diff options
Diffstat (limited to 'kernel/cpuset.c')
| -rw-r--r-- | kernel/cpuset.c | 427 |
1 files changed, 276 insertions, 151 deletions
diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 12815d3f1a0..72248d1b9e3 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -4,15 +4,14 @@ * Processor and Memory placement constraints for sets of tasks. * * Copyright (C) 2003 BULL SA. - * Copyright (C) 2004 Silicon Graphics, Inc. + * Copyright (C) 2004-2006 Silicon Graphics, Inc. * * Portions derived from Patrick Mochel's sysfs code. * sysfs is Copyright (c) 2001-3 Patrick Mochel - * Portions Copyright (c) 2004 Silicon Graphics, Inc. * - * 2003-10-10 Written by Simon Derr <simon.derr@bull.net> + * 2003-10-10 Written by Simon Derr. * 2003-10-22 Updates by Stephen Hemminger. - * 2004 May-July Rework by Paul Jackson <pj@sgi.com> + * 2004 May-July Rework by Paul Jackson. * * This file is subject to the terms and conditions of the GNU General Public * License. See the file COPYING in the main directory of the Linux @@ -53,7 +52,7 @@ #include <asm/uaccess.h> #include <asm/atomic.h> -#include <asm/semaphore.h> +#include <linux/mutex.h> #define CPUSET_SUPER_MAGIC 0x27e0eb @@ -108,37 +107,49 @@ typedef enum { CS_MEM_EXCLUSIVE, CS_MEMORY_MIGRATE, CS_REMOVED, - CS_NOTIFY_ON_RELEASE + CS_NOTIFY_ON_RELEASE, + CS_SPREAD_PAGE, + CS_SPREAD_SLAB, } cpuset_flagbits_t; /* convenient tests for these bits */ static inline int is_cpu_exclusive(const struct cpuset *cs) { - return !!test_bit(CS_CPU_EXCLUSIVE, &cs->flags); + return test_bit(CS_CPU_EXCLUSIVE, &cs->flags); } static inline int is_mem_exclusive(const struct cpuset *cs) { - return !!test_bit(CS_MEM_EXCLUSIVE, &cs->flags); + return test_bit(CS_MEM_EXCLUSIVE, &cs->flags); } static inline int is_removed(const struct cpuset *cs) { - return !!test_bit(CS_REMOVED, &cs->flags); + return test_bit(CS_REMOVED, &cs->flags); } static inline int notify_on_release(const struct cpuset *cs) { - return !!test_bit(CS_NOTIFY_ON_RELEASE, &cs->flags); + return test_bit(CS_NOTIFY_ON_RELEASE, &cs->flags); } static inline int is_memory_migrate(const struct cpuset *cs) { - return !!test_bit(CS_MEMORY_MIGRATE, &cs->flags); + return test_bit(CS_MEMORY_MIGRATE, &cs->flags); +} + +static inline int is_spread_page(const struct cpuset *cs) +{ + return test_bit(CS_SPREAD_PAGE, &cs->flags); +} + +static inline int is_spread_slab(const struct cpuset *cs) +{ + return test_bit(CS_SPREAD_SLAB, &cs->flags); } /* - * Increment this atomic integer everytime any cpuset changes its + * 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. @@ -152,8 +163,11 @@ static inline int is_memory_migrate(const struct cpuset *cs) * 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 cpuset_mems_generation is guarded by manage_mutex, + * there is no need to mark it atomic. */ -static atomic_t cpuset_mems_generation = ATOMIC_INIT(1); +static int cpuset_mems_generation; static struct cpuset top_cpuset = { .flags = ((1 << CS_CPU_EXCLUSIVE) | (1 << CS_MEM_EXCLUSIVE)), @@ -168,63 +182,57 @@ static struct vfsmount *cpuset_mount; static struct super_block *cpuset_sb; /* - * We have two global cpuset semaphores below. They can nest. - * It is ok to first take manage_sem, then nest callback_sem. We also + * We have two global cpuset mutexes below. They can nest. + * It is ok to first take manage_mutex, then nest callback_mutex. We also * require taking task_lock() when dereferencing a tasks cpuset pointer. * See "The task_lock() exception", at the end of this comment. * - * A task must hold both semaphores to modify cpusets. If a task - * holds manage_sem, then it blocks others wanting that semaphore, - * ensuring that it is the only task able to also acquire callback_sem + * A task must hold both mutexes to modify cpusets. If a task + * holds manage_mutex, then it blocks others wanting that mutex, + * ensuring that it is the only task able to also acquire callback_mutex * and be able to modify cpusets. It can perform various checks on * the cpuset structure first, knowing nothing will change. It can - * also allocate memory while just holding manage_sem. While it is + * also allocate memory while just holding manage_mutex. While it is * performing these checks, various callback routines can briefly - * acquire callback_sem to query cpusets. Once it is ready to make - * the changes, it takes callback_sem, blocking everyone else. + * acquire callback_mutex to query cpusets. Once it is ready to make + * the changes, it takes callback_mutex, blocking everyone else. * * Calls to the kernel memory allocator can not be made while holding - * callback_sem, as that would risk double tripping on callback_sem + * callback_mutex, as that would risk double tripping on callback_mutex * from one of the callbacks into the cpuset code from within * __alloc_pages(). * - * If a task is only holding callback_sem, then it has read-only + * 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. * * Any task can increment and decrement the count field without lock. - * So in general, code holding manage_sem or callback_sem can't rely + * So in general, code holding manage_mutex or callback_mutex can't rely * on the count field not changing. However, if the count goes to - * zero, then only attach_task(), which holds both semaphores, can + * zero, then only attach_task(), which holds both mutexes, can * increment it again. Because a count of zero means that no tasks * are currently attached, therefore there is no way a task attached * to that cpuset can fork (the other way to increment the count). - * So code holding manage_sem or callback_sem can safely assume that + * So code holding manage_mutex or callback_mutex can safely assume that * if the count is zero, it will stay zero. Similarly, if a task - * holds manage_sem or callback_sem on a cpuset with zero count, it + * holds manage_mutex or callback_mutex on a cpuset with zero count, it * knows that the cpuset won't be removed, as cpuset_rmdir() needs - * both of those semaphores. - * - * A possible optimization to improve parallelism would be to make - * callback_sem a R/W semaphore (rwsem), allowing the callback routines - * to proceed in parallel, with read access, until the holder of - * manage_sem needed to take this rwsem for exclusive write access - * and modify some cpusets. + * both of those mutexes. * * The cpuset_common_file_write handler for operations that modify - * the cpuset hierarchy holds manage_sem across the entire operation, + * the cpuset hierarchy holds manage_mutex across the entire operation, * single threading all such cpuset modifications across the system. * - * The cpuset_common_file_read() handlers only hold callback_sem across + * The cpuset_common_file_read() handlers only hold callback_mutex across * small pieces of code, such as when reading out possibly multi-word * cpumasks and nodemasks. * * The fork and exit callbacks cpuset_fork() and cpuset_exit(), don't - * (usually) take either semaphore. These are the two most performance + * (usually) take either mutex. These are the two most performance * critical pieces of code here. The exception occurs on cpuset_exit(), - * when a task in a notify_on_release cpuset exits. Then manage_sem + * when a task in a notify_on_release cpuset exits. Then manage_mutex * is taken, and if the cpuset count is zero, a usermode call made * to /sbin/cpuset_release_agent with the name of the cpuset (path * relative to the root of cpuset file system) as the argument. @@ -242,9 +250,9 @@ static struct super_block *cpuset_sb; * * The need for this exception arises from the action of attach_task(), * which overwrites one tasks cpuset pointer with another. It does - * so using both semaphores, however there are several performance + * so using both mutexes, however there are several performance * critical places that need to reference task->cpuset without the - * expense of grabbing a system global semaphore. Therefore except as + * expense of grabbing a system global mutex. Therefore except as * noted below, when dereferencing or, as in attach_task(), modifying * a tasks cpuset pointer we use task_lock(), which acts on a spinlock * (task->alloc_lock) already in the task_struct routinely used for @@ -256,8 +264,8 @@ static struct super_block *cpuset_sb; * the routine cpuset_update_task_memory_state(). */ -static DECLARE_MUTEX(manage_sem); -static DECLARE_MUTEX(callback_sem); +static DEFINE_MUTEX(manage_mutex); +static DEFINE_MUTEX(callback_mutex); /* * A couple of forward declarations required, due to cyclic reference loop: @@ -432,7 +440,7 @@ static inline struct cftype *__d_cft(struct dentry *dentry) } /* - * Call with manage_sem held. Writes path of cpuset into buf. + * Call with manage_mutex held. Writes path of cpuset into buf. * Returns 0 on success, -errno on error. */ @@ -484,11 +492,11 @@ static int cpuset_path(const struct cpuset *cs, char *buf, int buflen) * status of the /sbin/cpuset_release_agent task, so no sense holding * our caller up for that. * - * When we had only one cpuset semaphore, we had to call this + * When we had only one cpuset mutex, we had to call this * without holding it, to avoid deadlock when call_usermodehelper() * allocated memory. With two locks, we could now call this while - * holding manage_sem, but we still don't, so as to minimize - * the time manage_sem is held. + * holding manage_mutex, but we still don't, so as to minimize + * the time manage_mutex is held. */ static void cpuset_release_agent(const char *pathbuf) @@ -520,15 +528,15 @@ static void cpuset_release_agent(const char *pathbuf) * cs is notify_on_release() and now both the user count is zero and * the list of children is empty, prepare cpuset path in a kmalloc'd * buffer, to be returned via ppathbuf, so that the caller can invoke - * cpuset_release_agent() with it later on, once manage_sem is dropped. - * Call here with manage_sem held. + * cpuset_release_agent() with it later on, once manage_mutex is dropped. + * Call here with manage_mutex held. * * This check_for_release() routine is responsible for kmalloc'ing * pathbuf. The above cpuset_release_agent() is responsible for * kfree'ing pathbuf. The caller of these routines is responsible * for providing a pathbuf pointer, initialized to NULL, then - * calling check_for_release() with manage_sem held and the address - * of the pathbuf pointer, then dropping manage_sem, then calling + * calling check_for_release() with manage_mutex held and the address + * of the pathbuf pointer, then dropping manage_mutex, then calling * cpuset_release_agent() with pathbuf, as set by check_for_release(). */ @@ -559,7 +567,7 @@ static void check_for_release(struct cpuset *cs, char **ppathbuf) * One way or another, we guarantee to return some non-empty subset * of cpu_online_map. * - * Call with callback_sem held. + * Call with callback_mutex held. */ static void guarantee_online_cpus(const struct cpuset *cs, cpumask_t *pmask) @@ -583,7 +591,7 @@ static void guarantee_online_cpus(const struct cpuset *cs, cpumask_t *pmask) * One way or another, we guarantee to return some non-empty subset * of node_online_map. * - * Call with callback_sem held. + * Call with callback_mutex held. */ static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask) @@ -608,12 +616,10 @@ static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask) * current->cpuset if a task has its memory placement changed. * Do not call this routine if in_interrupt(). * - * Call without callback_sem or task_lock() held. May be called - * with or without manage_sem held. Doesn't need task_lock to guard - * against another task changing a non-NULL cpuset pointer to NULL, - * as that is only done by a task on itself, and if the current task - * is here, it is not simultaneously in the exit code NULL'ing its - * cpuset pointer. This routine also might acquire callback_sem and + * Call without callback_mutex or task_lock() held. May be + * called with or without manage_mutex held. Thanks in part to + * 'the_top_cpuset_hack', the tasks cpuset pointer will never + * be NULL. This routine also might acquire callback_mutex and * current->mm->mmap_sem during call. * * Reading current->cpuset->mems_generation doesn't need task_lock @@ -658,13 +664,21 @@ void cpuset_update_task_memory_state(void) } if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) { - down(&callback_sem); + mutex_lock(&callback_mutex); task_lock(tsk); cs = tsk->cpuset; /* 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); - up(&callback_sem); + mutex_unlock(&callback_mutex); mpol_rebind_task(tsk, &tsk->mems_allowed); } } @@ -674,7 +688,7 @@ void cpuset_update_task_memory_state(void) * * One cpuset is a subset of another if all its allowed CPUs and * Memory Nodes are a subset of the other, and its exclusive flags - * are only set if the other's are set. Call holding manage_sem. + * are only set if the other's are set. Call holding manage_mutex. */ static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q) @@ -692,7 +706,7 @@ static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q) * If we replaced the flag and mask values of the current cpuset * (cur) with those values in the trial cpuset (trial), would * our various subset and exclusive rules still be valid? Presumes - * manage_sem held. + * manage_mutex held. * * 'cur' is the address of an actual, in-use cpuset. Operations * such as list traversal that depend on the actual address of the @@ -746,7 +760,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) * exclusive child cpusets * Build these two partitions by calling partition_sched_domains * - * Call with manage_sem held. May nest a call to the + * Call with manage_mutex held. May nest a call to the * lock_cpu_hotplug()/unlock_cpu_hotplug() pair. */ @@ -792,7 +806,7 @@ static void update_cpu_domains(struct cpuset *cur) } /* - * Call with manage_sem held. May take callback_sem during call. + * Call with manage_mutex held. May take callback_mutex during call. */ static int update_cpumask(struct cpuset *cs, char *buf) @@ -811,15 +825,64 @@ static int update_cpumask(struct cpuset *cs, char *buf) if (retval < 0) return retval; cpus_unchanged = cpus_equal(cs->cpus_allowed, trialcs.cpus_allowed); - down(&callback_sem); + mutex_lock(&callback_mutex); cs->cpus_allowed = trialcs.cpus_allowed; - up(&callback_sem); + mutex_unlock(&callback_mutex); if (is_cpu_exclusive(cs) && !cpus_unchanged) update_cpu_domains(cs); return 0; } /* + * cpuset_migrate_mm + * + * Migrate memory region from one set of nodes to another. + * + * Temporarilly set tasks mems_allowed to target nodes of migration, + * so that the migration code can allocate pages on these nodes. + * + * Call holding manage_mutex, so our current->cpuset won't change + * during this call, as manage_mutex holds off any attach_task() + * calls. Therefore we don't need to take task_lock around the + * call to guarantee_online_mems(), as we know no one is changing + * our tasks cpuset. + * + * Hold callback_mutex around the two modifications of our tasks + * mems_allowed to synchronize with cpuset_mems_allowed(). + * + * While the mm_struct we are migrating is typically from some + * 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, + const nodemask_t *to) +{ + 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(tsk->cpuset, &tsk->mems_allowed); + mutex_unlock(&callback_mutex); +} + +/* * 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 @@ -827,7 +890,7 @@ static int update_cpumask(struct cpuset *cs, char *buf) * the cpuset is marked 'memory_migrate', migrate the tasks * pages to the new memory. * - * Call with manage_sem held. May take callback_sem during call. + * Call with manage_mutex held. May take callback_mutex during call. * Will take tasklist_lock, scan tasklist for tasks in cpuset cs, * lock each such tasks mm->mmap_sem, scan its vma's and rebind * their mempolicies to the cpusets new mems_allowed. @@ -862,11 +925,10 @@ static int update_nodemask(struct cpuset *cs, char *buf) if (retval < 0) goto done; - down(&callback_sem); + mutex_lock(&callback_mutex); cs->mems_allowed = trialcs.mems_allowed; - atomic_inc(&cpuset_mems_generation); - cs->mems_generation = atomic_read(&cpuset_mems_generation); - up(&callback_sem); + cs->mems_generation = cpuset_mems_generation++; + mutex_unlock(&callback_mutex); set_cpuset_being_rebound(cs); /* causes mpol_copy() rebind */ @@ -922,7 +984,7 @@ static int update_nodemask(struct cpuset *cs, char *buf) * tasklist_lock. Forks can happen again now - the mpol_copy() * cpuset_being_rebound check will catch such forks, and rebind * their vma mempolicies too. Because we still hold the global - * cpuset manage_sem, we know that no other rebind effort will + * cpuset manage_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. @@ -932,10 +994,8 @@ static int update_nodemask(struct cpuset *cs, char *buf) struct mm_struct *mm = mmarray[i]; mpol_rebind_mm(mm, &cs->mems_allowed); - if (migrate) { - do_migrate_pages(mm, &oldmem, &cs->mems_allowed, - MPOL_MF_MOVE_ALL); - } + if (migrate) + cpuset_migrate_mm(mm, &oldmem, &cs->mems_allowed); mmput(mm); } @@ -948,7 +1008,7 @@ done: } /* - * Call with manage_sem held. + * Call with manage_mutex held. */ static int update_memory_pressure_enabled(struct cpuset *cs, char *buf) @@ -963,11 +1023,12 @@ static int update_memory_pressure_enabled(struct cpuset *cs, char *buf) /* * update_flag - read a 0 or a 1 in a file and update associated flag * bit: the bit to update (CS_CPU_EXCLUSIVE, CS_MEM_EXCLUSIVE, - * CS_NOTIFY_ON_RELEASE, CS_MEMORY_MIGRATE) + * CS_NOTIFY_ON_RELEASE, CS_MEMORY_MIGRATE, + * CS_SPREAD_PAGE, CS_SPREAD_SLAB) * cs: the cpuset to update * buf: the buffer where we read the 0 or 1 * - * Call with manage_sem held. + * Call with manage_mutex held. */ static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, char *buf) @@ -989,12 +1050,12 @@ static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, char *buf) return err; cpu_exclusive_changed = (is_cpu_exclusive(cs) != is_cpu_exclusive(&trialcs)); - down(&callback_sem); + mutex_lock(&callback_mutex); if (turning_on) set_bit(bit, &cs->flags); else clear_bit(bit, &cs->flags); - up(&callback_sem); + mutex_unlock(&callback_mutex); if (cpu_exclusive_changed) update_cpu_domains(cs); @@ -1104,7 +1165,7 @@ static int fmeter_getrate(struct fmeter *fmp) * writing the path of the old cpuset in 'ppathbuf' if it needs to be * notified on release. * - * Call holding manage_sem. May take callback_sem and task_lock of + * Call holding manage_mutex. May take callback_mutex and task_lock of * the task 'pid' during call. */ @@ -1144,13 +1205,13 @@ static int attach_task(struct cpuset *cs, char *pidbuf, char **ppathbuf) get_task_struct(tsk); } - down(&callback_sem); + mutex_lock(&callback_mutex); task_lock(tsk); oldcs = tsk->cpuset; if (!oldcs) { task_unlock(tsk); - up(&callback_sem); + mutex_unlock(&callback_mutex); put_task_struct(tsk); return -ESRCH; } @@ -1164,16 +1225,16 @@ static int attach_task(struct cpuset *cs, char *pidbuf, char **ppathbuf) from = oldcs->mems_allowed; to = cs->mems_allowed; - up(&callback_sem); + mutex_unlock(&callback_mutex); mm = get_task_mm(tsk); if (mm) { mpol_rebind_mm(mm, &to); + if (is_memory_migrate(cs)) + cpuset_migrate_mm(mm, &from, &to); mmput(mm); } - if (is_memory_migrate(cs)) - do_migrate_pages(tsk->mm, &from, &to, MPOL_MF_MOVE_ALL); put_task_struct(tsk); synchronize_rcu(); if (atomic_dec_and_test(&oldcs->count)) @@ -1194,6 +1255,8 @@ typedef enum { FILE_NOTIFY_ON_RELEASE, FILE_MEMORY_PRESSURE_ENABLED, FILE_MEMORY_PRESSURE, + FILE_SPREAD_PAGE, + FILE_SPREAD_SLAB, FILE_TASKLIST, } cpuset_filetype_t; @@ -1221,7 +1284,7 @@ static ssize_t cpuset_common_file_write(struct file *file, const char __user *us } buffer[nbytes] = 0; /* nul-terminate */ - down(&manage_sem); + mutex_lock(&manage_mutex); if (is_removed(cs)) { retval = -ENODEV; @@ -1253,6 +1316,14 @@ static ssize_t cpuset_common_file_write(struct file *file, const char __user *us case FILE_MEMORY_PRESSURE: retval = -EACCES; break; + case FILE_SPREAD_PAGE: + retval = update_flag(CS_SPREAD_PAGE, cs, buffer); + cs->mems_generation = cpuset_mems_generation++; + break; + case FILE_SPREAD_SLAB: + retval = update_flag(CS_SPREAD_SLAB, cs, buffer); + cs->mems_generation = cpuset_mems_generation++; + break; case FILE_TASKLIST: retval = attach_task(cs, buffer, &pathbuf); break; @@ -1264,7 +1335,7 @@ static ssize_t cpuset_common_file_write(struct file *file, const char __user *us if (retval == 0) retval = nbytes; out2: - up(&manage_sem); + mutex_unlock(&manage_mutex); cpuset_release_agent(pathbuf); out1: kfree(buffer); @@ -1304,9 +1375,9 @@ static int cpuset_sprintf_cpulist(char *page, struct cpuset *cs) { cpumask_t mask; - down(&callback_sem); + mutex_lock(&callback_mutex); mask = cs->cpus_allowed; - up(&callback_sem); + mutex_unlock(&callback_mutex); return cpulist_scnprintf(page, PAGE_SIZE, mask); } @@ -1315,9 +1386,9 @@ static int cpuset_sprintf_memlist(char *page, struct cpuset *cs) { nodemask_t mask; - down(&callback_sem); + mutex_lock(&callback_mutex); mask = cs->mems_allowed; - up(&callback_sem); + mutex_unlock(&callback_mutex); return nodelist_scnprintf(page, PAGE_SIZE, mask); } @@ -1362,6 +1433,12 @@ static ssize_t cpuset_common_file_read(struct file *file, char __user *buf, case FILE_MEMORY_PRESSURE: s += sprintf(s, "%d", fmeter_getrate(&cs->fmeter)); break; + case FILE_SPREAD_PAGE: + *s++ = is_spread_page(cs) ? '1' : '0'; + break; + case FILE_SPREAD_SLAB: + *s++ = is_spread_slab(cs) ? '1' : '0'; + break; default: retval = -EINVAL; goto out; @@ -1598,7 +1675,7 @@ static int pid_array_to_buf(char *buf, int sz, pid_t *a, int npids) * Handle an open on 'tasks' file. Prepare a buffer listing the * process id's of tasks currently attached to the cpuset being opened. * - * Does not require any specific cpuset semaphores, and does not take any. + * Does not require any specific cpuset mutexes, and does not take any. */ static int cpuset_tasks_open(struct inode *unused, struct file *file) { @@ -1725,6 +1802,16 @@ static struct cftype cft_memory_pressure = { .private = FILE_MEMORY_PRESSURE, }; +static struct cftype cft_spread_page = { + .name = "memory_spread_page", + .private = FILE_SPREAD_PAGE, +}; + +static struct cftype cft_spread_slab = { + .name = "memory_spread_slab", + .private = FILE_SPREAD_SLAB, +}; + static int cpuset_populate_dir(struct dentry *cs_dentry) { int err; @@ -1743,6 +1830,10 @@ static int cpuset_populate_dir(struct dentry *cs_dentry) return err; if ((err = cpuset_add_file(cs_dentry, &cft_memory_pressure)) < 0) return err; + if ((err = cpuset_add_file(cs_dentry, &cft_spread_page)) < 0) + return err; + if ((err = cpuset_add_file(cs_dentry, &cft_spread_slab)) < 0) + return err; if ((err = cpuset_add_file(cs_dentry, &cft_tasks)) < 0) return err; return 0; @@ -1754,7 +1845,7 @@ static int cpuset_populate_dir(struct dentry *cs_dentry) * name: name of the new cpuset. Will be strcpy'ed. * mode: mode to set on new inode * - * Must be called with the semaphore on the parent inode held + * Must be called with the mutex on the parent inode held */ static long cpuset_create(struct cpuset *parent, const char *name, int mode) @@ -1766,44 +1857,47 @@ static long cpuset_create(struct cpuset *parent, const char *name, int mode) if (!cs) return -ENOMEM; - down(&manage_sem); + mutex_lock(&manage_mutex); cpuset_update_task_memory_state(); cs->flags = 0; if (notify_on_release(parent)) set_bit(CS_NOTIFY_ON_RELEASE, &cs->flags); + if (is_spread_page(parent)) + set_bit(CS_SPREAD_PAGE, &cs->flags); + if (is_spread_slab(parent)) + set_bit(CS_SPREAD_SLAB, &cs->flags); cs->cpus_allowed = CPU_MASK_NONE; cs->mems_allowed = NODE_MASK_NONE; atomic_set(&cs->count, 0); INIT_LIST_HEAD(&cs->sibling); INIT_LIST_HEAD(&cs->children); - atomic_inc(&cpuset_mems_generation); - cs->mems_generation = atomic_read(&cpuset_mems_generation); + cs->mems_generation = cpuset_mems_generation++; fmeter_init(&cs->fmeter); cs->parent = parent; - down(&callback_sem); + mutex_lock(&callback_mutex); list_add(&cs->sibling, &cs->parent->children); number_of_cpusets++; - up(&callback_sem); + mutex_unlock(&callback_mutex); err = cpuset_create_dir(cs, name, mode); if (err < 0) goto err; /* - * Release manage_sem before cpuset_populate_dir() because it + * Release manage_mutex before cpuset_populate_dir() because it * will down() this new directory's i_mutex and if we race with * another mkdir, we might deadlock. */ - up(&manage_sem); + mutex_unlock(&manage_mutex); err = cpuset_populate_dir(cs->dentry); /* If err < 0, we have a half-filled directory - oh well ;) */ return 0; err: list_del(&cs->sibling); - up(&manage_sem); + mutex_unlock(&manage_mutex); kfree(cs); return err; } @@ -1825,18 +1919,18 @@ static int cpuset_rmdir(struct inode *unused_dir, struct dentry *dentry) /* the vfs holds both inode->i_mutex already */ - down(&manage_sem); + mutex_lock(&manage_mutex); cpuset_update_task_memory_state(); if (atomic_read(&cs->count) > 0) { - up(&manage_sem); + mutex_unlock(&manage_mutex); return -EBUSY; } if (!list_empty(&cs->children)) { - up(&manage_sem); + mutex_unlock(&manage_mutex); return -EBUSY; } parent = cs->parent; - down(&callback_sem); + mutex_lock(&callback_mutex); set_bit(CS_REMOVED, &cs->flags); if (is_cpu_exclusive(cs)) update_cpu_domains(cs); @@ -1848,10 +1942,10 @@ static int cpuset_rmdir(struct inode *unused_dir, struct dentry *dentry) cpuset_d_remove_dir(d); dput(d); number_of_cpusets--; - up(&callback_sem); + mutex_unlock(&callback_mutex); if (list_empty(&parent->children)) check_for_release(parent, &pathbuf); - up(&manage_sem); + mutex_unlock(&manage_mutex); cpuset_release_agent(pathbuf); return 0; } @@ -1867,7 +1961,7 @@ int __init cpuset_init_early(void) struct task_struct *tsk = current; tsk->cpuset = &top_cpuset; - tsk->cpuset->mems_generation = atomic_read(&cpuset_mems_generation); + tsk->cpuset->mems_generation = cpuset_mems_generation++; return 0; } @@ -1886,8 +1980,7 @@ int __init cpuset_init(void) top_cpuset.mems_allowed = NODE_MASK_ALL; fmeter_init(&top_cpuset.fmeter); - atomic_inc(&cpuset_mems_generation); - top_cpuset.mems_generation = atomic_read(&cpuset_mems_generation); + top_cpuset.mems_generation = cpuset_mems_generation++; init_task.cpuset = &top_cpuset; @@ -1960,25 +2053,25 @@ void cpuset_fork(struct task_struct *child) * Description: Detach cpuset from @tsk and release it. * * Note that cpusets marked notify_on_release force every task in - * them to take the global manage_sem semaphore when exiting. + * them to take the global manage_mutex mutex when exiting. * This could impact scaling on very large systems. Be reluctant to * use notify_on_release cpusets where very high task exit scaling * is required on large systems. * * Don't even think about derefencing 'cs' after the cpuset use count - * goes to zero, except inside a critical section guarded by manage_sem - * or callback_sem. Otherwise a zero cpuset use count is a license to + * goes to zero, except inside a critical section guarded by manage_mutex + * or callback_mutex. Otherwise a zero cpuset use count is a license to * any other task to nuke the cpuset immediately, via cpuset_rmdir(). * - * This routine has to take manage_sem, not callback_sem, because - * it is holding that semaphore while calling check_for_release(), - * which calls kmalloc(), so can't be called holding callback__sem(). + * This routine has to take manage_mutex, not callback_mutex, because + * it is holding that mutex while calling check_for_release(), + * which calls kmalloc(), so can't be called holding callback_mutex(). * * We don't need to task_lock() this reference to tsk->cpuset, * because tsk is already marked PF_EXITING, so attach_task() won't * mess with it, or task is a failed fork, never visible to attach_task. * - * Hack: + * the_top_cpuset_hack: * * Set the exiting tasks cpuset to the root cpuset (top_cpuset). * @@ -2017,15 +2110,15 @@ void cpuset_exit(struct task_struct *tsk) struct cpuset *cs; cs = tsk->cpuset; - tsk->cpuset = &top_cpuset; /* Hack - see comment above */ + tsk->cpuset = &top_cpuset; /* the_top_cpuset_hack - see above */ if (notify_on_release(cs)) { char *pathbuf = NULL; - down(&manage_sem); + mutex_lock(&manage_mutex); if (atomic_dec_and_test(&cs->count)) check_for_release(cs, &pathbuf); - up(&manage_sem); + mutex_unlock(&manage_mutex); cpuset_release_agent(pathbuf); } else { atomic_dec(&cs->count); @@ -2046,11 +2139,11 @@ cpumask_t cpuset_cpus_allowed(struct task_struct *tsk) { cpumask_t mask; - down(&callback_sem); + mutex_lock(&callback_mutex); task_lock(tsk); guarantee_online_cpus(tsk->cpuset, &mask); task_unlock(tsk); - up(&callback_sem); + mutex_unlock(&callback_mutex); return mask; } @@ -2074,11 +2167,11 @@ nodemask_t cpuset_mems_allowed(struct task_struct *tsk) { nodemask_t mask; - down(&callback_sem); + mutex_lock(&callback_mutex); task_lock(tsk); guarantee_online_mems(tsk->cpuset, &mask); task_unlock(tsk); - up(&callback_sem); + mutex_unlock(&callback_mutex); return mask; } @@ -2104,7 +2197,7 @@ int cpuset_zonelist_valid_mems_allowed(struct zonelist *zl) /* * nearest_exclusive_ancestor() - Returns the nearest mem_exclusive - * ancestor to the specified cpuset. Call holding callback_sem. + * ancestor to the specified cpuset. Call holding callback_mutex. * If no ancestor is mem_exclusive (an unusual configuration), then * returns the root cpuset. */ @@ -2131,12 +2224,12 @@ static const struct cpuset *nearest_exclusive_ancestor(const struct cpuset *cs) * GFP_KERNEL allocations are not so marked, so can escape to the * nearest mem_exclusive ancestor cpuset. * - * Scanning up parent cpusets requires callback_sem. The __alloc_pages() + * Scanning up parent cpusets requires callback_mutex. The __alloc_pages() * routine only calls here with __GFP_HARDWALL bit _not_ set if * it's a GFP_KERNEL allocation, and all nodes in the current tasks * mems_allowed came up empty on the first pass over the zonelist. * So only GFP_KERNEL allocations, if all nodes in the cpuset are - * short of memory, might require taking the callback_sem semaphore. + * short of memory, might require taking the callback_mutex mutex. * * The first loop over the zonelist in mm/page_alloc.c:__alloc_pages() * calls here with __GFP_HARDWALL always set in gfp_mask, enforcing @@ -2157,7 +2250,7 @@ int __cpuset_zone_allowed(struct zone *z, gfp_t gfp_mask) { int node; /* node that zone z is on */ const struct cpuset *cs; /* current cpuset ancestors */ - int allowed = 1; /* is allocation in zone z allowed? */ + int allowed; /* is allocation in zone z allowed? */ if (in_interrupt()) return 1; @@ -2171,31 +2264,31 @@ int __cpuset_zone_allowed(struct zone *z, gfp_t gfp_mask) return 1; /* Not hardwall and node outside mems_allowed: scan up cpusets */ - down(&callback_sem); + mutex_lock(&callback_mutex); task_lock(current); cs = nearest_exclusive_ancestor(current->cpuset); task_unlock(current); allowed = node_isset(node, cs->mems_allowed); - up(&callback_sem); + mutex_unlock(&callback_mutex); return allowed; } /** * cpuset_lock - lock out any changes to cpuset structures * - * The out of memory (oom) code needs to lock down cpusets + * The out of memory (oom) code needs to mutex_lock cpusets * from being changed while it scans the tasklist looking for a - * task in an overlapping cpuset. Expose callback_sem via this + * task in an overlapping cpuset. Expose callback_mutex via this * cpuset_lock() routine, so the oom code can lock it, before * locking the task list. The tasklist_lock is a spinlock, so - * must be taken inside callback_sem. + * must be taken inside callback_mutex. */ void cpuset_lock(void) { - down(&callback_sem); + mutex_lock(&callback_mutex); } /** @@ -2206,10 +2299,48 @@ void cpuset_lock(void) void cpuset_unlock(void) { - up(&callback_sem); + mutex_unlock(&callback_mutex); } /** + * cpuset_mem_spread_node() - On which node to begin search for a page + * + * If a task is marked PF_SPREAD_PAGE or PF_SPREAD_SLAB (as for + * tasks in a cpuset with is_spread_page or is_spread_slab set), + * and if the memory allocation used cpuset_mem_spread_node() + * to determine on which node to start looking, as it will for + * certain page cache or slab cache pages such as used for file + * system buffers and inode caches, then instead of starting on the + * local node to look for a free page, rather spread the starting + * node around the tasks mems_allowed nodes. + * + * We don't have to worry about the returned node being offline + * because "it can't happen", and even if it did, it would be ok. + * + * The routines calling guarantee_online_mems() are careful to + * only set nodes in task->mems_allowed that are online. So it + * should not be possible for the following code to return an + * offline node. But if it did, that would be ok, as this routine + * is not returning the node where the allocation must be, only + * the node where the search should start. The zonelist passed to + * __alloc_pages() will include all nodes. If the slab allocator + * is passed an offline node, it will fall back to the local node. + * See kmem_cache_alloc_node(). + */ + +int cpuset_mem_spread_node(void) +{ + int node; + + node = next_node(current->cpuset_mem_spread_rotor, current->mems_allowed); + if (node == MAX_NUMNODES) + node = first_node(current->mems_allowed); + current->cpuset_mem_spread_rotor = node; + return node; +} +EXPORT_SYMBOL_GPL(cpuset_mem_spread_node); + +/** * cpuset_excl_nodes_overlap - Do we overlap @p's mem_exclusive ancestors? * @p: pointer to task_struct of some other task. * @@ -2218,7 +2349,7 @@ void cpuset_unlock(void) * determine if task @p's memory usage might impact the memory * available to the current task. * - * Call while holding callback_sem. + * Call while holding callback_mutex. **/ int cpuset_excl_nodes_overlap(const struct task_struct *p) @@ -2289,13 +2420,13 @@ void __cpuset_memory_pressure_bump(void) * - Used for /proc/<pid>/cpuset. * - No need to task_lock(tsk) on this tsk->cpuset reference, as it * doesn't really matter if tsk->cpuset changes after we read it, - * and we take manage_sem, keeping attach_task() from changing it - * anyway. + * and we take manage_mutex, keeping attach_task() from changing it + * anyway. No need to check that tsk->cpuset != NULL, thanks to + * the_top_cpuset_hack in cpuset_exit(), which sets an exiting tasks + * cpuset to top_cpuset. */ - static int proc_cpuset_show(struct seq_file *m, void *v) { - struct cpuset *cs; struct task_struct *tsk; char *buf; int retval = 0; @@ -2305,20 +2436,14 @@ static int proc_cpuset_show(struct seq_file *m, void *v) return -ENOMEM; tsk = m->private; - down(&manage_sem); - cs = tsk->cpuset; - if (!cs) { - retval = -EINVAL; - goto out; - } - - retval = cpuset_path(cs, buf, PAGE_SIZE); + mutex_lock(&manage_mutex); + retval = cpuset_path(tsk->cpuset, buf, PAGE_SIZE); if (retval < 0) goto out; seq_puts(m, buf); seq_putc(m, '\n'); out: - up(&manage_sem); + mutex_unlock(&manage_mutex); kfree(buf); return retval; } |