trace->depth) + clear_tsk_trace_graph(current); + atomic_dec(&data->disabled); + local_irq_restore(flags); +} + +static int graph_trace_init(struct trace_array *tr) +{ + int ret; + + graph_array = tr; ret = register_ftrace_graph(&trace_graph_return, - &trace_graph_entry); + &trace_graph_entry); if (ret) return ret; tracing_start_cmdline_record(); @@ -58,49 +290,30 @@ static int graph_trace_init(struct trace_array *tr) return 0; } +void set_graph_array(struct trace_array *tr) +{ + graph_array = tr; +} + static void graph_trace_reset(struct trace_array *tr) { tracing_stop_cmdline_record(); unregister_ftrace_graph(); } -static inline int log10_cpu(int nb) -{ - if (nb / 100) - return 3; - if (nb / 10) - return 2; - return 1; -} +static int max_bytes_for_cpu; static enum print_line_t print_graph_cpu(struct trace_seq *s, int cpu) { - int i; int ret; - int log10_this = log10_cpu(cpu); - int log10_all = log10_cpu(cpumask_weight(cpu_online_mask)); - /* * Start with a space character - to make it stand out * to the right a bit when trace output is pasted into * email: */ - ret = trace_seq_printf(s, " "); - - /* - * Tricky - we space the CPU field according to the max - * number of online CPUs. On a 2-cpu system it would take - * a maximum of 1 digit - on a 128 cpu system it would - * take up to 3 digits: - */ - for (i = 0; i < log10_all - log10_this; i++) { - ret = trace_seq_printf(s, " "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - ret = trace_seq_printf(s, "%d) ", cpu); + ret = trace_seq_printf(s, " %*d) ", max_bytes_for_cpu, cpu); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -112,15 +325,15 @@ print_graph_cpu(struct trace_seq *s, int cpu) static enum print_line_t print_graph_proc(struct trace_seq *s, pid_t pid) { - int i; - int ret; - int len; - char comm[8]; - int spaces = 0; + char comm[TASK_COMM_LEN]; /* sign + log10(MAX_INT) + '\0' */ char pid_str[11]; + int spaces = 0; + int ret; + int len; + int i; - strncpy(comm, trace_find_cmdline(pid), 7); + trace_find_cmdline(pid, comm); comm[7] = '\0'; sprintf(pid_str, "%d", pid); @@ -153,17 +366,25 @@ print_graph_proc(struct trace_seq *s, pid_t pid) /* If the pid changed since the last trace, output this event */ static enum print_line_t -verif_pid(struct trace_seq *s, pid_t pid, int cpu) +verif_pid(struct trace_seq *s, pid_t pid, int cpu, struct fgraph_data *data) { pid_t prev_pid; + pid_t *last_pid; int ret; - if (last_pid[cpu] != -1 && last_pid[cpu] == pid) + if (!data) + return TRACE_TYPE_HANDLED; + + last_pid = &(per_cpu_ptr(data, cpu)->last_pid); + + if (*last_pid == pid) return TRACE_TYPE_HANDLED; - prev_pid = last_pid[cpu]; - last_pid[cpu] = pid; + prev_pid = *last_pid; + *last_pid = pid; + if (prev_pid == -1) + return TRACE_TYPE_HANDLED; /* * Context-switch trace line: @@ -175,34 +396,34 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu) ret = trace_seq_printf(s, " ------------------------------------------\n"); if (!ret) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = print_graph_cpu(s, cpu); if (ret == TRACE_TYPE_PARTIAL_LINE) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = print_graph_proc(s, prev_pid); if (ret == TRACE_TYPE_PARTIAL_LINE) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = trace_seq_printf(s, " => "); if (!ret) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = print_graph_proc(s, pid); if (ret == TRACE_TYPE_PARTIAL_LINE) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = trace_seq_printf(s, "\n ------------------------------------------\n\n"); if (!ret) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; - return ret; + return TRACE_TYPE_HANDLED; } -static bool -trace_branch_is_leaf(struct trace_iterator *iter, +static struct ftrace_graph_ret_entry * +get_return_for_leaf(struct trace_iterator *iter, struct ftrace_graph_ent_entry *curr) { struct ring_buffer_iter *ring_iter; @@ -211,72 +432,130 @@ trace_branch_is_leaf(struct trace_iterator *iter, ring_iter = iter->buffer_iter[iter->cpu]; - if (!ring_iter) - return false; - - event = ring_buffer_iter_peek(ring_iter, NULL); + /* First peek to compare current entry and the next one */ + if (ring_iter) + event = ring_buffer_iter_peek(ring_iter, NULL); + else { + /* We need to consume the current entry to see the next one */ + ring_buffer_consume(iter->tr->buffer, iter->cpu, NULL); + event = ring_buffer_peek(iter->tr->buffer, iter->cpu, + NULL); + } if (!event) - return false; + return NULL; next = ring_buffer_event_data(event); if (next->ent.type != TRACE_GRAPH_RET) - return false; + return NULL; if (curr->ent.pid != next->ent.pid || curr->graph_ent.func != next->ret.func) - return false; + return NULL; + + /* this is a leaf, now advance the iterator */ + if (ring_iter) + ring_buffer_read(ring_iter, NULL); - return true; + return next; +} + +/* Signal a overhead of time execution to the output */ +static int +print_graph_overhead(unsigned long long duration, struct trace_seq *s) +{ + /* If duration disappear, we don't need anything */ + if (!(tracer_flags.val & TRACE_GRAPH_PRINT_DURATION)) + return 1; + + /* Non nested entry or return */ + if (duration == -1) + return trace_seq_printf(s, " "); + + if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { + /* Duration exceeded 100 msecs */ + if (duration > 100000ULL) + return trace_seq_printf(s, "! "); + + /* Duration exceeded 10 msecs */ + if (duration > 10000ULL) + return trace_seq_printf(s, "+ "); + } + + return trace_seq_printf(s, " "); +} + +static int print_graph_abs_time(u64 t, struct trace_seq *s) +{ + unsigned long usecs_rem; + + usecs_rem = do_div(t, NSEC_PER_SEC); + usecs_rem /= 1000; + + return trace_seq_printf(s, "%5lu.%06lu | ", + (unsigned long)t, usecs_rem); } static enum print_line_t -print_graph_irq(struct trace_seq *s, unsigned long addr, - enum trace_type type, int cpu, pid_t pid) +print_graph_irq(struct trace_iterator *iter, unsigned long addr, + enum trace_type type, int cpu, pid_t pid) { int ret; + struct trace_seq *s = &iter->seq; if (addr < (unsigned long)__irqentry_text_start || addr >= (unsigned long)__irqentry_text_end) return TRACE_TYPE_UNHANDLED; - if (type == TRACE_GRAPH_ENT) { - ret = trace_seq_printf(s, "==========> | "); - } else { - /* Cpu */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { - ret = print_graph_cpu(s, cpu); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - } - /* Proc */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { - ret = print_graph_proc(s, pid); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; + /* Absolute time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { + ret = print_graph_abs_time(iter->ts, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } - ret = trace_seq_printf(s, " | "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + /* Cpu */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { + ret = print_graph_cpu(s, cpu); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + } + /* Proc */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { + ret = print_graph_proc(s, pid); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + ret = trace_seq_printf(s, " | "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } - /* No overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = trace_seq_printf(s, " "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + /* No overhead */ + ret = print_graph_overhead(-1, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + if (type == TRACE_GRAPH_ENT) + ret = trace_seq_printf(s, "==========>"); + else + ret = trace_seq_printf(s, "<=========="); + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Don't close the duration column if haven't one */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) + trace_seq_printf(s, " |"); + ret = trace_seq_printf(s, "\n"); - ret = trace_seq_printf(s, "<========== |\n"); - } if (!ret) return TRACE_TYPE_PARTIAL_LINE; return TRACE_TYPE_HANDLED; } -static enum print_line_t -print_graph_duration(unsigned long long duration, struct trace_seq *s) +enum print_line_t +trace_print_graph_duration(unsigned long long duration, struct trace_seq *s) { unsigned long nsecs_rem = do_div(duration, 1000); /* log10(ULONG_MAX) + '\0' */ @@ -288,7 +567,7 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s) sprintf(msecs_str, "%lu", (unsigned long) duration); /* Print msecs */ - ret = trace_seq_printf(s, msecs_str); + ret = trace_seq_printf(s, "%s", msecs_str); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -313,60 +592,66 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s) if (!ret) return TRACE_TYPE_PARTIAL_LINE; } - - ret = trace_seq_printf(s, "| "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; return TRACE_TYPE_HANDLED; - } -/* Signal a overhead of time execution to the output */ -static int -print_graph_overhead(unsigned long long duration, struct trace_seq *s) +static enum print_line_t +print_graph_duration(unsigned long long duration, struct trace_seq *s) { - /* Duration exceeded 100 msecs */ - if (duration > 100000ULL) - return trace_seq_printf(s, "! "); + int ret; - /* Duration exceeded 10 msecs */ - if (duration > 10000ULL) - return trace_seq_printf(s, "+ "); + ret = trace_print_graph_duration(duration, s); + if (ret != TRACE_TYPE_HANDLED) + return ret; - return trace_seq_printf(s, " "); + ret = trace_seq_printf(s, "| "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; } /* Case of a leaf function on its call entry */ static enum print_line_t print_graph_entry_leaf(struct trace_iterator *iter, - struct ftrace_graph_ent_entry *entry, struct trace_seq *s) + struct ftrace_graph_ent_entry *entry, + struct ftrace_graph_ret_entry *ret_entry, struct trace_seq *s) { - struct ftrace_graph_ret_entry *ret_entry; + struct fgraph_data *data = iter->private; struct ftrace_graph_ret *graph_ret; - struct ring_buffer_event *event; struct ftrace_graph_ent *call; unsigned long long duration; int ret; int i; - event = ring_buffer_read(iter->buffer_iter[iter->cpu], NULL); - ret_entry = ring_buffer_event_data(event); graph_ret = &ret_entry->ret; call = &entry->graph_ent; duration = graph_ret->rettime - graph_ret->calltime; - /* Overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = print_graph_overhead(duration, s); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + if (data) { + int cpu = iter->cpu; + int *depth = &(per_cpu_ptr(data, cpu)->depth); + + /* + * Comments display at + 1 to depth. Since + * this is a leaf function, keep the comments + * equal to this depth. + */ + *depth = call->depth - 1; } - /* Duration */ - ret = print_graph_duration(duration, s); - if (ret == TRACE_TYPE_PARTIAL_LINE) + /* Overhead */ + ret = print_graph_overhead(duration, s); + if (!ret) return TRACE_TYPE_PARTIAL_LINE; + /* Duration */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) { + ret = print_graph_duration(duration, s); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + } + /* Function */ for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { ret = trace_seq_printf(s, " "); @@ -374,11 +659,7 @@ print_graph_entry_leaf(struct trace_iterator *iter, return TRACE_TYPE_PARTIAL_LINE; } - ret = seq_print_ip_sym(s, call->func, 0); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - ret = trace_seq_printf(s, "();\n"); + ret = trace_seq_printf(s, "%pf();\n", (void *)call->func); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -386,33 +667,34 @@ print_graph_entry_leaf(struct trace_iterator *iter, } static enum print_line_t -print_graph_entry_nested(struct ftrace_graph_ent_entry *entry, - struct trace_seq *s, pid_t pid, int cpu) +print_graph_entry_nested(struct trace_iterator *iter, + struct ftrace_graph_ent_entry *entry, + struct trace_seq *s, int cpu) { - int i; - int ret; struct ftrace_graph_ent *call = &entry->graph_ent; + struct fgraph_data *data = iter->private; + int ret; + int i; - /* No overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = trace_seq_printf(s, " "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + if (data) { + int cpu = iter->cpu; + int *depth = &(per_cpu_ptr(data, cpu)->depth); + + *depth = call->depth; } - /* Interrupt */ - ret = print_graph_irq(s, call->func, TRACE_GRAPH_ENT, cpu, pid); - if (ret == TRACE_TYPE_UNHANDLED) { - /* No time */ + /* No overhead */ + ret = print_graph_overhead(-1, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* No time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) { ret = trace_seq_printf(s, " | "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; - } else { - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; } - /* Function */ for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { ret = trace_seq_printf(s, " "); @@ -420,28 +702,44 @@ print_graph_entry_nested(struct ftrace_graph_ent_entry *entry, return TRACE_TYPE_PARTIAL_LINE; } - ret = seq_print_ip_sym(s, call->func, 0); + ret = trace_seq_printf(s, "%pf() {\n", (void *)call->func); if (!ret) return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, "() {\n"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - return TRACE_TYPE_HANDLED; + /* + * we already consumed the current entry to check the next one + * and see if this is a leaf. + */ + return TRACE_TYPE_NO_CONSUME; } static enum print_line_t -print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, - struct trace_iterator *iter, int cpu) +print_graph_prologue(struct trace_iterator *iter, struct trace_seq *s, + int type, unsigned long addr) { - int ret; + struct fgraph_data *data = iter->private; struct trace_entry *ent = iter->ent; + int cpu = iter->cpu; + int ret; /* Pid */ - if (verif_pid(s, ent->pid, cpu) == TRACE_TYPE_PARTIAL_LINE) + if (verif_pid(s, ent->pid, cpu, data) == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; + if (type) { + /* Interrupt */ + ret = print_graph_irq(iter, addr, type, cpu, ent->pid); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + } + + /* Absolute time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { + ret = print_graph_abs_time(iter->ts, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + /* Cpu */ if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { ret = print_graph_cpu(s, cpu); @@ -460,54 +758,65 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; } - if (trace_branch_is_leaf(iter, field)) - return print_graph_entry_leaf(iter, field, s); + return 0; +} + +static enum print_line_t +print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, + struct trace_iterator *iter) +{ + int cpu = iter->cpu; + struct ftrace_graph_ent *call = &field->graph_ent; + struct ftrace_graph_ret_entry *leaf_ret; + + if (print_graph_prologue(iter, s, TRACE_GRAPH_ENT, call->func)) + return TRACE_TYPE_PARTIAL_LINE; + + leaf_ret = get_return_for_leaf(iter, field); + if (leaf_ret) + return print_graph_entry_leaf(iter, field, leaf_ret, s); else - return print_graph_entry_nested(field, s, iter->ent->pid, cpu); + return print_graph_entry_nested(iter, field, s, cpu); } static enum print_line_t print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, - struct trace_entry *ent, int cpu) + struct trace_entry *ent, struct trace_iterator *iter) { - int i; - int ret; unsigned long long duration = trace->rettime - trace->calltime; + struct fgraph_data *data = iter->private; + pid_t pid = ent->pid; + int cpu = iter->cpu; + int ret; + int i; - /* Pid */ - if (verif_pid(s, ent->pid, cpu) == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; + if (data) { + int cpu = iter->cpu; + int *depth = &(per_cpu_ptr(data, cpu)->depth); - /* Cpu */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { - ret = print_graph_cpu(s, cpu); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; + /* + * Comments display at + 1 to depth. This is the + * return from a function, we now want the comments + * to display at the same level of the bracket. + */ + *depth = trace->depth - 1; } - /* Proc */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { - ret = print_graph_proc(s, ent->pid); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - - ret = trace_seq_printf(s, " | "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + if (print_graph_prologue(iter, s, 0, 0)) + return TRACE_TYPE_PARTIAL_LINE; /* Overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = print_graph_overhead(duration, s); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + ret = print_graph_overhead(duration, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; /* Duration */ - ret = print_graph_duration(duration, s); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) { + ret = print_graph_duration(duration, s); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + } /* Closing brace */ for (i = 0; i < trace->depth * TRACE_GRAPH_INDENT; i++) { @@ -528,7 +837,7 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; } - ret = print_graph_irq(s, trace->func, TRACE_GRAPH_RET, cpu, ent->pid); + ret = print_graph_irq(iter, trace->func, TRACE_GRAPH_RET, cpu, pid); if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; @@ -536,61 +845,73 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, } static enum print_line_t -print_graph_comment(struct print_entry *trace, struct trace_seq *s, - struct trace_entry *ent, struct trace_iterator *iter) +print_graph_comment(struct trace_seq *s, struct trace_entry *ent, + struct trace_iterator *iter) { - int i; + unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); + struct fgraph_data *data = iter->private; + struct trace_event *event; + int depth = 0; int ret; + int i; - /* Pid */ - if (verif_pid(s, ent->pid, iter->cpu) == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - - /* Cpu */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { - ret = print_graph_cpu(s, iter->cpu); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - } - - /* Proc */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { - ret = print_graph_proc(s, ent->pid); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; + if (data) + depth = per_cpu_ptr(data, iter->cpu)->depth; - ret = trace_seq_printf(s, " | "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + if (print_graph_prologue(iter, s, 0, 0)) + return TRACE_TYPE_PARTIAL_LINE; /* No overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = trace_seq_printf(s, " "); + ret = print_graph_overhead(-1, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* No time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) { + ret = trace_seq_printf(s, " | "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } - /* No time */ - ret = trace_seq_printf(s, " | "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - /* Indentation */ - if (trace->depth > 0) - for (i = 0; i < (trace->depth + 1) * TRACE_GRAPH_INDENT; i++) { + if (depth > 0) + for (i = 0; i < (depth + 1) * TRACE_GRAPH_INDENT; i++) { ret = trace_seq_printf(s, " "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } /* The comment */ - ret = trace_seq_printf(s, "/* %s", trace->buf); + ret = trace_seq_printf(s, "/* "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; - if (ent->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); + switch (iter->ent->type) { + case TRACE_BPRINT: + ret = trace_print_bprintk_msg_only(iter); + if (ret != TRACE_TYPE_HANDLED) + return ret; + break; + case TRACE_PRINT: + ret = trace_print_printk_msg_only(iter); + if (ret != TRACE_TYPE_HANDLED) + return ret; + break; + default: + event = ftrace_find_event(ent->type); + if (!event) + return TRACE_TYPE_UNHANDLED; + + ret = event->trace(iter, sym_flags); + if (ret != TRACE_TYPE_HANDLED) + return ret; + } + + /* Strip ending newline */ + if (s->buffer[s->len - 1] == '\n') { + s->buffer[s->len - 1] = '\0'; + s->len--; + } ret = trace_seq_printf(s, " */\n"); if (!ret) @@ -603,66 +924,104 @@ print_graph_comment(struct print_entry *trace, struct trace_seq *s, enum print_line_t print_graph_function(struct trace_iterator *iter) { - struct trace_seq *s = &iter->seq; struct trace_entry *entry = iter->ent; + struct trace_seq *s = &iter->seq; switch (entry->type) { case TRACE_GRAPH_ENT: { - struct ftrace_graph_ent_entry *field; + /* + * print_graph_entry() may consume the current event, + * thus @field may become invalid, so we need to save it. + * sizeof(struct ftrace_graph_ent_entry) is very small, + * it can be safely saved at the stack. + */ + struct ftrace_graph_ent_entry *field, saved; trace_assign_type(field, entry); - return print_graph_entry(field, s, iter, - iter->cpu); + saved = *field; + return print_graph_entry(&saved, s, iter); } case TRACE_GRAPH_RET: { struct ftrace_graph_ret_entry *field; trace_assign_type(field, entry); - return print_graph_return(&field->ret, s, entry, iter->cpu); - } - case TRACE_PRINT: { - struct print_entry *field; - trace_assign_type(field, entry); - return print_graph_comment(field, s, entry, iter); + return print_graph_return(&field->ret, s, entry, iter); } default: - return TRACE_TYPE_UNHANDLED; + return print_graph_comment(s, entry, iter); } + + return TRACE_TYPE_HANDLED; } static void print_graph_headers(struct seq_file *s) { /* 1st line */ seq_printf(s, "# "); + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) + seq_printf(s, " TIME "); if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) - seq_printf(s, "CPU "); + seq_printf(s, "CPU"); if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) - seq_printf(s, "TASK/PID "); - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) - seq_printf(s, "OVERHEAD/"); - seq_printf(s, "DURATION FUNCTION CALLS\n"); + seq_printf(s, " TASK/PID "); + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) + seq_printf(s, " DURATION "); + seq_printf(s, " FUNCTION CALLS\n"); /* 2nd line */ seq_printf(s, "# "); + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) + seq_printf(s, " | "); if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) - seq_printf(s, "| "); + seq_printf(s, "| "); if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) - seq_printf(s, "| | "); - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - seq_printf(s, "| "); - seq_printf(s, "| | | | |\n"); - } else - seq_printf(s, " | | | | |\n"); + seq_printf(s, " | | "); + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) + seq_printf(s, " | | "); + seq_printf(s, " | | | |\n"); } + +static void graph_trace_open(struct trace_iterator *iter) +{ + /* pid and depth on the last trace processed */ + struct fgraph_data *data = alloc_percpu(struct fgraph_data); + int cpu; + + if (!data) + pr_warning("function graph tracer: not enough memory\n"); + else + for_each_possible_cpu(cpu) { + pid_t *pid = &(per_cpu_ptr(data, cpu)->last_pid); + int *depth = &(per_cpu_ptr(data, cpu)->depth); + *pid = -1; + *depth = 0; + } + + iter->private = data; +} + +static void graph_trace_close(struct trace_iterator *iter) +{ + free_percpu(iter->private); +} + static struct tracer graph_trace __read_mostly = { - .name = "function_graph", - .init = graph_trace_init, - .reset = graph_trace_reset, + .name = "function_graph", + .open = graph_trace_open, + .close = graph_trace_close, + .wait_pipe = poll_wait_pipe, + .init = graph_trace_init, + .reset = graph_trace_reset, .print_line = print_graph_function, .print_header = print_graph_headers, .flags = &tracer_flags, +#ifdef CONFIG_FTRACE_SELFTEST + .selftest = trace_selftest_startup_function_graph, +#endif }; static __init int init_graph_trace(void) { + max_bytes_for_cpu = snprintf(NULL, 0, "%d", nr_cpu_ids - 1); + return register_tracer(&graph_trace); } diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index 649df22d435..ca7d7c4d0c2 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -1,89 +1,147 @@ /* - * h/w branch tracer for x86 based on bts - * - * Copyright (C) 2008 Markus Metzger <markus.t.metzger@gmail.com> + * h/w branch tracer for x86 based on BTS * + * Copyright (C) 2008-2009 Intel Corporation. + * Markus Metzger <markus.t.metzger@gmail.com>, 2008-2009 */ - -#include <linux/module.h> -#include <linux/fs.h> +#include <linux/kallsyms.h> #include <linux/debugfs.h> #include <linux/ftrace.h> -#include <linux/kallsyms.h> +#include <linux/module.h> +#include <linux/cpu.h> +#include <linux/smp.h> +#include <linux/fs.h> #include <asm/ds.h> +#include "trace_output.h" #include "trace.h" -#define SIZEOF_BTS (1 << 13) +#define BTS_BUFFER_SIZE (1 << 13) static DEFINE_PER_CPU(struct bts_tracer *, tracer); -static DEFINE_PER_CPU(unsigned char[SIZEOF_BTS], buffer); +static DEFINE_PER_CPU(unsigned char[BTS_BUFFER_SIZE], buffer); #define this_tracer per_cpu(tracer, smp_processor_id()) -#define this_buffer per_cpu(buffer, smp_processor_id()) +static int trace_hw_branches_enabled __read_mostly; +static int trace_hw_branches_suspended __read_mostly; +static struct trace_array *hw_branch_trace __read_mostly; -static void bts_trace_start_cpu(void *arg) + +static void bts_trace_init_cpu(int cpu) { - if (this_tracer) - ds_release_bts(this_tracer); + per_cpu(tracer, cpu) = + ds_request_bts_cpu(cpu, per_cpu(buffer, cpu), BTS_BUFFER_SIZE, + NULL, (size_t)-1, BTS_KERNEL); - this_tracer = - ds_request_bts(/* task = */ NULL, this_buffer, SIZEOF_BTS, - /* ovfl = */ NULL, /* th = */ (size_t)-1, - BTS_KERNEL); - if (IS_ERR(this_tracer)) { - this_tracer = NULL; - return; - } + if (IS_ERR(per_cpu(tracer, cpu))) + per_cpu(tracer, cpu) = NULL; } -static void bts_trace_start(struct trace_array *tr) +static int bts_trace_init(struct trace_array *tr) { int cpu; - tracing_reset_online_cpus(tr); + hw_branch_trace = tr; + trace_hw_branches_enabled = 0; + + get_online_cpus(); + for_each_online_cpu(cpu) { + bts_trace_init_cpu(cpu); + + if (likely(per_cpu(tracer, cpu))) + trace_hw_branches_enabled = 1; + } + trace_hw_branches_suspended = 0; + put_online_cpus(); - for_each_cpu(cpu, cpu_possible_mask) - smp_call_function_single(cpu, bts_trace_start_cpu, NULL, 1); + /* If we could not enable tracing on a single cpu, we fail. */ + return trace_hw_branches_enabled ? 0 : -EOPNOTSUPP; } -static void bts_trace_stop_cpu(void *arg) +static void bts_trace_reset(struct trace_array *tr) { - if (this_tracer) { - ds_release_bts(this_tracer); - this_tracer = NULL; + int cpu; + + get_online_cpus(); + for_each_online_cpu(cpu) { + if (likely(per_cpu(tracer, cpu))) { + ds_release_bts(per_cpu(tracer, cpu)); + per_cpu(tracer, cpu) = NULL; + } } + trace_hw_branches_enabled = 0; + trace_hw_branches_suspended = 0; + put_online_cpus(); +} + +static void bts_trace_start(struct trace_array *tr) +{ + int cpu; + + get_online_cpus(); + for_each_online_cpu(cpu) + if (likely(per_cpu(tracer, cpu))) + ds_resume_bts(per_cpu(tracer, cpu)); + trace_hw_branches_suspended = 0; + put_online_cpus(); } static void bts_trace_stop(struct trace_array *tr) { int cpu; - for_each_cpu(cpu, cpu_possible_mask) - smp_call_function_single(cpu, bts_trace_stop_cpu, NULL, 1); + get_online_cpus(); + for_each_online_cpu(cpu) + if (likely(per_cpu(tracer, cpu))) + ds_suspend_bts(per_cpu(tracer, cpu)); + trace_hw_branches_suspended = 1; + put_online_cpus(); } -static int bts_trace_init(struct trace_array *tr) +static int __cpuinit bts_hotcpu_handler(struct notifier_block *nfb, + unsigned long action, void *hcpu) { - tracing_reset_online_cpus(tr); - bts_trace_start(tr); + int cpu = (long)hcpu; + + switch (action) { + case CPU_ONLINE: + case CPU_DOWN_FAILED: + /* The notification is sent with interrupts enabled. */ + if (trace_hw_branches_enabled) { + bts_trace_init_cpu(cpu); + + if (trace_hw_branches_suspended && + likely(per_cpu(tracer, cpu))) + ds_suspend_bts(per_cpu(tracer, cpu)); + } + break; - return 0; + case CPU_DOWN_PREPARE: + /* The notification is sent with interrupts enabled. */ + if (likely(per_cpu(tracer, cpu))) { + ds_release_bts(per_cpu(tracer, cpu)); + per_cpu(tracer, cpu) = NULL; + } + } + + return NOTIFY_DONE; } +static struct notifier_block bts_hotcpu_notifier __cpuinitdata = { + .notifier_call = bts_hotcpu_handler +}; + static void bts_trace_print_header(struct seq_file *m) { - seq_puts(m, - "# CPU# FROM TO FUNCTION\n"); - seq_puts(m, - "# | | | |\n"); + seq_puts(m, "# CPU# TO <- FROM\n"); } static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) { + unsigned long symflags = TRACE_ITER_SYM_OFFSET; struct trace_entry *entry = iter->ent; struct trace_seq *seq = &iter->seq; struct hw_branch_entry *it; @@ -91,11 +149,10 @@ static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) trace_assign_type(it, entry); if (entry->type == TRACE_HW_BRANCHES) { - if (trace_seq_printf(seq, "%4d ", entry->cpu) && - trace_seq_printf(seq, "0x%016llx -> 0x%016llx ", - it->from, it->to) && - (!it->from || - seq_print_ip_sym(seq, it->from, /* sym_flags = */ 0)) && + if (trace_seq_printf(seq, "%4d ", iter->cpu) && + seq_print_ip_sym(seq, it->to, symflags) && + trace_seq_printf(seq, "\t <- ") && + seq_print_ip_sym(seq, it->from, symflags) && trace_seq_printf(seq, "\n")) return TRACE_TYPE_HANDLED; return TRACE_TYPE_PARTIAL_LINE;; @@ -103,26 +160,44 @@ static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) return TRACE_TYPE_UNHANDLED; } -void trace_hw_branch(struct trace_array *tr, u64 from, u64 to) +void trace_hw_branch(u64 from, u64 to) { + struct ftrace_event_call *call = &event_hw_branch; + struct trace_array *tr = hw_branch_trace; struct ring_buffer_event *event; struct hw_branch_entry *entry; - unsigned long irq; + unsigned long irq1; + int cpu; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), &irq); - if (!event) + if (unlikely(!tr)) + return; + + if (unlikely(!trace_hw_branches_enabled)) return; + + local_irq_save(irq1); + cpu = raw_smp_processor_id(); + if (atomic_inc_return(&tr->data[cpu]->disabled) != 1) + goto out; + + event = trace_buffer_lock_reserve(tr, TRACE_HW_BRANCHES, + sizeof(*entry), 0, 0); + if (!event) + goto out; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, from); entry->ent.type = TRACE_HW_BRANCHES; - entry->ent.cpu = smp_processor_id(); entry->from = from; entry->to = to; - ring_buffer_unlock_commit(tr->buffer, event, irq); + if (!filter_check_discard(call, entry, tr->buffer, event)) + trace_buffer_unlock_commit(tr, event, 0, 0); + + out: + atomic_dec(&tr->data[cpu]->disabled); + local_irq_restore(irq1); } -static void trace_bts_at(struct trace_array *tr, - const struct bts_trace *trace, void *at) +static void trace_bts_at(const struct bts_trace *trace, void *at) { struct bts_struct bts; int err = 0; @@ -137,59 +212,98 @@ static void trace_bts_at(struct trace_array *tr, switch (bts.qualifier) { case BTS_BRANCH: - trace_hw_branch(tr, bts.variant.lbr.from, bts.variant.lbr.to); + trace_hw_branch(bts.variant.lbr.from, bts.variant.lbr.to); break; } } +/* + * Collect the trace on the current cpu and write it into the ftrace buffer. + * + * pre: tracing must be suspended on the current cpu + */ static void trace_bts_cpu(void *arg) { - struct trace_array *tr = (struct trace_array *) arg; + struct trace_array *tr = (struct trace_array *)arg; const struct bts_trace *trace; unsigned char *at; - if (!this_tracer) + if (unlikely(!tr)) + return; + + if (unlikely(atomic_read(&tr->data[raw_smp_processor_id()]->disabled))) + return; + + if (unlikely(!this_tracer)) return; - ds_suspend_bts(this_tracer); trace = ds_read_bts(this_tracer); if (!trace) - goto out; + return; for (at = trace->ds.top; (void *)at < trace->ds.end; at += trace->ds.size) - trace_bts_at(tr, trace, at); + trace_bts_at(trace, at); for (at = trace->ds.begin; (void *)at < trace->ds.top; at += trace->ds.size) - trace_bts_at(tr, trace, at); - -out: - ds_resume_bts(this_tracer); + trace_bts_at(trace, at); } static void trace_bts_prepare(struct trace_iterator *iter) { int cpu; - for_each_cpu(cpu, cpu_possible_mask) - smp_call_function_single(cpu, trace_bts_cpu, iter->tr, 1); + get_online_cpus(); + for_each_online_cpu(cpu) + if (likely(per_cpu(tracer, cpu))) + ds_suspend_bts(per_cpu(tracer, cpu)); + /* + * We need to collect the trace on the respective cpu since ftrace + * implicitly adds the record for the current cpu. + * Once that is more flexible, we could collect the data from any cpu. + */ + on_each_cpu(trace_bts_cpu, iter->tr, 1); + + for_each_online_cpu(cpu) + if (likely(per_cpu(tracer, cpu))) + ds_resume_bts(per_cpu(tracer, cpu)); + put_online_cpus(); +} + +static void trace_bts_close(struct trace_iterator *iter) +{ + tracing_reset_online_cpus(iter->tr); +} + +void trace_hw_branch_oops(void) +{ + if (this_tracer) { + ds_suspend_bts_noirq(this_tracer); + trace_bts_cpu(hw_branch_trace); + ds_resume_bts_noirq(this_tracer); + } } struct tracer bts_tracer __read_mostly = { .name = "hw-branch-tracer", .init = bts_trace_init, - .reset = bts_trace_stop, + .reset = bts_trace_reset, .print_header = bts_trace_print_header, .print_line = bts_trace_print_line, .start = bts_trace_start, .stop = bts_trace_stop, - .open = trace_bts_prepare + .open = trace_bts_prepare, + .close = trace_bts_close, +#ifdef CONFIG_FTRACE_SELFTEST + .selftest = trace_selftest_startup_hw_branches, +#endif /* CONFIG_FTRACE_SELFTEST */ }; __init static int init_bts_trace(void) { + register_hotcpu_notifier(&bts_hotcpu_notifier); return register_tracer(&bts_tracer); } device_initcall(init_bts_trace); diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index 7c2e326bbc8..5555b75a0d1 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -1,5 +1,5 @@ /* - * trace irqs off criticall timings + * trace irqs off critical timings * * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com> * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com> @@ -32,6 +32,8 @@ enum { static int trace_type __read_mostly; +static int save_lat_flag; + #ifdef CONFIG_PREEMPT_TRACER static inline int preempt_trace(void) @@ -95,7 +97,7 @@ irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip) disabled = atomic_inc_return(&data->disabled); if (likely(disabled == 1)) - trace_function(tr, data, ip, parent_ip, flags, preempt_count()); + trace_function(tr, ip, parent_ip, flags, preempt_count()); atomic_dec(&data->disabled); } @@ -153,7 +155,7 @@ check_critical_timing(struct trace_array *tr, if (!report_latency(delta)) goto out_unlock; - trace_function(tr, data, CALLER_ADDR0, parent_ip, flags, pc); + trace_function(tr, CALLER_ADDR0, parent_ip, flags, pc); latency = nsecs_to_usecs(delta); @@ -176,8 +178,7 @@ out_unlock: out: data->critical_sequence = max_sequence; data->preempt_timestamp = ftrace_now(cpu); - tracing_reset(tr, cpu); - trace_function(tr, data, CALLER_ADDR0, parent_ip, flags, pc); + trace_function(tr, CALLER_ADDR0, parent_ip, flags, pc); } static inline void @@ -206,11 +207,10 @@ start_critical_timing(unsigned long ip, unsigned long parent_ip) data->critical_sequence = max_sequence; data->preempt_timestamp = ftrace_now(cpu); data->critical_start = parent_ip ? : ip; - tracing_reset(tr, cpu); local_save_flags(flags); - trace_function(tr, data, ip, parent_ip, flags, preempt_count()); + trace_function(tr, ip, parent_ip, flags, preempt_count()); per_cpu(tracing_cpu, cpu) = 1; @@ -244,7 +244,7 @@ stop_critical_timing(unsigned long ip, unsigned long parent_ip) atomic_inc(&data->disabled); local_save_flags(flags); - trace_function(tr, data, ip, parent_ip, flags, preempt_count()); + trace_function(tr, ip, parent_ip, flags, preempt_count()); check_critical_timing(tr, data, parent_ip ? : ip, cpu); data->critical_start = 0; atomic_dec(&data->disabled); @@ -353,66 +353,50 @@ void trace_preempt_off(unsigned long a0, unsigned long a1) } #endif /* CONFIG_PREEMPT_TRACER */ -/* - * save_tracer_enabled is used to save the state of the tracer_enabled - * variable when we disable it when we open a trace output file. - */ -static int save_tracer_enabled; - static void start_irqsoff_tracer(struct trace_array *tr) { register_ftrace_function(&trace_ops); - if (tracing_is_enabled()) { + if (tracing_is_enabled()) tracer_enabled = 1; - save_tracer_enabled = 1; - } else { + else tracer_enabled = 0; - save_tracer_enabled = 0; - } } static void stop_irqsoff_tracer(struct trace_array *tr) { tracer_enabled = 0; - save_tracer_enabled = 0; unregister_ftrace_function(&trace_ops); } static void __irqsoff_tracer_init(struct trace_array *tr) { + save_lat_flag = trace_flags & TRACE_ITER_LATENCY_FMT; + trace_flags |= TRACE_ITER_LATENCY_FMT; + + tracing_max_latency = 0; irqsoff_trace = tr; /* make sure that the tracer is visible */ smp_wmb(); + tracing_reset_online_cpus(tr); start_irqsoff_tracer(tr); } static void irqsoff_tracer_reset(struct trace_array *tr) { stop_irqsoff_tracer(tr); + + if (!save_lat_flag) + trace_flags &= ~TRACE_ITER_LATENCY_FMT; } static void irqsoff_tracer_start(struct trace_array *tr) { tracer_enabled = 1; - save_tracer_enabled = 1; } static void irqsoff_tracer_stop(struct trace_array *tr) { tracer_enabled = 0; - save_tracer_enabled = 0; -} - -static void irqsoff_tracer_open(struct trace_iterator *iter) -{ - /* stop the trace while dumping */ - tracer_enabled = 0; -} - -static void irqsoff_tracer_close(struct trace_iterator *iter) -{ - /* restart tracing */ - tracer_enabled = save_tracer_enabled; } #ifdef CONFIG_IRQSOFF_TRACER @@ -430,8 +414,6 @@ static struct tracer irqsoff_tracer __read_mostly = .reset = irqsoff_tracer_reset, .start = irqsoff_tracer_start, .stop = irqsoff_tracer_stop, - .open = irqsoff_tracer_open, - .close = irqsoff_tracer_close, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_irqsoff, @@ -458,8 +440,6 @@ static struct tracer preemptoff_tracer __read_mostly = .reset = irqsoff_tracer_reset, .start = irqsoff_tracer_start, .stop = irqsoff_tracer_stop, - .open = irqsoff_tracer_open, - .close = irqsoff_tracer_close, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_preemptoff, @@ -488,8 +468,6 @@ static struct tracer preemptirqsoff_tracer __read_mostly = .reset = irqsoff_tracer_reset, .start = irqsoff_tracer_start, .stop = irqsoff_tracer_stop, - .open = irqsoff_tracer_open, - .close = irqsoff_tracer_close, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_preemptirqsoff, diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index fffcb069f1d..c4c9bbda53d 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -9,8 +9,12 @@ #include <linux/kernel.h> #include <linux/mmiotrace.h> #include <linux/pci.h> +#include <linux/time.h> + +#include <asm/atomic.h> #include "trace.h" +#include "trace_output.h" struct header_iter { struct pci_dev *dev; @@ -19,6 +23,7 @@ struct header_iter { static struct trace_array *mmio_trace_array; static bool overrun_detected; static unsigned long prev_overruns; +static atomic_t dropped_count; static void mmio_reset_data(struct trace_array *tr) { @@ -121,11 +126,11 @@ static void mmio_close(struct trace_iterator *iter) static unsigned long count_overruns(struct trace_iterator *iter) { - unsigned long cnt = 0; + unsigned long cnt = atomic_xchg(&dropped_count, 0); unsigned long over = ring_buffer_overruns(iter->tr->buffer); if (over > prev_overruns) - cnt = over - prev_overruns; + cnt += over - prev_overruns; prev_overruns = over; return cnt; } @@ -171,7 +176,7 @@ static enum print_line_t mmio_print_rw(struct trace_iterator *iter) struct mmiotrace_rw *rw; struct trace_seq *s = &iter->seq; unsigned long long t = ns2usecs(iter->ts); - unsigned long usec_rem = do_div(t, 1000000ULL); + unsigned long usec_rem = do_div(t, USEC_PER_SEC); unsigned secs = (unsigned long)t; int ret = 1; @@ -181,21 +186,22 @@ static enum print_line_t mmio_print_rw(struct trace_iterator *iter) switch (rw->opcode) { case MMIO_READ: ret = trace_seq_printf(s, - "R %d %lu.%06lu %d 0x%llx 0x%lx 0x%lx %d\n", + "R %d %u.%06lu %d 0x%llx 0x%lx 0x%lx %d\n", rw->width, secs, usec_rem, rw->map_id, (unsigned long long)rw->phys, rw->value, rw->pc, 0); break; case MMIO_WRITE: ret = trace_seq_printf(s, - "W %d %lu.%06lu %d 0x%llx 0x%lx 0x%lx %d\n", + "W %d %u.%06lu %d 0x%llx 0x%lx 0x%lx %d\n", rw->width, secs, usec_rem, rw->map_id, (unsigned long long)rw->phys, rw->value, rw->pc, 0); break; case MMIO_UNKNOWN_OP: ret = trace_seq_printf(s, - "UNKNOWN %lu.%06lu %d 0x%llx %02x,%02x,%02x 0x%lx %d\n", + "UNKNOWN %u.%06lu %d 0x%llx %02lx,%02lx," + "%02lx 0x%lx %d\n", secs, usec_rem, rw->map_id, (unsigned long long)rw->phys, (rw->value >> 16) & 0xff, (rw->value >> 8) & 0xff, @@ -217,7 +223,7 @@ static enum print_line_t mmio_print_map(struct trace_iterator *iter) struct mmiotrace_map *m; struct trace_seq *s = &iter->seq; unsigned long long t = ns2usecs(iter->ts); - unsigned long usec_rem = do_div(t, 1000000ULL); + unsigned long usec_rem = do_div(t, USEC_PER_SEC); unsigned secs = (unsigned long)t; int ret; @@ -227,14 +233,14 @@ static enum print_line_t mmio_print_map(struct trace_iterator *iter) switch (m->opcode) { case MMIO_PROBE: ret = trace_seq_printf(s, - "MAP %lu.%06lu %d 0x%llx 0x%lx 0x%lx 0x%lx %d\n", + "MAP %u.%06lu %d 0x%llx 0x%lx 0x%lx 0x%lx %d\n", secs, usec_rem, m->map_id, (unsigned long long)m->phys, m->virt, m->len, 0UL, 0); break; case MMIO_UNPROBE: ret = trace_seq_printf(s, - "UNMAP %lu.%06lu %d 0x%lx %d\n", + "UNMAP %u.%06lu %d 0x%lx %d\n", secs, usec_rem, m->map_id, 0UL, 0); break; default: @@ -253,18 +259,15 @@ static enum print_line_t mmio_print_mark(struct trace_iterator *iter) const char *msg = print->buf; struct trace_seq *s = &iter->seq; unsigned long long t = ns2usecs(iter->ts); - unsigned long usec_rem = do_div(t, 1000000ULL); + unsigned long usec_rem = do_div(t, USEC_PER_SEC); unsigned secs = (unsigned long)t; int ret; /* The trailing newline must be in the message. */ - ret = trace_seq_printf(s, "MARK %lu.%06lu %s", secs, usec_rem, msg); + ret = trace_seq_printf(s, "MARK %u.%06lu %s", secs, usec_rem, msg); if (!ret) return TRACE_TYPE_PARTIAL_LINE; - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); - return TRACE_TYPE_HANDLED; } @@ -304,21 +307,20 @@ static void __trace_mmiotrace_rw(struct trace_array *tr, struct trace_array_cpu *data, struct mmiotrace_rw *rw) { + struct ring_buffer *buffer = tr->buffer; struct ring_buffer_event *event; struct trace_mmiotrace_rw *entry; - unsigned long irq_flags; + int pc = preempt_count(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); - if (!event) + event = trace_buffer_lock_reserve(buffer, TRACE_MMIO_RW, + sizeof(*entry), 0, pc); + if (!event) { + atomic_inc(&dropped_count); return; + } entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, preempt_count()); - entry->ent.type = TRACE_MMIO_RW; entry->rw = *rw; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - - trace_wake_up(); + trace_buffer_unlock_commit(buffer, event, 0, pc); } void mmio_trace_rw(struct mmiotrace_rw *rw) @@ -332,21 +334,20 @@ static void __trace_mmiotrace_map(struct trace_array *tr, struct trace_array_cpu *data, struct mmiotrace_map *map) { + struct ring_buffer *buffer = tr->buffer; struct ring_buffer_event *event; struct trace_mmiotrace_map *entry; - unsigned long irq_flags; + int pc = preempt_count(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); - if (!event) + event = trace_buffer_lock_reserve(buffer, TRACE_MMIO_MAP, + sizeof(*entry), 0, pc); + if (!event) { + atomic_inc(&dropped_count); return; + } entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, preempt_count()); - entry->ent.type = TRACE_MMIO_MAP; entry->map = *map; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - - trace_wake_up(); + trace_buffer_unlock_commit(buffer, event, 0, pc); } void mmio_trace_mapping(struct mmiotrace_map *map) @@ -362,5 +363,5 @@ void mmio_trace_mapping(struct mmiotrace_map *map) int mmio_trace_printk(const char *fmt, va_list args) { - return trace_vprintk(0, -1, fmt, args); + return trace_vprintk(0, fmt, args); } diff --git a/kernel/trace/trace_nop.c b/kernel/trace/trace_nop.c index b9767acd30a..394f94417e2 100644 --- a/kernel/trace/trace_nop.c +++ b/kernel/trace/trace_nop.c @@ -47,12 +47,7 @@ static void stop_nop_trace(struct trace_array *tr) static int nop_trace_init(struct trace_array *tr) { - int cpu; ctx_trace = tr; - - for_each_online_cpu(cpu) - tracing_reset(tr, cpu); - start_nop_trace(tr); return 0; } @@ -96,6 +91,7 @@ struct tracer nop_trace __read_mostly = .name = "nop", .init = nop_trace_init, .reset = nop_trace_reset, + .wait_pipe = poll_wait_pipe, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_nop, #endif diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c new file mode 100644 index 00000000000..e0c2545622e --- /dev/null +++ b/kernel/trace/trace_output.c @@ -0,0 +1,1202 @@ +/* + * trace_output.c + * + * Copyright (C) 2008 Red Hat Inc, Steven Rostedt <srostedt@redhat.com> + * + */ + +#include <linux/module.h> +#include <linux/mutex.h> +#include <linux/ftrace.h> + +#include "trace_output.h" + +/* must be a power of 2 */ +#define EVENT_HASHSIZE 128 + +DECLARE_RWSEM(trace_event_mutex); + +DEFINE_PER_CPU(struct trace_seq, ftrace_event_seq); +EXPORT_PER_CPU_SYMBOL(ftrace_event_seq); + +static struct hlist_head event_hash[EVENT_HASHSIZE] __read_mostly; + +static int next_event_type = __TRACE_LAST_TYPE + 1; + +void trace_print_seq(struct seq_file *m, struct trace_seq *s) +{ + int len = s->len >= PAGE_SIZE ? PAGE_SIZE - 1 : s->len; + + seq_write(m, s->buffer, len); + + trace_seq_init(s); +} + +enum print_line_t trace_print_bprintk_msg_only(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; + struct bprint_entry *field; + int ret; + + trace_assign_type(field, entry); + + ret = trace_seq_bprintf(s, field->fmt, field->buf); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +enum print_line_t trace_print_printk_msg_only(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; + struct print_entry *field; + int ret; + + trace_assign_type(field, entry); + + ret = trace_seq_printf(s, "%s", field->buf); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +/** + * trace_seq_printf - sequence printing of trace information + * @s: trace sequence descriptor + * @fmt: printf format string + * + * The tracer may use either sequence operations or its own + * copy to user routines. To simplify formating of a trace + * trace_seq_printf is used to store strings into a special + * buffer (@s). Then the output may be either used by + * the sequencer or pulled into another buffer. + */ +int +trace_seq_printf(struct trace_seq *s, const char *fmt, ...) +{ + int len = (PAGE_SIZE - 1) - s->len; + va_list ap; + int ret; + + if (!len) + return 0; + + va_start(ap, fmt); + ret = vsnprintf(s->buffer + s->len, len, fmt, ap); + va_end(ap); + + /* If we can't write it all, don't bother writing anything */ + if (ret >= len) + return 0; + + s->len += ret; + + return len; +} +EXPORT_SYMBOL_GPL(trace_seq_printf); + +/** + * trace_seq_vprintf - sequence printing of trace information + * @s: trace sequence descriptor + * @fmt: printf format string + * + * The tracer may use either sequence operations or its own + * copy to user routines. To simplify formating of a trace + * trace_seq_printf is used to store strings into a special + * buffer (@s). Then the output may be either used by + * the sequencer or pulled into another buffer. + */ +int +trace_seq_vprintf(struct trace_seq *s, const char *fmt, va_list args) +{ + int len = (PAGE_SIZE - 1) - s->len; + int ret; + + if (!len) + return 0; + + ret = vsnprintf(s->buffer + s->len, len, fmt, args); + + /* If we can't write it all, don't bother writing anything */ + if (ret >= len) + return 0; + + s->len += ret; + + return len; +} +EXPORT_SYMBOL_GPL(trace_seq_vprintf); + +int trace_seq_bprintf(struct trace_seq *s, const char *fmt, const u32 *binary) +{ + int len = (PAGE_SIZE - 1) - s->len; + int ret; + + if (!len) + return 0; + + ret = bstr_printf(s->buffer + s->len, len, fmt, binary); + + /* If we can't write it all, don't bother writing anything */ + if (ret >= len) + return 0; + + s->len += ret; + + return len; +} + +/** + * trace_seq_puts - trace sequence printing of simple string + * @s: trace sequence descriptor + * @str: simple string to record + * + * The tracer may use either the sequence operations or its own + * copy to user routines. This function records a simple string + * into a special buffer (@s) for later retrieval by a sequencer + * or other mechanism. + */ +int trace_seq_puts(struct trace_seq *s, const char *str) +{ + int len = strlen(str); + + if (len > ((PAGE_SIZE - 1) - s->len)) + return 0; + + memcpy(s->buffer + s->len, str, len); + s->len += len; + + return len; +} + +int trace_seq_putc(struct trace_seq *s, unsigned char c) +{ + if (s->len >= (PAGE_SIZE - 1)) + return 0; + + s->buffer[s->len++] = c; + + return 1; +} + +int trace_seq_putmem(struct trace_seq *s, const void *mem, size_t len) +{ + if (len > ((PAGE_SIZE - 1) - s->len)) + return 0; + + memcpy(s->buffer + s->len, mem, len); + s->len += len; + + return len; +} + +int trace_seq_putmem_hex(struct trace_seq *s, const void *mem, size_t len) +{ + unsigned char hex[HEX_CHARS]; + const unsigned char *data = mem; + int i, j; + +#ifdef __BIG_ENDIAN + for (i = 0, j = 0; i < len; i++) { +#else + for (i = len-1, j = 0; i >= 0; i--) { +#endif + hex[j++] = hex_asc_hi(data[i]); + hex[j++] = hex_asc_lo(data[i]); + } + hex[j++] = ' '; + + return trace_seq_putmem(s, hex, j); +} + +void *trace_seq_reserve(struct trace_seq *s, size_t len) +{ + void *ret; + + if (len > ((PAGE_SIZE - 1) - s->len)) + return NULL; + + ret = s->buffer + s->len; + s->len += len; + + return ret; +} + +int trace_seq_path(struct trace_seq *s, struct path *path) +{ + unsigned char *p; + + if (s->len >= (PAGE_SIZE - 1)) + return 0; + p = d_path(path, s->buffer + s->len, PAGE_SIZE - s->len); + if (!IS_ERR(p)) { + p = mangle_path(s->buffer + s->len, p, "\n"); + if (p) { + s->len = p - s->buffer; + return 1; + } + } else { + s->buffer[s->len++] = '?'; + return 1; + } + + return 0; +} + +const char * +ftrace_print_flags_seq(struct trace_seq *p, const char *delim, + unsigned long flags, + const struct trace_print_flags *flag_array) +{ + unsigned long mask; + const char *str; + const char *ret = p->buffer + p->len; + int i; + + for (i = 0; flag_array[i].name && flags; i++) { + + mask = flag_array[i].mask; + if ((flags & mask) != mask) + continue; + + str = flag_array[i].name; + flags &= ~mask; + if (p->len && delim) + trace_seq_puts(p, delim); + trace_seq_puts(p, str); + } + + /* check for left over flags */ + if (flags) { + if (p->len && delim) + trace_seq_puts(p, delim); + trace_seq_printf(p, "0x%lx", flags); + } + + trace_seq_putc(p, 0); + + return ret; +} +EXPORT_SYMBOL(ftrace_print_flags_seq); + +const char * +ftrace_print_symbols_seq(struct trace_seq *p, unsigned long val, + const struct trace_print_flags *symbol_array) +{ + int i; + const char *ret = p->buffer + p->len; + + for (i = 0; symbol_array[i].name; i++) { + + if (val != symbol_array[i].mask) + continue; + + trace_seq_puts(p, symbol_array[i].name); + break; + } + + if (!p->len) + trace_seq_printf(p, "0x%lx", val); + + trace_seq_putc(p, 0); + + return ret; +} +EXPORT_SYMBOL(ftrace_print_symbols_seq); + +#ifdef CONFIG_KRETPROBES +static inline const char *kretprobed(const char *name) +{ + static const char tramp_name[] = "kretprobe_trampoline"; + int size = sizeof(tramp_name); + + if (strncmp(tramp_name, name, size) == 0) + return "[unknown/kretprobe'd]"; + return name; +} +#else +static inline const char *kretprobed(const char *name) +{ + return name; +} +#endif /* CONFIG_KRETPROBES */ + +static int +seq_print_sym_short(struct trace_seq *s, const char *fmt, unsigned long address) +{ +#ifdef CONFIG_KALLSYMS + char str[KSYM_SYMBOL_LEN]; + const char *name; + + kallsyms_lookup(address, NULL, NULL, NULL, str); + + name = kretprobed(str); + + return trace_seq_printf(s, fmt, name); +#endif + return 1; +} + +static int +seq_print_sym_offset(struct trace_seq *s, const char *fmt, + unsigned long address) +{ +#ifdef CONFIG_KALLSYMS + char str[KSYM_SYMBOL_LEN]; + const char *name; + + sprint_symbol(str, address); + name = kretprobed(str); + + return trace_seq_printf(s, fmt, name); +#endif + return 1; +} + +#ifndef CONFIG_64BIT +# define IP_FMT "%08lx" +#else +# define IP_FMT "%016lx" +#endif + +int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, + unsigned long ip, unsigned long sym_flags) +{ + struct file *file = NULL; + unsigned long vmstart = 0; + int ret = 1; + + if (mm) { + const struct vm_area_struct *vma; + + down_read(&mm->mmap_sem); + vma = find_vma(mm, ip); + if (vma) { + file = vma->vm_file; + vmstart = vma->vm_start; + } + if (file) { + ret = trace_seq_path(s, &file->f_path); + if (ret) + ret = trace_seq_printf(s, "[+0x%lx]", + ip - vmstart); + } + up_read(&mm->mmap_sem); + } + if (ret && ((sym_flags & TRACE_ITER_SYM_ADDR) || !file)) + ret = trace_seq_printf(s, " <" IP_FMT ">", ip); + return ret; +} + +int +seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s, + unsigned long sym_flags) +{ + struct mm_struct *mm = NULL; + int ret = 1; + unsigned int i; + + if (trace_flags & TRACE_ITER_SYM_USEROBJ) { + struct task_struct *task; + /* + * we do the lookup on the thread group leader, + * since individual threads might have already quit! + */ + rcu_read_lock(); + task = find_task_by_vpid(entry->ent.tgid); + if (task) + mm = get_task_mm(task); + rcu_read_unlock(); + } + + for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { + unsigned long ip = entry->caller[i]; + + if (ip == ULONG_MAX || !ret) + break; + if (ret) + ret = trace_seq_puts(s, " => "); + if (!ip) { + if (ret) + ret = trace_seq_puts(s, "??"); + if (ret) + ret = trace_seq_puts(s, "\n"); + continue; + } + if (!ret) + break; + if (ret) + ret = seq_print_user_ip(s, mm, ip, sym_flags); + ret = trace_seq_puts(s, "\n"); + } + + if (mm) + mmput(mm); + return ret; +} + +int +seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) +{ + int ret; + + if (!ip) + return trace_seq_printf(s, "0"); + + if (sym_flags & TRACE_ITER_SYM_OFFSET) + ret = seq_print_sym_offset(s, "%s", ip); + else + ret = seq_print_sym_short(s, "%s", ip); + + if (!ret) + return 0; + + if (sym_flags & TRACE_ITER_SYM_ADDR) + ret = trace_seq_printf(s, " <" IP_FMT ">", ip); + return ret; +} + +static int +lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu) +{ + int hardirq, softirq; + char comm[TASK_COMM_LEN]; + + trace_find_cmdline(entry->pid, comm); + hardirq = entry->flags & TRACE_FLAG_HARDIRQ; + softirq = entry->flags & TRACE_FLAG_SOFTIRQ; + + if (!trace_seq_printf(s, "%8.8s-%-5d %3d%c%c%c", + comm, entry->pid, cpu, + (entry->flags & TRACE_FLAG_IRQS_OFF) ? 'd' : + (entry->flags & TRACE_FLAG_IRQS_NOSUPPORT) ? + 'X' : '.', + (entry->flags & TRACE_FLAG_NEED_RESCHED) ? + 'N' : '.', + (hardirq && softirq) ? 'H' : + hardirq ? 'h' : softirq ? 's' : '.')) + return 0; + + if (entry->preempt_count) + return trace_seq_printf(s, "%x", entry->preempt_count); + return trace_seq_puts(s, "."); +} + +static unsigned long preempt_mark_thresh = 100; + +static int +lat_print_timestamp(struct trace_seq *s, u64 abs_usecs, + unsigned long rel_usecs) +{ + return trace_seq_printf(s, " %4lldus%c: ", abs_usecs, + rel_usecs > preempt_mark_thresh ? '!' : + rel_usecs > 1 ? '+' : ' '); +} + +int trace_print_context(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; + unsigned long long t = ns2usecs(iter->ts); + unsigned long usec_rem = do_div(t, USEC_PER_SEC); + unsigned long secs = (unsigned long)t; + char comm[TASK_COMM_LEN]; + + trace_find_cmdline(entry->pid, comm); + + return trace_seq_printf(s, "%16s-%-5d [%03d] %5lu.%06lu: ", + comm, entry->pid, iter->cpu, secs, usec_rem); +} + +int trace_print_lat_context(struct trace_iterator *iter) +{ + u64 next_ts; + int ret; + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent, + *next_entry = trace_find_next_entry(iter, NULL, + &next_ts); + unsigned long verbose = (trace_flags & TRACE_ITER_VERBOSE); + unsigned long abs_usecs = ns2usecs(iter->ts - iter->tr->time_start); + unsigned long rel_usecs; + + if (!next_entry) + next_ts = iter->ts; + rel_usecs = ns2usecs(next_ts - iter->ts); + + if (verbose) { + char comm[TASK_COMM_LEN]; + + trace_find_cmdline(entry->pid, comm); + + ret = trace_seq_printf(s, "%16s %5d %3d %d %08x %08lx [%08llx]" + " %ld.%03ldms (+%ld.%03ldms): ", comm, + entry->pid, iter->cpu, entry->flags, + entry->preempt_count, iter->idx, + ns2usecs(iter->ts), + abs_usecs / USEC_PER_MSEC, + abs_usecs % USEC_PER_MSEC, + rel_usecs / USEC_PER_MSEC, + rel_usecs % USEC_PER_MSEC); + } else { + ret = lat_print_generic(s, entry, iter->cpu); + if (ret) + ret = lat_print_timestamp(s, abs_usecs, rel_usecs); + } + + return ret; +} + +static const char state_to_char[] = TASK_STATE_TO_CHAR_STR; + +static int task_state_char(unsigned long state) +{ + int bit = state ? __ffs(state) + 1 : 0; + + return bit < sizeof(state_to_char) - 1 ? state_to_char[bit] : '?'; +} + +/** + * ftrace_find_event - find a registered event + * @type: the type of event to look for + * + * Returns an event of type @type otherwise NULL + * Called with trace_event_read_lock() held. + */ +struct trace_event *ftrace_find_event(int type) +{ + struct trace_event *event; + struct hlist_node *n; + unsigned key; + + key = type & (EVENT_HASHSIZE - 1); + + hlist_for_each_entry(event, n, &event_hash[key], node) { + if (event->type == type) + return event; + } + + return NULL; +} + +static LIST_HEAD(ftrace_event_list); + +static int trace_search_list(struct list_head **list) +{ + struct trace_event *e; + int last = __TRACE_LAST_TYPE; + + if (list_empty(&ftrace_event_list)) { + *list = &ftrace_event_list; + return last + 1; + } + + /* + * We used up all possible max events, + * lets see if somebody freed one. + */ + list_for_each_entry(e, &ftrace_event_list, list) { + if (e->type != last + 1) + break; + last++; + } + + /* Did we used up all 65 thousand events??? */ + if ((last + 1) > FTRACE_MAX_EVENT) + return 0; + + *list = &e->list; + return last + 1; +} + +void trace_event_read_lock(void) +{ + down_read(&trace_event_mutex); +} + +void trace_event_read_unlock(void) +{ + up_read(&trace_event_mutex); +} + +/** + * register_ftrace_event - register output for an event type + * @event: the event type to register + * + * Event types are stored in a hash and this hash is used to + * find a way to print an event. If the @event->type is set + * then it will use that type, otherwise it will assign a + * type to use. + * + * If you assign your own type, please make sure it is added + * to the trace_type enum in trace.h, to avoid collisions + * with the dynamic types. + * + * Returns the event type number or zero on error. + */ +int register_ftrace_event(struct trace_event *event) +{ + unsigned key; + int ret = 0; + + down_write(&trace_event_mutex); + + if (WARN_ON(!event)) + goto out; + + INIT_LIST_HEAD(&event->list); + + if (!event->type) { + struct list_head *list = NULL; + + if (next_event_type > FTRACE_MAX_EVENT) { + + event->type = trace_search_list(&list); + if (!event->type) + goto out; + + } else { + + event->type = next_event_type++; + list = &ftrace_event_list; + } + + if (WARN_ON(ftrace_find_event(event->type))) + goto out; + + list_add_tail(&event->list, list); + + } else if (event->type > __TRACE_LAST_TYPE) { + printk(KERN_WARNING "Need to add type to trace.h\n"); + WARN_ON(1); + goto out; + } else { + /* Is this event already used */ + if (ftrace_find_event(event->type)) + goto out; + } + + if (event->trace == NULL) + event->trace = trace_nop_print; + if (event->raw == NULL) + event->raw = trace_nop_print; + if (event->hex == NULL) + event->hex = trace_nop_print; + if (event->binary == NULL) + event->binary = trace_nop_print; + + key = event->type & (EVENT_HASHSIZE - 1); + + hlist_add_head(&event->node, &event_hash[key]); + + ret = event->type; + out: + up_write(&trace_event_mutex); + + return ret; +} +EXPORT_SYMBOL_GPL(register_ftrace_event); + +/* + * Used by module code with the trace_event_mutex held for write. + */ +int __unregister_ftrace_event(struct trace_event *event) +{ + hlist_del(&event->node); + list_del(&event->list); + return 0; +} + +/** + * unregister_ftrace_event - remove a no longer used event + * @event: the event to remove + */ +int unregister_ftrace_event(struct trace_event *event) +{ + down_write(&trace_event_mutex); + __unregister_ftrace_event(event); + up_write(&trace_event_mutex); + + return 0; +} +EXPORT_SYMBOL_GPL(unregister_ftrace_event); + +/* + * Standard events + */ + +enum print_line_t trace_nop_print(struct trace_iterator *iter, int flags) +{ + return TRACE_TYPE_HANDLED; +} + +/* TRACE_FN */ +static enum print_line_t trace_fn_trace(struct trace_iterator *iter, int flags) +{ + struct ftrace_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + if (!seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if ((flags & TRACE_ITER_PRINT_PARENT) && field->parent_ip) { + if (!trace_seq_printf(s, " <-")) + goto partial; + if (!seq_print_ip_sym(s, + field->parent_ip, + flags)) + goto partial; + } + if (!trace_seq_printf(s, "\n")) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static enum print_line_t trace_fn_raw(struct trace_iterator *iter, int flags) +{ + struct ftrace_entry *field; + + trace_assign_type(field, iter->ent); + + if (!trace_seq_printf(&iter->seq, "%lx %lx\n", + field->ip, + field->parent_ip)) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t trace_fn_hex(struct trace_iterator *iter, int flags) +{ + struct ftrace_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + SEQ_PUT_HEX_FIELD_RET(s, field->ip); + SEQ_PUT_HEX_FIELD_RET(s, field->parent_ip); + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t trace_fn_bin(struct trace_iterator *iter, int flags) +{ + struct ftrace_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + SEQ_PUT_FIELD_RET(s, field->ip); + SEQ_PUT_FIELD_RET(s, field->parent_ip); + + return TRACE_TYPE_HANDLED; +} + +static struct trace_event trace_fn_event = { + .type = TRACE_FN, + .trace = trace_fn_trace, + .raw = trace_fn_raw, + .hex = trace_fn_hex, + .binary = trace_fn_bin, +}; + +/* TRACE_CTX an TRACE_WAKE */ +static enum print_line_t trace_ctxwake_print(struct trace_iterator *iter, + char *delim) +{ + struct ctx_switch_entry *field; + char comm[TASK_COMM_LEN]; + int S, T; + + + trace_assign_type(field, iter->ent); + + T = task_state_char(field->next_state); + S = task_state_char(field->prev_state); + trace_find_cmdline(field->next_pid, comm); + if (!trace_seq_printf(&iter->seq, + " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", + field->prev_pid, + field->prev_prio, + S, delim, + field->next_cpu, + field->next_pid, + field->next_prio, + T, comm)) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t trace_ctx_print(struct trace_iterator *iter, int flags) +{ + return trace_ctxwake_print(iter, "==>"); +} + +static enum print_line_t trace_wake_print(struct trace_iterator *iter, + int flags) +{ + return trace_ctxwake_print(iter, " +"); +} + +static int trace_ctxwake_raw(struct trace_iterator *iter, char S) +{ + struct ctx_switch_entry *field; + int T; + + trace_assign_type(field, iter->ent); + + if (!S) + task_state_char(field->prev_state); + T = task_state_char(field->next_state); + if (!trace_seq_printf(&iter->seq, "%d %d %c %d %d %d %c\n", + field->prev_pid, + field->prev_prio, + S, + field->next_cpu, + field->next_pid, + field->next_prio, + T)) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t trace_ctx_raw(struct trace_iterator *iter, int flags) +{ + return trace_ctxwake_raw(iter, 0); +} + +static enum print_line_t trace_wake_raw(struct trace_iterator *iter, int flags) +{ + return trace_ctxwake_raw(iter, '+'); +} + + +static int trace_ctxwake_hex(struct trace_iterator *iter, char S) +{ + struct ctx_switch_entry *field; + struct trace_seq *s = &iter->seq; + int T; + + trace_assign_type(field, iter->ent); + + if (!S) + task_state_char(field->prev_state); + T = task_state_char(field->next_state); + + SEQ_PUT_HEX_FIELD_RET(s, field->prev_pid); + SEQ_PUT_HEX_FIELD_RET(s, field->prev_prio); + SEQ_PUT_HEX_FIELD_RET(s, S); + SEQ_PUT_HEX_FIELD_RET(s, field->next_cpu); + SEQ_PUT_HEX_FIELD_RET(s, field->next_pid); + SEQ_PUT_HEX_FIELD_RET(s, field->next_prio); + SEQ_PUT_HEX_FIELD_RET(s, T); + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t trace_ctx_hex(struct trace_iterator *iter, int flags) +{ + return trace_ctxwake_hex(iter, 0); +} + +static enum print_line_t trace_wake_hex(struct trace_iterator *iter, int flags) +{ + return trace_ctxwake_hex(iter, '+'); +} + +static enum print_line_t trace_ctxwake_bin(struct trace_iterator *iter, + int flags) +{ + struct ctx_switch_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + SEQ_PUT_FIELD_RET(s, field->prev_pid); + SEQ_PUT_FIELD_RET(s, field->prev_prio); + SEQ_PUT_FIELD_RET(s, field->prev_state); + SEQ_PUT_FIELD_RET(s, field->next_pid); + SEQ_PUT_FIELD_RET(s, field->next_prio); + SEQ_PUT_FIELD_RET(s, field->next_state); + + return TRACE_TYPE_HANDLED; +} + +static struct trace_event trace_ctx_event = { + .type = TRACE_CTX, + .trace = trace_ctx_print, + .raw = trace_ctx_raw, + .hex = trace_ctx_hex, + .binary = trace_ctxwake_bin, +}; + +static struct trace_event trace_wake_event = { + .type = TRACE_WAKE, + .trace = trace_wake_print, + .raw = trace_wake_raw, + .hex = trace_wake_hex, + .binary = trace_ctxwake_bin, +}; + +/* TRACE_SPECIAL */ +static enum print_line_t trace_special_print(struct trace_iterator *iter, + int flags) +{ + struct special_entry *field; + + trace_assign_type(field, iter->ent); + + if (!trace_seq_printf(&iter->seq, "# %ld %ld %ld\n", + field->arg1, + field->arg2, + field->arg3)) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t trace_special_hex(struct trace_iterator *iter, + int flags) +{ + struct special_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + SEQ_PUT_HEX_FIELD_RET(s, field->arg1); + SEQ_PUT_HEX_FIELD_RET(s, field->arg2); + SEQ_PUT_HEX_FIELD_RET(s, field->arg3); + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t trace_special_bin(struct trace_iterator *iter, + int flags) +{ + struct special_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + SEQ_PUT_FIELD_RET(s, field->arg1); + SEQ_PUT_FIELD_RET(s, field->arg2); + SEQ_PUT_FIELD_RET(s, field->arg3); + + return TRACE_TYPE_HANDLED; +} + +static struct trace_event trace_special_event = { + .type = TRACE_SPECIAL, + .trace = trace_special_print, + .raw = trace_special_print, + .hex = trace_special_hex, + .binary = trace_special_bin, +}; + +/* TRACE_STACK */ + +static enum print_line_t trace_stack_print(struct trace_iterator *iter, + int flags) +{ + struct stack_entry *field; + struct trace_seq *s = &iter->seq; + int i; + + trace_assign_type(field, iter->ent); + + if (!trace_seq_puts(s, "<stack trace>\n")) + goto partial; + for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { + if (!field->caller[i] || (field->caller[i] == ULONG_MAX)) + break; + if (!trace_seq_puts(s, " => ")) + goto partial; + + if (!seq_print_ip_sym(s, field->caller[i], flags)) + goto partial; + if (!trace_seq_puts(s, "\n")) + goto partial; + } + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static struct trace_event trace_stack_event = { + .type = TRACE_STACK, + .trace = trace_stack_print, + .raw = trace_special_print, + .hex = trace_special_hex, + .binary = trace_special_bin, +}; + +/* TRACE_USER_STACK */ +static enum print_line_t trace_user_stack_print(struct trace_iterator *iter, + int flags) +{ + struct userstack_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + if (!trace_seq_puts(s, "<user stack trace>\n")) + goto partial; + + if (!seq_print_userip_objs(field, s, flags)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static struct trace_event trace_user_stack_event = { + .type = TRACE_USER_STACK, + .trace = trace_user_stack_print, + .raw = trace_special_print, + .hex = trace_special_hex, + .binary = trace_special_bin, +}; + +/* TRACE_BPRINT */ +static enum print_line_t +trace_bprint_print(struct trace_iterator *iter, int flags) +{ + struct trace_entry *entry = iter->ent; + struct trace_seq *s = &iter->seq; + struct bprint_entry *field; + + trace_assign_type(field, entry); + + if (!seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if (!trace_seq_puts(s, ": ")) + goto partial; + + if (!trace_seq_bprintf(s, field->fmt, field->buf)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + + +static enum print_line_t +trace_bprint_raw(struct trace_iterator *iter, int flags) +{ + struct bprint_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + if (!trace_seq_printf(s, ": %lx : ", field->ip)) + goto partial; + + if (!trace_seq_bprintf(s, field->fmt, field->buf)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + + +static struct trace_event trace_bprint_event = { + .type = TRACE_BPRINT, + .trace = trace_bprint_print, + .raw = trace_bprint_raw, +}; + +/* TRACE_PRINT */ +static enum print_line_t trace_print_print(struct trace_iterator *iter, + int flags) +{ + struct print_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + if (!seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if (!trace_seq_printf(s, ": %s", field->buf)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) +{ + struct print_entry *field; + + trace_assign_type(field, iter->ent); + + if (!trace_seq_printf(&iter->seq, "# %lx %s", field->ip, field->buf)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static struct trace_event trace_print_event = { + .type = TRACE_PRINT, + .trace = trace_print_print, + .raw = trace_print_raw, +}; + + +static struct trace_event *events[] __initdata = { + &trace_fn_event, + &trace_ctx_event, + &trace_wake_event, + &trace_special_event, + &trace_stack_event, + &trace_user_stack_event, + &trace_bprint_event, + &trace_print_event, + NULL +}; + +__init static int init_events(void) +{ + struct trace_event *event; + int i, ret; + + for (i = 0; events[i]; i++) { + event = events[i]; + + ret = register_ftrace_event(event); + if (!ret) { + printk(KERN_WARNING "event %d failed to register\n", + event->type); + WARN_ON_ONCE(1); + } + } + + return 0; +} +device_initcall(init_events); diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h new file mode 100644 index 00000000000..d38bec4a9c3 --- /dev/null +++ b/kernel/trace/trace_output.h @@ -0,0 +1,51 @@ +#ifndef __TRACE_EVENTS_H +#define __TRACE_EVENTS_H + +#include <linux/trace_seq.h> +#include "trace.h" + +extern enum print_line_t +trace_print_bprintk_msg_only(struct trace_iterator *iter); +extern enum print_line_t +trace_print_printk_msg_only(struct trace_iterator *iter); + +extern int +seq_print_ip_sym(struct trace_seq *s, unsigned long ip, + unsigned long sym_flags); +extern int seq_print_userip_objs(const struct userstack_entry *entry, + struct trace_seq *s, unsigned long sym_flags); +extern int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, + unsigned long ip, unsigned long sym_flags); + +extern int trace_print_context(struct trace_iterator *iter); +extern int trace_print_lat_context(struct trace_iterator *iter); + +extern void trace_event_read_lock(void); +extern void trace_event_read_unlock(void); +extern struct trace_event *ftrace_find_event(int type); + +extern enum print_line_t trace_nop_print(struct trace_iterator *iter, + int flags); + +/* used by module unregistering */ +extern int __unregister_ftrace_event(struct trace_event *event); +extern struct rw_semaphore trace_event_mutex; + +#define MAX_MEMHEX_BYTES 8 +#define HEX_CHARS (MAX_MEMHEX_BYTES*2 + 1) + +#define SEQ_PUT_FIELD_RET(s, x) \ +do { \ + if (!trace_seq_putmem(s, &(x), sizeof(x))) \ + return TRACE_TYPE_PARTIAL_LINE; \ +} while (0) + +#define SEQ_PUT_HEX_FIELD_RET(s, x) \ +do { \ + BUILD_BUG_ON(sizeof(x) > MAX_MEMHEX_BYTES); \ + if (!trace_seq_putmem_hex(s, &(x), sizeof(x))) \ + return TRACE_TYPE_PARTIAL_LINE; \ +} while (0) + +#endif + diff --git a/kernel/trace/trace_power.c b/kernel/trace/trace_power.c index 7bda248daf5..fe1a00f1445 100644 --- a/kernel/trace/trace_power.c +++ b/kernel/trace/trace_power.c @@ -11,15 +11,123 @@ #include <linux/init.h> #include <linux/debugfs.h> -#include <linux/ftrace.h> +#include <trace/power.h> #include <linux/kallsyms.h> #include <linux/module.h> #include "trace.h" +#include "trace_output.h" static struct trace_array *power_trace; static int __read_mostly trace_power_enabled; +static void probe_power_start(struct power_trace *it, unsigned int type, + unsigned int level) +{ + if (!trace_power_enabled) + return; + + memset(it, 0, sizeof(struct power_trace)); + it->state = level; + it->type = type; + it->stamp = ktime_get(); +} + + +static void probe_power_end(struct power_trace *it) +{ + struct ftrace_event_call *call = &event_power; + struct ring_buffer_event *event; + struct ring_buffer *buffer; + struct trace_power *entry; + struct trace_array_cpu *data; + struct trace_array *tr = power_trace; + + if (!trace_power_enabled) + return; + + buffer = tr->buffer; + + preempt_disable(); + it->end = ktime_get(); + data = tr->data[smp_processor_id()]; + + event = trace_buffer_lock_reserve(buffer, TRACE_POWER, + sizeof(*entry), 0, 0); + if (!event) + goto out; + entry = ring_buffer_event_data(event); + entry->state_data = *it; + if (!filter_check_discard(call, entry, buffer, event)) + trace_buffer_unlock_commit(buffer, event, 0, 0); + out: + preempt_enable(); +} + +static void probe_power_mark(struct power_trace *it, unsigned int type, + unsigned int level) +{ + struct ftrace_event_call *call = &event_power; + struct ring_buffer_event *event; + struct ring_buffer *buffer; + struct trace_power *entry; + struct trace_array_cpu *data; + struct trace_array *tr = power_trace; + + if (!trace_power_enabled) + return; + + buffer = tr->buffer; + + memset(it, 0, sizeof(struct power_trace)); + it->state = level; + it->type = type; + it->stamp = ktime_get(); + preempt_disable(); + it->end = it->stamp; + data = tr->data[smp_processor_id()]; + + event = trace_buffer_lock_reserve(buffer, TRACE_POWER, + sizeof(*entry), 0, 0); + if (!event) + goto out; + entry = ring_buffer_event_data(event); + entry->state_data = *it; + if (!filter_check_discard(call, entry, buffer, event)) + trace_buffer_unlock_commit(buffer, event, 0, 0); + out: + preempt_enable(); +} + +static int tracing_power_register(void) +{ + int ret; + + ret = register_trace_power_start(probe_power_start); + if (ret) { + pr_info("power trace: Couldn't activate tracepoint" + " probe to trace_power_start\n"); + return ret; + } + ret = register_trace_power_end(probe_power_end); + if (ret) { + pr_info("power trace: Couldn't activate tracepoint" + " probe to trace_power_end\n"); + goto fail_start; + } + ret = register_trace_power_mark(probe_power_mark); + if (ret) { + pr_info("power trace: Couldn't activate tracepoint" + " probe to trace_power_mark\n"); + goto fail_end; + } + return ret; +fail_end: + unregister_trace_power_end(probe_power_end); +fail_start: + unregister_trace_power_start(probe_power_start); + return ret; +} static void start_power_trace(struct trace_array *tr) { @@ -31,16 +139,23 @@ static void stop_power_trace(struct trace_array *tr) trace_power_enabled = 0; } +static void power_trace_reset(struct trace_array *tr) +{ + trace_power_enabled = 0; + unregister_trace_power_start(probe_power_start); + unregister_trace_power_end(probe_power_end); + unregister_trace_power_mark(probe_power_mark); +} + static int power_trace_init(struct trace_array *tr) { - int cpu; power_trace = tr; trace_power_enabled = 1; + tracing_power_register(); - for_each_cpu(cpu, cpu_possible_mask) - tracing_reset(tr, cpu); + tracing_reset_online_cpus(tr); return 0; } @@ -79,14 +194,21 @@ static enum print_line_t power_print_line(struct trace_iterator *iter) return TRACE_TYPE_UNHANDLED; } +static void power_print_header(struct seq_file *s) +{ + seq_puts(s, "# TIMESTAMP STATE EVENT\n"); + seq_puts(s, "# | | |\n"); +} + static struct tracer power_tracer __read_mostly = { .name = "power", .init = power_trace_init, .start = start_power_trace, .stop = stop_power_trace, - .reset = stop_power_trace, + .reset = power_trace_reset, .print_line = power_print_line, + .print_header = power_print_header, }; static int init_power_trace(void) @@ -94,86 +216,3 @@ static int init_power_trace(void) return register_tracer(&power_tracer); } device_initcall(init_power_trace); - -void trace_power_start(struct power_trace *it, unsigned int type, - unsigned int level) -{ - if (!trace_power_enabled) - return; - - memset(it, 0, sizeof(struct power_trace)); - it->state = level; - it->type = type; - it->stamp = ktime_get(); -} -EXPORT_SYMBOL_GPL(trace_power_start); - - -void trace_power_end(struct power_trace *it) -{ - struct ring_buffer_event *event; - struct trace_power *entry; - struct trace_array_cpu *data; - unsigned long irq_flags; - struct trace_array *tr = power_trace; - - if (!trace_power_enabled) - return; - - preempt_disable(); - it->end = ktime_get(); - data = tr->data[smp_processor_id()]; - - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); - if (!event) - goto out; - entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_POWER; - entry->state_data = *it; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - - trace_wake_up(); - - out: - preempt_enable(); -} -EXPORT_SYMBOL_GPL(trace_power_end); - -void trace_power_mark(struct power_trace *it, unsigned int type, - unsigned int level) -{ - struct ring_buffer_event *event; - struct trace_power *entry; - struct trace_array_cpu *data; - unsigned long irq_flags; - struct trace_array *tr = power_trace; - - if (!trace_power_enabled) - return; - - memset(it, 0, sizeof(struct power_trace)); - it->state = level; - it->type = type; - it->stamp = ktime_get(); - preempt_disable(); - it->end = it->stamp; - data = tr->data[smp_processor_id()]; - - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); - if (!event) - goto out; - entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_POWER; - entry->state_data = *it; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - - trace_wake_up(); - - out: - preempt_enable(); -} -EXPORT_SYMBOL_GPL(trace_power_mark); diff --git a/kernel/trace/trace_printk.c b/kernel/trace/trace_printk.c new file mode 100644 index 00000000000..687699d365a --- /dev/null +++ b/kernel/trace/trace_printk.c @@ -0,0 +1,252 @@ +/* + * trace binary printk + * + * Copyright (C) 2008 Lai Jiangshan <laijs@cn.fujitsu.com> + * + */ +#include <linux/seq_file.h> +#include <linux/debugfs.h> +#include <linux/uaccess.h> +#include <linux/kernel.h> +#include <linux/ftrace.h> +#include <linux/string.h> +#include <linux/module.h> +#include <linux/marker.h> +#include <linux/mutex.h> +#include <linux/ctype.h> +#include <linux/list.h> +#include <linux/slab.h> +#include <linux/fs.h> + +#include "trace.h" + +#ifdef CONFIG_MODULES + +/* + * modules trace_printk()'s formats are autosaved in struct trace_bprintk_fmt + * which are queued on trace_bprintk_fmt_list. + */ +static LIST_HEAD(trace_bprintk_fmt_list); + +/* serialize accesses to trace_bprintk_fmt_list */ +static DEFINE_MUTEX(btrace_mutex); + +struct trace_bprintk_fmt { + struct list_head list; + char fmt[0]; +}; + +static inline struct trace_bprintk_fmt *lookup_format(const char *fmt) +{ + struct trace_bprintk_fmt *pos; + list_for_each_entry(pos, &trace_bprintk_fmt_list, list) { + if (!strcmp(pos->fmt, fmt)) + return pos; + } + return NULL; +} + +static +void hold_module_trace_bprintk_format(const char **start, const char **end) +{ + const char **iter; + + mutex_lock(&btrace_mutex); + for (iter = start; iter < end; iter++) { + struct trace_bprintk_fmt *tb_fmt = lookup_format(*iter); + if (tb_fmt) { + *iter = tb_fmt->fmt; + continue; + } + + tb_fmt = kmalloc(offsetof(struct trace_bprintk_fmt, fmt) + + strlen(*iter) + 1, GFP_KERNEL); + if (tb_fmt) { + list_add_tail(&tb_fmt->list, &trace_bprintk_fmt_list); + strcpy(tb_fmt->fmt, *iter); + *iter = tb_fmt->fmt; + } else + *iter = NULL; + } + mutex_unlock(&btrace_mutex); +} + +static int module_trace_bprintk_format_notify(struct notifier_block *self, + unsigned long val, void *data) +{ + struct module *mod = data; + if (mod->num_trace_bprintk_fmt) { + const char **start = mod->trace_bprintk_fmt_start; + const char **end = start + mod->num_trace_bprintk_fmt; + + if (val == MODULE_STATE_COMING) + hold_module_trace_bprintk_format(start, end); + } + return 0; +} + +#else /* !CONFIG_MODULES */ +__init static int +module_trace_bprintk_format_notify(struct notifier_block *self, + unsigned long val, void *data) +{ + return 0; +} +#endif /* CONFIG_MODULES */ + + +__initdata_or_module static +struct notifier_block module_trace_bprintk_format_nb = { + .notifier_call = module_trace_bprintk_format_notify, +}; + +int __trace_bprintk(unsigned long ip, const char *fmt, ...) + { + int ret; + va_list ap; + + if (unlikely(!fmt)) + return 0; + + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; + + va_start(ap, fmt); + ret = trace_vbprintk(ip, fmt, ap); + va_end(ap); + return ret; +} +EXPORT_SYMBOL_GPL(__trace_bprintk); + +int __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap) + { + if (unlikely(!fmt)) + return 0; + + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; + + return trace_vbprintk(ip, fmt, ap); +} +EXPORT_SYMBOL_GPL(__ftrace_vbprintk); + +int __trace_printk(unsigned long ip, const char *fmt, ...) +{ + int ret; + va_list ap; + + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; + + va_start(ap, fmt); + ret = trace_vprintk(ip, fmt, ap); + va_end(ap); + return ret; +} +EXPORT_SYMBOL_GPL(__trace_printk); + +int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) +{ + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; + + return trace_vprintk(ip, fmt, ap); +} +EXPORT_SYMBOL_GPL(__ftrace_vprintk); + +static void * +t_start(struct seq_file *m, loff_t *pos) +{ + const char **fmt = __start___trace_bprintk_fmt + *pos; + + if ((unsigned long)fmt >= (unsigned long)__stop___trace_bprintk_fmt) + return NULL; + return fmt; +} + +static void *t_next(struct seq_file *m, void * v, loff_t *pos) +{ + (*pos)++; + return t_start(m, pos); +} + +static int t_show(struct seq_file *m, void *v) +{ + const char **fmt = v; + const char *str = *fmt; + int i; + + seq_printf(m, "0x%lx : \"", *(unsigned long *)fmt); + + /* + * Tabs and new lines need to be converted. + */ + for (i = 0; str[i]; i++) { + switch (str[i]) { + case '\n': + seq_puts(m, "\\n"); + break; + case '\t': + seq_puts(m, "\\t"); + break; + case '\\': + seq_puts(m, "\\"); + break; + case '"': + seq_puts(m, "\\\""); + break; + default: + seq_putc(m, str[i]); + } + } + seq_puts(m, "\"\n"); + + return 0; +} + +static void t_stop(struct seq_file *m, void *p) +{ +} + +static const struct seq_operations show_format_seq_ops = { + .start = t_start, + .next = t_next, + .show = t_show, + .stop = t_stop, +}; + +static int +ftrace_formats_open(struct inode *inode, struct file *file) +{ + return seq_open(file, &show_format_seq_ops); +} + +static const struct file_operations ftrace_formats_fops = { + .open = ftrace_formats_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + +static __init int init_trace_printk_function_export(void) +{ + struct dentry *d_tracer; + + d_tracer = tracing_init_dentry(); + if (!d_tracer) + return 0; + + trace_create_file("printk_formats", 0444, d_tracer, + NULL, &ftrace_formats_fops); + + return 0; +} + +fs_initcall(init_trace_printk_function_export); + +static __init int init_trace_printk(void) +{ + return register_module_notifier(&module_trace_bprintk_format_nb); +} + +early_initcall(init_trace_printk); diff --git a/kernel/trace/trace_sched_switch.c b/kernel/trace/trace_sched_switch.c index df175cb4564..5fca0f51fde 100644 --- a/kernel/trace/trace_sched_switch.c +++ b/kernel/trace/trace_sched_switch.c @@ -10,7 +10,7 @@ #include <linux/kallsyms.h> #include <linux/uaccess.h> #include <linux/ftrace.h> -#include <trace/sched.h> +#include <trace/events/sched.h> #include "trace.h" @@ -18,6 +18,36 @@ static struct trace_array *ctx_trace; static int __read_mostly tracer_enabled; static int sched_ref; static DEFINE_MUTEX(sched_register_mutex); +static int sched_stopped; + + +void +tracing_sched_switch_trace(struct trace_array *tr, + struct task_struct *prev, + struct task_struct *next, + unsigned long flags, int pc) +{ + struct ftrace_event_call *call = &event_context_switch; + struct ring_buffer *buffer = tr->buffer; + struct ring_buffer_event *event; + struct ctx_switch_entry *entry; + + event = trace_buffer_lock_reserve(buffer, TRACE_CTX, + sizeof(*entry), flags, pc); + if (!event) + return; + entry = ring_buffer_event_data(event); + entry->prev_pid = prev->pid; + entry->prev_prio = prev->prio; + entry->prev_state = prev->state; + entry->next_pid = next->pid; + entry->next_prio = next->prio; + entry->next_state = next->state; + entry->next_cpu = task_cpu(next); + + if (!filter_check_discard(call, entry, buffer, event)) + trace_buffer_unlock_commit(buffer, event, flags, pc); +} static void probe_sched_switch(struct rq *__rq, struct task_struct *prev, @@ -28,13 +58,13 @@ probe_sched_switch(struct rq *__rq, struct task_struct *prev, int cpu; int pc; - if (!sched_ref) + if (unlikely(!sched_ref)) return; tracing_record_cmdline(prev); tracing_record_cmdline(next); - if (!tracer_enabled) + if (!tracer_enabled || sched_stopped) return; pc = preempt_count(); @@ -43,11 +73,41 @@ probe_sched_switch(struct rq *__rq, struct task_struct *prev, data = ctx_trace->data[cpu]; if (likely(!atomic_read(&data->disabled))) - tracing_sched_switch_trace(ctx_trace, data, prev, next, flags, pc); + tracing_sched_switch_trace(ctx_trace, prev, next, flags, pc); local_irq_restore(flags); } +void +tracing_sched_wakeup_trace(struct trace_array *tr, + struct task_struct *wakee, + struct task_struct *curr, + unsigned long flags, int pc) +{ + struct ftrace_event_call *call = &event_wakeup; + struct ring_buffer_event *event; + struct ctx_switch_entry *entry; + struct ring_buffer *buffer = tr->buffer; + + event = trace_buffer_lock_reserve(buffer, TRACE_WAKE, + sizeof(*entry), flags, pc); + if (!event) + return; + entry = ring_buffer_event_data(event); + entry->prev_pid = curr->pid; + entry->prev_prio = curr->prio; + entry->prev_state = curr->state; + entry->next_pid = wakee->pid; + entry->next_prio = wakee->prio; + entry->next_state = wakee->state; + entry->next_cpu = task_cpu(wakee); + + if (!filter_check_discard(call, entry, buffer, event)) + ring_buffer_unlock_commit(buffer, event); + ftrace_trace_stack(tr->buffer, flags, 6, pc); + ftrace_trace_userstack(tr->buffer, flags, pc); +} + static void probe_sched_wakeup(struct rq *__rq, struct task_struct *wakee, int success) { @@ -55,18 +115,21 @@ probe_sched_wakeup(struct rq *__rq, struct task_struct *wakee, int success) unsigned long flags; int cpu, pc; - if (!likely(tracer_enabled)) + if (unlikely(!sched_ref)) return; - pc = preempt_count(); tracing_record_cmdline(current); + if (!tracer_enabled || sched_stopped) + return; + + pc = preempt_count(); local_irq_save(flags); cpu = raw_smp_processor_id(); data = ctx_trace->data[cpu]; if (likely(!atomic_read(&data->disabled))) - tracing_sched_wakeup_trace(ctx_trace, data, wakee, current, + tracing_sched_wakeup_trace(ctx_trace, wakee, current, flags, pc); local_irq_restore(flags); @@ -93,7 +156,7 @@ static int tracing_sched_register(void) ret = register_trace_sched_switch(probe_sched_switch); if (ret) { pr_info("sched trace: Couldn't activate tracepoint" - " probe to kernel_sched_schedule\n"); + " probe to kernel_sched_switch\n"); goto fail_deprobe_wake_new; } @@ -185,12 +248,6 @@ void tracing_sched_switch_assign_trace(struct trace_array *tr) ctx_trace = tr; } -static void start_sched_trace(struct trace_array *tr) -{ - tracing_reset_online_cpus(tr); - tracing_start_sched_switch_record(); -} - static void stop_sched_trace(struct trace_array *tr) { tracing_stop_sched_switch_record(); @@ -199,7 +256,8 @@ static void stop_sched_trace(struct trace_array *tr) static int sched_switch_trace_init(struct trace_array *tr) { ctx_trace = tr; - start_sched_trace(tr); + tracing_reset_online_cpus(tr); + tracing_start_sched_switch_record(); return 0; } @@ -211,13 +269,12 @@ static void sched_switch_trace_reset(struct trace_array *tr) static void sched_switch_trace_start(struct trace_array *tr) { - tracing_reset_online_cpus(tr); - tracing_start_sched_switch(); + sched_stopped = 0; } static void sched_switch_trace_stop(struct trace_array *tr) { - tracing_stop_sched_switch(); + sched_stopped = 1; } static struct tracer sched_switch_trace __read_mostly = @@ -227,6 +284,7 @@ static struct tracer sched_switch_trace __read_mostly = .reset = sched_switch_trace_reset, .start = sched_switch_trace_start, .stop = sched_switch_trace_stop, + .wait_pipe = poll_wait_pipe, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_sched_switch, #endif diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 43586b689e3..ad69f105a7c 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -15,7 +15,7 @@ #include <linux/kallsyms.h> #include <linux/uaccess.h> #include <linux/ftrace.h> -#include <trace/sched.h> +#include <trace/events/sched.h> #include "trace.h" @@ -25,12 +25,15 @@ static int __read_mostly tracer_enabled; static struct task_struct *wakeup_task; static int wakeup_cpu; static unsigned wakeup_prio = -1; +static int wakeup_rt; static raw_spinlock_t wakeup_lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; static void __wakeup_reset(struct trace_array *tr); +static int save_lat_flag; + #ifdef CONFIG_FUNCTION_TRACER /* * irqsoff uses its own tracer function to keep the overhead down: @@ -71,7 +74,7 @@ wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) if (task_cpu(wakeup_task) != cpu) goto unlock; - trace_function(tr, data, ip, parent_ip, flags, pc); + trace_function(tr, ip, parent_ip, flags, pc); unlock: __raw_spin_unlock(&wakeup_lock); @@ -135,9 +138,6 @@ probe_wakeup_sched_switch(struct rq *rq, struct task_struct *prev, pc = preempt_count(); - /* The task we are waiting for is waking up */ - data = wakeup_trace->data[wakeup_cpu]; - /* disable local data, not wakeup_cpu data */ cpu = raw_smp_processor_id(); disabled = atomic_inc_return(&wakeup_trace->data[cpu]->disabled); @@ -151,7 +151,11 @@ probe_wakeup_sched_switch(struct rq *rq, struct task_struct *prev, if (unlikely(!tracer_enabled || next != wakeup_task)) goto out_unlock; - trace_function(wakeup_trace, data, CALLER_ADDR1, CALLER_ADDR2, flags, pc); + /* The task we are waiting for is waking up */ + data = wakeup_trace->data[wakeup_cpu]; + + trace_function(wakeup_trace, CALLER_ADDR0, CALLER_ADDR1, flags, pc); + tracing_sched_switch_trace(wakeup_trace, prev, next, flags, pc); /* * usecs conversion is slow so we try to delay the conversion @@ -182,14 +186,6 @@ out: static void __wakeup_reset(struct trace_array *tr) { - struct trace_array_cpu *data; - int cpu; - - for_each_possible_cpu(cpu) { - data = tr->data[cpu]; - tracing_reset(tr, cpu); - } - wakeup_cpu = -1; wakeup_prio = -1; @@ -203,6 +199,8 @@ static void wakeup_reset(struct trace_array *tr) { unsigned long flags; + tracing_reset_online_cpus(tr); + local_irq_save(flags); __raw_spin_lock(&wakeup_lock); __wakeup_reset(tr); @@ -213,6 +211,7 @@ static void wakeup_reset(struct trace_array *tr) static void probe_wakeup(struct rq *rq, struct task_struct *p, int success) { + struct trace_array_cpu *data; int cpu = smp_processor_id(); unsigned long flags; long disabled; @@ -224,7 +223,7 @@ probe_wakeup(struct rq *rq, struct task_struct *p, int success) tracing_record_cmdline(p); tracing_record_cmdline(current); - if (likely(!rt_task(p)) || + if ((wakeup_rt && !rt_task(p)) || p->prio >= wakeup_prio || p->prio >= current->prio) return; @@ -252,9 +251,16 @@ probe_wakeup(struct rq *rq, struct task_struct *p, int success) local_save_flags(flags); - wakeup_trace->data[wakeup_cpu]->preempt_timestamp = ftrace_now(cpu); - trace_function(wakeup_trace, wakeup_trace->data[wakeup_cpu], - CALLER_ADDR1, CALLER_ADDR2, flags, pc); + data = wakeup_trace->data[wakeup_cpu]; + data->preempt_timestamp = ftrace_now(cpu); + tracing_sched_wakeup_trace(wakeup_trace, p, current, flags, pc); + + /* + * We must be careful in using CALLER_ADDR2. But since wake_up + * is not called by an assembly function (where as schedule is) + * it should be safe to use it here. + */ + trace_function(wakeup_trace, CALLER_ADDR1, CALLER_ADDR2, flags, pc); out_locked: __raw_spin_unlock(&wakeup_lock); @@ -262,12 +268,6 @@ out: atomic_dec(&wakeup_trace->data[cpu]->disabled); } -/* - * save_tracer_enabled is used to save the state of the tracer_enabled - * variable when we disable it when we open a trace output file. - */ -static int save_tracer_enabled; - static void start_wakeup_tracer(struct trace_array *tr) { int ret; @@ -289,7 +289,7 @@ static void start_wakeup_tracer(struct trace_array *tr) ret = register_trace_sched_switch(probe_wakeup_sched_switch); if (ret) { pr_info("sched trace: Couldn't activate tracepoint" - " probe to kernel_sched_schedule\n"); + " probe to kernel_sched_switch\n"); goto fail_deprobe_wake_new; } @@ -306,13 +306,10 @@ static void start_wakeup_tracer(struct trace_array *tr) register_ftrace_function(&trace_ops); - if (tracing_is_enabled()) { + if (tracing_is_enabled()) tracer_enabled = 1; - save_tracer_enabled = 1; - } else { + else tracer_enabled = 0; - save_tracer_enabled = 0; - } return; fail_deprobe_wake_new: @@ -324,53 +321,54 @@ fail_deprobe: static void stop_wakeup_tracer(struct trace_array *tr) { tracer_enabled = 0; - save_tracer_enabled = 0; unregister_ftrace_function(&trace_ops); unregister_trace_sched_switch(probe_wakeup_sched_switch); unregister_trace_sched_wakeup_new(probe_wakeup); unregister_trace_sched_wakeup(probe_wakeup); } -static int wakeup_tracer_init(struct trace_array *tr) +static int __wakeup_tracer_init(struct trace_array *tr) { + save_lat_flag = trace_flags & TRACE_ITER_LATENCY_FMT; + trace_flags |= TRACE_ITER_LATENCY_FMT; + + tracing_max_latency = 0; wakeup_trace = tr; start_wakeup_tracer(tr); return 0; } +static int wakeup_tracer_init(struct trace_array *tr) +{ + wakeup_rt = 0; + return __wakeup_tracer_init(tr); +} + +static int wakeup_rt_tracer_init(struct trace_array *tr) +{ + wakeup_rt = 1; + return __wakeup_tracer_init(tr); +} + static void wakeup_tracer_reset(struct trace_array *tr) { stop_wakeup_tracer(tr); /* make sure we put back any tasks we are tracing */ wakeup_reset(tr); + + if (!save_lat_flag) + trace_flags &= ~TRACE_ITER_LATENCY_FMT; } static void wakeup_tracer_start(struct trace_array *tr) { wakeup_reset(tr); tracer_enabled = 1; - save_tracer_enabled = 1; } static void wakeup_tracer_stop(struct trace_array *tr) { tracer_enabled = 0; - save_tracer_enabled = 0; -} - -static void wakeup_tracer_open(struct trace_iterator *iter) -{ - /* stop the trace while dumping */ - tracer_enabled = 0; -} - -static void wakeup_tracer_close(struct trace_iterator *iter) -{ - /* forget about any processes we were recording */ - if (save_tracer_enabled) { - wakeup_reset(iter->tr); - tracer_enabled = 1; - } } static struct tracer wakeup_tracer __read_mostly = @@ -380,8 +378,20 @@ static struct tracer wakeup_tracer __read_mostly = .reset = wakeup_tracer_reset, .start = wakeup_tracer_start, .stop = wakeup_tracer_stop, - .open = wakeup_tracer_open, - .close = wakeup_tracer_close, + .print_max = 1, +#ifdef CONFIG_FTRACE_SELFTEST + .selftest = trace_selftest_startup_wakeup, +#endif +}; + +static struct tracer wakeup_rt_tracer __read_mostly = +{ + .name = "wakeup_rt", + .init = wakeup_rt_tracer_init, + .reset = wakeup_tracer_reset, + .start = wakeup_tracer_start, + .stop = wakeup_tracer_stop, + .wait_pipe = poll_wait_pipe, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_wakeup, @@ -396,6 +406,10 @@ __init static int init_wakeup_tracer(void) if (ret) return ret; + ret = register_tracer(&wakeup_rt_tracer); + if (ret) + return ret; + return 0; } device_initcall(init_wakeup_tracer); diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 88c8eb70f54..d2cdbabb4ea 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -1,5 +1,6 @@ /* Include in trace.c */ +#include <linux/stringify.h> #include <linux/kthread.h> #include <linux/delay.h> @@ -9,11 +10,13 @@ static inline int trace_valid_entry(struct trace_entry *entry) case TRACE_FN: case TRACE_CTX: case TRACE_WAKE: - case TRACE_CONT: case TRACE_STACK: case TRACE_PRINT: case TRACE_SPECIAL: case TRACE_BRANCH: + case TRACE_GRAPH_ENT: + case TRACE_GRAPH_RET: + case TRACE_HW_BRANCHES: return 1; } return 0; @@ -23,10 +26,20 @@ static int trace_test_buffer_cpu(struct trace_array *tr, int cpu) { struct ring_buffer_event *event; struct trace_entry *entry; + unsigned int loops = 0; while ((event = ring_buffer_consume(tr->buffer, cpu, NULL))) { entry = ring_buffer_event_data(event); + /* + * The ring buffer is a size of trace_buf_size, if + * we loop more than the size, there's something wrong + * with the ring buffer. + */ + if (loops++ > trace_buf_size) { + printk(KERN_CONT ".. bad ring buffer "); + goto failed; + } if (!trace_valid_entry(entry)) { printk(KERN_CONT ".. invalid entry %d ", entry->type); @@ -57,11 +70,20 @@ static int trace_test_buffer(struct trace_array *tr, unsigned long *count) cnt = ring_buffer_entries(tr->buffer); + /* + * The trace_test_buffer_cpu runs a while loop to consume all data. + * If the calling tracer is broken, and is constantly filling + * the buffer, this will run forever, and hard lock the box. + * We disable the ring buffer while we do this test to prevent + * a hard lock up. + */ + tracing_off(); for_each_possible_cpu(cpu) { ret = trace_test_buffer_cpu(tr, cpu); if (ret) break; } + tracing_on(); __raw_spin_unlock(&ftrace_max_lock); local_irq_restore(flags); @@ -80,9 +102,6 @@ static inline void warn_failed_init_tracer(struct tracer *trace, int init_ret) #ifdef CONFIG_DYNAMIC_FTRACE -#define __STR(x) #x -#define STR(x) __STR(x) - /* Test dynamic code modification and ftrace filters */ int trace_selftest_startup_dynamic_tracing(struct tracer *trace, struct trace_array *tr, @@ -106,17 +125,17 @@ int trace_selftest_startup_dynamic_tracing(struct tracer *trace, func(); /* - * Some archs *cough*PowerPC*cough* add charachters to the + * Some archs *cough*PowerPC*cough* add characters to the * start of the function names. We simply put a '*' to - * accomodate them. + * accommodate them. */ - func_name = "*" STR(DYN_FTRACE_TEST_NAME); + func_name = "*" __stringify(DYN_FTRACE_TEST_NAME); /* filter only on our function */ ftrace_set_filter(func_name, strlen(func_name), 1); /* enable tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); goto out; @@ -170,6 +189,7 @@ int trace_selftest_startup_dynamic_tracing(struct tracer *trace, #else # define trace_selftest_startup_dynamic_tracing(trace, tr, func) ({ 0; }) #endif /* CONFIG_DYNAMIC_FTRACE */ + /* * Simple verification test of ftrace function tracer. * Enable ftrace, sleep 1/10 second, and then read the trace @@ -190,7 +210,7 @@ trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr) ftrace_enabled = 1; tracer_enabled = 1; - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); goto out; @@ -228,6 +248,91 @@ trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr) } #endif /* CONFIG_FUNCTION_TRACER */ + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + +/* Maximum number of functions to trace before diagnosing a hang */ +#define GRAPH_MAX_FUNC_TEST 100000000 + +static void __ftrace_dump(bool disable_tracing); +static unsigned int graph_hang_thresh; + +/* Wrap the real function entry probe to avoid possible hanging */ +static int trace_graph_entry_watchdog(struct ftrace_graph_ent *trace) +{ + /* This is harmlessly racy, we want to approximately detect a hang */ + if (unlikely(++graph_hang_thresh > GRAPH_MAX_FUNC_TEST)) { + ftrace_graph_stop(); + printk(KERN_WARNING "BUG: Function graph tracer hang!\n"); + if (ftrace_dump_on_oops) + __ftrace_dump(false); + return 0; + } + + return trace_graph_entry(trace); +} + +/* + * Pretty much the same than for the function tracer from which the selftest + * has been borrowed. + */ +int +trace_selftest_startup_function_graph(struct tracer *trace, + struct trace_array *tr) +{ + int ret; + unsigned long count; + + /* + * Simulate the init() callback but we attach a watchdog callback + * to detect and recover from possible hangs + */ + tracing_reset_online_cpus(tr); + set_graph_array(tr); + ret = register_ftrace_graph(&trace_graph_return, + &trace_graph_entry_watchdog); + if (ret) { + warn_failed_init_tracer(trace, ret); + goto out; + } + tracing_start_cmdline_record(); + + /* Sleep for a 1/10 of a second */ + msleep(100); + + /* Have we just recovered from a hang? */ + if (graph_hang_thresh > GRAPH_MAX_FUNC_TEST) { + tracing_selftest_disabled = true; + ret = -1; + goto out; + } + + tracing_stop(); + + /* check the trace buffer */ + ret = trace_test_buffer(tr, &count); + + trace->reset(tr); + tracing_start(); + + if (!ret && !count) { + printk(KERN_CONT ".. no entries found .."); + ret = -1; + goto out; + } + + /* Don't test dynamic tracing, the function tracer already did */ + +out: + /* Stop it if we failed */ + if (ret) + ftrace_graph_stop(); + + return ret; +} +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ + + #ifdef CONFIG_IRQSOFF_TRACER int trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr) @@ -237,7 +342,7 @@ trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr) int ret; /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -249,6 +354,14 @@ trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr) local_irq_disable(); udelay(100); local_irq_enable(); + + /* + * Stop the tracer to avoid a warning subsequent + * to buffer flipping failure because tracing_stop() + * disables the tr and max buffers, making flipping impossible + * in case of parallels max irqs off latencies. + */ + trace->stop(tr); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ @@ -291,7 +404,7 @@ trace_selftest_startup_preemptoff(struct tracer *trace, struct trace_array *tr) } /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -303,6 +416,14 @@ trace_selftest_startup_preemptoff(struct tracer *trace, struct trace_array *tr) preempt_disable(); udelay(100); preempt_enable(); + + /* + * Stop the tracer to avoid a warning subsequent + * to buffer flipping failure because tracing_stop() + * disables the tr and max buffers, making flipping impossible + * in case of parallels max preempt off latencies. + */ + trace->stop(tr); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ @@ -345,10 +466,10 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * } /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); - goto out; + goto out_no_start; } /* reset the max latency */ @@ -362,31 +483,35 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * /* reverse the order of preempt vs irqs */ local_irq_enable(); + /* + * Stop the tracer to avoid a warning subsequent + * to buffer flipping failure because tracing_stop() + * disables the tr and max buffers, making flipping impossible + * in case of parallels max irqs/preempt off latencies. + */ + trace->stop(tr); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ ret = trace_test_buffer(tr, NULL); - if (ret) { - tracing_start(); + if (ret) goto out; - } ret = trace_test_buffer(&max_tr, &count); - if (ret) { - tracing_start(); + if (ret) goto out; - } if (!ret && !count) { printk(KERN_CONT ".. no entries found .."); ret = -1; - tracing_start(); goto out; } /* do the test by disabling interrupts first this time */ tracing_max_latency = 0; tracing_start(); + trace->start(tr); + preempt_disable(); local_irq_disable(); udelay(100); @@ -394,6 +519,7 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * /* reverse the order of preempt vs irqs */ local_irq_enable(); + trace->stop(tr); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ @@ -409,9 +535,10 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * goto out; } - out: - trace->reset(tr); +out: tracing_start(); +out_no_start: + trace->reset(tr); tracing_max_latency = save_max; return ret; @@ -477,7 +604,7 @@ trace_selftest_startup_wakeup(struct tracer *trace, struct trace_array *tr) wait_for_completion(&isrt); /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -538,7 +665,7 @@ trace_selftest_startup_sched_switch(struct tracer *trace, struct trace_array *tr int ret; /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -570,10 +697,10 @@ trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr) int ret; /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); - return 0; + return ret; } /* Sleep for a 1/10 of a second */ @@ -585,6 +712,11 @@ trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr) trace->reset(tr); tracing_start(); + if (!ret && !count) { + printk(KERN_CONT ".. no entries found .."); + ret = -1; + } + return ret; } #endif /* CONFIG_SYSPROF_TRACER */ @@ -597,7 +729,7 @@ trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr) int ret; /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -612,6 +744,67 @@ trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr) trace->reset(tr); tracing_start(); + if (!ret && !count) { + printk(KERN_CONT ".. no entries found .."); + ret = -1; + } + return ret; } #endif /* CONFIG_BRANCH_TRACER */ + +#ifdef CONFIG_HW_BRANCH_TRACER +int +trace_selftest_startup_hw_branches(struct tracer *trace, + struct trace_array *tr) +{ + struct trace_iterator *iter; + struct tracer tracer; + unsigned long count; + int ret; + + if (!trace->open) { + printk(KERN_CONT "missing open function..."); + return -1; + } + + ret = tracer_init(trace, tr); + if (ret) { + warn_failed_init_tracer(trace, ret); + return ret; + } + + /* + * The hw-branch tracer needs to collect the trace from the various + * cpu trace buffers - before tracing is stopped. + */ + iter = kzalloc(sizeof(*iter), GFP_KERNEL); + if (!iter) + return -ENOMEM; + + memcpy(&tracer, trace, sizeof(tracer)); + + iter->trace = &tracer; + iter->tr = tr; + iter->pos = -1; + mutex_init(&iter->mutex); + + trace->open(iter); + + mutex_destroy(&iter->mutex); + kfree(iter); + + tracing_stop(); + + ret = trace_test_buffer(tr, &count); + trace->reset(tr); + tracing_start(); + + if (!ret && !count) { + printk(KERN_CONT "no entries found.."); + ret = -1; + } + + return ret; +} +#endif /* CONFIG_HW_BRANCH_TRACER */ diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index d0871bc0aca..0f6facb050a 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -186,43 +186,33 @@ static const struct file_operations stack_max_size_fops = { }; static void * -t_next(struct seq_file *m, void *v, loff_t *pos) +__next(struct seq_file *m, loff_t *pos) { - long i; - - (*pos)++; + long n = *pos - 1; - if (v == SEQ_START_TOKEN) - i = 0; - else { - i = *(long *)v; - i++; - } - - if (i >= max_stack_trace.nr_entries || - stack_dump_trace[i] == ULONG_MAX) + if (n >= max_stack_trace.nr_entries || stack_dump_trace[n] == ULONG_MAX) return NULL; - m->private = (void *)i; - + m->private = (void *)n; return &m->private; } -static void *t_start(struct seq_file *m, loff_t *pos) +static void * +t_next(struct seq_file *m, void *v, loff_t *pos) { - void *t = SEQ_START_TOKEN; - loff_t l = 0; + (*pos)++; + return __next(m, pos); +} +static void *t_start(struct seq_file *m, loff_t *pos) +{ local_irq_disable(); __raw_spin_lock(&max_stack_lock); if (*pos == 0) return SEQ_START_TOKEN; - for (; t && l < *pos; t = t_next(m, t, &l)) - ; - - return t; + return __next(m, pos); } static void t_stop(struct seq_file *m, void *p) @@ -234,15 +224,19 @@ static void t_stop(struct seq_file *m, void *p) static int trace_lookup_stack(struct seq_file *m, long i) { unsigned long addr = stack_dump_trace[i]; -#ifdef CONFIG_KALLSYMS - char str[KSYM_SYMBOL_LEN]; - sprint_symbol(str, addr); + return seq_printf(m, "%pF\n", (void *)addr); +} - return seq_printf(m, "%s\n", str); -#else - return seq_printf(m, "%p\n", (void*)addr); -#endif +static void print_disabled(struct seq_file *m) +{ + seq_puts(m, "#\n" + "# Stack tracer disabled\n" + "#\n" + "# To enable the stack tracer, either add 'stacktrace' to the\n" + "# kernel command line\n" + "# or 'echo 1 > /proc/sys/kernel/stack_tracer_enabled'\n" + "#\n"); } static int t_show(struct seq_file *m, void *v) @@ -251,10 +245,14 @@ static int t_show(struct seq_file *m, void *v) int size; if (v == SEQ_START_TOKEN) { - seq_printf(m, " Depth Size Location" + seq_printf(m, " Depth Size Location" " (%d entries)\n" - " ----- ---- --------\n", - max_stack_trace.nr_entries); + " ----- ---- --------\n", + max_stack_trace.nr_entries - 1); + + if (!stack_tracer_enabled && !max_stack_size) + print_disabled(m); + return 0; } @@ -286,17 +284,14 @@ static const struct seq_operations stack_trace_seq_ops = { static int stack_trace_open(struct inode *inode, struct file *file) { - int ret; - - ret = seq_open(file, &stack_trace_seq_ops); - - return ret; + return seq_open(file, &stack_trace_seq_ops); } static const struct file_operations stack_trace_fops = { .open = stack_trace_open, .read = seq_read, .llseek = seq_lseek, + .release = seq_release, }; int @@ -311,10 +306,10 @@ stack_trace_sysctl(struct ctl_table *table, int write, ret = proc_dointvec(table, write, file, buffer, lenp, ppos); if (ret || !write || - (last_stack_tracer_enabled == stack_tracer_enabled)) + (last_stack_tracer_enabled == !!stack_tracer_enabled)) goto out; - last_stack_tracer_enabled = stack_tracer_enabled; + last_stack_tracer_enabled = !!stack_tracer_enabled; if (stack_tracer_enabled) register_ftrace_function(&trace_ops); @@ -337,19 +332,14 @@ __setup("stacktrace", enable_stacktrace); static __init int stack_trace_init(void) { struct dentry *d_tracer; - struct dentry *entry; d_tracer = tracing_init_dentry(); - entry = debugfs_create_file("stack_max_size", 0644, d_tracer, - &max_stack_size, &stack_max_size_fops); - if (!entry) - pr_warning("Could not create debugfs 'stack_max_size' entry\n"); + trace_create_file("stack_max_size", 0644, d_tracer, + &max_stack_size, &stack_max_size_fops); - entry = debugfs_create_file("stack_trace", 0444, d_tracer, - NULL, &stack_trace_fops); - if (!entry) - pr_warning("Could not create debugfs 'stack_trace' entry\n"); + trace_create_file("stack_trace", 0444, d_tracer, + NULL, &stack_trace_fops); if (stack_tracer_enabled) register_ftrace_function(&trace_ops); diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c new file mode 100644 index 00000000000..a4bb239eb98 --- /dev/null +++ b/kernel/trace/trace_stat.c @@ -0,0 +1,387 @@ +/* + * Infrastructure for statistic tracing (histogram output). + * + * Copyright (C) 2008-2009 Frederic Weisbecker <fweisbec@gmail.com> + * + * Based on the code from trace_branch.c which is + * Copyright (C) 2008 Steven Rostedt <srostedt@redhat.com> + * + */ + + +#include <linux/list.h> +#include <linux/rbtree.h> +#include <linux/debugfs.h> +#include "trace_stat.h" +#include "trace.h" + + +/* + * List of stat red-black nodes from a tracer + * We use a such tree to sort quickly the stat + * entries from the tracer. + */ +struct stat_node { + struct rb_node node; + void *stat; +}; + +/* A stat session is the stats output in one file */ +struct stat_session { + struct list_head session_list; + struct tracer_stat *ts; + struct rb_root stat_root; + struct mutex stat_mutex; + struct dentry *file; +}; + +/* All of the sessions currently in use. Each stat file embed one session */ +static LIST_HEAD(all_stat_sessions); +static DEFINE_MUTEX(all_stat_sessions_mutex); + +/* The root directory for all stat files */ +static struct dentry *stat_dir; + +/* + * Iterate through the rbtree using a post order traversal path + * to release the next node. + * It won't necessary release one at each iteration + * but it will at least advance closer to the next one + * to be released. + */ +static struct rb_node *release_next(struct tracer_stat *ts, + struct rb_node *node) +{ + struct stat_node *snode; + struct rb_node *parent = rb_parent(node); + + if (node->rb_left) + return node->rb_left; + else if (node->rb_right) + return node->rb_right; + else { + if (!parent) + ; + else if (parent->rb_left == node) + parent->rb_left = NULL; + else + parent->rb_right = NULL; + + snode = container_of(node, struct stat_node, node); + if (ts->stat_release) + ts->stat_release(snode->stat); + kfree(snode); + + return parent; + } +} + +static void __reset_stat_session(struct stat_session *session) +{ + struct rb_node *node = session->stat_root.rb_node; + + while (node) + node = release_next(session->ts, node); + + session->stat_root = RB_ROOT; +} + +static void reset_stat_session(struct stat_session *session) +{ + mutex_lock(&session->stat_mutex); + __reset_stat_session(session); + mutex_unlock(&session->stat_mutex); +} + +static void destroy_session(struct stat_session *session) +{ + debugfs_remove(session->file); + __reset_stat_session(session); + mutex_destroy(&session->stat_mutex); + kfree(session); +} + +typedef int (*cmp_stat_t)(void *, void *); + +static int insert_stat(struct rb_root *root, void *stat, cmp_stat_t cmp) +{ + struct rb_node **new = &(root->rb_node), *parent = NULL; + struct stat_node *data; + + data = kzalloc(sizeof(*data), GFP_KERNEL); + if (!data) + return -ENOMEM; + data->stat = stat; + + /* + * Figure out where to put new node + * This is a descendent sorting + */ + while (*new) { + struct stat_node *this; + int result; + + this = container_of(*new, struct stat_node, node); + result = cmp(data->stat, this->stat); + + parent = *new; + if (result >= 0) + new = &((*new)->rb_left); + else + new = &((*new)->rb_right); + } + + rb_link_node(&data->node, parent, new); + rb_insert_color(&data->node, root); + return 0; +} + +/* + * For tracers that don't provide a stat_cmp callback. + * This one will force an insertion as right-most node + * in the rbtree. + */ +static int dummy_cmp(void *p1, void *p2) +{ + return -1; +} + +/* + * Initialize the stat rbtree at each trace_stat file opening. + * All of these copies and sorting are required on all opening + * since the stats could have changed between two file sessions. + */ +static int stat_seq_init(struct stat_session *session) +{ + struct tracer_stat *ts = session->ts; + struct rb_root *root = &session->stat_root; + void *stat; + int ret = 0; + int i; + + mutex_lock(&session->stat_mutex); + __reset_stat_session(session); + + if (!ts->stat_cmp) + ts->stat_cmp = dummy_cmp; + + stat = ts->stat_start(ts); + if (!stat) + goto exit; + + ret = insert_stat(root, stat, ts->stat_cmp); + if (ret) + goto exit; + + /* + * Iterate over the tracer stat entries and store them in an rbtree. + */ + for (i = 1; ; i++) { + stat = ts->stat_next(stat, i); + + /* End of insertion */ + if (!stat) + break; + + ret = insert_stat(root, stat, ts->stat_cmp); + if (ret) + goto exit_free_rbtree; + } + +exit: + mutex_unlock(&session->stat_mutex); + return ret; + +exit_free_rbtree: + __reset_stat_session(session); + mutex_unlock(&session->stat_mutex); + return ret; +} + + +static void *stat_seq_start(struct seq_file *s, loff_t *pos) +{ + struct stat_session *session = s->private; + struct rb_node *node; + int n = *pos; + int i; + + /* Prevent from tracer switch or rbtree modification */ + mutex_lock(&session->stat_mutex); + + /* If we are in the beginning of the file, print the headers */ + if (session->ts->stat_headers) { + if (n == 0) + return SEQ_START_TOKEN; + n--; + } + + node = rb_first(&session->stat_root); + for (i = 0; node && i < n; i++) + node = rb_next(node); + + return node; +} + +static void *stat_seq_next(struct seq_file *s, void *p, loff_t *pos) +{ + struct stat_session *session = s->private; + struct rb_node *node = p; + + (*pos)++; + + if (p == SEQ_START_TOKEN) + return rb_first(&session->stat_root); + + return rb_next(node); +} + +static void stat_seq_stop(struct seq_file *s, void *p) +{ + struct stat_session *session = s->private; + mutex_unlock(&session->stat_mutex); +} + +static int stat_seq_show(struct seq_file *s, void *v) +{ + struct stat_session *session = s->private; + struct stat_node *l = container_of(v, struct stat_node, node); + + if (v == SEQ_START_TOKEN) + return session->ts->stat_headers(s); + + return session->ts->stat_show(s, l->stat); +} + +static const struct seq_operations trace_stat_seq_ops = { + .start = stat_seq_start, + .next = stat_seq_next, + .stop = stat_seq_stop, + .show = stat_seq_show +}; + +/* The session stat is refilled and resorted at each stat file opening */ +static int tracing_stat_open(struct inode *inode, struct file *file) +{ + int ret; + struct seq_file *m; + struct stat_session *session = inode->i_private; + + ret = stat_seq_init(session); + if (ret) + return ret; + + ret = seq_open(file, &trace_stat_seq_ops); + if (ret) { + reset_stat_session(session); + return ret; + } + + m = file->private_data; + m->private = session; + return ret; +} + +/* + * Avoid consuming memory with our now useless rbtree. + */ +static int tracing_stat_release(struct inode *i, struct file *f) +{ + struct stat_session *session = i->i_private; + + reset_stat_session(session); + + return seq_release(i, f); +} + +static const struct file_operations tracing_stat_fops = { + .open = tracing_stat_open, + .read = seq_read, + .llseek = seq_lseek, + .release = tracing_stat_release +}; + +static int tracing_stat_init(void) +{ + struct dentry *d_tracing; + + d_tracing = tracing_init_dentry(); + + stat_dir = debugfs_create_dir("trace_stat", d_tracing); + if (!stat_dir) + pr_warning("Could not create debugfs " + "'trace_stat' entry\n"); + return 0; +} + +static int init_stat_file(struct stat_session *session) +{ + if (!stat_dir && tracing_stat_init()) + return -ENODEV; + + session->file = debugfs_create_file(session->ts->name, 0644, + stat_dir, + session, &tracing_stat_fops); + if (!session->file) + return -ENOMEM; + return 0; +} + +int register_stat_tracer(struct tracer_stat *trace) +{ + struct stat_session *session, *node; + int ret; + + if (!trace) + return -EINVAL; + + if (!trace->stat_start || !trace->stat_next || !trace->stat_show) + return -EINVAL; + + /* Already registered? */ + mutex_lock(&all_stat_sessions_mutex); + list_for_each_entry(node, &all_stat_sessions, session_list) { + if (node->ts == trace) { + mutex_unlock(&all_stat_sessions_mutex); + return -EINVAL; + } + } + mutex_unlock(&all_stat_sessions_mutex); + + /* Init the session */ + session = kzalloc(sizeof(*session), GFP_KERNEL); + if (!session) + return -ENOMEM; + + session->ts = trace; + INIT_LIST_HEAD(&session->session_list); + mutex_init(&session->stat_mutex); + + ret = init_stat_file(session); + if (ret) { + destroy_session(session); + return ret; + } + + /* Register */ + mutex_lock(&all_stat_sessions_mutex); + list_add_tail(&session->session_list, &all_stat_sessions); + mutex_unlock(&all_stat_sessions_mutex); + + return 0; +} + +void unregister_stat_tracer(struct tracer_stat *trace) +{ + struct stat_session *node, *tmp; + + mutex_lock(&all_stat_sessions_mutex); + list_for_each_entry_safe(node, tmp, &all_stat_sessions, session_list) { + if (node->ts == trace) { + list_del(&node->session_list); + destroy_session(node); + break; + } + } + mutex_unlock(&all_stat_sessions_mutex); +} diff --git a/kernel/trace/trace_stat.h b/kernel/trace/trace_stat.h new file mode 100644 index 00000000000..8f03914b9a6 --- /dev/null +++ b/kernel/trace/trace_stat.h @@ -0,0 +1,33 @@ +#ifndef __TRACE_STAT_H +#define __TRACE_STAT_H + +#include <linux/seq_file.h> + +/* + * If you want to provide a stat file (one-shot statistics), fill + * an iterator with stat_start/stat_next and a stat_show callbacks. + * The others callbacks are optional. + */ +struct tracer_stat { + /* The name of your stat file */ + const char *name; + /* Iteration over statistic entries */ + void *(*stat_start)(struct tracer_stat *trace); + void *(*stat_next)(void *prev, int idx); + /* Compare two entries for stats sorting */ + int (*stat_cmp)(void *p1, void *p2); + /* Print a stat entry */ + int (*stat_show)(struct seq_file *s, void *p); + /* Release an entry */ + void (*stat_release)(void *stat); + /* Print the headers of your stat entries */ + int (*stat_headers)(struct seq_file *s); +}; + +/* + * Destroy or create a stat file + */ +extern int register_stat_tracer(struct tracer_stat *trace); +extern void unregister_stat_tracer(struct tracer_stat *trace); + +#endif /* __TRACE_STAT_H */ diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c new file mode 100644 index 00000000000..8712ce3c6a0 --- /dev/null +++ b/kernel/trace/trace_syscalls.c @@ -0,0 +1,523 @@ +#include <trace/syscall.h> +#include <trace/events/syscalls.h> +#include <linux/kernel.h> +#include <linux/ftrace.h> +#include <linux/perf_counter.h> +#include <asm/syscall.h> + +#include "trace_output.h" +#include "trace.h" + +static DEFINE_MUTEX(syscall_trace_lock); +static int sys_refcount_enter; +static int sys_refcount_exit; +static DECLARE_BITMAP(enabled_enter_syscalls, NR_syscalls); +static DECLARE_BITMAP(enabled_exit_syscalls, NR_syscalls); + +enum print_line_t +print_syscall_enter(struct trace_iterator *iter, int flags) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *ent = iter->ent; + struct syscall_trace_enter *trace; + struct syscall_metadata *entry; + int i, ret, syscall; + + trace = (typeof(trace))ent; + syscall = trace->nr; + entry = syscall_nr_to_meta(syscall); + + if (!entry) + goto end; + + if (entry->enter_id != ent->type) { + WARN_ON_ONCE(1); + goto end; + } + + ret = trace_seq_printf(s, "%s(", entry->name); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + for (i = 0; i < entry->nb_args; i++) { + /* parameter types */ + if (trace_flags & TRACE_ITER_VERBOSE) { + ret = trace_seq_printf(s, "%s ", entry->types[i]); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + /* parameter values */ + ret = trace_seq_printf(s, "%s: %lx%s", entry->args[i], + trace->args[i], + i == entry->nb_args - 1 ? "" : ", "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + + ret = trace_seq_putc(s, ')'); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + +end: + ret = trace_seq_putc(s, '\n'); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +enum print_line_t +print_syscall_exit(struct trace_iterator *iter, int flags) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *ent = iter->ent; + struct syscall_trace_exit *trace; + int syscall; + struct syscall_metadata *entry; + int ret; + + trace = (typeof(trace))ent; + syscall = trace->nr; + entry = syscall_nr_to_meta(syscall); + + if (!entry) { + trace_seq_printf(s, "\n"); + return TRACE_TYPE_HANDLED; + } + + if (entry->exit_id != ent->type) { + WARN_ON_ONCE(1); + return TRACE_TYPE_UNHANDLED; + } + + ret = trace_seq_printf(s, "%s -> 0x%lx\n", entry->name, + trace->ret); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +extern char *__bad_type_size(void); + +#define SYSCALL_FIELD(type, name) \ + sizeof(type) != sizeof(trace.name) ? \ + __bad_type_size() : \ + #type, #name, offsetof(typeof(trace), name), sizeof(trace.name) + +int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s) +{ + int i; + int nr; + int ret; + struct syscall_metadata *entry; + struct syscall_trace_enter trace; + int offset = offsetof(struct syscall_trace_enter, args); + + nr = syscall_name_to_nr(call->data); + entry = syscall_nr_to_meta(nr); + + if (!entry) + return 0; + + ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n", + SYSCALL_FIELD(int, nr)); + if (!ret) + return 0; + + for (i = 0; i < entry->nb_args; i++) { + ret = trace_seq_printf(s, "\tfield:%s %s;", entry->types[i], + entry->args[i]); + if (!ret) + return 0; + ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;\n", offset, + sizeof(unsigned long)); + if (!ret) + return 0; + offset += sizeof(unsigned long); + } + + trace_seq_puts(s, "\nprint fmt: \""); + for (i = 0; i < entry->nb_args; i++) { + ret = trace_seq_printf(s, "%s: 0x%%0%zulx%s", entry->args[i], + sizeof(unsigned long), + i == entry->nb_args - 1 ? "" : ", "); + if (!ret) + return 0; + } + trace_seq_putc(s, '"'); + + for (i = 0; i < entry->nb_args; i++) { + ret = trace_seq_printf(s, ", ((unsigned long)(REC->%s))", + entry->args[i]); + if (!ret) + return 0; + } + + return trace_seq_putc(s, '\n'); +} + +int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) +{ + int ret; + struct syscall_trace_exit trace; + + ret = trace_seq_printf(s, + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n", + SYSCALL_FIELD(int, nr), + SYSCALL_FIELD(unsigned long, ret)); + if (!ret) + return 0; + + return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n"); +} + +int syscall_enter_define_fields(struct ftrace_event_call *call) +{ + struct syscall_trace_enter trace; + struct syscall_metadata *meta; + int ret; + int nr; + int i; + int offset = offsetof(typeof(trace), args); + + nr = syscall_name_to_nr(call->data); + meta = syscall_nr_to_meta(nr); + + if (!meta) + return 0; + + ret = trace_define_common_fields(call); + if (ret) + return ret; + + for (i = 0; i < meta->nb_args; i++) { + ret = trace_define_field(call, meta->types[i], + meta->args[i], offset, + sizeof(unsigned long), 0, + FILTER_OTHER); + offset += sizeof(unsigned long); + } + + return ret; +} + +int syscall_exit_define_fields(struct ftrace_event_call *call) +{ + struct syscall_trace_exit trace; + int ret; + + ret = trace_define_common_fields(call); + if (ret) + return ret; + + ret = trace_define_field(call, SYSCALL_FIELD(unsigned long, ret), 0, + FILTER_OTHER); + + return ret; +} + +void ftrace_syscall_enter(struct pt_regs *regs, long id) +{ + struct syscall_trace_enter *entry; + struct syscall_metadata *sys_data; + struct ring_buffer_event *event; + struct ring_buffer *buffer; + int size; + int syscall_nr; + + syscall_nr = syscall_get_nr(current, regs); + if (syscall_nr < 0) + return; + if (!test_bit(syscall_nr, enabled_enter_syscalls)) + return; + + sys_data = syscall_nr_to_meta(syscall_nr); + if (!sys_data) + return; + + size = sizeof(*entry) + sizeof(unsigned long) * sys_data->nb_args; + + event = trace_current_buffer_lock_reserve(&buffer, sys_data->enter_id, + size, 0, 0); + if (!event) + return; + + entry = ring_buffer_event_data(event); + entry->nr = syscall_nr; + syscall_get_arguments(current, regs, 0, sys_data->nb_args, entry->args); + + if (!filter_current_check_discard(buffer, sys_data->enter_event, + entry, event)) + trace_current_buffer_unlock_commit(buffer, event, 0, 0); +} + +void ftrace_syscall_exit(struct pt_regs *regs, long ret) +{ + struct syscall_trace_exit *entry; + struct syscall_metadata *sys_data; + struct ring_buffer_event *event; + struct ring_buffer *buffer; + int syscall_nr; + + syscall_nr = syscall_get_nr(current, regs); + if (syscall_nr < 0) + return; + if (!test_bit(syscall_nr, enabled_exit_syscalls)) + return; + + sys_data = syscall_nr_to_meta(syscall_nr); + if (!sys_data) + return; + + event = trace_current_buffer_lock_reserve(&buffer, sys_data->exit_id, + sizeof(*entry), 0, 0); + if (!event) + return; + + entry = ring_buffer_event_data(event); + entry->nr = syscall_nr; + entry->ret = syscall_get_return_value(current, regs); + + if (!filter_current_check_discard(buffer, sys_data->exit_event, + entry, event)) + trace_current_buffer_unlock_commit(buffer, event, 0, 0); +} + +int reg_event_syscall_enter(void *ptr) +{ + int ret = 0; + int num; + char *name; + + name = (char *)ptr; + num = syscall_name_to_nr(name); + if (num < 0 || num >= NR_syscalls) + return -ENOSYS; + mutex_lock(&syscall_trace_lock); + if (!sys_refcount_enter) + ret = register_trace_sys_enter(ftrace_syscall_enter); + if (ret) { + pr_info("event trace: Could not activate" + "syscall entry trace point"); + } else { + set_bit(num, enabled_enter_syscalls); + sys_refcount_enter++; + } + mutex_unlock(&syscall_trace_lock); + return ret; +} + +void unreg_event_syscall_enter(void *ptr) +{ + int num; + char *name; + + name = (char *)ptr; + num = syscall_name_to_nr(name); + if (num < 0 || num >= NR_syscalls) + return; + mutex_lock(&syscall_trace_lock); + sys_refcount_enter--; + clear_bit(num, enabled_enter_syscalls); + if (!sys_refcount_enter) + unregister_trace_sys_enter(ftrace_syscall_enter); + mutex_unlock(&syscall_trace_lock); +} + +int reg_event_syscall_exit(void *ptr) +{ + int ret = 0; + int num; + char *name; + + name = (char *)ptr; + num = syscall_name_to_nr(name); + if (num < 0 || num >= NR_syscalls) + return -ENOSYS; + mutex_lock(&syscall_trace_lock); + if (!sys_refcount_exit) + ret = register_trace_sys_exit(ftrace_syscall_exit); + if (ret) { + pr_info("event trace: Could not activate" + "syscall exit trace point"); + } else { + set_bit(num, enabled_exit_syscalls); + sys_refcount_exit++; + } + mutex_unlock(&syscall_trace_lock); + return ret; +} + +void unreg_event_syscall_exit(void *ptr) +{ + int num; + char *name; + + name = (char *)ptr; + num = syscall_name_to_nr(name); + if (num < 0 || num >= NR_syscalls) + return; + mutex_lock(&syscall_trace_lock); + sys_refcount_exit--; + clear_bit(num, enabled_exit_syscalls); + if (!sys_refcount_exit) + unregister_trace_sys_exit(ftrace_syscall_exit); + mutex_unlock(&syscall_trace_lock); +} + +struct trace_event event_syscall_enter = { + .trace = print_syscall_enter, +}; + +struct trace_event event_syscall_exit = { + .trace = print_syscall_exit, +}; + +#ifdef CONFIG_EVENT_PROFILE + +static DECLARE_BITMAP(enabled_prof_enter_syscalls, NR_syscalls); +static DECLARE_BITMAP(enabled_prof_exit_syscalls, NR_syscalls); +static int sys_prof_refcount_enter; +static int sys_prof_refcount_exit; + +static void prof_syscall_enter(struct pt_regs *regs, long id) +{ + struct syscall_trace_enter *rec; + struct syscall_metadata *sys_data; + int syscall_nr; + int size; + + syscall_nr = syscall_get_nr(current, regs); + if (!test_bit(syscall_nr, enabled_prof_enter_syscalls)) + return; + + sys_data = syscall_nr_to_meta(syscall_nr); + if (!sys_data) + return; + + /* get the size after alignment with the u32 buffer size field */ + size = sizeof(unsigned long) * sys_data->nb_args + sizeof(*rec); + size = ALIGN(size + sizeof(u32), sizeof(u64)); + size -= sizeof(u32); + + do { + char raw_data[size]; + + /* zero the dead bytes from align to not leak stack to user */ + *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; + + rec = (struct syscall_trace_enter *) raw_data; + tracing_generic_entry_update(&rec->ent, 0, 0); + rec->ent.type = sys_data->enter_id; + rec->nr = syscall_nr; + syscall_get_arguments(current, regs, 0, sys_data->nb_args, + (unsigned long *)&rec->args); + perf_tpcounter_event(sys_data->enter_id, 0, 1, rec, size); + } while(0); +} + +int reg_prof_syscall_enter(char *name) +{ + int ret = 0; + int num; + + num = syscall_name_to_nr(name); + if (num < 0 || num >= NR_syscalls) + return -ENOSYS; + + mutex_lock(&syscall_trace_lock); + if (!sys_prof_refcount_enter) + ret = register_trace_sys_enter(prof_syscall_enter); + if (ret) { + pr_info("event trace: Could not activate" + "syscall entry trace point"); + } else { + set_bit(num, enabled_prof_enter_syscalls); + sys_prof_refcount_enter++; + } + mutex_unlock(&syscall_trace_lock); + return ret; +} + +void unreg_prof_syscall_enter(char *name) +{ + int num; + + num = syscall_name_to_nr(name); + if (num < 0 || num >= NR_syscalls) + return; + + mutex_lock(&syscall_trace_lock); + sys_prof_refcount_enter--; + clear_bit(num, enabled_prof_enter_syscalls); + if (!sys_prof_refcount_enter) + unregister_trace_sys_enter(prof_syscall_enter); + mutex_unlock(&syscall_trace_lock); +} + +static void prof_syscall_exit(struct pt_regs *regs, long ret) +{ + struct syscall_metadata *sys_data; + struct syscall_trace_exit rec; + int syscall_nr; + + syscall_nr = syscall_get_nr(current, regs); + if (!test_bit(syscall_nr, enabled_prof_exit_syscalls)) + return; + + sys_data = syscall_nr_to_meta(syscall_nr); + if (!sys_data) + return; + + tracing_generic_entry_update(&rec.ent, 0, 0); + rec.ent.type = sys_data->exit_id; + rec.nr = syscall_nr; + rec.ret = syscall_get_return_value(current, regs); + + perf_tpcounter_event(sys_data->exit_id, 0, 1, &rec, sizeof(rec)); +} + +int reg_prof_syscall_exit(char *name) +{ + int ret = 0; + int num; + + num = syscall_name_to_nr(name); + if (num < 0 || num >= NR_syscalls) + return -ENOSYS; + + mutex_lock(&syscall_trace_lock); + if (!sys_prof_refcount_exit) + ret = register_trace_sys_exit(prof_syscall_exit); + if (ret) { + pr_info("event trace: Could not activate" + "syscall entry trace point"); + } else { + set_bit(num, enabled_prof_exit_syscalls); + sys_prof_refcount_exit++; + } + mutex_unlock(&syscall_trace_lock); + return ret; +} + +void unreg_prof_syscall_exit(char *name) +{ + int num; + + num = syscall_name_to_nr(name); + if (num < 0 || num >= NR_syscalls) + return; + + mutex_lock(&syscall_trace_lock); + sys_prof_refcount_exit--; + clear_bit(num, enabled_prof_exit_syscalls); + if (!sys_prof_refcount_exit) + unregister_trace_sys_exit(prof_syscall_exit); + mutex_unlock(&syscall_trace_lock); +} + +#endif + + diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c index eaca5ad803f..f6693969287 100644 --- a/kernel/trace/trace_sysprof.c +++ b/kernel/trace/trace_sysprof.c @@ -88,7 +88,7 @@ static void backtrace_address(void *data, unsigned long addr, int reliable) } } -const static struct stacktrace_ops backtrace_ops = { +static const struct stacktrace_ops backtrace_ops = { .warning = backtrace_warning, .warning_symbol = backtrace_warning_symbol, .stack = backtrace_stack, @@ -203,7 +203,8 @@ static void start_stack_timer(void *unused) hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); hrtimer->function = stack_trace_timer_fn; - hrtimer_start(hrtimer, ns_to_ktime(sample_period), HRTIMER_MODE_REL); + hrtimer_start(hrtimer, ns_to_ktime(sample_period), + HRTIMER_MODE_REL_PINNED); } static void start_stack_timers(void) @@ -226,15 +227,6 @@ static void stop_stack_timers(void) stop_stack_timer(cpu); } -static void start_stack_trace(struct trace_array *tr) -{ - mutex_lock(&sample_timer_lock); - tracing_reset_online_cpus(tr); - start_stack_timers(); - tracer_enabled = 1; - mutex_unlock(&sample_timer_lock); -} - static void stop_stack_trace(struct trace_array *tr) { mutex_lock(&sample_timer_lock); @@ -247,12 +239,18 @@ static int stack_trace_init(struct trace_array *tr) { sysprof_trace = tr; - start_stack_trace(tr); + tracing_start_cmdline_record(); + + mutex_lock(&sample_timer_lock); + start_stack_timers(); + tracer_enabled = 1; + mutex_unlock(&sample_timer_lock); return 0; } static void stack_trace_reset(struct trace_array *tr) { + tracing_stop_cmdline_record(); stop_stack_trace(tr); } @@ -317,18 +315,14 @@ sysprof_sample_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations sysprof_sample_fops = { +static const struct file_operations sysprof_sample_fops = { .read = sysprof_sample_read, .write = sysprof_sample_write, }; void init_tracer_sysprof_debugfs(struct dentry *d_tracer) { - struct dentry *entry; - entry = debugfs_create_file("sysprof_sample_period", 0644, + trace_create_file("sysprof_sample_period", 0644, d_tracer, NULL, &sysprof_sample_fops); - if (entry) - return; - pr_warning("Could not create debugfs 'dyn_ftrace_total_info' entry\n"); } diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c new file mode 100644 index 00000000000..40cafb07dff --- /dev/null +++ b/kernel/trace/trace_workqueue.c @@ -0,0 +1,295 @@ +/* + * Workqueue statistical tracer. + * + * Copyright (C) 2008 Frederic Weisbecker <fweisbec@gmail.com> + * + */ + + +#include <trace/events/workqueue.h> +#include <linux/list.h> +#include <linux/percpu.h> +#include <linux/kref.h> +#include "trace_stat.h" +#include "trace.h" + + +/* A cpu workqueue thread */ +struct cpu_workqueue_stats { + struct list_head list; + struct kref kref; + int cpu; + pid_t pid; +/* Can be inserted from interrupt or user context, need to be atomic */ + atomic_t inserted; +/* + * Don't need to be atomic, works are serialized in a single workqueue thread + * on a single CPU. + */ + unsigned int executed; +}; + +/* List of workqueue threads on one cpu */ +struct workqueue_global_stats { + struct list_head list; + spinlock_t lock; +}; + +/* Don't need a global lock because allocated before the workqueues, and + * never freed. + */ +static DEFINE_PER_CPU(struct workqueue_global_stats, all_workqueue_stat); +#define workqueue_cpu_stat(cpu) (&per_cpu(all_workqueue_stat, cpu)) + +static void cpu_workqueue_stat_free(struct kref *kref) +{ + kfree(container_of(kref, struct cpu_workqueue_stats, kref)); +} + +/* Insertion of a work */ +static void +probe_workqueue_insertion(struct task_struct *wq_thread, + struct work_struct *work) +{ + int cpu = cpumask_first(&wq_thread->cpus_allowed); + struct cpu_workqueue_stats *node; + unsigned long flags; + + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + list_for_each_entry(node, &workqueue_cpu_stat(cpu)->list, list) { + if (node->pid == wq_thread->pid) { + atomic_inc(&node->inserted); + goto found; + } + } + pr_debug("trace_workqueue: entry not found\n"); +found: + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); +} + +/* Execution of a work */ +static void +probe_workqueue_execution(struct task_struct *wq_thread, + struct work_struct *work) +{ + int cpu = cpumask_first(&wq_thread->cpus_allowed); + struct cpu_workqueue_stats *node; + unsigned long flags; + + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + list_for_each_entry(node, &workqueue_cpu_stat(cpu)->list, list) { + if (node->pid == wq_thread->pid) { + node->executed++; + goto found; + } + } + pr_debug("trace_workqueue: entry not found\n"); +found: + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); +} + +/* Creation of a cpu workqueue thread */ +static void probe_workqueue_creation(struct task_struct *wq_thread, int cpu) +{ + struct cpu_workqueue_stats *cws; + unsigned long flags; + + WARN_ON(cpu < 0); + + /* Workqueues are sometimes created in atomic context */ + cws = kzalloc(sizeof(struct cpu_workqueue_stats), GFP_ATOMIC); + if (!cws) { + pr_warning("trace_workqueue: not enough memory\n"); + return; + } + INIT_LIST_HEAD(&cws->list); + kref_init(&cws->kref); + cws->cpu = cpu; + cws->pid = wq_thread->pid; + + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + list_add_tail(&cws->list, &workqueue_cpu_stat(cpu)->list); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); +} + +/* Destruction of a cpu workqueue thread */ +static void probe_workqueue_destruction(struct task_struct *wq_thread) +{ + /* Workqueue only execute on one cpu */ + int cpu = cpumask_first(&wq_thread->cpus_allowed); + struct cpu_workqueue_stats *node, *next; + unsigned long flags; + + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + list_for_each_entry_safe(node, next, &workqueue_cpu_stat(cpu)->list, + list) { + if (node->pid == wq_thread->pid) { + list_del(&node->list); + kref_put(&node->kref, cpu_workqueue_stat_free); + goto found; + } + } + + pr_debug("trace_workqueue: don't find workqueue to destroy\n"); +found: + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); + +} + +static struct cpu_workqueue_stats *workqueue_stat_start_cpu(int cpu) +{ + unsigned long flags; + struct cpu_workqueue_stats *ret = NULL; + + + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + + if (!list_empty(&workqueue_cpu_stat(cpu)->list)) { + ret = list_entry(workqueue_cpu_stat(cpu)->list.next, + struct cpu_workqueue_stats, list); + kref_get(&ret->kref); + } + + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); + + return ret; +} + +static void *workqueue_stat_start(struct tracer_stat *trace) +{ + int cpu; + void *ret = NULL; + + for_each_possible_cpu(cpu) { + ret = workqueue_stat_start_cpu(cpu); + if (ret) + return ret; + } + return NULL; +} + +static void *workqueue_stat_next(void *prev, int idx) +{ + struct cpu_workqueue_stats *prev_cws = prev; + struct cpu_workqueue_stats *ret; + int cpu = prev_cws->cpu; + unsigned long flags; + + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + if (list_is_last(&prev_cws->list, &workqueue_cpu_stat(cpu)->list)) { + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); + do { + cpu = cpumask_next(cpu, cpu_possible_mask); + if (cpu >= nr_cpu_ids) + return NULL; + } while (!(ret = workqueue_stat_start_cpu(cpu))); + return ret; + } else { + ret = list_entry(prev_cws->list.next, + struct cpu_workqueue_stats, list); + kref_get(&ret->kref); + } + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); + + return ret; +} + +static int workqueue_stat_show(struct seq_file *s, void *p) +{ + struct cpu_workqueue_stats *cws = p; + struct pid *pid; + struct task_struct *tsk; + + pid = find_get_pid(cws->pid); + if (pid) { + tsk = get_pid_task(pid, PIDTYPE_PID); + if (tsk) { + seq_printf(s, "%3d %6d %6u %s\n", cws->cpu, + atomic_read(&cws->inserted), cws->executed, + tsk->comm); + put_task_struct(tsk); + } + put_pid(pid); + } + + return 0; +} + +static void workqueue_stat_release(void *stat) +{ + struct cpu_workqueue_stats *node = stat; + + kref_put(&node->kref, cpu_workqueue_stat_free); +} + +static int workqueue_stat_headers(struct seq_file *s) +{ + seq_printf(s, "# CPU INSERTED EXECUTED NAME\n"); + seq_printf(s, "# | | | |\n"); + return 0; +} + +struct tracer_stat workqueue_stats __read_mostly = { + .name = "workqueues", + .stat_start = workqueue_stat_start, + .stat_next = workqueue_stat_next, + .stat_show = workqueue_stat_show, + .stat_release = workqueue_stat_release, + .stat_headers = workqueue_stat_headers +}; + + +int __init stat_workqueue_init(void) +{ + if (register_stat_tracer(&workqueue_stats)) { + pr_warning("Unable to register workqueue stat tracer\n"); + return 1; + } + + return 0; +} +fs_initcall(stat_workqueue_init); + +/* + * Workqueues are created very early, just after pre-smp initcalls. + * So we must register our tracepoints at this stage. + */ +int __init trace_workqueue_early_init(void) +{ + int ret, cpu; + + ret = register_trace_workqueue_insertion(probe_workqueue_insertion); + if (ret) + goto out; + + ret = register_trace_workqueue_execution(probe_workqueue_execution); + if (ret) + goto no_insertion; + + ret = register_trace_workqueue_creation(probe_workqueue_creation); + if (ret) + goto no_execution; + + ret = register_trace_workqueue_destruction(probe_workqueue_destruction); + if (ret) + goto no_creation; + + for_each_possible_cpu(cpu) { + spin_lock_init(&workqueue_cpu_stat(cpu)->lock); + INIT_LIST_HEAD(&workqueue_cpu_stat(cpu)->list); + } + + return 0; + +no_creation: + unregister_trace_workqueue_creation(probe_workqueue_creation); +no_execution: + unregister_trace_workqueue_execution(probe_workqueue_execution); +no_insertion: + unregister_trace_workqueue_insertion(probe_workqueue_insertion); +out: + pr_warning("trace_workqueue: unable to trace workqueues\n"); + + return 1; +} +early_initcall(trace_workqueue_early_init); diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c index 79602740bbb..9489a0a9b1b 100644 --- a/kernel/tracepoint.c +++ b/kernel/tracepoint.c @@ -24,6 +24,7 @@ #include <linux/tracepoint.h> #include <linux/err.h> #include <linux/slab.h> +#include <linux/sched.h> extern struct tracepoint __start___tracepoints[]; extern struct tracepoint __stop___tracepoints[]; @@ -242,6 +243,11 @@ static void set_tracepoint(struct tracepoint_entry **entry, { WARN_ON(strcmp((*entry)->name, elem->name) != 0); + if (elem->regfunc && !elem->state && active) + elem->regfunc(); + else if (elem->unregfunc && elem->state && !active) + elem->unregfunc(); + /* * rcu_assign_pointer has a smp_wmb() which makes sure that the new * probe callbacks array is consistent before setting a pointer to it. @@ -261,6 +267,9 @@ static void set_tracepoint(struct tracepoint_entry **entry, */ static void disable_tracepoint(struct tracepoint *elem) { + if (elem->unregfunc && elem->state) + elem->unregfunc(); + elem->state = 0; rcu_assign_pointer(elem->funcs, NULL); } @@ -272,12 +281,15 @@ static void disable_tracepoint(struct tracepoint *elem) * * Updates the probe callback corresponding to a range of tracepoints. */ -void tracepoint_update_probe_range(struct tracepoint *begin, - struct tracepoint *end) +void +tracepoint_update_probe_range(struct tracepoint *begin, struct tracepoint *end) { struct tracepoint *iter; struct tracepoint_entry *mark_entry; + if (!begin) + return; + mutex_lock(&tracepoints_mutex); for (iter = begin; iter < end; iter++) { mark_entry = get_tracepoint(iter->name); @@ -551,9 +563,6 @@ int tracepoint_module_notify(struct notifier_block *self, switch (val) { case MODULE_STATE_COMING: - tracepoint_update_probe_range(mod->tracepoints, - mod->tracepoints + mod->num_tracepoints); - break; case MODULE_STATE_GOING: tracepoint_update_probe_range(mod->tracepoints, mod->tracepoints + mod->num_tracepoints); @@ -574,3 +583,41 @@ static int init_tracepoints(void) __initcall(init_tracepoints); #endif /* CONFIG_MODULES */ + +#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS + +/* NB: reg/unreg are called while guarded with the tracepoints_mutex */ +static int sys_tracepoint_refcount; + +void syscall_regfunc(void) +{ + unsigned long flags; + struct task_struct *g, *t; + + if (!sys_tracepoint_refcount) { + read_lock_irqsave(&tasklist_lock, flags); + do_each_thread(g, t) { + /* Skip kernel threads. */ + if (t->mm) + set_tsk_thread_flag(t, TIF_SYSCALL_TRACEPOINT); + } while_each_thread(g, t); + read_unlock_irqrestore(&tasklist_lock, flags); + } + sys_tracepoint_refcount++; +} + +void syscall_unregfunc(void) +{ + unsigned long flags; + struct task_struct *g, *t; + + sys_tracepoint_refcount--; + if (!sys_tracepoint_refcount) { + read_lock_irqsave(&tasklist_lock, flags); + do_each_thread(g, t) { + clear_tsk_thread_flag(t, TIF_SYSCALL_TRACEPOINT); + } while_each_thread(g, t); + read_unlock_irqrestore(&tasklist_lock, flags); + } +} +#endif diff --git a/kernel/tsacct.c b/kernel/tsacct.c index 43f891b05a4..00d59d048ed 100644 --- a/kernel/tsacct.c +++ b/kernel/tsacct.c @@ -122,8 +122,10 @@ void acct_update_integrals(struct task_struct *tsk) if (likely(tsk->mm)) { cputime_t time, dtime; struct timeval value; + unsigned long flags; u64 delta; + local_irq_save(flags); time = tsk->stime + tsk->utime; dtime = cputime_sub(time, tsk->acct_timexpd); jiffies_to_timeval(cputime_to_jiffies(dtime), &value); @@ -131,10 +133,12 @@ void acct_update_integrals(struct task_struct *tsk) delta = delta * USEC_PER_SEC + value.tv_usec; if (delta == 0) - return; + goto out; tsk->acct_timexpd = time; tsk->acct_rss_mem1 += delta * get_mm_rss(tsk->mm); tsk->acct_vm_mem1 += delta * tsk->mm->total_vm; + out: + local_irq_restore(flags); } } diff --git a/kernel/user.c b/kernel/user.c index 477b6660f44..2c000e7132a 100644 --- a/kernel/user.c +++ b/kernel/user.c @@ -20,7 +20,7 @@ struct user_namespace init_user_ns = { .kref = { - .refcount = ATOMIC_INIT(1), + .refcount = ATOMIC_INIT(2), }, .creator = &root_user, }; @@ -72,21 +72,7 @@ static void uid_hash_insert(struct user_struct *up, struct hlist_head *hashent) static void uid_hash_remove(struct user_struct *up) { hlist_del_init(&up->uidhash_node); -} - -static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) -{ - struct user_struct *user; - struct hlist_node *h; - - hlist_for_each_entry(user, h, hashent, uidhash_node) { - if (user->uid == uid) { - atomic_inc(&user->__count); - return user; - } - } - - return NULL; + put_user_ns(up->user_ns); } #ifdef CONFIG_USER_SCHED @@ -118,6 +104,23 @@ static int sched_create_user(struct user_struct *up) { return 0; } #if defined(CONFIG_USER_SCHED) && defined(CONFIG_SYSFS) +static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) +{ + struct user_struct *user; + struct hlist_node *h; + + hlist_for_each_entry(user, h, hashent, uidhash_node) { + if (user->uid == uid) { + /* possibly resurrect an "almost deleted" object */ + if (atomic_inc_return(&user->__count) == 1) + cancel_delayed_work(&user->work); + return user; + } + } + + return NULL; +} + static struct kset *uids_kset; /* represents the /sys/kernel/uids/ directory */ static DEFINE_MUTEX(uids_mutex); @@ -282,38 +285,35 @@ int __init uids_sysfs_init(void) return uids_user_create(&root_user); } -/* work function to remove sysfs directory for a user and free up +/* delayed work function to remove sysfs directory for a user and free up * corresponding structures. */ -static void remove_user_sysfs_dir(struct work_struct *w) +static void cleanup_user_struct(struct work_struct *w) { - struct user_struct *up = container_of(w, struct user_struct, work); + struct user_struct *up = container_of(w, struct user_struct, work.work); unsigned long flags; int remove_user = 0; - if (up->user_ns != &init_user_ns) - return; /* Make uid_hash_remove() + sysfs_remove_file() + kobject_del() * atomic. */ uids_mutex_lock(); - local_irq_save(flags); - - if (atomic_dec_and_lock(&up->__count, &uidhash_lock)) { + spin_lock_irqsave(&uidhash_lock, flags); + if (atomic_read(&up->__count) == 0) { uid_hash_remove(up); remove_user = 1; - spin_unlock_irqrestore(&uidhash_lock, flags); - } else { - local_irq_restore(flags); } + spin_unlock_irqrestore(&uidhash_lock, flags); if (!remove_user) goto done; - kobject_uevent(&up->kobj, KOBJ_REMOVE); - kobject_del(&up->kobj); - kobject_put(&up->kobj); + if (up->user_ns == &init_user_ns) { + kobject_uevent(&up->kobj, KOBJ_REMOVE); + kobject_del(&up->kobj); + kobject_put(&up->kobj); + } sched_destroy_user(up); key_put(up->uid_keyring); @@ -330,17 +330,28 @@ done: */ static void free_user(struct user_struct *up, unsigned long flags) { - /* restore back the count */ - atomic_inc(&up->__count); spin_unlock_irqrestore(&uidhash_lock, flags); - - put_user_ns(up->user_ns); - INIT_WORK(&up->work, remove_user_sysfs_dir); - schedule_work(&up->work); + INIT_DELAYED_WORK(&up->work, cleanup_user_struct); + schedule_delayed_work(&up->work, msecs_to_jiffies(1000)); } #else /* CONFIG_USER_SCHED && CONFIG_SYSFS */ +static struct user_struct *uid_hash_find(uid_t uid, struct hlist_head *hashent) +{ + struct user_struct *user; + struct hlist_node *h; + + hlist_for_each_entry(user, h, hashent, uidhash_node) { + if (user->uid == uid) { + atomic_inc(&user->__count); + return user; + } + } + + return NULL; +} + int uids_sysfs_init(void) { return 0; } static inline int uids_user_create(struct user_struct *up) { return 0; } static inline void uids_mutex_lock(void) { } @@ -357,12 +368,29 @@ static void free_user(struct user_struct *up, unsigned long flags) sched_destroy_user(up); key_put(up->uid_keyring); key_put(up->session_keyring); - put_user_ns(up->user_ns); kmem_cache_free(uid_cachep, up); } #endif +#if defined(CONFIG_RT_GROUP_SCHED) && defined(CONFIG_USER_SCHED) +/* + * We need to check if a setuid can take place. This function should be called + * before successfully completing the setuid. + */ +int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) +{ + + return sched_rt_can_attach(up->tg, tsk); + +} +#else +int task_can_switch_user(struct user_struct *up, struct task_struct *tsk) +{ + return 1; +} +#endif + /* * Locate the user_struct for the passed UID. If found, take a ref on it. The * caller must undo that ref with free_uid(). diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index 79084311ee5..076c7c8215b 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -60,12 +60,25 @@ int create_user_ns(struct cred *new) return 0; } -void free_user_ns(struct kref *kref) +/* + * Deferred destructor for a user namespace. This is required because + * free_user_ns() may be called with uidhash_lock held, but we need to call + * back to free_uid() which will want to take the lock again. + */ +static void free_user_ns_work(struct work_struct *work) { - struct user_namespace *ns; - - ns = container_of(kref, struct user_namespace, kref); + struct user_namespace *ns = + container_of(work, struct user_namespace, destroyer); free_uid(ns->creator); kfree(ns); } + +void free_user_ns(struct kref *kref) +{ + struct user_namespace *ns = + container_of(kref, struct user_namespace, kref); + + INIT_WORK(&ns->destroyer, free_user_ns_work); + schedule_work(&ns->destroyer); +} EXPORT_SYMBOL(free_user_ns); diff --git a/kernel/utsname.c b/kernel/utsname.c index 815237a55af..8a82b4b8ea5 100644 --- a/kernel/utsname.c +++ b/kernel/utsname.c @@ -15,6 +15,16 @@ #include <linux/err.h> #include <linux/slab.h> +static struct uts_namespace *create_uts_ns(void) +{ + struct uts_namespace *uts_ns; + + uts_ns = kmalloc(sizeof(struct uts_namespace), GFP_KERNEL); + if (uts_ns) + kref_init(&uts_ns->kref); + return uts_ns; +} + /* * Clone a new ns copying an original utsname, setting refcount to 1 * @old_ns: namespace to clone @@ -24,14 +34,13 @@ static struct uts_namespace *clone_uts_ns(struct uts_namespace *old_ns) { struct uts_namespace *ns; - ns = kmalloc(sizeof(struct uts_namespace), GFP_KERNEL); + ns = create_uts_ns(); if (!ns) return ERR_PTR(-ENOMEM); down_read(&uts_sem); memcpy(&ns->name, &old_ns->name, sizeof(ns->name)); up_read(&uts_sem); - kref_init(&ns->kref); return ns; } diff --git a/kernel/utsname_sysctl.c b/kernel/utsname_sysctl.c index 3b34b354593..92359cc747a 100644 --- a/kernel/utsname_sysctl.c +++ b/kernel/utsname_sysctl.c @@ -37,7 +37,7 @@ static void put_uts(ctl_table *table, int write, void *which) up_write(&uts_sem); } -#ifdef CONFIG_PROC_FS +#ifdef CONFIG_PROC_SYSCTL /* * Special case of dostring for the UTS structure. This has locks * to observe. Should this be in kernel/sys.c ???? diff --git a/kernel/wait.c b/kernel/wait.c index cd87131f2fc..c4bd3d825f3 100644 --- a/kernel/wait.c +++ b/kernel/wait.c @@ -10,13 +10,14 @@ #include <linux/wait.h> #include <linux/hash.h> -void init_waitqueue_head(wait_queue_head_t *q) +void __init_waitqueue_head(wait_queue_head_t *q, struct lock_class_key *key) { spin_lock_init(&q->lock); + lockdep_set_class(&q->lock, key); INIT_LIST_HEAD(&q->task_list); } -EXPORT_SYMBOL(init_waitqueue_head); +EXPORT_SYMBOL(__init_waitqueue_head); void add_wait_queue(wait_queue_head_t *q, wait_queue_t *wait) { @@ -91,6 +92,15 @@ prepare_to_wait_exclusive(wait_queue_head_t *q, wait_queue_t *wait, int state) } EXPORT_SYMBOL(prepare_to_wait_exclusive); +/* + * finish_wait - clean up after waiting in a queue + * @q: waitqueue waited on + * @wait: wait descriptor + * + * Sets current thread back to running state and removes + * the wait descriptor from the given waitqueue if still + * queued. + */ void finish_wait(wait_queue_head_t *q, wait_queue_t *wait) { unsigned long flags; @@ -117,6 +127,39 @@ void finish_wait(wait_queue_head_t *q, wait_queue_t *wait) } EXPORT_SYMBOL(finish_wait); +/* + * abort_exclusive_wait - abort exclusive waiting in a queue + * @q: waitqueue waited on + * @wait: wait descriptor + * @state: runstate of the waiter to be woken + * @key: key to identify a wait bit queue or %NULL + * + * Sets current thread back to running state and removes + * the wait descriptor from the given waitqueue if still + * queued. + * + * Wakes up the next waiter if the caller is concurrently + * woken up through the queue. + * + * This prevents waiter starvation where an exclusive waiter + * aborts and is woken up concurrently and noone wakes up + * the next waiter. + */ +void abort_exclusive_wait(wait_queue_head_t *q, wait_queue_t *wait, + unsigned int mode, void *key) +{ + unsigned long flags; + + __set_current_state(TASK_RUNNING); + spin_lock_irqsave(&q->lock, flags); + if (!list_empty(&wait->task_list)) + list_del_init(&wait->task_list); + else if (waitqueue_active(q)) + __wake_up_locked_key(q, mode, key); + spin_unlock_irqrestore(&q->lock, flags); +} +EXPORT_SYMBOL(abort_exclusive_wait); + int autoremove_wake_function(wait_queue_t *wait, unsigned mode, int sync, void *key) { int ret = default_wake_function(wait, mode, sync, key); @@ -177,17 +220,20 @@ int __sched __wait_on_bit_lock(wait_queue_head_t *wq, struct wait_bit_queue *q, int (*action)(void *), unsigned mode) { - int ret = 0; - do { + int ret; + prepare_to_wait_exclusive(wq, &q->wait, mode); - if (test_bit(q->key.bit_nr, q->key.flags)) { - if ((ret = (*action)(q->key.flags))) - break; - } + if (!test_bit(q->key.bit_nr, q->key.flags)) + continue; + ret = action(q->key.flags); + if (!ret) + continue; + abort_exclusive_wait(wq, &q->wait, mode, &q->key); + return ret; } while (test_and_set_bit(q->key.bit_nr, q->key.flags)); finish_wait(wq, &q->wait); - return ret; + return 0; } EXPORT_SYMBOL(__wait_on_bit_lock); diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 2f445833ae3..addfe2df93b 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -33,6 +33,8 @@ #include <linux/kallsyms.h> #include <linux/debug_locks.h> #include <linux/lockdep.h> +#define CREATE_TRACE_POINTS +#include <trace/events/workqueue.h> /* * The per-CPU workqueue (if single thread, we always use the first @@ -48,8 +50,6 @@ struct cpu_workqueue_struct { struct workqueue_struct *wq; struct task_struct *thread; - - int run_depth; /* Detect run_workqueue() recursion depth */ } ____cacheline_aligned; /* @@ -128,6 +128,8 @@ struct cpu_workqueue_struct *get_wq_data(struct work_struct *work) static void insert_work(struct cpu_workqueue_struct *cwq, struct work_struct *work, struct list_head *head) { + trace_workqueue_insertion(cwq->thread, work); + set_wq_data(work, cwq); /* * Ensure that we get the right work->data if we see the @@ -262,13 +264,6 @@ EXPORT_SYMBOL_GPL(queue_delayed_work_on); static void run_workqueue(struct cpu_workqueue_struct *cwq) { spin_lock_irq(&cwq->lock); - cwq->run_depth++; - if (cwq->run_depth > 3) { - /* morton gets to eat his hat */ - printk("%s: recursion depth exceeded: %d\n", - __func__, cwq->run_depth); - dump_stack(); - } while (!list_empty(&cwq->worklist)) { struct work_struct *work = list_entry(cwq->worklist.next, struct work_struct, entry); @@ -284,7 +279,7 @@ static void run_workqueue(struct cpu_workqueue_struct *cwq) */ struct lockdep_map lockdep_map = work->lockdep_map; #endif - + trace_workqueue_execution(cwq->thread, work); cwq->current_work = work; list_del_init(cwq->worklist.next); spin_unlock_irq(&cwq->lock); @@ -311,7 +306,6 @@ static void run_workqueue(struct cpu_workqueue_struct *cwq) spin_lock_irq(&cwq->lock); cwq->current_work = NULL; } - cwq->run_depth--; spin_unlock_irq(&cwq->lock); } @@ -323,8 +317,6 @@ static int worker_thread(void *__cwq) if (cwq->wq->freezeable) set_freezable(); - set_user_nice(current, -5); - for (;;) { prepare_to_wait(&cwq->more_work, &wait, TASK_INTERRUPTIBLE); if (!freezing(current) && @@ -368,29 +360,20 @@ static void insert_wq_barrier(struct cpu_workqueue_struct *cwq, static int flush_cpu_workqueue(struct cpu_workqueue_struct *cwq) { - int active; - - if (cwq->thread == current) { - /* - * Probably keventd trying to flush its own queue. So simply run - * it by hand rather than deadlocking. - */ - run_workqueue(cwq); - active = 1; - } else { - struct wq_barrier barr; + int active = 0; + struct wq_barrier barr; - active = 0; - spin_lock_irq(&cwq->lock); - if (!list_empty(&cwq->worklist) || cwq->current_work != NULL) { - insert_wq_barrier(cwq, &barr, &cwq->worklist); - active = 1; - } - spin_unlock_irq(&cwq->lock); + WARN_ON(cwq->thread == current); - if (active) - wait_for_completion(&barr.done); + spin_lock_irq(&cwq->lock); + if (!list_empty(&cwq->worklist) || cwq->current_work != NULL) { + insert_wq_barrier(cwq, &barr, &cwq->worklist); + active = 1; } + spin_unlock_irq(&cwq->lock); + + if (active) + wait_for_completion(&barr.done); return active; } @@ -416,7 +399,7 @@ void flush_workqueue(struct workqueue_struct *wq) might_sleep(); lock_map_acquire(&wq->lockdep_map); lock_map_release(&wq->lockdep_map); - for_each_cpu_mask_nr(cpu, *cpu_map) + for_each_cpu(cpu, cpu_map) flush_cpu_workqueue(per_cpu_ptr(wq->cpu_wq, cpu)); } EXPORT_SYMBOL_GPL(flush_workqueue); @@ -547,7 +530,7 @@ static void wait_on_work(struct work_struct *work) wq = cwq->wq; cpu_map = wq_cpu_map(wq); - for_each_cpu_mask_nr(cpu, *cpu_map) + for_each_cpu(cpu, cpu_map) wait_on_cpu_work(per_cpu_ptr(wq->cpu_wq, cpu), work); } @@ -615,7 +598,12 @@ static struct workqueue_struct *keventd_wq __read_mostly; * schedule_work - put work task in global workqueue * @work: job to be done * - * This puts a job in the kernel-global workqueue. + * Returns zero if @work was already on the kernel-global workqueue and + * non-zero otherwise. + * + * This puts a job in the kernel-global workqueue if it was not already + * queued and leaves it in the same position on the kernel-global + * workqueue otherwise. */ int schedule_work(struct work_struct *work) { @@ -787,6 +775,8 @@ static int create_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu) sched_setscheduler_nocheck(p, SCHED_FIFO, ¶m); cwq->thread = p; + trace_workqueue_creation(cwq->thread, cpu); + return 0; } @@ -891,6 +881,7 @@ static void cleanup_workqueue_thread(struct cpu_workqueue_struct *cwq) * checks list_empty(), and a "normal" queue_work() can't use * a dead CPU. */ + trace_workqueue_destruction(cwq->thread); kthread_stop(cwq->thread); cwq->thread = NULL; } @@ -911,7 +902,7 @@ void destroy_workqueue(struct workqueue_struct *wq) list_del(&wq->list); spin_unlock(&workqueue_lock); - for_each_cpu_mask_nr(cpu, *cpu_map) + for_each_cpu(cpu, cpu_map) cleanup_workqueue_thread(per_cpu_ptr(wq->cpu_wq, cpu)); cpu_maps_update_done(); @@ -971,18 +962,20 @@ undo: } #ifdef CONFIG_SMP + struct work_for_cpu { - struct work_struct work; + struct completion completion; long (*fn)(void *); void *arg; long ret; }; -static void do_work_for_cpu(struct work_struct *w) +static int do_work_for_cpu(void *_wfc) { - struct work_for_cpu *wfc = container_of(w, struct work_for_cpu, work); - + struct work_for_cpu *wfc = _wfc; wfc->ret = wfc->fn(wfc->arg); + complete(&wfc->completion); + return 0; } /** @@ -991,25 +984,25 @@ static void do_work_for_cpu(struct work_struct *w) * @fn: the function to run * @arg: the function arg * - * This will return -EINVAL in the cpu is not online, or the return value - * of @fn otherwise. + * This will return the value @fn returns. + * It is up to the caller to ensure that the cpu doesn't go offline. + * The caller must not hold any locks which would prevent @fn from completing. */ long work_on_cpu(unsigned int cpu, long (*fn)(void *), void *arg) { - struct work_for_cpu wfc; - - INIT_WORK(&wfc.work, do_work_for_cpu); - wfc.fn = fn; - wfc.arg = arg; - get_online_cpus(); - if (unlikely(!cpu_online(cpu))) - wfc.ret = -EINVAL; - else { - schedule_work_on(cpu, &wfc.work); - flush_work(&wfc.work); - } - put_online_cpus(); - + struct task_struct *sub_thread; + struct work_for_cpu wfc = { + .completion = COMPLETION_INITIALIZER_ONSTACK(wfc.completion), + .fn = fn, + .arg = arg, + }; + + sub_thread = kthread_create(do_work_for_cpu, &wfc, "work_for_cpu"); + if (IS_ERR(sub_thread)) + return PTR_ERR(sub_thread); + kthread_bind(sub_thread, cpu); + wake_up_process(sub_thread); + wait_for_completion(&wfc.completion); return wfc.ret; } EXPORT_SYMBOL_GPL(work_on_cpu); |