diff options
149 files changed, 9456 insertions, 3129 deletions
diff --git a/Documentation/DocBook/tracepoint.tmpl b/Documentation/DocBook/tracepoint.tmpl index b0756d0fd57..8bca1d5cec0 100644 --- a/Documentation/DocBook/tracepoint.tmpl +++ b/Documentation/DocBook/tracepoint.tmpl @@ -86,4 +86,9 @@ !Iinclude/trace/events/irq.h </chapter> + <chapter id="signal"> + <title>SIGNAL</title> +!Iinclude/trace/events/signal.h + </chapter> + </book> diff --git a/arch/Kconfig b/arch/Kconfig index 7f418bbc261..eef3bbb9707 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -126,4 +126,11 @@ config HAVE_DMA_API_DEBUG config HAVE_DEFAULT_NO_SPIN_MUTEXES bool +config HAVE_HW_BREAKPOINT + bool + depends on HAVE_PERF_EVENTS + select ANON_INODES + select PERF_EVENTS + + source "kernel/gcov/Kconfig" diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 72ace9515a0..178084b4377 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -49,6 +49,7 @@ config X86 select HAVE_KERNEL_GZIP select HAVE_KERNEL_BZIP2 select HAVE_KERNEL_LZMA + select HAVE_HW_BREAKPOINT select HAVE_ARCH_KMEMCHECK config OUTPUT_FORMAT diff --git a/arch/x86/include/asm/Kbuild b/arch/x86/include/asm/Kbuild index 4a8e80cdcfa..9f828f87ca3 100644 --- a/arch/x86/include/asm/Kbuild +++ b/arch/x86/include/asm/Kbuild @@ -10,6 +10,7 @@ header-y += ptrace-abi.h header-y += sigcontext32.h header-y += ucontext.h header-y += processor-flags.h +header-y += hw_breakpoint.h unifdef-y += e820.h unifdef-y += ist.h diff --git a/arch/x86/include/asm/a.out-core.h b/arch/x86/include/asm/a.out-core.h index bb70e397aa8..7a15588e45d 100644 --- a/arch/x86/include/asm/a.out-core.h +++ b/arch/x86/include/asm/a.out-core.h @@ -17,6 +17,7 @@ #include <linux/user.h> #include <linux/elfcore.h> +#include <asm/debugreg.h> /* * fill in the user structure for an a.out core dump @@ -32,14 +33,7 @@ static inline void aout_dump_thread(struct pt_regs *regs, struct user *dump) >> PAGE_SHIFT; dump->u_dsize -= dump->u_tsize; dump->u_ssize = 0; - dump->u_debugreg[0] = current->thread.debugreg0; - dump->u_debugreg[1] = current->thread.debugreg1; - dump->u_debugreg[2] = current->thread.debugreg2; - dump->u_debugreg[3] = current->thread.debugreg3; - dump->u_debugreg[4] = 0; - dump->u_debugreg[5] = 0; - dump->u_debugreg[6] = current->thread.debugreg6; - dump->u_debugreg[7] = current->thread.debugreg7; + aout_dump_debugregs(dump); if (dump->start_stack < TASK_SIZE) dump->u_ssize = ((unsigned long)(TASK_SIZE - dump->start_stack)) diff --git a/arch/x86/include/asm/debugreg.h b/arch/x86/include/asm/debugreg.h index 3ea6f37be9e..8240f76b531 100644 --- a/arch/x86/include/asm/debugreg.h +++ b/arch/x86/include/asm/debugreg.h @@ -18,6 +18,7 @@ #define DR_TRAP1 (0x2) /* db1 */ #define DR_TRAP2 (0x4) /* db2 */ #define DR_TRAP3 (0x8) /* db3 */ +#define DR_TRAP_BITS (DR_TRAP0|DR_TRAP1|DR_TRAP2|DR_TRAP3) #define DR_STEP (0x4000) /* single-step */ #define DR_SWITCH (0x8000) /* task switch */ @@ -49,6 +50,8 @@ #define DR_LOCAL_ENABLE_SHIFT 0 /* Extra shift to the local enable bit */ #define DR_GLOBAL_ENABLE_SHIFT 1 /* Extra shift to the global enable bit */ +#define DR_LOCAL_ENABLE (0x1) /* Local enable for reg 0 */ +#define DR_GLOBAL_ENABLE (0x2) /* Global enable for reg 0 */ #define DR_ENABLE_SIZE 2 /* 2 enable bits per register */ #define DR_LOCAL_ENABLE_MASK (0x55) /* Set local bits for all 4 regs */ @@ -67,4 +70,34 @@ #define DR_LOCAL_SLOWDOWN (0x100) /* Local slow the pipeline */ #define DR_GLOBAL_SLOWDOWN (0x200) /* Global slow the pipeline */ +/* + * HW breakpoint additions + */ +#ifdef __KERNEL__ + +DECLARE_PER_CPU(unsigned long, cpu_dr7); + +static inline void hw_breakpoint_disable(void) +{ + /* Zero the control register for HW Breakpoint */ + set_debugreg(0UL, 7); + + /* Zero-out the individual HW breakpoint address registers */ + set_debugreg(0UL, 0); + set_debugreg(0UL, 1); + set_debugreg(0UL, 2); + set_debugreg(0UL, 3); +} + +static inline int hw_breakpoint_active(void) +{ + return __get_cpu_var(cpu_dr7) & DR_GLOBAL_ENABLE_MASK; +} + +extern void aout_dump_debugregs(struct user *dump); + +extern void hw_breakpoint_restore(void); + +#endif /* __KERNEL__ */ + #endif /* _ASM_X86_DEBUGREG_H */ diff --git a/arch/x86/include/asm/hardirq.h b/arch/x86/include/asm/hardirq.h index 82e3e8f0104..108eb6fd1ae 100644 --- a/arch/x86/include/asm/hardirq.h +++ b/arch/x86/include/asm/hardirq.h @@ -20,11 +20,11 @@ typedef struct { unsigned int irq_call_count; unsigned int irq_tlb_count; #endif -#ifdef CONFIG_X86_MCE +#ifdef CONFIG_X86_THERMAL_VECTOR unsigned int irq_thermal_count; -# ifdef CONFIG_X86_MCE_THRESHOLD +#endif +#ifdef CONFIG_X86_MCE_THRESHOLD unsigned int irq_threshold_count; -# endif #endif } ____cacheline_aligned irq_cpustat_t; diff --git a/arch/x86/include/asm/hw_breakpoint.h b/arch/x86/include/asm/hw_breakpoint.h new file mode 100644 index 00000000000..0675a7c4c20 --- /dev/null +++ b/arch/x86/include/asm/hw_breakpoint.h @@ -0,0 +1,73 @@ +#ifndef _I386_HW_BREAKPOINT_H +#define _I386_HW_BREAKPOINT_H + +#ifdef __KERNEL__ +#define __ARCH_HW_BREAKPOINT_H + +/* + * The name should probably be something dealt in + * a higher level. While dealing with the user + * (display/resolving) + */ +struct arch_hw_breakpoint { + char *name; /* Contains name of the symbol to set bkpt */ + unsigned long address; + u8 len; + u8 type; +}; + +#include <linux/kdebug.h> +#include <linux/percpu.h> +#include <linux/list.h> + +/* Available HW breakpoint length encodings */ +#define X86_BREAKPOINT_LEN_1 0x40 +#define X86_BREAKPOINT_LEN_2 0x44 +#define X86_BREAKPOINT_LEN_4 0x4c +#define X86_BREAKPOINT_LEN_EXECUTE 0x40 + +#ifdef CONFIG_X86_64 +#define X86_BREAKPOINT_LEN_8 0x48 +#endif + +/* Available HW breakpoint type encodings */ + +/* trigger on instruction execute */ +#define X86_BREAKPOINT_EXECUTE 0x80 +/* trigger on memory write */ +#define X86_BREAKPOINT_WRITE 0x81 +/* trigger on memory read or write */ +#define X86_BREAKPOINT_RW 0x83 + +/* Total number of available HW breakpoint registers */ +#define HBP_NUM 4 + +struct perf_event; +struct pmu; + +extern int arch_check_va_in_userspace(unsigned long va, u8 hbp_len); +extern int arch_validate_hwbkpt_settings(struct perf_event *bp, + struct task_struct *tsk); +extern int hw_breakpoint_exceptions_notify(struct notifier_block *unused, + unsigned long val, void *data); + + +int arch_install_hw_breakpoint(struct perf_event *bp); +void arch_uninstall_hw_breakpoint(struct perf_event *bp); +void hw_breakpoint_pmu_read(struct perf_event *bp); +void hw_breakpoint_pmu_unthrottle(struct perf_event *bp); + +extern void +arch_fill_perf_breakpoint(struct perf_event *bp); + +unsigned long encode_dr7(int drnum, unsigned int len, unsigned int type); +int decode_dr7(unsigned long dr7, int bpnum, unsigned *len, unsigned *type); + +extern int arch_bp_generic_fields(int x86_len, int x86_type, + int *gen_len, int *gen_type); + +extern struct pmu perf_ops_bp; + +#endif /* __KERNEL__ */ +#endif /* _I386_HW_BREAKPOINT_H */ + diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h index f1363b72364..858baa061cf 100644 --- a/arch/x86/include/asm/mce.h +++ b/arch/x86/include/asm/mce.h @@ -108,6 +108,8 @@ struct mce_log { #define K8_MCE_THRESHOLD_BANK_5 (MCE_THRESHOLD_BASE + 5 * 9) #define K8_MCE_THRESHOLD_DRAM_ECC (MCE_THRESHOLD_BANK_4 + 0) +extern struct atomic_notifier_head x86_mce_decoder_chain; + #ifdef __KERNEL__ #include <linux/percpu.h> @@ -118,9 +120,11 @@ extern int mce_disabled; extern int mce_p5_enabled; #ifdef CONFIG_X86_MCE -void mcheck_init(struct cpuinfo_x86 *c); +int mcheck_init(void); +void mcheck_cpu_init(struct cpuinfo_x86 *c); #else -static inline void mcheck_init(struct cpuinfo_x86 *c) {} +static inline int mcheck_init(void) { return 0; } +static inline void mcheck_cpu_init(struct cpuinfo_x86 *c) {} #endif #ifdef CONFIG_X86_ANCIENT_MCE @@ -214,5 +218,11 @@ void intel_init_thermal(struct cpuinfo_x86 *c); void mce_log_therm_throt_event(__u64 status); +#ifdef CONFIG_X86_THERMAL_VECTOR +extern void mcheck_intel_therm_init(void); +#else +static inline void mcheck_intel_therm_init(void) { } +#endif + #endif /* __KERNEL__ */ #endif /* _ASM_X86_MCE_H */ diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index c9786480f0f..6f8ec1c37e0 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -30,6 +30,7 @@ struct mm_struct; #include <linux/math64.h> #include <linux/init.h> +#define HBP_NUM 4 /* * Default implementation of macro that returns current * instruction pointer ("program counter"). @@ -422,6 +423,8 @@ extern unsigned int xstate_size; extern void free_thread_xstate(struct task_struct *); extern struct kmem_cache *task_xstate_cachep; +struct perf_event; + struct thread_struct { /* Cached TLS descriptors: */ struct desc_struct tls_array[GDT_ENTRY_TLS_ENTRIES]; @@ -443,13 +446,10 @@ struct thread_struct { unsigned long fs; #endif unsigned long gs; - /* Hardware debugging registers: */ - unsigned long debugreg0; - unsigned long debugreg1; - unsigned long debugreg2; - unsigned long debugreg3; - unsigned long debugreg6; - unsigned long debugreg7; + /* Save middle states of ptrace breakpoints */ + struct perf_event *ptrace_bps[HBP_NUM]; + /* Debug status used for traps, single steps, etc... */ + unsigned long debugreg6; /* Fault info: */ unsigned long cr2; unsigned long trap_no; diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h index a3d49dd7d26..3d11fd0f44c 100644 --- a/arch/x86/include/asm/ptrace.h +++ b/arch/x86/include/asm/ptrace.h @@ -227,8 +227,8 @@ extern const char *regs_query_register_name(unsigned int offset); * @regs: pt_regs from which register value is gotten. * @offset: offset number of the register. * - * regs_get_register returns the value of a register whose offset from @regs - * is @offset. The @offset is the offset of the register in struct pt_regs. + * regs_get_register returns the value of a register. The @offset is the + * offset of the register in struct pt_regs address which specified by @regs. * If @offset is bigger than MAX_REG_OFFSET, this returns 0. */ static inline unsigned long regs_get_register(struct pt_regs *regs, @@ -244,7 +244,7 @@ static inline unsigned long regs_get_register(struct pt_regs *regs, * @regs: pt_regs which contains kernel stack pointer. * @addr: address which is checked. * - * regs_within_kenel_stack() checks @addr is within the kernel stack page(s). + * regs_within_kernel_stack() checks @addr is within the kernel stack page(s). * If @addr is within the kernel stack, it returns true. If not, returns false. */ static inline int regs_within_kernel_stack(struct pt_regs *regs, @@ -260,7 +260,7 @@ static inline int regs_within_kernel_stack(struct pt_regs *regs, * @n: stack entry number. * * regs_get_kernel_stack_nth() returns @n th entry of the kernel stack which - * is specifined by @regs. If the @n th entry is NOT in the kernel stack, + * is specified by @regs. If the @n th entry is NOT in the kernel stack, * this returns 0. */ static inline unsigned long regs_get_kernel_stack_nth(struct pt_regs *regs, diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index d8e5d0cdd67..4f2e66e29ec 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -40,7 +40,7 @@ obj-$(CONFIG_X86_64) += sys_x86_64.o x8664_ksyms_64.o obj-$(CONFIG_X86_64) += syscall_64.o vsyscall_64.o obj-y += bootflag.o e820.o obj-y += pci-dma.o quirks.o i8237.o topology.o kdebugfs.o -obj-y += alternative.o i8253.o pci-nommu.o +obj-y += alternative.o i8253.o pci-nommu.o hw_breakpoint.o obj-y += tsc.o io_delay.o rtc.o obj-$(CONFIG_X86_TRAMPOLINE) += trampoline.o diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile index 68537e957a9..1d2cb383410 100644 --- a/arch/x86/kernel/cpu/Makefile +++ b/arch/x86/kernel/cpu/Makefile @@ -5,6 +5,7 @@ # Don't trace early stages of a secondary CPU boot ifdef CONFIG_FUNCTION_TRACER CFLAGS_REMOVE_common.o = -pg +CFLAGS_REMOVE_perf_event.o = -pg endif # Make sure load_percpu_segment has no stackprotector diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index cc25c2b4a56..9053be5d95c 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -837,10 +837,8 @@ static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) boot_cpu_data.x86_capability[i] &= c->x86_capability[i]; } -#ifdef CONFIG_X86_MCE /* Init Machine Check Exception if available. */ - mcheck_init(c); -#endif + mcheck_cpu_init(c); select_idle_routine(c); diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index 721a77ca811..0bcaa387586 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -46,6 +46,9 @@ #include "mce-internal.h" +#define CREATE_TRACE_POINTS +#include <trace/events/mce.h> + int mce_disabled __read_mostly; #define MISC_MCELOG_MINOR 227 @@ -85,18 +88,26 @@ static DECLARE_WAIT_QUEUE_HEAD(mce_wait); static DEFINE_PER_CPU(struct mce, mces_seen); static int cpu_missing; -static void default_decode_mce(struct mce *m) +/* + * CPU/chipset specific EDAC code can register a notifier call here to print + * MCE errors in a human-readable form. + */ +ATOMIC_NOTIFIER_HEAD(x86_mce_decoder_chain); +EXPORT_SYMBOL_GPL(x86_mce_decoder_chain); + +static int default_decode_mce(struct notifier_block *nb, unsigned long val, + void *data) { pr_emerg("No human readable MCE decoding support on this CPU type.\n"); pr_emerg("Run the message through 'mcelog --ascii' to decode.\n"); + + return NOTIFY_STOP; } -/* - * CPU/chipset specific EDAC code can register a callback here to print - * MCE errors in a human-readable form: - */ -void (*x86_mce_decode_callback)(struct mce *m) = default_decode_mce; -EXPORT_SYMBOL(x86_mce_decode_callback); +static struct notifier_block mce_dec_nb = { + .notifier_call = default_decode_mce, + .priority = -1, +}; /* MCA banks polled by the period polling timer for corrected events */ DEFINE_PER_CPU(mce_banks_t, mce_poll_banks) = { @@ -141,6 +152,9 @@ void mce_log(struct mce *mce) { unsigned next, entry; + /* Emit the trace record: */ + trace_mce_record(mce); + mce->finished = 0; wmb(); for (;;) { @@ -204,9 +218,9 @@ static void print_mce(struct mce *m) /* * Print out human-readable details about the MCE error, - * (if the CPU has an implementation for that): + * (if the CPU has an implementation for that) */ - x86_mce_decode_callback(m); + atomic_notifier_call_chain(&x86_mce_decoder_chain, 0, m); } static void print_mce_head(void) @@ -1122,7 +1136,7 @@ static int check_interval = 5 * 60; /* 5 minutes */ static DEFINE_PER_CPU(int, mce_next_interval); /* in jiffies */ static DEFINE_PER_CPU(struct timer_list, mce_timer); -static void mcheck_timer(unsigned long data) +static void mce_start_timer(unsigned long data) { struct timer_list *t = &per_cpu(mce_timer, data); int *n; @@ -1187,7 +1201,7 @@ int mce_notify_irq(void) } EXPORT_SYMBOL_GPL(mce_notify_irq); -static int mce_banks_init(void) +static int __cpuinit __mcheck_cpu_mce_banks_init(void) { int i; @@ -1206,7 +1220,7 @@ static int mce_banks_init(void) /* * Initialize Machine Checks for a CPU. */ -static int __cpuinit mce_cap_init(void) +static int __cpuinit __mcheck_cpu_cap_init(void) { unsigned b; u64 cap; @@ -1228,7 +1242,7 @@ static int __cpuinit mce_cap_init(void) WARN_ON(banks != 0 && b != banks); banks = b; if (!mce_banks) { - int err = mce_banks_init(); + int err = __mcheck_cpu_mce_banks_init(); if (err) return err; @@ -1244,7 +1258,7 @@ static int __cpuinit mce_cap_init(void) return 0; } -static void mce_init(void) +static void __mcheck_cpu_init_generic(void) { mce_banks_t all_banks; u64 cap; @@ -1273,7 +1287,7 @@ static void mce_init(void) } /* Add per CPU specific workarounds here */ -static int __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c) +static int __cpuinit __mcheck_cpu_apply_quirks(struct cpuinfo_x86 *c) { if (c->x86_vendor == X86_VENDOR_UNKNOWN) { pr_info("MCE: unknown CPU type - not enabling MCE support.\n"); @@ -1341,7 +1355,7 @@ static int __cpuinit mce_cpu_quirks(struct cpuinfo_x86 *c) return 0; } -static void __cpuinit mce_ancient_init(struct cpuinfo_x86 *c) +static void __cpuinit __mcheck_cpu_ancient_init(struct cpuinfo_x86 *c) { if (c->x86 != 5) return; @@ -1355,7 +1369,7 @@ static void __cpuinit mce_ancient_init(struct cpuinfo_x86 *c) } } -static void mce_cpu_features(struct cpuinfo_x86 *c) +static void __mcheck_cpu_init_vendor(struct cpuinfo_x86 *c) { switch (c->x86_vendor) { case X86_VENDOR_INTEL: @@ -1369,7 +1383,7 @@ static void mce_cpu_features(struct cpuinfo_x86 *c) } } -static void mce_init_timer(void) +static void __mcheck_cpu_init_timer(void) { struct timer_list *t = &__get_cpu_var(mce_timer); int *n = &__get_cpu_var(mce_next_interval); @@ -1380,7 +1394,7 @@ static void mce_init_timer(void) *n = check_interval * HZ; if (!*n) return; - setup_timer(t, mcheck_timer, smp_processor_id()); + setup_timer(t, mce_start_timer, smp_processor_id()); t->expires = round_jiffies(jiffies + *n); add_timer_on(t, smp_processor_id()); } @@ -1400,27 +1414,28 @@ void (*machine_check_vector)(struct pt_regs *, long error_code) = * Called for each booted CPU to set up machine checks. * Must be called with preempt off: */ -void __cpuinit mcheck_init(struct cpuinfo_x86 *c) +void __cpuinit mcheck_cpu_init(struct cpuinfo_x86 *c) { if (mce_disabled) return; - mce_ancient_init(c); + __mcheck_cpu_ancient_init(c); if (!mce_available(c)) return; - if (mce_cap_init() < 0 || mce_cpu_quirks(c) < 0) { + if (__mcheck_cpu_cap_init() < 0 || __mcheck_cpu_apply_quirks(c) < 0) { mce_disabled = 1; return; } machine_check_vector = do_machine_check; - mce_init(); - mce_cpu_features(c); - mce_init_timer(); + __mcheck_cpu_init_generic(); + __mcheck_cpu_init_vendor(c); + __mcheck_cpu_init_timer(); INIT_WORK(&__get_cpu_var(mce_work), mce_process_work); + } /* @@ -1640,6 +1655,15 @@ static int __init mcheck_enable(char *str) } __setup("mce", mcheck_enable); +int __init mcheck_init(void) +{ + atomic_notifier_chain_register(&x86_mce_decoder_chain, &mce_dec_nb); + + mcheck_intel_therm_init(); + + return 0; +} + /* * Sysfs support */ @@ -1648,7 +1672,7 @@ __setup("mce", mcheck_enable); * Disable machine checks on suspend and shutdown. We can't really handle * them later. */ -static int mce_disable(void) +static int mce_disable_error_reporting(void) { int i; @@ -1663,12 +1687,12 @@ static int mce_disable(void) static int mce_suspend(struct sys_device *dev, pm_message_t state) { - return mce_disable(); + return mce_disable_error_reporting(); } static int mce_shutdown(struct sys_device *dev) { - return mce_disable(); + return mce_disable_error_reporting(); } /* @@ -1678,8 +1702,8 @@ static int mce_shutdown(struct sys_device *dev) */ static int mce_resume(struct sys_device *dev) { - mce_init(); - mce_cpu_features(¤t_cpu_data); + __mcheck_cpu_init_generic(); + __mcheck_cpu_init_vendor(¤t_cpu_data); return 0; } @@ -1689,8 +1713,8 @@ static void mce_cpu_restart(void *data) del_timer_sync(&__get_cpu_var(mce_timer)); if (!mce_available(¤t_cpu_data)) return; - mce_init(); - mce_init_timer(); + __mcheck_cpu_init_generic(); + __mcheck_cpu_init_timer(); } /* Reinit MCEs after user configuration changes */ @@ -1716,7 +1740,7 @@ static void mce_enable_ce(void *all) cmci_reenable(); cmci_recheck(); if (all) - mce_init_timer(); + __mcheck_cpu_init_timer(); } static struct sysdev_class mce_sysclass = { @@ -1929,13 +1953,14 @@ static __cpuinit void mce_remove_device(unsigned int cpu) } /* Make sure there are no machine checks on offlined CPUs. */ -static void mce_disable_cpu(void *h) +static void __cpuinit mce_disable_cpu(void *h) { unsigned long action = *(unsigned long *)h; int i; if (!mce_available(¤t_cpu_data)) return; + if (!(action & CPU_TASKS_FROZEN)) cmci_clear(); for (i = 0; i < banks; i++) { @@ -1946,7 +1971,7 @@ static void mce_disable_cpu(void *h) } } -static void mce_reenable_cpu(void *h) +static void __cpuinit mce_reenable_cpu(void *h) { unsigned long action = *(unsigned long *)h; int i; @@ -2025,7 +2050,7 @@ static __init void mce_init_banks(void) } } -static __init int mce_init_device(void) +static __init int mcheck_init_device(void) { int err; int i = 0; @@ -2053,7 +2078,7 @@ static __init int mce_init_device(void) return err; } -device_initcall(mce_init_device); +device_initcall(mcheck_init_device); /* * Old style boot options parsing. Only for compatibility. @@ -2101,7 +2126,7 @@ static int fake_panic_set(void *data, u64 val) DEFINE_SIMPLE_ATTRIBUTE(fake_panic_fops, fake_panic_get, fake_panic_set, "%llu\n"); -static int __init mce_debugfs_init(void) +static int __init mcheck_debugfs_init(void) { struct dentry *dmce, *ffake_panic; @@ -2115,5 +2140,5 @@ static int __init mce_debugfs_init(void) return 0; } -late_initcall(mce_debugfs_init); +late_initcall(mcheck_debugfs_init); #endif diff --git a/arch/x86/kernel/cpu/mcheck/therm_throt.c b/arch/x86/kernel/cpu/mcheck/therm_throt.c index b3a1dba7533..4fef985fc22 100644 --- a/arch/x86/kernel/cpu/mcheck/therm_throt.c +++ b/arch/x86/kernel/cpu/mcheck/therm_throt.c @@ -49,6 +49,8 @@ static DEFINE_PER_CPU(struct thermal_state, thermal_state); static atomic_t therm_throt_en = ATOMIC_INIT(0); +static u32 lvtthmr_init __read_mostly; + #ifdef CONFIG_SYSFS #define define_therm_throt_sysdev_one_ro(_name) \ static SYSDEV_ATTR(_name, 0444, therm_throt_sysdev_show_##_name, NULL) @@ -254,6 +256,18 @@ asmlinkage void smp_thermal_interrupt(struct pt_regs *regs) ack_APIC_irq(); } +void __init mcheck_intel_therm_init(void) +{ + /* + * This function is only called on boot CPU. Save the init thermal + * LVT value on BSP and use that value to restore APs' thermal LVT + * entry BIOS programmed later + */ + if (cpu_has(&boot_cpu_data, X86_FEATURE_ACPI) && + cpu_has(&boot_cpu_data, X86_FEATURE_ACC)) + lvtthmr_init = apic_read(APIC_LVTTHMR); +} + void intel_init_thermal(struct cpuinfo_x86 *c) { unsigned int cpu = smp_processor_id(); @@ -270,7 +284,20 @@ void intel_init_thermal(struct cpuinfo_x86 *c) * since it might be delivered via SMI already: */ rdmsr(MSR_IA32_MISC_ENABLE, l, h); - h = apic_read(APIC_LVTTHMR); + + /* + * The initial value of thermal LVT entries on all APs always reads + * 0x10000 because APs are woken up by BSP issuing INIT-SIPI-SIPI + * sequence to them and LVT registers are reset to 0s except for + * the mask bits which are set to 1s when APs receive INIT IPI. + * Always restore the value that BIOS has programmed on AP based on + * BSP's info we saved since BIOS is always setting the same value + * for all threads/cores + */ + apic_write(APIC_LVTTHMR, lvtthmr_init); + + h = lvtthmr_init; + if ((l & MSR_IA32_MISC_ENABLE_TM1) && (h & APIC_DM_SMI)) { printk(KERN_DEBUG "CPU%d: Thermal monitoring handled by SMI\n", cpu); diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index bd874302420..c1bbed1021d 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -2229,10 +2229,10 @@ validate_event(struct cpu_hw_events *cpuc, struct perf_event *event) { struct hw_perf_event fake_event = event->hw; - if (event->pmu != &pmu) + if (event->pmu && event->pmu != &pmu) return 0; - return x86_schedule_event(cpuc, &fake_event); + return x86_schedule_event(cpuc, &fake_event) >= 0; } static int validate_group(struct perf_event *event) diff --git a/arch/x86/kernel/hw_breakpoint.c b/arch/x86/kernel/hw_breakpoint.c new file mode 100644 index 00000000000..d42f65ac492 --- /dev/null +++ b/arch/x86/kernel/hw_breakpoint.c @@ -0,0 +1,555 @@ +/* + * 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) 2007 Alan Stern + * Copyright (C) 2009 IBM Corporation + * Copyright (C) 2009 Frederic Weisbecker <fweisbec@gmail.com> + * + * Authors: Alan Stern <stern@rowland.harvard.edu> + * K.Prasad <prasad@linux.vnet.ibm.com> + * Frederic Weisbecker <fweisbec@gmail.com> + */ + +/* + * HW_breakpoint: a unified kernel/user-space hardware breakpoint facility, + * using the CPU's debug registers. + */ + +#include <linux/perf_event.h> +#include <linux/hw_breakpoint.h> +#include <linux/irqflags.h> +#include <linux/notifier.h> +#include <linux/kallsyms.h> +#include <linux/kprobes.h> +#include <linux/percpu.h> +#include <linux/kdebug.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/sched.h> +#include <linux/init.h> +#include <linux/smp.h> + +#include <asm/hw_breakpoint.h> +#include <asm/processor.h> +#include <asm/debugreg.h> + +/* Per cpu debug control register value */ +DEFINE_PER_CPU(unsigned long, cpu_dr7); +EXPORT_PER_CPU_SYMBOL(cpu_dr7); + +/* Per cpu debug address registers values */ +static DEFINE_PER_CPU(unsigned long, cpu_debugreg[HBP_NUM]); + +/* + * Stores the breakpoints currently in use on each breakpoint address + * register for each cpus + */ +static DEFINE_PER_CPU(struct perf_event *, bp_per_reg[HBP_NUM]); + + +static inline unsigned long +__encode_dr7(int drnum, unsigned int len, unsigned int type) +{ + unsigned long bp_info; + + bp_info = (len | type) & 0xf; + bp_info <<= (DR_CONTROL_SHIFT + drnum * DR_CONTROL_SIZE); + bp_info |= (DR_GLOBAL_ENABLE << (drnum * DR_ENABLE_SIZE)); + + return bp_info; +} + +/* + * Encode the length, type, Exact, and Enable bits for a particular breakpoint + * as stored in debug register 7. + */ +unsigned long encode_dr7(int drnum, unsigned int len, unsigned int type) +{ + return __encode_dr7(drnum, len, type) | DR_GLOBAL_SLOWDOWN; +} + +/* + * Decode the length and type bits for a particular breakpoint as + * stored in debug register 7. Return the "enabled" status. + */ +int decode_dr7(unsigned long dr7, int bpnum, unsigned *len, unsigned *type) +{ + int bp_info = dr7 >> (DR_CONTROL_SHIFT + bpnum * DR_CONTROL_SIZE); + + *len = (bp_info & 0xc) | 0x40; + *type = (bp_info & 0x3) | 0x80; + + return (dr7 >> (bpnum * DR_ENABLE_SIZE)) & 0x3; +} + +/* + * Install a perf counter breakpoint. + * + * We seek a free debug address register and use it for this + * breakpoint. Eventually we enable it in the debug control register. + * + * Atomic: we hold the counter->ctx->lock and we only handle variables + * and registers local to this cpu. + */ +int arch_install_hw_breakpoint(struct perf_event *bp) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + unsigned long *dr7; + int i; + + for (i = 0; i < HBP_NUM; i++) { + struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + + if (!*slot) { + *slot = bp; + break; + } + } + + if (WARN_ONCE(i == HBP_NUM, "Can't find any breakpoint slot")) + return -EBUSY; + + set_debugreg(info->address, i); + __get_cpu_var(cpu_debugreg[i]) = info->address; + + dr7 = &__get_cpu_var(cpu_dr7); + *dr7 |= encode_dr7(i, info->len, info->type); + + set_debugreg(*dr7, 7); + + return 0; +} + +/* + * Uninstall the breakpoint contained in the given counter. + * + * First we search the debug address register it uses and then we disable + * it. + * + * Atomic: we hold the counter->ctx->lock and we only handle variables + * and registers local to this cpu. + */ +void arch_uninstall_hw_breakpoint(struct perf_event *bp) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + unsigned long *dr7; + int i; + + for (i = 0; i < HBP_NUM; i++) { + struct perf_event **slot = &__get_cpu_var(bp_per_reg[i]); + + if (*slot == bp) { + *slot = NULL; + break; + } + } + + if (WARN_ONCE(i == HBP_NUM, "Can't find any breakpoint slot")) + return; + + dr7 = &__get_cpu_var(cpu_dr7); + *dr7 &= ~__encode_dr7(i, info->len, info->type); + + set_debugreg(*dr7, 7); +} + +static int get_hbp_len(u8 hbp_len) +{ + unsigned int len_in_bytes = 0; + + switch (hbp_len) { + case X86_BREAKPOINT_LEN_1: + len_in_bytes = 1; + break; + case X86_BREAKPOINT_LEN_2: + len_in_bytes = 2; + break; + case X86_BREAKPOINT_LEN_4: + len_in_bytes = 4; + break; +#ifdef CONFIG_X86_64 + case X86_BREAKPOINT_LEN_8: + len_in_bytes = 8; + break; +#endif + } + return len_in_bytes; +} + +/* + * Check for virtual address in user space. + */ +int arch_check_va_in_userspace(unsigned long va, u8 hbp_len) +{ + unsigned int len; + + len = get_hbp_len(hbp_len); + + return (va <= TASK_SIZE - len); +} + +/* + * Check for virtual address in kernel space. + */ +static int arch_check_va_in_kernelspace(unsigned long va, u8 hbp_len) +{ + unsigned int len; + + len = get_hbp_len(hbp_len); + + return (va >= TASK_SIZE) && ((va + len - 1) >= TASK_SIZE); +} + +/* + * Store a breakpoint's encoded address, length, and type. + */ +static int arch_store_info(struct perf_event *bp) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + /* + * For kernel-addresses, either the address or symbol name can be + * specified. + */ + if (info->name) + info->address = (unsigned long) + kallsyms_lookup_name(info->name); + if (info->address) + return 0; + + return -EINVAL; +} + +int arch_bp_generic_fields(int x86_len, int x86_type, + int *gen_len, int *gen_type) +{ + /* Len */ + switch (x86_len) { + case X86_BREAKPOINT_LEN_1: + *gen_len = HW_BREAKPOINT_LEN_1; + break; + case X86_BREAKPOINT_LEN_2: + *gen_len = HW_BREAKPOINT_LEN_2; + break; + case X86_BREAKPOINT_LEN_4: + *gen_len = HW_BREAKPOINT_LEN_4; + break; +#ifdef CONFIG_X86_64 + case X86_BREAKPOINT_LEN_8: + *gen_len = HW_BREAKPOINT_LEN_8; + break; +#endif + default: + return -EINVAL; + } + + /* Type */ + switch (x86_type) { + case X86_BREAKPOINT_EXECUTE: + *gen_type = HW_BREAKPOINT_X; + break; + case X86_BREAKPOINT_WRITE: + *gen_type = HW_BREAKPOINT_W; + break; + case X86_BREAKPOINT_RW: + *gen_type = HW_BREAKPOINT_W | HW_BREAKPOINT_R; + break; + default: + return -EINVAL; + } + + return 0; +} + + +static int arch_build_bp_info(struct perf_event *bp) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + + info->address = bp->attr.bp_addr; + + /* Len */ + switch (bp->attr.bp_len) { + case HW_BREAKPOINT_LEN_1: + info->len = X86_BREAKPOINT_LEN_1; + break; + case HW_BREAKPOINT_LEN_2: + info->len = X86_BREAKPOINT_LEN_2; + break; + case HW_BREAKPOINT_LEN_4: + info->len = X86_BREAKPOINT_LEN_4; + break; +#ifdef CONFIG_X86_64 + case HW_BREAKPOINT_LEN_8: + info->len = X86_BREAKPOINT_LEN_8; + break; +#endif + default: + return -EINVAL; + } + + /* Type */ + switch (bp->attr.bp_type) { + case HW_BREAKPOINT_W: + info->type = X86_BREAKPOINT_WRITE; + break; + case HW_BREAKPOINT_W | HW_BREAKPOINT_R: + info->type = X86_BREAKPOINT_RW; + break; + case HW_BREAKPOINT_X: + info->type = X86_BREAKPOINT_EXECUTE; + break; + default: + return -EINVAL; + } + + return 0; +} +/* + * Validate the arch-specific HW Breakpoint register settings + */ +int arch_validate_hwbkpt_settings(struct perf_event *bp, + struct task_struct *tsk) +{ + struct arch_hw_breakpoint *info = counter_arch_bp(bp); + unsigned int align; + int ret; + + + ret = arch_build_bp_info(bp); + if (ret) + return ret; + + ret = -EINVAL; + + if (info->type == X86_BREAKPOINT_EXECUTE) + /* + * Ptrace-refactoring code + * For now, we'll allow instruction breakpoint only for user-space + * addresses + */ + if ((!arch_check_va_in_userspace(info->address, info->len)) && + info->len != X86_BREAKPOINT_EXECUTE) + return ret; + + switch (info->len) { + case X86_BREAKPOINT_LEN_1: + align = 0; + break; + case X86_BREAKPOINT_LEN_2: + align = 1; + break; + case X86_BREAKPOINT_LEN_4: + align = 3; + break; +#ifdef CONFIG_X86_64 + case X86_BREAKPOINT_LEN_8: + align = 7; + break; +#endif + default: + return ret; + } + + if (bp->callback) + ret = arch_store_info(bp); + + if (ret < 0) + return ret; + /* + * Check that the low-order bits of the address are appropriate + * for the alignment implied by len. + */ + if (info->address & align) + return -EINVAL; + + /* Check that the virtual address is in the proper range */ + if (tsk) { + if (!arch_check_va_in_userspace(info->address, info->len)) + return -EFAULT; + } else { + if (!arch_check_va_in_kernelspace(info->address, info->len)) + return -EFAULT; + } + + return 0; +} + +/* + * Dump the debug register contents to the user. + * We can't dump our per cpu values because it + * may contain cpu wide breakpoint, something that + * doesn't belong to the current task. + * + * TODO: include non-ptrace user breakpoints (perf) + */ +void aout_dump_debugregs(struct user *dump) +{ + int i; + int dr7 = 0; + struct perf_event *bp; + struct arch_hw_breakpoint *info; + struct thread_struct *thread = ¤t->thread; + + for (i = 0; i < HBP_NUM; i++) { + bp = thread->ptrace_bps[i]; + + if (bp && !bp->attr.disabled) { + dump->u_debugreg[i] = bp->attr.bp_addr; + info = counter_arch_bp(bp); + dr7 |= encode_dr7(i, info->len, info->type); + } else { + dump->u_debugreg[i] = 0; + } + } + + dump->u_debugreg[4] = 0; + dump->u_debugreg[5] = 0; + dump->u_debugreg[6] = current->thread.debugreg6; + + dump->u_debugreg[7] = dr7; +} +EXPORT_SYMBOL_GPL(aout_dump_debugregs); + +/* + * Release the user breakpoints used by ptrace + */ +void flush_ptrace_hw_breakpoint(struct task_struct *tsk) +{ + int i; + struct thread_struct *t = &tsk->thread; + + for (i = 0; i < HBP_NUM; i++) { + unregister_hw_breakpoint(t->ptrace_bps[i]); + t->ptrace_bps[i] = NULL; + } +} + +void hw_breakpoint_restore(void) +{ + set_debugreg(__get_cpu_var(cpu_debugreg[0]), 0); + set_debugreg(__get_cpu_var(cpu_debugreg[1]), 1); + set_debugreg(__get_cpu_var(cpu_debugreg[2]), 2); + set_debugreg(__get_cpu_var(cpu_debugreg[3]), 3); + set_debugreg(current->thread.debugreg6, 6); + set_debugreg(__get_cpu_var(cpu_dr7), 7); +} +EXPORT_SYMBOL_GPL(hw_breakpoint_restore); + +/* + * Handle debug exception notifications. + * + * Return value is either NOTIFY_STOP or NOTIFY_DONE as explained below. + * + * NOTIFY_DONE returned if one of the following conditions is true. + * i) When the causative address is from user-space and the exception + * is a valid one, i.e. not triggered as a result of lazy debug register + * switching + * ii) When there are more bits than trap<n> set in DR6 register (such + * as BD, BS or BT) indicating that more than one debug condition is + * met and requires some more action in do_debug(). + * + * NOTIFY_STOP returned for all other cases + * + */ +static int __kprobes hw_breakpoint_handler(struct die_args *args) +{ + int i, cpu, rc = NOTIFY_STOP; + struct perf_event *bp; + unsigned long dr7, dr6; + unsigned long *dr6_p; + + /* The DR6 value is pointed by args->err */ + dr6_p = (unsigned long *)ERR_PTR(args->err); + dr6 = *dr6_p; + + /* Do an early return if no trap bits are set in DR6 */ + if ((dr6 & DR_TRAP_BITS) == 0) + return NOTIFY_DONE; + + get_debugreg(dr7, 7); + /* Disable breakpoints during exception handling */ + set_debugreg(0UL, 7); + /* + * Assert that local interrupts are disabled + * Reset the DRn bits in the virtualized register value. + * The ptrace trigger routine will add in whatever is needed. + */ + current->thread.debugreg6 &= ~DR_TRAP_BITS; + cpu = get_cpu(); + + /* Handle all the breakpoints that were triggered */ + for (i = 0; i < HBP_NUM; ++i) { + if (likely(!(dr6 & (DR_TRAP0 << i)))) + continue; + + /* + * The counter may be concurrently released but that can only + * occur from a call_rcu() path. We can then safely fetch + * the breakpoint, use its callback, touch its counter + * while we are in an rcu_read_lock() path. + */ + rcu_read_lock(); + + bp = per_cpu(bp_per_reg[i], cpu); + if (bp) + rc = NOTIFY_DONE; + /* + * Reset the 'i'th TRAP bit in dr6 to denote completion of + * exception handling + */ + (*dr6_p) &= ~(DR_TRAP0 << i); + /* + * bp can be NULL due to lazy debug register switching + * or due to concurrent perf counter removing. + */ + if (!bp) { + rcu_read_unlock(); + break; + } + + (bp->callback)(bp, args->regs); + + rcu_read_unlock(); + } + if (dr6 & (~DR_TRAP_BITS)) + rc = NOTIFY_DONE; + + set_debugreg(dr7, 7); + put_cpu(); + + return rc; +} + +/* + * Handle debug exception notifications. + */ +int __kprobes hw_breakpoint_exceptions_notify( + struct notifier_block *unused, unsigned long val, void *data) +{ + if (val != DIE_DEBUG) + return NOTIFY_DONE; + + return hw_breakpoint_handler(data); +} + +void hw_breakpoint_pmu_read(struct perf_event *bp) +{ + /* TODO */ +} + +void hw_breakpoint_pmu_unthrottle(struct perf_event *bp) +{ + /* TODO */ +} diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index 04bbd527856..19212cb0155 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -92,17 +92,17 @@ static int show_other_interrupts(struct seq_file *p, int prec) seq_printf(p, "%10u ", irq_stats(j)->irq_tlb_count); seq_printf(p, " TLB shootdowns\n"); #endif -#ifdef CONFIG_X86_MCE +#ifdef CONFIG_X86_THERMAL_VECTOR seq_printf(p, "%*s: ", prec, "TRM"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_thermal_count); seq_printf(p, " Thermal event interrupts\n"); -# ifdef CONFIG_X86_MCE_THRESHOLD +#endif +#ifdef CONFIG_X86_MCE_THRESHOLD seq_printf(p, "%*s: ", prec, "THR"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_threshold_count); seq_printf(p, " Threshold APIC interrupts\n"); -# endif #endif #ifdef CONFIG_X86_MCE seq_printf(p, "%*s: ", prec, "MCE"); @@ -194,11 +194,11 @@ u64 arch_irq_stat_cpu(unsigned int cpu) sum += irq_stats(cpu)->irq_call_count; sum += irq_stats(cpu)->irq_tlb_count; #endif -#ifdef CONFIG_X86_MCE +#ifdef CONFIG_X86_THERMAL_VECTOR sum += irq_stats(cpu)->irq_thermal_count; -# ifdef CONFIG_X86_MCE_THRESHOLD +#endif +#ifdef CONFIG_X86_MCE_THRESHOLD sum += irq_stats(cpu)->irq_threshold_count; -# endif #endif #ifdef CONFIG_X86_MCE sum += per_cpu(mce_exception_count, cpu); diff --git a/arch/x86/kernel/kgdb.c b/arch/x86/kernel/kgdb.c index 8d82a77a3f3..34e86b67550 100644 --- a/arch/x86/kernel/kgdb.c +++ b/arch/x86/kernel/kgdb.c @@ -43,6 +43,7 @@ #include <linux/smp.h> #include <linux/nmi.h> +#include <asm/debugreg.h> #include <asm/apicdef.h> #include <asm/system.h> @@ -434,6 +435,11 @@ single_step_cont(struct pt_regs *regs, struct die_args *args) "resuming...\n"); kgdb_arch_handle_exception(args->trapnr, args->signr, args->err, "c", "", regs); + /* + * Reset the BS bit in dr6 (pointed by args->err) to + * denote completion of processing + */ + (*(unsigned long *)ERR_PTR(args->err)) &= ~DR_STEP; return NOTIFY_STOP; } diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c index c5f1f117e0c..3fe86d706a1 100644 --- a/arch/x86/kernel/kprobes.c +++ b/arch/x86/kernel/kprobes.c @@ -56,6 +56,7 @@ #include <asm/uaccess.h> #include <asm/alternative.h> #include <asm/insn.h> +#include <asm/debugreg.h> void jprobe_return_end(void); @@ -945,8 +946,14 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self, ret = NOTIFY_STOP; break; case DIE_DEBUG: - if (post_kprobe_handler(args->regs)) + if (post_kprobe_handler(args->regs)) { + /* + * Reset the BS bit in dr6 (pointed by args->err) to + * denote completion of processing + */ + (*(unsigned long *)ERR_PTR(args->err)) &= ~DR_STEP; ret = NOTIFY_STOP; + } break; case DIE_GPF: /* diff --git a/arch/x86/kernel/machine_kexec_32.c b/arch/x86/kernel/machine_kexec_32.c index c1c429d0013..c843f8406da 100644 --- a/arch/x86/kernel/machine_kexec_32.c +++ b/arch/x86/kernel/machine_kexec_32.c @@ -25,6 +25,7 @@ #include <asm/desc.h> #include <asm/system.h> #include <asm/cacheflush.h> +#include <asm/debugreg.h> static void set_idt(void *newidt, __u16 limit) { @@ -202,6 +203,7 @@ void machine_kexec(struct kimage *image) /* Interrupts aren't acceptable while we reboot */ local_irq_disable(); + hw_breakpoint_disable(); if (image->preserve_context) { #ifdef CONFIG_X86_IO_APIC diff --git a/arch/x86/kernel/machine_kexec_64.c b/arch/x86/kernel/machine_kexec_64.c index 84c3bf209e9..4a8bb82248a 100644 --- a/arch/x86/kernel/machine_kexec_64.c +++ b/arch/x86/kernel/machine_kexec_64.c @@ -18,6 +18,7 @@ #include <asm/pgtable.h> #include <asm/tlbflush.h> #include <asm/mmu_context.h> +#include <asm/debugreg.h> static int init_one_level2_page(struct kimage *image, pgd_t *pgd, unsigned long addr) @@ -282,6 +283,7 @@ void machine_kexec(struct kimage *image) /* Interrupts aren't acceptable while we reboot */ local_irq_disable(); + hw_breakpoint_disable(); if (image->preserve_context) { #ifdef CONFIG_X86_IO_APIC diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index 5284cd2b577..744508e7cfd 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -10,6 +10,7 @@ #include <linux/clockchips.h> #include <linux/random.h> #include <trace/events/power.h> +#include <linux/hw_breakpoint.h> #include <asm/system.h> #include <asm/apic.h> #include <asm/syscalls.h> @@ -17,6 +18,7 @@ #include <asm/uaccess.h> #include <asm/i387.h> #include <asm/ds.h> +#include <asm/debugreg.h> unsigned long idle_halt; EXPORT_SYMBOL(idle_halt); @@ -103,14 +105,7 @@ void flush_thread(void) } #endif - clear_tsk_thread_flag(tsk, TIF_DEBUG); - - tsk->thread.debugreg0 = 0; - tsk->thread.debugreg1 = 0; - tsk->thread.debugreg2 = 0; - tsk->thread.debugreg3 = 0; - tsk->thread.debugreg6 = 0; - tsk->thread.debugreg7 = 0; + flush_ptrace_hw_breakpoint(tsk); memset(tsk->thread.tls_array, 0, sizeof(tsk->thread.tls_array)); /* * Forget coprocessor state.. @@ -192,16 +187,6 @@ void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p, else if (next->debugctlmsr != prev->debugctlmsr) update_debugctlmsr(next->debugctlmsr); - if (test_tsk_thread_flag(next_p, TIF_DEBUG)) { - set_debugreg(next->debugreg0, 0); - set_debugreg(next->debugreg1, 1); - set_debugreg(next->debugreg2, 2); - set_debugreg(next->debugreg3, 3); - /* no 4 and 5 */ - set_debugreg(next->debugreg6, 6); - set_debugreg(next->debugreg7, 7); - } - if (test_tsk_thread_flag(prev_p, TIF_NOTSC) ^ test_tsk_thread_flag(next_p, TIF_NOTSC)) { /* prev and next are different */ diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 4cf79567cda..d5bd3132ee7 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -58,6 +58,7 @@ #include <asm/idle.h> #include <asm/syscalls.h> #include <asm/ds.h> +#include <asm/debugreg.h> asmlinkage void ret_from_fork(void) __asm__("ret_from_fork"); @@ -259,7 +260,12 @@ int copy_thread(unsigned long clone_flags, unsigned long sp, task_user_gs(p) = get_user_gs(regs); + p->thread.io_bitmap_ptr = NULL; tsk = current; + err = -ENOMEM; + + memset(p->thread.ptrace_bps, 0, sizeof(p->thread.ptrace_bps)); + if (unlikely(test_tsk_thread_flag(tsk, TIF_IO_BITMAP))) { p->thread.io_bitmap_ptr = kmemdup(tsk->thread.io_bitmap_ptr, IO_BITMAP_BYTES, GFP_KERNEL); diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index eb62cbcaa49..70cf15873f3 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -52,6 +52,7 @@ #include <asm/idle.h> #include <asm/syscalls.h> #include <asm/ds.h> +#include <asm/debugreg.h> asmlinkage extern void ret_from_fork(void); @@ -297,12 +298,16 @@ int copy_thread(unsigned long clone_flags, unsigned long sp, p->thread.fs = me->thread.fs; p->thread.gs = me->thread.gs; + p->thread.io_bitmap_ptr = NULL; savesegment(gs, p->thread.gsindex); savesegment(fs, p->thread.fsindex); savesegment(es, p->thread.es); savesegment(ds, p->thread.ds); + err = -ENOMEM; + memset(p->thread.ptrace_bps, 0, sizeof(p->thread.ptrace_bps)); + if (unlikely(test_tsk_thread_flag(me, TIF_IO_BITMAP))) { p->thread.io_bitmap_ptr = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL); if (!p->thread.io_bitmap_ptr) { @@ -341,6 +346,7 @@ out: kfree(p->thread.io_bitmap_ptr); p->thread.io_bitmap_max = 0; } + return err; } @@ -495,6 +501,7 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p) */ if (preload_fpu) __math_state_restore(); + return prev_p; } diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index c4f76d275ee..04d182a7cfd 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -22,6 +22,8 @@ #include <linux/seccomp.h> #include <linux/signal.h> #include <linux/workqueue.h> +#include <linux/perf_event.h> +#include <linux/hw_breakpoint.h> #include <asm/uaccess.h> #include <asm/pgtable.h> @@ -34,6 +36,7 @@ #include <asm/prctl.h> #include <asm/proto.h> #include <asm/ds.h> +#include <asm/hw_breakpoint.h> #include "tls.h" @@ -249,11 +252,6 @@ static int set_segment_reg(struct task_struct *task, return 0; } -static unsigned long debugreg_addr_limit(struct task_struct *task) -{ - return TASK_SIZE - 3; -} - #else /* CONFIG_X86_64 */ #define FLAG_MASK (FLAG_MASK_32 | X86_EFLAGS_NT) @@ -378,15 +376,6 @@ static int set_segment_reg(struct task_struct *task, return 0; } -static unsigned long debugreg_addr_limit(struct task_struct *task) -{ -#ifdef CONFIG_IA32_EMULATION - if (test_tsk_thread_flag(task, TIF_IA32)) - return IA32_PAGE_OFFSET - 3; -#endif - return TASK_SIZE_MAX - 7; -} - #endif /* CONFIG_X86_32 */ static unsigned long get_flags(struct task_struct *task) @@ -566,99 +555,239 @@ static int genregs_set(struct task_struct *target, return ret; } +static void ptrace_triggered(struct perf_event *bp, void *data) +{ + int i; + struct thread_struct *thread = &(current->thread); + + /* + * Store in the virtual DR6 register the fact that the breakpoint + * was hit so the thread's debugger will see it. + */ + for (i = 0; i < HBP_NUM; i++) { + if (thread->ptrace_bps[i] == bp) + break; + } + + thread->debugreg6 |= (DR_TRAP0 << i); +} + /* - * This function is trivial and will be inlined by the compiler. - * Having it separates the implementation details of debug - * registers from the interface details of ptrace. + * Walk through every ptrace breakpoints for this thread and + * build the dr7 value on top of their attributes. + * */ -static unsigned long ptrace_get_debugreg(struct task_struct *child, int n) +static unsigned long ptrace_get_dr7(struct perf_event *bp[]) { - switch (n) { - case 0: return child->thread.debugreg0; - case 1: return child->thread.debugreg1; - case 2: return child->thread.debugreg2; - case 3: return child->thread.debugreg3; - case 6: return child->thread.debugreg6; - case 7: return child->thread.debugreg7; + int i; + int dr7 = 0; + struct arch_hw_breakpoint *info; + + for (i = 0; i < HBP_NUM; i++) { + if (bp[i] && !bp[i]->attr.disabled) { + info = counter_arch_bp(bp[i]); + dr7 |= encode_dr7(i, info->len, info->type); + } } - return 0; + + return dr7; } -static int ptrace_set_debugreg(struct task_struct *child, - int n, unsigned long data) +static struct perf_event * +ptrace_modify_breakpoint(struct perf_event *bp, int len, int type, + struct task_struct *tsk, int disabled) { - int i; + int err; + int gen_len, gen_type; + DEFINE_BREAKPOINT_ATTR(attr); - if (unlikely(n == 4 || n == 5)) - return -EIO; + /* + * We shoud have at least an inactive breakpoint at this + * slot. It means the user is writing dr7 without having + * written the address register first + */ + if (!bp) + return ERR_PTR(-EINVAL); - if (n < 4 && unlikely(data >= debugreg_addr_limit(child))) - return -EIO; + err = arch_bp_generic_fields(len, type, &gen_len, &gen_type); + if (err) + return ERR_PTR(err); - switch (n) { - case 0: child->thread.debugreg0 = data; break; - case 1: child->thread.debugreg1 = data; break; - case 2: child->thread.debugreg2 = data; break; - case 3: child->thread.debugreg3 = data; break; + attr = bp->attr; + attr.bp_len = gen_len; + attr.bp_type = gen_type; + attr.disabled = disabled; - case 6: - if ((data & ~0xffffffffUL) != 0) - return -EIO; - child->thread.debugreg6 = data; - break; + return modify_user_hw_breakpoint(bp, &attr, bp->callback, tsk); +} + +/* + * Handle ptrace writes to debug register 7. + */ +static int ptrace_write_dr7(struct task_struct *tsk, unsigned long data) +{ + struct thread_struct *thread = &(tsk->thread); + unsigned long old_dr7; + int i, orig_ret = 0, rc = 0; + int enabled, second_pass = 0; + unsigned len, type; + struct perf_event *bp; + + data &= ~DR_CONTROL_RESERVED; + old_dr7 = ptrace_get_dr7(thread->ptrace_bps); +restore: + /* + * Loop through all the hardware breakpoints, making the + * appropriate changes to each. + */ + for (i = 0; i < HBP_NUM; i++) { + enabled = decode_dr7(data, i, &len, &type); + bp = thread->ptrace_bps[i]; + + if (!enabled) { + if (bp) { + /* + * Don't unregister the breakpoints right-away, + * unless all register_user_hw_breakpoint() + * requests have succeeded. This prevents + * any window of opportunity for debug + * register grabbing by other users. + */ + if (!second_pass) + continue; + + thread->ptrace_bps[i] = NULL; + bp = ptrace_modify_breakpoint(bp, len, type, + tsk, 1); + if (IS_ERR(bp)) { + rc = PTR_ERR(bp); + thread->ptrace_bps[i] = NULL; + break; + } + thread->ptrace_bps[i] = bp; + } + continue; + } + + bp = ptrace_modify_breakpoint(bp, len, type, tsk, 0); - case 7: + /* Incorrect bp, or we have a bug in bp API */ + if (IS_ERR(bp)) { + rc = PTR_ERR(bp); + thread->ptrace_bps[i] = NULL; + break; + } + thread->ptrace_bps[i] = bp; + } + /* + * Make a second pass to free the remaining unused breakpoints + * or to restore the original breakpoints if an error occurred. + */ + if (!second_pass) { + second_pass = 1; + if (rc < 0) { + orig_ret = rc; + data = old_dr7; + } + goto restore; + } + return ((orig_ret < 0) ? orig_ret : rc); +} + +/* + * Handle PTRACE_PEEKUSR calls for the debug register area. + */ +static unsigned long ptrace_get_debugreg(struct task_struct *tsk, int n) +{ + struct thread_struct *thread = &(tsk->thread); + unsigned long val = 0; + + if (n < HBP_NUM) { + struct perf_event *bp; + bp = thread->ptrace_bps[n]; + if (!bp) + return 0; + val = bp->hw.info.address; + } else if (n == 6) { + val = thread->debugreg6; + } else if (n == 7) { + val = ptrace_get_dr7(thread->ptrace_bps); + } + return val; +} + +static int ptrace_set_breakpoint_addr(struct task_struct *tsk, int nr, + unsigned long addr) +{ + struct perf_event *bp; + struct thread_struct *t = &tsk->thread; + DEFINE_BREAKPOINT_ATTR(attr); + + if (!t->ptrace_bps[nr]) { /* - * Sanity-check data. Take one half-byte at once with - * check = (val >> (16 + 4*i)) & 0xf. It contains the - * R/Wi and LENi bits; bits 0 and 1 are R/Wi, and bits - * 2 and 3 are LENi. Given a list of invalid values, - * we do mask |= 1 << invalid_value, so that - * (mask >> check) & 1 is a correct test for invalid - * values. - * - * R/Wi contains the type of the breakpoint / - * watchpoint, LENi contains the length of the watched - * data in the watchpoint case. - * - * The invalid values are: - * - LENi == 0x10 (undefined), so mask |= 0x0f00. [32-bit] - * - R/Wi == 0x10 (break on I/O reads or writes), so - * mask |= 0x4444. - * - R/Wi == 0x00 && LENi != 0x00, so we have mask |= - * 0x1110. - * - * Finally, mask = 0x0f00 | 0x4444 | 0x1110 == 0x5f54. - * - * See the Intel Manual "System Programming Guide", - * 15.2.4 - * - * Note that LENi == 0x10 is defined on x86_64 in long - * mode (i.e. even for 32-bit userspace software, but - * 64-bit kernel), so the x86_64 mask value is 0x5454. - * See the AMD manual no. 24593 (AMD64 System Programming) + * Put stub len and type to register (reserve) an inactive but + * correct bp */ -#ifdef CONFIG_X86_32 -#define DR7_MASK 0x5f54 -#else -#define DR7_MASK 0x5554 -#endif - data &= ~DR_CONTROL_RESERVED; - for (i = 0; i < 4; i++) - if ((DR7_MASK >> ((data >> (16 + 4*i)) & 0xf)) & 1) - return -EIO; - child->thread.debugreg7 = data; - if (data) - set_tsk_thread_flag(child, TIF_DEBUG); - else - clear_tsk_thread_flag(child, TIF_DEBUG); - break; + attr.bp_addr = addr; + attr.bp_len = HW_BREAKPOINT_LEN_1; + attr.bp_type = HW_BREAKPOINT_W; + attr.disabled = 1; + + bp = register_user_hw_breakpoint(&attr, ptrace_triggered, tsk); + } else { + bp = t->ptrace_bps[nr]; + t->ptrace_bps[nr] = NULL; + + attr = bp->attr; + attr.bp_addr = addr; + bp = modify_user_hw_breakpoint(bp, &attr, bp->callback, tsk); } + /* + * CHECKME: the previous code returned -EIO if the addr wasn't a + * valid task virtual addr. The new one will return -EINVAL in this + * case. + * -EINVAL may be what we want for in-kernel breakpoints users, but + * -EIO looks better for ptrace, since we refuse a register writing + * for the user. And anyway this is the previous behaviour. + */ + if (IS_ERR(bp)) + return PTR_ERR(bp); + + t->ptrace_bps[nr] = bp; return 0; } /* + * Handle PTRACE_POKEUSR calls for the debug register area. + */ +int ptrace_set_debugreg(struct task_struct *tsk, int n, unsigned long val) +{ + struct thread_struct *thread = &(tsk->thread); + int rc = 0; + + /* There are no DR4 or DR5 registers */ + if (n == 4 || n == 5) + return -EIO; + + if (n == 6) { + thread->debugreg6 = val; + goto ret_path; + } + if (n < HBP_NUM) { + rc = ptrace_set_breakpoint_addr(tsk, n, val); + if (rc) + return rc; + } + /* All that's left is DR7 */ + if (n == 7) + rc = ptrace_write_dr7(tsk, val); + +ret_path: + return rc; +} + +/* * These access the current or another (stopped) task's io permission * bitmap for debugging or core dump. */ diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 2a34f9c5be2..c0ca8f921c9 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -109,6 +109,7 @@ #ifdef CONFIG_X86_64 #include <asm/numa_64.h> #endif +#include <asm/mce.h> /* * end_pfn only includes RAM, while max_pfn_mapped includes all e820 entries. @@ -1031,6 +1032,8 @@ void __init setup_arch(char **cmdline_p) #endif #endif x86_init.oem.banner(); + + mcheck_init(); } #ifdef CONFIG_X86_32 diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c index 6a44a76055a..fbf3b07c856 100644 --- a/arch/x86/kernel/signal.c +++ b/arch/x86/kernel/signal.c @@ -799,15 +799,6 @@ static void do_signal(struct pt_regs *regs) signr = get_signal_to_deliver(&info, &ka, regs, NULL); if (signr > 0) { - /* - * Re-enable any watchpoints before delivering the - * signal to user space. The processor register will - * have been cleared if the watchpoint triggered - * inside the kernel. - */ - if (current->thread.debugreg7) - set_debugreg(current->thread.debugreg7, 7); - /* Whee! Actually deliver the signal. */ if (handle_signal(signr, &info, &ka, oldset, regs) == 0) { /* diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 7e37dcee0cc..33399176512 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -529,77 +529,56 @@ asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs) dotraplinkage void __kprobes do_debug(struct pt_regs *regs, long error_code) { struct task_struct *tsk = current; - unsigned long condition; + unsigned long dr6; int si_code; - get_debugreg(condition, 6); + get_debugreg(dr6, 6); /* Catch kmemcheck conditions first of all! */ - if (condition & DR_STEP && kmemcheck_trap(regs)) + if ((dr6 & DR_STEP) && kmemcheck_trap(regs)) return; + /* DR6 may or may not be cleared by the CPU */ + set_debugreg(0, 6); /* * The processor cleared BTF, so don't mark that we need it set. */ clear_tsk_thread_flag(tsk, TIF_DEBUGCTLMSR); tsk->thread.debugctlmsr = 0; - if (notify_die(DIE_DEBUG, "debug", regs, condition, error_code, - SIGTRAP) == NOTIFY_STOP) + /* Store the virtualized DR6 value */ + tsk->thread.debugreg6 = dr6; + + if (notify_die(DIE_DEBUG, "debug", regs, PTR_ERR(&dr6), error_code, + SIGTRAP) == NOTIFY_STOP) return; /* It's safe to allow irq's after DR6 has been saved */ preempt_conditional_sti(regs); - /* Mask out spurious debug traps due to lazy DR7 setting */ - if (condition & (DR_TRAP0|DR_TRAP1|DR_TRAP2|DR_TRAP3)) { - if (!tsk->thread.debugreg7) - goto clear_dr7; + if (regs->flags & X86_VM_MASK) { + handle_vm86_trap((struct kernel_vm86_regs *) regs, + error_code, 1); + return; } -#ifdef CONFIG_X86_32 - if (regs->flags & X86_VM_MASK) - goto debug_vm86; -#endif - - /* Save debug status register where ptrace can see it */ - tsk->thread.debugreg6 = condition; - /* - * Single-stepping through TF: make sure we ignore any events in - * kernel space (but re-enable TF when returning to user mode). + * Single-stepping through system calls: ignore any exceptions in + * kernel space, but re-enable TF when returning to user mode. + * + * We already checked v86 mode above, so we can check for kernel mode + * by just checking the CPL of CS. */ - if (condition & DR_STEP) { - if (!user_mode(regs)) - goto clear_TF_reenable; + if ((dr6 & DR_STEP) && !user_mode(regs)) { + tsk->thread.debugreg6 &= ~DR_STEP; + set_tsk_thread_flag(tsk, TIF_SINGLESTEP); + regs->flags &= ~X86_EFLAGS_TF; } - - si_code = get_si_code(condition); - /* Ok, finally something we can handle */ - send_sigtrap(tsk, regs, error_code, si_code); - - /* - * Disable additional traps. They'll be re-enabled when - * the signal is delivered. - */ -clear_dr7: - set_debugreg(0, 7); + si_code = get_si_code(tsk->thread.debugreg6); + if (tsk->thread.debugreg6 & (DR_STEP | DR_TRAP_BITS)) + send_sigtrap(tsk, regs, error_code, si_code); preempt_conditional_cli(regs); - return; -#ifdef CONFIG_X86_32 -debug_vm86: - /* reenable preemption: handle_vm86_trap() might sleep */ - dec_preempt_count(); - handle_vm86_trap((struct kernel_vm86_regs *) regs, error_code, 1); - conditional_cli(regs); - return; -#endif - -clear_TF_reenable: - set_tsk_thread_flag(tsk, TIF_SINGLESTEP); - regs->flags &= ~X86_EFLAGS_TF; - preempt_conditional_cli(regs); return; } diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index ae07d261527..4fc80174191 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -42,6 +42,7 @@ #define CREATE_TRACE_POINTS #include "trace.h" +#include <asm/debugreg.h> #include <asm/uaccess.h> #include <asm/msr.h> #include <asm/desc.h> @@ -3643,14 +3644,15 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) trace_kvm_entry(vcpu->vcpu_id); kvm_x86_ops->run(vcpu, kvm_run); - if (unlikely(vcpu->arch.switch_db_regs || test_thread_flag(TIF_DEBUG))) { - set_debugreg(current->thread.debugreg0, 0); - set_debugreg(current->thread.debugreg1, 1); - set_debugreg(current->thread.debugreg2, 2); - set_debugreg(current->thread.debugreg3, 3); - set_debugreg(current->thread.debugreg6, 6); - set_debugreg(current->thread.debugreg7, 7); - } + /* + * If the guest has used debug registers, at least dr7 + * will be disabled while returning to the host. + * If we don't have active breakpoints in the host, we don't + * care about the messed up debug address registers. But if + * we have some of them active, restore the old state. + */ + if (hw_breakpoint_active()) + hw_breakpoint_restore(); set_bit(KVM_REQ_KICK, &vcpu->requests); local_irq_enable(); diff --git a/arch/x86/mm/kmmio.c b/arch/x86/mm/kmmio.c index 16ccbd77917..11a4ad4d625 100644 --- a/arch/x86/mm/kmmio.c +++ b/arch/x86/mm/kmmio.c @@ -540,8 +540,14 @@ kmmio_die_notifier(struct notifier_block *nb, unsigned long val, void *args) struct die_args *arg = args; if (val == DIE_DEBUG && (arg->err & DR_STEP)) - if (post_kmmio_handler(arg->err, arg->regs) == 1) + if (post_kmmio_handler(arg->err, arg->regs) == 1) { + /* + * Reset the BS bit in dr6 (pointed by args->err) to + * denote completion of processing + */ + (*(unsigned long *)ERR_PTR(arg->err)) &= ~DR_STEP; return NOTIFY_STOP; + } return NOTIFY_DONE; } diff --git a/arch/x86/power/cpu.c b/arch/x86/power/cpu.c index 8aa85f17667..0a979f3e5b8 100644 --- a/arch/x86/power/cpu.c +++ b/arch/x86/power/cpu.c @@ -18,6 +18,7 @@ #include <asm/mce.h> #include <asm/xcr.h> #include <asm/suspend.h> +#include <asm/debugreg.h> #ifdef CONFIG_X86_32 static struct saved_context saved_context; @@ -142,31 +143,6 @@ static void fix_processor_context(void) #endif load_TR_desc(); /* This does ltr */ load_LDT(¤t->active_mm->context); /* This does lldt */ - - /* - * Now maybe reload the debug registers - */ - if (current->thread.debugreg7) { -#ifdef CONFIG_X86_32 - set_debugreg(current->thread.debugreg0, 0); - set_debugreg(current->thread.debugreg1, 1); - set_debugreg(current->thread.debugreg2, 2); - set_debugreg(current->thread.debugreg3, 3); - /* no 4 and 5 */ - set_debugreg(current->thread.debugreg6, 6); - set_debugreg(current->thread.debugreg7, 7); -#else - /* CONFIG_X86_64 */ - loaddebug(¤t->thread, 0); - loaddebug(¤t->thread, 1); - loaddebug(¤t->thread, 2); - loaddebug(¤t->thread, 3); - /* no 4 and 5 */ - loaddebug(¤t->thread, 6); - loaddebug(¤t->thread, 7); -#endif - } - } /** diff --git a/arch/x86/tools/test_get_len.c b/arch/x86/tools/test_get_len.c index af75e07217b..d8214dc03fa 100644 --- a/arch/x86/tools/test_get_len.c +++ b/arch/x86/tools/test_get_len.c @@ -114,6 +114,7 @@ int main(int argc, char **argv) unsigned char insn_buf[16]; struct insn insn; int insns = 0, c; + int warnings = 0; parse_args(argc, argv); @@ -151,18 +152,22 @@ int main(int argc, char **argv) insn_init(&insn, insn_buf, x86_64); insn_get_length(&insn); if (insn.length != nb) { - fprintf(stderr, "Error: %s found a difference at %s\n", + warnings++; + fprintf(stderr, "Warning: %s found difference at %s\n", prog, sym); - fprintf(stderr, "Error: %s", line); - fprintf(stderr, "Error: objdump says %d bytes, but " + fprintf(stderr, "Warning: %s", line); + fprintf(stderr, "Warning: objdump says %d bytes, but " "insn_get_length() says %d\n", nb, insn.length); if (verbose) dump_insn(stderr, &insn); - exit(2); } } - fprintf(stderr, "Succeed: decoded and checked %d instructions\n", - insns); + if (warnings) + fprintf(stderr, "Warning: decoded and checked %d" + " instructions with %d warnings\n", insns, warnings); + else + fprintf(stderr, "Succeed: decoded and checked %d" + " instructions\n", insns); return 0; } diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c index 713ed7d3724..689cc6a6214 100644 --- a/drivers/edac/edac_mce_amd.c +++ b/drivers/edac/edac_mce_amd.c @@ -3,7 +3,6 @@ static bool report_gart_errors; static void (*nb_bus_decoder)(int node_id, struct err_regs *regs); -static void (*orig_mce_callback)(struct mce *m); void amd_report_gart_errors(bool v) { @@ -363,8 +362,10 @@ static inline void amd_decode_err_code(unsigned int ec) pr_warning("Huh? Unknown MCE error 0x%x\n", ec); } -static void amd_decode_mce(struct mce *m) +static int amd_decode_mce(struct notifier_block *nb, unsigned long val, + void *data) { + struct mce *m = (struct mce *)data; struct err_regs regs; int node, ecc; @@ -420,20 +421,22 @@ static void amd_decode_mce(struct mce *m) } amd_decode_err_code(m->status & 0xffff); + + return NOTIFY_STOP; } +static struct notifier_block amd_mce_dec_nb = { + .notifier_call = amd_decode_mce, +}; + static int __init mce_amd_init(void) { /* * We can decode MCEs for Opteron and later CPUs: */ if ((boot_cpu_data.x86_vendor == X86_VENDOR_AMD) && - (boot_cpu_data.x86 >= 0xf)) { - /* safe the default decode mce callback */ - orig_mce_callback = x86_mce_decode_callback; - - x86_mce_decode_callback = amd_decode_mce; - } + (boot_cpu_data.x86 >= 0xf)) + atomic_notifier_chain_register(&x86_mce_decoder_chain, &amd_mce_dec_nb); return 0; } @@ -442,7 +445,7 @@ early_initcall(mce_amd_init); #ifdef MODULE static void __exit mce_amd_exit(void) { - x86_mce_decode_callback = orig_mce_callback; + atomic_notifier_chain_unregister(&x86_mce_decoder_chain, &amd_mce_dec_nb); } MODULE_DESCRIPTION("AMD MCE decoder"); diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h index 43360c1d8f7..47bbdf9c38d 100644 --- a/include/linux/ftrace_event.h +++ b/include/linux/ftrace_event.h @@ -137,13 +137,8 @@ struct ftrace_event_call { #define FTRACE_MAX_PROFILE_SIZE 2048 -struct perf_trace_buf { - char buf[FTRACE_MAX_PROFILE_SIZE]; - int recursion; -}; - -extern struct perf_trace_buf *perf_trace_buf; -extern struct perf_trace_buf *perf_trace_buf_nmi; +extern char *perf_trace_buf; +extern char *perf_trace_buf_nmi; #define MAX_FILTER_PRED 32 #define MAX_FILTER_STR_VAL 256 /* Should handle KSYM_SYMBOL_LEN */ diff --git a/include/linux/hw_breakpoint.h b/include/linux/hw_breakpoint.h new file mode 100644 index 00000000000..a03daed08c5 --- /dev/null +++ b/include/linux/hw_breakpoint.h @@ -0,0 +1,131 @@ +#ifndef _LINUX_HW_BREAKPOINT_H +#define _LINUX_HW_BREAKPOINT_H + +enum { + HW_BREAKPOINT_LEN_1 = 1, + HW_BREAKPOINT_LEN_2 = 2, + HW_BREAKPOINT_LEN_4 = 4, + HW_BREAKPOINT_LEN_8 = 8, +}; + +enum { + HW_BREAKPOINT_R = 1, + HW_BREAKPOINT_W = 2, + HW_BREAKPOINT_X = 4, +}; + +#ifdef __KERNEL__ + +#include <linux/perf_event.h> + +#ifdef CONFIG_HAVE_HW_BREAKPOINT + +/* As it's for in-kernel or ptrace use, we want it to be pinned */ +#define DEFINE_BREAKPOINT_ATTR(name) \ +struct perf_event_attr name = { \ + .type = PERF_TYPE_BREAKPOINT, \ + .size = sizeof(name), \ + .pinned = 1, \ +}; + +static inline void hw_breakpoint_init(struct perf_event_attr *attr) +{ + attr->type = PERF_TYPE_BREAKPOINT; + attr->size = sizeof(*attr); + attr->pinned = 1; +} + +static inline unsigned long hw_breakpoint_addr(struct perf_event *bp) +{ + return bp->attr.bp_addr; +} + +static inline int hw_breakpoint_type(struct perf_event *bp) +{ + return bp->attr.bp_type; +} + +static inline int hw_breakpoint_len(struct perf_event *bp) +{ + return bp->attr.bp_len; +} + +extern struct perf_event * +register_user_hw_breakpoint(struct perf_event_attr *attr, + perf_callback_t triggered, + struct task_struct *tsk); + +/* FIXME: only change from the attr, and don't unregister */ +extern struct perf_event * +modify_user_hw_breakpoint(struct perf_event *bp, + struct perf_event_attr *attr, + perf_callback_t triggered, + struct task_struct *tsk); + +/* + * Kernel breakpoints are not associated with any particular thread. + */ +extern struct perf_event * +register_wide_hw_breakpoint_cpu(struct perf_event_attr *attr, + perf_callback_t triggered, + int cpu); + +extern struct perf_event ** +register_wide_hw_breakpoint(struct perf_event_attr *attr, + perf_callback_t triggered); + +extern int register_perf_hw_breakpoint(struct perf_event *bp); +extern int __register_perf_hw_breakpoint(struct perf_event *bp); +extern void unregister_hw_breakpoint(struct perf_event *bp); +extern void unregister_wide_hw_breakpoint(struct perf_event **cpu_events); + +extern int reserve_bp_slot(struct perf_event *bp); +extern void release_bp_slot(struct perf_event *bp); + +extern void flush_ptrace_hw_breakpoint(struct task_struct *tsk); + +static inline struct arch_hw_breakpoint *counter_arch_bp(struct perf_event *bp) +{ + return &bp->hw.info; +} + +#else /* !CONFIG_HAVE_HW_BREAKPOINT */ + +static inline struct perf_event * +register_user_hw_breakpoint(struct perf_event_attr *attr, + perf_callback_t triggered, + struct task_struct *tsk) { return NULL; } +static inline struct perf_event * +modify_user_hw_breakpoint(struct perf_event *bp, + struct perf_event_attr *attr, + perf_callback_t triggered, + struct task_struct *tsk) { return NULL; } +static inline struct perf_event * +register_wide_hw_breakpoint_cpu(struct perf_event_attr *attr, + perf_callback_t triggered, + int cpu) { return NULL; } +static inline struct perf_event ** +register_wide_hw_breakpoint(struct perf_event_attr *attr, + perf_callback_t triggered) { return NULL; } +static inline int +register_perf_hw_breakpoint(struct perf_event *bp) { return -ENOSYS; } +static inline int +__register_perf_hw_breakpoint(struct perf_event *bp) { return -ENOSYS; } +static inline void unregister_hw_breakpoint(struct perf_event *bp) { } +static inline void +unregister_wide_hw_breakpoint(struct perf_event **cpu_events) { } +static inline int +reserve_bp_slot(struct perf_event *bp) {return -ENOSYS; } +static inline void release_bp_slot(struct perf_event *bp) { } + +static inline void flush_ptrace_hw_breakpoint(struct task_struct *tsk) { } + +static inline struct arch_hw_breakpoint *counter_arch_bp(struct perf_event *bp) +{ + return NULL; +} + +#endif /* CONFIG_HAVE_HW_BREAKPOINT */ +#endif /* __KERNEL__ */ + +#endif /* _LINUX_HW_BREAKPOINT_H */ diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7f87563c848..43adbd7f001 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -18,6 +18,10 @@ #include <linux/ioctl.h> #include <asm/byteorder.h> +#ifdef CONFIG_HAVE_HW_BREAKPOINT +#include <asm/hw_breakpoint.h> +#endif + /* * User-space ABI bits: */ @@ -31,6 +35,7 @@ enum perf_type_id { PERF_TYPE_TRACEPOINT = 2, PERF_TYPE_HW_CACHE = 3, PERF_TYPE_RAW = 4, + PERF_TYPE_BREAKPOINT = 5, PERF_TYPE_MAX, /* non-ABI */ }; @@ -209,6 +214,15 @@ struct perf_event_attr { __u32 wakeup_events; /* wakeup every n events */ __u32 wakeup_watermark; /* bytes before wakeup */ }; + + union { + struct { /* Hardware breakpoint info */ + __u64 bp_addr; + __u32 bp_type; + __u32 bp_len; + }; + }; + __u32 __reserved_2; __u64 __reserved_3; @@ -478,6 +492,11 @@ struct hw_perf_event { s64 remaining; struct hrtimer hrtimer; }; +#ifdef CONFIG_HAVE_HW_BREAKPOINT + union { /* breakpoint */ + struct arch_hw_breakpoint info; + }; +#endif }; atomic64_t prev_count; u64 sample_period; @@ -546,6 +565,10 @@ struct perf_pending_entry { void (*func)(struct perf_pending_entry *); }; +typedef void (*perf_callback_t)(struct perf_event *, void *); + +struct perf_sample_data; + /** * struct perf_event - performance event kernel representation: */ @@ -588,7 +611,7 @@ struct perf_event { u64 tstamp_running; u64 tstamp_stopped; - struct perf_event_attr attr; + struct perf_event_attr attr; struct hw_perf_event hw; struct perf_event_context *ctx; @@ -637,10 +660,18 @@ struct perf_event { struct pid_namespace *ns; u64 id; + void (*overflow_handler)(struct perf_event *event, + int nmi, struct perf_sample_data *data, + struct pt_regs *regs); + #ifdef CONFIG_EVENT_PROFILE struct event_filter *filter; #endif + perf_callback_t callback; + + perf_callback_t event_callback; + #endif /* CONFIG_PERF_EVENTS */ }; @@ -745,6 +776,14 @@ extern int hw_perf_group_sched_in(struct perf_event *group_leader, struct perf_cpu_context *cpuctx, struct perf_event_context *ctx, int cpu); extern void perf_event_update_userpage(struct perf_event *event); +extern int perf_event_release_kernel(struct perf_event *event); +extern struct perf_event * +perf_event_create_kernel_counter(struct perf_event_attr *attr, + int cpu, + pid_t pid, + perf_callback_t callback); +extern u64 perf_event_read_value(struct perf_event *event, + u64 *enabled, u64 *running); struct perf_sample_data { u64 type; @@ -821,6 +860,7 @@ extern int sysctl_perf_event_sample_rate; extern void perf_event_init(void); extern void perf_tp_event(int event_id, u64 addr, u64 count, void *record, int entry_size); +extern void perf_bp_event(struct perf_event *event, void *data); #ifndef perf_misc_flags #define perf_misc_flags(regs) (user_mode(regs) ? PERF_RECORD_MISC_USER : \ @@ -834,6 +874,8 @@ extern int perf_output_begin(struct perf_output_handle *handle, extern void perf_output_end(struct perf_output_handle *handle); extern void perf_output_copy(struct perf_output_handle *handle, const void *buf, unsigned int len); +extern int perf_swevent_get_recursion_context(void); +extern void perf_swevent_put_recursion_context(int rctx); #else static inline void perf_event_task_sched_in(struct task_struct *task, int cpu) { } @@ -855,11 +897,15 @@ static inline int perf_event_task_enable(void) { return -EINVAL; } static inline void perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr) { } +static inline void +perf_bp_event(struct perf_event *event, void *data) { } static inline void perf_event_mmap(struct vm_area_struct *vma) { } static inline void perf_event_comm(struct task_struct *tsk) { } static inline void perf_event_fork(struct task_struct *tsk) { } static inline void perf_event_init(void) { } +static inline int perf_swevent_get_recursion_context(void) { return -1; } +static inline void perf_swevent_put_recursion_context(int rctx) { } #endif diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index b50974a93af..e79e2f3ccc5 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -99,37 +99,16 @@ struct perf_event_attr; #define __SC_TEST6(t6, a6, ...) __SC_TEST(t6); __SC_TEST5(__VA_ARGS__) #ifdef CONFIG_EVENT_PROFILE -#define TRACE_SYS_ENTER_PROFILE(sname) \ -static int prof_sysenter_enable_##sname(struct ftrace_event_call *unused) \ -{ \ - return reg_prof_syscall_enter("sys"#sname); \ -} \ - \ -static void prof_sysenter_disable_##sname(struct ftrace_event_call *unused) \ -{ \ - unreg_prof_syscall_enter("sys"#sname); \ -} - -#define TRACE_SYS_EXIT_PROFILE(sname) \ -static int prof_sysexit_enable_##sname(struct ftrace_event_call *unused) \ -{ \ - return reg_prof_syscall_exit("sys"#sname); \ -} \ - \ -static void prof_sysexit_disable_##sname(struct ftrace_event_call *unused) \ -{ \ - unreg_prof_syscall_exit("sys"#sname); \ -} #define TRACE_SYS_ENTER_PROFILE_INIT(sname) \ .profile_count = ATOMIC_INIT(-1), \ - .profile_enable = prof_sysenter_enable_##sname, \ - .profile_disable = prof_sysenter_disable_##sname, + .profile_enable = prof_sysenter_enable, \ + .profile_disable = prof_sysenter_disable, #define TRACE_SYS_EXIT_PROFILE_INIT(sname) \ .profile_count = ATOMIC_INIT(-1), \ - .profile_enable = prof_sysexit_enable_##sname, \ - .profile_disable = prof_sysexit_disable_##sname, + .profile_enable = prof_sysexit_enable, \ + .profile_disable = prof_sysexit_disable, #else #define TRACE_SYS_ENTER_PROFILE(sname) #define TRACE_SYS_ENTER_PROFILE_INIT(sname) @@ -153,74 +132,46 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *unused) \ #define __SC_STR_TDECL6(t, a, ...) #t, __SC_STR_TDECL5(__VA_ARGS__) #define SYSCALL_TRACE_ENTER_EVENT(sname) \ + static const struct syscall_metadata __syscall_meta_##sname; \ static struct ftrace_event_call event_enter_##sname; \ - struct trace_event enter_syscall_print_##sname = { \ + static struct trace_event enter_syscall_print_##sname = { \ .trace = print_syscall_enter, \ }; \ - static int init_enter_##sname(struct ftrace_event_call *call) \ - { \ - int num, id; \ - num = syscall_name_to_nr("sys"#sname); \ - if (num < 0) \ - return -ENOSYS; \ - id = register_ftrace_event(&enter_syscall_print_##sname);\ - if (!id) \ - return -ENODEV; \ - event_enter_##sname.id = id; \ - set_syscall_enter_id(num, id); \ - INIT_LIST_HEAD(&event_enter_##sname.fields); \ - return 0; \ - } \ - TRACE_SYS_ENTER_PROFILE(sname); \ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) \ event_enter_##sname = { \ .name = "sys_enter"#sname, \ .system = "syscalls", \ - .event = &event_syscall_enter, \ - .raw_init = init_enter_##sname, \ + .event = &enter_syscall_print_##sname, \ + .raw_init = init_syscall_trace, \ .show_format = syscall_enter_format, \ .define_fields = syscall_enter_define_fields, \ .regfunc = reg_event_syscall_enter, \ .unregfunc = unreg_event_syscall_enter, \ - .data = "sys"#sname, \ + .data = (void *)&__syscall_meta_##sname,\ TRACE_SYS_ENTER_PROFILE_INIT(sname) \ } #define SYSCALL_TRACE_EXIT_EVENT(sname) \ + static const struct syscall_metadata __syscall_meta_##sname; \ static struct ftrace_event_call event_exit_##sname; \ - struct trace_event exit_syscall_print_##sname = { \ + static struct trace_event exit_syscall_print_##sname = { \ .trace = print_syscall_exit, \ }; \ - static int init_exit_##sname(struct ftrace_event_call *call) \ - { \ - int num, id; \ - num = syscall_name_to_nr("sys"#sname); \ - if (num < 0) \ - return -ENOSYS; \ - id = register_ftrace_event(&exit_syscall_print_##sname);\ - if (!id) \ - return -ENODEV; \ - event_exit_##sname.id = id; \ - set_syscall_exit_id(num, id); \ - INIT_LIST_HEAD(&event_exit_##sname.fields); \ - return 0; \ - } \ - TRACE_SYS_EXIT_PROFILE(sname); \ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) \ event_exit_##sname = { \ .name = "sys_exit"#sname, \ .system = "syscalls", \ - .event = &event_syscall_exit, \ - .raw_init = init_exit_##sname, \ + .event = &exit_syscall_print_##sname, \ + .raw_init = init_syscall_trace, \ .show_format = syscall_exit_format, \ .define_fields = syscall_exit_define_fields, \ .regfunc = reg_event_syscall_exit, \ .unregfunc = unreg_event_syscall_exit, \ - .data = "sys"#sname, \ + .data = (void *)&__syscall_meta_##sname,\ TRACE_SYS_EXIT_PROFILE_INIT(sname) \ } diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 2aac8a83e89..f59604ed0ec 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -280,6 +280,12 @@ static inline void tracepoint_synchronize_unregister(void) * TRACE_EVENT_FN to perform any (un)registration work. */ +#define DECLARE_EVENT_CLASS(name, proto, args, tstruct, assign, print) +#define DEFINE_EVENT(template, name, proto, args) \ + DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) + #define TRACE_EVENT(name, proto, args, struct, assign, print) \ DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) #define TRACE_EVENT_FN(name, proto, args, struct, \ diff --git a/include/trace/define_trace.h b/include/trace/define_trace.h index 2a4b3bf7403..5acfb1eb4df 100644 --- a/include/trace/define_trace.h +++ b/include/trace/define_trace.h @@ -31,6 +31,14 @@ assign, print, reg, unreg) \ DEFINE_TRACE_FN(name, reg, unreg) +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, name, proto, args) \ + DEFINE_TRACE(name) + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DEFINE_TRACE(name) + #undef DECLARE_TRACE #define DECLARE_TRACE(name, proto, args) \ DEFINE_TRACE(name) @@ -63,6 +71,9 @@ #undef TRACE_EVENT #undef TRACE_EVENT_FN +#undef DECLARE_EVENT_CLASS +#undef DEFINE_EVENT +#undef DEFINE_EVENT_PRINT #undef TRACE_HEADER_MULTI_READ /* Only undef what we defined in this file */ diff --git a/include/trace/events/block.h b/include/trace/events/block.h index 00405b5f624..5fb72733331 100644 --- a/include/trace/events/block.h +++ b/include/trace/events/block.h @@ -8,7 +8,7 @@ #include <linux/blkdev.h> #include <linux/tracepoint.h> -TRACE_EVENT(block_rq_abort, +DECLARE_EVENT_CLASS(block_rq_with_error, TP_PROTO(struct request_queue *q, struct request *rq), @@ -40,41 +40,28 @@ TRACE_EVENT(block_rq_abort, __entry->nr_sector, __entry->errors) ); -TRACE_EVENT(block_rq_insert, +DEFINE_EVENT(block_rq_with_error, block_rq_abort, TP_PROTO(struct request_queue *q, struct request *rq), - TP_ARGS(q, rq), + TP_ARGS(q, rq) +); - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( sector_t, sector ) - __field( unsigned int, nr_sector ) - __field( unsigned int, bytes ) - __array( char, rwbs, 6 ) - __array( char, comm, TASK_COMM_LEN ) - __dynamic_array( char, cmd, blk_cmd_buf_len(rq) ) - ), +DEFINE_EVENT(block_rq_with_error, block_rq_requeue, - TP_fast_assign( - __entry->dev = rq->rq_disk ? disk_devt(rq->rq_disk) : 0; - __entry->sector = blk_pc_request(rq) ? 0 : blk_rq_pos(rq); - __entry->nr_sector = blk_pc_request(rq) ? 0 : blk_rq_sectors(rq); - __entry->bytes = blk_pc_request(rq) ? blk_rq_bytes(rq) : 0; + TP_PROTO(struct request_queue *q, struct request *rq), - blk_fill_rwbs_rq(__entry->rwbs, rq); - blk_dump_cmd(__get_str(cmd), rq); - memcpy(__entry->comm, current->comm, TASK_COMM_LEN); - ), + TP_ARGS(q, rq) +); - TP_printk("%d,%d %s %u (%s) %llu + %u [%s]", - MAJOR(__entry->dev), MINOR(__entry->dev), - __entry->rwbs, __entry->bytes, __get_str(cmd), - (unsigned long long)__entry->sector, - __entry->nr_sector, __entry->comm) +DEFINE_EVENT(block_rq_with_error, block_rq_complete, + + TP_PROTO(struct request_queue *q, struct request *rq), + + TP_ARGS(q, rq) ); -TRACE_EVENT(block_rq_issue, +DECLARE_EVENT_CLASS(block_rq, TP_PROTO(struct request_queue *q, struct request *rq), @@ -86,7 +73,7 @@ TRACE_EVENT(block_rq_issue, __field( unsigned int, nr_sector ) __field( unsigned int, bytes ) __array( char, rwbs, 6 ) - __array( char, comm, TASK_COMM_LEN ) + __array( char, comm, TASK_COMM_LEN ) __dynamic_array( char, cmd, blk_cmd_buf_len(rq) ) ), @@ -108,68 +95,18 @@ TRACE_EVENT(block_rq_issue, __entry->nr_sector, __entry->comm) ); -TRACE_EVENT(block_rq_requeue, +DEFINE_EVENT(block_rq, block_rq_insert, TP_PROTO(struct request_queue *q, struct request *rq), - TP_ARGS(q, rq), - - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( sector_t, sector ) - __field( unsigned int, nr_sector ) - __field( int, errors ) - __array( char, rwbs, 6 ) - __dynamic_array( char, cmd, blk_cmd_buf_len(rq) ) - ), - - TP_fast_assign( - __entry->dev = rq->rq_disk ? disk_devt(rq->rq_disk) : 0; - __entry->sector = blk_pc_request(rq) ? 0 : blk_rq_pos(rq); - __entry->nr_sector = blk_pc_request(rq) ? 0 : blk_rq_sectors(rq); - __entry->errors = rq->errors; - - blk_fill_rwbs_rq(__entry->rwbs, rq); - blk_dump_cmd(__get_str(cmd), rq); - ), - - TP_printk("%d,%d %s (%s) %llu + %u [%d]", - MAJOR(__entry->dev), MINOR(__entry->dev), - __entry->rwbs, __get_str(cmd), - (unsigned long long)__entry->sector, - __entry->nr_sector, __entry->errors) + TP_ARGS(q, rq) ); -TRACE_EVENT(block_rq_complete, +DEFINE_EVENT(block_rq, block_rq_issue, TP_PROTO(struct request_queue *q, struct request *rq), - TP_ARGS(q, rq), - - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( sector_t, sector ) - __field( unsigned int, nr_sector ) - __field( int, errors ) - __array( char, rwbs, 6 ) - __dynamic_array( char, cmd, blk_cmd_buf_len(rq) ) - ), - - TP_fast_assign( - __entry->dev = rq->rq_disk ? disk_devt(rq->rq_disk) : 0; - __entry->sector = blk_pc_request(rq) ? 0 : blk_rq_pos(rq); - __entry->nr_sector = blk_pc_request(rq) ? 0 : blk_rq_sectors(rq); - __entry->errors = rq->errors; - - blk_fill_rwbs_rq(__entry->rwbs, rq); - blk_dump_cmd(__get_str(cmd), rq); - ), - - TP_printk("%d,%d %s (%s) %llu + %u [%d]", - MAJOR(__entry->dev), MINOR(__entry->dev), - __entry->rwbs, __get_str(cmd), - (unsigned long long)__entry->sector, - __entry->nr_sector, __entry->errors) + TP_ARGS(q, rq) ); TRACE_EVENT(block_bio_bounce, @@ -228,7 +165,7 @@ TRACE_EVENT(block_bio_complete, __entry->nr_sector, __entry->error) ); -TRACE_EVENT(block_bio_backmerge, +DECLARE_EVENT_CLASS(block_bio, TP_PROTO(struct request_queue *q, struct bio *bio), @@ -256,63 +193,28 @@ TRACE_EVENT(block_bio_backmerge, __entry->nr_sector, __entry->comm) ); -TRACE_EVENT(block_bio_frontmerge, +DEFINE_EVENT(block_bio, block_bio_backmerge, TP_PROTO(struct request_queue *q, struct bio *bio), - TP_ARGS(q, bio), - - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( sector_t, sector ) - __field( unsigned, nr_sector ) - __array( char, rwbs, 6 ) - __array( char, comm, TASK_COMM_LEN ) - ), - - TP_fast_assign( - __entry->dev = bio->bi_bdev->bd_dev; - __entry->sector = bio->bi_sector; - __entry->nr_sector = bio->bi_size >> 9; - blk_fill_rwbs(__entry->rwbs, bio->bi_rw, bio->bi_size); - memcpy(__entry->comm, current->comm, TASK_COMM_LEN); - ), - - TP_printk("%d,%d %s %llu + %u [%s]", - MAJOR(__entry->dev), MINOR(__entry->dev), __entry->rwbs, - (unsigned long long)__entry->sector, - __entry->nr_sector, __entry->comm) + TP_ARGS(q, bio) ); -TRACE_EVENT(block_bio_queue, +DEFINE_EVENT(block_bio, block_bio_frontmerge, TP_PROTO(struct request_queue *q, struct bio *bio), - TP_ARGS(q, bio), + TP_ARGS(q, bio) +); - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( sector_t, sector ) - __field( unsigned int, nr_sector ) - __array( char, rwbs, 6 ) - __array( char, comm, TASK_COMM_LEN ) - ), +DEFINE_EVENT(block_bio, block_bio_queue, - TP_fast_assign( - __entry->dev = bio->bi_bdev->bd_dev; - __entry->sector = bio->bi_sector; - __entry->nr_sector = bio->bi_size >> 9; - blk_fill_rwbs(__entry->rwbs, bio->bi_rw, bio->bi_size); - memcpy(__entry->comm, current->comm, TASK_COMM_LEN); - ), + TP_PROTO(struct request_queue *q, struct bio *bio), - TP_printk("%d,%d %s %llu + %u [%s]", - MAJOR(__entry->dev), MINOR(__entry->dev), __entry->rwbs, - (unsigned long long)__entry->sector, - __entry->nr_sector, __entry->comm) + TP_ARGS(q, bio) ); -TRACE_EVENT(block_getrq, +DECLARE_EVENT_CLASS(block_get_rq, TP_PROTO(struct request_queue *q, struct bio *bio, int rw), @@ -341,33 +243,18 @@ TRACE_EVENT(block_getrq, __entry->nr_sector, __entry->comm) ); -TRACE_EVENT(block_sleeprq, +DEFINE_EVENT(block_get_rq, block_getrq, TP_PROTO(struct request_queue *q, struct bio *bio, int rw), - TP_ARGS(q, bio, rw), + TP_ARGS(q, bio, rw) +); - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( sector_t, sector ) - __field( unsigned int, nr_sector ) - __array( char, rwbs, 6 ) - __array( char, comm, TASK_COMM_LEN ) - ), +DEFINE_EVENT(block_get_rq, block_sleeprq, - TP_fast_assign( - __entry->dev = bio ? bio->bi_bdev->bd_dev : 0; - __entry->sector = bio ? bio->bi_sector : 0; - __entry->nr_sector = bio ? bio->bi_size >> 9 : 0; - blk_fill_rwbs(__entry->rwbs, - bio ? bio->bi_rw : 0, __entry->nr_sector); - memcpy(__entry->comm, current->comm, TASK_COMM_LEN); - ), + TP_PROTO(struct request_queue *q, struct bio *bio, int rw), - TP_printk("%d,%d %s %llu + %u [%s]", - MAJOR(__entry->dev), MINOR(__entry->dev), __entry->rwbs, - (unsigned long long)__entry->sector, - __entry->nr_sector, __entry->comm) + TP_ARGS(q, bio, rw) ); TRACE_EVENT(block_plug, @@ -387,7 +274,7 @@ TRACE_EVENT(block_plug, TP_printk("[%s]", __entry->comm) ); -TRACE_EVENT(block_unplug_timer, +DECLARE_EVENT_CLASS(block_unplug, TP_PROTO(struct request_queue *q), @@ -406,23 +293,18 @@ TRACE_EVENT(block_unplug_timer, TP_printk("[%s] %d", __entry->comm, __entry->nr_rq) ); -TRACE_EVENT(block_unplug_io, +DEFINE_EVENT(block_unplug, block_unplug_timer, TP_PROTO(struct request_queue *q), - TP_ARGS(q), + TP_ARGS(q) +); - TP_STRUCT__entry( - __field( int, nr_rq ) - __array( char, comm, TASK_COMM_LEN ) - ), +DEFINE_EVENT(block_unplug, block_unplug_io, - TP_fast_assign( - __entry->nr_rq = q->rq.count[READ] + q->rq.count[WRITE]; - memcpy(__entry->comm, current->comm, TASK_COMM_LEN); - ), + TP_PROTO(struct request_queue *q), - TP_printk("[%s] %d", __entry->comm, __entry->nr_rq) + TP_ARGS(q) ); TRACE_EVENT(block_split, diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h index d09550bf3f9..318f76535bd 100644 --- a/include/trace/events/ext4.h +++ b/include/trace/events/ext4.h @@ -90,7 +90,7 @@ TRACE_EVENT(ext4_allocate_inode, (unsigned long) __entry->dir, __entry->mode) ); -TRACE_EVENT(ext4_write_begin, +DECLARE_EVENT_CLASS(ext4__write_begin, TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, unsigned int flags), @@ -118,7 +118,23 @@ TRACE_EVENT(ext4_write_begin, __entry->pos, __entry->len, __entry->flags) ); -TRACE_EVENT(ext4_ordered_write_end, +DEFINE_EVENT(ext4__write_begin, ext4_write_begin, + + TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, + unsigned int flags), + + TP_ARGS(inode, pos, len, flags) +); + +DEFINE_EVENT(ext4__write_begin, ext4_da_write_begin, + + TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, + unsigned int flags), + + TP_ARGS(inode, pos, len, flags) +); + +DECLARE_EVENT_CLASS(ext4__write_end, TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, unsigned int copied), @@ -145,57 +161,36 @@ TRACE_EVENT(ext4_ordered_write_end, __entry->pos, __entry->len, __entry->copied) ); -TRACE_EVENT(ext4_writeback_write_end, +DEFINE_EVENT(ext4__write_end, ext4_ordered_write_end, + TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, unsigned int copied), - TP_ARGS(inode, pos, len, copied), + TP_ARGS(inode, pos, len, copied) +); - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( ino_t, ino ) - __field( loff_t, pos ) - __field( unsigned int, len ) - __field( unsigned int, copied ) - ), +DEFINE_EVENT(ext4__write_end, ext4_writeback_write_end, - TP_fast_assign( - __entry->dev = inode->i_sb->s_dev; - __entry->ino = inode->i_ino; - __entry->pos = pos; - __entry->len = len; - __entry->copied = copied; - ), + TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, + unsigned int copied), - TP_printk("dev %s ino %lu pos %llu len %u copied %u", - jbd2_dev_to_name(__entry->dev), (unsigned long) __entry->ino, - __entry->pos, __entry->len, __entry->copied) + TP_ARGS(inode, pos, len, copied) ); -TRACE_EVENT(ext4_journalled_write_end, +DEFINE_EVENT(ext4__write_end, ext4_journalled_write_end, + TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, unsigned int copied), - TP_ARGS(inode, pos, len, copied), - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( ino_t, ino ) - __field( loff_t, pos ) - __field( unsigned int, len ) - __field( unsigned int, copied ) - ), + TP_ARGS(inode, pos, len, copied) +); - TP_fast_assign( - __entry->dev = inode->i_sb->s_dev; - __entry->ino = inode->i_ino; - __entry->pos = pos; - __entry->len = len; - __entry->copied = copied; - ), +DEFINE_EVENT(ext4__write_end, ext4_da_write_end, - TP_printk("dev %s ino %lu pos %llu len %u copied %u", - jbd2_dev_to_name(__entry->dev), (unsigned long) __entry->ino, - __entry->pos, __entry->len, __entry->copied) + TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, + unsigned int copied), + + TP_ARGS(inode, pos, len, copied) ); TRACE_EVENT(ext4_writepage, @@ -337,60 +332,6 @@ TRACE_EVENT(ext4_da_writepages_result, (unsigned long) __entry->writeback_index) ); -TRACE_EVENT(ext4_da_write_begin, - TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, - unsigned int flags), - - TP_ARGS(inode, pos, len, flags), - - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( ino_t, ino ) - __field( loff_t, pos ) - __field( unsigned int, len ) - __field( unsigned int, flags ) - ), - - TP_fast_assign( - __entry->dev = inode->i_sb->s_dev; - __entry->ino = inode->i_ino; - __entry->pos = pos; - __entry->len = len; - __entry->flags = flags; - ), - - TP_printk("dev %s ino %lu pos %llu len %u flags %u", - jbd2_dev_to_name(__entry->dev), (unsigned long) __entry->ino, - __entry->pos, __entry->len, __entry->flags) -); - -TRACE_EVENT(ext4_da_write_end, - TP_PROTO(struct inode *inode, loff_t pos, unsigned int len, - unsigned int copied), - - TP_ARGS(inode, pos, len, copied), - - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( ino_t, ino ) - __field( loff_t, pos ) - __field( unsigned int, len ) - __field( unsigned int, copied ) - ), - - TP_fast_assign( - __entry->dev = inode->i_sb->s_dev; - __entry->ino = inode->i_ino; - __entry->pos = pos; - __entry->len = len; - __entry->copied = copied; - ), - - TP_printk("dev %s ino %lu pos %llu len %u copied %u", - jbd2_dev_to_name(__entry->dev), (unsigned long) __entry->ino, - __entry->pos, __entry->len, __entry->copied) -); - TRACE_EVENT(ext4_discard_blocks, TP_PROTO(struct super_block *sb, unsigned long long blk, unsigned long long count), diff --git a/include/trace/events/irq.h b/include/trace/events/irq.h index dcfcd440762..0e4cfb694fe 100644 --- a/include/trace/events/irq.h +++ b/include/trace/events/irq.h @@ -82,18 +82,7 @@ TRACE_EVENT(irq_handler_exit, __entry->irq, __entry->ret ? "handled" : "unhandled") ); -/** - * softirq_entry - called immediately before the softirq handler - * @h: pointer to struct softirq_action - * @vec: pointer to first struct softirq_action in softirq_vec array - * - * The @h parameter, contains a pointer to the struct softirq_action - * which has a pointer to the action handler that is called. By subtracting - * the @vec pointer from the @h pointer, we can determine the softirq - * number. Also, when used in combination with the softirq_exit tracepoint - * we can determine the softirq latency. - */ -TRACE_EVENT(softirq_entry, +DECLARE_EVENT_CLASS(softirq, TP_PROTO(struct softirq_action *h, struct softirq_action *vec), @@ -112,6 +101,24 @@ TRACE_EVENT(softirq_entry, ); /** + * softirq_entry - called immediately before the softirq handler + * @h: pointer to struct softirq_action + * @vec: pointer to first struct softirq_action in softirq_vec array + * + * The @h parameter, contains a pointer to the struct softirq_action + * which has a pointer to the action handler that is called. By subtracting + * the @vec pointer from the @h pointer, we can determine the softirq + * number. Also, when used in combination with the softirq_exit tracepoint + * we can determine the softirq latency. + */ +DEFINE_EVENT(softirq, softirq_entry, + + TP_PROTO(struct softirq_action *h, struct softirq_action *vec), + + TP_ARGS(h, vec) +); + +/** * softirq_exit - called immediately after the softirq handler returns * @h: pointer to struct softirq_action * @vec: pointer to first struct softirq_action in softirq_vec array @@ -122,22 +129,11 @@ TRACE_EVENT(softirq_entry, * combination with the softirq_entry tracepoint we can determine the softirq * latency. */ -TRACE_EVENT(softirq_exit, +DEFINE_EVENT(softirq, softirq_exit, TP_PROTO(struct softirq_action *h, struct softirq_action *vec), - TP_ARGS(h, vec), - - TP_STRUCT__entry( - __field( int, vec ) - ), - - TP_fast_assign( - __entry->vec = (int)(h - vec); - ), - - TP_printk("vec=%d [action=%s]", __entry->vec, - show_softirq_name(__entry->vec)) + TP_ARGS(h, vec) ); #endif /* _TRACE_IRQ_H */ diff --git a/include/trace/events/jbd2.h b/include/trace/events/jbd2.h index 3c60b75adb9..96b370a050d 100644 --- a/include/trace/events/jbd2.h +++ b/include/trace/events/jbd2.h @@ -30,7 +30,7 @@ TRACE_EVENT(jbd2_checkpoint, jbd2_dev_to_name(__entry->dev), __entry->result) ); -TRACE_EVENT(jbd2_start_commit, +DECLARE_EVENT_CLASS(jbd2_commit, TP_PROTO(journal_t *journal, transaction_t *commit_transaction), @@ -53,73 +53,32 @@ TRACE_EVENT(jbd2_start_commit, __entry->sync_commit) ); -TRACE_EVENT(jbd2_commit_locking, +DEFINE_EVENT(jbd2_commit, jbd2_start_commit, TP_PROTO(journal_t *journal, transaction_t *commit_transaction), - TP_ARGS(journal, commit_transaction), - - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( char, sync_commit ) - __field( int, transaction ) - ), - - TP_fast_assign( - __entry->dev = journal->j_fs_dev->bd_dev; - __entry->sync_commit = commit_transaction->t_synchronous_commit; - __entry->transaction = commit_transaction->t_tid; - ), - - TP_printk("dev %s transaction %d sync %d", - jbd2_dev_to_name(__entry->dev), __entry->transaction, - __entry->sync_commit) + TP_ARGS(journal, commit_transaction) ); -TRACE_EVENT(jbd2_commit_flushing, +DEFINE_EVENT(jbd2_commit, jbd2_commit_locking, TP_PROTO(journal_t *journal, transaction_t *commit_transaction), - TP_ARGS(journal, commit_transaction), - - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( char, sync_commit ) - __field( int, transaction ) - ), - - TP_fast_assign( - __entry->dev = journal->j_fs_dev->bd_dev; - __entry->sync_commit = commit_transaction->t_synchronous_commit; - __entry->transaction = commit_transaction->t_tid; - ), - - TP_printk("dev %s transaction %d sync %d", - jbd2_dev_to_name(__entry->dev), __entry->transaction, - __entry->sync_commit) + TP_ARGS(journal, commit_transaction) ); -TRACE_EVENT(jbd2_commit_logging, +DEFINE_EVENT(jbd2_commit, jbd2_commit_flushing, TP_PROTO(journal_t *journal, transaction_t *commit_transaction), - TP_ARGS(journal, commit_transaction), + TP_ARGS(journal, commit_transaction) +); - TP_STRUCT__entry( - __field( dev_t, dev ) - __field( char, sync_commit ) - __field( int, transaction ) - ), +DEFINE_EVENT(jbd2_commit, jbd2_commit_logging, - TP_fast_assign( - __entry->dev = journal->j_fs_dev->bd_dev; - __entry->sync_commit = commit_transaction->t_synchronous_commit; - __entry->transaction = commit_transaction->t_tid; - ), + TP_PROTO(journal_t *journal, transaction_t *commit_transaction), - TP_printk("dev %s transaction %d sync %d", - jbd2_dev_to_name(__entry->dev), __entry->transaction, - __entry->sync_commit) + TP_ARGS(journal, commit_transaction) ); TRACE_EVENT(jbd2_end_commit, diff --git a/include/trace/events/kmem.h b/include/trace/events/kmem.h index eaf46bdd18a..3adca0ca9db 100644 --- a/include/trace/events/kmem.h +++ b/include/trace/events/kmem.h @@ -44,7 +44,7 @@ {(unsigned long)__GFP_MOVABLE, "GFP_MOVABLE"} \ ) : "GFP_NOWAIT" -TRACE_EVENT(kmalloc, +DECLARE_EVENT_CLASS(kmem_alloc, TP_PROTO(unsigned long call_site, const void *ptr, @@ -78,41 +78,23 @@ TRACE_EVENT(kmalloc, show_gfp_flags(__entry->gfp_flags)) ); -TRACE_EVENT(kmem_cache_alloc, +DEFINE_EVENT(kmem_alloc, kmalloc, - TP_PROTO(unsigned long call_site, - const void *ptr, - size_t bytes_req, - size_t bytes_alloc, - gfp_t gfp_flags), + TP_PROTO(unsigned long call_site, const void *ptr, + size_t bytes_req, size_t bytes_alloc, gfp_t gfp_flags), - TP_ARGS(call_site, ptr, bytes_req, bytes_alloc, gfp_flags), + TP_ARGS(call_site, ptr, bytes_req, bytes_alloc, gfp_flags) +); - TP_STRUCT__entry( - __field( unsigned long, call_site ) - __field( const void *, ptr ) - __field( size_t, bytes_req ) - __field( size_t, bytes_alloc ) - __field( gfp_t, gfp_flags ) - ), +DEFINE_EVENT(kmem_alloc, kmem_cache_alloc, - TP_fast_assign( - __entry->call_site = call_site; - __entry->ptr = ptr; - __entry->bytes_req = bytes_req; - __entry->bytes_alloc = bytes_alloc; - __entry->gfp_flags = gfp_flags; - ), + TP_PROTO(unsigned long call_site, const void *ptr, + size_t bytes_req, size_t bytes_alloc, gfp_t gfp_flags), - TP_printk("call_site=%lx ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s", - __entry->call_site, - __entry->ptr, - __entry->bytes_req, - __entry->bytes_alloc, - show_gfp_flags(__entry->gfp_flags)) + TP_ARGS(call_site, ptr, bytes_req, bytes_alloc, gfp_flags) ); -TRACE_EVENT(kmalloc_node, +DECLARE_EVENT_CLASS(kmem_alloc_node, TP_PROTO(unsigned long call_site, const void *ptr, @@ -150,45 +132,25 @@ TRACE_EVENT(kmalloc_node, __entry->node) ); -TRACE_EVENT(kmem_cache_alloc_node, +DEFINE_EVENT(kmem_alloc_node, kmalloc_node, - TP_PROTO(unsigned long call_site, - const void *ptr, - size_t bytes_req, - size_t bytes_alloc, - gfp_t gfp_flags, - int node), + TP_PROTO(unsigned long call_site, const void *ptr, + size_t bytes_req, size_t bytes_alloc, + gfp_t gfp_flags, int node), - TP_ARGS(call_site, ptr, bytes_req, bytes_alloc, gfp_flags, node), + TP_ARGS(call_site, ptr, bytes_req, bytes_alloc, gfp_flags, node) +); - TP_STRUCT__entry( - __field( unsigned long, call_site ) - __field( const void *, ptr ) - __field( size_t, bytes_req ) - __field( size_t, bytes_alloc ) - __field( gfp_t, gfp_flags ) - __field( int, node ) - ), +DEFINE_EVENT(kmem_alloc_node, kmem_cache_alloc_node, - TP_fast_assign( - __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; - ), + TP_PROTO(unsigned long call_site, const void *ptr, + size_t bytes_req, size_t bytes_alloc, + gfp_t gfp_flags, int node), - TP_printk("call_site=%lx ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s node=%d", - __entry->call_site, - __entry->ptr, - __entry->bytes_req, - __entry->bytes_alloc, - show_gfp_flags(__entry->gfp_flags), - __entry->node) + TP_ARGS(call_site, ptr, bytes_req, bytes_alloc, gfp_flags, node) ); -TRACE_EVENT(kfree, +DECLARE_EVENT_CLASS(kmem_free, TP_PROTO(unsigned long call_site, const void *ptr), @@ -207,23 +169,18 @@ TRACE_EVENT(kfree, TP_printk("call_site=%lx ptr=%p", __entry->call_site, __entry->ptr) ); -TRACE_EVENT(kmem_cache_free, +DEFINE_EVENT(kmem_free, kfree, TP_PROTO(unsigned long call_site, const void *ptr), - TP_ARGS(call_site, ptr), + TP_ARGS(call_site, ptr) +); - TP_STRUCT__entry( - __field( unsigned long, call_site ) - __field( const void *, ptr ) - ), +DEFINE_EVENT(kmem_free, kmem_cache_free, - TP_fast_assign( - __entry->call_site = call_site; - __entry->ptr = ptr; - ), + TP_PROTO(unsigned long call_site, const void *ptr), - TP_printk("call_site=%lx ptr=%p", __entry->call_site, __entry->ptr) + TP_ARGS(call_site, ptr) ); TRACE_EVENT(mm_page_free_direct, @@ -299,7 +256,7 @@ TRACE_EVENT(mm_page_alloc, show_gfp_flags(__entry->gfp_flags)) ); -TRACE_EVENT(mm_page_alloc_zone_locked, +DECLARE_EVENT_CLASS(mm_page, TP_PROTO(struct page *page, unsigned int order, int migratetype), @@ -325,29 +282,22 @@ TRACE_EVENT(mm_page_alloc_zone_locked, __entry->order == 0) ); -TRACE_EVENT(mm_page_pcpu_drain, +DEFINE_EVENT(mm_page, mm_page_alloc_zone_locked, - TP_PROTO(struct page *page, int order, int migratetype), + TP_PROTO(struct page *page, unsigned int order, int migratetype), - TP_ARGS(page, order, migratetype), + TP_ARGS(page, order, migratetype) +); - TP_STRUCT__entry( - __field( struct page *, page ) - __field( int, order ) - __field( int, migratetype ) - ), +DEFINE_EVENT_PRINT(mm_page, mm_page_pcpu_drain, - TP_fast_assign( - __entry->page = page; - __entry->order = order; - __entry->migratetype = migratetype; - ), + TP_PROTO(struct page *page, unsigned int order, int migratetype), + + TP_ARGS(page, order, migratetype), TP_printk("page=%p pfn=%lu order=%d migratetype=%d", - __entry->page, - page_to_pfn(__entry->page), - __entry->order, - __entry->migratetype) + __entry->page, page_to_pfn(__entry->page), + __entry->order, __entry->migratetype) ); TRACE_EVENT(mm_page_alloc_extfrag, diff --git a/include/trace/events/mce.h b/include/trace/events/mce.h new file mode 100644 index 00000000000..7eee77895cb --- /dev/null +++ b/include/trace/events/mce.h @@ -0,0 +1,69 @@ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM mce + +#if !defined(_TRACE_MCE_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_MCE_H + +#include <linux/ktime.h> +#include <linux/tracepoint.h> +#include <asm/mce.h> + +TRACE_EVENT(mce_record, + + TP_PROTO(struct mce *m), + + TP_ARGS(m), + + TP_STRUCT__entry( + __field( u64, mcgcap ) + __field( u64, mcgstatus ) + __field( u8, bank ) + __field( u64, status ) + __field( u64, addr ) + __field( u64, misc ) + __field( u64, ip ) + __field( u8, cs ) + __field( u64, tsc ) + __field( u64, walltime ) + __field( u32, cpu ) + __field( u32, cpuid ) + __field( u32, apicid ) + __field( u32, socketid ) + __field( u8, cpuvendor ) + ), + + TP_fast_assign( + __entry->mcgcap = m->mcgcap; + __entry->mcgstatus = m->mcgstatus; + __entry->bank = m->bank; + __entry->status = m->status; + __entry->addr = m->addr; + __entry->misc = m->misc; + __entry->ip = m->ip; + __entry->cs = m->cs; + __entry->tsc = m->tsc; + __entry->walltime = m->time; + __entry->cpu = m->extcpu; + __entry->cpuid = m->cpuid; + __entry->apicid = m->apicid; + __entry->socketid = m->socketid; + __entry->cpuvendor = m->cpuvendor; + ), + + TP_printk("CPU: %d, MCGc/s: %llx/%llx, MC%d: %016Lx, ADDR/MISC: %016Lx/%016Lx, RIP: %02x:<%016Lx>, TSC: %llx, PROCESSOR: %u:%x, TIME: %llu, SOCKET: %u, APIC: %x", + __entry->cpu, + __entry->mcgcap, __entry->mcgstatus, + __entry->bank, __entry->status, + __entry->addr, __entry->misc, + __entry->cs, __entry->ip, + __entry->tsc, + __entry->cpuvendor, __entry->cpuid, + __entry->walltime, + __entry->socketid, + __entry->apicid) +); + +#endif /* _TRACE_MCE_H */ + +/* This part must be outside protection */ +#include <trace/define_trace.h> diff --git a/include/trace/events/module.h b/include/trace/events/module.h index 84160fb1847..4b0f48ba16a 100644 --- a/include/trace/events/module.h +++ b/include/trace/events/module.h @@ -51,7 +51,7 @@ TRACE_EVENT(module_free, TP_printk("%s", __get_str(name)) ); -TRACE_EVENT(module_get, +DECLARE_EVENT_CLASS(module_refcnt, TP_PROTO(struct module *mod, unsigned long ip, int refcnt), @@ -73,26 +73,18 @@ TRACE_EVENT(module_get, __get_str(name), (void *)__entry->ip, __entry->refcnt) ); -TRACE_EVENT(module_put, +DEFINE_EVENT(module_refcnt, module_get, TP_PROTO(struct module *mod, unsigned long ip, int refcnt), - TP_ARGS(mod, ip, refcnt), + TP_ARGS(mod, ip, refcnt) +); - TP_STRUCT__entry( - __field( unsigned long, ip ) - __field( int, refcnt ) - __string( name, mod->name ) - ), +DEFINE_EVENT(module_refcnt, module_put, - TP_fast_assign( - __entry->ip = ip; - __entry->refcnt = refcnt; - __assign_str(name, mod->name); - ), + TP_PROTO(struct module *mod, unsigned long ip, int refcnt), - TP_printk("%s call_site=%pf refcnt=%d", - __get_str(name), (void *)__entry->ip, __entry->refcnt) + TP_ARGS(mod, ip, refcnt) ); TRACE_EVENT(module_request, diff --git a/include/trace/events/power.h b/include/trace/events/power.h index 9bb96e5a284..c4efe9b8280 100644 --- a/include/trace/events/power.h +++ b/include/trace/events/power.h @@ -16,7 +16,7 @@ enum { }; #endif -TRACE_EVENT(power_start, +DECLARE_EVENT_CLASS(power, TP_PROTO(unsigned int type, unsigned int state), @@ -35,42 +35,36 @@ TRACE_EVENT(power_start, TP_printk("type=%lu state=%lu", (unsigned long)__entry->type, (unsigned long)__entry->state) ); -TRACE_EVENT(power_end, - - TP_PROTO(int dummy), +DEFINE_EVENT(power, power_start, - TP_ARGS(dummy), + TP_PROTO(unsigned int type, unsigned int state), - TP_STRUCT__entry( - __field( u64, dummy ) - ), + TP_ARGS(type, state) +); - TP_fast_assign( - __entry->dummy = 0xffff; - ), +DEFINE_EVENT(power, power_frequency, - TP_printk("dummy=%lu", (unsigned long)__entry->dummy) + TP_PROTO(unsigned int type, unsigned int state), + TP_ARGS(type, state) ); +TRACE_EVENT(power_end, -TRACE_EVENT(power_frequency, - - TP_PROTO(unsigned int type, unsigned int state), + TP_PROTO(int dummy), - TP_ARGS(type, state), + TP_ARGS(dummy), TP_STRUCT__entry( - __field( u64, type ) - __field( u64, state ) + __field( u64, dummy ) ), TP_fast_assign( - __entry->type = type; - __entry->state = state; + __entry->dummy = 0xffff; ), - TP_printk("type=%lu state=%lu", (unsigned long)__entry->type, (unsigned long) __entry->state) + TP_printk("dummy=%lu", (unsigned long)__entry->dummy) + ); #endif /* _TRACE_POWER_H */ diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h index b50b9856c59..cfceb0b73e2 100644 --- a/include/trace/events/sched.h +++ b/include/trace/events/sched.h @@ -83,7 +83,7 @@ TRACE_EVENT(sched_wait_task, * (NOTE: the 'rq' argument is not used by generic trace events, * but used by the latency tracer plugin. ) */ -TRACE_EVENT(sched_wakeup, +DECLARE_EVENT_CLASS(sched_wakeup_template, TP_PROTO(struct rq *rq, struct task_struct *p, int success), @@ -110,38 +110,19 @@ TRACE_EVENT(sched_wakeup, __entry->success, __entry->target_cpu) ); +DEFINE_EVENT(sched_wakeup_template, sched_wakeup, + TP_PROTO(struct rq *rq, struct task_struct *p, int success), + TP_ARGS(rq, p, success)); + /* * Tracepoint for waking up a new task: * * (NOTE: the 'rq' argument is not used by generic trace events, * but used by the latency tracer plugin. ) */ -TRACE_EVENT(sched_wakeup_new, - - TP_PROTO(struct rq *rq, struct task_struct *p, int success), - - TP_ARGS(rq, p, success), - - TP_STRUCT__entry( - __array( char, comm, TASK_COMM_LEN ) - __field( pid_t, pid ) - __field( int, prio ) - __field( int, success ) - __field( int, target_cpu ) - ), - - TP_fast_assign( - memcpy(__entry->comm, p->comm, TASK_COMM_LEN); - __entry->pid = p->pid; - __entry->prio = p->prio; - __entry->success = success; - __entry->target_cpu = task_cpu(p); - ), - - TP_printk("comm=%s pid=%d prio=%d success=%d target_cpu=%03d", - __entry->comm, __entry->pid, __entry->prio, - __entry->success, __entry->target_cpu) -); +DEFINE_EVENT(sched_wakeup_template, sched_wakeup_new, + TP_PROTO(struct rq *rq, struct task_struct *p, int success), + TP_ARGS(rq, p, success)); /* * Tracepoint for task switches, performed by the scheduler: @@ -216,10 +197,7 @@ TRACE_EVENT(sched_migrate_task, __entry->orig_cpu, __entry->dest_cpu) ); -/* - * Tracepoint for freeing a task: - */ -TRACE_EVENT(sched_process_free, +DECLARE_EVENT_CLASS(sched_process_template, TP_PROTO(struct task_struct *p), @@ -242,29 +220,19 @@ TRACE_EVENT(sched_process_free, ); /* - * Tracepoint for a task exiting: + * Tracepoint for freeing a task: */ -TRACE_EVENT(sched_process_exit, +DEFINE_EVENT(sched_process_template, sched_process_free, + TP_PROTO(struct task_struct *p), + TP_ARGS(p)); + - TP_PROTO(struct task_struct *p), - - TP_ARGS(p), - - TP_STRUCT__entry( - __array( char, comm, TASK_COMM_LEN ) - __field( pid_t, pid ) - __field( int, prio ) - ), - - TP_fast_assign( - memcpy(__entry->comm, p->comm, TASK_COMM_LEN); - __entry->pid = p->pid; - __entry->prio = p->prio; - ), - - TP_printk("comm=%s pid=%d prio=%d", - __entry->comm, __entry->pid, __entry->prio) -); +/* + * Tracepoint for a task exiting: + */ +DEFINE_EVENT(sched_process_template, sched_process_exit, + TP_PROTO(struct task_struct *p), + TP_ARGS(p)); /* * Tracepoint for a waiting task: @@ -320,40 +288,10 @@ TRACE_EVENT(sched_process_fork, ); /* - * Tracepoint for sending a signal: - */ -TRACE_EVENT(sched_signal_send, - - TP_PROTO(int sig, struct task_struct *p), - - TP_ARGS(sig, p), - - TP_STRUCT__entry( - __field( int, sig ) - __array( char, comm, TASK_COMM_LEN ) - __field( pid_t, pid ) - ), - - TP_fast_assign( - memcpy(__entry->comm, p->comm, TASK_COMM_LEN); - __entry->pid = p->pid; - __entry->sig = sig; - ), - - TP_printk("sig=%d comm=%s pid=%d", - __entry->sig, __entry->comm, __entry->pid) -); - -/* * XXX the below sched_stat tracepoints only apply to SCHED_OTHER/BATCH/IDLE * adding sched_stat support to SCHED_FIFO/RR would be welcome. */ - -/* - * Tracepoint for accounting wait time (time the task is runnable - * but not actually running due to scheduler contention). - */ -TRACE_EVENT(sched_stat_wait, +DECLARE_EVENT_CLASS(sched_stat_template, TP_PROTO(struct task_struct *tsk, u64 delay), @@ -379,6 +317,31 @@ TRACE_EVENT(sched_stat_wait, (unsigned long long)__entry->delay) ); + +/* + * Tracepoint for accounting wait time (time the task is runnable + * but not actually running due to scheduler contention). + */ +DEFINE_EVENT(sched_stat_template, sched_stat_wait, + TP_PROTO(struct task_struct *tsk, u64 delay), + TP_ARGS(tsk, delay)); + +/* + * Tracepoint for accounting sleep time (time the task is not runnable, + * including iowait, see below). + */ +DEFINE_EVENT(sched_stat_template, sched_stat_sleep, + TP_PROTO(struct task_struct *tsk, u64 delay), + TP_ARGS(tsk, delay)); + +/* + * Tracepoint for accounting iowait time (time the task is not runnable + * due to waiting on IO to complete). + */ +DEFINE_EVENT(sched_stat_template, sched_stat_iowait, + TP_PROTO(struct task_struct *tsk, u64 delay), + TP_ARGS(tsk, delay)); + /* * Tracepoint for accounting runtime (time the task is executing * on a CPU). @@ -412,66 +375,6 @@ TRACE_EVENT(sched_stat_runtime, (unsigned long long)__entry->vruntime) ); -/* - * Tracepoint for accounting sleep time (time the task is not runnable, - * including iowait, see below). - */ -TRACE_EVENT(sched_stat_sleep, - - TP_PROTO(struct task_struct *tsk, u64 delay), - - TP_ARGS(tsk, delay), - - TP_STRUCT__entry( - __array( char, comm, TASK_COMM_LEN ) - __field( pid_t, pid ) - __field( u64, delay ) - ), - - TP_fast_assign( - memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN); - __entry->pid = tsk->pid; - __entry->delay = delay; - ) - TP_perf_assign( - __perf_count(delay); - ), - - TP_printk("comm=%s pid=%d delay=%Lu [ns]", - __entry->comm, __entry->pid, - (unsigned long long)__entry->delay) -); - -/* - * Tracepoint for accounting iowait time (time the task is not runnable - * due to waiting on IO to complete). - */ -TRACE_EVENT(sched_stat_iowait, - - TP_PROTO(struct task_struct *tsk, u64 delay), - - TP_ARGS(tsk, delay), - - TP_STRUCT__entry( - __array( char, comm, TASK_COMM_LEN ) - __field( pid_t, pid ) - __field( u64, delay ) - ), - - TP_fast_assign( - memcpy(__entry->comm, tsk->comm, TASK_COMM_LEN); - __entry->pid = tsk->pid; - __entry->delay = delay; - ) - TP_perf_assign( - __perf_count(delay); - ), - - TP_printk("comm=%s pid=%d delay=%Lu [ns]", - __entry->comm, __entry->pid, - (unsigned long long)__entry->delay) -); - #endif /* _TRACE_SCHED_H */ /* This part must be outside protection */ diff --git a/include/trace/events/signal.h b/include/trace/events/signal.h new file mode 100644 index 00000000000..a510b75ac30 --- /dev/null +++ b/include/trace/events/signal.h @@ -0,0 +1,173 @@ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM signal + +#if !defined(_TRACE_SIGNAL_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_SIGNAL_H + +#include <linux/signal.h> +#include <linux/sched.h> +#include <linux/tracepoint.h> + +#define TP_STORE_SIGINFO(__entry, info) \ + do { \ + if (info == SEND_SIG_NOINFO) { \ + __entry->errno = 0; \ + __entry->code = SI_USER; \ + } else if (info == SEND_SIG_PRIV) { \ + __entry->errno = 0; \ + __entry->code = SI_KERNEL; \ + } else { \ + __entry->errno = info->si_errno; \ + __entry->code = info->si_code; \ + } \ + } while (0) + +/** + * signal_generate - called when a signal is generated + * @sig: signal number + * @info: pointer to struct siginfo + * @task: pointer to struct task_struct + * + * Current process sends a 'sig' signal to 'task' process with + * 'info' siginfo. If 'info' is SEND_SIG_NOINFO or SEND_SIG_PRIV, + * 'info' is not a pointer and you can't access its field. Instead, + * SEND_SIG_NOINFO means that si_code is SI_USER, and SEND_SIG_PRIV + * means that si_code is SI_KERNEL. + */ +TRACE_EVENT(signal_generate, + + TP_PROTO(int sig, struct siginfo *info, struct task_struct *task), + + TP_ARGS(sig, info, task), + + TP_STRUCT__entry( + __field( int, sig ) + __field( int, errno ) + __field( int, code ) + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + ), + + TP_fast_assign( + __entry->sig = sig; + TP_STORE_SIGINFO(__entry, info); + memcpy(__entry->comm, task->comm, TASK_COMM_LEN); + __entry->pid = task->pid; + ), + + TP_printk("sig=%d errno=%d code=%d comm=%s pid=%d", + __entry->sig, __entry->errno, __entry->code, + __entry->comm, __entry->pid) +); + +/** + * signal_deliver - called when a signal is delivered + * @sig: signal number + * @info: pointer to struct siginfo + * @ka: pointer to struct k_sigaction + * + * A 'sig' signal is delivered to current process with 'info' siginfo, + * and it will be handled by 'ka'. ka->sa.sa_handler can be SIG_IGN or + * SIG_DFL. + * Note that some signals reported by signal_generate tracepoint can be + * lost, ignored or modified (by debugger) before hitting this tracepoint. + * This means, this can show which signals are actually delivered, but + * matching generated signals and delivered signals may not be correct. + */ +TRACE_EVENT(signal_deliver, + + TP_PROTO(int sig, struct siginfo *info, struct k_sigaction *ka), + + TP_ARGS(sig, info, ka), + + TP_STRUCT__entry( + __field( int, sig ) + __field( int, errno ) + __field( int, code ) + __field( unsigned long, sa_handler ) + __field( unsigned long, sa_flags ) + ), + + TP_fast_assign( + __entry->sig = sig; + TP_STORE_SIGINFO(__entry, info); + __entry->sa_handler = (unsigned long)ka->sa.sa_handler; + __entry->sa_flags = ka->sa.sa_flags; + ), + + TP_printk("sig=%d errno=%d code=%d sa_handler=%lx sa_flags=%lx", + __entry->sig, __entry->errno, __entry->code, + __entry->sa_handler, __entry->sa_flags) +); + +/** + * signal_overflow_fail - called when signal queue is overflow + * @sig: signal number + * @group: signal to process group or not (bool) + * @info: pointer to struct siginfo + * + * Kernel fails to generate 'sig' signal with 'info' siginfo, because + * siginfo queue is overflow, and the signal is dropped. + * 'group' is not 0 if the signal will be sent to a process group. + * 'sig' is always one of RT signals. + */ +TRACE_EVENT(signal_overflow_fail, + + TP_PROTO(int sig, int group, struct siginfo *info), + + TP_ARGS(sig, group, info), + + TP_STRUCT__entry( + __field( int, sig ) + __field( int, group ) + __field( int, errno ) + __field( int, code ) + ), + + TP_fast_assign( + __entry->sig = sig; + __entry->group = group; + TP_STORE_SIGINFO(__entry, info); + ), + + TP_printk("sig=%d group=%d errno=%d code=%d", + __entry->sig, __entry->group, __entry->errno, __entry->code) +); + +/** + * signal_lose_info - called when siginfo is lost + * @sig: signal number + * @group: signal to process group or not (bool) + * @info: pointer to struct siginfo + * + * Kernel generates 'sig' signal but loses 'info' siginfo, because siginfo + * queue is overflow. + * 'group' is not 0 if the signal will be sent to a process group. + * 'sig' is always one of non-RT signals. + */ +TRACE_EVENT(signal_lose_info, + + TP_PROTO(int sig, int group, struct siginfo *info), + + TP_ARGS(sig, group, info), + + TP_STRUCT__entry( + __field( int, sig ) + __field( int, group ) + __field( int, errno ) + __field( int, code ) + ), + + TP_fast_assign( + __entry->sig = sig; + __entry->group = group; + TP_STORE_SIGINFO(__entry, info); + ), + + TP_printk("sig=%d group=%d errno=%d code=%d", + __entry->sig, __entry->group, __entry->errno, __entry->code) +); +#endif /* _TRACE_SIGNAL_H */ + +/* This part must be outside protection */ +#include <trace/define_trace.h> diff --git a/include/trace/events/workqueue.h b/include/trace/events/workqueue.h index e4612dbd7ba..d6c974474e7 100644 --- a/include/trace/events/workqueue.h +++ b/include/trace/events/workqueue.h @@ -8,7 +8,7 @@ #include <linux/sched.h> #include <linux/tracepoint.h> -TRACE_EVENT(workqueue_insertion, +DECLARE_EVENT_CLASS(workqueue, TP_PROTO(struct task_struct *wq_thread, struct work_struct *work), @@ -30,26 +30,18 @@ TRACE_EVENT(workqueue_insertion, __entry->thread_pid, __entry->func) ); -TRACE_EVENT(workqueue_execution, +DEFINE_EVENT(workqueue, workqueue_insertion, TP_PROTO(struct task_struct *wq_thread, struct work_struct *work), - TP_ARGS(wq_thread, work), + TP_ARGS(wq_thread, work) +); - TP_STRUCT__entry( - __array(char, thread_comm, TASK_COMM_LEN) - __field(pid_t, thread_pid) - __field(work_func_t, func) - ), +DEFINE_EVENT(workqueue, workqueue_execution, - TP_fast_assign( - memcpy(__entry->thread_comm, wq_thread->comm, TASK_COMM_LEN); - __entry->thread_pid = wq_thread->pid; - __entry->func = work->func; - ), + TP_PROTO(struct task_struct *wq_thread, struct work_struct *work), - TP_printk("thread=%s:%d func=%pf", __entry->thread_comm, - __entry->thread_pid, __entry->func) + TP_ARGS(wq_thread, work) ); /* Trace the creation of one workqueue thread on a cpu */ diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h index 4945d1c9986..2c9c073e45a 100644 --- a/include/trace/ftrace.h +++ b/include/trace/ftrace.h @@ -18,6 +18,26 @@ #include <linux/ftrace_event.h> +/* + * DECLARE_EVENT_CLASS can be used to add a generic function + * handlers for events. That is, if all events have the same + * parameters and just have distinct trace points. + * Each tracepoint can be defined with DEFINE_EVENT and that + * will map the DECLARE_EVENT_CLASS to the tracepoint. + * + * TRACE_EVENT is a one to one mapping between tracepoint and template. + */ +#undef TRACE_EVENT +#define TRACE_EVENT(name, proto, args, tstruct, assign, print) \ + DECLARE_EVENT_CLASS(name, \ + PARAMS(proto), \ + PARAMS(args), \ + PARAMS(tstruct), \ + PARAMS(assign), \ + PARAMS(print)); \ + DEFINE_EVENT(name, name, PARAMS(proto), PARAMS(args)); + + #undef __field #define __field(type, item) type item; @@ -36,15 +56,21 @@ #undef TP_STRUCT__entry #define TP_STRUCT__entry(args...) args -#undef TRACE_EVENT -#define TRACE_EVENT(name, proto, args, tstruct, assign, print) \ - struct ftrace_raw_##name { \ - struct trace_entry ent; \ - tstruct \ - char __data[0]; \ - }; \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(name, proto, args, tstruct, assign, print) \ + struct ftrace_raw_##name { \ + struct trace_entry ent; \ + tstruct \ + char __data[0]; \ + }; +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, name, proto, args) \ static struct ftrace_event_call event_##name +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) + #undef __cpparg #define __cpparg(arg...) arg @@ -89,12 +115,19 @@ #undef __string #define __string(item, src) __dynamic_array(char, item, -1) -#undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ struct ftrace_data_offsets_##call { \ tstruct; \ }; +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, name, proto, args) + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) + #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) /* @@ -170,17 +203,50 @@ #undef TP_perf_assign #define TP_perf_assign(args...) -#undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, func, print) \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, func, print) \ static int \ -ftrace_format_##call(struct ftrace_event_call *unused, \ - struct trace_seq *s) \ +ftrace_format_setup_##call(struct ftrace_event_call *unused, \ + struct trace_seq *s) \ { \ struct ftrace_raw_##call field __attribute__((unused)); \ int ret = 0; \ \ tstruct; \ \ + return ret; \ +} \ + \ +static int \ +ftrace_format_##call(struct ftrace_event_call *unused, \ + struct trace_seq *s) \ +{ \ + int ret = 0; \ + \ + ret = ftrace_format_setup_##call(unused, s); \ + if (!ret) \ + return ret; \ + \ + ret = trace_seq_printf(s, "\nprint fmt: " print); \ + \ + return ret; \ +} + +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, name, proto, args) + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ +static int \ +ftrace_format_##name(struct ftrace_event_call *unused, \ + struct trace_seq *s) \ +{ \ + int ret = 0; \ + \ + ret = ftrace_format_setup_##template(unused, s); \ + if (!ret) \ + return ret; \ + \ trace_seq_printf(s, "\nprint fmt: " print); \ \ return ret; \ @@ -255,10 +321,11 @@ ftrace_format_##call(struct ftrace_event_call *unused, \ ftrace_print_symbols_seq(p, value, symbols); \ }) -#undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ static enum print_line_t \ -ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ +ftrace_raw_output_id_##call(int event_id, const char *name, \ + struct trace_iterator *iter, int flags) \ { \ struct trace_seq *s = &iter->seq; \ struct ftrace_raw_##call *field; \ @@ -268,6 +335,47 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ \ entry = iter->ent; \ \ + if (entry->type != event_id) { \ + WARN_ON_ONCE(1); \ + return TRACE_TYPE_UNHANDLED; \ + } \ + \ + field = (typeof(field))entry; \ + \ + p = &get_cpu_var(ftrace_event_seq); \ + trace_seq_init(p); \ + ret = trace_seq_printf(s, "%s: ", name); \ + if (ret) \ + ret = trace_seq_printf(s, print); \ + put_cpu(); \ + if (!ret) \ + return TRACE_TYPE_PARTIAL_LINE; \ + \ + return TRACE_TYPE_HANDLED; \ +} + +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, name, proto, args) \ +static enum print_line_t \ +ftrace_raw_output_##name(struct trace_iterator *iter, int flags) \ +{ \ + return ftrace_raw_output_id_##template(event_##name.id, \ + #name, iter, flags); \ +} + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, call, proto, args, print) \ +static enum print_line_t \ +ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ +{ \ + struct trace_seq *s = &iter->seq; \ + struct ftrace_raw_##template *field; \ + struct trace_entry *entry; \ + struct trace_seq *p; \ + int ret; \ + \ + entry = iter->ent; \ + \ if (entry->type != event_##call.id) { \ WARN_ON_ONCE(1); \ return TRACE_TYPE_UNHANDLED; \ @@ -277,14 +385,16 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ \ p = &get_cpu_var(ftrace_event_seq); \ trace_seq_init(p); \ - ret = trace_seq_printf(s, #call ": " print); \ + ret = trace_seq_printf(s, "%s: ", #call); \ + if (ret) \ + ret = trace_seq_printf(s, print); \ put_cpu(); \ if (!ret) \ return TRACE_TYPE_PARTIAL_LINE; \ \ return TRACE_TYPE_HANDLED; \ } - + #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) #undef __field_ext @@ -318,8 +428,8 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ #undef __string #define __string(item, src) __dynamic_array(char, item, -1) -#undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, func, print) \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, func, print) \ static int \ ftrace_define_fields_##call(struct ftrace_event_call *event_call) \ { \ @@ -335,6 +445,13 @@ ftrace_define_fields_##call(struct ftrace_event_call *event_call) \ return ret; \ } +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, name, proto, args) + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) + #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) /* @@ -361,10 +478,10 @@ ftrace_define_fields_##call(struct ftrace_event_call *event_call) \ __data_size += (len) * sizeof(type); #undef __string -#define __string(item, src) __dynamic_array(char, item, strlen(src) + 1) \ +#define __string(item, src) __dynamic_array(char, item, strlen(src) + 1) -#undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ static inline int ftrace_get_offsets_##call( \ struct ftrace_data_offsets_##call *__data_offsets, proto) \ { \ @@ -376,6 +493,13 @@ static inline int ftrace_get_offsets_##call( \ return __data_size; \ } +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, name, proto, args) + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) + #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) #ifdef CONFIG_EVENT_PROFILE @@ -397,21 +521,28 @@ static inline int ftrace_get_offsets_##call( \ * */ -#undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) + +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, name, proto, args) \ \ -static void ftrace_profile_##call(proto); \ +static void ftrace_profile_##name(proto); \ \ -static int ftrace_profile_enable_##call(struct ftrace_event_call *unused)\ +static int ftrace_profile_enable_##name(struct ftrace_event_call *unused)\ { \ - return register_trace_##call(ftrace_profile_##call); \ + return register_trace_##name(ftrace_profile_##name); \ } \ \ -static void ftrace_profile_disable_##call(struct ftrace_event_call *unused)\ +static void ftrace_profile_disable_##name(struct ftrace_event_call *unused)\ { \ - unregister_trace_##call(ftrace_profile_##call); \ + unregister_trace_##name(ftrace_profile_##name); \ } +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) + #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) #endif @@ -550,15 +681,13 @@ static void ftrace_profile_disable_##call(struct ftrace_event_call *unused)\ #define __assign_str(dst, src) \ strcpy(__get_str(dst), src); -#undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ \ -static struct ftrace_event_call event_##call; \ - \ -static void ftrace_raw_event_##call(proto) \ +static void ftrace_raw_event_id_##call(struct ftrace_event_call *event_call, \ + proto) \ { \ struct ftrace_data_offsets_##call __maybe_unused __data_offsets;\ - struct ftrace_event_call *event_call = &event_##call; \ struct ring_buffer_event *event; \ struct ftrace_raw_##call *entry; \ struct ring_buffer *buffer; \ @@ -572,7 +701,7 @@ static void ftrace_raw_event_##call(proto) \ __data_size = ftrace_get_offsets_##call(&__data_offsets, args); \ \ event = trace_current_buffer_lock_reserve(&buffer, \ - event_##call.id, \ + event_call->id, \ sizeof(*entry) + __data_size, \ irq_flags, pc); \ if (!event) \ @@ -587,6 +716,14 @@ static void ftrace_raw_event_##call(proto) \ if (!filter_current_check_discard(buffer, event_call, entry, event)) \ trace_nowake_buffer_unlock_commit(buffer, \ event, irq_flags, pc); \ +} + +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, call, proto, args) \ + \ +static void ftrace_raw_event_##call(proto) \ +{ \ + ftrace_raw_event_id_##template(&event_##call, args); \ } \ \ static int ftrace_raw_reg_event_##call(struct ftrace_event_call *unused)\ @@ -619,7 +756,36 @@ static int ftrace_raw_init_event_##call(struct ftrace_event_call *unused)\ event_##call.id = id; \ INIT_LIST_HEAD(&event_##call.fields); \ return 0; \ -} \ +} + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) + +#include TRACE_INCLUDE(TRACE_INCLUDE_FILE) + +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) + +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, call, proto, args) \ + \ +static struct ftrace_event_call __used \ +__attribute__((__aligned__(4))) \ +__attribute__((section("_ftrace_events"))) event_##call = { \ + .name = #call, \ + .system = __stringify(TRACE_SYSTEM), \ + .event = &ftrace_event_type_##call, \ + .raw_init = ftrace_raw_init_event_##call, \ + .regfunc = ftrace_raw_reg_event_##call, \ + .unregfunc = ftrace_raw_unreg_event_##call, \ + .show_format = ftrace_format_##template, \ + .define_fields = ftrace_define_fields_##template, \ + _TRACE_PROFILE_INIT(call) \ +} + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, call, proto, args, print) \ \ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ @@ -631,7 +797,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .regfunc = ftrace_raw_reg_event_##call, \ .unregfunc = ftrace_raw_unreg_event_##call, \ .show_format = ftrace_format_##call, \ - .define_fields = ftrace_define_fields_##call, \ + .define_fields = ftrace_define_fields_##template, \ _TRACE_PROFILE_INIT(call) \ } @@ -719,22 +885,26 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ #undef __perf_count #define __perf_count(c) __count = (c) -#undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ -static void ftrace_profile_##call(proto) \ +#undef DECLARE_EVENT_CLASS +#define DECLARE_EVENT_CLASS(call, proto, args, tstruct, assign, print) \ +static void \ +ftrace_profile_templ_##call(struct ftrace_event_call *event_call, \ + proto) \ { \ struct ftrace_data_offsets_##call __maybe_unused __data_offsets;\ - struct ftrace_event_call *event_call = &event_##call; \ + extern int perf_swevent_get_recursion_context(void); \ + extern void perf_swevent_put_recursion_context(int rctx); \ extern void perf_tp_event(int, u64, u64, void *, int); \ struct ftrace_raw_##call *entry; \ - struct perf_trace_buf *trace_buf; \ u64 __addr = 0, __count = 1; \ unsigned long irq_flags; \ struct trace_entry *ent; \ int __entry_size; \ int __data_size; \ + char *trace_buf; \ char *raw_data; \ int __cpu; \ + int rctx; \ int pc; \ \ pc = preempt_count(); \ @@ -749,6 +919,11 @@ static void ftrace_profile_##call(proto) \ return; \ \ local_irq_save(irq_flags); \ + \ + rctx = perf_swevent_get_recursion_context(); \ + if (rctx < 0) \ + goto end_recursion; \ + \ __cpu = smp_processor_id(); \ \ if (in_nmi()) \ @@ -759,13 +934,7 @@ static void ftrace_profile_##call(proto) \ if (!trace_buf) \ goto end; \ \ - trace_buf = per_cpu_ptr(trace_buf, __cpu); \ - if (trace_buf->recursion++) \ - goto end_recursion; \ - \ - barrier(); \ - \ - raw_data = trace_buf->buf; \ + raw_data = per_cpu_ptr(trace_buf, __cpu); \ \ *(u64 *)(&raw_data[__entry_size - sizeof(u64)]) = 0ULL; \ entry = (struct ftrace_raw_##call *)raw_data; \ @@ -780,13 +949,26 @@ static void ftrace_profile_##call(proto) \ perf_tp_event(event_call->id, __addr, __count, entry, \ __entry_size); \ \ -end_recursion: \ - trace_buf->recursion--; \ end: \ + perf_swevent_put_recursion_context(rctx); \ +end_recursion: \ local_irq_restore(irq_flags); \ \ } +#undef DEFINE_EVENT +#define DEFINE_EVENT(template, call, proto, args) \ +static void ftrace_profile_##call(proto) \ +{ \ + struct ftrace_event_call *event_call = &event_##call; \ + \ + ftrace_profile_templ_##template(event_call, args); \ +} + +#undef DEFINE_EVENT_PRINT +#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \ + DEFINE_EVENT(template, name, PARAMS(proto), PARAMS(args)) + #include TRACE_INCLUDE(TRACE_INCLUDE_FILE) #endif /* CONFIG_EVENT_PROFILE */ diff --git a/include/trace/syscall.h b/include/trace/syscall.h index 51ee17d3632..961fda3556b 100644 --- a/include/trace/syscall.h +++ b/include/trace/syscall.h @@ -12,21 +12,19 @@ * A syscall entry in the ftrace syscalls array. * * @name: name of the syscall + * @syscall_nr: number of the syscall * @nb_args: number of parameters it takes * @types: list of types as strings * @args: list of args as strings (args[i] matches types[i]) - * @enter_id: associated ftrace enter event id - * @exit_id: associated ftrace exit event id * @enter_event: associated syscall_enter trace event * @exit_event: associated syscall_exit trace event */ struct syscall_metadata { const char *name; + int syscall_nr; int nb_args; const char **types; const char **args; - int enter_id; - int exit_id; struct ftrace_event_call *enter_event; struct ftrace_event_call *exit_event; @@ -34,11 +32,7 @@ struct syscall_metadata { #ifdef CONFIG_FTRACE_SYSCALLS extern unsigned long arch_syscall_addr(int nr); -extern int syscall_name_to_nr(char *name); -void set_syscall_enter_id(int num, int id); -void set_syscall_exit_id(int num, int id); -extern struct trace_event event_syscall_enter; -extern struct trace_event event_syscall_exit; +extern int init_syscall_trace(struct ftrace_event_call *call); extern int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s); @@ -56,10 +50,10 @@ enum print_line_t print_syscall_enter(struct trace_iterator *iter, int flags); enum print_line_t print_syscall_exit(struct trace_iterator *iter, int flags); #endif #ifdef CONFIG_EVENT_PROFILE -int reg_prof_syscall_enter(char *name); -void unreg_prof_syscall_enter(char *name); -int reg_prof_syscall_exit(char *name); -void unreg_prof_syscall_exit(char *name); +int prof_sysenter_enable(struct ftrace_event_call *call); +void prof_sysenter_disable(struct ftrace_event_call *call); +int prof_sysexit_enable(struct ftrace_event_call *call); +void prof_sysexit_disable(struct ftrace_event_call *call); #endif diff --git a/kernel/Makefile b/kernel/Makefile index b8d4cd8ac0b..6b7ce8173df 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -21,6 +21,7 @@ CFLAGS_REMOVE_mutex-debug.o = -pg CFLAGS_REMOVE_rtmutex-debug.o = -pg CFLAGS_REMOVE_cgroup-debug.o = -pg CFLAGS_REMOVE_sched_clock.o = -pg +CFLAGS_REMOVE_perf_event.o = -pg endif obj-$(CONFIG_FREEZER) += freezer.o @@ -95,6 +96,7 @@ obj-$(CONFIG_RING_BUFFER) += trace/ obj-$(CONFIG_SMP) += sched_cpupri.o obj-$(CONFIG_SLOW_WORK) += slow-work.o obj-$(CONFIG_PERF_EVENTS) += perf_event.o +obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) # According to Alan Modra <alan@linuxcare.com.au>, the -fno-omit-frame-pointer is diff --git a/kernel/exit.c b/kernel/exit.c index f7864ac2ecc..3f45e3cf931 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -49,6 +49,7 @@ #include <linux/init_task.h> #include <linux/perf_event.h> #include <trace/events/sched.h> +#include <linux/hw_breakpoint.h> #include <asm/uaccess.h> #include <asm/unistd.h> @@ -978,6 +979,10 @@ NORET_TYPE void do_exit(long code) proc_exit_connector(tsk); /* + * FIXME: do that only when needed, using sched_exit tracepoint + */ + flush_ptrace_hw_breakpoint(tsk); + /* * Flush inherited counters to the parent - before the parent * gets woken up by child-exit notifications. */ diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c new file mode 100644 index 00000000000..cf5ee162841 --- /dev/null +++ b/kernel/hw_breakpoint.c @@ -0,0 +1,423 @@ +/* + * 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) 2007 Alan Stern + * Copyright (C) IBM Corporation, 2009 + * Copyright (C) 2009, Frederic Weisbecker <fweisbec@gmail.com> + * + * Thanks to Ingo Molnar for his many suggestions. + * + * Authors: Alan Stern <stern@rowland.harvard.edu> + * K.Prasad <prasad@linux.vnet.ibm.com> + * Frederic Weisbecker <fweisbec@gmail.com> + */ + +/* + * HW_breakpoint: a unified kernel/user-space hardware breakpoint facility, + * using the CPU's debug registers. + * This file contains the arch-independent routines. + */ + +#include <linux/irqflags.h> +#include <linux/kallsyms.h> +#include <linux/notifier.h> +#include <linux/kprobes.h> +#include <linux/kdebug.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/percpu.h> +#include <linux/sched.h> +#include <linux/init.h> +#include <linux/smp.h> + +#include <linux/hw_breakpoint.h> + +/* + * Constraints data + */ + +/* Number of pinned cpu breakpoints in a cpu */ +static DEFINE_PER_CPU(unsigned int, nr_cpu_bp_pinned); + +/* Number of pinned task breakpoints in a cpu */ +static DEFINE_PER_CPU(unsigned int, task_bp_pinned[HBP_NUM]); + +/* Number of non-pinned cpu/task breakpoints in a cpu */ +static DEFINE_PER_CPU(unsigned int, nr_bp_flexible); + +/* Gather the number of total pinned and un-pinned bp in a cpuset */ +struct bp_busy_slots { + unsigned int pinned; + unsigned int flexible; +}; + +/* Serialize accesses to the above constraints */ +static DEFINE_MUTEX(nr_bp_mutex); + +/* + * Report the maximum number of pinned breakpoints a task + * have in this cpu + */ +static unsigned int max_task_bp_pinned(int cpu) +{ + int i; + unsigned int *tsk_pinned = per_cpu(task_bp_pinned, cpu); + + for (i = HBP_NUM -1; i >= 0; i--) { + if (tsk_pinned[i] > 0) + return i + 1; + } + + return 0; +} + +/* + * Report the number of pinned/un-pinned breakpoints we have in + * a given cpu (cpu > -1) or in all of them (cpu = -1). + */ +static void fetch_bp_busy_slots(struct bp_busy_slots *slots, int cpu) +{ + if (cpu >= 0) { + slots->pinned = per_cpu(nr_cpu_bp_pinned, cpu); + slots->pinned += max_task_bp_pinned(cpu); + slots->flexible = per_cpu(nr_bp_flexible, cpu); + + return; + } + + for_each_online_cpu(cpu) { + unsigned int nr; + + nr = per_cpu(nr_cpu_bp_pinned, cpu); + nr += max_task_bp_pinned(cpu); + + if (nr > slots->pinned) + slots->pinned = nr; + + nr = per_cpu(nr_bp_flexible, cpu); + + if (nr > slots->flexible) + slots->flexible = nr; + } +} + +/* + * Add a pinned breakpoint for the given task in our constraint table + */ +static void toggle_bp_task_slot(struct task_struct *tsk, int cpu, bool enable) +{ + int count = 0; + struct perf_event *bp; + struct perf_event_context *ctx = tsk->perf_event_ctxp; + unsigned int *tsk_pinned; + struct list_head *list; + unsigned long flags; + + if (WARN_ONCE(!ctx, "No perf context for this task")) + return; + + list = &ctx->event_list; + + spin_lock_irqsave(&ctx->lock, flags); + + /* + * The current breakpoint counter is not included in the list + * at the open() callback time + */ + list_for_each_entry(bp, list, event_entry) { + if (bp->attr.type == PERF_TYPE_BREAKPOINT) + count++; + } + + spin_unlock_irqrestore(&ctx->lock, flags); + + if (WARN_ONCE(count < 0, "No breakpoint counter found in the counter list")) + return; + + tsk_pinned = per_cpu(task_bp_pinned, cpu); + if (enable) { + tsk_pinned[count]++; + if (count > 0) + tsk_pinned[count-1]--; + } else { + tsk_pinned[count]--; + if (count > 0) + tsk_pinned[count-1]++; + } +} + +/* + * Add/remove the given breakpoint in our constraint table + */ +static void toggle_bp_slot(struct perf_event *bp, bool enable) +{ + int cpu = bp->cpu; + struct task_struct *tsk = bp->ctx->task; + + /* Pinned counter task profiling */ + if (tsk) { + if (cpu >= 0) { + toggle_bp_task_slot(tsk, cpu, enable); + return; + } + + for_each_online_cpu(cpu) + toggle_bp_task_slot(tsk, cpu, enable); + return; + } + + /* Pinned counter cpu profiling */ + if (enable) + per_cpu(nr_cpu_bp_pinned, bp->cpu)++; + else + per_cpu(nr_cpu_bp_pinned, bp->cpu)--; +} + +/* + * Contraints to check before allowing this new breakpoint counter: + * + * == Non-pinned counter == (Considered as pinned for now) + * + * - If attached to a single cpu, check: + * + * (per_cpu(nr_bp_flexible, cpu) || (per_cpu(nr_cpu_bp_pinned, cpu) + * + max(per_cpu(task_bp_pinned, cpu)))) < HBP_NUM + * + * -> If there are already non-pinned counters in this cpu, it means + * there is already a free slot for them. + * Otherwise, we check that the maximum number of per task + * breakpoints (for this cpu) plus the number of per cpu breakpoint + * (for this cpu) doesn't cover every registers. + * + * - If attached to every cpus, check: + * + * (per_cpu(nr_bp_flexible, *) || (max(per_cpu(nr_cpu_bp_pinned, *)) + * + max(per_cpu(task_bp_pinned, *)))) < HBP_NUM + * + * -> This is roughly the same, except we check the number of per cpu + * bp for every cpu and we keep the max one. Same for the per tasks + * breakpoints. + * + * + * == Pinned counter == + * + * - If attached to a single cpu, check: + * + * ((per_cpu(nr_bp_flexible, cpu) > 1) + per_cpu(nr_cpu_bp_pinned, cpu) + * + max(per_cpu(task_bp_pinned, cpu))) < HBP_NUM + * + * -> Same checks as before. But now the nr_bp_flexible, if any, must keep + * one register at least (or they will never be fed). + * + * - If attached to every cpus, check: + * + * ((per_cpu(nr_bp_flexible, *) > 1) + max(per_cpu(nr_cpu_bp_pinned, *)) + * + max(per_cpu(task_bp_pinned, *))) < HBP_NUM + */ +int reserve_bp_slot(struct perf_event *bp) +{ + struct bp_busy_slots slots = {0}; + int ret = 0; + + mutex_lock(&nr_bp_mutex); + + fetch_bp_busy_slots(&slots, bp->cpu); + + /* Flexible counters need to keep at least one slot */ + if (slots.pinned + (!!slots.flexible) == HBP_NUM) { + ret = -ENOSPC; + goto end; + } + + toggle_bp_slot(bp, true); + +end: + mutex_unlock(&nr_bp_mutex); + + return ret; +} + +void release_bp_slot(struct perf_event *bp) +{ + mutex_lock(&nr_bp_mutex); + + toggle_bp_slot(bp, false); + + mutex_unlock(&nr_bp_mutex); +} + + +int __register_perf_hw_breakpoint(struct perf_event *bp) +{ + int ret; + + ret = reserve_bp_slot(bp); + if (ret) + return ret; + + /* + * Ptrace breakpoints can be temporary perf events only + * meant to reserve a slot. In this case, it is created disabled and + * we don't want to check the params right now (as we put a null addr) + * But perf tools create events as disabled and we want to check + * the params for them. + * This is a quick hack that will be removed soon, once we remove + * the tmp breakpoints from ptrace + */ + if (!bp->attr.disabled || bp->callback == perf_bp_event) + ret = arch_validate_hwbkpt_settings(bp, bp->ctx->task); + + return ret; +} + +int register_perf_hw_breakpoint(struct perf_event *bp) +{ + bp->callback = perf_bp_event; + + return __register_perf_hw_breakpoint(bp); +} + +/** + * register_user_hw_breakpoint - register a hardware breakpoint for user space + * @attr: breakpoint attributes + * @triggered: callback to trigger when we hit the breakpoint + * @tsk: pointer to 'task_struct' of the process to which the address belongs + */ +struct perf_event * +register_user_hw_breakpoint(struct perf_event_attr *attr, + perf_callback_t triggered, + struct task_struct *tsk) +{ + return perf_event_create_kernel_counter(attr, -1, tsk->pid, triggered); +} +EXPORT_SYMBOL_GPL(register_user_hw_breakpoint); + +/** + * modify_user_hw_breakpoint - modify a user-space hardware breakpoint + * @bp: the breakpoint structure to modify + * @attr: new breakpoint attributes + * @triggered: callback to trigger when we hit the breakpoint + * @tsk: pointer to 'task_struct' of the process to which the address belongs + */ +struct perf_event * +modify_user_hw_breakpoint(struct perf_event *bp, struct perf_event_attr *attr, + perf_callback_t triggered, + struct task_struct *tsk) +{ + /* + * FIXME: do it without unregistering + * - We don't want to lose our slot + * - If the new bp is incorrect, don't lose the older one + */ + unregister_hw_breakpoint(bp); + + return perf_event_create_kernel_counter(attr, -1, tsk->pid, triggered); +} +EXPORT_SYMBOL_GPL(modify_user_hw_breakpoint); + +/** + * unregister_hw_breakpoint - unregister a user-space hardware breakpoint + * @bp: the breakpoint structure to unregister + */ +void unregister_hw_breakpoint(struct perf_event *bp) +{ + if (!bp) + return; + perf_event_release_kernel(bp); +} +EXPORT_SYMBOL_GPL(unregister_hw_breakpoint); + +/** + * register_wide_hw_breakpoint - register a wide breakpoint in the kernel + * @attr: breakpoint attributes + * @triggered: callback to trigger when we hit the breakpoint + * + * @return a set of per_cpu pointers to perf events + */ +struct perf_event ** +register_wide_hw_breakpoint(struct perf_event_attr *attr, + perf_callback_t triggered) +{ + struct perf_event **cpu_events, **pevent, *bp; + long err; + int cpu; + + cpu_events = alloc_percpu(typeof(*cpu_events)); + if (!cpu_events) + return ERR_PTR(-ENOMEM); + + for_each_possible_cpu(cpu) { + pevent = per_cpu_ptr(cpu_events, cpu); + bp = perf_event_create_kernel_counter(attr, cpu, -1, triggered); + + *pevent = bp; + + if (IS_ERR(bp)) { + err = PTR_ERR(bp); + goto fail; + } + } + + return cpu_events; + +fail: + for_each_possible_cpu(cpu) { + pevent = per_cpu_ptr(cpu_events, cpu); + if (IS_ERR(*pevent)) + break; + unregister_hw_breakpoint(*pevent); + } + free_percpu(cpu_events); + /* return the error if any */ + return ERR_PTR(err); +} +EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); + +/** + * unregister_wide_hw_breakpoint - unregister a wide breakpoint in the kernel + * @cpu_events: the per cpu set of events to unregister + */ +void unregister_wide_hw_breakpoint(struct perf_event **cpu_events) +{ + int cpu; + struct perf_event **pevent; + + for_each_possible_cpu(cpu) { + pevent = per_cpu_ptr(cpu_events, cpu); + unregister_hw_breakpoint(*pevent); + } + free_percpu(cpu_events); +} +EXPORT_SYMBOL_GPL(unregister_wide_hw_breakpoint); + +static struct notifier_block hw_breakpoint_exceptions_nb = { + .notifier_call = hw_breakpoint_exceptions_notify, + /* we need to be notified first */ + .priority = 0x7fffffff +}; + +static int __init init_hw_breakpoint(void) +{ + return register_die_notifier(&hw_breakpoint_exceptions_nb); +} +core_initcall(init_hw_breakpoint); + + +struct pmu perf_ops_bp = { + .enable = arch_install_hw_breakpoint, + .disable = arch_uninstall_hw_breakpoint, + .read = hw_breakpoint_pmu_read, + .unthrottle = hw_breakpoint_pmu_unthrottle +}; diff --git a/kernel/kallsyms.c b/kernel/kallsyms.c index 8b6b8b697c6..8e5288a8a35 100644 --- a/kernel/kallsyms.c +++ b/kernel/kallsyms.c @@ -181,6 +181,7 @@ unsigned long kallsyms_lookup_name(const char *name) } return module_kallsyms_lookup_name(name); } +EXPORT_SYMBOL_GPL(kallsyms_lookup_name); int kallsyms_on_each_symbol(int (*fn)(void *, const char *, struct module *, unsigned long), diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 3256e36ad25..6b7ddba1dd6 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -29,6 +29,7 @@ #include <linux/kernel_stat.h> #include <linux/perf_event.h> #include <linux/ftrace_event.h> +#include <linux/hw_breakpoint.h> #include <asm/irq_regs.h> @@ -245,6 +246,49 @@ static void perf_unpin_context(struct perf_event_context *ctx) put_ctx(ctx); } +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_event_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 event. + */ +static void update_event_times(struct perf_event *event) +{ + struct perf_event_context *ctx = event->ctx; + u64 run_end; + + if (event->state < PERF_EVENT_STATE_INACTIVE || + event->group_leader->state < PERF_EVENT_STATE_INACTIVE) + return; + + if (ctx->is_active) + run_end = ctx->time; + else + run_end = event->tstamp_stopped; + + event->total_time_enabled = run_end - event->tstamp_enabled; + + if (event->state == PERF_EVENT_STATE_INACTIVE) + run_end = event->tstamp_stopped; + else + run_end = ctx->time; + + event->total_time_running = run_end - event->tstamp_running; +} + /* * Add a event from the lists for its context. * Must be called with ctx->mutex and ctx->lock held. @@ -293,6 +337,18 @@ list_del_event(struct perf_event *event, struct perf_event_context *ctx) if (event->group_leader != event) event->group_leader->nr_siblings--; + update_event_times(event); + + /* + * If event was in error state, then keep it + * that way, otherwise bogus counts will be + * returned on read(). The only way to get out + * of error state is by explicit re-enabling + * of the event + */ + if (event->state > PERF_EVENT_STATE_OFF) + event->state = PERF_EVENT_STATE_OFF; + /* * If this was a group event with sibling events then * upgrade the siblings to singleton events by adding them @@ -446,50 +502,11 @@ retry: * can remove the event safely, if the call above did not * succeed. */ - if (!list_empty(&event->group_entry)) { + if (!list_empty(&event->group_entry)) list_del_event(event, 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_event_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 event. - */ -static void update_event_times(struct perf_event *event) -{ - struct perf_event_context *ctx = event->ctx; - u64 run_end; - - if (event->state < PERF_EVENT_STATE_INACTIVE || - event->group_leader->state < PERF_EVENT_STATE_INACTIVE) - return; - - event->total_time_enabled = ctx->time - event->tstamp_enabled; - - if (event->state == PERF_EVENT_STATE_INACTIVE) - run_end = event->tstamp_stopped; - else - run_end = ctx->time; - - event->total_time_running = run_end - event->tstamp_running; -} - /* * Update total_time_enabled and total_time_running for all events in a group. */ @@ -1032,10 +1049,10 @@ void __perf_event_sched_out(struct perf_event_context *ctx, update_context_time(ctx); perf_disable(); - if (ctx->nr_active) + if (ctx->nr_active) { list_for_each_entry(event, &ctx->group_list, group_entry) group_sched_out(event, cpuctx, ctx); - + } perf_enable(); out: spin_unlock(&ctx->lock); @@ -1060,8 +1077,6 @@ static int context_equiv(struct perf_event_context *ctx1, && !ctx1->pin_count && !ctx2->pin_count; } -static void __perf_event_read(void *event); - static void __perf_event_sync_stat(struct perf_event *event, struct perf_event *next_event) { @@ -1079,8 +1094,8 @@ static void __perf_event_sync_stat(struct perf_event *event, */ switch (event->state) { case PERF_EVENT_STATE_ACTIVE: - __perf_event_read(event); - break; + event->pmu->read(event); + /* fall-through */ case PERF_EVENT_STATE_INACTIVE: update_event_times(event); @@ -1119,6 +1134,8 @@ static void perf_event_sync_stat(struct perf_event_context *ctx, if (!ctx->nr_stat) return; + update_context_time(ctx); + event = list_first_entry(&ctx->event_list, struct perf_event, event_entry); @@ -1162,8 +1179,6 @@ void perf_event_task_sched_out(struct task_struct *task, if (likely(!ctx || !cpuctx->task_ctx)) return; - update_context_time(ctx); - rcu_read_lock(); parent = rcu_dereference(ctx->parent_ctx); next_ctx = next->perf_event_ctxp; @@ -1516,7 +1531,6 @@ static void __perf_event_read(void *info) struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_event *event = info; struct perf_event_context *ctx = event->ctx; - unsigned long flags; /* * If this is a task context, we need to check whether it is @@ -1528,12 +1542,12 @@ static void __perf_event_read(void *info) if (ctx->task && cpuctx->task_ctx != ctx) return; - local_irq_save(flags); - if (ctx->is_active) - update_context_time(ctx); - event->pmu->read(event); + spin_lock(&ctx->lock); + update_context_time(ctx); update_event_times(event); - local_irq_restore(flags); + spin_unlock(&ctx->lock); + + event->pmu->read(event); } static u64 perf_event_read(struct perf_event *event) @@ -1546,7 +1560,13 @@ static u64 perf_event_read(struct perf_event *event) smp_call_function_single(event->oncpu, __perf_event_read, event, 1); } else if (event->state == PERF_EVENT_STATE_INACTIVE) { + struct perf_event_context *ctx = event->ctx; + unsigned long flags; + + spin_lock_irqsave(&ctx->lock, flags); + update_context_time(ctx); update_event_times(event); + spin_unlock_irqrestore(&ctx->lock, flags); } return atomic64_read(&event->count); @@ -1700,16 +1720,10 @@ static void free_event(struct perf_event *event) call_rcu(&event->rcu_head, free_event_rcu); } -/* - * Called when the last reference to the file is gone. - */ -static int perf_release(struct inode *inode, struct file *file) +int perf_event_release_kernel(struct perf_event *event) { - struct perf_event *event = file->private_data; struct perf_event_context *ctx = event->ctx; - file->private_data = NULL; - WARN_ON_ONCE(ctx->parent_ctx); mutex_lock(&ctx->mutex); perf_event_remove_from_context(event); @@ -1724,6 +1738,19 @@ static int perf_release(struct inode *inode, struct file *file) return 0; } +EXPORT_SYMBOL_GPL(perf_event_release_kernel); + +/* + * Called when the last reference to the file is gone. + */ +static int perf_release(struct inode *inode, struct file *file) +{ + struct perf_event *event = file->private_data; + + file->private_data = NULL; + + return perf_event_release_kernel(event); +} static int perf_event_read_size(struct perf_event *event) { @@ -1750,91 +1777,94 @@ static int perf_event_read_size(struct perf_event *event) return size; } -static u64 perf_event_read_value(struct perf_event *event) +u64 perf_event_read_value(struct perf_event *event, u64 *enabled, u64 *running) { struct perf_event *child; u64 total = 0; + *enabled = 0; + *running = 0; + + mutex_lock(&event->child_mutex); total += perf_event_read(event); - list_for_each_entry(child, &event->child_list, child_list) + *enabled += event->total_time_enabled + + atomic64_read(&event->child_total_time_enabled); + *running += event->total_time_running + + atomic64_read(&event->child_total_time_running); + + list_for_each_entry(child, &event->child_list, child_list) { total += perf_event_read(child); + *enabled += child->total_time_enabled; + *running += child->total_time_running; + } + mutex_unlock(&event->child_mutex); return total; } - -static int perf_event_read_entry(struct perf_event *event, - u64 read_format, char __user *buf) -{ - int n = 0, count = 0; - u64 values[2]; - - values[n++] = perf_event_read_value(event); - if (read_format & PERF_FORMAT_ID) - values[n++] = primary_event_id(event); - - count = n * sizeof(u64); - - if (copy_to_user(buf, values, count)) - return -EFAULT; - - return count; -} +EXPORT_SYMBOL_GPL(perf_event_read_value); static int perf_event_read_group(struct perf_event *event, u64 read_format, char __user *buf) { struct perf_event *leader = event->group_leader, *sub; - int n = 0, size = 0, err = -EFAULT; - u64 values[3]; + int n = 0, size = 0, ret = -EFAULT; + struct perf_event_context *ctx = leader->ctx; + u64 values[5]; + u64 count, enabled, running; + + mutex_lock(&ctx->mutex); + count = perf_event_read_value(leader, &enabled, &running); 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); - } + if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) + values[n++] = enabled; + if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) + values[n++] = running; + values[n++] = count; + if (read_format & PERF_FORMAT_ID) + values[n++] = primary_event_id(leader); size = n * sizeof(u64); if (copy_to_user(buf, values, size)) - return -EFAULT; - - err = perf_event_read_entry(leader, read_format, buf + size); - if (err < 0) - return err; + goto unlock; - size += err; + ret = size; list_for_each_entry(sub, &leader->sibling_list, group_entry) { - err = perf_event_read_entry(sub, read_format, - buf + size); - if (err < 0) - return err; + n = 0; + + values[n++] = perf_event_read_value(sub, &enabled, &running); + if (read_format & PERF_FORMAT_ID) + values[n++] = primary_event_id(sub); + + size = n * sizeof(u64); + + if (copy_to_user(buf + ret, values, size)) { + ret = -EFAULT; + goto unlock; + } - size += err; + ret += size; } +unlock: + mutex_unlock(&ctx->mutex); - return size; + return ret; } static int perf_event_read_one(struct perf_event *event, u64 read_format, char __user *buf) { + u64 enabled, running; u64 values[4]; int n = 0; - values[n++] = perf_event_read_value(event); - if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) { - values[n++] = event->total_time_enabled + - atomic64_read(&event->child_total_time_enabled); - } - if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) { - values[n++] = event->total_time_running + - atomic64_read(&event->child_total_time_running); - } + values[n++] = perf_event_read_value(event, &enabled, &running); + if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED) + values[n++] = enabled; + if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) + values[n++] = running; if (read_format & PERF_FORMAT_ID) values[n++] = primary_event_id(event); @@ -1865,12 +1895,10 @@ perf_read_hw(struct perf_event *event, char __user *buf, size_t count) return -ENOSPC; WARN_ON_ONCE(event->ctx->parent_ctx); - mutex_lock(&event->child_mutex); if (read_format & PERF_FORMAT_GROUP) ret = perf_event_read_group(event, read_format, buf); else ret = perf_event_read_one(event, read_format, buf); - mutex_unlock(&event->child_mutex); return ret; } @@ -2182,6 +2210,7 @@ static void perf_mmap_data_free(struct perf_mmap_data *data) 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); } #else @@ -2222,6 +2251,7 @@ static void perf_mmap_data_free_work(struct work_struct *work) perf_mmap_unmark_page(base + (i * PAGE_SIZE)); vfree(base); + kfree(data); } static void perf_mmap_data_free(struct perf_mmap_data *data) @@ -2315,7 +2345,7 @@ perf_mmap_data_init(struct perf_event *event, struct perf_mmap_data *data) } if (!data->watermark) - data->watermark = max_t(long, PAGE_SIZE, max_size / 2); + data->watermark = max_size / 2; rcu_assign_pointer(event->data, data); @@ -2327,7 +2357,6 @@ static void perf_mmap_data_free_rcu(struct rcu_head *rcu_head) data = container_of(rcu_head, struct perf_mmap_data, rcu_head); perf_mmap_data_free(data); - kfree(data); } static void perf_mmap_data_release(struct perf_event *event) @@ -3245,15 +3274,10 @@ static void perf_event_task_ctx(struct perf_event_context *ctx, { struct perf_event *event; - if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list)) - return; - - rcu_read_lock(); list_for_each_entry_rcu(event, &ctx->event_list, event_entry) { if (perf_event_task_match(event)) perf_event_task_output(event, task_event); } - rcu_read_unlock(); } static void perf_event_task_event(struct perf_task_event *task_event) @@ -3261,11 +3285,11 @@ static void perf_event_task_event(struct perf_task_event *task_event) struct perf_cpu_context *cpuctx; struct perf_event_context *ctx = task_event->task_ctx; + rcu_read_lock(); cpuctx = &get_cpu_var(perf_cpu_context); perf_event_task_ctx(&cpuctx->ctx, task_event); put_cpu_var(perf_cpu_context); - rcu_read_lock(); if (!ctx) ctx = rcu_dereference(task_event->task->perf_event_ctxp); if (ctx) @@ -3357,15 +3381,10 @@ static void perf_event_comm_ctx(struct perf_event_context *ctx, { struct perf_event *event; - if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list)) - return; - - rcu_read_lock(); list_for_each_entry_rcu(event, &ctx->event_list, event_entry) { if (perf_event_comm_match(event)) perf_event_comm_output(event, comm_event); } - rcu_read_unlock(); } static void perf_event_comm_event(struct perf_comm_event *comm_event) @@ -3376,7 +3395,7 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event) char comm[TASK_COMM_LEN]; memset(comm, 0, sizeof(comm)); - strncpy(comm, comm_event->task->comm, sizeof(comm)); + strlcpy(comm, comm_event->task->comm, sizeof(comm)); size = ALIGN(strlen(comm)+1, sizeof(u64)); comm_event->comm = comm; @@ -3384,11 +3403,11 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event) comm_event->event_id.header.size = sizeof(comm_event->event_id) + size; + rcu_read_lock(); cpuctx = &get_cpu_var(perf_cpu_context); perf_event_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. @@ -3481,15 +3500,10 @@ static void perf_event_mmap_ctx(struct perf_event_context *ctx, { struct perf_event *event; - if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list)) - return; - - rcu_read_lock(); list_for_each_entry_rcu(event, &ctx->event_list, event_entry) { if (perf_event_mmap_match(event, mmap_event)) perf_event_mmap_output(event, mmap_event); } - rcu_read_unlock(); } static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) @@ -3545,11 +3559,11 @@ got_name: mmap_event->event_id.header.size = sizeof(mmap_event->event_id) + size; + rcu_read_lock(); cpuctx = &get_cpu_var(perf_cpu_context); perf_event_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. @@ -3688,7 +3702,11 @@ static int __perf_event_overflow(struct perf_event *event, int nmi, perf_event_disable(event); } - perf_event_output(event, nmi, data, regs); + if (event->overflow_handler) + event->overflow_handler(event, nmi, data, regs); + else + perf_event_output(event, nmi, data, regs); + return ret; } @@ -3733,16 +3751,16 @@ again: return nr; } -static void perf_swevent_overflow(struct perf_event *event, +static void perf_swevent_overflow(struct perf_event *event, u64 overflow, int nmi, struct perf_sample_data *data, struct pt_regs *regs) { struct hw_perf_event *hwc = &event->hw; int throttle = 0; - u64 overflow; data->period = event->hw.last_period; - overflow = perf_swevent_set_period(event); + if (!overflow) + overflow = perf_swevent_set_period(event); if (hwc->interrupts == MAX_INTERRUPTS) return; @@ -3775,14 +3793,19 @@ static void perf_swevent_add(struct perf_event *event, u64 nr, atomic64_add(nr, &event->count); + if (!regs) + return; + if (!hwc->sample_period) return; - if (!regs) + if (nr == 1 && hwc->sample_period == 1 && !event->attr.freq) + return perf_swevent_overflow(event, 1, nmi, data, regs); + + if (atomic64_add_negative(nr, &hwc->period_left)) return; - if (!atomic64_add_negative(nr, &hwc->period_left)) - perf_swevent_overflow(event, nmi, data, regs); + perf_swevent_overflow(event, 0, nmi, data, regs); } static int perf_swevent_is_counting(struct perf_event *event) @@ -3818,6 +3841,20 @@ static int perf_swevent_is_counting(struct perf_event *event) static int perf_tp_event_match(struct perf_event *event, struct perf_sample_data *data); +static int perf_exclude_event(struct perf_event *event, + struct pt_regs *regs) +{ + if (regs) { + if (event->attr.exclude_user && user_mode(regs)) + return 1; + + if (event->attr.exclude_kernel && !user_mode(regs)) + return 1; + } + + return 0; +} + static int perf_swevent_match(struct perf_event *event, enum perf_type_id type, u32 event_id, @@ -3829,16 +3866,12 @@ static int perf_swevent_match(struct perf_event *event, if (event->attr.type != type) return 0; + if (event->attr.config != event_id) return 0; - if (regs) { - if (event->attr.exclude_user && user_mode(regs)) - return 0; - - if (event->attr.exclude_kernel && !user_mode(regs)) - return 0; - } + if (perf_exclude_event(event, regs)) + return 0; if (event->attr.type == PERF_TYPE_TRACEPOINT && !perf_tp_event_match(event, data)) @@ -3855,49 +3888,59 @@ static void perf_swevent_ctx_event(struct perf_event_context *ctx, { struct perf_event *event; - if (system_state != SYSTEM_RUNNING || list_empty(&ctx->event_list)) - return; - - rcu_read_lock(); list_for_each_entry_rcu(event, &ctx->event_list, event_entry) { if (perf_swevent_match(event, type, event_id, data, regs)) perf_swevent_add(event, nr, nmi, data, regs); } - rcu_read_unlock(); } -static int *perf_swevent_recursion_context(struct perf_cpu_context *cpuctx) +int perf_swevent_get_recursion_context(void) { + struct perf_cpu_context *cpuctx = &get_cpu_var(perf_cpu_context); + int rctx; + if (in_nmi()) - return &cpuctx->recursion[3]; + rctx = 3; + else if (in_irq()) + rctx = 2; + else if (in_softirq()) + rctx = 1; + else + rctx = 0; - if (in_irq()) - return &cpuctx->recursion[2]; + if (cpuctx->recursion[rctx]) { + put_cpu_var(perf_cpu_context); + return -1; + } - if (in_softirq()) - return &cpuctx->recursion[1]; + cpuctx->recursion[rctx]++; + barrier(); - return &cpuctx->recursion[0]; + return rctx; } +EXPORT_SYMBOL_GPL(perf_swevent_get_recursion_context); + +void perf_swevent_put_recursion_context(int rctx) +{ + struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + barrier(); + cpuctx->recursion[rctx]--; + put_cpu_var(perf_cpu_context); +} +EXPORT_SYMBOL_GPL(perf_swevent_put_recursion_context); static void do_perf_sw_event(enum perf_type_id type, u32 event_id, u64 nr, int nmi, struct perf_sample_data *data, struct pt_regs *regs) { - struct perf_cpu_context *cpuctx = &get_cpu_var(perf_cpu_context); - int *recursion = perf_swevent_recursion_context(cpuctx); + struct perf_cpu_context *cpuctx; struct perf_event_context *ctx; - if (*recursion) - goto out; - - (*recursion)++; - barrier(); - + cpuctx = &__get_cpu_var(perf_cpu_context); + rcu_read_lock(); perf_swevent_ctx_event(&cpuctx->ctx, type, event_id, nr, nmi, data, regs); - rcu_read_lock(); /* * doesn't really matter which of the child contexts the * events ends up in. @@ -3906,23 +3949,24 @@ static void do_perf_sw_event(enum perf_type_id type, u32 event_id, if (ctx) perf_swevent_ctx_event(ctx, type, event_id, nr, nmi, data, regs); rcu_read_unlock(); - - barrier(); - (*recursion)--; - -out: - put_cpu_var(perf_cpu_context); } void __perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr) { - struct perf_sample_data data = { - .addr = addr, - }; + struct perf_sample_data data; + int rctx; - do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, - &data, regs); + rctx = perf_swevent_get_recursion_context(); + if (rctx < 0) + return; + + data.addr = addr; + data.raw = NULL; + + do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, &data, regs); + + perf_swevent_put_recursion_context(rctx); } static void perf_swevent_read(struct perf_event *event) @@ -3967,6 +4011,7 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer) event->pmu->read(event); data.addr = 0; + data.period = event->hw.last_period; regs = get_irq_regs(); /* * In case we exclude kernel IPs or are somehow not in interrupt @@ -4145,6 +4190,7 @@ void perf_tp_event(int event_id, u64 addr, u64 count, void *record, if (!regs) regs = task_pt_regs(current); + /* Trace events already protected against recursion */ do_perf_sw_event(PERF_TYPE_TRACEPOINT, event_id, count, 1, &data, regs); } @@ -4231,6 +4277,53 @@ static void perf_event_free_filter(struct perf_event *event) #endif /* CONFIG_EVENT_PROFILE */ +#ifdef CONFIG_HAVE_HW_BREAKPOINT +static void bp_perf_event_destroy(struct perf_event *event) +{ + release_bp_slot(event); +} + +static const struct pmu *bp_perf_event_init(struct perf_event *bp) +{ + int err; + /* + * The breakpoint is already filled if we haven't created the counter + * through perf syscall + * FIXME: manage to get trigerred to NULL if it comes from syscalls + */ + if (!bp->callback) + err = register_perf_hw_breakpoint(bp); + else + err = __register_perf_hw_breakpoint(bp); + if (err) + return ERR_PTR(err); + + bp->destroy = bp_perf_event_destroy; + + return &perf_ops_bp; +} + +void perf_bp_event(struct perf_event *bp, void *data) +{ + struct perf_sample_data sample; + struct pt_regs *regs = data; + + sample.addr = bp->attr.bp_addr; + + if (!perf_exclude_event(bp, regs)) + perf_swevent_add(bp, 1, 1, &sample, regs); +} +#else +static const struct pmu *bp_perf_event_init(struct perf_event *bp) +{ + return NULL; +} + +void perf_bp_event(struct perf_event *bp, void *regs) +{ +} +#endif + atomic_t perf_swevent_enabled[PERF_COUNT_SW_MAX]; static void sw_perf_event_destroy(struct perf_event *event) @@ -4297,6 +4390,7 @@ perf_event_alloc(struct perf_event_attr *attr, struct perf_event_context *ctx, struct perf_event *group_leader, struct perf_event *parent_event, + perf_callback_t callback, gfp_t gfpflags) { const struct pmu *pmu; @@ -4339,6 +4433,11 @@ perf_event_alloc(struct perf_event_attr *attr, event->state = PERF_EVENT_STATE_INACTIVE; + if (!callback && parent_event) + callback = parent_event->callback; + + event->callback = callback; + if (attr->disabled) event->state = PERF_EVENT_STATE_OFF; @@ -4373,6 +4472,11 @@ perf_event_alloc(struct perf_event_attr *attr, pmu = tp_perf_event_init(event); break; + case PERF_TYPE_BREAKPOINT: + pmu = bp_perf_event_init(event); + break; + + default: break; } @@ -4615,7 +4719,7 @@ SYSCALL_DEFINE5(perf_event_open, } event = perf_event_alloc(&attr, cpu, ctx, group_leader, - NULL, GFP_KERNEL); + NULL, NULL, GFP_KERNEL); err = PTR_ERR(event); if (IS_ERR(event)) goto err_put_context; @@ -4663,6 +4767,60 @@ err_put_context: return err; } +/** + * perf_event_create_kernel_counter + * + * @attr: attributes of the counter to create + * @cpu: cpu in which the counter is bound + * @pid: task to profile + */ +struct perf_event * +perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, + pid_t pid, perf_callback_t callback) +{ + struct perf_event *event; + struct perf_event_context *ctx; + int err; + + /* + * Get the target context (task or percpu): + */ + + ctx = find_get_context(pid, cpu); + if (IS_ERR(ctx)) { + err = PTR_ERR(ctx); + goto err_exit; + } + + event = perf_event_alloc(attr, cpu, ctx, NULL, + NULL, callback, GFP_KERNEL); + if (IS_ERR(event)) { + err = PTR_ERR(event); + goto err_put_context; + } + + event->filp = NULL; + WARN_ON_ONCE(ctx->parent_ctx); + mutex_lock(&ctx->mutex); + perf_install_in_context(ctx, event, cpu); + ++ctx->generation; + mutex_unlock(&ctx->mutex); + + event->owner = current; + get_task_struct(current); + mutex_lock(¤t->perf_event_mutex); + list_add_tail(&event->owner_entry, ¤t->perf_event_list); + mutex_unlock(¤t->perf_event_mutex); + + return event; + + err_put_context: + put_ctx(ctx); + err_exit: + return ERR_PTR(err); +} +EXPORT_SYMBOL_GPL(perf_event_create_kernel_counter); + /* * inherit a event from parent task to child task: */ @@ -4688,7 +4846,7 @@ inherit_event(struct perf_event *parent_event, child_event = perf_event_alloc(&parent_event->attr, parent_event->cpu, child_ctx, group_leader, parent_event, - GFP_KERNEL); + NULL, GFP_KERNEL); if (IS_ERR(child_event)) return child_event; get_ctx(child_ctx); @@ -4706,6 +4864,8 @@ inherit_event(struct perf_event *parent_event, if (parent_event->attr.freq) child_event->hw.sample_period = parent_event->hw.sample_period; + child_event->overflow_handler = parent_event->overflow_handler; + /* * Link it up in the child's context: */ @@ -4795,7 +4955,6 @@ __perf_event_exit_task(struct perf_event *child_event, { struct perf_event *parent_event; - update_event_times(child_event); perf_event_remove_from_context(child_event); parent_event = child_event->parent; @@ -4847,6 +5006,7 @@ void perf_event_exit_task(struct task_struct *child) * the events from it. */ unclone_ctx(child_ctx); + update_context_time(child_ctx); spin_unlock_irqrestore(&child_ctx->lock, flags); /* diff --git a/kernel/signal.c b/kernel/signal.c index 6705320784f..93e72e5feae 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -27,7 +27,8 @@ #include <linux/freezer.h> #include <linux/pid_namespace.h> #include <linux/nsproxy.h> -#include <trace/events/sched.h> +#define CREATE_TRACE_POINTS +#include <trace/events/signal.h> #include <asm/param.h> #include <asm/uaccess.h> @@ -834,7 +835,7 @@ static int __send_signal(int sig, struct siginfo *info, struct task_struct *t, struct sigqueue *q; int override_rlimit; - trace_sched_signal_send(sig, t); + trace_signal_generate(sig, info, t); assert_spin_locked(&t->sighand->siglock); @@ -896,12 +897,21 @@ static int __send_signal(int sig, struct siginfo *info, struct task_struct *t, break; } } else if (!is_si_special(info)) { - if (sig >= SIGRTMIN && info->si_code != SI_USER) - /* - * Queue overflow, abort. We may abort if the signal was rt - * and sent by user using something other than kill(). - */ + if (sig >= SIGRTMIN && info->si_code != SI_USER) { + /* + * Queue overflow, abort. We may abort if the + * signal was rt and sent by user using something + * other than kill(). + */ + trace_signal_overflow_fail(sig, group, info); return -EAGAIN; + } else { + /* + * This is a silent loss of information. We still + * send the signal, but the *info bits are lost. + */ + trace_signal_lose_info(sig, group, info); + } } out_set: @@ -1839,6 +1849,9 @@ relock: ka = &sighand->action[signr-1]; } + /* Trace actually delivered signals. */ + trace_signal_deliver(signr, info, ka); + if (ka->sa.sa_handler == SIG_IGN) /* Do nothing. */ continue; if (ka->sa.sa_handler != SIG_DFL) { diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index f05671609a8..d006554888d 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -339,6 +339,27 @@ config POWER_TRACER power management decisions, specifically the C-state and P-state behavior. +config KSYM_TRACER + bool "Trace read and write access on kernel memory locations" + depends on HAVE_HW_BREAKPOINT + select TRACING + help + This tracer helps find read and write operations on any given kernel + symbol i.e. /proc/kallsyms. + +config PROFILE_KSYM_TRACER + bool "Profile all kernel memory accesses on 'watched' variables" + depends on KSYM_TRACER + help + This tracer profiles kernel accesses on variables watched through the + ksym tracer ftrace plugin. Depending upon the hardware, all read + and write operations on kernel variables can be monitored for + accesses. + + The results will be displayed in: + /debugfs/tracing/profile_ksym + + Say N if unsure. config STACK_TRACER bool "Trace max stack" diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index edc3a3cca1a..cd9ecd89ec7 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -54,6 +54,7 @@ obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o obj-$(CONFIG_EVENT_PROFILE) += trace_event_profile.o obj-$(CONFIG_EVENT_TRACING) += trace_events_filter.o obj-$(CONFIG_KPROBE_EVENT) += trace_kprobe.o +obj-$(CONFIG_KSYM_TRACER) += trace_ksym.o obj-$(CONFIG_EVENT_TRACING) += power-traces.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index b4e4212e66d..1d7f4830a80 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -11,6 +11,7 @@ #include <linux/ftrace.h> #include <trace/boot.h> #include <linux/kmemtrace.h> +#include <linux/hw_breakpoint.h> #include <linux/trace_seq.h> #include <linux/ftrace_event.h> @@ -37,6 +38,7 @@ enum trace_type { TRACE_KMEM_ALLOC, TRACE_KMEM_FREE, TRACE_BLK, + TRACE_KSYM, __TRACE_LAST_TYPE, }; @@ -98,7 +100,7 @@ struct syscall_trace_enter { struct syscall_trace_exit { struct trace_entry ent; int nr; - unsigned long ret; + long ret; }; struct kprobe_trace_entry { @@ -232,6 +234,7 @@ extern void __ftrace_bad_type(void); TRACE_KMEM_ALLOC); \ IF_ASSIGN(var, ent, struct kmemtrace_free_entry, \ TRACE_KMEM_FREE); \ + IF_ASSIGN(var, ent, struct ksym_trace_entry, TRACE_KSYM);\ __ftrace_bad_type(); \ } while (0) @@ -387,6 +390,8 @@ int register_tracer(struct tracer *type); void unregister_tracer(struct tracer *type); int is_tracing_stopped(void); +extern int process_new_ksym_entry(char *ksymname, int op, unsigned long addr); + extern unsigned long nsecs_to_usecs(unsigned long nsecs); #ifdef CONFIG_TRACER_MAX_TRACE @@ -461,6 +466,8 @@ 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); +extern int trace_selftest_startup_ksym(struct tracer *trace, + struct trace_array *tr); #endif /* CONFIG_FTRACE_STARTUP_TEST */ extern void *head_page(struct trace_array_cpu *data); diff --git a/kernel/trace/trace_entries.h b/kernel/trace/trace_entries.h index ead3d724599..c16a08f399d 100644 --- a/kernel/trace/trace_entries.h +++ b/kernel/trace/trace_entries.h @@ -364,3 +364,19 @@ FTRACE_ENTRY(kmem_free, kmemtrace_free_entry, F_printk("type:%u call_site:%lx ptr:%p", __entry->type_id, __entry->call_site, __entry->ptr) ); + +FTRACE_ENTRY(ksym_trace, ksym_trace_entry, + + TRACE_KSYM, + + F_STRUCT( + __field( unsigned long, ip ) + __field( unsigned char, type ) + __array( char , cmd, TASK_COMM_LEN ) + __field( unsigned long, addr ) + ), + + F_printk("ip: %pF type: %d ksym_name: %pS cmd: %s", + (void *)__entry->ip, (unsigned int)__entry->type, + (void *)__entry->addr, __entry->cmd) +); diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c index e0d351b01f5..d9c60f80aa0 100644 --- a/kernel/trace/trace_event_profile.c +++ b/kernel/trace/trace_event_profile.c @@ -9,31 +9,33 @@ #include "trace.h" -struct perf_trace_buf *perf_trace_buf; +char *perf_trace_buf; EXPORT_SYMBOL_GPL(perf_trace_buf); -struct perf_trace_buf *perf_trace_buf_nmi; +char *perf_trace_buf_nmi; EXPORT_SYMBOL_GPL(perf_trace_buf_nmi); +typedef typeof(char [FTRACE_MAX_PROFILE_SIZE]) perf_trace_t ; + /* Count the events in use (per event id, not per instance) */ static int total_profile_count; static int ftrace_profile_enable_event(struct ftrace_event_call *event) { - struct perf_trace_buf *buf; + char *buf; int ret = -ENOMEM; if (atomic_inc_return(&event->profile_count)) return 0; if (!total_profile_count) { - buf = alloc_percpu(struct perf_trace_buf); + buf = (char *)alloc_percpu(perf_trace_t); if (!buf) goto fail_buf; rcu_assign_pointer(perf_trace_buf, buf); - buf = alloc_percpu(struct perf_trace_buf); + buf = (char *)alloc_percpu(perf_trace_t); if (!buf) goto fail_buf_nmi; @@ -79,7 +81,7 @@ int ftrace_profile_enable(int event_id) static void ftrace_profile_disable_event(struct ftrace_event_call *event) { - struct perf_trace_buf *buf, *nmi_buf; + char *buf, *nmi_buf; if (!atomic_add_negative(-1, &event->profile_count)) return; diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 3696476f307..aff5f80b59b 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -243,7 +243,11 @@ static int probe_arg_string(char *buf, size_t n, struct fetch_func *ff) ret = snprintf(buf, n, "@0x%p", ff->data); else if (ff->func == fetch_symbol) { struct symbol_cache *sc = ff->data; - ret = snprintf(buf, n, "@%s%+ld", sc->symbol, sc->offset); + if (sc->offset) + ret = snprintf(buf, n, "@%s%+ld", sc->symbol, + sc->offset); + else + ret = snprintf(buf, n, "@%s", sc->symbol); } else if (ff->func == fetch_retvalue) ret = snprintf(buf, n, "$retval"); else if (ff->func == fetch_stack_address) @@ -479,7 +483,8 @@ static int parse_probe_vars(char *arg, struct fetch_func *ff, int is_return) return ret; } -static int parse_probe_arg(char *arg, struct fetch_func *ff, int is_return) +/* Recursive argument parser */ +static int __parse_probe_arg(char *arg, struct fetch_func *ff, int is_return) { int ret = 0; unsigned long param; @@ -539,7 +544,7 @@ static int parse_probe_arg(char *arg, struct fetch_func *ff, int is_return) if (!id) return -ENOMEM; id->offset = offset; - ret = parse_probe_arg(arg, &id->orig, is_return); + ret = __parse_probe_arg(arg, &id->orig, is_return); if (ret) kfree(id); else { @@ -556,6 +561,16 @@ static int parse_probe_arg(char *arg, struct fetch_func *ff, int is_return) return ret; } +/* String length checking wrapper */ +static int parse_probe_arg(char *arg, struct fetch_func *ff, int is_return) +{ + if (strlen(arg) > MAX_ARGSTR_LEN) { + pr_info("Argument is too long.: %s\n", arg); + return -ENOSPC; + } + return __parse_probe_arg(arg, ff, is_return); +} + /* Return 1 if name is reserved or already used by another argument */ static int conflict_field_name(const char *name, struct probe_arg *args, int narg) @@ -694,20 +709,23 @@ static int create_trace_probe(int argc, char **argv) } tp->args[i].name = kstrdup(argv[i], GFP_KERNEL); - - /* Parse fetch argument */ - if (strlen(arg) > MAX_ARGSTR_LEN) { - pr_info("Argument%d(%s) is too long.\n", i, arg); - ret = -ENOSPC; + if (!tp->args[i].name) { + pr_info("Failed to allocate argument%d name '%s'.\n", + i, argv[i]); + ret = -ENOMEM; goto error; } + + /* Parse fetch argument */ ret = parse_probe_arg(arg, &tp->args[i].fetch, is_return); if (ret) { pr_info("Parse error at argument%d. (%d)\n", i, ret); + kfree(tp->args[i].name); goto error; } + + tp->nr_args++; } - tp->nr_args = i; ret = register_trace_probe(tp); if (ret) @@ -758,12 +776,14 @@ static int probes_seq_show(struct seq_file *m, void *v) char buf[MAX_ARGSTR_LEN + 1]; seq_printf(m, "%c", probe_is_return(tp) ? 'r' : 'p'); - seq_printf(m, ":%s", tp->call.name); + seq_printf(m, ":%s/%s", tp->call.system, tp->call.name); - if (tp->symbol) + if (!tp->symbol) + seq_printf(m, " 0x%p", tp->rp.kp.addr); + else if (tp->rp.kp.offset) seq_printf(m, " %s+%u", probe_symbol(tp), tp->rp.kp.offset); else - seq_printf(m, " 0x%p", tp->rp.kp.addr); + seq_printf(m, " %s", probe_symbol(tp)); for (i = 0; i < tp->nr_args; i++) { ret = probe_arg_string(buf, MAX_ARGSTR_LEN, &tp->args[i].fetch); @@ -1208,11 +1228,12 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, struct trace_probe *tp = container_of(kp, struct trace_probe, rp.kp); struct ftrace_event_call *call = &tp->call; struct kprobe_trace_entry *entry; - struct perf_trace_buf *trace_buf; struct trace_entry *ent; int size, __size, i, pc, __cpu; unsigned long irq_flags; + char *trace_buf; char *raw_data; + int rctx; pc = preempt_count(); __size = SIZEOF_KPROBE_TRACE_ENTRY(tp->nr_args); @@ -1227,6 +1248,11 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, * This also protects the rcu read side */ local_irq_save(irq_flags); + + rctx = perf_swevent_get_recursion_context(); + if (rctx < 0) + goto end_recursion; + __cpu = smp_processor_id(); if (in_nmi()) @@ -1237,18 +1263,7 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, if (!trace_buf) goto end; - trace_buf = per_cpu_ptr(trace_buf, __cpu); - - if (trace_buf->recursion++) - goto end_recursion; - - /* - * Make recursion update visible before entering perf_tp_event - * so that we protect from perf recursions. - */ - barrier(); - - raw_data = trace_buf->buf; + raw_data = per_cpu_ptr(trace_buf, __cpu); /* Zero dead bytes from alignment to avoid buffer leak to userspace */ *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; @@ -1263,9 +1278,9 @@ static __kprobes int kprobe_profile_func(struct kprobe *kp, entry->args[i] = call_fetch(&tp->args[i].fetch, regs); perf_tp_event(call->id, entry->ip, 1, entry, size); -end_recursion: - trace_buf->recursion--; end: + perf_swevent_put_recursion_context(rctx); +end_recursion: local_irq_restore(irq_flags); return 0; @@ -1278,11 +1293,12 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, struct trace_probe *tp = container_of(ri->rp, struct trace_probe, rp); struct ftrace_event_call *call = &tp->call; struct kretprobe_trace_entry *entry; - struct perf_trace_buf *trace_buf; struct trace_entry *ent; int size, __size, i, pc, __cpu; unsigned long irq_flags; + char *trace_buf; char *raw_data; + int rctx; pc = preempt_count(); __size = SIZEOF_KRETPROBE_TRACE_ENTRY(tp->nr_args); @@ -1297,6 +1313,11 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, * This also protects the rcu read side */ local_irq_save(irq_flags); + + rctx = perf_swevent_get_recursion_context(); + if (rctx < 0) + goto end_recursion; + __cpu = smp_processor_id(); if (in_nmi()) @@ -1307,18 +1328,7 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, if (!trace_buf) goto end; - trace_buf = per_cpu_ptr(trace_buf, __cpu); - - if (trace_buf->recursion++) - goto end_recursion; - - /* - * Make recursion update visible before entering perf_tp_event - * so that we protect from perf recursions. - */ - barrier(); - - raw_data = trace_buf->buf; + raw_data = per_cpu_ptr(trace_buf, __cpu); /* Zero dead bytes from alignment to avoid buffer leak to userspace */ *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; @@ -1334,9 +1344,9 @@ static __kprobes int kretprobe_profile_func(struct kretprobe_instance *ri, entry->args[i] = call_fetch(&tp->args[i].fetch, regs); perf_tp_event(call->id, entry->ret_ip, 1, entry, size); -end_recursion: - trace_buf->recursion--; end: + perf_swevent_put_recursion_context(rctx); +end_recursion: local_irq_restore(irq_flags); return 0; diff --git a/kernel/trace/trace_ksym.c b/kernel/trace/trace_ksym.c new file mode 100644 index 00000000000..ddfa0fd43bc --- /dev/null +++ b/kernel/trace/trace_ksym.c @@ -0,0 +1,550 @@ +/* + * trace_ksym.c - Kernel Symbol Tracer + * + * 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) IBM Corporation, 2009 + */ + +#include <linux/kallsyms.h> +#include <linux/uaccess.h> +#include <linux/debugfs.h> +#include <linux/ftrace.h> +#include <linux/module.h> +#include <linux/fs.h> + +#include "trace_output.h" +#include "trace_stat.h" +#include "trace.h" + +#include <linux/hw_breakpoint.h> +#include <asm/hw_breakpoint.h> + +/* + * For now, let us restrict the no. of symbols traced simultaneously to number + * of available hardware breakpoint registers. + */ +#define KSYM_TRACER_MAX HBP_NUM + +#define KSYM_TRACER_OP_LEN 3 /* rw- */ + +struct trace_ksym { + struct perf_event **ksym_hbp; + struct perf_event_attr attr; +#ifdef CONFIG_PROFILE_KSYM_TRACER + unsigned long counter; +#endif + struct hlist_node ksym_hlist; +}; + +static struct trace_array *ksym_trace_array; + +static unsigned int ksym_filter_entry_count; +static unsigned int ksym_tracing_enabled; + +static HLIST_HEAD(ksym_filter_head); + +static DEFINE_MUTEX(ksym_tracer_mutex); + +#ifdef CONFIG_PROFILE_KSYM_TRACER + +#define MAX_UL_INT 0xffffffff + +void ksym_collect_stats(unsigned long hbp_hit_addr) +{ + struct hlist_node *node; + struct trace_ksym *entry; + + rcu_read_lock(); + hlist_for_each_entry_rcu(entry, node, &ksym_filter_head, ksym_hlist) { + if ((entry->attr.bp_addr == hbp_hit_addr) && + (entry->counter <= MAX_UL_INT)) { + entry->counter++; + break; + } + } + rcu_read_unlock(); +} +#endif /* CONFIG_PROFILE_KSYM_TRACER */ + +void ksym_hbp_handler(struct perf_event *hbp, void *data) +{ + struct ring_buffer_event *event; + struct ksym_trace_entry *entry; + struct pt_regs *regs = data; + struct ring_buffer *buffer; + int pc; + + if (!ksym_tracing_enabled) + return; + + buffer = ksym_trace_array->buffer; + + pc = preempt_count(); + + event = trace_buffer_lock_reserve(buffer, TRACE_KSYM, + sizeof(*entry), 0, pc); + if (!event) + return; + + entry = ring_buffer_event_data(event); + entry->ip = instruction_pointer(regs); + entry->type = hw_breakpoint_type(hbp); + entry->addr = hw_breakpoint_addr(hbp); + strlcpy(entry->cmd, current->comm, TASK_COMM_LEN); + +#ifdef CONFIG_PROFILE_KSYM_TRACER + ksym_collect_stats(hw_breakpoint_addr(hbp)); +#endif /* CONFIG_PROFILE_KSYM_TRACER */ + + trace_buffer_unlock_commit(buffer, event, 0, pc); +} + +/* Valid access types are represented as + * + * rw- : Set Read/Write Access Breakpoint + * -w- : Set Write Access Breakpoint + * --- : Clear Breakpoints + * --x : Set Execution Break points (Not available yet) + * + */ +static int ksym_trace_get_access_type(char *str) +{ + int access = 0; + + if (str[0] == 'r') + access |= HW_BREAKPOINT_R; + + if (str[1] == 'w') + access |= HW_BREAKPOINT_W; + + if (str[2] == 'x') + access |= HW_BREAKPOINT_X; + + switch (access) { + case HW_BREAKPOINT_R: + case HW_BREAKPOINT_W: + case HW_BREAKPOINT_W | HW_BREAKPOINT_R: + return access; + default: + return -EINVAL; + } +} + +/* + * There can be several possible malformed requests and we attempt to capture + * all of them. We enumerate some of the rules + * 1. We will not allow kernel symbols with ':' since it is used as a delimiter. + * i.e. multiple ':' symbols disallowed. Possible uses are of the form + * <module>:<ksym_name>:<op>. + * 2. No delimiter symbol ':' in the input string + * 3. Spurious operator symbols or symbols not in their respective positions + * 4. <ksym_name>:--- i.e. clear breakpoint request when ksym_name not in file + * 5. Kernel symbol not a part of /proc/kallsyms + * 6. Duplicate requests + */ +static int parse_ksym_trace_str(char *input_string, char **ksymname, + unsigned long *addr) +{ + int ret; + + *ksymname = strsep(&input_string, ":"); + *addr = kallsyms_lookup_name(*ksymname); + + /* Check for malformed request: (2), (1) and (5) */ + if ((!input_string) || + (strlen(input_string) != KSYM_TRACER_OP_LEN) || + (*addr == 0)) + return -EINVAL;; + + ret = ksym_trace_get_access_type(input_string); + + return ret; +} + +int process_new_ksym_entry(char *ksymname, int op, unsigned long addr) +{ + struct trace_ksym *entry; + int ret = -ENOMEM; + + if (ksym_filter_entry_count >= KSYM_TRACER_MAX) { + printk(KERN_ERR "ksym_tracer: Maximum limit:(%d) reached. No" + " new requests for tracing can be accepted now.\n", + KSYM_TRACER_MAX); + return -ENOSPC; + } + + entry = kzalloc(sizeof(struct trace_ksym), GFP_KERNEL); + if (!entry) + return -ENOMEM; + + hw_breakpoint_init(&entry->attr); + + entry->attr.bp_type = op; + entry->attr.bp_addr = addr; + entry->attr.bp_len = HW_BREAKPOINT_LEN_4; + + ret = -EAGAIN; + entry->ksym_hbp = register_wide_hw_breakpoint(&entry->attr, + ksym_hbp_handler); + + if (IS_ERR(entry->ksym_hbp)) { + ret = PTR_ERR(entry->ksym_hbp); + printk(KERN_INFO "ksym_tracer request failed. Try again" + " later!!\n"); + goto err; + } + + hlist_add_head_rcu(&(entry->ksym_hlist), &ksym_filter_head); + ksym_filter_entry_count++; + + return 0; + +err: + kfree(entry); + + return ret; +} + +static ssize_t ksym_trace_filter_read(struct file *filp, char __user *ubuf, + size_t count, loff_t *ppos) +{ + struct trace_ksym *entry; + struct hlist_node *node; + struct trace_seq *s; + ssize_t cnt = 0; + int ret; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + trace_seq_init(s); + + mutex_lock(&ksym_tracer_mutex); + + hlist_for_each_entry(entry, node, &ksym_filter_head, ksym_hlist) { + ret = trace_seq_printf(s, "%pS:", (void *)entry->attr.bp_addr); + if (entry->attr.bp_type == HW_BREAKPOINT_R) + ret = trace_seq_puts(s, "r--\n"); + else if (entry->attr.bp_type == HW_BREAKPOINT_W) + ret = trace_seq_puts(s, "-w-\n"); + else if (entry->attr.bp_type == (HW_BREAKPOINT_W | HW_BREAKPOINT_R)) + ret = trace_seq_puts(s, "rw-\n"); + WARN_ON_ONCE(!ret); + } + + cnt = simple_read_from_buffer(ubuf, count, ppos, s->buffer, s->len); + + mutex_unlock(&ksym_tracer_mutex); + + kfree(s); + + return cnt; +} + +static void __ksym_trace_reset(void) +{ + struct trace_ksym *entry; + struct hlist_node *node, *node1; + + mutex_lock(&ksym_tracer_mutex); + hlist_for_each_entry_safe(entry, node, node1, &ksym_filter_head, + ksym_hlist) { + unregister_wide_hw_breakpoint(entry->ksym_hbp); + ksym_filter_entry_count--; + hlist_del_rcu(&(entry->ksym_hlist)); + synchronize_rcu(); + kfree(entry); + } + mutex_unlock(&ksym_tracer_mutex); +} + +static ssize_t ksym_trace_filter_write(struct file *file, + const char __user *buffer, + size_t count, loff_t *ppos) +{ + struct trace_ksym *entry; + struct hlist_node *node; + char *input_string, *ksymname = NULL; + unsigned long ksym_addr = 0; + int ret, op, changed = 0; + + input_string = kzalloc(count + 1, GFP_KERNEL); + if (!input_string) + return -ENOMEM; + + if (copy_from_user(input_string, buffer, count)) { + kfree(input_string); + return -EFAULT; + } + input_string[count] = '\0'; + + strstrip(input_string); + + /* + * Clear all breakpoints if: + * 1: echo > ksym_trace_filter + * 2: echo 0 > ksym_trace_filter + * 3: echo "*:---" > ksym_trace_filter + */ + if (!input_string[0] || !strcmp(input_string, "0") || + !strcmp(input_string, "*:---")) { + __ksym_trace_reset(); + kfree(input_string); + return count; + } + + ret = op = parse_ksym_trace_str(input_string, &ksymname, &ksym_addr); + if (ret < 0) { + kfree(input_string); + return ret; + } + + mutex_lock(&ksym_tracer_mutex); + + ret = -EINVAL; + hlist_for_each_entry(entry, node, &ksym_filter_head, ksym_hlist) { + if (entry->attr.bp_addr == ksym_addr) { + /* Check for malformed request: (6) */ + if (entry->attr.bp_type != op) + changed = 1; + else + goto out; + break; + } + } + if (changed) { + unregister_wide_hw_breakpoint(entry->ksym_hbp); + entry->attr.bp_type = op; + ret = 0; + if (op > 0) { + entry->ksym_hbp = + register_wide_hw_breakpoint(&entry->attr, + ksym_hbp_handler); + if (IS_ERR(entry->ksym_hbp)) + ret = PTR_ERR(entry->ksym_hbp); + else + goto out; + } + /* Error or "symbol:---" case: drop it */ + ksym_filter_entry_count--; + hlist_del_rcu(&(entry->ksym_hlist)); + synchronize_rcu(); + kfree(entry); + goto out; + } else { + /* Check for malformed request: (4) */ + if (op == 0) + goto out; + ret = process_new_ksym_entry(ksymname, op, ksym_addr); + } +out: + mutex_unlock(&ksym_tracer_mutex); + + kfree(input_string); + + if (!ret) + ret = count; + return ret; +} + +static const struct file_operations ksym_tracing_fops = { + .open = tracing_open_generic, + .read = ksym_trace_filter_read, + .write = ksym_trace_filter_write, +}; + +static void ksym_trace_reset(struct trace_array *tr) +{ + ksym_tracing_enabled = 0; + __ksym_trace_reset(); +} + +static int ksym_trace_init(struct trace_array *tr) +{ + int cpu, ret = 0; + + for_each_online_cpu(cpu) + tracing_reset(tr, cpu); + ksym_tracing_enabled = 1; + ksym_trace_array = tr; + + return ret; +} + +static void ksym_trace_print_header(struct seq_file *m) +{ + seq_puts(m, + "# TASK-PID CPU# Symbol " + "Type Function\n"); + seq_puts(m, + "# | | | " + " | |\n"); +} + +static enum print_line_t ksym_trace_output(struct trace_iterator *iter) +{ + struct trace_entry *entry = iter->ent; + struct trace_seq *s = &iter->seq; + struct ksym_trace_entry *field; + char str[KSYM_SYMBOL_LEN]; + int ret; + + if (entry->type != TRACE_KSYM) + return TRACE_TYPE_UNHANDLED; + + trace_assign_type(field, entry); + + ret = trace_seq_printf(s, "%11s-%-5d [%03d] %pS", field->cmd, + entry->pid, iter->cpu, (char *)field->addr); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + switch (field->type) { + case HW_BREAKPOINT_R: + ret = trace_seq_printf(s, " R "); + break; + case HW_BREAKPOINT_W: + ret = trace_seq_printf(s, " W "); + break; + case HW_BREAKPOINT_R | HW_BREAKPOINT_W: + ret = trace_seq_printf(s, " RW "); + break; + default: + return TRACE_TYPE_PARTIAL_LINE; + } + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + sprint_symbol(str, field->ip); + ret = trace_seq_printf(s, "%s\n", str); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +struct tracer ksym_tracer __read_mostly = +{ + .name = "ksym_tracer", + .init = ksym_trace_init, + .reset = ksym_trace_reset, +#ifdef CONFIG_FTRACE_SELFTEST + .selftest = trace_selftest_startup_ksym, +#endif + .print_header = ksym_trace_print_header, + .print_line = ksym_trace_output +}; + +__init static int init_ksym_trace(void) +{ + struct dentry *d_tracer; + struct dentry *entry; + + d_tracer = tracing_init_dentry(); + ksym_filter_entry_count = 0; + + entry = debugfs_create_file("ksym_trace_filter", 0644, d_tracer, + NULL, &ksym_tracing_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'ksym_trace_filter' file\n"); + + return register_tracer(&ksym_tracer); +} +device_initcall(init_ksym_trace); + + +#ifdef CONFIG_PROFILE_KSYM_TRACER +static int ksym_tracer_stat_headers(struct seq_file *m) +{ + seq_puts(m, " Access Type "); + seq_puts(m, " Symbol Counter\n"); + seq_puts(m, " ----------- "); + seq_puts(m, " ------ -------\n"); + return 0; +} + +static int ksym_tracer_stat_show(struct seq_file *m, void *v) +{ + struct hlist_node *stat = v; + struct trace_ksym *entry; + int access_type = 0; + char fn_name[KSYM_NAME_LEN]; + + entry = hlist_entry(stat, struct trace_ksym, ksym_hlist); + + access_type = entry->attr.bp_type; + + switch (access_type) { + case HW_BREAKPOINT_R: + seq_puts(m, " R "); + break; + case HW_BREAKPOINT_W: + seq_puts(m, " W "); + break; + case HW_BREAKPOINT_R | HW_BREAKPOINT_W: + seq_puts(m, " RW "); + break; + default: + seq_puts(m, " NA "); + } + + if (lookup_symbol_name(entry->attr.bp_addr, fn_name) >= 0) + seq_printf(m, " %-36s", fn_name); + else + seq_printf(m, " %-36s", "<NA>"); + seq_printf(m, " %15lu\n", entry->counter); + + return 0; +} + +static void *ksym_tracer_stat_start(struct tracer_stat *trace) +{ + return ksym_filter_head.first; +} + +static void * +ksym_tracer_stat_next(void *v, int idx) +{ + struct hlist_node *stat = v; + + return stat->next; +} + +static struct tracer_stat ksym_tracer_stats = { + .name = "ksym_tracer", + .stat_start = ksym_tracer_stat_start, + .stat_next = ksym_tracer_stat_next, + .stat_headers = ksym_tracer_stat_headers, + .stat_show = ksym_tracer_stat_show +}; + +__init static int ksym_tracer_stat_init(void) +{ + int ret; + + ret = register_stat_tracer(&ksym_tracer_stats); + if (ret) { + printk(KERN_WARNING "Warning: could not register " + "ksym tracer stats\n"); + return 1; + } + + return 0; +} +fs_initcall(ksym_tracer_stat_init); +#endif /* CONFIG_PROFILE_KSYM_TRACER */ diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index d2cdbabb4ea..dc98309e839 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -17,6 +17,7 @@ static inline int trace_valid_entry(struct trace_entry *entry) case TRACE_GRAPH_ENT: case TRACE_GRAPH_RET: case TRACE_HW_BRANCHES: + case TRACE_KSYM: return 1; } return 0; @@ -808,3 +809,57 @@ trace_selftest_startup_hw_branches(struct tracer *trace, return ret; } #endif /* CONFIG_HW_BRANCH_TRACER */ + +#ifdef CONFIG_KSYM_TRACER +static int ksym_selftest_dummy; + +int +trace_selftest_startup_ksym(struct tracer *trace, struct trace_array *tr) +{ + unsigned long count; + int ret; + + /* start the tracing */ + ret = tracer_init(trace, tr); + if (ret) { + warn_failed_init_tracer(trace, ret); + return ret; + } + + ksym_selftest_dummy = 0; + /* Register the read-write tracing request */ + + ret = process_new_ksym_entry("ksym_selftest_dummy", + HW_BREAKPOINT_R | HW_BREAKPOINT_W, + (unsigned long)(&ksym_selftest_dummy)); + + if (ret < 0) { + printk(KERN_CONT "ksym_trace read-write startup test failed\n"); + goto ret_path; + } + /* Perform a read and a write operation over the dummy variable to + * trigger the tracer + */ + if (ksym_selftest_dummy == 0) + ksym_selftest_dummy++; + + /* stop the tracing. */ + tracing_stop(); + /* check the trace buffer */ + ret = trace_test_buffer(tr, &count); + trace->reset(tr); + tracing_start(); + + /* read & write operations - one each is performed on the dummy variable + * triggering two entries in the trace buffer + */ + if (!ret && count != 2) { + printk(KERN_CONT "Ksym tracer startup test failed"); + ret = -1; + } + +ret_path: + return ret; +} +#endif /* CONFIG_KSYM_TRACER */ + diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 51213b0aa81..57501d90096 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -51,32 +51,6 @@ static struct syscall_metadata *syscall_nr_to_meta(int nr) return syscalls_metadata[nr]; } -int syscall_name_to_nr(char *name) -{ - int i; - - if (!syscalls_metadata) - return -1; - - for (i = 0; i < NR_syscalls; i++) { - if (syscalls_metadata[i]) { - if (!strcmp(syscalls_metadata[i]->name, name)) - return i; - } - } - return -1; -} - -void set_syscall_enter_id(int num, int id) -{ - syscalls_metadata[num]->enter_id = id; -} - -void set_syscall_exit_id(int num, int id) -{ - syscalls_metadata[num]->exit_id = id; -} - enum print_line_t print_syscall_enter(struct trace_iterator *iter, int flags) { @@ -93,7 +67,7 @@ print_syscall_enter(struct trace_iterator *iter, int flags) if (!entry) goto end; - if (entry->enter_id != ent->type) { + if (entry->enter_event->id != ent->type) { WARN_ON_ONCE(1); goto end; } @@ -148,7 +122,7 @@ print_syscall_exit(struct trace_iterator *iter, int flags) return TRACE_TYPE_HANDLED; } - if (entry->exit_id != ent->type) { + if (entry->exit_event->id != ent->type) { WARN_ON_ONCE(1); return TRACE_TYPE_UNHANDLED; } @@ -172,18 +146,11 @@ extern char *__bad_type_size(void); 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_metadata *entry = call->data; 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;" "\tsigned:%u;\n", SYSCALL_FIELD(int, nr)); @@ -245,22 +212,19 @@ int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s) int syscall_enter_define_fields(struct ftrace_event_call *call) { struct syscall_trace_enter trace; - struct syscall_metadata *meta; + struct syscall_metadata *meta = call->data; 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; + ret = trace_define_field(call, SYSCALL_FIELD(int, nr), FILTER_OTHER); + if (ret) + return ret; + for (i = 0; i < meta->nb_args; i++) { ret = trace_define_field(call, meta->types[i], meta->args[i], offset, @@ -281,6 +245,10 @@ int syscall_exit_define_fields(struct ftrace_event_call *call) if (ret) return ret; + ret = trace_define_field(call, SYSCALL_FIELD(int, nr), FILTER_OTHER); + if (ret) + return ret; + ret = trace_define_field(call, SYSCALL_FIELD(long, ret), FILTER_OTHER); @@ -308,8 +276,8 @@ void ftrace_syscall_enter(struct pt_regs *regs, long id) size = sizeof(*entry) + sizeof(unsigned long) * sys_data->nb_args; - event = trace_current_buffer_lock_reserve(&buffer, sys_data->enter_id, - size, 0, 0); + event = trace_current_buffer_lock_reserve(&buffer, + sys_data->enter_event->id, size, 0, 0); if (!event) return; @@ -340,8 +308,8 @@ void ftrace_syscall_exit(struct pt_regs *regs, long ret) if (!sys_data) return; - event = trace_current_buffer_lock_reserve(&buffer, sys_data->exit_id, - sizeof(*entry), 0, 0); + event = trace_current_buffer_lock_reserve(&buffer, + sys_data->exit_event->id, sizeof(*entry), 0, 0); if (!event) return; @@ -358,10 +326,8 @@ int reg_event_syscall_enter(struct ftrace_event_call *call) { int ret = 0; int num; - char *name; - name = (char *)call->data; - num = syscall_name_to_nr(name); + num = ((struct syscall_metadata *)call->data)->syscall_nr; if (num < 0 || num >= NR_syscalls) return -ENOSYS; mutex_lock(&syscall_trace_lock); @@ -381,10 +347,8 @@ int reg_event_syscall_enter(struct ftrace_event_call *call) void unreg_event_syscall_enter(struct ftrace_event_call *call) { int num; - char *name; - name = (char *)call->data; - num = syscall_name_to_nr(name); + num = ((struct syscall_metadata *)call->data)->syscall_nr; if (num < 0 || num >= NR_syscalls) return; mutex_lock(&syscall_trace_lock); @@ -399,10 +363,8 @@ int reg_event_syscall_exit(struct ftrace_event_call *call) { int ret = 0; int num; - char *name; - name = call->data; - num = syscall_name_to_nr(name); + num = ((struct syscall_metadata *)call->data)->syscall_nr; if (num < 0 || num >= NR_syscalls) return -ENOSYS; mutex_lock(&syscall_trace_lock); @@ -422,10 +384,8 @@ int reg_event_syscall_exit(struct ftrace_event_call *call) void unreg_event_syscall_exit(struct ftrace_event_call *call) { int num; - char *name; - name = call->data; - num = syscall_name_to_nr(name); + num = ((struct syscall_metadata *)call->data)->syscall_nr; if (num < 0 || num >= NR_syscalls) return; mutex_lock(&syscall_trace_lock); @@ -436,13 +396,17 @@ void unreg_event_syscall_exit(struct ftrace_event_call *call) mutex_unlock(&syscall_trace_lock); } -struct trace_event event_syscall_enter = { - .trace = print_syscall_enter, -}; +int init_syscall_trace(struct ftrace_event_call *call) +{ + int id; -struct trace_event event_syscall_exit = { - .trace = print_syscall_exit, -}; + id = register_ftrace_event(call->event); + if (!id) + return -ENODEV; + call->id = id; + INIT_LIST_HEAD(&call->fields); + return 0; +} int __init init_ftrace_syscalls(void) { @@ -460,6 +424,10 @@ int __init init_ftrace_syscalls(void) for (i = 0; i < NR_syscalls; i++) { addr = arch_syscall_addr(i); meta = find_syscall_meta(addr); + if (!meta) + continue; + + meta->syscall_nr = i; syscalls_metadata[i] = meta; } @@ -477,11 +445,12 @@ static int sys_prof_refcount_exit; static void prof_syscall_enter(struct pt_regs *regs, long id) { struct syscall_metadata *sys_data; - struct perf_trace_buf *trace_buf; struct syscall_trace_enter *rec; unsigned long flags; + char *trace_buf; char *raw_data; int syscall_nr; + int rctx; int size; int cpu; @@ -505,54 +474,42 @@ static void prof_syscall_enter(struct pt_regs *regs, long id) /* Protect the per cpu buffer, begin the rcu read side */ local_irq_save(flags); + rctx = perf_swevent_get_recursion_context(); + if (rctx < 0) + goto end_recursion; + cpu = smp_processor_id(); - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); + trace_buf = rcu_dereference(perf_trace_buf); if (!trace_buf) goto end; - trace_buf = per_cpu_ptr(trace_buf, cpu); - - if (trace_buf->recursion++) - goto end_recursion; - - /* - * Make recursion update visible before entering perf_tp_event - * so that we protect from perf recursions. - */ - barrier(); - - raw_data = trace_buf->buf; + raw_data = per_cpu_ptr(trace_buf, cpu); /* 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->ent.type = sys_data->enter_event->id; rec->nr = syscall_nr; syscall_get_arguments(current, regs, 0, sys_data->nb_args, (unsigned long *)&rec->args); - perf_tp_event(sys_data->enter_id, 0, 1, rec, size); + perf_tp_event(sys_data->enter_event->id, 0, 1, rec, size); -end_recursion: - trace_buf->recursion--; end: + perf_swevent_put_recursion_context(rctx); +end_recursion: local_irq_restore(flags); } -int reg_prof_syscall_enter(char *name) +int prof_sysenter_enable(struct ftrace_event_call *call) { int ret = 0; int num; - num = syscall_name_to_nr(name); - if (num < 0 || num >= NR_syscalls) - return -ENOSYS; + num = ((struct syscall_metadata *)call->data)->syscall_nr; mutex_lock(&syscall_trace_lock); if (!sys_prof_refcount_enter) @@ -568,13 +525,11 @@ int reg_prof_syscall_enter(char *name) return ret; } -void unreg_prof_syscall_enter(char *name) +void prof_sysenter_disable(struct ftrace_event_call *call) { int num; - num = syscall_name_to_nr(name); - if (num < 0 || num >= NR_syscalls) - return; + num = ((struct syscall_metadata *)call->data)->syscall_nr; mutex_lock(&syscall_trace_lock); sys_prof_refcount_enter--; @@ -588,10 +543,11 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) { struct syscall_metadata *sys_data; struct syscall_trace_exit *rec; - struct perf_trace_buf *trace_buf; unsigned long flags; int syscall_nr; + char *trace_buf; char *raw_data; + int rctx; int size; int cpu; @@ -617,28 +573,19 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) /* Protect the per cpu buffer, begin the rcu read side */ local_irq_save(flags); + + rctx = perf_swevent_get_recursion_context(); + if (rctx < 0) + goto end_recursion; + cpu = smp_processor_id(); - if (in_nmi()) - trace_buf = rcu_dereference(perf_trace_buf_nmi); - else - trace_buf = rcu_dereference(perf_trace_buf); + trace_buf = rcu_dereference(perf_trace_buf); if (!trace_buf) goto end; - trace_buf = per_cpu_ptr(trace_buf, cpu); - - if (trace_buf->recursion++) - goto end_recursion; - - /* - * Make recursion update visible before entering perf_tp_event - * so that we protect from perf recursions. - */ - barrier(); - - raw_data = trace_buf->buf; + raw_data = per_cpu_ptr(trace_buf, cpu); /* zero the dead bytes from align to not leak stack to user */ *(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL; @@ -646,26 +593,24 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret) rec = (struct syscall_trace_exit *)raw_data; tracing_generic_entry_update(&rec->ent, 0, 0); - rec->ent.type = sys_data->exit_id; + rec->ent.type = sys_data->exit_event->id; rec->nr = syscall_nr; rec->ret = syscall_get_return_value(current, regs); - perf_tp_event(sys_data->exit_id, 0, 1, rec, size); + perf_tp_event(sys_data->exit_event->id, 0, 1, rec, size); -end_recursion: - trace_buf->recursion--; end: + perf_swevent_put_recursion_context(rctx); +end_recursion: local_irq_restore(flags); } -int reg_prof_syscall_exit(char *name) +int prof_sysexit_enable(struct ftrace_event_call *call) { int ret = 0; int num; - num = syscall_name_to_nr(name); - if (num < 0 || num >= NR_syscalls) - return -ENOSYS; + num = ((struct syscall_metadata *)call->data)->syscall_nr; mutex_lock(&syscall_trace_lock); if (!sys_prof_refcount_exit) @@ -681,13 +626,11 @@ int reg_prof_syscall_exit(char *name) return ret; } -void unreg_prof_syscall_exit(char *name) +void prof_sysexit_disable(struct ftrace_event_call *call) { int num; - num = syscall_name_to_nr(name); - if (num < 0 || num >= NR_syscalls) - return; + num = ((struct syscall_metadata *)call->data)->syscall_nr; mutex_lock(&syscall_trace_lock); sys_prof_refcount_exit--; diff --git a/samples/Kconfig b/samples/Kconfig index b92bde3c6a8..e4be84ac3d3 100644 --- a/samples/Kconfig +++ b/samples/Kconfig @@ -40,5 +40,11 @@ config SAMPLE_KRETPROBES default m depends on SAMPLE_KPROBES && KRETPROBES +config SAMPLE_HW_BREAKPOINT + tristate "Build kernel hardware breakpoint examples -- loadable module only" + depends on HAVE_HW_BREAKPOINT && m + help + This builds kernel hardware breakpoint example modules. + endif # SAMPLES diff --git a/samples/Makefile b/samples/Makefile index 43343a03b1f..0f15e6d77fd 100644 --- a/samples/Makefile +++ b/samples/Makefile @@ -1,3 +1,4 @@ # Makefile for Linux samples code -obj-$(CONFIG_SAMPLES) += kobject/ kprobes/ tracepoints/ trace_events/ +obj-$(CONFIG_SAMPLES) += kobject/ kprobes/ tracepoints/ trace_events/ \ + hw_breakpoint/ diff --git a/samples/hw_breakpoint/Makefile b/samples/hw_breakpoint/Makefile new file mode 100644 index 00000000000..0f5c31c2fc4 --- /dev/null +++ b/samples/hw_breakpoint/Makefile @@ -0,0 +1 @@ +obj-$(CONFIG_SAMPLE_HW_BREAKPOINT) += data_breakpoint.o diff --git a/samples/hw_breakpoint/data_breakpoint.c b/samples/hw_breakpoint/data_breakpoint.c new file mode 100644 index 00000000000..29525500df0 --- /dev/null +++ b/samples/hw_breakpoint/data_breakpoint.c @@ -0,0 +1,87 @@ +/* + * data_breakpoint.c - Sample HW Breakpoint file to watch kernel data address + * + * 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. + * + * usage: insmod data_breakpoint.ko ksym=<ksym_name> + * + * This file is a kernel module that places a breakpoint over ksym_name kernel + * variable using Hardware Breakpoint register. The corresponding handler which + * prints a backtrace is invoked everytime a write operation is performed on + * that variable. + * + * Copyright (C) IBM Corporation, 2009 + * + * Author: K.Prasad <prasad@linux.vnet.ibm.com> + */ +#include <linux/module.h> /* Needed by all modules */ +#include <linux/kernel.h> /* Needed for KERN_INFO */ +#include <linux/init.h> /* Needed for the macros */ +#include <linux/kallsyms.h> + +#include <linux/perf_event.h> +#include <linux/hw_breakpoint.h> + +struct perf_event **sample_hbp; + +static char ksym_name[KSYM_NAME_LEN] = "pid_max"; +module_param_string(ksym, ksym_name, KSYM_NAME_LEN, S_IRUGO); +MODULE_PARM_DESC(ksym, "Kernel symbol to monitor; this module will report any" + " write operations on the kernel symbol"); + +static void sample_hbp_handler(struct perf_event *temp, void *data) +{ + printk(KERN_INFO "%s value is changed\n", ksym_name); + dump_stack(); + printk(KERN_INFO "Dump stack from sample_hbp_handler\n"); +} + +static int __init hw_break_module_init(void) +{ + int ret; + DEFINE_BREAKPOINT_ATTR(attr); + + attr.bp_addr = kallsyms_lookup_name(ksym_name); + attr.bp_len = HW_BREAKPOINT_LEN_4; + attr.bp_type = HW_BREAKPOINT_W | HW_BREAKPOINT_R; + + sample_hbp = register_wide_hw_breakpoint(&attr, sample_hbp_handler); + if (IS_ERR(sample_hbp)) { + ret = PTR_ERR(sample_hbp); + goto fail; + } + + printk(KERN_INFO "HW Breakpoint for %s write installed\n", ksym_name); + + return 0; + +fail: + printk(KERN_INFO "Breakpoint registration failed\n"); + + return ret; +} + +static void __exit hw_break_module_exit(void) +{ + unregister_wide_hw_breakpoint(sample_hbp); + printk(KERN_INFO "HW Breakpoint for %s write uninstalled\n", ksym_name); +} + +module_init(hw_break_module_init); +module_exit(hw_break_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("K.Prasad"); +MODULE_DESCRIPTION("ksym breakpoint"); diff --git a/scripts/kernel-doc b/scripts/kernel-doc index ea9f8a58678..241310e59cd 100755 --- a/scripts/kernel-doc +++ b/scripts/kernel-doc @@ -1852,10 +1852,17 @@ sub tracepoint_munge($) { my $tracepointname = 0; my $tracepointargs = 0; - if($prototype =~ m/TRACE_EVENT\((.*?),/) { + if ($prototype =~ m/TRACE_EVENT\((.*?),/) { $tracepointname = $1; } - if($prototype =~ m/TP_PROTO\((.*?)\)/) { + if ($prototype =~ m/DEFINE_SINGLE_EVENT\((.*?),/) { + $tracepointname = $1; + } + if ($prototype =~ m/DEFINE_EVENT\((.*?),(.*?),/) { + $tracepointname = $2; + } + $tracepointname =~ s/^\s+//; #strip leading whitespace + if ($prototype =~ m/TP_PROTO\((.*?)\)/) { $tracepointargs = $1; } if (($tracepointname eq 0) || ($tracepointargs eq 0)) { @@ -1920,7 +1927,9 @@ sub process_state3_function($$) { if ($prototype =~ /SYSCALL_DEFINE/) { syscall_munge(); } - if ($prototype =~ /TRACE_EVENT/) { + if ($prototype =~ /TRACE_EVENT/ || $prototype =~ /DEFINE_EVENT/ || + $prototype =~ /DEFINE_SINGLE_EVENT/) + { tracepoint_munge($file); } dump_function($prototype, $file); diff --git a/tools/perf/.gitignore b/tools/perf/.gitignore index 0854f110bf7..fe08660ce0b 100644 --- a/tools/perf/.gitignore +++ b/tools/perf/.gitignore @@ -12,6 +12,7 @@ perf*.1 perf*.xml perf*.html common-cmds.h +perf.data tags TAGS cscope* diff --git a/tools/perf/Documentation/perf-kmem.txt b/tools/perf/Documentation/perf-kmem.txt new file mode 100644 index 00000000000..44b0ce35c28 --- /dev/null +++ b/tools/perf/Documentation/perf-kmem.txt @@ -0,0 +1,44 @@ +perf-kmem(1) +============== + +NAME +---- +perf-kmem - Tool to trace/measure kernel memory(slab) properties + +SYNOPSIS +-------- +[verse] +'perf kmem' {record} [<options>] + +DESCRIPTION +----------- +There's two variants of perf kmem: + + 'perf kmem record <command>' to record the kmem events + of an arbitrary workload. + + 'perf kmem' to report kernel memory statistics. + +OPTIONS +------- +-i <file>:: +--input=<file>:: + Select the input file (default: perf.data) + +--stat=<caller|alloc>:: + Select per callsite or per allocation statistics + +-s <key[,key2...]>:: +--sort=<key[,key2...]>:: + Sort the output (default: frag,hit,bytes) + +-l <num>:: +--line=<num>:: + Print n lines only + +--raw-ip:: + Print raw ip instead of symbol + +SEE ALSO +-------- +linkperf:perf-record[1] diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt index 0ff23de9e45..fc46c0b40f6 100644 --- a/tools/perf/Documentation/perf-record.txt +++ b/tools/perf/Documentation/perf-record.txt @@ -26,11 +26,19 @@ OPTIONS -e:: --event=:: - Select the PMU event. Selection can be a symbolic event name - (use 'perf list' to list all events) or a raw PMU - event (eventsel+umask) in the form of rNNN where NNN is a - hexadecimal event descriptor. + Select the PMU event. Selection can be: + - a symbolic event name (use 'perf list' to list all events) + + - a raw PMU event (eventsel+umask) in the form of rNNN where NNN is a + hexadecimal event descriptor. + + - a hardware breakpoint event in the form of '\mem:addr[:access]' + where addr is the address in memory you want to break in. + Access is the memory access type (read, write, execute) it can + be passed as follows: '\mem:addr[:[r][w][x]]'. + If you want to profile read-write accesses in 0x1000, just set + 'mem:0x1000:rw'. -a:: System-wide collection. diff --git a/tools/perf/Documentation/perf-trace-perl.txt b/tools/perf/Documentation/perf-trace-perl.txt new file mode 100644 index 00000000000..c5f55f43909 --- /dev/null +++ b/tools/perf/Documentation/perf-trace-perl.txt @@ -0,0 +1,219 @@ +perf-trace-perl(1) +================== + +NAME +---- +perf-trace-perl - Process trace data with a Perl script + +SYNOPSIS +-------- +[verse] +'perf trace' [-s [lang]:script[.ext] ] + +DESCRIPTION +----------- + +This perf trace option is used to process perf trace data using perf's +built-in Perl interpreter. It reads and processes the input file and +displays the results of the trace analysis implemented in the given +Perl script, if any. + +STARTER SCRIPTS +--------------- + +You can avoid reading the rest of this document by running 'perf trace +-g perl' in the same directory as an existing perf.data trace file. +That will generate a starter script containing a handler for each of +the event types in the trace file; it simply prints every available +field for each event in the trace file. + +You can also look at the existing scripts in +~/libexec/perf-core/scripts/perl for typical examples showing how to +do basic things like aggregate event data, print results, etc. Also, +the check-perf-trace.pl script, while not interesting for its results, +attempts to exercise all of the main scripting features. + +EVENT HANDLERS +-------------- + +When perf trace is invoked using a trace script, a user-defined +'handler function' is called for each event in the trace. If there's +no handler function defined for a given event type, the event is +ignored (or passed to a 'trace_handled' function, see below) and the +next event is processed. + +Most of the event's field values are passed as arguments to the +handler function; some of the less common ones aren't - those are +available as calls back into the perf executable (see below). + +As an example, the following perf record command can be used to record +all sched_wakeup events in the system: + + # perf record -c 1 -f -a -M -R -e sched:sched_wakeup + +Traces meant to be processed using a script should be recorded with +the above options: -c 1 says to sample every event, -a to enable +system-wide collection, -M to multiplex the output, and -R to collect +raw samples. + +The format file for the sched_wakep event defines the following fields +(see /sys/kernel/debug/tracing/events/sched/sched_wakeup/format): + +---- + format: + field:unsigned short common_type; + field:unsigned char common_flags; + field:unsigned char common_preempt_count; + field:int common_pid; + field:int common_lock_depth; + + field:char comm[TASK_COMM_LEN]; + field:pid_t pid; + field:int prio; + field:int success; + field:int target_cpu; +---- + +The handler function for this event would be defined as: + +---- +sub sched::sched_wakeup +{ + my ($event_name, $context, $common_cpu, $common_secs, + $common_nsecs, $common_pid, $common_comm, + $comm, $pid, $prio, $success, $target_cpu) = @_; +} +---- + +The handler function takes the form subsystem::event_name. + +The $common_* arguments in the handler's argument list are the set of +arguments passed to all event handlers; some of the fields correspond +to the common_* fields in the format file, but some are synthesized, +and some of the common_* fields aren't common enough to to be passed +to every event as arguments but are available as library functions. + +Here's a brief description of each of the invariant event args: + + $event_name the name of the event as text + $context an opaque 'cookie' used in calls back into perf + $common_cpu the cpu the event occurred on + $common_secs the secs portion of the event timestamp + $common_nsecs the nsecs portion of the event timestamp + $common_pid the pid of the current task + $common_comm the name of the current process + +All of the remaining fields in the event's format file have +counterparts as handler function arguments of the same name, as can be +seen in the example above. + +The above provides the basics needed to directly access every field of +every event in a trace, which covers 90% of what you need to know to +write a useful trace script. The sections below cover the rest. + +SCRIPT LAYOUT +------------- + +Every perf trace Perl script should start by setting up a Perl module +search path and 'use'ing a few support modules (see module +descriptions below): + +---- + use lib "$ENV{'PERF_EXEC_PATH'}/scripts/perl/Perf-Trace-Util/lib"; + use lib "./Perf-Trace-Util/lib"; + use Perf::Trace::Core; + use Perf::Trace::Context; + use Perf::Trace::Util; +---- + +The rest of the script can contain handler functions and support +functions in any order. + +Aside from the event handler functions discussed above, every script +can implement a set of optional functions: + +*trace_begin*, if defined, is called before any event is processed and +gives scripts a chance to do setup tasks: + +---- + sub trace_begin + { + } +---- + +*trace_end*, if defined, is called after all events have been + processed and gives scripts a chance to do end-of-script tasks, such + as display results: + +---- +sub trace_end +{ +} +---- + +*trace_unhandled*, if defined, is called after for any event that + doesn't have a handler explicitly defined for it. The standard set + of common arguments are passed into it: + +---- +sub trace_unhandled +{ + my ($event_name, $context, $common_cpu, $common_secs, + $common_nsecs, $common_pid, $common_comm) = @_; +} +---- + +The remaining sections provide descriptions of each of the available +built-in perf trace Perl modules and their associated functions. + +AVAILABLE MODULES AND FUNCTIONS +------------------------------- + +The following sections describe the functions and variables available +via the various Perf::Trace::* Perl modules. To use the functions and +variables from the given module, add the corresponding 'use +Perf::Trace::XXX' line to your perf trace script. + +Perf::Trace::Core Module +~~~~~~~~~~~~~~~~~~~~~~~~ + +These functions provide some essential functions to user scripts. + +The *flag_str* and *symbol_str* functions provide human-readable +strings for flag and symbolic fields. These correspond to the strings +and values parsed from the 'print fmt' fields of the event format +files: + + flag_str($event_name, $field_name, $field_value) - returns the string represention corresponding to $field_value for the flag field $field_name of event $event_name + symbol_str($event_name, $field_name, $field_value) - returns the string represention corresponding to $field_value for the symbolic field $field_name of event $event_name + +Perf::Trace::Context Module +~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Some of the 'common' fields in the event format file aren't all that +common, but need to be made accessible to user scripts nonetheless. + +Perf::Trace::Context defines a set of functions that can be used to +access this data in the context of the current event. Each of these +functions expects a $context variable, which is the same as the +$context variable passed into every event handler as the second +argument. + + common_pc($context) - returns common_preempt count for the current event + common_flags($context) - returns common_flags for the current event + common_lock_depth($context) - returns common_lock_depth for the current event + +Perf::Trace::Util Module +~~~~~~~~~~~~~~~~~~~~~~~~ + +Various utility functions for use with perf trace: + + nsecs($secs, $nsecs) - returns total nsecs given secs/nsecs pair + nsecs_secs($nsecs) - returns whole secs portion given nsecs + nsecs_nsecs($nsecs) - returns nsecs remainder given nsecs + nsecs_str($nsecs) - returns printable string in the form secs.nsecs + avg($total, $n) - returns average given a sum and a total number of values + +SEE ALSO +-------- +linkperf:perf-trace[1] diff --git a/tools/perf/Documentation/perf-trace.txt b/tools/perf/Documentation/perf-trace.txt index 41ed75398ca..07065efa60e 100644 --- a/tools/perf/Documentation/perf-trace.txt +++ b/tools/perf/Documentation/perf-trace.txt @@ -20,6 +20,15 @@ OPTIONS --dump-raw-trace=:: Display verbose dump of the trace data. +-s:: +--script=:: + Process trace data with the given script ([lang]:script[.ext]). + +-g:: +--gen-script=:: + Generate perf-trace.[ext] starter script for given language, + using current perf.data. + SEE ALSO -------- -linkperf:perf-record[1] +linkperf:perf-record[1], linkperf:perf-trace-perl[1] diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 5d1a8b0dff8..23ec66098bd 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -2,6 +2,7 @@ all:: # Define V=1 to have a more verbose compile. +# Define V=2 to have an even more verbose compile. # # Define SNPRINTF_RETURNS_BOGUS if your are on a system which snprintf() # or vsnprintf() return -1 instead of number of characters which would @@ -147,6 +148,8 @@ all:: # broken, or spawning external process is slower than built-in grep perf has). # # Define LDFLAGS=-static to build a static binary. +# +# Define EXTRA_CFLAGS=-m64 or EXTRA_CFLAGS=-m32 as appropriate for cross-builds. PERF-VERSION-FILE: .FORCE-PERF-VERSION-FILE @$(SHELL_PATH) util/PERF-VERSION-GEN @@ -159,20 +162,6 @@ uname_R := $(shell sh -c 'uname -r 2>/dev/null || echo not') uname_P := $(shell sh -c 'uname -p 2>/dev/null || echo not') uname_V := $(shell sh -c 'uname -v 2>/dev/null || echo not') -# -# Add -m32 for cross-builds: -# -ifdef NO_64BIT - MBITS := -m32 -else - # - # If we're on a 64-bit kernel, use -m64: - # - ifneq ($(patsubst %64,%,$(uname_M)),$(uname_M)) - MBITS := -m64 - endif -endif - # CFLAGS and LDFLAGS are for the users to override from the command line. # @@ -209,7 +198,7 @@ ifndef PERF_DEBUG CFLAGS_OPTIMIZE = -O6 endif -CFLAGS = $(MBITS) -ggdb3 -Wall -Wextra -std=gnu99 -Werror $(CFLAGS_OPTIMIZE) -D_FORTIFY_SOURCE=2 $(EXTRA_WARNINGS) +CFLAGS = -ggdb3 -Wall -Wextra -std=gnu99 -Werror $(CFLAGS_OPTIMIZE) -D_FORTIFY_SOURCE=2 $(EXTRA_WARNINGS) $(EXTRA_CFLAGS) EXTLIBS = -lpthread -lrt -lelf -lm ALL_CFLAGS = $(CFLAGS) ALL_LDFLAGS = $(LDFLAGS) @@ -261,7 +250,7 @@ PTHREAD_LIBS = -lpthread # explicitly what architecture to check for. Fix this up for yours.. SPARSE_FLAGS = -D__BIG_ENDIAN__ -D__powerpc__ -ifeq ($(shell sh -c "echo 'int foo(void) {char X[2]; return 3;}' | $(CC) -x c -c -Werror -fstack-protector-all - -o /dev/null >/dev/null 2>&1 && echo y"), y) +ifeq ($(shell sh -c "echo 'int foo(void) {char X[2]; return 3;}' | $(CC) -x c -c -Werror -fstack-protector-all - -o /dev/null "$(QUIET_STDERR)" && echo y"), y) CFLAGS := $(CFLAGS) -fstack-protector-all endif @@ -380,6 +369,8 @@ LIB_H += util/sort.h LIB_H += util/hist.h LIB_H += util/thread.h LIB_H += util/data_map.h +LIB_H += util/probe-finder.h +LIB_H += util/probe-event.h LIB_OBJS += util/abspath.o LIB_OBJS += util/alias.o @@ -418,10 +409,12 @@ LIB_OBJS += util/thread.o LIB_OBJS += util/trace-event-parse.o LIB_OBJS += util/trace-event-read.o LIB_OBJS += util/trace-event-info.o +LIB_OBJS += util/trace-event-perl.o LIB_OBJS += util/svghelper.o LIB_OBJS += util/sort.o LIB_OBJS += util/hist.o LIB_OBJS += util/data_map.o +LIB_OBJS += util/probe-event.o BUILTIN_OBJS += builtin-annotate.o @@ -430,6 +423,7 @@ BUILTIN_OBJS += builtin-bench.o # Benchmark modules BUILTIN_OBJS += bench/sched-messaging.o BUILTIN_OBJS += bench/sched-pipe.o +BUILTIN_OBJS += bench/mem-memcpy.o BUILTIN_OBJS += builtin-help.o BUILTIN_OBJS += builtin-sched.o @@ -442,9 +436,15 @@ BUILTIN_OBJS += builtin-timechart.o BUILTIN_OBJS += builtin-top.o BUILTIN_OBJS += builtin-trace.o BUILTIN_OBJS += builtin-probe.o +BUILTIN_OBJS += builtin-kmem.o PERFLIBS = $(LIB_FILE) +ifeq ($(V), 2) + QUIET_STDERR = ">/dev/null" +else + QUIET_STDERR = ">/dev/null 2>&1" +endif # # Platform specific tweaks # @@ -472,49 +472,58 @@ ifeq ($(uname_S),Darwin) PTHREAD_LIBS = endif -ifeq ($(shell sh -c "(echo '\#include <libelf.h>'; echo 'int main(void) { Elf * elf = elf_begin(0, ELF_C_READ, 0); return (long)elf; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) > /dev/null 2>&1 && echo y"), y) -ifneq ($(shell sh -c "(echo '\#include <gnu/libc-version.h>'; echo 'int main(void) { const char * version = gnu_get_libc_version(); return (long)version; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) > /dev/null 2>&1 && echo y"), y) - msg := $(error No gnu/libc-version.h found, please install glibc-dev[el]); +ifeq ($(shell sh -c "(echo '\#include <libelf.h>'; echo 'int main(void) { Elf * elf = elf_begin(0, ELF_C_READ, 0); return (long)elf; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) "$(QUIET_STDERR)" && echo y"), y) +ifneq ($(shell sh -c "(echo '\#include <gnu/libc-version.h>'; echo 'int main(void) { const char * version = gnu_get_libc_version(); return (long)version; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) "$(QUIET_STDERR)" && echo y"), y) + msg := $(error No gnu/libc-version.h found, please install glibc-dev[el]/glibc-static); endif - ifneq ($(shell sh -c "(echo '\#include <libelf.h>'; echo 'int main(void) { Elf * elf = elf_begin(0, ELF_C_READ_MMAP, 0); return (long)elf; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) > /dev/null 2>&1 && echo y"), y) + ifneq ($(shell sh -c "(echo '\#include <libelf.h>'; echo 'int main(void) { Elf * elf = elf_begin(0, ELF_C_READ_MMAP, 0); return (long)elf; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) "$(QUIET_STDERR)" && echo y"), y) BASIC_CFLAGS += -DLIBELF_NO_MMAP endif else msg := $(error No libelf.h/libelf found, please install libelf-dev/elfutils-libelf-devel and glibc-dev[el]); endif -ifneq ($(shell sh -c "(echo '\#include <libdwarf/dwarf.h>'; echo '\#include <libdwarf/libdwarf.h>'; echo 'int main(void) { Dwarf_Debug dbg; Dwarf_Error err; Dwarf_Ranges *rng; dwarf_init(0, DW_DLC_READ, 0, 0, &dbg, &err); dwarf_get_ranges(dbg, 0, &rng, 0, 0, &err); return (long)dbg; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -ldwarf -lelf -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) > /dev/null 2>&1 && echo y"), y) +ifneq ($(shell sh -c "(echo '\#include <libdwarf/dwarf.h>'; echo '\#include <libdwarf/libdwarf.h>'; echo 'int main(void) { Dwarf_Debug dbg; Dwarf_Error err; Dwarf_Ranges *rng; dwarf_init(0, DW_DLC_READ, 0, 0, &dbg, &err); dwarf_get_ranges(dbg, 0, &rng, 0, 0, &err); return (long)dbg; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -ldwarf -lelf -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) "$(QUIET_STDERR)" && echo y"), y) msg := $(warning No libdwarf.h found or old libdwarf.h found, disables dwarf support. Please install libdwarf-dev/libdwarf-devel >= 20081231); BASIC_CFLAGS += -DNO_LIBDWARF else EXTLIBS += -lelf -ldwarf - LIB_H += util/probe-finder.h LIB_OBJS += util/probe-finder.o endif +PERL_EMBED_LDOPTS = `perl -MExtUtils::Embed -e ldopts 2>/dev/null` +PERL_EMBED_CCOPTS = `perl -MExtUtils::Embed -e ccopts 2>/dev/null` + +ifneq ($(shell sh -c "(echo '\#include <EXTERN.h>'; echo '\#include <perl.h>'; echo 'int main(void) { perl_alloc(); return 0; }') | $(CC) -x c - $(PERL_EMBED_CCOPTS) -o /dev/null $(PERL_EMBED_LDOPTS) > /dev/null 2>&1 && echo y"), y) + BASIC_CFLAGS += -DNO_LIBPERL +else + ALL_LDFLAGS += $(PERL_EMBED_LDOPTS) + LIB_OBJS += scripts/perl/Perf-Trace-Util/Context.o +endif + ifdef NO_DEMANGLE BASIC_CFLAGS += -DNO_DEMANGLE else - has_bfd := $(shell sh -c "(echo '\#include <bfd.h>'; echo 'int main(void) { bfd_demangle(0, 0, 0); return 0; }') | $(CC) -x c - $(ALL_CFLAGS) -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) -lbfd > /dev/null 2>&1 && echo y") + has_bfd := $(shell sh -c "(echo '\#include <bfd.h>'; echo 'int main(void) { bfd_demangle(0, 0, 0); return 0; }') | $(CC) -x c - $(ALL_CFLAGS) -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) -lbfd "$(QUIET_STDERR)" && echo y") ifeq ($(has_bfd),y) EXTLIBS += -lbfd else - has_bfd_iberty := $(shell sh -c "(echo '\#include <bfd.h>'; echo 'int main(void) { bfd_demangle(0, 0, 0); return 0; }') | $(CC) -x c - $(ALL_CFLAGS) -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) -lbfd -liberty > /dev/null 2>&1 && echo y") + has_bfd_iberty := $(shell sh -c "(echo '\#include <bfd.h>'; echo 'int main(void) { bfd_demangle(0, 0, 0); return 0; }') | $(CC) -x c - $(ALL_CFLAGS) -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) -lbfd -liberty "$(QUIET_STDERR)" && echo y") ifeq ($(has_bfd_iberty),y) EXTLIBS += -lbfd -liberty else - has_bfd_iberty_z := $(shell sh -c "(echo '\#include <bfd.h>'; echo 'int main(void) { bfd_demangle(0, 0, 0); return 0; }') | $(CC) -x c - $(ALL_CFLAGS) -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) -lbfd -liberty -lz > /dev/null 2>&1 && echo y") + has_bfd_iberty_z := $(shell sh -c "(echo '\#include <bfd.h>'; echo 'int main(void) { bfd_demangle(0, 0, 0); return 0; }') | $(CC) -x c - $(ALL_CFLAGS) -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) -lbfd -liberty -lz "$(QUIET_STDERR)" && echo y") ifeq ($(has_bfd_iberty_z),y) EXTLIBS += -lbfd -liberty -lz else - has_cplus_demangle := $(shell sh -c "(echo 'extern char *cplus_demangle(const char *, int);'; echo 'int main(void) { cplus_demangle(0, 0); return 0; }') | $(CC) -x c - $(ALL_CFLAGS) -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) -liberty > /dev/null 2>&1 && echo y") + has_cplus_demangle := $(shell sh -c "(echo 'extern char *cplus_demangle(const char *, int);'; echo 'int main(void) { cplus_demangle(0, 0); return 0; }') | $(CC) -x c - $(ALL_CFLAGS) -o /dev/null $(ALL_LDFLAGS) $(EXTLIBS) -liberty "$(QUIET_STDERR)" && echo y") ifeq ($(has_cplus_demangle),y) EXTLIBS += -liberty BASIC_CFLAGS += -DHAVE_CPLUS_DEMANGLE else - msg := $(warning No bfd.h/libbfd found, install binutils-dev[el] to gain symbol demangling) + msg := $(warning No bfd.h/libbfd found, install binutils-dev[el]/zlib-static to gain symbol demangling) BASIC_CFLAGS += -DNO_DEMANGLE endif endif @@ -864,6 +873,12 @@ util/hweight.o: ../../lib/hweight.c PERF-CFLAGS util/find_next_bit.o: ../../lib/find_next_bit.c PERF-CFLAGS $(QUIET_CC)$(CC) -o util/find_next_bit.o -c $(ALL_CFLAGS) -DETC_PERFCONFIG='"$(ETC_PERFCONFIG_SQ)"' $< +util/trace-event-perl.o: util/trace-event-perl.c PERF-CFLAGS + $(QUIET_CC)$(CC) -o util/trace-event-perl.o -c $(ALL_CFLAGS) $(PERL_EMBED_CCOPTS) -Wno-redundant-decls -Wno-strict-prototypes -Wno-unused-parameter -Wno-shadow $< + +scripts/perl/Perf-Trace-Util/Context.o: scripts/perl/Perf-Trace-Util/Context.c PERF-CFLAGS + $(QUIET_CC)$(CC) -o scripts/perl/Perf-Trace-Util/Context.o -c $(ALL_CFLAGS) $(PERL_EMBED_CCOPTS) -Wno-redundant-decls -Wno-strict-prototypes -Wno-unused-parameter -Wno-nested-externs $< + perf-%$X: %.o $(PERFLIBS) $(QUIET_LINK)$(CC) $(ALL_CFLAGS) -o $@ $(ALL_LDFLAGS) $(filter %.o,$^) $(LIBS) @@ -971,6 +986,13 @@ export perfexec_instdir install: all $(INSTALL) -d -m 755 '$(DESTDIR_SQ)$(bindir_SQ)' $(INSTALL) perf$X '$(DESTDIR_SQ)$(bindir_SQ)' + $(INSTALL) -d -m 755 '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util/lib/Perf/Trace' + $(INSTALL) -d -m 755 '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/bin' + $(INSTALL) scripts/perl/Perf-Trace-Util/lib/Perf/Trace/* -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util/lib/Perf/Trace' + $(INSTALL) scripts/perl/*.pl -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl' + $(INSTALL) scripts/perl/bin/* -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/bin' + $(INSTALL) scripts/perl/Perf-Trace-Util/Makefile.PL -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util' + $(INSTALL) scripts/perl/Perf-Trace-Util/README -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util' ifdef BUILT_INS $(INSTALL) -d -m 755 '$(DESTDIR_SQ)$(perfexec_instdir_SQ)' $(INSTALL) $(BUILT_INS) '$(DESTDIR_SQ)$(perfexec_instdir_SQ)' @@ -1056,7 +1078,7 @@ distclean: clean # $(RM) configure clean: - $(RM) *.o */*.o $(LIB_FILE) + $(RM) *.o */*.o */*/*.o */*/*/*.o $(LIB_FILE) $(RM) $(ALL_PROGRAMS) $(BUILT_INS) perf$X $(RM) $(TEST_PROGRAMS) $(RM) *.spec *.pyc *.pyo */*.pyc */*.pyo common-cmds.h TAGS tags cscope* diff --git a/tools/perf/bench/bench.h b/tools/perf/bench/bench.h index 9fbd8d745fa..f7781c6267c 100644 --- a/tools/perf/bench/bench.h +++ b/tools/perf/bench/bench.h @@ -3,6 +3,7 @@ extern int bench_sched_messaging(int argc, const char **argv, const char *prefix); extern int bench_sched_pipe(int argc, const char **argv, const char *prefix); +extern int bench_mem_memcpy(int argc, const char **argv, const char *prefix __used); #define BENCH_FORMAT_DEFAULT_STR "default" #define BENCH_FORMAT_DEFAULT 0 diff --git a/tools/perf/bench/mem-memcpy.c b/tools/perf/bench/mem-memcpy.c new file mode 100644 index 00000000000..89773178e89 --- /dev/null +++ b/tools/perf/bench/mem-memcpy.c @@ -0,0 +1,193 @@ +/* + * mem-memcpy.c + * + * memcpy: Simple memory copy in various ways + * + * Written by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp> + */ +#include <ctype.h> + +#include "../perf.h" +#include "../util/util.h" +#include "../util/parse-options.h" +#include "../util/string.h" +#include "../util/header.h" +#include "bench.h" + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <sys/time.h> +#include <errno.h> + +#define K 1024 + +static const char *length_str = "1MB"; +static const char *routine = "default"; +static int use_clock = 0; +static int clock_fd; + +static const struct option options[] = { + OPT_STRING('l', "length", &length_str, "1MB", + "Specify length of memory to copy. " + "available unit: B, MB, GB (upper and lower)"), + OPT_STRING('r', "routine", &routine, "default", + "Specify routine to copy"), + OPT_BOOLEAN('c', "clock", &use_clock, + "Use CPU clock for measuring"), + OPT_END() +}; + +struct routine { + const char *name; + const char *desc; + void * (*fn)(void *dst, const void *src, size_t len); +}; + +struct routine routines[] = { + { "default", + "Default memcpy() provided by glibc", + memcpy }, + { NULL, + NULL, + NULL } +}; + +static const char * const bench_mem_memcpy_usage[] = { + "perf bench mem memcpy <options>", + NULL +}; + +static struct perf_event_attr clock_attr = { + .type = PERF_TYPE_HARDWARE, + .config = PERF_COUNT_HW_CPU_CYCLES +}; + +static void init_clock(void) +{ + clock_fd = sys_perf_event_open(&clock_attr, getpid(), -1, -1, 0); + + if (clock_fd < 0 && errno == ENOSYS) + die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); + else + BUG_ON(clock_fd < 0); +} + +static u64 get_clock(void) +{ + int ret; + u64 clk; + + ret = read(clock_fd, &clk, sizeof(u64)); + BUG_ON(ret != sizeof(u64)); + + return clk; +} + +static double timeval2double(struct timeval *ts) +{ + return (double)ts->tv_sec + + (double)ts->tv_usec / (double)1000000; +} + +int bench_mem_memcpy(int argc, const char **argv, + const char *prefix __used) +{ + int i; + void *dst, *src; + size_t length; + double bps = 0.0; + struct timeval tv_start, tv_end, tv_diff; + u64 clock_start, clock_end, clock_diff; + + clock_start = clock_end = clock_diff = 0ULL; + argc = parse_options(argc, argv, options, + bench_mem_memcpy_usage, 0); + + tv_diff.tv_sec = 0; + tv_diff.tv_usec = 0; + length = (size_t)perf_atoll((char *)length_str); + + if ((s64)length <= 0) { + fprintf(stderr, "Invalid length:%s\n", length_str); + return 1; + } + + for (i = 0; routines[i].name; i++) { + if (!strcmp(routines[i].name, routine)) + break; + } + if (!routines[i].name) { + printf("Unknown routine:%s\n", routine); + printf("Available routines...\n"); + for (i = 0; routines[i].name; i++) { + printf("\t%s ... %s\n", + routines[i].name, routines[i].desc); + } + return 1; + } + + dst = zalloc(length); + if (!dst) + die("memory allocation failed - maybe length is too large?\n"); + + src = zalloc(length); + if (!src) + die("memory allocation failed - maybe length is too large?\n"); + + if (bench_format == BENCH_FORMAT_DEFAULT) { + printf("# Copying %s Bytes from %p to %p ...\n\n", + length_str, src, dst); + } + + if (use_clock) { + init_clock(); + clock_start = get_clock(); + } else { + BUG_ON(gettimeofday(&tv_start, NULL)); + } + + routines[i].fn(dst, src, length); + + if (use_clock) { + clock_end = get_clock(); + clock_diff = clock_end - clock_start; + } else { + BUG_ON(gettimeofday(&tv_end, NULL)); + timersub(&tv_end, &tv_start, &tv_diff); + bps = (double)((double)length / timeval2double(&tv_diff)); + } + + switch (bench_format) { + case BENCH_FORMAT_DEFAULT: + if (use_clock) { + printf(" %14lf Clock/Byte\n", + (double)clock_diff / (double)length); + } else { + if (bps < K) + printf(" %14lf B/Sec\n", bps); + else if (bps < K * K) + printf(" %14lfd KB/Sec\n", bps / 1024); + else if (bps < K * K * K) + printf(" %14lf MB/Sec\n", bps / 1024 / 1024); + else { + printf(" %14lf GB/Sec\n", + bps / 1024 / 1024 / 1024); + } + } + break; + case BENCH_FORMAT_SIMPLE: + if (use_clock) { + printf("%14lf\n", + (double)clock_diff / (double)length); + } else + printf("%lf\n", bps); + break; + default: + /* reaching this means there's some disaster: */ + die("unknown format: %d\n", bench_format); + break; + } + + return 0; +} diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 77d50a6d680..0bf2e8f9af5 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -19,24 +19,22 @@ #include "perf.h" #include "util/debug.h" +#include "util/event.h" #include "util/parse-options.h" #include "util/parse-events.h" #include "util/thread.h" #include "util/sort.h" #include "util/hist.h" +#include "util/data_map.h" static char const *input_name = "perf.data"; static int force; -static int input; static int full_paths; static int print_line; -static unsigned long page_size; -static unsigned long mmap_window = 32; - struct sym_hist { u64 sum; u64 ip[0]; @@ -53,6 +51,11 @@ struct sym_priv { struct sym_ext *ext; }; +static struct symbol_conf symbol_conf = { + .priv_size = sizeof(struct sym_priv), + .try_vmlinux_path = true, +}; + static const char *sym_hist_filter; static int symbol_filter(struct map *map __used, struct symbol *sym) @@ -118,186 +121,34 @@ static void hist_hit(struct hist_entry *he, u64 ip) h->ip[offset]); } -static int hist_entry__add(struct thread *thread, struct map *map, - struct symbol *sym, u64 ip, u64 count, char level) +static int hist_entry__add(struct addr_location *al, u64 count) { bool hit; - struct hist_entry *he = __hist_entry__add(thread, map, sym, NULL, ip, - count, level, &hit); + struct hist_entry *he = __hist_entry__add(al, NULL, count, &hit); if (he == NULL) return -ENOMEM; - hist_hit(he, ip); + hist_hit(he, al->addr); return 0; } -static int -process_sample_event(event_t *event, unsigned long offset, unsigned long head) +static int process_sample_event(event_t *event) { - char level; - u64 ip = event->ip.ip; - struct map *map = NULL; - struct symbol *sym = NULL; - struct thread *thread = threads__findnew(event->ip.pid); - - dump_printf("%p [%p]: PERF_EVENT (IP, %d): %d: %p\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->header.misc, - event->ip.pid, - (void *)(long)ip); - - if (thread == NULL) { + struct addr_location al; + + dump_printf("(IP, %d): %d: %p\n", event->header.misc, + event->ip.pid, (void *)(long)event->ip.ip); + + if (event__preprocess_sample(event, &al, symbol_filter) < 0) { fprintf(stderr, "problem processing %d event, skipping it.\n", event->header.type); return -1; } - dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); - - if (event->header.misc & PERF_RECORD_MISC_KERNEL) { - level = 'k'; - sym = kernel_maps__find_symbol(ip, &map); - dump_printf(" ...... dso: %s\n", - map ? map->dso->long_name : "<not found>"); - } else if (event->header.misc & PERF_RECORD_MISC_USER) { - level = '.'; - map = thread__find_map(thread, ip); - if (map != NULL) { -got_map: - ip = map->map_ip(map, ip); - sym = map__find_symbol(map, ip, symbol_filter); - } else { - /* - * If this is outside of all known maps, - * and is a negative address, try to look it - * up in the kernel dso, as it might be a - * vsyscall or vdso (which executes in user-mode). - * - * XXX This is nasty, we should have a symbol list in - * the "[vdso]" dso, but for now lets use the old - * trick of looking in the whole kernel symbol list. - */ - if ((long long)ip < 0) { - map = kernel_map; - goto got_map; - } - } - dump_printf(" ...... dso: %s\n", - map ? map->dso->long_name : "<not found>"); - } else { - level = 'H'; - dump_printf(" ...... dso: [hypervisor]\n"); - } - - if (hist_entry__add(thread, map, sym, ip, 1, level)) { + if (hist_entry__add(&al, 1)) { fprintf(stderr, "problem incrementing symbol count, " "skipping event\n"); return -1; } - total++; - - return 0; -} - -static int -process_mmap_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct map *map = map__new(&event->mmap, NULL, 0); - struct thread *thread = threads__findnew(event->mmap.pid); - - dump_printf("%p [%p]: PERF_RECORD_MMAP %d: [%p(%p) @ %p]: %s\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->mmap.pid, - (void *)(long)event->mmap.start, - (void *)(long)event->mmap.len, - (void *)(long)event->mmap.pgoff, - event->mmap.filename); - - if (thread == NULL || map == NULL) { - dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); - return 0; - } - - thread__insert_map(thread, map); - total_mmap++; - - return 0; -} - -static int -process_comm_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct thread *thread = threads__findnew(event->comm.pid); - - dump_printf("%p [%p]: PERF_RECORD_COMM: %s:%d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->comm.comm, event->comm.pid); - - if (thread == NULL || - thread__set_comm(thread, event->comm.comm)) { - dump_printf("problem processing PERF_RECORD_COMM, skipping event.\n"); - return -1; - } - total_comm++; - - return 0; -} - -static int -process_fork_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct thread *thread = threads__findnew(event->fork.pid); - struct thread *parent = threads__findnew(event->fork.ppid); - - dump_printf("%p [%p]: PERF_RECORD_FORK: %d:%d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->fork.pid, event->fork.ppid); - - /* - * A thread clone will have the same PID for both - * parent and child. - */ - if (thread == parent) - return 0; - - if (!thread || !parent || thread__fork(thread, parent)) { - dump_printf("problem processing PERF_RECORD_FORK, skipping event.\n"); - return -1; - } - total_fork++; - - return 0; -} - -static int -process_event(event_t *event, unsigned long offset, unsigned long head) -{ - switch (event->header.type) { - case PERF_RECORD_SAMPLE: - return process_sample_event(event, offset, head); - - case PERF_RECORD_MMAP: - return process_mmap_event(event, offset, head); - - case PERF_RECORD_COMM: - return process_comm_event(event, offset, head); - - case PERF_RECORD_FORK: - return process_fork_event(event, offset, head); - /* - * We dont process them right now but they are fine: - */ - - case PERF_RECORD_THROTTLE: - case PERF_RECORD_UNTHROTTLE: - return 0; - - default: - return -1; - } return 0; } @@ -602,115 +453,31 @@ static void find_annotations(void) } } +static struct perf_file_handler file_handler = { + .process_sample_event = process_sample_event, + .process_mmap_event = event__process_mmap, + .process_comm_event = event__process_comm, + .process_fork_event = event__process_task, +}; + static int __cmd_annotate(void) { - int ret, rc = EXIT_FAILURE; - unsigned long offset = 0; - unsigned long head = 0; - struct stat input_stat; - event_t *event; - uint32_t size; - char *buf; - - register_idle_thread(); - - input = open(input_name, O_RDONLY); - if (input < 0) { - perror("failed to open file"); - exit(-1); - } - - ret = fstat(input, &input_stat); - if (ret < 0) { - perror("failed to stat file"); - exit(-1); - } - - if (!force && input_stat.st_uid && (input_stat.st_uid != geteuid())) { - fprintf(stderr, "file: %s not owned by current user or root\n", input_name); - exit(-1); - } - - if (!input_stat.st_size) { - fprintf(stderr, "zero-sized file, nothing to do!\n"); - exit(0); - } - - if (load_kernel(symbol_filter) < 0) { - perror("failed to load kernel symbols"); - return EXIT_FAILURE; - } - -remap: - buf = (char *)mmap(NULL, page_size * mmap_window, PROT_READ, - MAP_SHARED, input, offset); - if (buf == MAP_FAILED) { - perror("failed to mmap file"); - exit(-1); - } - -more: - event = (event_t *)(buf + head); - - size = event->header.size; - if (!size) - size = 8; - - if (head + event->header.size >= page_size * mmap_window) { - unsigned long shift = page_size * (head / page_size); - int munmap_ret; - - munmap_ret = munmap(buf, page_size * mmap_window); - assert(munmap_ret == 0); - - offset += shift; - head -= shift; - goto remap; - } - - size = event->header.size; - - dump_printf("%p [%p]: event: %d\n", - (void *)(offset + head), - (void *)(long)event->header.size, - event->header.type); - - if (!size || process_event(event, offset, head) < 0) { - - dump_printf("%p [%p]: skipping unknown header type: %d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->header.type); - - total_unknown++; - - /* - * assume we lost track of the stream, check alignment, and - * increment a single u64 in the hope to catch on again 'soon'. - */ - - if (unlikely(head & 7)) - head &= ~7ULL; - - size = 8; - } - - head += size; - - if (offset + head < (unsigned long)input_stat.st_size) - goto more; + struct perf_header *header; + struct thread *idle; + int ret; - rc = EXIT_SUCCESS; - close(input); + idle = register_idle_thread(); + register_perf_file_handler(&file_handler); - dump_printf(" IP events: %10ld\n", total); - dump_printf(" mmap events: %10ld\n", total_mmap); - dump_printf(" comm events: %10ld\n", total_comm); - dump_printf(" fork events: %10ld\n", total_fork); - dump_printf(" unknown events: %10ld\n", total_unknown); + ret = mmap_dispatch_perf_file(&header, input_name, 0, 0, + &event__cwdlen, &event__cwd); + if (ret) + return ret; - if (dump_trace) + if (dump_trace) { + event__print_totals(); return 0; + } if (verbose > 3) threads__fprintf(stdout); @@ -719,11 +486,11 @@ more: dsos__fprintf(stdout); collapse__resort(); - output__resort(total); + output__resort(event__total[0]); find_annotations(); - return rc; + return ret; } static const char * const annotate_usage[] = { @@ -741,8 +508,9 @@ static const struct option options[] = { "be more verbose (show symbol address, etc)"), OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, "dump raw trace in ASCII"), - OPT_STRING('k', "vmlinux", &vmlinux_name, "file", "vmlinux pathname"), - OPT_BOOLEAN('m', "modules", &modules, + OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, + "file", "vmlinux pathname"), + OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules, "load module symbols - WARNING: use only with -k and LIVE kernel"), OPT_BOOLEAN('l', "print-line", &print_line, "print matching source lines (may be slow)"), @@ -768,9 +536,8 @@ static void setup_sorting(void) int cmd_annotate(int argc, const char **argv, const char *prefix __used) { - symbol__init(sizeof(struct sym_priv)); - - page_size = getpagesize(); + if (symbol__init(&symbol_conf) < 0) + return -1; argc = parse_options(argc, argv, options, annotate_usage, 0); diff --git a/tools/perf/builtin-bench.c b/tools/perf/builtin-bench.c index 90c39baae0d..e043eb83092 100644 --- a/tools/perf/builtin-bench.c +++ b/tools/perf/builtin-bench.c @@ -12,6 +12,7 @@ * * Available subsystem list: * sched ... scheduler and IPC mechanism + * mem ... memory access performance * */ @@ -43,6 +44,15 @@ static struct bench_suite sched_suites[] = { NULL } }; +static struct bench_suite mem_suites[] = { + { "memcpy", + "Simple memory copy in various ways", + bench_mem_memcpy }, + { NULL, + NULL, + NULL } +}; + struct bench_subsys { const char *name; const char *summary; @@ -53,9 +63,12 @@ static struct bench_subsys subsystems[] = { { "sched", "scheduler and IPC mechanism", sched_suites }, + { "mem", + "memory access performance", + mem_suites }, { NULL, NULL, - NULL } + NULL } }; static void dump_suites(int subsys_index) diff --git a/tools/perf/builtin-help.c b/tools/perf/builtin-help.c index 768f9c82631..9f810b17c25 100644 --- a/tools/perf/builtin-help.c +++ b/tools/perf/builtin-help.c @@ -179,7 +179,7 @@ static void add_man_viewer(const char *name) while (*p) p = &((*p)->next); - *p = calloc(1, (sizeof(**p) + len + 1)); + *p = zalloc(sizeof(**p) + len + 1); strncpy((*p)->name, name, len); } @@ -194,7 +194,7 @@ static void do_add_man_viewer_info(const char *name, size_t len, const char *value) { - struct man_viewer_info_list *new = calloc(1, sizeof(*new) + len + 1); + struct man_viewer_info_list *new = zalloc(sizeof(*new) + len + 1); strncpy(new->name, name, len); new->info = strdup(value); diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c new file mode 100644 index 00000000000..047fef74bd5 --- /dev/null +++ b/tools/perf/builtin-kmem.c @@ -0,0 +1,807 @@ +#include "builtin.h" +#include "perf.h" + +#include "util/util.h" +#include "util/cache.h" +#include "util/symbol.h" +#include "util/thread.h" +#include "util/header.h" + +#include "util/parse-options.h" +#include "util/trace-event.h" + +#include "util/debug.h" +#include "util/data_map.h" + +#include <linux/rbtree.h> + +struct alloc_stat; +typedef int (*sort_fn_t)(struct alloc_stat *, struct alloc_stat *); + +static char const *input_name = "perf.data"; + +static struct perf_header *header; +static u64 sample_type; + +static int alloc_flag; +static int caller_flag; + +static int alloc_lines = -1; +static int caller_lines = -1; + +static bool raw_ip; + +static char default_sort_order[] = "frag,hit,bytes"; + +static int *cpunode_map; +static int max_cpu_num; + +struct alloc_stat { + u64 call_site; + u64 ptr; + u64 bytes_req; + u64 bytes_alloc; + u32 hit; + u32 pingpong; + + short alloc_cpu; + + struct rb_node node; +}; + +static struct rb_root root_alloc_stat; +static struct rb_root root_alloc_sorted; +static struct rb_root root_caller_stat; +static struct rb_root root_caller_sorted; + +static unsigned long total_requested, total_allocated; +static unsigned long nr_allocs, nr_cross_allocs; + +struct raw_event_sample { + u32 size; + char data[0]; +}; + +#define PATH_SYS_NODE "/sys/devices/system/node" + +static void init_cpunode_map(void) +{ + FILE *fp; + int i; + + fp = fopen("/sys/devices/system/cpu/kernel_max", "r"); + if (!fp) { + max_cpu_num = 4096; + return; + } + + if (fscanf(fp, "%d", &max_cpu_num) < 1) + die("Failed to read 'kernel_max' from sysfs"); + max_cpu_num++; + + cpunode_map = calloc(max_cpu_num, sizeof(int)); + if (!cpunode_map) + die("calloc"); + for (i = 0; i < max_cpu_num; i++) + cpunode_map[i] = -1; + fclose(fp); +} + +static void setup_cpunode_map(void) +{ + struct dirent *dent1, *dent2; + DIR *dir1, *dir2; + unsigned int cpu, mem; + char buf[PATH_MAX]; + + init_cpunode_map(); + + dir1 = opendir(PATH_SYS_NODE); + if (!dir1) + return; + + while (true) { + dent1 = readdir(dir1); + if (!dent1) + break; + + if (sscanf(dent1->d_name, "node%u", &mem) < 1) + continue; + + snprintf(buf, PATH_MAX, "%s/%s", PATH_SYS_NODE, dent1->d_name); + dir2 = opendir(buf); + if (!dir2) + continue; + while (true) { + dent2 = readdir(dir2); + if (!dent2) + break; + if (sscanf(dent2->d_name, "cpu%u", &cpu) < 1) + continue; + cpunode_map[cpu] = mem; + } + } +} + +static void insert_alloc_stat(unsigned long call_site, unsigned long ptr, + int bytes_req, int bytes_alloc, int cpu) +{ + struct rb_node **node = &root_alloc_stat.rb_node; + struct rb_node *parent = NULL; + struct alloc_stat *data = NULL; + + while (*node) { + parent = *node; + data = rb_entry(*node, struct alloc_stat, node); + + if (ptr > data->ptr) + node = &(*node)->rb_right; + else if (ptr < data->ptr) + node = &(*node)->rb_left; + else + break; + } + + if (data && data->ptr == ptr) { + data->hit++; + data->bytes_req += bytes_req; + data->bytes_alloc += bytes_req; + } else { + data = malloc(sizeof(*data)); + if (!data) + die("malloc"); + data->ptr = ptr; + data->pingpong = 0; + data->hit = 1; + data->bytes_req = bytes_req; + data->bytes_alloc = bytes_alloc; + + rb_link_node(&data->node, parent, node); + rb_insert_color(&data->node, &root_alloc_stat); + } + data->call_site = call_site; + data->alloc_cpu = cpu; +} + +static void insert_caller_stat(unsigned long call_site, + int bytes_req, int bytes_alloc) +{ + struct rb_node **node = &root_caller_stat.rb_node; + struct rb_node *parent = NULL; + struct alloc_stat *data = NULL; + + while (*node) { + parent = *node; + data = rb_entry(*node, struct alloc_stat, node); + + if (call_site > data->call_site) + node = &(*node)->rb_right; + else if (call_site < data->call_site) + node = &(*node)->rb_left; + else + break; + } + + if (data && data->call_site == call_site) { + data->hit++; + data->bytes_req += bytes_req; + data->bytes_alloc += bytes_req; + } else { + data = malloc(sizeof(*data)); + if (!data) + die("malloc"); + data->call_site = call_site; + data->pingpong = 0; + data->hit = 1; + data->bytes_req = bytes_req; + data->bytes_alloc = bytes_alloc; + + rb_link_node(&data->node, parent, node); + rb_insert_color(&data->node, &root_caller_stat); + } +} + +static void process_alloc_event(struct raw_event_sample *raw, + struct event *event, + int cpu, + u64 timestamp __used, + struct thread *thread __used, + int node) +{ + unsigned long call_site; + unsigned long ptr; + int bytes_req; + int bytes_alloc; + int node1, node2; + + ptr = raw_field_value(event, "ptr", raw->data); + call_site = raw_field_value(event, "call_site", raw->data); + bytes_req = raw_field_value(event, "bytes_req", raw->data); + bytes_alloc = raw_field_value(event, "bytes_alloc", raw->data); + + insert_alloc_stat(call_site, ptr, bytes_req, bytes_alloc, cpu); + insert_caller_stat(call_site, bytes_req, bytes_alloc); + + total_requested += bytes_req; + total_allocated += bytes_alloc; + + if (node) { + node1 = cpunode_map[cpu]; + node2 = raw_field_value(event, "node", raw->data); + if (node1 != node2) + nr_cross_allocs++; + } + nr_allocs++; +} + +static int ptr_cmp(struct alloc_stat *, struct alloc_stat *); +static int callsite_cmp(struct alloc_stat *, struct alloc_stat *); + +static struct alloc_stat *search_alloc_stat(unsigned long ptr, + unsigned long call_site, + struct rb_root *root, + sort_fn_t sort_fn) +{ + struct rb_node *node = root->rb_node; + struct alloc_stat key = { .ptr = ptr, .call_site = call_site }; + + while (node) { + struct alloc_stat *data; + int cmp; + + data = rb_entry(node, struct alloc_stat, node); + + cmp = sort_fn(&key, data); + if (cmp < 0) + node = node->rb_left; + else if (cmp > 0) + node = node->rb_right; + else + return data; + } + return NULL; +} + +static void process_free_event(struct raw_event_sample *raw, + struct event *event, + int cpu, + u64 timestamp __used, + struct thread *thread __used) +{ + unsigned long ptr; + struct alloc_stat *s_alloc, *s_caller; + + ptr = raw_field_value(event, "ptr", raw->data); + + s_alloc = search_alloc_stat(ptr, 0, &root_alloc_stat, ptr_cmp); + if (!s_alloc) + return; + + if (cpu != s_alloc->alloc_cpu) { + s_alloc->pingpong++; + + s_caller = search_alloc_stat(0, s_alloc->call_site, + &root_caller_stat, callsite_cmp); + assert(s_caller); + s_caller->pingpong++; + } + s_alloc->alloc_cpu = -1; +} + +static void +process_raw_event(event_t *raw_event __used, void *more_data, + int cpu, u64 timestamp, struct thread *thread) +{ + struct raw_event_sample *raw = more_data; + struct event *event; + int type; + + type = trace_parse_common_type(raw->data); + event = trace_find_event(type); + + if (!strcmp(event->name, "kmalloc") || + !strcmp(event->name, "kmem_cache_alloc")) { + process_alloc_event(raw, event, cpu, timestamp, thread, 0); + return; + } + + if (!strcmp(event->name, "kmalloc_node") || + !strcmp(event->name, "kmem_cache_alloc_node")) { + process_alloc_event(raw, event, cpu, timestamp, thread, 1); + return; + } + + if (!strcmp(event->name, "kfree") || + !strcmp(event->name, "kmem_cache_free")) { + process_free_event(raw, event, cpu, timestamp, thread); + return; + } +} + +static int process_sample_event(event_t *event) +{ + u64 ip = event->ip.ip; + u64 timestamp = -1; + u32 cpu = -1; + u64 period = 1; + void *more_data = event->ip.__more_data; + struct thread *thread = threads__findnew(event->ip.pid); + + if (sample_type & PERF_SAMPLE_TIME) { + timestamp = *(u64 *)more_data; + more_data += sizeof(u64); + } + + if (sample_type & PERF_SAMPLE_CPU) { + cpu = *(u32 *)more_data; + more_data += sizeof(u32); + more_data += sizeof(u32); /* reserved */ + } + + if (sample_type & PERF_SAMPLE_PERIOD) { + period = *(u64 *)more_data; + more_data += sizeof(u64); + } + + dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", + event->header.misc, + event->ip.pid, event->ip.tid, + (void *)(long)ip, + (long long)period); + + if (thread == NULL) { + pr_debug("problem processing %d event, skipping it.\n", + event->header.type); + return -1; + } + + dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); + + process_raw_event(event, more_data, cpu, timestamp, thread); + + return 0; +} + +static int sample_type_check(u64 type) +{ + sample_type = type; + + if (!(sample_type & PERF_SAMPLE_RAW)) { + fprintf(stderr, + "No trace sample to read. Did you call perf record " + "without -R?"); + return -1; + } + + return 0; +} + +static struct perf_file_handler file_handler = { + .process_sample_event = process_sample_event, + .process_comm_event = event__process_comm, + .sample_type_check = sample_type_check, +}; + +static int read_events(void) +{ + register_idle_thread(); + register_perf_file_handler(&file_handler); + + return mmap_dispatch_perf_file(&header, input_name, 0, 0, + &event__cwdlen, &event__cwd); +} + +static double fragmentation(unsigned long n_req, unsigned long n_alloc) +{ + if (n_alloc == 0) + return 0.0; + else + return 100.0 - (100.0 * n_req / n_alloc); +} + +static void __print_result(struct rb_root *root, int n_lines, int is_caller) +{ + struct rb_node *next; + + printf("%.102s\n", graph_dotted_line); + printf(" %-34s |", is_caller ? "Callsite": "Alloc Ptr"); + printf(" Total_alloc/Per | Total_req/Per | Hit | Ping-pong | Frag\n"); + printf("%.102s\n", graph_dotted_line); + + next = rb_first(root); + + while (next && n_lines--) { + struct alloc_stat *data = rb_entry(next, struct alloc_stat, + node); + struct symbol *sym = NULL; + char buf[BUFSIZ]; + u64 addr; + + if (is_caller) { + addr = data->call_site; + if (!raw_ip) + sym = thread__find_function(kthread, addr, NULL); + } else + addr = data->ptr; + + if (sym != NULL) + snprintf(buf, sizeof(buf), "%s+%Lx", sym->name, + addr - sym->start); + else + snprintf(buf, sizeof(buf), "%#Lx", addr); + printf(" %-34s |", buf); + + printf(" %9llu/%-5lu | %9llu/%-5lu | %6lu | %8lu | %6.3f%%\n", + (unsigned long long)data->bytes_alloc, + (unsigned long)data->bytes_alloc / data->hit, + (unsigned long long)data->bytes_req, + (unsigned long)data->bytes_req / data->hit, + (unsigned long)data->hit, + (unsigned long)data->pingpong, + fragmentation(data->bytes_req, data->bytes_alloc)); + + next = rb_next(next); + } + + if (n_lines == -1) + printf(" ... | ... | ... | ... | ... | ... \n"); + + printf("%.102s\n", graph_dotted_line); +} + +static void print_summary(void) +{ + printf("\nSUMMARY\n=======\n"); + printf("Total bytes requested: %lu\n", total_requested); + printf("Total bytes allocated: %lu\n", total_allocated); + printf("Total bytes wasted on internal fragmentation: %lu\n", + total_allocated - total_requested); + printf("Internal fragmentation: %f%%\n", + fragmentation(total_requested, total_allocated)); + printf("Cross CPU allocations: %lu/%lu\n", nr_cross_allocs, nr_allocs); +} + +static void print_result(void) +{ + if (caller_flag) + __print_result(&root_caller_sorted, caller_lines, 1); + if (alloc_flag) + __print_result(&root_alloc_sorted, alloc_lines, 0); + print_summary(); +} + +struct sort_dimension { + const char name[20]; + sort_fn_t cmp; + struct list_head list; +}; + +static LIST_HEAD(caller_sort); +static LIST_HEAD(alloc_sort); + +static void sort_insert(struct rb_root *root, struct alloc_stat *data, + struct list_head *sort_list) +{ + struct rb_node **new = &(root->rb_node); + struct rb_node *parent = NULL; + struct sort_dimension *sort; + + while (*new) { + struct alloc_stat *this; + int cmp = 0; + + this = rb_entry(*new, struct alloc_stat, node); + parent = *new; + + list_for_each_entry(sort, sort_list, list) { + cmp = sort->cmp(data, this); + if (cmp) + break; + } + + if (cmp > 0) + new = &((*new)->rb_left); + else + new = &((*new)->rb_right); + } + + rb_link_node(&data->node, parent, new); + rb_insert_color(&data->node, root); +} + +static void __sort_result(struct rb_root *root, struct rb_root *root_sorted, + struct list_head *sort_list) +{ + struct rb_node *node; + struct alloc_stat *data; + + for (;;) { + node = rb_first(root); + if (!node) + break; + + rb_erase(node, root); + data = rb_entry(node, struct alloc_stat, node); + sort_insert(root_sorted, data, sort_list); + } +} + +static void sort_result(void) +{ + __sort_result(&root_alloc_stat, &root_alloc_sorted, &alloc_sort); + __sort_result(&root_caller_stat, &root_caller_sorted, &caller_sort); +} + +static int __cmd_kmem(void) +{ + setup_pager(); + read_events(); + sort_result(); + print_result(); + + return 0; +} + +static const char * const kmem_usage[] = { + "perf kmem [<options>] {record}", + NULL +}; + +static int ptr_cmp(struct alloc_stat *l, struct alloc_stat *r) +{ + if (l->ptr < r->ptr) + return -1; + else if (l->ptr > r->ptr) + return 1; + return 0; +} + +static struct sort_dimension ptr_sort_dimension = { + .name = "ptr", + .cmp = ptr_cmp, +}; + +static int callsite_cmp(struct alloc_stat *l, struct alloc_stat *r) +{ + if (l->call_site < r->call_site) + return -1; + else if (l->call_site > r->call_site) + return 1; + return 0; +} + +static struct sort_dimension callsite_sort_dimension = { + .name = "callsite", + .cmp = callsite_cmp, +}; + +static int hit_cmp(struct alloc_stat *l, struct alloc_stat *r) +{ + if (l->hit < r->hit) + return -1; + else if (l->hit > r->hit) + return 1; + return 0; +} + +static struct sort_dimension hit_sort_dimension = { + .name = "hit", + .cmp = hit_cmp, +}; + +static int bytes_cmp(struct alloc_stat *l, struct alloc_stat *r) +{ + if (l->bytes_alloc < r->bytes_alloc) + return -1; + else if (l->bytes_alloc > r->bytes_alloc) + return 1; + return 0; +} + +static struct sort_dimension bytes_sort_dimension = { + .name = "bytes", + .cmp = bytes_cmp, +}; + +static int frag_cmp(struct alloc_stat *l, struct alloc_stat *r) +{ + double x, y; + + x = fragmentation(l->bytes_req, l->bytes_alloc); + y = fragmentation(r->bytes_req, r->bytes_alloc); + + if (x < y) + return -1; + else if (x > y) + return 1; + return 0; +} + +static struct sort_dimension frag_sort_dimension = { + .name = "frag", + .cmp = frag_cmp, +}; + +static int pingpong_cmp(struct alloc_stat *l, struct alloc_stat *r) +{ + if (l->pingpong < r->pingpong) + return -1; + else if (l->pingpong > r->pingpong) + return 1; + return 0; +} + +static struct sort_dimension pingpong_sort_dimension = { + .name = "pingpong", + .cmp = pingpong_cmp, +}; + +static struct sort_dimension *avail_sorts[] = { + &ptr_sort_dimension, + &callsite_sort_dimension, + &hit_sort_dimension, + &bytes_sort_dimension, + &frag_sort_dimension, + &pingpong_sort_dimension, +}; + +#define NUM_AVAIL_SORTS \ + (int)(sizeof(avail_sorts) / sizeof(struct sort_dimension *)) + +static int sort_dimension__add(const char *tok, struct list_head *list) +{ + struct sort_dimension *sort; + int i; + + for (i = 0; i < NUM_AVAIL_SORTS; i++) { + if (!strcmp(avail_sorts[i]->name, tok)) { + sort = malloc(sizeof(*sort)); + if (!sort) + die("malloc"); + memcpy(sort, avail_sorts[i], sizeof(*sort)); + list_add_tail(&sort->list, list); + return 0; + } + } + + return -1; +} + +static int setup_sorting(struct list_head *sort_list, const char *arg) +{ + char *tok; + char *str = strdup(arg); + + if (!str) + die("strdup"); + + while (true) { + tok = strsep(&str, ","); + if (!tok) + break; + if (sort_dimension__add(tok, sort_list) < 0) { + error("Unknown --sort key: '%s'", tok); + return -1; + } + } + + free(str); + return 0; +} + +static int parse_sort_opt(const struct option *opt __used, + const char *arg, int unset __used) +{ + if (!arg) + return -1; + + if (caller_flag > alloc_flag) + return setup_sorting(&caller_sort, arg); + else + return setup_sorting(&alloc_sort, arg); + + return 0; +} + +static int parse_stat_opt(const struct option *opt __used, + const char *arg, int unset __used) +{ + if (!arg) + return -1; + + if (strcmp(arg, "alloc") == 0) + alloc_flag = (caller_flag + 1); + else if (strcmp(arg, "caller") == 0) + caller_flag = (alloc_flag + 1); + else + return -1; + return 0; +} + +static int parse_line_opt(const struct option *opt __used, + const char *arg, int unset __used) +{ + int lines; + + if (!arg) + return -1; + + lines = strtoul(arg, NULL, 10); + + if (caller_flag > alloc_flag) + caller_lines = lines; + else + alloc_lines = lines; + + return 0; +} + +static const struct option kmem_options[] = { + OPT_STRING('i', "input", &input_name, "file", + "input file name"), + OPT_CALLBACK(0, "stat", NULL, "<alloc>|<caller>", + "stat selector, Pass 'alloc' or 'caller'.", + parse_stat_opt), + OPT_CALLBACK('s', "sort", NULL, "key[,key2...]", + "sort by keys: ptr, call_site, bytes, hit, pingpong, frag", + parse_sort_opt), + OPT_CALLBACK('l', "line", NULL, "num", + "show n lins", + parse_line_opt), + OPT_BOOLEAN(0, "raw-ip", &raw_ip, "show raw ip instead of symbol"), + OPT_END() +}; + +static const char *record_args[] = { + "record", + "-a", + "-R", + "-M", + "-f", + "-c", "1", + "-e", "kmem:kmalloc", + "-e", "kmem:kmalloc_node", + "-e", "kmem:kfree", + "-e", "kmem:kmem_cache_alloc", + "-e", "kmem:kmem_cache_alloc_node", + "-e", "kmem:kmem_cache_free", +}; + +static int __cmd_record(int argc, const char **argv) +{ + unsigned int rec_argc, i, j; + const char **rec_argv; + + rec_argc = ARRAY_SIZE(record_args) + argc - 1; + rec_argv = calloc(rec_argc + 1, sizeof(char *)); + + for (i = 0; i < ARRAY_SIZE(record_args); i++) + rec_argv[i] = strdup(record_args[i]); + + for (j = 1; j < (unsigned int)argc; j++, i++) + rec_argv[i] = argv[j]; + + return cmd_record(i, rec_argv, NULL); +} + +int cmd_kmem(int argc, const char **argv, const char *prefix __used) +{ + symbol__init(0); + + argc = parse_options(argc, argv, kmem_options, kmem_usage, 0); + + if (argc && !strncmp(argv[0], "rec", 3)) + return __cmd_record(argc, argv); + else if (argc) + usage_with_options(kmem_usage, kmem_options); + + if (list_empty(&caller_sort)) + setup_sorting(&caller_sort, default_sort_order); + if (list_empty(&alloc_sort)) + setup_sorting(&alloc_sort, default_sort_order); + + setup_cpunode_map(); + + return __cmd_kmem(); +} + diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index d78a3d94549..a58e11b7ea8 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -40,6 +40,7 @@ #include "util/parse-options.h" #include "util/parse-events.h" /* For debugfs_path */ #include "util/probe-finder.h" +#include "util/probe-event.h" /* Default vmlinux search paths */ #define NR_SEARCH_PATH 3 @@ -51,8 +52,6 @@ const char *default_search_path[NR_SEARCH_PATH] = { #define MAX_PATH_LEN 256 #define MAX_PROBES 128 -#define MAX_PROBE_ARGS 128 -#define PERFPROBE_GROUP "probe" /* Session management structure */ static struct { @@ -63,152 +62,19 @@ static struct { struct probe_point probes[MAX_PROBES]; } session; -#define semantic_error(msg ...) die("Semantic error :" msg) - -/* Parse probe point. Return 1 if return probe */ -static void parse_probe_point(char *arg, struct probe_point *pp) -{ - char *ptr, *tmp; - char c, nc = 0; - /* - * <Syntax> - * perf probe SRC:LN - * perf probe FUNC[+OFFS|%return][@SRC] - */ - - ptr = strpbrk(arg, ":+@%"); - if (ptr) { - nc = *ptr; - *ptr++ = '\0'; - } - - /* Check arg is function or file and copy it */ - if (strchr(arg, '.')) /* File */ - pp->file = strdup(arg); - else /* Function */ - pp->function = strdup(arg); - DIE_IF(pp->file == NULL && pp->function == NULL); - - /* Parse other options */ - while (ptr) { - arg = ptr; - c = nc; - ptr = strpbrk(arg, ":+@%"); - if (ptr) { - nc = *ptr; - *ptr++ = '\0'; - } - switch (c) { - case ':': /* Line number */ - pp->line = strtoul(arg, &tmp, 0); - if (*tmp != '\0') - semantic_error("There is non-digit charactor" - " in line number."); - break; - case '+': /* Byte offset from a symbol */ - pp->offset = strtoul(arg, &tmp, 0); - if (*tmp != '\0') - semantic_error("There is non-digit charactor" - " in offset."); - break; - case '@': /* File name */ - if (pp->file) - semantic_error("SRC@SRC is not allowed."); - pp->file = strdup(arg); - DIE_IF(pp->file == NULL); - if (ptr) - semantic_error("@SRC must be the last " - "option."); - break; - case '%': /* Probe places */ - if (strcmp(arg, "return") == 0) { - pp->retprobe = 1; - } else /* Others not supported yet */ - semantic_error("%%%s is not supported.", arg); - break; - default: - DIE_IF("Program has a bug."); - break; - } - } - - /* Exclusion check */ - if (pp->line && pp->offset) - semantic_error("Offset can't be used with line number."); - if (!pp->line && pp->file && !pp->function) - semantic_error("File always requires line number."); - if (pp->offset && !pp->function) - semantic_error("Offset requires an entry function."); - if (pp->retprobe && !pp->function) - semantic_error("Return probe requires an entry function."); - if ((pp->offset || pp->line) && pp->retprobe) - semantic_error("Offset/Line can't be used with return probe."); - - pr_debug("symbol:%s file:%s line:%d offset:%d, return:%d\n", - pp->function, pp->file, pp->line, pp->offset, pp->retprobe); -} +static bool listing; /* Parse an event definition. Note that any error must die. */ static void parse_probe_event(const char *str) { - char *argv[MAX_PROBE_ARGS + 2]; /* Event + probe + args */ - int argc, i; struct probe_point *pp = &session.probes[session.nr_probe]; pr_debug("probe-definition(%d): %s\n", session.nr_probe, str); if (++session.nr_probe == MAX_PROBES) - semantic_error("Too many probes"); - - /* Separate arguments, similar to argv_split */ - argc = 0; - do { - /* Skip separators */ - while (isspace(*str)) - str++; - - /* Add an argument */ - if (*str != '\0') { - const char *s = str; - - /* Skip the argument */ - while (!isspace(*str) && *str != '\0') - str++; - - /* Duplicate the argument */ - argv[argc] = strndup(s, str - s); - if (argv[argc] == NULL) - die("strndup"); - if (++argc == MAX_PROBE_ARGS) - semantic_error("Too many arguments"); - pr_debug("argv[%d]=%s\n", argc, argv[argc - 1]); - } - } while (*str != '\0'); - if (!argc) - semantic_error("An empty argument."); - - /* Parse probe point */ - parse_probe_point(argv[0], pp); - free(argv[0]); - if (pp->file || pp->line) - session.need_dwarf = 1; - - /* Copy arguments */ - pp->nr_args = argc - 1; - if (pp->nr_args > 0) { - pp->args = (char **)malloc(sizeof(char *) * pp->nr_args); - if (!pp->args) - die("malloc"); - memcpy(pp->args, &argv[1], sizeof(char *) * pp->nr_args); - } + die("Too many probes (> %d) are specified.", MAX_PROBES); - /* Ensure return probe has no C argument */ - for (i = 0; i < pp->nr_args; i++) - if (is_c_varname(pp->args[i])) { - if (pp->retprobe) - semantic_error("You can't specify local" - " variable for kretprobe"); - session.need_dwarf = 1; - } + /* Parse perf-probe event into probe_point */ + session.need_dwarf = parse_perf_probe_event(str, pp); pr_debug("%d arguments\n", pp->nr_args); } @@ -255,6 +121,7 @@ static int open_default_vmlinux(void) static const char * const probe_usage[] = { "perf probe [<options>] 'PROBEDEF' ['PROBEDEF' ...]", "perf probe [<options>] --add 'PROBEDEF' [--add 'PROBEDEF' ...]", + "perf probe --list", NULL }; @@ -265,6 +132,7 @@ static const struct option options[] = { OPT_STRING('k', "vmlinux", &session.vmlinux, "file", "vmlinux/module pathname"), #endif + OPT_BOOLEAN('l', "list", &listing, "list up current probes"), OPT_CALLBACK('a', "add", NULL, #ifdef NO_LIBDWARF "FUNC[+OFFS|%return] [ARG ...]", @@ -285,73 +153,38 @@ static const struct option options[] = { "\t\tALN:\tAbsolute line number in file.\n" "\t\tARG:\tProbe argument (local variable name or\n" #endif - "\t\t\tkprobe-tracer argument format is supported.)\n", + "\t\t\tkprobe-tracer argument format.)\n", opt_add_probe_event), OPT_END() }; -static int write_new_event(int fd, const char *buf) -{ - int ret; - - ret = write(fd, buf, strlen(buf)); - if (ret <= 0) - die("Failed to create event."); - else - printf("Added new event: %s\n", buf); - - return ret; -} - -#define MAX_CMDLEN 256 - -static int synthesize_probe_event(struct probe_point *pp) -{ - char *buf; - int i, len, ret; - pp->probes[0] = buf = (char *)calloc(MAX_CMDLEN, sizeof(char)); - if (!buf) - die("Failed to allocate memory by calloc."); - ret = snprintf(buf, MAX_CMDLEN, "%s+%d", pp->function, pp->offset); - if (ret <= 0 || ret >= MAX_CMDLEN) - goto error; - len = ret; - - for (i = 0; i < pp->nr_args; i++) { - ret = snprintf(&buf[len], MAX_CMDLEN - len, " %s", - pp->args[i]); - if (ret <= 0 || ret >= MAX_CMDLEN - len) - goto error; - len += ret; - } - pp->found = 1; - return pp->found; -error: - free(pp->probes[0]); - if (ret > 0) - ret = -E2BIG; - return ret; -} - int cmd_probe(int argc, const char **argv, const char *prefix __used) { - int i, j, fd, ret; + int i, j, ret; +#ifndef NO_LIBDWARF + int fd; +#endif struct probe_point *pp; - char buf[MAX_CMDLEN]; argc = parse_options(argc, argv, options, probe_usage, PARSE_OPT_STOP_AT_NON_OPTION); for (i = 0; i < argc; i++) parse_probe_event(argv[i]); - if (session.nr_probe == 0) + if ((session.nr_probe == 0 && !listing) || + (session.nr_probe != 0 && listing)) usage_with_options(probe_usage, options); + if (listing) { + show_perf_probe_events(); + return 0; + } + if (session.need_dwarf) #ifdef NO_LIBDWARF - semantic_error("Debuginfo-analysis is not supported"); + die("Debuginfo-analysis is not supported"); #else /* !NO_LIBDWARF */ - pr_info("Some probes require debuginfo.\n"); + pr_debug("Some probes require debuginfo.\n"); if (session.vmlinux) fd = open(session.vmlinux, O_RDONLY); @@ -395,41 +228,15 @@ end_dwarf: if (pp->found) /* This probe is already found. */ continue; - ret = synthesize_probe_event(pp); + ret = synthesize_trace_kprobe_event(pp); if (ret == -E2BIG) - semantic_error("probe point is too long."); + die("probe point definition becomes too long."); else if (ret < 0) die("Failed to synthesize a probe point."); } /* Settng up probe points */ - snprintf(buf, MAX_CMDLEN, "%s/../kprobe_events", debugfs_path); - fd = open(buf, O_WRONLY, O_APPEND); - if (fd < 0) { - if (errno == ENOENT) - die("kprobe_events file does not exist - please rebuild with CONFIG_KPROBE_TRACER."); - else - die("Could not open kprobe_events file: %s", - strerror(errno)); - } - for (j = 0; j < session.nr_probe; j++) { - pp = &session.probes[j]; - if (pp->found == 1) { - snprintf(buf, MAX_CMDLEN, "%c:%s/%s_%x %s\n", - pp->retprobe ? 'r' : 'p', PERFPROBE_GROUP, - pp->function, pp->offset, pp->probes[0]); - write_new_event(fd, buf); - } else - for (i = 0; i < pp->found; i++) { - snprintf(buf, MAX_CMDLEN, "%c:%s/%s_%x_%d %s\n", - pp->retprobe ? 'r' : 'p', - PERFPROBE_GROUP, - pp->function, pp->offset, i, - pp->probes[0]); - write_new_event(fd, buf); - } - } - close(fd); + add_trace_kprobe_events(session.probes, session.nr_probe); return 0; } diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 82260c56db3..0e519c667e3 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -307,6 +307,12 @@ try_again: printf("\n"); error("perfcounter syscall returned with %d (%s)\n", fd[nr_cpu][counter], strerror(err)); + +#if defined(__i386__) || defined(__x86_64__) + if (attr->type == PERF_TYPE_HARDWARE && err == EOPNOTSUPP) + die("No hardware sampling interrupt available. No APIC? If so then you can boot the kernel with the \"lapic\" boot parameter to force-enable it.\n"); +#endif + die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); exit(-1); } @@ -400,7 +406,7 @@ static int __cmd_record(int argc, const char **argv) struct stat st; pid_t pid = 0; int flags; - int ret; + int err; unsigned long waking = 0; page_size = sysconf(_SC_PAGE_SIZE); @@ -434,16 +440,18 @@ static int __cmd_record(int argc, const char **argv) exit(-1); } - if (!file_new) - header = perf_header__read(output); - else - header = perf_header__new(); - + header = perf_header__new(); if (header == NULL) { pr_err("Not enough memory for reading perf file header\n"); return -1; } + if (!file_new) { + err = perf_header__read(header, output); + if (err < 0) + return err; + } + if (raw_samples) { perf_header__set_feat(header, HEADER_TRACE_INFO); } else { @@ -472,8 +480,11 @@ static int __cmd_record(int argc, const char **argv) } } - if (file_new) - perf_header__write(header, output, false); + if (file_new) { + err = perf_header__write(header, output, false); + if (err < 0) + return err; + } if (!system_wide) event__synthesize_thread(pid, process_synthesized_event); @@ -527,7 +538,7 @@ static int __cmd_record(int argc, const char **argv) if (hits == samples) { if (done) break; - ret = poll(event_array, nr_poll, -1); + err = poll(event_array, nr_poll, -1); waking++; } diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 1a806d5f05c..383c4ab4f9a 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -52,13 +52,12 @@ static int exclude_other = 1; static char callchain_default_opt[] = "fractal,0.5"; -static char *cwd; -static int cwdlen; - static struct perf_header *header; static u64 sample_type; +struct symbol_conf symbol_conf; + static size_t callchain__fprintf_left_margin(FILE *fp, int left_margin) @@ -409,55 +408,6 @@ static int thread__set_comm_adjust(struct thread *self, const char *comm) return 0; } - -static struct symbol * -resolve_symbol(struct thread *thread, struct map **mapp, u64 *ipp) -{ - struct map *map = mapp ? *mapp : NULL; - u64 ip = *ipp; - - if (map) - goto got_map; - - if (!thread) - return NULL; - - map = thread__find_map(thread, ip); - if (map != NULL) { - /* - * We have to do this here as we may have a dso - * with no symbol hit that has a name longer than - * the ones with symbols sampled. - */ - if (!sort_dso.elide && !map->dso->slen_calculated) - dso__calc_col_width(map->dso); - - if (mapp) - *mapp = map; -got_map: - ip = map->map_ip(map, ip); - } else { - /* - * If this is outside of all known maps, - * and is a negative address, try to look it - * up in the kernel dso, as it might be a - * vsyscall or vdso (which executes in user-mode). - * - * XXX This is nasty, we should have a symbol list in - * the "[vdso]" dso, but for now lets use the old - * trick of looking in the whole kernel symbol list. - */ - if ((long long)ip < 0) - return kernel_maps__find_symbol(ip, mapp); - } - dump_printf(" ...... dso: %s\n", - map ? map->dso->long_name : "<not found>"); - dump_printf(" ...... map: %Lx -> %Lx\n", *ipp, ip); - *ipp = ip; - - return map ? map__find_symbol(map, ip, NULL) : NULL; -} - static int call__match(struct symbol *sym) { if (sym->name && !regexec(&parent_regex, sym->name, 0, NULL, 0)) @@ -466,11 +416,11 @@ static int call__match(struct symbol *sym) return 0; } -static struct symbol **resolve_callchain(struct thread *thread, struct map *map, +static struct symbol **resolve_callchain(struct thread *thread, struct ip_callchain *chain, struct symbol **parent) { - u64 context = PERF_CONTEXT_MAX; + u8 cpumode = PERF_RECORD_MISC_USER; struct symbol **syms = NULL; unsigned int i; @@ -484,30 +434,31 @@ static struct symbol **resolve_callchain(struct thread *thread, struct map *map, for (i = 0; i < chain->nr; i++) { u64 ip = chain->ips[i]; - struct symbol *sym = NULL; + struct addr_location al; if (ip >= PERF_CONTEXT_MAX) { - context = ip; + switch (ip) { + case PERF_CONTEXT_HV: + cpumode = PERF_RECORD_MISC_HYPERVISOR; break; + case PERF_CONTEXT_KERNEL: + cpumode = PERF_RECORD_MISC_KERNEL; break; + case PERF_CONTEXT_USER: + cpumode = PERF_RECORD_MISC_USER; break; + default: + break; + } continue; } - switch (context) { - case PERF_CONTEXT_HV: - break; - case PERF_CONTEXT_KERNEL: - sym = kernel_maps__find_symbol(ip, &map); - break; - default: - sym = resolve_symbol(thread, &map, &ip); - break; - } - - if (sym) { - if (sort__has_parent && !*parent && call__match(sym)) - *parent = sym; + thread__find_addr_location(thread, cpumode, MAP__FUNCTION, + ip, &al, NULL); + if (al.sym != NULL) { + if (sort__has_parent && !*parent && + call__match(al.sym)) + *parent = al.sym; if (!callchain) break; - syms[i] = sym; + syms[i] = al.sym; } } @@ -518,20 +469,17 @@ static struct symbol **resolve_callchain(struct thread *thread, struct map *map, * collect histogram counts */ -static int -hist_entry__add(struct thread *thread, struct map *map, - struct symbol *sym, u64 ip, struct ip_callchain *chain, - char level, u64 count) +static int hist_entry__add(struct addr_location *al, + struct ip_callchain *chain, u64 count) { struct symbol **syms = NULL, *parent = NULL; bool hit; struct hist_entry *he; if ((sort__has_parent || callchain) && chain) - syms = resolve_callchain(thread, map, chain, &parent); + syms = resolve_callchain(al->thread, chain, &parent); - he = __hist_entry__add(thread, map, sym, parent, - ip, count, level, &hit); + he = __hist_entry__add(al, parent, count, &hit); if (he == NULL) return -ENOMEM; @@ -655,17 +603,14 @@ static int validate_chain(struct ip_callchain *chain, event_t *event) return 0; } -static int -process_sample_event(event_t *event, unsigned long offset, unsigned long head) +static int process_sample_event(event_t *event) { - char level; - struct symbol *sym = NULL; u64 ip = event->ip.ip; u64 period = 1; - struct map *map = NULL; void *more_data = event->ip.__more_data; struct ip_callchain *chain = NULL; int cpumode; + struct addr_location al; struct thread *thread = threads__findnew(event->ip.pid); if (sample_type & PERF_SAMPLE_PERIOD) { @@ -673,9 +618,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) more_data += sizeof(u64); } - dump_printf("%p [%p]: PERF_RECORD_SAMPLE (IP, %d): %d/%d: %p period: %Ld\n", - (void *)(offset + head), - (void *)(long)(event->header.size), + dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, event->ip.pid, event->ip.tid, (void *)(long)ip, @@ -713,136 +656,51 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; - if (cpumode == PERF_RECORD_MISC_KERNEL) { - level = 'k'; - sym = kernel_maps__find_symbol(ip, &map); - dump_printf(" ...... dso: %s\n", - map ? map->dso->long_name : "<not found>"); - } else if (cpumode == PERF_RECORD_MISC_USER) { - level = '.'; - sym = resolve_symbol(thread, &map, &ip); - - } else { - level = 'H'; - dump_printf(" ...... dso: [hypervisor]\n"); - } + thread__find_addr_location(thread, cpumode, + MAP__FUNCTION, ip, &al, NULL); + /* + * We have to do this here as we may have a dso with no symbol hit that + * has a name longer than the ones with symbols sampled. + */ + if (al.map && !sort_dso.elide && !al.map->dso->slen_calculated) + dso__calc_col_width(al.map->dso); if (dso_list && - (!map || !map->dso || - !(strlist__has_entry(dso_list, map->dso->short_name) || - (map->dso->short_name != map->dso->long_name && - strlist__has_entry(dso_list, map->dso->long_name))))) + (!al.map || !al.map->dso || + !(strlist__has_entry(dso_list, al.map->dso->short_name) || + (al.map->dso->short_name != al.map->dso->long_name && + strlist__has_entry(dso_list, al.map->dso->long_name))))) return 0; - if (sym_list && sym && !strlist__has_entry(sym_list, sym->name)) + if (sym_list && al.sym && !strlist__has_entry(sym_list, al.sym->name)) return 0; - if (hist_entry__add(thread, map, sym, ip, - chain, level, period)) { + if (hist_entry__add(&al, chain, period)) { pr_debug("problem incrementing symbol count, skipping event\n"); return -1; } - total += period; + event__stats.total += period; return 0; } -static int -process_mmap_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct map *map = map__new(&event->mmap, cwd, cwdlen); - struct thread *thread = threads__findnew(event->mmap.pid); - - dump_printf("%p [%p]: PERF_RECORD_MMAP %d/%d: [%p(%p) @ %p]: %s\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->mmap.pid, - event->mmap.tid, - (void *)(long)event->mmap.start, - (void *)(long)event->mmap.len, - (void *)(long)event->mmap.pgoff, - event->mmap.filename); - - if (thread == NULL || map == NULL) { - dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); - return 0; - } - - thread__insert_map(thread, map); - total_mmap++; - - return 0; -} - -static int -process_comm_event(event_t *event, unsigned long offset, unsigned long head) +static int process_comm_event(event_t *event) { struct thread *thread = threads__findnew(event->comm.pid); - dump_printf("%p [%p]: PERF_RECORD_COMM: %s:%d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->comm.comm, event->comm.pid); + dump_printf(": %s:%d\n", event->comm.comm, event->comm.pid); if (thread == NULL || thread__set_comm_adjust(thread, event->comm.comm)) { dump_printf("problem processing PERF_RECORD_COMM, skipping event.\n"); return -1; } - total_comm++; - - return 0; -} - -static int -process_task_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct thread *thread = threads__findnew(event->fork.pid); - struct thread *parent = threads__findnew(event->fork.ppid); - - dump_printf("%p [%p]: PERF_RECORD_%s: (%d:%d):(%d:%d)\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->header.type == PERF_RECORD_FORK ? "FORK" : "EXIT", - event->fork.pid, event->fork.tid, - event->fork.ppid, event->fork.ptid); - - /* - * A thread clone will have the same PID for both - * parent and child. - */ - if (thread == parent) - return 0; - - if (event->header.type == PERF_RECORD_EXIT) - return 0; - - if (!thread || !parent || thread__fork(thread, parent)) { - dump_printf("problem processing PERF_RECORD_FORK, skipping event.\n"); - return -1; - } - total_fork++; - - return 0; -} - -static int -process_lost_event(event_t *event, unsigned long offset, unsigned long head) -{ - dump_printf("%p [%p]: PERF_RECORD_LOST: id:%Ld: lost:%Ld\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->lost.id, - event->lost.lost); - - total_lost += event->lost.lost; return 0; } -static int -process_read_event(event_t *event, unsigned long offset, unsigned long head) +static int process_read_event(event_t *event) { struct perf_event_attr *attr; @@ -858,14 +716,9 @@ process_read_event(event_t *event, unsigned long offset, unsigned long head) event->read.value); } - dump_printf("%p [%p]: PERF_RECORD_READ: %d %d %s %Lu\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->read.pid, - event->read.tid, - attr ? __event_name(attr->type, attr->config) - : "FAIL", - event->read.value); + dump_printf(": %d %d %s %Lu\n", event->read.pid, event->read.tid, + attr ? __event_name(attr->type, attr->config) : "FAIL", + event->read.value); return 0; } @@ -901,11 +754,11 @@ static int sample_type_check(u64 type) static struct perf_file_handler file_handler = { .process_sample_event = process_sample_event, - .process_mmap_event = process_mmap_event, + .process_mmap_event = event__process_mmap, .process_comm_event = process_comm_event, - .process_exit_event = process_task_event, - .process_fork_event = process_task_event, - .process_lost_event = process_lost_event, + .process_exit_event = event__process_task, + .process_fork_event = event__process_task, + .process_lost_event = event__process_lost, .process_read_event = process_read_event, .sample_type_check = sample_type_check, }; @@ -924,20 +777,15 @@ static int __cmd_report(void) register_perf_file_handler(&file_handler); - ret = mmap_dispatch_perf_file(&header, input_name, force, full_paths, - &cwdlen, &cwd); + ret = mmap_dispatch_perf_file(&header, input_name, force, + full_paths, &event__cwdlen, &event__cwd); if (ret) return ret; - dump_printf(" IP events: %10ld\n", total); - dump_printf(" mmap events: %10ld\n", total_mmap); - dump_printf(" comm events: %10ld\n", total_comm); - dump_printf(" fork events: %10ld\n", total_fork); - dump_printf(" lost events: %10ld\n", total_lost); - dump_printf(" unknown events: %10ld\n", file_handler.total_unknown); - - if (dump_trace) + if (dump_trace) { + event__print_totals(); return 0; + } if (verbose > 3) threads__fprintf(stdout); @@ -946,8 +794,8 @@ static int __cmd_report(void) dsos__fprintf(stdout); collapse__resort(); - output__resort(total); - output__fprintf(stdout, total); + output__resort(event__stats.total); + output__fprintf(stdout, event__stats.total); if (show_threads) perf_read_values_destroy(&show_threads_values); @@ -1021,9 +869,10 @@ static const struct option options[] = { "be more verbose (show symbol address, etc)"), OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, "dump raw trace in ASCII"), - OPT_STRING('k', "vmlinux", &vmlinux_name, "file", "vmlinux pathname"), + OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, + "file", "vmlinux pathname"), OPT_BOOLEAN('f', "force", &force, "don't complain, do it"), - OPT_BOOLEAN('m', "modules", &modules, + OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules, "load module symbols - WARNING: use only with -k and LIVE kernel"), OPT_BOOLEAN('n', "show-nr-samples", &show_nr_samples, "Show a column with the number of samples"), @@ -1093,7 +942,8 @@ static void setup_list(struct strlist **list, const char *list_str, int cmd_report(int argc, const char **argv, const char *prefix __used) { - symbol__init(0); + if (symbol__init(&symbol_conf) < 0) + return -1; argc = parse_options(argc, argv, options, report_usage, 0); diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index df44b756cec..26b782f26ee 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -22,8 +22,6 @@ static char const *input_name = "perf.data"; -static unsigned long total_comm = 0; - static struct perf_header *header; static u64 sample_type; @@ -32,9 +30,6 @@ static char *sort_order = default_sort_order; static int profile_cpu = -1; -static char *cwd; -static int cwdlen; - #define PR_SET_NAME 15 /* Set process name */ #define MAX_CPUS 4096 @@ -225,7 +220,7 @@ static void calibrate_sleep_measurement_overhead(void) static struct sched_atom * get_new_event(struct task_desc *task, u64 timestamp) { - struct sched_atom *event = calloc(1, sizeof(*event)); + struct sched_atom *event = zalloc(sizeof(*event)); unsigned long idx = task->nr_events; size_t size; @@ -293,7 +288,7 @@ add_sched_event_wakeup(struct task_desc *task, u64 timestamp, return; } - wakee_event->wait_sem = calloc(1, sizeof(*wakee_event->wait_sem)); + wakee_event->wait_sem = zalloc(sizeof(*wakee_event->wait_sem)); sem_init(wakee_event->wait_sem, 0, 0); wakee_event->specific_wait = 1; event->wait_sem = wakee_event->wait_sem; @@ -323,7 +318,7 @@ static struct task_desc *register_pid(unsigned long pid, const char *comm) if (task) return task; - task = calloc(1, sizeof(*task)); + task = zalloc(sizeof(*task)); task->pid = pid; task->nr = nr_tasks; strcpy(task->comm, comm); @@ -633,27 +628,6 @@ static void test_calibrations(void) printf("the sleep test took %Ld nsecs\n", T1-T0); } -static int -process_comm_event(event_t *event, unsigned long offset, unsigned long head) -{ - struct thread *thread = threads__findnew(event->comm.tid); - - dump_printf("%p [%p]: perf_event_comm: %s:%d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->comm.comm, event->comm.pid); - - if (thread == NULL || - thread__set_comm(thread, event->comm.comm)) { - dump_printf("problem processing perf_event_comm, skipping event.\n"); - return -1; - } - total_comm++; - - return 0; -} - - struct raw_event_sample { u32 size; char data[0]; @@ -962,9 +936,7 @@ __thread_latency_insert(struct rb_root *root, struct work_atoms *data, static void thread_atoms_insert(struct thread *thread) { - struct work_atoms *atoms; - - atoms = calloc(sizeof(*atoms), 1); + struct work_atoms *atoms = zalloc(sizeof(*atoms)); if (!atoms) die("No memory"); @@ -996,9 +968,7 @@ add_sched_out_event(struct work_atoms *atoms, char run_state, u64 timestamp) { - struct work_atom *atom; - - atom = calloc(sizeof(*atom), 1); + struct work_atom *atom = zalloc(sizeof(*atom)); if (!atom) die("Non memory"); @@ -1626,8 +1596,7 @@ process_raw_event(event_t *raw_event __used, void *more_data, process_sched_migrate_task_event(raw, event, cpu, timestamp, thread); } -static int -process_sample_event(event_t *event, unsigned long offset, unsigned long head) +static int process_sample_event(event_t *event) { struct thread *thread; u64 ip = event->ip.ip; @@ -1657,9 +1626,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) more_data += sizeof(u64); } - dump_printf("%p [%p]: PERF_RECORD_SAMPLE (IP, %d): %d/%d: %p period: %Ld\n", - (void *)(offset + head), - (void *)(long)(event->header.size), + dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, event->ip.pid, event->ip.tid, (void *)(long)ip, @@ -1681,10 +1648,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) return 0; } -static int -process_lost_event(event_t *event __used, - unsigned long offset __used, - unsigned long head __used) +static int process_lost_event(event_t *event __used) { nr_lost_chunks++; nr_lost_events += event->lost.lost; @@ -1708,7 +1672,7 @@ static int sample_type_check(u64 type) static struct perf_file_handler file_handler = { .process_sample_event = process_sample_event, - .process_comm_event = process_comm_event, + .process_comm_event = event__process_comm, .process_lost_event = process_lost_event, .sample_type_check = sample_type_check, }; @@ -1718,7 +1682,8 @@ static int read_events(void) register_idle_thread(); register_perf_file_handler(&file_handler); - return mmap_dispatch_perf_file(&header, input_name, 0, 0, &cwdlen, &cwd); + return mmap_dispatch_perf_file(&header, input_name, 0, 0, + &event__cwdlen, &event__cwd); } static void print_bad_events(void) diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 665877e4a94..cb58b6605fc 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -29,14 +29,14 @@ #include "util/header.h" #include "util/parse-options.h" #include "util/parse-events.h" +#include "util/event.h" +#include "util/data_map.h" #include "util/svghelper.h" static char const *input_name = "perf.data"; static char const *output_name = "output.svg"; -static unsigned long page_size; -static unsigned long mmap_window = 32; static u64 sample_type; static unsigned int numcpus; @@ -49,8 +49,6 @@ static u64 first_time, last_time; static int power_only; -static struct perf_header *header; - struct per_pid; struct per_pidcomm; @@ -156,9 +154,9 @@ struct sample_wrapper *all_samples; struct process_filter; struct process_filter { - char *name; - int pid; - struct process_filter *next; + char *name; + int pid; + struct process_filter *next; }; static struct process_filter *process_filter; @@ -1045,36 +1043,6 @@ static void write_svg_file(const char *filename) svg_close(); } -static int -process_event(event_t *event) -{ - - switch (event->header.type) { - - case PERF_RECORD_COMM: - return process_comm_event(event); - case PERF_RECORD_FORK: - return process_fork_event(event); - case PERF_RECORD_EXIT: - return process_exit_event(event); - case PERF_RECORD_SAMPLE: - return queue_sample_event(event); - - /* - * We dont process them right now but they are fine: - */ - case PERF_RECORD_MMAP: - case PERF_RECORD_THROTTLE: - case PERF_RECORD_UNTHROTTLE: - return 0; - - default: - return -1; - } - - return 0; -} - static void process_samples(void) { struct sample_wrapper *cursor; @@ -1090,105 +1058,38 @@ static void process_samples(void) } } - -static int __cmd_timechart(void) +static int sample_type_check(u64 type) { - int ret, rc = EXIT_FAILURE; - unsigned long offset = 0; - unsigned long head, shift; - struct stat statbuf; - event_t *event; - uint32_t size; - char *buf; - int input; - - input = open(input_name, O_RDONLY); - if (input < 0) { - fprintf(stderr, " failed to open file: %s", input_name); - if (!strcmp(input_name, "perf.data")) - fprintf(stderr, " (try 'perf record' first)"); - fprintf(stderr, "\n"); - exit(-1); - } + sample_type = type; - ret = fstat(input, &statbuf); - if (ret < 0) { - perror("failed to stat file"); - exit(-1); - } - - if (!statbuf.st_size) { - fprintf(stderr, "zero-sized file, nothing to do!\n"); - exit(0); - } - - header = perf_header__read(input); - head = header->data_offset; - - sample_type = perf_header__sample_type(header); - - shift = page_size * (head / page_size); - offset += shift; - head -= shift; - -remap: - buf = (char *)mmap(NULL, page_size * mmap_window, PROT_READ, - MAP_SHARED, input, offset); - if (buf == MAP_FAILED) { - perror("failed to mmap file"); - exit(-1); - } - -more: - event = (event_t *)(buf + head); - - size = event->header.size; - if (!size) - size = 8; - - if (head + event->header.size >= page_size * mmap_window) { - int ret2; - - shift = page_size * (head / page_size); - - ret2 = munmap(buf, page_size * mmap_window); - assert(ret2 == 0); - - offset += shift; - head -= shift; - goto remap; - } - - size = event->header.size; - - if (!size || process_event(event) < 0) { - pr_warning("%p [%p]: skipping unknown header type: %d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->header.type); - /* - * assume we lost track of the stream, check alignment, and - * increment a single u64 in the hope to catch on again 'soon'. - */ - - if (unlikely(head & 7)) - head &= ~7ULL; - - size = 8; + if (!(sample_type & PERF_SAMPLE_RAW)) { + fprintf(stderr, "No trace samples found in the file.\n" + "Have you used 'perf timechart record' to record it?\n"); + return -1; } - head += size; + return 0; +} - if (offset + head >= header->data_offset + header->data_size) - goto done; +static struct perf_file_handler file_handler = { + .process_comm_event = process_comm_event, + .process_fork_event = process_fork_event, + .process_exit_event = process_exit_event, + .process_sample_event = queue_sample_event, + .sample_type_check = sample_type_check, +}; - if (offset + head < (unsigned long)statbuf.st_size) - goto more; +static int __cmd_timechart(void) +{ + struct perf_header *header; + int ret; -done: - rc = EXIT_SUCCESS; - close(input); + register_perf_file_handler(&file_handler); + ret = mmap_dispatch_perf_file(&header, input_name, 0, 0, + &event__cwdlen, &event__cwd); + if (ret) + return EXIT_FAILURE; process_samples(); @@ -1201,7 +1102,7 @@ done: pr_info("Written %2.1f seconds of trace to %s.\n", (last_time - first_time) / 1000000000.0, output_name); - return rc; + return EXIT_SUCCESS; } static const char * const timechart_usage[] = { @@ -1268,8 +1169,6 @@ int cmd_timechart(int argc, const char **argv, const char *prefix __used) { symbol__init(0); - page_size = getpagesize(); - argc = parse_options(argc, argv, options, timechart_usage, PARSE_OPT_STOP_AT_NON_OPTION); diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 89b7f68a179..e0a374d0e43 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -78,6 +78,8 @@ static int dump_symtab = 0; static bool hide_kernel_symbols = false; static bool hide_user_symbols = false; +static struct winsize winsize; +struct symbol_conf symbol_conf; /* * Source @@ -100,58 +102,75 @@ static int display_weighted = -1; * Symbols */ +struct sym_entry_source { + struct source_line *source; + struct source_line *lines; + struct source_line **lines_tail; + pthread_mutex_t lock; +}; + struct sym_entry { struct rb_node rb_node; struct list_head node; - unsigned long count[MAX_COUNTERS]; unsigned long snap_count; double weight; int skip; + u16 name_len; u8 origin; struct map *map; - struct source_line *source; - struct source_line *lines; - struct source_line **lines_tail; - pthread_mutex_t source_lock; + struct sym_entry_source *src; + unsigned long count[0]; }; /* * Source functions */ -/* most GUI terminals set LINES (although some don't export it) */ -static int term_rows(void) +static inline struct symbol *sym_entry__symbol(struct sym_entry *self) { - char *lines_string = getenv("LINES"); - int n_lines; + return ((void *)self) + symbol_conf.priv_size; +} - if (lines_string && (n_lines = atoi(lines_string)) > 0) - return n_lines; -#ifdef TIOCGWINSZ - else { - struct winsize ws; - if (!ioctl(1, TIOCGWINSZ, &ws) && ws.ws_row) - return ws.ws_row; +static void get_term_dimensions(struct winsize *ws) +{ + char *s = getenv("LINES"); + + if (s != NULL) { + ws->ws_row = atoi(s); + s = getenv("COLUMNS"); + if (s != NULL) { + ws->ws_col = atoi(s); + if (ws->ws_row && ws->ws_col) + return; + } } +#ifdef TIOCGWINSZ + if (ioctl(1, TIOCGWINSZ, ws) == 0 && + ws->ws_row && ws->ws_col) + return; #endif - return 25; + ws->ws_row = 25; + ws->ws_col = 80; } -static void update_print_entries(void) +static void update_print_entries(struct winsize *ws) { - print_entries = term_rows(); + print_entries = ws->ws_row; + if (print_entries > 9) print_entries -= 9; } static void sig_winch_handler(int sig __used) { - update_print_entries(); + get_term_dimensions(&winsize); + update_print_entries(&winsize); } static void parse_source(struct sym_entry *syme) { struct symbol *sym; + struct sym_entry_source *source; struct map *map; FILE *file; char command[PATH_MAX*2]; @@ -161,12 +180,21 @@ static void parse_source(struct sym_entry *syme) if (!syme) return; - if (syme->lines) { - pthread_mutex_lock(&syme->source_lock); + if (syme->src == NULL) { + syme->src = zalloc(sizeof(*source)); + if (syme->src == NULL) + return; + pthread_mutex_init(&syme->src->lock, NULL); + } + + source = syme->src; + + if (source->lines) { + pthread_mutex_lock(&source->lock); goto out_assign; } - sym = (struct symbol *)(syme + 1); + sym = sym_entry__symbol(syme); map = syme->map; path = map->dso->long_name; @@ -182,8 +210,8 @@ static void parse_source(struct sym_entry *syme) if (!file) return; - pthread_mutex_lock(&syme->source_lock); - syme->lines_tail = &syme->lines; + pthread_mutex_lock(&source->lock); + source->lines_tail = &source->lines; while (!feof(file)) { struct source_line *src; size_t dummy = 0; @@ -203,8 +231,8 @@ static void parse_source(struct sym_entry *syme) *c = 0; src->next = NULL; - *syme->lines_tail = src; - syme->lines_tail = &src->next; + *source->lines_tail = src; + source->lines_tail = &src->next; if (strlen(src->line)>8 && src->line[8] == ':') { src->eip = strtoull(src->line, NULL, 16); @@ -218,7 +246,7 @@ static void parse_source(struct sym_entry *syme) pclose(file); out_assign: sym_filter_entry = syme; - pthread_mutex_unlock(&syme->source_lock); + pthread_mutex_unlock(&source->lock); } static void __zero_source_counters(struct sym_entry *syme) @@ -226,7 +254,7 @@ static void __zero_source_counters(struct sym_entry *syme) int i; struct source_line *line; - line = syme->lines; + line = syme->src->lines; while (line) { for (i = 0; i < nr_counters; i++) line->count[i] = 0; @@ -241,13 +269,13 @@ static void record_precise_ip(struct sym_entry *syme, int counter, u64 ip) if (syme != sym_filter_entry) return; - if (pthread_mutex_trylock(&syme->source_lock)) + if (pthread_mutex_trylock(&syme->src->lock)) return; - if (!syme->source) + if (syme->src == NULL || syme->src->source == NULL) goto out_unlock; - for (line = syme->lines; line; line = line->next) { + for (line = syme->src->lines; line; line = line->next) { if (line->eip == ip) { line->count[counter]++; break; @@ -256,25 +284,25 @@ static void record_precise_ip(struct sym_entry *syme, int counter, u64 ip) break; } out_unlock: - pthread_mutex_unlock(&syme->source_lock); + pthread_mutex_unlock(&syme->src->lock); } static void lookup_sym_source(struct sym_entry *syme) { - struct symbol *symbol = (struct symbol *)(syme + 1); + struct symbol *symbol = sym_entry__symbol(syme); struct source_line *line; char pattern[PATH_MAX]; sprintf(pattern, "<%s>:", symbol->name); - pthread_mutex_lock(&syme->source_lock); - for (line = syme->lines; line; line = line->next) { + pthread_mutex_lock(&syme->src->lock); + for (line = syme->src->lines; line; line = line->next) { if (strstr(line->line, pattern)) { - syme->source = line; + syme->src->source = line; break; } } - pthread_mutex_unlock(&syme->source_lock); + pthread_mutex_unlock(&syme->src->lock); } static void show_lines(struct source_line *queue, int count, int total) @@ -304,24 +332,24 @@ static void show_details(struct sym_entry *syme) if (!syme) return; - if (!syme->source) + if (!syme->src->source) lookup_sym_source(syme); - if (!syme->source) + if (!syme->src->source) return; - symbol = (struct symbol *)(syme + 1); + symbol = sym_entry__symbol(syme); printf("Showing %s for %s\n", event_name(sym_counter), symbol->name); printf(" Events Pcnt (>=%d%%)\n", sym_pcnt_filter); - pthread_mutex_lock(&syme->source_lock); - line = syme->source; + pthread_mutex_lock(&syme->src->lock); + line = syme->src->source; while (line) { total += line->count[sym_counter]; line = line->next; } - line = syme->source; + line = syme->src->source; while (line) { float pcnt = 0.0; @@ -346,7 +374,7 @@ static void show_details(struct sym_entry *syme) line->count[sym_counter] = zero ? 0 : line->count[sym_counter] * 7 / 8; line = line->next; } - pthread_mutex_unlock(&syme->source_lock); + pthread_mutex_unlock(&syme->src->lock); if (more) printf("%d lines not displayed, maybe increase display entries [e]\n", more); } @@ -423,6 +451,8 @@ static void print_sym_table(void) struct sym_entry *syme, *n; struct rb_root tmp = RB_ROOT; struct rb_node *nd; + int sym_width = 0, dso_width = 0, max_dso_width; + const int win_width = winsize.ws_col - 1; samples = userspace_samples = 0; @@ -434,6 +464,7 @@ static void print_sym_table(void) list_for_each_entry_safe_from(syme, n, &active_symbols, node) { syme->snap_count = syme->count[snap]; if (syme->snap_count != 0) { + if ((hide_user_symbols && syme->origin == PERF_RECORD_MISC_USER) || (hide_kernel_symbols && @@ -453,8 +484,7 @@ static void print_sym_table(void) puts(CONSOLE_CLEAR); - printf( -"------------------------------------------------------------------------------\n"); + printf("%-*.*s\n", win_width, win_width, graph_dotted_line); printf( " PerfTop:%8.0f irqs/sec kernel:%4.1f%% [", samples_per_sec, 100.0 - (100.0*((samples_per_sec-ksamples_per_sec)/samples_per_sec))); @@ -492,13 +522,35 @@ static void print_sym_table(void) printf(", %d CPUs)\n", nr_cpus); } - printf("------------------------------------------------------------------------------\n\n"); + printf("%-*.*s\n", win_width, win_width, graph_dotted_line); if (sym_filter_entry) { show_details(sym_filter_entry); return; } + /* + * Find the longest symbol name that will be displayed + */ + for (nd = rb_first(&tmp); nd; nd = rb_next(nd)) { + syme = rb_entry(nd, struct sym_entry, rb_node); + if (++printed > print_entries || + (int)syme->snap_count < count_filter) + continue; + + if (syme->map->dso->long_name_len > dso_width) + dso_width = syme->map->dso->long_name_len; + + if (syme->name_len > sym_width) + sym_width = syme->name_len; + } + + printed = 0; + + max_dso_width = winsize.ws_col - sym_width - 29; + if (dso_width > max_dso_width) + dso_width = max_dso_width; + putchar('\n'); if (nr_counters == 1) printf(" samples pcnt"); else @@ -506,19 +558,21 @@ static void print_sym_table(void) if (verbose) printf(" RIP "); - printf(" function DSO\n"); + printf(" %-*.*s DSO\n", sym_width, sym_width, "function"); printf(" %s _______ _____", nr_counters == 1 ? " " : "______"); if (verbose) printf(" ________________"); - printf(" ________________________________ ________________\n\n"); + printf(" %-*.*s", sym_width, sym_width, graph_line); + printf(" %-*.*s", dso_width, dso_width, graph_line); + puts("\n"); for (nd = rb_first(&tmp); nd; nd = rb_next(nd)) { struct symbol *sym; double pcnt; syme = rb_entry(nd, struct sym_entry, rb_node); - sym = (struct symbol *)(syme + 1); + sym = sym_entry__symbol(syme); if (++printed > print_entries || (int)syme->snap_count < count_filter) continue; @@ -534,9 +588,11 @@ static void print_sym_table(void) percent_color_fprintf(stdout, "%4.1f%%", pcnt); if (verbose) printf(" %016llx", sym->start); - printf(" %-32s", sym->name); - printf(" %s", syme->map->dso->short_name); - printf("\n"); + printf(" %-*.*s", sym_width, sym_width, sym->name); + printf(" %-*.*s\n", dso_width, dso_width, + dso_width >= syme->map->dso->long_name_len ? + syme->map->dso->long_name : + syme->map->dso->short_name); } } @@ -583,10 +639,10 @@ static void prompt_symbol(struct sym_entry **target, const char *msg) /* zero counters of active symbol */ if (syme) { - pthread_mutex_lock(&syme->source_lock); + pthread_mutex_lock(&syme->src->lock); __zero_source_counters(syme); *target = NULL; - pthread_mutex_unlock(&syme->source_lock); + pthread_mutex_unlock(&syme->src->lock); } fprintf(stdout, "\n%s: ", msg); @@ -602,7 +658,7 @@ static void prompt_symbol(struct sym_entry **target, const char *msg) pthread_mutex_unlock(&active_symbols_lock); list_for_each_entry_safe_from(syme, n, &active_symbols, node) { - struct symbol *sym = (struct symbol *)(syme + 1); + struct symbol *sym = sym_entry__symbol(syme); if (!strcmp(buf, sym->name)) { found = syme; @@ -626,7 +682,7 @@ static void print_mapped_keys(void) char *name = NULL; if (sym_filter_entry) { - struct symbol *sym = (struct symbol *)(sym_filter_entry+1); + struct symbol *sym = sym_entry__symbol(sym_filter_entry); name = sym->name; } @@ -639,7 +695,7 @@ static void print_mapped_keys(void) fprintf(stdout, "\t[f] profile display filter (count). \t(%d)\n", count_filter); - if (vmlinux_name) { + if (symbol_conf.vmlinux_name) { fprintf(stdout, "\t[F] annotate display filter (percent). \t(%d%%)\n", sym_pcnt_filter); fprintf(stdout, "\t[s] annotate symbol. \t(%s)\n", name?: "NULL"); fprintf(stdout, "\t[S] stop annotation.\n"); @@ -676,7 +732,7 @@ static int key_mapped(int c) case 'F': case 's': case 'S': - return vmlinux_name ? 1 : 0; + return symbol_conf.vmlinux_name ? 1 : 0; default: break; } @@ -718,7 +774,7 @@ static void handle_keypress(int c) case 'e': prompt_integer(&print_entries, "Enter display entries (lines)"); if (print_entries == 0) { - update_print_entries(); + sig_winch_handler(SIGWINCH); signal(SIGWINCH, sig_winch_handler); } else signal(SIGWINCH, SIG_DFL); @@ -752,6 +808,8 @@ static void handle_keypress(int c) case 'q': case 'Q': printf("exiting.\n"); + if (dump_symtab) + dsos__fprintf(stderr); exit(0); case 's': prompt_symbol(&sym_filter_entry, "Enter details symbol"); @@ -762,10 +820,10 @@ static void handle_keypress(int c) else { struct sym_entry *syme = sym_filter_entry; - pthread_mutex_lock(&syme->source_lock); + pthread_mutex_lock(&syme->src->lock); sym_filter_entry = NULL; __zero_source_counters(syme); - pthread_mutex_unlock(&syme->source_lock); + pthread_mutex_unlock(&syme->src->lock); } break; case 'U': @@ -851,7 +909,7 @@ static int symbol_filter(struct map *map, struct symbol *sym) syme = symbol__priv(sym); syme->map = map; - pthread_mutex_init(&syme->source_lock, NULL); + syme->src = NULL; if (!sym_filter_entry && sym_filter && !strcmp(name, sym_filter)) sym_filter_entry = syme; @@ -862,16 +920,8 @@ static int symbol_filter(struct map *map, struct symbol *sym) } } - return 0; -} - -static int parse_symbols(void) -{ - if (dsos__load_kernel(vmlinux_name, symbol_filter, 1) <= 0) - return -1; - - if (dump_symtab) - dsos__fprintf(stderr); + if (!syme->skip) + syme->name_len = strlen(sym->name); return 0; } @@ -879,55 +929,28 @@ static int parse_symbols(void) static void event__process_sample(const event_t *self, int counter) { u64 ip = self->ip.ip; - struct map *map; struct sym_entry *syme; - struct symbol *sym; + struct addr_location al; u8 origin = self->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; switch (origin) { - case PERF_RECORD_MISC_USER: { - struct thread *thread; - + case PERF_RECORD_MISC_USER: if (hide_user_symbols) return; - - thread = threads__findnew(self->ip.pid); - if (thread == NULL) - return; - - map = thread__find_map(thread, ip); - if (map != NULL) { - ip = map->map_ip(map, ip); - sym = map__find_symbol(map, ip, symbol_filter); - if (sym == NULL) - return; - userspace_samples++; - break; - } - } - /* - * If this is outside of all known maps, - * and is a negative address, try to look it - * up in the kernel dso, as it might be a - * vsyscall or vdso (which executes in user-mode). - */ - if ((long long)ip >= 0) - return; - /* Fall thru */ + break; case PERF_RECORD_MISC_KERNEL: if (hide_kernel_symbols) return; - - sym = kernel_maps__find_symbol(ip, &map); - if (sym == NULL) - return; break; default: return; } - syme = symbol__priv(sym); + if (event__preprocess_sample(self, &al, symbol_filter) < 0 || + al.sym == NULL) + return; + syme = symbol__priv(al.sym); if (!syme->skip) { syme->count[counter]++; syme->origin = origin; @@ -936,30 +959,12 @@ static void event__process_sample(const event_t *self, int counter) if (list_empty(&syme->node) || !syme->node.next) __list_insert_active_sym(syme); pthread_mutex_unlock(&active_symbols_lock); + if (origin == PERF_RECORD_MISC_USER) + ++userspace_samples; ++samples; - return; - } -} - -static void event__process_mmap(event_t *self) -{ - struct thread *thread = threads__findnew(self->mmap.pid); - - if (thread != NULL) { - struct map *map = map__new(&self->mmap, NULL, 0); - if (map != NULL) - thread__insert_map(thread, map); } } -static void event__process_comm(event_t *self) -{ - struct thread *thread = threads__findnew(self->comm.pid); - - if (thread != NULL) - thread__set_comm(thread, self->comm.comm); -} - static int event__process(event_t *event) { switch (event->header.type) { @@ -1211,7 +1216,8 @@ static const struct option options[] = { "system-wide collection from all CPUs"), OPT_INTEGER('C', "CPU", &profile_cpu, "CPU to profile on"), - OPT_STRING('k', "vmlinux", &vmlinux_name, "file", "vmlinux pathname"), + OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, + "file", "vmlinux pathname"), OPT_BOOLEAN('K', "hide_kernel_symbols", &hide_kernel_symbols, "hide kernel symbols"), OPT_INTEGER('m', "mmap-pages", &mmap_pages, @@ -1247,8 +1253,6 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) { int counter; - symbol__init(sizeof(struct sym_entry)); - page_size = sysconf(_SC_PAGE_SIZE); argc = parse_options(argc, argv, options, top_usage, 0); @@ -1265,13 +1269,18 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) if (!nr_counters) nr_counters = 1; + symbol_conf.priv_size = (sizeof(struct sym_entry) + + (nr_counters + 1) * sizeof(unsigned long)); + if (symbol_conf.vmlinux_name == NULL) + symbol_conf.try_vmlinux_path = true; + if (symbol__init(&symbol_conf) < 0) + return -1; + if (delay_secs < 1) delay_secs = 1; - parse_symbols(); parse_source(sym_filter_entry); - /* * User specified count overrides default frequency. */ @@ -1301,8 +1310,9 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) if (target_pid != -1 || profile_cpu != -1) nr_cpus = 1; + get_term_dimensions(&winsize); if (print_entries == 0) { - update_print_entries(); + update_print_entries(&winsize); signal(SIGWINCH, sig_winch_handler); } diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index d042d656c56..abb914aa7be 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -5,49 +5,66 @@ #include "util/symbol.h" #include "util/thread.h" #include "util/header.h" +#include "util/exec_cmd.h" +#include "util/trace-event.h" -#include "util/parse-options.h" +static char const *script_name; +static char const *generate_script_lang; -#include "perf.h" -#include "util/debug.h" +static int default_start_script(const char *script __attribute((unused))) +{ + return 0; +} -#include "util/trace-event.h" -#include "util/data_map.h" +static int default_stop_script(void) +{ + return 0; +} -static char const *input_name = "perf.data"; +static int default_generate_script(const char *outfile __attribute ((unused))) +{ + return 0; +} -static unsigned long total = 0; -static unsigned long total_comm = 0; +static struct scripting_ops default_scripting_ops = { + .start_script = default_start_script, + .stop_script = default_stop_script, + .process_event = print_event, + .generate_script = default_generate_script, +}; + +static struct scripting_ops *scripting_ops; -static struct perf_header *header; -static u64 sample_type; +static void setup_scripting(void) +{ + /* make sure PERF_EXEC_PATH is set for scripts */ + perf_set_argv_exec_path(perf_exec_path()); -static char *cwd; -static int cwdlen; + setup_perl_scripting(); + scripting_ops = &default_scripting_ops; +} -static int -process_comm_event(event_t *event, unsigned long offset, unsigned long head) +static int cleanup_scripting(void) { - struct thread *thread = threads__findnew(event->comm.pid); + return scripting_ops->stop_script(); +} - dump_printf("%p [%p]: PERF_RECORD_COMM: %s:%d\n", - (void *)(offset + head), - (void *)(long)(event->header.size), - event->comm.comm, event->comm.pid); +#include "util/parse-options.h" - if (thread == NULL || - thread__set_comm(thread, event->comm.comm)) { - dump_printf("problem processing PERF_RECORD_COMM, skipping event.\n"); - return -1; - } - total_comm++; +#include "perf.h" +#include "util/debug.h" - return 0; -} +#include "util/trace-event.h" +#include "util/data_map.h" +#include "util/exec_cmd.h" + +static char const *input_name = "perf.data"; -static int -process_sample_event(event_t *event, unsigned long offset, unsigned long head) +static struct perf_header *header; +static u64 sample_type; + +static int process_sample_event(event_t *event) { u64 ip = event->ip.ip; u64 timestamp = -1; @@ -72,9 +89,7 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) more_data += sizeof(u64); } - dump_printf("%p [%p]: PERF_RECORD_SAMPLE (IP, %d): %d/%d: %p period: %Ld\n", - (void *)(offset + head), - (void *)(long)(event->header.size), + dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, event->ip.pid, event->ip.tid, (void *)(long)ip, @@ -99,9 +114,10 @@ process_sample_event(event_t *event, unsigned long offset, unsigned long head) * field, although it should be the same than this perf * event pid */ - print_event(cpu, raw->data, raw->size, timestamp, thread->comm); + scripting_ops->process_event(cpu, raw->data, raw->size, + timestamp, thread->comm); } - total += period; + event__stats.total += period; return 0; } @@ -122,7 +138,7 @@ static int sample_type_check(u64 type) static struct perf_file_handler file_handler = { .process_sample_event = process_sample_event, - .process_comm_event = process_comm_event, + .process_comm_event = event__process_comm, .sample_type_check = sample_type_check, }; @@ -131,7 +147,156 @@ static int __cmd_trace(void) register_idle_thread(); register_perf_file_handler(&file_handler); - return mmap_dispatch_perf_file(&header, input_name, 0, 0, &cwdlen, &cwd); + return mmap_dispatch_perf_file(&header, input_name, + 0, 0, &event__cwdlen, &event__cwd); +} + +struct script_spec { + struct list_head node; + struct scripting_ops *ops; + char spec[0]; +}; + +LIST_HEAD(script_specs); + +static struct script_spec *script_spec__new(const char *spec, + struct scripting_ops *ops) +{ + struct script_spec *s = malloc(sizeof(*s) + strlen(spec) + 1); + + if (s != NULL) { + strcpy(s->spec, spec); + s->ops = ops; + } + + return s; +} + +static void script_spec__delete(struct script_spec *s) +{ + free(s->spec); + free(s); +} + +static void script_spec__add(struct script_spec *s) +{ + list_add_tail(&s->node, &script_specs); +} + +static struct script_spec *script_spec__find(const char *spec) +{ + struct script_spec *s; + + list_for_each_entry(s, &script_specs, node) + if (strcasecmp(s->spec, spec) == 0) + return s; + return NULL; +} + +static struct script_spec *script_spec__findnew(const char *spec, + struct scripting_ops *ops) +{ + struct script_spec *s = script_spec__find(spec); + + if (s) + return s; + + s = script_spec__new(spec, ops); + if (!s) + goto out_delete_spec; + + script_spec__add(s); + + return s; + +out_delete_spec: + script_spec__delete(s); + + return NULL; +} + +int script_spec_register(const char *spec, struct scripting_ops *ops) +{ + struct script_spec *s; + + s = script_spec__find(spec); + if (s) + return -1; + + s = script_spec__findnew(spec, ops); + if (!s) + return -1; + + return 0; +} + +static struct scripting_ops *script_spec__lookup(const char *spec) +{ + struct script_spec *s = script_spec__find(spec); + if (!s) + return NULL; + + return s->ops; +} + +static void list_available_languages(void) +{ + struct script_spec *s; + + fprintf(stderr, "\n"); + fprintf(stderr, "Scripting language extensions (used in " + "perf trace -s [spec:]script.[spec]):\n\n"); + + list_for_each_entry(s, &script_specs, node) + fprintf(stderr, " %-42s [%s]\n", s->spec, s->ops->name); + + fprintf(stderr, "\n"); +} + +static int parse_scriptname(const struct option *opt __used, + const char *str, int unset __used) +{ + char spec[PATH_MAX]; + const char *script, *ext; + int len; + + if (strcmp(str, "list") == 0) { + list_available_languages(); + return 0; + } + + script = strchr(str, ':'); + if (script) { + len = script - str; + if (len >= PATH_MAX) { + fprintf(stderr, "invalid language specifier"); + return -1; + } + strncpy(spec, str, len); + spec[len] = '\0'; + scripting_ops = script_spec__lookup(spec); + if (!scripting_ops) { + fprintf(stderr, "invalid language specifier"); + return -1; + } + script++; + } else { + script = str; + ext = strchr(script, '.'); + if (!ext) { + fprintf(stderr, "invalid script extension"); + return -1; + } + scripting_ops = script_spec__lookup(++ext); + if (!scripting_ops) { + fprintf(stderr, "invalid script extension"); + return -1; + } + } + + script_name = strdup(script); + + return 0; } static const char * const annotate_usage[] = { @@ -146,13 +311,23 @@ static const struct option options[] = { "be more verbose (show symbol address, etc)"), OPT_BOOLEAN('l', "latency", &latency_format, "show latency attributes (irqs/preemption disabled, etc)"), + OPT_CALLBACK('s', "script", NULL, "name", + "script file name (lang:script name, script name, or *)", + parse_scriptname), + OPT_STRING('g', "gen-script", &generate_script_lang, "lang", + "generate perf-trace.xx script in specified language"), + OPT_END() }; int cmd_trace(int argc, const char **argv, const char *prefix __used) { + int err; + symbol__init(0); + setup_scripting(); + argc = parse_options(argc, argv, options, annotate_usage, 0); if (argc) { /* @@ -165,5 +340,50 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) setup_pager(); - return __cmd_trace(); + if (generate_script_lang) { + struct stat perf_stat; + + int input = open(input_name, O_RDONLY); + if (input < 0) { + perror("failed to open file"); + exit(-1); + } + + err = fstat(input, &perf_stat); + if (err < 0) { + perror("failed to stat file"); + exit(-1); + } + + if (!perf_stat.st_size) { + fprintf(stderr, "zero-sized file, nothing to do!\n"); + exit(0); + } + + scripting_ops = script_spec__lookup(generate_script_lang); + if (!scripting_ops) { + fprintf(stderr, "invalid language specifier"); + return -1; + } + + header = perf_header__new(); + if (header == NULL) + return -1; + + perf_header__read(header, input); + err = scripting_ops->generate_script("perf-trace"); + goto out; + } + + if (script_name) { + err = scripting_ops->start_script(script_name); + if (err) + goto out; + } + + err = __cmd_trace(); + + cleanup_scripting(); +out: + return err; } diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h index 9b02d85091f..a3d8bf65f26 100644 --- a/tools/perf/builtin.h +++ b/tools/perf/builtin.h @@ -28,5 +28,6 @@ extern int cmd_top(int argc, const char **argv, const char *prefix); extern int cmd_trace(int argc, const char **argv, const char *prefix); extern int cmd_version(int argc, const char **argv, const char *prefix); extern int cmd_probe(int argc, const char **argv, const char *prefix); +extern int cmd_kmem(int argc, const char **argv, const char *prefix); #endif diff --git a/tools/perf/command-list.txt b/tools/perf/command-list.txt index d3a6e18e4a5..02b09ea17a3 100644 --- a/tools/perf/command-list.txt +++ b/tools/perf/command-list.txt @@ -14,3 +14,4 @@ perf-timechart mainporcelain common perf-top mainporcelain common perf-trace mainporcelain common perf-probe mainporcelain common +perf-kmem mainporcelain common diff --git a/tools/perf/perf.c b/tools/perf/perf.c index 89b82acac7d..cf64049bc9b 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -285,20 +285,21 @@ static void handle_internal_command(int argc, const char **argv) { const char *cmd = argv[0]; static struct cmd_struct commands[] = { - { "help", cmd_help, 0 }, - { "list", cmd_list, 0 }, { "buildid-list", cmd_buildid_list, 0 }, - { "record", cmd_record, 0 }, - { "report", cmd_report, 0 }, - { "bench", cmd_bench, 0 }, - { "stat", cmd_stat, 0 }, - { "timechart", cmd_timechart, 0 }, - { "top", cmd_top, 0 }, - { "annotate", cmd_annotate, 0 }, - { "version", cmd_version, 0 }, - { "trace", cmd_trace, 0 }, - { "sched", cmd_sched, 0 }, - { "probe", cmd_probe, 0 }, + { "help", cmd_help, 0 }, + { "list", cmd_list, 0 }, + { "record", cmd_record, 0 }, + { "report", cmd_report, 0 }, + { "bench", cmd_bench, 0 }, + { "stat", cmd_stat, 0 }, + { "timechart", cmd_timechart, 0 }, + { "top", cmd_top, 0 }, + { "annotate", cmd_annotate, 0 }, + { "version", cmd_version, 0 }, + { "trace", cmd_trace, 0 }, + { "sched", cmd_sched, 0 }, + { "probe", cmd_probe, 0 }, + { "kmem", cmd_kmem, 0 }, }; unsigned int i; static const char ext[] = STRIP_EXTENSION; diff --git a/tools/perf/perf.h b/tools/perf/perf.h index 216bdb223f6..454d5d55f32 100644 --- a/tools/perf/perf.h +++ b/tools/perf/perf.h @@ -53,6 +53,12 @@ #define cpu_relax() asm volatile("" ::: "memory") #endif +#ifdef __ia64__ +#include "../../arch/ia64/include/asm/unistd.h" +#define rmb() asm volatile ("mf" ::: "memory") +#define cpu_relax() asm volatile ("hint @pause" ::: "memory") +#endif + #include <time.h> #include <unistd.h> #include <sys/types.h> diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/Context.c b/tools/perf/scripts/perl/Perf-Trace-Util/Context.c new file mode 100644 index 00000000000..af78d9a52a7 --- /dev/null +++ b/tools/perf/scripts/perl/Perf-Trace-Util/Context.c @@ -0,0 +1,134 @@ +/* + * This file was generated automatically by ExtUtils::ParseXS version 2.18_02 from the + * contents of Context.xs. Do not edit this file, edit Context.xs instead. + * + * ANY CHANGES MADE HERE WILL BE LOST! + * + */ + +#line 1 "Context.xs" +/* + * Context.xs. XS interfaces for perf trace. + * + * Copyright (C) 2009 Tom Zanussi <tzanussi@gmail.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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include "EXTERN.h" +#include "perl.h" +#include "XSUB.h" +#include "../../../util/trace-event-perl.h" + +#ifndef PERL_UNUSED_VAR +# define PERL_UNUSED_VAR(var) if (0) var = var +#endif + +#line 41 "Context.c" + +XS(XS_Perf__Trace__Context_common_pc); /* prototype to pass -Wmissing-prototypes */ +XS(XS_Perf__Trace__Context_common_pc) +{ +#ifdef dVAR + dVAR; dXSARGS; +#else + dXSARGS; +#endif + if (items != 1) + Perl_croak(aTHX_ "Usage: %s(%s)", "Perf::Trace::Context::common_pc", "context"); + PERL_UNUSED_VAR(cv); /* -W */ + { + struct scripting_context * context = INT2PTR(struct scripting_context *,SvIV(ST(0))); + int RETVAL; + dXSTARG; + + RETVAL = common_pc(context); + XSprePUSH; PUSHi((IV)RETVAL); + } + XSRETURN(1); +} + + +XS(XS_Perf__Trace__Context_common_flags); /* prototype to pass -Wmissing-prototypes */ +XS(XS_Perf__Trace__Context_common_flags) +{ +#ifdef dVAR + dVAR; dXSARGS; +#else + dXSARGS; +#endif + if (items != 1) + Perl_croak(aTHX_ "Usage: %s(%s)", "Perf::Trace::Context::common_flags", "context"); + PERL_UNUSED_VAR(cv); /* -W */ + { + struct scripting_context * context = INT2PTR(struct scripting_context *,SvIV(ST(0))); + int RETVAL; + dXSTARG; + + RETVAL = common_flags(context); + XSprePUSH; PUSHi((IV)RETVAL); + } + XSRETURN(1); +} + + +XS(XS_Perf__Trace__Context_common_lock_depth); /* prototype to pass -Wmissing-prototypes */ +XS(XS_Perf__Trace__Context_common_lock_depth) +{ +#ifdef dVAR + dVAR; dXSARGS; +#else + dXSARGS; +#endif + if (items != 1) + Perl_croak(aTHX_ "Usage: %s(%s)", "Perf::Trace::Context::common_lock_depth", "context"); + PERL_UNUSED_VAR(cv); /* -W */ + { + struct scripting_context * context = INT2PTR(struct scripting_context *,SvIV(ST(0))); + int RETVAL; + dXSTARG; + + RETVAL = common_lock_depth(context); + XSprePUSH; PUSHi((IV)RETVAL); + } + XSRETURN(1); +} + +#ifdef __cplusplus +extern "C" +#endif +XS(boot_Perf__Trace__Context); /* prototype to pass -Wmissing-prototypes */ +XS(boot_Perf__Trace__Context) +{ +#ifdef dVAR + dVAR; dXSARGS; +#else + dXSARGS; +#endif + const char* file = __FILE__; + + PERL_UNUSED_VAR(cv); /* -W */ + PERL_UNUSED_VAR(items); /* -W */ + XS_VERSION_BOOTCHECK ; + + newXSproto("Perf::Trace::Context::common_pc", XS_Perf__Trace__Context_common_pc, file, "$"); + newXSproto("Perf::Trace::Context::common_flags", XS_Perf__Trace__Context_common_flags, file, "$"); + newXSproto("Perf::Trace::Context::common_lock_depth", XS_Perf__Trace__Context_common_lock_depth, file, "$"); + if (PL_unitcheckav) + call_list(PL_scopestack_ix, PL_unitcheckav); + XSRETURN_YES; +} + diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/Context.xs b/tools/perf/scripts/perl/Perf-Trace-Util/Context.xs new file mode 100644 index 00000000000..fb78006c165 --- /dev/null +++ b/tools/perf/scripts/perl/Perf-Trace-Util/Context.xs @@ -0,0 +1,41 @@ +/* + * Context.xs. XS interfaces for perf trace. + * + * Copyright (C) 2009 Tom Zanussi <tzanussi@gmail.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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include "EXTERN.h" +#include "perl.h" +#include "XSUB.h" +#include "../../../util/trace-event-perl.h" + +MODULE = Perf::Trace::Context PACKAGE = Perf::Trace::Context +PROTOTYPES: ENABLE + +int +common_pc(context) + struct scripting_context * context + +int +common_flags(context) + struct scripting_context * context + +int +common_lock_depth(context) + struct scripting_context * context + diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/Makefile.PL b/tools/perf/scripts/perl/Perf-Trace-Util/Makefile.PL new file mode 100644 index 00000000000..decdeb0f678 --- /dev/null +++ b/tools/perf/scripts/perl/Perf-Trace-Util/Makefile.PL @@ -0,0 +1,17 @@ +use 5.010000; +use ExtUtils::MakeMaker; +# See lib/ExtUtils/MakeMaker.pm for details of how to influence +# the contents of the Makefile that is written. +WriteMakefile( + NAME => 'Perf::Trace::Context', + VERSION_FROM => 'lib/Perf/Trace/Context.pm', # finds $VERSION + PREREQ_PM => {}, # e.g., Module::Name => 1.1 + ($] >= 5.005 ? ## Add these new keywords supported since 5.005 + (ABSTRACT_FROM => 'lib/Perf/Trace/Context.pm', # retrieve abstract from module + AUTHOR => 'Tom Zanussi <tzanussi@gmail.com>') : ()), + LIBS => [''], # e.g., '-lm' + DEFINE => '-I ../..', # e.g., '-DHAVE_SOMETHING' + INC => '-I.', # e.g., '-I. -I/usr/include/other' + # Un-comment this if you add C files to link with later: + OBJECT => 'Context.o', # link all the C files too +); diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/README b/tools/perf/scripts/perl/Perf-Trace-Util/README new file mode 100644 index 00000000000..9a970763079 --- /dev/null +++ b/tools/perf/scripts/perl/Perf-Trace-Util/README @@ -0,0 +1,59 @@ +Perf-Trace-Util version 0.01 +============================ + +This module contains utility functions for use with perf trace. + +Core.pm and Util.pm are pure Perl modules; Core.pm contains routines +that the core perf support for Perl calls on and should always be +'used', while Util.pm contains useful but optional utility functions +that scripts may want to use. Context.pm contains the Perl->C +interface that allows scripts to access data in the embedding perf +executable; scripts wishing to do that should 'use Context.pm'. + +The Perl->C perf interface is completely driven by Context.xs. If you +want to add new Perl functions that end up accessing C data in the +perf executable, you add desciptions of the new functions here. +scripting_context is a pointer to the perf data in the perf executable +that you want to access - it's passed as the second parameter, +$context, to all handler functions. + +After you do that: + + perl Makefile.PL # to create a Makefile for the next step + make # to create Context.c + + edit Context.c to add const to the char* file = __FILE__ line in + XS(boot_Perf__Trace__Context) to silence a warning/error. + + You can delete the Makefile, object files and anything else that was + generated e.g. blib and shared library, etc, except for of course + Context.c + + You should then be able to run the normal perf make as usual. + +INSTALLATION + +Building perf with perf trace Perl scripting should install this +module in the right place. + +You should make sure libperl and ExtUtils/Embed.pm are installed first +e.g. apt-get install libperl-dev or yum install perl-ExtUtils-Embed. + +DEPENDENCIES + +This module requires these other modules and libraries: + + None + +COPYRIGHT AND LICENCE + +Copyright (C) 2009 by Tom Zanussi <tzanussi@gmail.com> + +This library is free software; you can redistribute it and/or modify +it under the same terms as Perl itself, either Perl version 5.10.0 or, +at your option, any later version of Perl 5 you may have available. + +Alternatively, this software may be distributed under the terms of the +GNU General Public License ("GPL") version 2 as published by the Free +Software Foundation. + diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Context.pm b/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Context.pm new file mode 100644 index 00000000000..6c7f3659cb1 --- /dev/null +++ b/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Context.pm @@ -0,0 +1,55 @@ +package Perf::Trace::Context; + +use 5.010000; +use strict; +use warnings; + +require Exporter; + +our @ISA = qw(Exporter); + +our %EXPORT_TAGS = ( 'all' => [ qw( +) ] ); + +our @EXPORT_OK = ( @{ $EXPORT_TAGS{'all'} } ); + +our @EXPORT = qw( + common_pc common_flags common_lock_depth +); + +our $VERSION = '0.01'; + +require XSLoader; +XSLoader::load('Perf::Trace::Context', $VERSION); + +1; +__END__ +=head1 NAME + +Perf::Trace::Context - Perl extension for accessing functions in perf. + +=head1 SYNOPSIS + + use Perf::Trace::Context; + +=head1 SEE ALSO + +Perf (trace) documentation + +=head1 AUTHOR + +Tom Zanussi, E<lt>tzanussi@gmail.com<gt> + +=head1 COPYRIGHT AND LICENSE + +Copyright (C) 2009 by Tom Zanussi + +This library is free software; you can redistribute it and/or modify +it under the same terms as Perl itself, either Perl version 5.10.0 or, +at your option, any later version of Perl 5 you may have available. + +Alternatively, this software may be distributed under the terms of the +GNU General Public License ("GPL") version 2 as published by the Free +Software Foundation. + +=cut diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Core.pm b/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Core.pm new file mode 100644 index 00000000000..9df376a9f62 --- /dev/null +++ b/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Core.pm @@ -0,0 +1,192 @@ +package Perf::Trace::Core; + +use 5.010000; +use strict; +use warnings; + +require Exporter; + +our @ISA = qw(Exporter); + +our %EXPORT_TAGS = ( 'all' => [ qw( +) ] ); + +our @EXPORT_OK = ( @{ $EXPORT_TAGS{'all'} } ); + +our @EXPORT = qw( +define_flag_field define_flag_value flag_str dump_flag_fields +define_symbolic_field define_symbolic_value symbol_str dump_symbolic_fields +trace_flag_str +); + +our $VERSION = '0.01'; + +my %trace_flags = (0x00 => "NONE", + 0x01 => "IRQS_OFF", + 0x02 => "IRQS_NOSUPPORT", + 0x04 => "NEED_RESCHED", + 0x08 => "HARDIRQ", + 0x10 => "SOFTIRQ"); + +sub trace_flag_str +{ + my ($value) = @_; + + my $string; + + my $print_delim = 0; + + foreach my $idx (sort {$a <=> $b} keys %trace_flags) { + if (!$value && !$idx) { + $string .= "NONE"; + last; + } + + if ($idx && ($value & $idx) == $idx) { + if ($print_delim) { + $string .= " | "; + } + $string .= "$trace_flags{$idx}"; + $print_delim = 1; + $value &= ~$idx; + } + } + + return $string; +} + +my %flag_fields; +my %symbolic_fields; + +sub flag_str +{ + my ($event_name, $field_name, $value) = @_; + + my $string; + + if ($flag_fields{$event_name}{$field_name}) { + my $print_delim = 0; + foreach my $idx (sort {$a <=> $b} keys %{$flag_fields{$event_name}{$field_name}{"values"}}) { + if (!$value && !$idx) { + $string .= "$flag_fields{$event_name}{$field_name}{'values'}{$idx}"; + last; + } + if ($idx && ($value & $idx) == $idx) { + if ($print_delim && $flag_fields{$event_name}{$field_name}{'delim'}) { + $string .= " $flag_fields{$event_name}{$field_name}{'delim'} "; + } + $string .= "$flag_fields{$event_name}{$field_name}{'values'}{$idx}"; + $print_delim = 1; + $value &= ~$idx; + } + } + } + + return $string; +} + +sub define_flag_field +{ + my ($event_name, $field_name, $delim) = @_; + + $flag_fields{$event_name}{$field_name}{"delim"} = $delim; +} + +sub define_flag_value +{ + my ($event_name, $field_name, $value, $field_str) = @_; + + $flag_fields{$event_name}{$field_name}{"values"}{$value} = $field_str; +} + +sub dump_flag_fields +{ + for my $event (keys %flag_fields) { + print "event $event:\n"; + for my $field (keys %{$flag_fields{$event}}) { + print " field: $field:\n"; + print " delim: $flag_fields{$event}{$field}{'delim'}\n"; + foreach my $idx (sort {$a <=> $b} keys %{$flag_fields{$event}{$field}{"values"}}) { + print " value $idx: $flag_fields{$event}{$field}{'values'}{$idx}\n"; + } + } + } +} + +sub symbol_str +{ + my ($event_name, $field_name, $value) = @_; + + if ($symbolic_fields{$event_name}{$field_name}) { + foreach my $idx (sort {$a <=> $b} keys %{$symbolic_fields{$event_name}{$field_name}{"values"}}) { + if (!$value && !$idx) { + return "$symbolic_fields{$event_name}{$field_name}{'values'}{$idx}"; + last; + } + if ($value == $idx) { + return "$symbolic_fields{$event_name}{$field_name}{'values'}{$idx}"; + } + } + } + + return undef; +} + +sub define_symbolic_field +{ + my ($event_name, $field_name) = @_; + + # nothing to do, really +} + +sub define_symbolic_value +{ + my ($event_name, $field_name, $value, $field_str) = @_; + + $symbolic_fields{$event_name}{$field_name}{"values"}{$value} = $field_str; +} + +sub dump_symbolic_fields +{ + for my $event (keys %symbolic_fields) { + print "event $event:\n"; + for my $field (keys %{$symbolic_fields{$event}}) { + print " field: $field:\n"; + foreach my $idx (sort {$a <=> $b} keys %{$symbolic_fields{$event}{$field}{"values"}}) { + print " value $idx: $symbolic_fields{$event}{$field}{'values'}{$idx}\n"; + } + } + } +} + +1; +__END__ +=head1 NAME + +Perf::Trace::Core - Perl extension for perf trace + +=head1 SYNOPSIS + + use Perf::Trace::Core + +=head1 SEE ALSO + +Perf (trace) documentation + +=head1 AUTHOR + +Tom Zanussi, E<lt>tzanussi@gmail.com<gt> + +=head1 COPYRIGHT AND LICENSE + +Copyright (C) 2009 by Tom Zanussi + +This library is free software; you can redistribute it and/or modify +it under the same terms as Perl itself, either Perl version 5.10.0 or, +at your option, any later version of Perl 5 you may have available. + +Alternatively, this software may be distributed under the terms of the +GNU General Public License ("GPL") version 2 as published by the Free +Software Foundation. + +=cut diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Util.pm b/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Util.pm new file mode 100644 index 00000000000..052f132ced2 --- /dev/null +++ b/tools/perf/scripts/perl/Perf-Trace-Util/lib/Perf/Trace/Util.pm @@ -0,0 +1,88 @@ +package Perf::Trace::Util; + +use 5.010000; +use strict; +use warnings; + +require Exporter; + +our @ISA = qw(Exporter); + +our %EXPORT_TAGS = ( 'all' => [ qw( +) ] ); + +our @EXPORT_OK = ( @{ $EXPORT_TAGS{'all'} } ); + +our @EXPORT = qw( +avg nsecs nsecs_secs nsecs_nsecs nsecs_usecs print_nsecs +); + +our $VERSION = '0.01'; + +sub avg +{ + my ($total, $n) = @_; + + return $total / $n; +} + +my $NSECS_PER_SEC = 1000000000; + +sub nsecs +{ + my ($secs, $nsecs) = @_; + + return $secs * $NSECS_PER_SEC + $nsecs; +} + +sub nsecs_secs { + my ($nsecs) = @_; + + return $nsecs / $NSECS_PER_SEC; +} + +sub nsecs_nsecs { + my ($nsecs) = @_; + + return $nsecs - nsecs_secs($nsecs); +} + +sub nsecs_str { + my ($nsecs) = @_; + + my $str = sprintf("%5u.%09u", nsecs_secs($nsecs), nsecs_nsecs($nsecs)); + + return $str; +} + +1; +__END__ +=head1 NAME + +Perf::Trace::Util - Perl extension for perf trace + +=head1 SYNOPSIS + + use Perf::Trace::Util; + +=head1 SEE ALSO + +Perf (trace) documentation + +=head1 AUTHOR + +Tom Zanussi, E<lt>tzanussi@gmail.com<gt> + +=head1 COPYRIGHT AND LICENSE + +Copyright (C) 2009 by Tom Zanussi + +This library is free software; you can redistribute it and/or modify +it under the same terms as Perl itself, either Perl version 5.10.0 or, +at your option, any later version of Perl 5 you may have available. + +Alternatively, this software may be distributed under the terms of the +GNU General Public License ("GPL") version 2 as published by the Free +Software Foundation. + +=cut diff --git a/tools/perf/scripts/perl/Perf-Trace-Util/typemap b/tools/perf/scripts/perl/Perf-Trace-Util/typemap new file mode 100644 index 00000000000..840836804aa --- /dev/null +++ b/tools/perf/scripts/perl/Perf-Trace-Util/typemap @@ -0,0 +1 @@ +struct scripting_context * T_PTR diff --git a/tools/perf/scripts/perl/bin/check-perf-trace-record b/tools/perf/scripts/perl/bin/check-perf-trace-record new file mode 100644 index 00000000000..c7ec5de2f53 --- /dev/null +++ b/tools/perf/scripts/perl/bin/check-perf-trace-record @@ -0,0 +1,7 @@ +#!/bin/bash +perf record -c 1 -f -a -M -R -e kmem:kmalloc -e irq:softirq_entry + + + + + diff --git a/tools/perf/scripts/perl/bin/check-perf-trace-report b/tools/perf/scripts/perl/bin/check-perf-trace-report new file mode 100644 index 00000000000..89948b01502 --- /dev/null +++ b/tools/perf/scripts/perl/bin/check-perf-trace-report @@ -0,0 +1,5 @@ +#!/bin/bash +perf trace -s ~/libexec/perf-core/scripts/perl/check-perf-trace.pl + + + diff --git a/tools/perf/scripts/perl/bin/rw-by-file-record b/tools/perf/scripts/perl/bin/rw-by-file-record new file mode 100644 index 00000000000..b25056ebf96 --- /dev/null +++ b/tools/perf/scripts/perl/bin/rw-by-file-record @@ -0,0 +1,2 @@ +#!/bin/bash +perf record -c 1 -f -a -M -R -e syscalls:sys_enter_read -e syscalls:sys_enter_write diff --git a/tools/perf/scripts/perl/bin/rw-by-file-report b/tools/perf/scripts/perl/bin/rw-by-file-report new file mode 100644 index 00000000000..f5dcf9cb5bd --- /dev/null +++ b/tools/perf/scripts/perl/bin/rw-by-file-report @@ -0,0 +1,5 @@ +#!/bin/bash +perf trace -s ~/libexec/perf-core/scripts/perl/rw-by-file.pl + + + diff --git a/tools/perf/scripts/perl/bin/rw-by-pid-record b/tools/perf/scripts/perl/bin/rw-by-pid-record new file mode 100644 index 00000000000..8903979c5b6 --- /dev/null +++ b/tools/perf/scripts/perl/bin/rw-by-pid-record @@ -0,0 +1,2 @@ +#!/bin/bash +perf record -c 1 -f -a -M -R -e syscalls:sys_enter_read -e syscalls:sys_exit_read -e syscalls:sys_enter_write -e syscalls:sys_exit_write diff --git a/tools/perf/scripts/perl/bin/rw-by-pid-report b/tools/perf/scripts/perl/bin/rw-by-pid-report new file mode 100644 index 00000000000..cea16f78a3a --- /dev/null +++ b/tools/perf/scripts/perl/bin/rw-by-pid-report @@ -0,0 +1,5 @@ +#!/bin/bash +perf trace -s ~/libexec/perf-core/scripts/perl/rw-by-pid.pl + + + diff --git a/tools/perf/scripts/perl/bin/wakeup-latency-record b/tools/perf/scripts/perl/bin/wakeup-latency-record new file mode 100644 index 00000000000..6abedda911a --- /dev/null +++ b/tools/perf/scripts/perl/bin/wakeup-latency-record @@ -0,0 +1,6 @@ +#!/bin/bash +perf record -c 1 -f -a -M -R -e sched:sched_switch -e sched:sched_wakeup + + + + diff --git a/tools/perf/scripts/perl/bin/wakeup-latency-report b/tools/perf/scripts/perl/bin/wakeup-latency-report new file mode 100644 index 00000000000..85769dc456e --- /dev/null +++ b/tools/perf/scripts/perl/bin/wakeup-latency-report @@ -0,0 +1,5 @@ +#!/bin/bash +perf trace -s ~/libexec/perf-core/scripts/perl/wakeup-latency.pl + + + diff --git a/tools/perf/scripts/perl/bin/workqueue-stats-record b/tools/perf/scripts/perl/bin/workqueue-stats-record new file mode 100644 index 00000000000..fce6637b19b --- /dev/null +++ b/tools/perf/scripts/perl/bin/workqueue-stats-record @@ -0,0 +1,2 @@ +#!/bin/bash +perf record -c 1 -f -a -M -R -e workqueue:workqueue_creation -e workqueue:workqueue_destruction -e workqueue:workqueue_execution -e workqueue:workqueue_insertion diff --git a/tools/perf/scripts/perl/bin/workqueue-stats-report b/tools/perf/scripts/perl/bin/workqueue-stats-report new file mode 100644 index 00000000000..aa68435be92 --- /dev/null +++ b/tools/perf/scripts/perl/bin/workqueue-stats-report @@ -0,0 +1,6 @@ +#!/bin/bash +perf trace -s ~/libexec/perf-core/scripts/perl/workqueue-stats.pl + + + + diff --git a/tools/perf/scripts/perl/check-perf-trace.pl b/tools/perf/scripts/perl/check-perf-trace.pl new file mode 100644 index 00000000000..4e7dc0a407a --- /dev/null +++ b/tools/perf/scripts/perl/check-perf-trace.pl @@ -0,0 +1,106 @@ +# perf trace event handlers, generated by perf trace -g perl +# (c) 2009, Tom Zanussi <tzanussi@gmail.com> +# Licensed under the terms of the GNU GPL License version 2 + +# This script tests basic functionality such as flag and symbol +# strings, common_xxx() calls back into perf, begin, end, unhandled +# events, etc. Basically, if this script runs successfully and +# displays expected results, perl scripting support should be ok. + +use lib "$ENV{'PERF_EXEC_PATH'}/scripts/perl/Perf-Trace-Util/lib"; +use lib "./Perf-Trace-Util/lib"; +use Perf::Trace::Core; +use Perf::Trace::Context; +use Perf::Trace::Util; + +sub trace_begin +{ + print "trace_begin\n"; +} + +sub trace_end +{ + print "trace_end\n"; + + print_unhandled(); +} + +sub irq::softirq_entry +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $vec) = @_; + + print_header($event_name, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm); + + print_uncommon($context); + + printf("vec=%s\n", + symbol_str("irq::softirq_entry", "vec", $vec)); +} + +sub kmem::kmalloc +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $call_site, $ptr, $bytes_req, $bytes_alloc, + $gfp_flags) = @_; + + print_header($event_name, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm); + + print_uncommon($context); + + printf("call_site=%p, ptr=%p, bytes_req=%u, bytes_alloc=%u, ". + "gfp_flags=%s\n", + $call_site, $ptr, $bytes_req, $bytes_alloc, + + flag_str("kmem::kmalloc", "gfp_flags", $gfp_flags)); +} + +# print trace fields not included in handler args +sub print_uncommon +{ + my ($context) = @_; + + printf("common_preempt_count=%d, common_flags=%s, common_lock_depth=%d, ", + common_pc($context), trace_flag_str(common_flags($context)), + common_lock_depth($context)); + +} + +my %unhandled; + +sub print_unhandled +{ + if ((scalar keys %unhandled) == 0) { + return; + } + + print "\nunhandled events:\n\n"; + + printf("%-40s %10s\n", "event", "count"); + printf("%-40s %10s\n", "----------------------------------------", + "-----------"); + + foreach my $event_name (keys %unhandled) { + printf("%-40s %10d\n", $event_name, $unhandled{$event_name}); + } +} + +sub trace_unhandled +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm) = @_; + + $unhandled{$event_name}++; +} + +sub print_header +{ + my ($event_name, $cpu, $secs, $nsecs, $pid, $comm) = @_; + + printf("%-20s %5u %05u.%09u %8u %-20s ", + $event_name, $cpu, $secs, $nsecs, $pid, $comm); +} diff --git a/tools/perf/scripts/perl/rw-by-file.pl b/tools/perf/scripts/perl/rw-by-file.pl new file mode 100644 index 00000000000..61f91561d84 --- /dev/null +++ b/tools/perf/scripts/perl/rw-by-file.pl @@ -0,0 +1,105 @@ +#!/usr/bin/perl -w +# (c) 2009, Tom Zanussi <tzanussi@gmail.com> +# Licensed under the terms of the GNU GPL License version 2 + +# Display r/w activity for files read/written to for a given program + +# The common_* event handler fields are the most useful fields common to +# all events. They don't necessarily correspond to the 'common_*' fields +# in the status files. Those fields not available as handler params can +# be retrieved via script functions of the form get_common_*(). + +use 5.010000; +use strict; +use warnings; + +use lib "$ENV{'PERF_EXEC_PATH'}/scripts/perl/Perf-Trace-Util/lib"; +use lib "./Perf-Trace-Util/lib"; +use Perf::Trace::Core; +use Perf::Trace::Util; + +# change this to the comm of the program you're interested in +my $for_comm = "perf"; + +my %reads; +my %writes; + +sub syscalls::sys_enter_read +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, $nr, $fd, $buf, $count) = @_; + + if ($common_comm eq $for_comm) { + $reads{$fd}{bytes_requested} += $count; + $reads{$fd}{total_reads}++; + } +} + +sub syscalls::sys_enter_write +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, $nr, $fd, $buf, $count) = @_; + + if ($common_comm eq $for_comm) { + $writes{$fd}{bytes_written} += $count; + $writes{$fd}{total_writes}++; + } +} + +sub trace_end +{ + printf("file read counts for $for_comm:\n\n"); + + printf("%6s %10s %10s\n", "fd", "# reads", "bytes_requested"); + printf("%6s %10s %10s\n", "------", "----------", "-----------"); + + foreach my $fd (sort {$reads{$b}{bytes_requested} <=> + $reads{$a}{bytes_requested}} keys %reads) { + my $total_reads = $reads{$fd}{total_reads}; + my $bytes_requested = $reads{$fd}{bytes_requested}; + printf("%6u %10u %10u\n", $fd, $total_reads, $bytes_requested); + } + + printf("\nfile write counts for $for_comm:\n\n"); + + printf("%6s %10s %10s\n", "fd", "# writes", "bytes_written"); + printf("%6s %10s %10s\n", "------", "----------", "-----------"); + + foreach my $fd (sort {$writes{$b}{bytes_written} <=> + $writes{$a}{bytes_written}} keys %writes) { + my $total_writes = $writes{$fd}{total_writes}; + my $bytes_written = $writes{$fd}{bytes_written}; + printf("%6u %10u %10u\n", $fd, $total_writes, $bytes_written); + } + + print_unhandled(); +} + +my %unhandled; + +sub print_unhandled +{ + if ((scalar keys %unhandled) == 0) { + return; + } + + print "\nunhandled events:\n\n"; + + printf("%-40s %10s\n", "event", "count"); + printf("%-40s %10s\n", "----------------------------------------", + "-----------"); + + foreach my $event_name (keys %unhandled) { + printf("%-40s %10d\n", $event_name, $unhandled{$event_name}); + } +} + +sub trace_unhandled +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm) = @_; + + $unhandled{$event_name}++; +} + + diff --git a/tools/perf/scripts/perl/rw-by-pid.pl b/tools/perf/scripts/perl/rw-by-pid.pl new file mode 100644 index 00000000000..da601fae1a0 --- /dev/null +++ b/tools/perf/scripts/perl/rw-by-pid.pl @@ -0,0 +1,170 @@ +#!/usr/bin/perl -w +# (c) 2009, Tom Zanussi <tzanussi@gmail.com> +# Licensed under the terms of the GNU GPL License version 2 + +# Display r/w activity for all processes + +# The common_* event handler fields are the most useful fields common to +# all events. They don't necessarily correspond to the 'common_*' fields +# in the status files. Those fields not available as handler params can +# be retrieved via script functions of the form get_common_*(). + +use 5.010000; +use strict; +use warnings; + +use lib "$ENV{'PERF_EXEC_PATH'}/scripts/perl/Perf-Trace-Util/lib"; +use lib "./Perf-Trace-Util/lib"; +use Perf::Trace::Core; +use Perf::Trace::Util; + +my %reads; +my %writes; + +sub syscalls::sys_exit_read +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $nr, $ret) = @_; + + if ($ret > 0) { + $reads{$common_pid}{bytes_read} += $ret; + } else { + if (!defined ($reads{$common_pid}{bytes_read})) { + $reads{$common_pid}{bytes_read} = 0; + } + $reads{$common_pid}{errors}{$ret}++; + } +} + +sub syscalls::sys_enter_read +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $nr, $fd, $buf, $count) = @_; + + $reads{$common_pid}{bytes_requested} += $count; + $reads{$common_pid}{total_reads}++; + $reads{$common_pid}{comm} = $common_comm; +} + +sub syscalls::sys_exit_write +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $nr, $ret) = @_; + + if ($ret <= 0) { + $writes{$common_pid}{errors}{$ret}++; + } +} + +sub syscalls::sys_enter_write +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $nr, $fd, $buf, $count) = @_; + + $writes{$common_pid}{bytes_written} += $count; + $writes{$common_pid}{total_writes}++; + $writes{$common_pid}{comm} = $common_comm; +} + +sub trace_end +{ + printf("read counts by pid:\n\n"); + + printf("%6s %20s %10s %10s %10s\n", "pid", "comm", + "# reads", "bytes_requested", "bytes_read"); + printf("%6s %-20s %10s %10s %10s\n", "------", "--------------------", + "-----------", "----------", "----------"); + + foreach my $pid (sort {$reads{$b}{bytes_read} <=> + $reads{$a}{bytes_read}} keys %reads) { + my $comm = $reads{$pid}{comm}; + my $total_reads = $reads{$pid}{total_reads}; + my $bytes_requested = $reads{$pid}{bytes_requested}; + my $bytes_read = $reads{$pid}{bytes_read}; + + printf("%6s %-20s %10s %10s %10s\n", $pid, $comm, + $total_reads, $bytes_requested, $bytes_read); + } + + printf("\nfailed reads by pid:\n\n"); + + printf("%6s %20s %6s %10s\n", "pid", "comm", "error #", "# errors"); + printf("%6s %20s %6s %10s\n", "------", "--------------------", + "------", "----------"); + + foreach my $pid (keys %reads) { + my $comm = $reads{$pid}{comm}; + foreach my $err (sort {$reads{$b}{comm} cmp $reads{$a}{comm}} + keys %{$reads{$pid}{errors}}) { + my $errors = $reads{$pid}{errors}{$err}; + + printf("%6d %-20s %6d %10s\n", $pid, $comm, $err, $errors); + } + } + + printf("\nwrite counts by pid:\n\n"); + + printf("%6s %20s %10s %10s\n", "pid", "comm", + "# writes", "bytes_written"); + printf("%6s %-20s %10s %10s\n", "------", "--------------------", + "-----------", "----------"); + + foreach my $pid (sort {$writes{$b}{bytes_written} <=> + $writes{$a}{bytes_written}} keys %writes) { + my $comm = $writes{$pid}{comm}; + my $total_writes = $writes{$pid}{total_writes}; + my $bytes_written = $writes{$pid}{bytes_written}; + + printf("%6s %-20s %10s %10s\n", $pid, $comm, + $total_writes, $bytes_written); + } + + printf("\nfailed writes by pid:\n\n"); + + printf("%6s %20s %6s %10s\n", "pid", "comm", "error #", "# errors"); + printf("%6s %20s %6s %10s\n", "------", "--------------------", + "------", "----------"); + + foreach my $pid (keys %writes) { + my $comm = $writes{$pid}{comm}; + foreach my $err (sort {$writes{$b}{comm} cmp $writes{$a}{comm}} + keys %{$writes{$pid}{errors}}) { + my $errors = $writes{$pid}{errors}{$err}; + + printf("%6d %-20s %6d %10s\n", $pid, $comm, $err, $errors); + } + } + + print_unhandled(); +} + +my %unhandled; + +sub print_unhandled +{ + if ((scalar keys %unhandled) == 0) { + return; + } + + print "\nunhandled events:\n\n"; + + printf("%-40s %10s\n", "event", "count"); + printf("%-40s %10s\n", "----------------------------------------", + "-----------"); + + foreach my $event_name (keys %unhandled) { + printf("%-40s %10d\n", $event_name, $unhandled{$event_name}); + } +} + +sub trace_unhandled +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm) = @_; + + $unhandled{$event_name}++; +} diff --git a/tools/perf/scripts/perl/wakeup-latency.pl b/tools/perf/scripts/perl/wakeup-latency.pl new file mode 100644 index 00000000000..ed58ef284e2 --- /dev/null +++ b/tools/perf/scripts/perl/wakeup-latency.pl @@ -0,0 +1,103 @@ +#!/usr/bin/perl -w +# (c) 2009, Tom Zanussi <tzanussi@gmail.com> +# Licensed under the terms of the GNU GPL License version 2 + +# Display avg/min/max wakeup latency + +# The common_* event handler fields are the most useful fields common to +# all events. They don't necessarily correspond to the 'common_*' fields +# in the status files. Those fields not available as handler params can +# be retrieved via script functions of the form get_common_*(). + +use 5.010000; +use strict; +use warnings; + +use lib "$ENV{'PERF_EXEC_PATH'}/scripts/perl/Perf-Trace-Util/lib"; +use lib "./Perf-Trace-Util/lib"; +use Perf::Trace::Core; +use Perf::Trace::Util; + +my %last_wakeup; + +my $max_wakeup_latency; +my $min_wakeup_latency; +my $total_wakeup_latency; +my $total_wakeups; + +sub sched::sched_switch +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $prev_comm, $prev_pid, $prev_prio, $prev_state, $next_comm, $next_pid, + $next_prio) = @_; + + my $wakeup_ts = $last_wakeup{$common_cpu}{ts}; + if ($wakeup_ts) { + my $switch_ts = nsecs($common_secs, $common_nsecs); + my $wakeup_latency = $switch_ts - $wakeup_ts; + if ($wakeup_latency > $max_wakeup_latency) { + $max_wakeup_latency = $wakeup_latency; + } + if ($wakeup_latency < $min_wakeup_latency) { + $min_wakeup_latency = $wakeup_latency; + } + $total_wakeup_latency += $wakeup_latency; + $total_wakeups++; + } + $last_wakeup{$common_cpu}{ts} = 0; +} + +sub sched::sched_wakeup +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $comm, $pid, $prio, $success, $target_cpu) = @_; + + $last_wakeup{$target_cpu}{ts} = nsecs($common_secs, $common_nsecs); +} + +sub trace_begin +{ + $min_wakeup_latency = 1000000000; + $max_wakeup_latency = 0; +} + +sub trace_end +{ + printf("wakeup_latency stats:\n\n"); + print "total_wakeups: $total_wakeups\n"; + printf("avg_wakeup_latency (ns): %u\n", + avg($total_wakeup_latency, $total_wakeups)); + printf("min_wakeup_latency (ns): %u\n", $min_wakeup_latency); + printf("max_wakeup_latency (ns): %u\n", $max_wakeup_latency); + + print_unhandled(); +} + +my %unhandled; + +sub print_unhandled +{ + if ((scalar keys %unhandled) == 0) { + return; + } + + print "\nunhandled events:\n\n"; + + printf("%-40s %10s\n", "event", "count"); + printf("%-40s %10s\n", "----------------------------------------", + "-----------"); + + foreach my $event_name (keys %unhandled) { + printf("%-40s %10d\n", $event_name, $unhandled{$event_name}); + } +} + +sub trace_unhandled +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm) = @_; + + $unhandled{$event_name}++; +} diff --git a/tools/perf/scripts/perl/workqueue-stats.pl b/tools/perf/scripts/perl/workqueue-stats.pl new file mode 100644 index 00000000000..511302c8a49 --- /dev/null +++ b/tools/perf/scripts/perl/workqueue-stats.pl @@ -0,0 +1,129 @@ +#!/usr/bin/perl -w +# (c) 2009, Tom Zanussi <tzanussi@gmail.com> +# Licensed under the terms of the GNU GPL License version 2 + +# Displays workqueue stats +# +# Usage: +# +# perf record -c 1 -f -a -R -e workqueue:workqueue_creation -e +# workqueue:workqueue_destruction -e workqueue:workqueue_execution +# -e workqueue:workqueue_insertion +# +# perf trace -p -s tools/perf/scripts/perl/workqueue-stats.pl + +use 5.010000; +use strict; +use warnings; + +use lib "$ENV{'PERF_EXEC_PATH'}/scripts/perl/Perf-Trace-Util/lib"; +use lib "./Perf-Trace-Util/lib"; +use Perf::Trace::Core; +use Perf::Trace::Util; + +my @cpus; + +sub workqueue::workqueue_destruction +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $thread_comm, $thread_pid) = @_; + + $cpus[$common_cpu]{$thread_pid}{destroyed}++; + $cpus[$common_cpu]{$thread_pid}{comm} = $thread_comm; +} + +sub workqueue::workqueue_creation +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $thread_comm, $thread_pid, $cpu) = @_; + + $cpus[$common_cpu]{$thread_pid}{created}++; + $cpus[$common_cpu]{$thread_pid}{comm} = $thread_comm; +} + +sub workqueue::workqueue_execution +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $thread_comm, $thread_pid, $func) = @_; + + $cpus[$common_cpu]{$thread_pid}{executed}++; + $cpus[$common_cpu]{$thread_pid}{comm} = $thread_comm; +} + +sub workqueue::workqueue_insertion +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm, + $thread_comm, $thread_pid, $func) = @_; + + $cpus[$common_cpu]{$thread_pid}{inserted}++; + $cpus[$common_cpu]{$thread_pid}{comm} = $thread_comm; +} + +sub trace_end +{ + print "workqueue work stats:\n\n"; + my $cpu = 0; + printf("%3s %6s %6s\t%-20s\n", "cpu", "ins", "exec", "name"); + printf("%3s %6s %6s\t%-20s\n", "---", "---", "----", "----"); + foreach my $pidhash (@cpus) { + while ((my $pid, my $wqhash) = each %$pidhash) { + my $ins = $$wqhash{'inserted'}; + my $exe = $$wqhash{'executed'}; + my $comm = $$wqhash{'comm'}; + if ($ins || $exe) { + printf("%3u %6u %6u\t%-20s\n", $cpu, $ins, $exe, $comm); + } + } + $cpu++; + } + + $cpu = 0; + print "\nworkqueue lifecycle stats:\n\n"; + printf("%3s %6s %6s\t%-20s\n", "cpu", "created", "destroyed", "name"); + printf("%3s %6s %6s\t%-20s\n", "---", "-------", "---------", "----"); + foreach my $pidhash (@cpus) { + while ((my $pid, my $wqhash) = each %$pidhash) { + my $created = $$wqhash{'created'}; + my $destroyed = $$wqhash{'destroyed'}; + my $comm = $$wqhash{'comm'}; + if ($created || $destroyed) { + printf("%3u %6u %6u\t%-20s\n", $cpu, $created, $destroyed, + $comm); + } + } + $cpu++; + } + + print_unhandled(); +} + +my %unhandled; + +sub print_unhandled +{ + if ((scalar keys %unhandled) == 0) { + return; + } + + print "\nunhandled events:\n\n"; + + printf("%-40s %10s\n", "event", "count"); + printf("%-40s %10s\n", "----------------------------------------", + "-----------"); + + foreach my $event_name (keys %unhandled) { + printf("%-40s %10d\n", $event_name, $unhandled{$event_name}); + } +} + +sub trace_unhandled +{ + my ($event_name, $context, $common_cpu, $common_secs, $common_nsecs, + $common_pid, $common_comm) = @_; + + $unhandled{$event_name}++; +} diff --git a/tools/perf/util/ctype.c b/tools/perf/util/ctype.c index 0b791bd346b..35073621e5d 100644 --- a/tools/perf/util/ctype.c +++ b/tools/perf/util/ctype.c @@ -29,3 +29,11 @@ unsigned char sane_ctype[256] = { A, A, A, A, A, A, A, A, A, A, A, R, R, P, P, 0, /* 112..127 */ /* Nothing in the 128.. range */ }; + +const char *graph_line = + "_____________________________________________________________________" + "_____________________________________________________________________"; +const char *graph_dotted_line = + "---------------------------------------------------------------------" + "---------------------------------------------------------------------" + "---------------------------------------------------------------------"; diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index 14cb8465eb0..ca0bedf637c 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -8,11 +8,9 @@ static struct perf_file_handler *curr_handler; static unsigned long mmap_window = 32; static char __cwd[PATH_MAX]; -static int -process_event_stub(event_t *event __used, - unsigned long offset __used, - unsigned long head __used) +static int process_event_stub(event_t *event __used) { + dump_printf(": unhandled!\n"); return 0; } @@ -40,30 +38,62 @@ void register_perf_file_handler(struct perf_file_handler *handler) curr_handler = handler; } +static const char *event__name[] = { + [0] = "TOTAL", + [PERF_RECORD_MMAP] = "MMAP", + [PERF_RECORD_LOST] = "LOST", + [PERF_RECORD_COMM] = "COMM", + [PERF_RECORD_EXIT] = "EXIT", + [PERF_RECORD_THROTTLE] = "THROTTLE", + [PERF_RECORD_UNTHROTTLE] = "UNTHROTTLE", + [PERF_RECORD_FORK] = "FORK", + [PERF_RECORD_READ] = "READ", + [PERF_RECORD_SAMPLE] = "SAMPLE", +}; + +unsigned long event__total[PERF_RECORD_MAX]; + +void event__print_totals(void) +{ + int i; + for (i = 0; i < PERF_RECORD_MAX; ++i) + pr_info("%10s events: %10ld\n", + event__name[i], event__total[i]); +} + static int process_event(event_t *event, unsigned long offset, unsigned long head) { trace_event(event); + if (event->header.type < PERF_RECORD_MAX) { + dump_printf("%p [%p]: PERF_RECORD_%s", + (void *)(offset + head), + (void *)(long)(event->header.size), + event__name[event->header.type]); + ++event__total[0]; + ++event__total[event->header.type]; + } + switch (event->header.type) { case PERF_RECORD_SAMPLE: - return curr_handler->process_sample_event(event, offset, head); + return curr_handler->process_sample_event(event); case PERF_RECORD_MMAP: - return curr_handler->process_mmap_event(event, offset, head); + return curr_handler->process_mmap_event(event); case PERF_RECORD_COMM: - return curr_handler->process_comm_event(event, offset, head); + return curr_handler->process_comm_event(event); case PERF_RECORD_FORK: - return curr_handler->process_fork_event(event, offset, head); + return curr_handler->process_fork_event(event); case PERF_RECORD_EXIT: - return curr_handler->process_exit_event(event, offset, head); + return curr_handler->process_exit_event(event); case PERF_RECORD_LOST: - return curr_handler->process_lost_event(event, offset, head); + return curr_handler->process_lost_event(event); case PERF_RECORD_READ: - return curr_handler->process_read_event(event, offset, head); + return curr_handler->process_read_event(event); case PERF_RECORD_THROTTLE: - return curr_handler->process_throttle_event(event, offset, head); + return curr_handler->process_throttle_event(event); case PERF_RECORD_UNTHROTTLE: - return curr_handler->process_unthrottle_event(event, offset, head); + return curr_handler->process_unthrottle_event(event); default: curr_handler->total_unknown++; return -1; @@ -106,7 +136,7 @@ int mmap_dispatch_perf_file(struct perf_header **pheader, int *cwdlen, char **cwd) { - int ret, rc = EXIT_FAILURE; + int err; struct perf_header *header; unsigned long head, shift; unsigned long offset = 0; @@ -118,56 +148,63 @@ int mmap_dispatch_perf_file(struct perf_header **pheader, int input; char *buf; - if (!curr_handler) - die("Forgot to register perf file handler"); + if (curr_handler == NULL) { + pr_debug("Forgot to register perf file handler\n"); + return -EINVAL; + } page_size = getpagesize(); input = open(input_name, O_RDONLY); if (input < 0) { - fprintf(stderr, " failed to open file: %s", input_name); + pr_err("Failed to open file: %s", input_name); if (!strcmp(input_name, "perf.data")) - fprintf(stderr, " (try 'perf record' first)"); - fprintf(stderr, "\n"); - exit(-1); + pr_err(" (try 'perf record' first)"); + pr_err("\n"); + return -errno; } - ret = fstat(input, &input_stat); - if (ret < 0) { - perror("failed to stat file"); - exit(-1); + if (fstat(input, &input_stat) < 0) { + pr_err("failed to stat file"); + err = -errno; + goto out_close; } + err = -EACCES; if (!force && input_stat.st_uid && (input_stat.st_uid != geteuid())) { - fprintf(stderr, "file: %s not owned by current user or root\n", + pr_err("file: %s not owned by current user or root\n", input_name); - exit(-1); + goto out_close; } - if (!input_stat.st_size) { - fprintf(stderr, "zero-sized file, nothing to do!\n"); - exit(0); + if (input_stat.st_size == 0) { + pr_info("zero-sized file, nothing to do!\n"); + goto done; } - *pheader = perf_header__read(input); - header = *pheader; + err = -ENOMEM; + header = perf_header__new(); + if (header == NULL) + goto out_close; + + err = perf_header__read(header, input); + if (err < 0) + goto out_delete; + *pheader = header; head = header->data_offset; sample_type = perf_header__sample_type(header); - if (curr_handler->sample_type_check) - if (curr_handler->sample_type_check(sample_type) < 0) - exit(-1); - - if (load_kernel(NULL) < 0) { - perror("failed to load kernel symbols"); - return EXIT_FAILURE; - } + err = -EINVAL; + if (curr_handler->sample_type_check && + curr_handler->sample_type_check(sample_type) < 0) + goto out_delete; if (!full_paths) { if (getcwd(__cwd, sizeof(__cwd)) == NULL) { - perror("failed to get the current directory"); - return EXIT_FAILURE; + pr_err("failed to get the current directory\n"); + err = -errno; + goto out_delete; } *cwd = __cwd; *cwdlen = strlen(*cwd); @@ -181,11 +218,12 @@ int mmap_dispatch_perf_file(struct perf_header **pheader, head -= shift; remap: - buf = (char *)mmap(NULL, page_size * mmap_window, PROT_READ, - MAP_SHARED, input, offset); + buf = mmap(NULL, page_size * mmap_window, PROT_READ, + MAP_SHARED, input, offset); if (buf == MAP_FAILED) { - perror("failed to mmap file"); - exit(-1); + pr_err("failed to mmap file\n"); + err = -errno; + goto out_delete; } more: @@ -242,10 +280,12 @@ more: goto more; done: - rc = EXIT_SUCCESS; + err = 0; +out_close: close(input); - return rc; + return err; +out_delete: + perf_header__delete(header); + goto out_close; } - - diff --git a/tools/perf/util/data_map.h b/tools/perf/util/data_map.h index ae036ecd762..3180ff7e363 100644 --- a/tools/perf/util/data_map.h +++ b/tools/perf/util/data_map.h @@ -4,7 +4,7 @@ #include "event.h" #include "header.h" -typedef int (*event_type_handler_t)(event_t *, unsigned long, unsigned long); +typedef int (*event_type_handler_t)(event_t *); struct perf_file_handler { event_type_handler_t process_sample_event; diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 1dae7e3b400..414b89d1bde 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -2,6 +2,7 @@ #include "event.h" #include "debug.h" #include "string.h" +#include "thread.h" static pid_t event__synthesize_comm(pid_t pid, int full, int (*process)(event_t *event)) @@ -175,3 +176,137 @@ void event__synthesize_threads(int (*process)(event_t *event)) closedir(proc); } + +char *event__cwd; +int event__cwdlen; + +struct events_stats event__stats; + +int event__process_comm(event_t *self) +{ + struct thread *thread = threads__findnew(self->comm.pid); + + dump_printf(": %s:%d\n", self->comm.comm, self->comm.pid); + + if (thread == NULL || thread__set_comm(thread, self->comm.comm)) { + dump_printf("problem processing PERF_RECORD_COMM, skipping event.\n"); + return -1; + } + + return 0; +} + +int event__process_lost(event_t *self) +{ + dump_printf(": id:%Ld: lost:%Ld\n", self->lost.id, self->lost.lost); + event__stats.lost += self->lost.lost; + return 0; +} + +int event__process_mmap(event_t *self) +{ + struct thread *thread = threads__findnew(self->mmap.pid); + struct map *map = map__new(&self->mmap, MAP__FUNCTION, + event__cwd, event__cwdlen); + + dump_printf(" %d/%d: [%p(%p) @ %p]: %s\n", + self->mmap.pid, self->mmap.tid, + (void *)(long)self->mmap.start, + (void *)(long)self->mmap.len, + (void *)(long)self->mmap.pgoff, + self->mmap.filename); + + if (thread == NULL || map == NULL) + dump_printf("problem processing PERF_RECORD_MMAP, skipping event.\n"); + else + thread__insert_map(thread, map); + + return 0; +} + +int event__process_task(event_t *self) +{ + struct thread *thread = threads__findnew(self->fork.pid); + struct thread *parent = threads__findnew(self->fork.ppid); + + dump_printf("(%d:%d):(%d:%d)\n", self->fork.pid, self->fork.tid, + self->fork.ppid, self->fork.ptid); + /* + * A thread clone will have the same PID for both parent and child. + */ + if (thread == parent) + return 0; + + if (self->header.type == PERF_RECORD_EXIT) + return 0; + + if (thread == NULL || parent == NULL || + thread__fork(thread, parent) < 0) { + dump_printf("problem processing PERF_RECORD_FORK, skipping event.\n"); + return -1; + } + + return 0; +} + +void thread__find_addr_location(struct thread *self, u8 cpumode, + enum map_type type, u64 addr, + struct addr_location *al, + symbol_filter_t filter) +{ + struct thread *thread = al->thread = self; + + al->addr = addr; + + if (cpumode & PERF_RECORD_MISC_KERNEL) { + al->level = 'k'; + thread = kthread; + } else if (cpumode & PERF_RECORD_MISC_USER) + al->level = '.'; + else { + al->level = 'H'; + al->map = NULL; + al->sym = NULL; + return; + } +try_again: + al->map = thread__find_map(thread, type, al->addr); + if (al->map == NULL) { + /* + * If this is outside of all known maps, and is a negative + * address, try to look it up in the kernel dso, as it might be + * a vsyscall or vdso (which executes in user-mode). + * + * XXX This is nasty, we should have a symbol list in the + * "[vdso]" dso, but for now lets use the old trick of looking + * in the whole kernel symbol list. + */ + if ((long long)al->addr < 0 && thread != kthread) { + thread = kthread; + goto try_again; + } + al->sym = NULL; + } else { + al->addr = al->map->map_ip(al->map, al->addr); + al->sym = map__find_symbol(al->map, al->addr, filter); + } +} + +int event__preprocess_sample(const event_t *self, struct addr_location *al, + symbol_filter_t filter) +{ + u8 cpumode = self->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; + struct thread *thread = threads__findnew(self->ip.pid); + + if (thread == NULL) + return -1; + + dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); + + thread__find_addr_location(thread, cpumode, MAP__FUNCTION, + self->ip.ip, al, filter); + dump_printf(" ...... dso: %s\n", + al->map ? al->map->dso->long_name : + al->level == 'H' ? "[hypervisor]" : "<not found>"); + return 0; +} diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 1f771ce3a95..a4cc8105cf6 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -69,13 +69,6 @@ struct build_id_event { char filename[]; }; -struct build_id_list { - struct build_id_event event; - struct list_head list; - const char *dso_name; - int len; -}; - typedef union event_union { struct perf_event_header header; struct ip_event ip; @@ -87,6 +80,19 @@ typedef union event_union { struct sample_event sample; } event_t; +struct events_stats { + unsigned long total; + unsigned long lost; +}; + +void event__print_totals(void); + +enum map_type { + MAP__FUNCTION = 0, + + MAP__NR_TYPES, +}; + struct map { union { struct rb_node rb_node; @@ -94,6 +100,7 @@ struct map { }; u64 start; u64 end; + enum map_type type; u64 pgoff; u64 (*map_ip)(struct map *, u64); u64 (*unmap_ip)(struct map *, u64); @@ -119,15 +126,34 @@ struct symbol; typedef int (*symbol_filter_t)(struct map *map, struct symbol *sym); -void map__init(struct map *self, u64 start, u64 end, u64 pgoff, - struct dso *dso); -struct map *map__new(struct mmap_event *event, char *cwd, int cwdlen); +void map__init(struct map *self, enum map_type type, + u64 start, u64 end, u64 pgoff, struct dso *dso); +struct map *map__new(struct mmap_event *event, enum map_type, + char *cwd, int cwdlen); +void map__delete(struct map *self); struct map *map__clone(struct map *self); int map__overlap(struct map *l, struct map *r); size_t map__fprintf(struct map *self, FILE *fp); -struct symbol *map__find_symbol(struct map *self, u64 ip, symbol_filter_t filter); +struct symbol *map__find_symbol(struct map *self, u64 addr, + symbol_filter_t filter); +void map__fixup_start(struct map *self); +void map__fixup_end(struct map *self); int event__synthesize_thread(pid_t pid, int (*process)(event_t *event)); void event__synthesize_threads(int (*process)(event_t *event)); +extern char *event__cwd; +extern int event__cwdlen; +extern struct events_stats event__stats; +extern unsigned long event__total[PERF_RECORD_MAX]; + +int event__process_comm(event_t *self); +int event__process_lost(event_t *self); +int event__process_mmap(event_t *self); +int event__process_task(event_t *self); + +struct addr_location; +int event__preprocess_sample(const event_t *self, struct addr_location *al, + symbol_filter_t filter); + #endif /* __PERF_RECORD_H */ diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index b01a9537977..4805e6dfd23 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -63,7 +63,7 @@ int perf_header_attr__add_id(struct perf_header_attr *self, u64 id) */ struct perf_header *perf_header__new(void) { - struct perf_header *self = calloc(sizeof(*self), 1); + struct perf_header *self = zalloc(sizeof(*self)); if (self != NULL) { self->size = 1; @@ -78,16 +78,24 @@ struct perf_header *perf_header__new(void) return self; } +void perf_header__delete(struct perf_header *self) +{ + int i; + + for (i = 0; i < self->attrs; ++i) + perf_header_attr__delete(self->attr[i]); + + free(self->attr); + free(self); +} + int perf_header__add_attr(struct perf_header *self, struct perf_header_attr *attr) { - int pos = self->attrs; - if (self->frozen) return -1; - self->attrs++; - if (self->attrs > self->size) { + if (self->attrs == self->size) { int nsize = self->size * 2; struct perf_header_attr **nattr; @@ -98,7 +106,8 @@ int perf_header__add_attr(struct perf_header *self, self->size = nsize; self->attr = nattr; } - self->attr[pos] = attr; + + self->attr[self->attrs++] = attr; return 0; } @@ -167,7 +176,7 @@ static int do_write(int fd, const void *buf, size_t size) int ret = write(fd, buf, size); if (ret < 0) - return -1; + return -errno; size -= ret; buf += ret; @@ -176,43 +185,59 @@ static int do_write(int fd, const void *buf, size_t size) return 0; } -static int write_buildid_table(int fd, struct list_head *id_head) +static int __dsos__write_buildid_table(struct list_head *head, int fd) { - struct build_id_list *iter, *next; - - list_for_each_entry_safe(iter, next, id_head, list) { - struct build_id_event *b = &iter->event; - - if (do_write(fd, b, sizeof(*b)) < 0 || - do_write(fd, iter->dso_name, iter->len) < 0) - return -1; - list_del(&iter->list); - free(iter); + struct dso *pos; + + list_for_each_entry(pos, head, node) { + int err; + struct build_id_event b; + size_t len; + + if (!pos->has_build_id) + continue; + len = pos->long_name_len + 1; + len = ALIGN(len, 64); + memset(&b, 0, sizeof(b)); + memcpy(&b.build_id, pos->build_id, sizeof(pos->build_id)); + b.header.size = sizeof(b) + len; + err = do_write(fd, &b, sizeof(b)); + if (err < 0) + return err; + err = do_write(fd, pos->long_name, len); + if (err < 0) + return err; } return 0; } -static void -perf_header__adds_write(struct perf_header *self, int fd) +static int dsos__write_buildid_table(int fd) +{ + int err = __dsos__write_buildid_table(&dsos__kernel, fd); + if (err == 0) + err = __dsos__write_buildid_table(&dsos__user, fd); + return err; +} + +static int perf_header__adds_write(struct perf_header *self, int fd) { - LIST_HEAD(id_list); int nr_sections; struct perf_file_section *feat_sec; int sec_size; u64 sec_start; - int idx = 0; + int idx = 0, err; - if (fetch_build_id_table(&id_list)) + if (dsos__read_build_ids()) perf_header__set_feat(self, HEADER_BUILD_ID); nr_sections = bitmap_weight(self->adds_features, HEADER_FEAT_BITS); if (!nr_sections) - return; + return 0; feat_sec = calloc(sizeof(*feat_sec), nr_sections); - if (!feat_sec) - die("No memory"); + if (feat_sec == NULL) + return -ENOMEM; sec_size = sizeof(*feat_sec) * nr_sections; @@ -238,23 +263,29 @@ perf_header__adds_write(struct perf_header *self, int fd) /* Write build-ids */ buildid_sec->offset = lseek(fd, 0, SEEK_CUR); - if (write_buildid_table(fd, &id_list) < 0) - die("failed to write buildid table"); + err = dsos__write_buildid_table(fd); + if (err < 0) { + pr_debug("failed to write buildid table\n"); + goto out_free; + } buildid_sec->size = lseek(fd, 0, SEEK_CUR) - buildid_sec->offset; } lseek(fd, sec_start, SEEK_SET); - if (do_write(fd, feat_sec, sec_size) < 0) - die("failed to write feature section"); + err = do_write(fd, feat_sec, sec_size); + if (err < 0) + pr_debug("failed to write feature section\n"); +out_free: free(feat_sec); + return err; } -void perf_header__write(struct perf_header *self, int fd, bool at_exit) +int perf_header__write(struct perf_header *self, int fd, bool at_exit) { struct perf_file_header f_header; struct perf_file_attr f_attr; struct perf_header_attr *attr; - int i; + int i, err; lseek(fd, sizeof(f_header), SEEK_SET); @@ -263,8 +294,11 @@ void perf_header__write(struct perf_header *self, int fd, bool at_exit) attr = self->attr[i]; attr->id_offset = lseek(fd, 0, SEEK_CUR); - if (do_write(fd, attr->id, attr->ids * sizeof(u64)) < 0) - die("failed to write perf header"); + err = do_write(fd, attr->id, attr->ids * sizeof(u64)); + if (err < 0) { + pr_debug("failed to write perf header\n"); + return err; + } } @@ -280,20 +314,30 @@ void perf_header__write(struct perf_header *self, int fd, bool at_exit) .size = attr->ids * sizeof(u64), } }; - if (do_write(fd, &f_attr, sizeof(f_attr)) < 0) - die("failed to write perf header attribute"); + err = do_write(fd, &f_attr, sizeof(f_attr)); + if (err < 0) { + pr_debug("failed to write perf header attribute\n"); + return err; + } } self->event_offset = lseek(fd, 0, SEEK_CUR); self->event_size = event_count * sizeof(struct perf_trace_event_type); - if (events) - if (do_write(fd, events, self->event_size) < 0) - die("failed to write perf header events"); + if (events) { + err = do_write(fd, events, self->event_size); + if (err < 0) { + pr_debug("failed to write perf header events\n"); + return err; + } + } self->data_offset = lseek(fd, 0, SEEK_CUR); - if (at_exit) - perf_header__adds_write(self, fd); + if (at_exit) { + err = perf_header__adds_write(self, fd); + if (err < 0) + return err; + } f_header = (struct perf_file_header){ .magic = PERF_MAGIC, @@ -316,11 +360,15 @@ void perf_header__write(struct perf_header *self, int fd, bool at_exit) memcpy(&f_header.adds_features, &self->adds_features, sizeof(self->adds_features)); lseek(fd, 0, SEEK_SET); - if (do_write(fd, &f_header, sizeof(f_header)) < 0) - die("failed to write perf header"); + err = do_write(fd, &f_header, sizeof(f_header)); + if (err < 0) { + pr_debug("failed to write perf header\n"); + return err; + } lseek(fd, self->data_offset + self->data_size, SEEK_SET); self->frozen = 1; + return 0; } static void do_read(int fd, void *buf, size_t size) @@ -430,19 +478,17 @@ static int perf_file_section__process(struct perf_file_section *self, return 0; } -struct perf_header *perf_header__read(int fd) +int perf_header__read(struct perf_header *self, int fd) { - struct perf_header *self = perf_header__new(); struct perf_file_header f_header; struct perf_file_attr f_attr; u64 f_id; int nr_attrs, nr_ids, i, j; - if (self == NULL) - die("nomem"); - - if (perf_file_header__read(&f_header, self, fd) < 0) - die("incompatible file format"); + if (perf_file_header__read(&f_header, self, fd) < 0) { + pr_debug("incompatible file format\n"); + return -EINVAL; + } nr_attrs = f_header.attrs.size / sizeof(f_attr); lseek(fd, f_header.attrs.offset, SEEK_SET); @@ -456,7 +502,7 @@ struct perf_header *perf_header__read(int fd) attr = perf_header_attr__new(&f_attr.attr); if (attr == NULL) - die("nomem"); + return -ENOMEM; nr_ids = f_attr.ids.size / sizeof(u64); lseek(fd, f_attr.ids.offset, SEEK_SET); @@ -464,11 +510,15 @@ struct perf_header *perf_header__read(int fd) for (j = 0; j < nr_ids; j++) { do_read(fd, &f_id, sizeof(f_id)); - if (perf_header_attr__add_id(attr, f_id) < 0) - die("nomem"); + if (perf_header_attr__add_id(attr, f_id) < 0) { + perf_header_attr__delete(attr); + return -ENOMEM; + } + } + if (perf_header__add_attr(self, attr) < 0) { + perf_header_attr__delete(attr); + return -ENOMEM; } - if (perf_header__add_attr(self, attr) < 0) - die("nomem"); lseek(fd, tmp, SEEK_SET); } @@ -476,8 +526,8 @@ struct perf_header *perf_header__read(int fd) if (f_header.event_types.size) { lseek(fd, f_header.event_types.offset, SEEK_SET); events = malloc(f_header.event_types.size); - if (!events) - die("nomem"); + if (events == NULL) + return -ENOMEM; do_read(fd, events, f_header.event_types.size); event_count = f_header.event_types.size / sizeof(struct perf_trace_event_type); } @@ -487,8 +537,7 @@ struct perf_header *perf_header__read(int fd) lseek(fd, self->data_offset, SEEK_SET); self->frozen = 1; - - return self; + return 0; } u64 perf_header__sample_type(struct perf_header *header) diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index f46a94e09ee..d1dbe2b79c4 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -55,8 +55,11 @@ struct perf_header { DECLARE_BITMAP(adds_features, HEADER_FEAT_BITS); }; -struct perf_header *perf_header__read(int fd); -void perf_header__write(struct perf_header *self, int fd, bool at_exit); +struct perf_header *perf_header__new(void); +void perf_header__delete(struct perf_header *self); + +int perf_header__read(struct perf_header *self, int fd); +int perf_header__write(struct perf_header *self, int fd, bool at_exit); int perf_header__add_attr(struct perf_header *self, struct perf_header_attr *attr); @@ -75,8 +78,6 @@ perf_header__find_attr(u64 id, struct perf_header *header); void perf_header__set_feat(struct perf_header *self, int feat); bool perf_header__has_feat(const struct perf_header *self, int feat); -struct perf_header *perf_header__new(void); - int perf_header__process_sections(struct perf_header *self, int fd, int (*process)(struct perf_file_section *self, int feat, int fd)); diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 7393a02fd8d..0ebf6ee16ca 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -10,31 +10,23 @@ struct callchain_param callchain_param = { .min_percent = 0.5 }; -unsigned long total; -unsigned long total_mmap; -unsigned long total_comm; -unsigned long total_fork; -unsigned long total_unknown; -unsigned long total_lost; - /* * histogram, sorted on item, collects counts */ -struct hist_entry *__hist_entry__add(struct thread *thread, struct map *map, - struct symbol *sym, +struct hist_entry *__hist_entry__add(struct addr_location *al, struct symbol *sym_parent, - u64 ip, u64 count, char level, bool *hit) + u64 count, bool *hit) { struct rb_node **p = &hist.rb_node; struct rb_node *parent = NULL; struct hist_entry *he; struct hist_entry entry = { - .thread = thread, - .map = map, - .sym = sym, - .ip = ip, - .level = level, + .thread = al->thread, + .map = al->map, + .sym = al->sym, + .ip = al->addr, + .level = al->level, .count = count, .parent = sym_parent, }; diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index ac2149c559b..3020db0c929 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -36,9 +36,9 @@ extern unsigned long total_fork; extern unsigned long total_unknown; extern unsigned long total_lost; -struct hist_entry *__hist_entry__add(struct thread *thread, struct map *map, - struct symbol *sym, struct symbol *parent, - u64 ip, u64 count, char level, bool *hit); +struct hist_entry *__hist_entry__add(struct addr_location *al, + struct symbol *parent, + u64 count, bool *hit); extern int64_t hist_entry__cmp(struct hist_entry *, struct hist_entry *); extern int64_t hist_entry__collapse(struct hist_entry *, struct hist_entry *); extern void hist_entry__free(struct hist_entry *); diff --git a/tools/perf/util/include/asm/bug.h b/tools/perf/util/include/asm/bug.h new file mode 100644 index 00000000000..7fcc6810adc --- /dev/null +++ b/tools/perf/util/include/asm/bug.h @@ -0,0 +1,22 @@ +#ifndef _PERF_ASM_GENERIC_BUG_H +#define _PERF_ASM_GENERIC_BUG_H + +#define __WARN_printf(arg...) do { fprintf(stderr, arg); } while (0) + +#define WARN(condition, format...) ({ \ + int __ret_warn_on = !!(condition); \ + if (unlikely(__ret_warn_on)) \ + __WARN_printf(format); \ + unlikely(__ret_warn_on); \ +}) + +#define WARN_ONCE(condition, format...) ({ \ + static int __warned; \ + int __ret_warn_once = !!(condition); \ + \ + if (unlikely(__ret_warn_once)) \ + if (WARN(!__warned, format)) \ + __warned = 1; \ + unlikely(__ret_warn_once); \ +}) +#endif diff --git a/tools/perf/util/include/linux/bitops.h b/tools/perf/util/include/linux/bitops.h index ace57c36d1d..8d63116e943 100644 --- a/tools/perf/util/include/linux/bitops.h +++ b/tools/perf/util/include/linux/bitops.h @@ -7,6 +7,8 @@ #define CONFIG_GENERIC_FIND_FIRST_BIT #include "../../../../include/linux/bitops.h" +#undef __KERNEL__ + static inline void set_bit(int nr, unsigned long *addr) { addr[nr / BITS_PER_LONG] |= 1UL << (nr % BITS_PER_LONG); diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 94ca95073c4..69f94fe9db2 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -20,9 +20,10 @@ static int strcommon(const char *pathname, char *cwd, int cwdlen) return n; } -void map__init(struct map *self, u64 start, u64 end, u64 pgoff, - struct dso *dso) +void map__init(struct map *self, enum map_type type, + u64 start, u64 end, u64 pgoff, struct dso *dso) { + self->type = type; self->start = start; self->end = end; self->pgoff = pgoff; @@ -32,7 +33,8 @@ void map__init(struct map *self, u64 start, u64 end, u64 pgoff, RB_CLEAR_NODE(&self->rb_node); } -struct map *map__new(struct mmap_event *event, char *cwd, int cwdlen) +struct map *map__new(struct mmap_event *event, enum map_type type, + char *cwd, int cwdlen) { struct map *self = malloc(sizeof(*self)); @@ -63,7 +65,7 @@ struct map *map__new(struct mmap_event *event, char *cwd, int cwdlen) if (dso == NULL) goto out_delete; - map__init(self, event->start, event->start + event->len, + map__init(self, type, event->start, event->start + event->len, event->pgoff, dso); if (self->dso == vdso || anon) @@ -75,12 +77,37 @@ out_delete: return NULL; } +void map__delete(struct map *self) +{ + free(self); +} + +void map__fixup_start(struct map *self) +{ + struct rb_root *symbols = &self->dso->symbols[self->type]; + struct rb_node *nd = rb_first(symbols); + if (nd != NULL) { + struct symbol *sym = rb_entry(nd, struct symbol, rb_node); + self->start = sym->start; + } +} + +void map__fixup_end(struct map *self) +{ + struct rb_root *symbols = &self->dso->symbols[self->type]; + struct rb_node *nd = rb_last(symbols); + if (nd != NULL) { + struct symbol *sym = rb_entry(nd, struct symbol, rb_node); + self->end = sym->end; + } +} + #define DSO__DELETED "(deleted)" -struct symbol * -map__find_symbol(struct map *self, u64 ip, symbol_filter_t filter) +struct symbol *map__find_symbol(struct map *self, u64 addr, + symbol_filter_t filter) { - if (!self->dso->loaded) { + if (!dso__loaded(self->dso, self->type)) { int nr = dso__load(self->dso, self, filter); if (nr < 0) { @@ -113,7 +140,7 @@ map__find_symbol(struct map *self, u64 ip, symbol_filter_t filter) } } - return self->dso->find_symbol(self->dso, ip); + return self->dso->find_symbol(self->dso, self->type, addr); } struct map *map__clone(struct map *self) diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 0faf4f2bb5c..9e5dbd66d34 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -1,4 +1,4 @@ - +#include "../../../include/linux/hw_breakpoint.h" #include "util.h" #include "../perf.h" #include "parse-options.h" @@ -197,7 +197,7 @@ struct tracepoint_path *tracepoint_id_to_path(u64 config) if (id == config) { closedir(evt_dir); closedir(sys_dir); - path = calloc(1, sizeof(path)); + path = zalloc(sizeof(path)); path->system = malloc(MAX_EVENT_LENGTH); if (!path->system) { free(path); @@ -540,6 +540,81 @@ static enum event_result parse_tracepoint_event(const char **strp, attr, strp); } +static enum event_result +parse_breakpoint_type(const char *type, const char **strp, + struct perf_event_attr *attr) +{ + int i; + + for (i = 0; i < 3; i++) { + if (!type[i]) + break; + + switch (type[i]) { + case 'r': + attr->bp_type |= HW_BREAKPOINT_R; + break; + case 'w': + attr->bp_type |= HW_BREAKPOINT_W; + break; + case 'x': + attr->bp_type |= HW_BREAKPOINT_X; + break; + default: + return EVT_FAILED; + } + } + if (!attr->bp_type) /* Default */ + attr->bp_type = HW_BREAKPOINT_R | HW_BREAKPOINT_W; + + *strp = type + i; + + return EVT_HANDLED; +} + +static enum event_result +parse_breakpoint_event(const char **strp, struct perf_event_attr *attr) +{ + const char *target; + const char *type; + char *endaddr; + u64 addr; + enum event_result err; + + target = strchr(*strp, ':'); + if (!target) + return EVT_FAILED; + + if (strncmp(*strp, "mem", target - *strp) != 0) + return EVT_FAILED; + + target++; + + addr = strtoull(target, &endaddr, 0); + if (target == endaddr) + return EVT_FAILED; + + attr->bp_addr = addr; + *strp = endaddr; + + type = strchr(target, ':'); + + /* If no type is defined, just rw as default */ + if (!type) { + attr->bp_type = HW_BREAKPOINT_R | HW_BREAKPOINT_W; + } else { + err = parse_breakpoint_type(++type, strp, attr); + if (err == EVT_FAILED) + return EVT_FAILED; + } + + /* We should find a nice way to override the access type */ + attr->bp_len = HW_BREAKPOINT_LEN_4; + attr->type = PERF_TYPE_BREAKPOINT; + + return EVT_HANDLED; +} + static int check_events(const char *str, unsigned int i) { int n; @@ -673,6 +748,10 @@ parse_event_symbols(const char **str, struct perf_event_attr *attr) if (ret != EVT_FAILED) goto modifier; + ret = parse_breakpoint_event(str, attr); + if (ret != EVT_FAILED) + goto modifier; + fprintf(stderr, "invalid or unsupported event: '%s'\n", *str); fprintf(stderr, "Run 'perf list' for a list of valid events\n"); return EVT_FAILED; @@ -859,6 +938,9 @@ void print_events(void) "rNNN"); printf("\n"); + printf(" %-42s [hardware breakpoint]\n", "mem:<addr>[:access]"); + printf("\n"); + print_tracepoint_events(); exit(129); diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c new file mode 100644 index 00000000000..cd7fbda5e2a --- /dev/null +++ b/tools/perf/util/probe-event.c @@ -0,0 +1,484 @@ +/* + * probe-event.c : perf-probe definition to kprobe_events format converter + * + * Written by Masami Hiramatsu <mhiramat@redhat.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., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + */ + +#define _GNU_SOURCE +#include <sys/utsname.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <fcntl.h> +#include <errno.h> +#include <stdio.h> +#include <unistd.h> +#include <stdlib.h> +#include <string.h> +#include <stdarg.h> +#include <limits.h> + +#undef _GNU_SOURCE +#include "event.h" +#include "string.h" +#include "strlist.h" +#include "debug.h" +#include "parse-events.h" /* For debugfs_path */ +#include "probe-event.h" + +#define MAX_CMDLEN 256 +#define MAX_PROBE_ARGS 128 +#define PERFPROBE_GROUP "probe" + +#define semantic_error(msg ...) die("Semantic error :" msg) + +/* If there is no space to write, returns -E2BIG. */ +static int e_snprintf(char *str, size_t size, const char *format, ...) +{ + int ret; + va_list ap; + va_start(ap, format); + ret = vsnprintf(str, size, format, ap); + va_end(ap); + if (ret >= (int)size) + ret = -E2BIG; + return ret; +} + +/* Parse probepoint definition. */ +static void parse_perf_probe_probepoint(char *arg, struct probe_point *pp) +{ + char *ptr, *tmp; + char c, nc = 0; + /* + * <Syntax> + * perf probe SRC:LN + * perf probe FUNC[+OFFS|%return][@SRC] + */ + + ptr = strpbrk(arg, ":+@%"); + if (ptr) { + nc = *ptr; + *ptr++ = '\0'; + } + + /* Check arg is function or file and copy it */ + if (strchr(arg, '.')) /* File */ + pp->file = strdup(arg); + else /* Function */ + pp->function = strdup(arg); + DIE_IF(pp->file == NULL && pp->function == NULL); + + /* Parse other options */ + while (ptr) { + arg = ptr; + c = nc; + ptr = strpbrk(arg, ":+@%"); + if (ptr) { + nc = *ptr; + *ptr++ = '\0'; + } + switch (c) { + case ':': /* Line number */ + pp->line = strtoul(arg, &tmp, 0); + if (*tmp != '\0') + semantic_error("There is non-digit charactor" + " in line number."); + break; + case '+': /* Byte offset from a symbol */ + pp->offset = strtoul(arg, &tmp, 0); + if (*tmp != '\0') + semantic_error("There is non-digit charactor" + " in offset."); + break; + case '@': /* File name */ + if (pp->file) + semantic_error("SRC@SRC is not allowed."); + pp->file = strdup(arg); + DIE_IF(pp->file == NULL); + if (ptr) + semantic_error("@SRC must be the last " + "option."); + break; + case '%': /* Probe places */ + if (strcmp(arg, "return") == 0) { + pp->retprobe = 1; + } else /* Others not supported yet */ + semantic_error("%%%s is not supported.", arg); + break; + default: + DIE_IF("Program has a bug."); + break; + } + } + + /* Exclusion check */ + if (pp->line && pp->offset) + semantic_error("Offset can't be used with line number."); + + if (!pp->line && pp->file && !pp->function) + semantic_error("File always requires line number."); + + if (pp->offset && !pp->function) + semantic_error("Offset requires an entry function."); + + if (pp->retprobe && !pp->function) + semantic_error("Return probe requires an entry function."); + + if ((pp->offset || pp->line) && pp->retprobe) + semantic_error("Offset/Line can't be used with return probe."); + + pr_debug("symbol:%s file:%s line:%d offset:%d, return:%d\n", + pp->function, pp->file, pp->line, pp->offset, pp->retprobe); +} + +/* Parse perf-probe event definition */ +int parse_perf_probe_event(const char *str, struct probe_point *pp) +{ + char **argv; + int argc, i, need_dwarf = 0; + + argv = argv_split(str, &argc); + if (!argv) + die("argv_split failed."); + if (argc > MAX_PROBE_ARGS + 1) + semantic_error("Too many arguments"); + + /* Parse probe point */ + parse_perf_probe_probepoint(argv[0], pp); + if (pp->file || pp->line) + need_dwarf = 1; + + /* Copy arguments and ensure return probe has no C argument */ + pp->nr_args = argc - 1; + pp->args = zalloc(sizeof(char *) * pp->nr_args); + for (i = 0; i < pp->nr_args; i++) { + pp->args[i] = strdup(argv[i + 1]); + if (!pp->args[i]) + die("Failed to copy argument."); + if (is_c_varname(pp->args[i])) { + if (pp->retprobe) + semantic_error("You can't specify local" + " variable for kretprobe"); + need_dwarf = 1; + } + } + + argv_free(argv); + return need_dwarf; +} + +/* Parse kprobe_events event into struct probe_point */ +void parse_trace_kprobe_event(const char *str, char **group, char **event, + struct probe_point *pp) +{ + char pr; + char *p; + int ret, i, argc; + char **argv; + + pr_debug("Parsing kprobe_events: %s\n", str); + argv = argv_split(str, &argc); + if (!argv) + die("argv_split failed."); + if (argc < 2) + semantic_error("Too less arguments."); + + /* Scan event and group name. */ + ret = sscanf(argv[0], "%c:%a[^/ \t]/%a[^ \t]", + &pr, (float *)(void *)group, (float *)(void *)event); + if (ret != 3) + semantic_error("Failed to parse event name: %s", argv[0]); + pr_debug("Group:%s Event:%s probe:%c\n", *group, *event, pr); + + if (!pp) + goto end; + + pp->retprobe = (pr == 'r'); + + /* Scan function name and offset */ + ret = sscanf(argv[1], "%a[^+]+%d", (float *)(void *)&pp->function, &pp->offset); + if (ret == 1) + pp->offset = 0; + + /* kprobe_events doesn't have this information */ + pp->line = 0; + pp->file = NULL; + + pp->nr_args = argc - 2; + pp->args = zalloc(sizeof(char *) * pp->nr_args); + for (i = 0; i < pp->nr_args; i++) { + p = strchr(argv[i + 2], '='); + if (p) /* We don't need which register is assigned. */ + *p = '\0'; + pp->args[i] = strdup(argv[i + 2]); + if (!pp->args[i]) + die("Failed to copy argument."); + } + +end: + argv_free(argv); +} + +int synthesize_perf_probe_event(struct probe_point *pp) +{ + char *buf; + char offs[64] = "", line[64] = ""; + int i, len, ret; + + pp->probes[0] = buf = zalloc(MAX_CMDLEN); + if (!buf) + die("Failed to allocate memory by zalloc."); + if (pp->offset) { + ret = e_snprintf(offs, 64, "+%d", pp->offset); + if (ret <= 0) + goto error; + } + if (pp->line) { + ret = e_snprintf(line, 64, ":%d", pp->line); + if (ret <= 0) + goto error; + } + + if (pp->function) + ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s", pp->function, + offs, pp->retprobe ? "%return" : "", line); + else + ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s", pp->file, line); + if (ret <= 0) + goto error; + len = ret; + + for (i = 0; i < pp->nr_args; i++) { + ret = e_snprintf(&buf[len], MAX_CMDLEN - len, " %s", + pp->args[i]); + if (ret <= 0) + goto error; + len += ret; + } + pp->found = 1; + + return pp->found; +error: + free(pp->probes[0]); + + return ret; +} + +int synthesize_trace_kprobe_event(struct probe_point *pp) +{ + char *buf; + int i, len, ret; + + pp->probes[0] = buf = zalloc(MAX_CMDLEN); + if (!buf) + die("Failed to allocate memory by zalloc."); + ret = e_snprintf(buf, MAX_CMDLEN, "%s+%d", pp->function, pp->offset); + if (ret <= 0) + goto error; + len = ret; + + for (i = 0; i < pp->nr_args; i++) { + ret = e_snprintf(&buf[len], MAX_CMDLEN - len, " %s", + pp->args[i]); + if (ret <= 0) + goto error; + len += ret; + } + pp->found = 1; + + return pp->found; +error: + free(pp->probes[0]); + + return ret; +} + +static int open_kprobe_events(int flags, int mode) +{ + char buf[PATH_MAX]; + int ret; + + ret = e_snprintf(buf, PATH_MAX, "%s/../kprobe_events", debugfs_path); + if (ret < 0) + die("Failed to make kprobe_events path."); + + ret = open(buf, flags, mode); + if (ret < 0) { + if (errno == ENOENT) + die("kprobe_events file does not exist -" + " please rebuild with CONFIG_KPROBE_TRACER."); + else + die("Could not open kprobe_events file: %s", + strerror(errno)); + } + return ret; +} + +/* Get raw string list of current kprobe_events */ +static struct strlist *get_trace_kprobe_event_rawlist(int fd) +{ + int ret, idx; + FILE *fp; + char buf[MAX_CMDLEN]; + char *p; + struct strlist *sl; + + sl = strlist__new(true, NULL); + + fp = fdopen(dup(fd), "r"); + while (!feof(fp)) { + p = fgets(buf, MAX_CMDLEN, fp); + if (!p) + break; + + idx = strlen(p) - 1; + if (p[idx] == '\n') + p[idx] = '\0'; + ret = strlist__add(sl, buf); + if (ret < 0) + die("strlist__add failed: %s", strerror(-ret)); + } + fclose(fp); + + return sl; +} + +/* Free and zero clear probe_point */ +static void clear_probe_point(struct probe_point *pp) +{ + int i; + + if (pp->function) + free(pp->function); + if (pp->file) + free(pp->file); + for (i = 0; i < pp->nr_args; i++) + free(pp->args[i]); + if (pp->args) + free(pp->args); + for (i = 0; i < pp->found; i++) + free(pp->probes[i]); + memset(pp, 0, sizeof(pp)); +} + +/* List up current perf-probe events */ +void show_perf_probe_events(void) +{ + unsigned int i; + int fd; + char *group, *event; + struct probe_point pp; + struct strlist *rawlist; + struct str_node *ent; + + fd = open_kprobe_events(O_RDONLY, 0); + rawlist = get_trace_kprobe_event_rawlist(fd); + close(fd); + + for (i = 0; i < strlist__nr_entries(rawlist); i++) { + ent = strlist__entry(rawlist, i); + parse_trace_kprobe_event(ent->s, &group, &event, &pp); + synthesize_perf_probe_event(&pp); + printf("[%s:%s]\t%s\n", group, event, pp.probes[0]); + free(group); + free(event); + clear_probe_point(&pp); + } + + strlist__delete(rawlist); +} + +/* Get current perf-probe event names */ +static struct strlist *get_perf_event_names(int fd) +{ + unsigned int i; + char *group, *event; + struct strlist *sl, *rawlist; + struct str_node *ent; + + rawlist = get_trace_kprobe_event_rawlist(fd); + + sl = strlist__new(false, NULL); + for (i = 0; i < strlist__nr_entries(rawlist); i++) { + ent = strlist__entry(rawlist, i); + parse_trace_kprobe_event(ent->s, &group, &event, NULL); + strlist__add(sl, event); + free(group); + } + + strlist__delete(rawlist); + + return sl; +} + +static int write_trace_kprobe_event(int fd, const char *buf) +{ + int ret; + + ret = write(fd, buf, strlen(buf)); + if (ret <= 0) + die("Failed to create event."); + else + printf("Added new event: %s\n", buf); + + return ret; +} + +static void get_new_event_name(char *buf, size_t len, const char *base, + struct strlist *namelist) +{ + int i, ret; + for (i = 0; i < MAX_EVENT_INDEX; i++) { + ret = e_snprintf(buf, len, "%s_%d", base, i); + if (ret < 0) + die("snprintf() failed: %s", strerror(-ret)); + if (!strlist__has_entry(namelist, buf)) + break; + } + if (i == MAX_EVENT_INDEX) + die("Too many events are on the same function."); +} + +void add_trace_kprobe_events(struct probe_point *probes, int nr_probes) +{ + int i, j, fd; + struct probe_point *pp; + char buf[MAX_CMDLEN]; + char event[64]; + struct strlist *namelist; + + fd = open_kprobe_events(O_RDWR, O_APPEND); + /* Get current event names */ + namelist = get_perf_event_names(fd); + + for (j = 0; j < nr_probes; j++) { + pp = probes + j; + for (i = 0; i < pp->found; i++) { + /* Get an unused new event name */ + get_new_event_name(event, 64, pp->function, namelist); + snprintf(buf, MAX_CMDLEN, "%c:%s/%s %s\n", + pp->retprobe ? 'r' : 'p', + PERFPROBE_GROUP, event, + pp->probes[i]); + write_trace_kprobe_event(fd, buf); + /* Add added event name to namelist */ + strlist__add(namelist, event); + } + } + close(fd); +} diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h new file mode 100644 index 00000000000..0c6fe56fe38 --- /dev/null +++ b/tools/perf/util/probe-event.h @@ -0,0 +1,18 @@ +#ifndef _PROBE_EVENT_H +#define _PROBE_EVENT_H + +#include "probe-finder.h" +#include "strlist.h" + +extern int parse_perf_probe_event(const char *str, struct probe_point *pp); +extern int synthesize_perf_probe_event(struct probe_point *pp); +extern void parse_trace_kprobe_event(const char *str, char **group, + char **event, struct probe_point *pp); +extern int synthesize_trace_kprobe_event(struct probe_point *pp); +extern void add_trace_kprobe_events(struct probe_point *probes, int nr_probes); +extern void show_perf_probe_events(void); + +/* Maximum index number of event-name postfix */ +#define MAX_EVENT_INDEX 1024 + +#endif /*_PROBE_EVENT_H */ diff --git a/tools/perf/util/string.c b/tools/perf/util/string.c index 227043577e0..f24a8cc933d 100644 --- a/tools/perf/util/string.c +++ b/tools/perf/util/string.c @@ -1,5 +1,3 @@ -#include <string.h> -#include <stdlib.h> #include "string.h" #include "util.h" @@ -127,3 +125,104 @@ out_err: out: return length; } + +/* + * Helper function for splitting a string into an argv-like array. + * originaly copied from lib/argv_split.c + */ +static const char *skip_sep(const char *cp) +{ + while (*cp && isspace(*cp)) + cp++; + + return cp; +} + +static const char *skip_arg(const char *cp) +{ + while (*cp && !isspace(*cp)) + cp++; + + return cp; +} + +static int count_argc(const char *str) +{ + int count = 0; + + while (*str) { + str = skip_sep(str); + if (*str) { + count++; + str = skip_arg(str); + } + } + + return count; +} + +/** + * argv_free - free an argv + * @argv - the argument vector to be freed + * + * Frees an argv and the strings it points to. + */ +void argv_free(char **argv) +{ + char **p; + for (p = argv; *p; p++) + free(*p); + + free(argv); +} + +/** + * argv_split - split a string at whitespace, returning an argv + * @str: the string to be split + * @argcp: returned argument count + * + * Returns an array of pointers to strings which are split out from + * @str. This is performed by strictly splitting on white-space; no + * quote processing is performed. Multiple whitespace characters are + * considered to be a single argument separator. The returned array + * is always NULL-terminated. Returns NULL on memory allocation + * failure. + */ +char **argv_split(const char *str, int *argcp) +{ + int argc = count_argc(str); + char **argv = zalloc(sizeof(*argv) * (argc+1)); + char **argvp; + + if (argv == NULL) + goto out; + + if (argcp) + *argcp = argc; + + argvp = argv; + + while (*str) { + str = skip_sep(str); + + if (*str) { + const char *p = str; + char *t; + + str = skip_arg(str); + + t = strndup(p, str-p); + if (t == NULL) + goto fail; + *argvp++ = t; + } + } + *argvp = NULL; + +out: + return argv; + +fail: + argv_free(argv); + return NULL; +} diff --git a/tools/perf/util/string.h b/tools/perf/util/string.h index e50b07f8082..bfecec265a1 100644 --- a/tools/perf/util/string.h +++ b/tools/perf/util/string.h @@ -6,6 +6,8 @@ int hex2u64(const char *ptr, u64 *val); char *strxfrchar(char *s, char from, char to); s64 perf_atoll(const char *str); +char **argv_split(const char *str, int *argcp); +void argv_free(char **argv); #define _STR(x) #x #define STR(x) _STR(x) diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 1b77e81b38d..fffcb937cdc 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -6,11 +6,17 @@ #include "debug.h" +#include <asm/bug.h> #include <libelf.h> #include <gelf.h> #include <elf.h> +#include <limits.h> #include <sys/utsname.h> +#ifndef NT_GNU_BUILD_ID +#define NT_GNU_BUILD_ID 3 +#endif + enum dso_origin { DSO__ORIG_KERNEL = 0, DSO__ORIG_JAVA_JIT, @@ -22,17 +28,37 @@ enum dso_origin { DSO__ORIG_NOT_FOUND, }; -static void dsos__add(struct dso *dso); -static struct dso *dsos__find(const char *name); -static struct map *map__new2(u64 start, struct dso *dso); -static void kernel_maps__insert(struct map *map); +static void dsos__add(struct list_head *head, struct dso *dso); +static struct map *thread__find_map_by_name(struct thread *self, char *name); +static struct map *map__new2(u64 start, struct dso *dso, enum map_type type); +struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr); +static int dso__load_kernel_sym(struct dso *self, struct map *map, + struct thread *thread, symbol_filter_t filter); unsigned int symbol__priv_size; +static int vmlinux_path__nr_entries; +static char **vmlinux_path; + +static struct symbol_conf symbol_conf__defaults = { + .use_modules = true, + .try_vmlinux_path = true, +}; + +static struct thread kthread_mem; +struct thread *kthread = &kthread_mem; + +bool dso__loaded(const struct dso *self, enum map_type type) +{ + return self->loaded & (1 << type); +} -static struct rb_root kernel_maps; +static void dso__set_loaded(struct dso *self, enum map_type type) +{ + self->loaded |= (1 << type); +} -static void dso__fixup_sym_end(struct dso *self) +static void symbols__fixup_end(struct rb_root *self) { - struct rb_node *nd, *prevnd = rb_first(&self->syms); + struct rb_node *nd, *prevnd = rb_first(self); struct symbol *curr, *prev; if (prevnd == NULL) @@ -53,10 +79,10 @@ static void dso__fixup_sym_end(struct dso *self) curr->end = roundup(curr->start, 4096); } -static void kernel_maps__fixup_end(void) +static void __thread__fixup_maps_end(struct thread *self, enum map_type type) { struct map *prev, *curr; - struct rb_node *nd, *prevnd = rb_first(&kernel_maps); + struct rb_node *nd, *prevnd = rb_first(&self->maps[type]); if (prevnd == NULL) return; @@ -69,25 +95,31 @@ static void kernel_maps__fixup_end(void) prev->end = curr->start - 1; } - nd = rb_last(&curr->dso->syms); - if (nd) { - struct symbol *sym = rb_entry(nd, struct symbol, rb_node); - curr->end = sym->end; - } + /* + * We still haven't the actual symbols, so guess the + * last map final address. + */ + curr->end = ~0UL; +} + +static void thread__fixup_maps_end(struct thread *self) +{ + int i; + for (i = 0; i < MAP__NR_TYPES; ++i) + __thread__fixup_maps_end(self, i); } static struct symbol *symbol__new(u64 start, u64 len, const char *name) { size_t namelen = strlen(name) + 1; - struct symbol *self = calloc(1, (symbol__priv_size + - sizeof(*self) + namelen)); - if (!self) + struct symbol *self = zalloc(symbol__priv_size + + sizeof(*self) + namelen); + if (self == NULL) return NULL; - if (symbol__priv_size) { - memset(self, 0, symbol__priv_size); + if (symbol__priv_size) self = ((void *)self) + symbol__priv_size; - } + self->start = start; self->end = len ? start + len - 1 : start; @@ -109,15 +141,30 @@ static size_t symbol__fprintf(struct symbol *self, FILE *fp) self->start, self->end, self->name); } +static void dso__set_long_name(struct dso *self, char *name) +{ + if (name == NULL) + return; + self->long_name = name; + self->long_name_len = strlen(name); +} + +static void dso__set_basename(struct dso *self) +{ + self->short_name = basename(self->long_name); +} + struct dso *dso__new(const char *name) { struct dso *self = malloc(sizeof(*self) + strlen(name) + 1); if (self != NULL) { + int i; strcpy(self->name, name); - self->long_name = self->name; + dso__set_long_name(self, self->name); self->short_name = self->name; - self->syms = RB_ROOT; + for (i = 0; i < MAP__NR_TYPES; ++i) + self->symbols[i] = RB_ROOT; self->find_symbol = dso__find_symbol; self->slen_calculated = 0; self->origin = DSO__ORIG_NOT_FOUND; @@ -128,22 +175,24 @@ struct dso *dso__new(const char *name) return self; } -static void dso__delete_symbols(struct dso *self) +static void symbols__delete(struct rb_root *self) { struct symbol *pos; - struct rb_node *next = rb_first(&self->syms); + struct rb_node *next = rb_first(self); while (next) { pos = rb_entry(next, struct symbol, rb_node); next = rb_next(&pos->rb_node); - rb_erase(&pos->rb_node, &self->syms); + rb_erase(&pos->rb_node, self); symbol__delete(pos); } } void dso__delete(struct dso *self) { - dso__delete_symbols(self); + int i; + for (i = 0; i < MAP__NR_TYPES; ++i) + symbols__delete(&self->symbols[i]); if (self->long_name != self->name) free(self->long_name); free(self); @@ -155,9 +204,9 @@ void dso__set_build_id(struct dso *self, void *build_id) self->has_build_id = 1; } -static void dso__insert_symbol(struct dso *self, struct symbol *sym) +static void symbols__insert(struct rb_root *self, struct symbol *sym) { - struct rb_node **p = &self->syms.rb_node; + struct rb_node **p = &self->rb_node; struct rb_node *parent = NULL; const u64 ip = sym->start; struct symbol *s; @@ -171,17 +220,17 @@ static void dso__insert_symbol(struct dso *self, struct symbol *sym) p = &(*p)->rb_right; } rb_link_node(&sym->rb_node, parent, p); - rb_insert_color(&sym->rb_node, &self->syms); + rb_insert_color(&sym->rb_node, self); } -struct symbol *dso__find_symbol(struct dso *self, u64 ip) +static struct symbol *symbols__find(struct rb_root *self, u64 ip) { struct rb_node *n; if (self == NULL) return NULL; - n = self->syms.rb_node; + n = self->rb_node; while (n) { struct symbol *s = rb_entry(n, struct symbol, rb_node); @@ -197,6 +246,11 @@ struct symbol *dso__find_symbol(struct dso *self, u64 ip) return NULL; } +struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr) +{ + return symbols__find(&self->symbols[type], addr); +} + int build_id__sprintf(u8 *self, int len, char *bf) { char *bid = bf; @@ -220,15 +274,14 @@ size_t dso__fprintf_buildid(struct dso *self, FILE *fp) return fprintf(fp, "%s", sbuild_id); } -size_t dso__fprintf(struct dso *self, FILE *fp) +size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp) { struct rb_node *nd; size_t ret = fprintf(fp, "dso: %s (", self->short_name); ret += dso__fprintf_buildid(self, fp); ret += fprintf(fp, ")\n"); - - for (nd = rb_first(&self->syms); nd; nd = rb_next(nd)) { + for (nd = rb_first(&self->symbols[type]); nd; nd = rb_next(nd)) { struct symbol *pos = rb_entry(nd, struct symbol, rb_node); ret += symbol__fprintf(pos, fp); } @@ -241,10 +294,11 @@ size_t dso__fprintf(struct dso *self, FILE *fp) * so that we can in the next step set the symbol ->end address and then * call kernel_maps__split_kallsyms. */ -static int kernel_maps__load_all_kallsyms(void) +static int dso__load_all_kallsyms(struct dso *self, struct map *map) { char *line = NULL; size_t n; + struct rb_root *root = &self->symbols[map->type]; FILE *file = fopen("/proc/kallsyms", "r"); if (file == NULL) @@ -287,13 +341,11 @@ static int kernel_maps__load_all_kallsyms(void) if (sym == NULL) goto out_delete_line; - /* * We will pass the symbols to the filter later, in - * kernel_maps__split_kallsyms, when we have split the - * maps per module + * map__split_kallsyms, when we have split the maps per module */ - dso__insert_symbol(kernel_map->dso, sym); + symbols__insert(root, sym); } free(line); @@ -312,12 +364,14 @@ out_failure: * kernel range is broken in several maps, named [kernel].N, as we don't have * the original ELF section names vmlinux have. */ -static int kernel_maps__split_kallsyms(symbol_filter_t filter, int use_modules) +static int dso__split_kallsyms(struct dso *self, struct map *map, struct thread *thread, + symbol_filter_t filter) { - struct map *map = kernel_map; + struct map *curr_map = map; struct symbol *pos; int count = 0; - struct rb_node *next = rb_first(&kernel_map->dso->syms); + struct rb_root *root = &self->symbols[map->type]; + struct rb_node *next = rb_first(root); int kernel_range = 0; while (next) { @@ -328,16 +382,16 @@ static int kernel_maps__split_kallsyms(symbol_filter_t filter, int use_modules) module = strchr(pos->name, '\t'); if (module) { - if (!use_modules) - goto delete_symbol; + if (!thread->use_modules) + goto discard_symbol; *module++ = '\0'; - if (strcmp(map->dso->name, module)) { - map = kernel_maps__find_by_dso_name(module); - if (!map) { - pr_err("/proc/{kallsyms,modules} " - "inconsistency!\n"); + if (strcmp(self->name, module)) { + curr_map = thread__find_map_by_name(thread, module); + if (curr_map == NULL) { + pr_debug("/proc/{kallsyms,modules} " + "inconsistency!\n"); return -1; } } @@ -345,9 +399,9 @@ static int kernel_maps__split_kallsyms(symbol_filter_t filter, int use_modules) * So that we look just like we get from .ko files, * i.e. not prelinked, relative to map->start. */ - pos->start = map->map_ip(map, pos->start); - pos->end = map->map_ip(map, pos->end); - } else if (map != kernel_map) { + pos->start = curr_map->map_ip(curr_map, pos->start); + pos->end = curr_map->map_ip(curr_map, pos->end); + } else if (curr_map != map) { char dso_name[PATH_MAX]; struct dso *dso; @@ -358,25 +412,24 @@ static int kernel_maps__split_kallsyms(symbol_filter_t filter, int use_modules) if (dso == NULL) return -1; - map = map__new2(pos->start, dso); + curr_map = map__new2(pos->start, dso, map->type); if (map == NULL) { dso__delete(dso); return -1; } - map->map_ip = map->unmap_ip = identity__map_ip; - kernel_maps__insert(map); + curr_map->map_ip = curr_map->unmap_ip = identity__map_ip; + __thread__insert_map(thread, curr_map); ++kernel_range; } - if (filter && filter(map, pos)) { -delete_symbol: - rb_erase(&pos->rb_node, &kernel_map->dso->syms); + if (filter && filter(curr_map, pos)) { +discard_symbol: rb_erase(&pos->rb_node, root); symbol__delete(pos); } else { - if (map != kernel_map) { - rb_erase(&pos->rb_node, &kernel_map->dso->syms); - dso__insert_symbol(map->dso, pos); + if (curr_map != map) { + rb_erase(&pos->rb_node, root); + symbols__insert(&curr_map->dso->symbols[curr_map->type], pos); } count++; } @@ -386,32 +439,22 @@ delete_symbol: } -static int kernel_maps__load_kallsyms(symbol_filter_t filter, int use_modules) +static int dso__load_kallsyms(struct dso *self, struct map *map, + struct thread *thread, symbol_filter_t filter) { - if (kernel_maps__load_all_kallsyms()) + if (dso__load_all_kallsyms(self, map) < 0) return -1; - dso__fixup_sym_end(kernel_map->dso); + symbols__fixup_end(&self->symbols[map->type]); + self->origin = DSO__ORIG_KERNEL; - return kernel_maps__split_kallsyms(filter, use_modules); + return dso__split_kallsyms(self, map, thread, filter); } -static size_t kernel_maps__fprintf(FILE *fp) +size_t kernel_maps__fprintf(FILE *fp) { size_t printed = fprintf(fp, "Kernel maps:\n"); - struct rb_node *nd; - - for (nd = rb_first(&kernel_maps); nd; nd = rb_next(nd)) { - struct map *pos = rb_entry(nd, struct map, rb_node); - - printed += fprintf(fp, "Map:"); - printed += map__fprintf(pos, fp); - if (verbose > 1) { - printed += dso__fprintf(pos->dso, fp); - printed += fprintf(fp, "--\n"); - } - } - + printed += thread__fprintf_maps(kthread, fp); return printed + fprintf(fp, "END kernel maps\n"); } @@ -461,7 +504,7 @@ static int dso__load_perf_map(struct dso *self, struct map *map, if (filter && filter(map, sym)) symbol__delete(sym); else { - dso__insert_symbol(self, sym); + symbols__insert(&self->symbols[map->type], sym); nr_syms++; } } @@ -659,7 +702,7 @@ static int dso__synthesize_plt_symbols(struct dso *self, struct map *map, if (filter && filter(map, f)) symbol__delete(f); else { - dso__insert_symbol(self, f); + symbols__insert(&self->symbols[map->type], f); ++nr; } } @@ -681,7 +724,7 @@ static int dso__synthesize_plt_symbols(struct dso *self, struct map *map, if (filter && filter(map, f)) symbol__delete(f); else { - dso__insert_symbol(self, f); + symbols__insert(&self->symbols[map->type], f); ++nr; } } @@ -701,9 +744,9 @@ out: return 0; } -static int dso__load_sym(struct dso *self, struct map *map, const char *name, - int fd, symbol_filter_t filter, int kernel, - int kmodule) +static int dso__load_sym(struct dso *self, struct map *map, + struct thread *thread, const char *name, int fd, + symbol_filter_t filter, int kernel, int kmodule) { struct map *curr_map = map; struct dso *curr_dso = self; @@ -806,7 +849,7 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name, snprintf(dso_name, sizeof(dso_name), "%s%s", self->short_name, section_name); - curr_map = kernel_maps__find_by_dso_name(dso_name); + curr_map = thread__find_map_by_name(thread, dso_name); if (curr_map == NULL) { u64 start = sym.st_value; @@ -816,7 +859,8 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name, curr_dso = dso__new(dso_name); if (curr_dso == NULL) goto out_elf_end; - curr_map = map__new2(start, curr_dso); + curr_map = map__new2(start, curr_dso, + MAP__FUNCTION); if (curr_map == NULL) { dso__delete(curr_dso); goto out_elf_end; @@ -824,8 +868,8 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name, curr_map->map_ip = identity__map_ip; curr_map->unmap_ip = identity__map_ip; curr_dso->origin = DSO__ORIG_KERNEL; - kernel_maps__insert(curr_map); - dsos__add(curr_dso); + __thread__insert_map(kthread, curr_map); + dsos__add(&dsos__kernel, curr_dso); } else curr_dso = curr_map->dso; @@ -855,7 +899,7 @@ new_symbol: if (filter && filter(curr_map, f)) symbol__delete(f); else { - dso__insert_symbol(curr_dso, f); + symbols__insert(&curr_dso->symbols[curr_map->type], f); nr++; } } @@ -864,7 +908,7 @@ new_symbol: * For misannotated, zeroed, ASM function sizes. */ if (nr > 0) - dso__fixup_sym_end(self); + symbols__fixup_end(&self->symbols[map->type]); err = nr; out_elf_end: elf_end(elf); @@ -872,47 +916,46 @@ out_close: return err; } -bool fetch_build_id_table(struct list_head *head) +static bool dso__build_id_equal(const struct dso *self, u8 *build_id) { - bool have_buildid = false; - struct dso *pos; - - list_for_each_entry(pos, &dsos, node) { - struct build_id_list *new; - struct build_id_event b; - size_t len; - - if (filename__read_build_id(pos->long_name, - &b.build_id, - sizeof(b.build_id)) < 0) - continue; - have_buildid = true; - memset(&b.header, 0, sizeof(b.header)); - len = strlen(pos->long_name) + 1; - len = ALIGN(len, 64); - b.header.size = sizeof(b) + len; + return memcmp(self->build_id, build_id, sizeof(self->build_id)) == 0; +} - new = malloc(sizeof(*new)); - if (!new) - die("No memory\n"); +static bool __dsos__read_build_ids(struct list_head *head) +{ + bool have_build_id = false; + struct dso *pos; - memcpy(&new->event, &b, sizeof(b)); - new->dso_name = pos->long_name; - new->len = len; + list_for_each_entry(pos, head, node) + if (filename__read_build_id(pos->long_name, pos->build_id, + sizeof(pos->build_id)) > 0) { + have_build_id = true; + pos->has_build_id = true; + } - list_add_tail(&new->list, head); - } + return have_build_id; +} - return have_buildid; +bool dsos__read_build_ids(void) +{ + return __dsos__read_build_ids(&dsos__kernel) || + __dsos__read_build_ids(&dsos__user); } +/* + * Align offset to 4 bytes as needed for note name and descriptor data. + */ +#define NOTE_ALIGN(n) (((n) + 3) & -4U) + int filename__read_build_id(const char *filename, void *bf, size_t size) { int fd, err = -1; GElf_Ehdr ehdr; GElf_Shdr shdr; - Elf_Data *build_id_data; + Elf_Data *data; Elf_Scn *sec; + Elf_Kind ek; + void *ptr; Elf *elf; if (size < BUILD_ID_SIZE) @@ -928,6 +971,10 @@ int filename__read_build_id(const char *filename, void *bf, size_t size) goto out_close; } + ek = elf_kind(elf); + if (ek != ELF_K_ELF) + goto out_elf_end; + if (gelf_getehdr(elf, &ehdr) == NULL) { pr_err("%s: cannot get elf header.\n", __func__); goto out_elf_end; @@ -935,14 +982,37 @@ int filename__read_build_id(const char *filename, void *bf, size_t size) sec = elf_section_by_name(elf, &ehdr, &shdr, ".note.gnu.build-id", NULL); - if (sec == NULL) - goto out_elf_end; + if (sec == NULL) { + sec = elf_section_by_name(elf, &ehdr, &shdr, + ".notes", NULL); + if (sec == NULL) + goto out_elf_end; + } - build_id_data = elf_getdata(sec, NULL); - if (build_id_data == NULL) + data = elf_getdata(sec, NULL); + if (data == NULL) goto out_elf_end; - memcpy(bf, build_id_data->d_buf + 16, BUILD_ID_SIZE); - err = BUILD_ID_SIZE; + + ptr = data->d_buf; + while (ptr < (data->d_buf + data->d_size)) { + GElf_Nhdr *nhdr = ptr; + int namesz = NOTE_ALIGN(nhdr->n_namesz), + descsz = NOTE_ALIGN(nhdr->n_descsz); + const char *name; + + ptr += sizeof(*nhdr); + name = ptr; + ptr += namesz; + if (nhdr->n_type == NT_GNU_BUILD_ID && + nhdr->n_namesz == sizeof("GNU")) { + if (memcmp(name, "GNU", sizeof("GNU")) == 0) { + memcpy(bf, ptr, BUILD_ID_SIZE); + err = BUILD_ID_SIZE; + break; + } + } + ptr += descsz; + } out_elf_end: elf_end(elf); out_close: @@ -951,23 +1021,48 @@ out: return err; } -static char *dso__read_build_id(struct dso *self) +int sysfs__read_build_id(const char *filename, void *build_id, size_t size) { - int len; - char *build_id = NULL; - unsigned char rawbf[BUILD_ID_SIZE]; + int fd, err = -1; - len = filename__read_build_id(self->long_name, rawbf, sizeof(rawbf)); - if (len < 0) + if (size < BUILD_ID_SIZE) goto out; - build_id = malloc(len * 2 + 1); - if (build_id == NULL) + fd = open(filename, O_RDONLY); + if (fd < 0) goto out; - build_id__sprintf(rawbf, len, build_id); + while (1) { + char bf[BUFSIZ]; + GElf_Nhdr nhdr; + int namesz, descsz; + + if (read(fd, &nhdr, sizeof(nhdr)) != sizeof(nhdr)) + break; + + namesz = NOTE_ALIGN(nhdr.n_namesz); + descsz = NOTE_ALIGN(nhdr.n_descsz); + if (nhdr.n_type == NT_GNU_BUILD_ID && + nhdr.n_namesz == sizeof("GNU")) { + if (read(fd, bf, namesz) != namesz) + break; + if (memcmp(bf, "GNU", sizeof("GNU")) == 0) { + if (read(fd, build_id, + BUILD_ID_SIZE) == BUILD_ID_SIZE) { + err = 0; + break; + } + } else if (read(fd, bf, descsz) != descsz) + break; + } else { + int n = namesz + descsz; + if (read(fd, bf, n) != n) + break; + } + } + close(fd); out: - return build_id; + return err; } char dso__symtab_origin(const struct dso *self) @@ -990,12 +1085,17 @@ char dso__symtab_origin(const struct dso *self) int dso__load(struct dso *self, struct map *map, symbol_filter_t filter) { int size = PATH_MAX; - char *name = malloc(size), *build_id = NULL; + char *name; + u8 build_id[BUILD_ID_SIZE]; int ret = -1; int fd; - self->loaded = 1; + dso__set_loaded(self, map->type); + + if (self->kernel) + return dso__load_kernel_sym(self, map, kthread, filter); + name = malloc(size); if (!name) return -1; @@ -1012,8 +1112,6 @@ int dso__load(struct dso *self, struct map *map, symbol_filter_t filter) more: do { - int berr = 0; - self->origin++; switch (self->origin) { case DSO__ORIG_FEDORA: @@ -1025,12 +1123,18 @@ more: self->long_name); break; case DSO__ORIG_BUILDID: - build_id = dso__read_build_id(self); - if (build_id != NULL) { + if (filename__read_build_id(self->long_name, build_id, + sizeof(build_id))) { + char build_id_hex[BUILD_ID_SIZE * 2 + 1]; + + build_id__sprintf(build_id, sizeof(build_id), + build_id_hex); snprintf(name, size, "/usr/lib/debug/.build-id/%.2s/%s.debug", - build_id, build_id + 2); - goto compare_build_id; + build_id_hex, build_id_hex + 2); + if (self->has_build_id) + goto compare_build_id; + break; } self->origin++; /* Fall thru */ @@ -1043,25 +1147,18 @@ more: } if (self->has_build_id) { - bool match; - build_id = malloc(BUILD_ID_SIZE); - if (build_id == NULL) + if (filename__read_build_id(name, build_id, + sizeof(build_id)) < 0) goto more; - berr = filename__read_build_id(name, build_id, - BUILD_ID_SIZE); compare_build_id: - match = berr > 0 && memcmp(build_id, self->build_id, - sizeof(self->build_id)) == 0; - free(build_id); - build_id = NULL; - if (!match) + if (!dso__build_id_equal(self, build_id)) goto more; } fd = open(name, O_RDONLY); } while (fd < 0); - ret = dso__load_sym(self, map, name, fd, filter, 0, 0); + ret = dso__load_sym(self, map, NULL, name, fd, filter, 0, 0); close(fd); /* @@ -1082,33 +1179,11 @@ out: return ret; } -struct map *kernel_map; - -static void kernel_maps__insert(struct map *map) -{ - maps__insert(&kernel_maps, map); -} - -struct symbol *kernel_maps__find_symbol(u64 ip, struct map **mapp) -{ - struct map *map = maps__find(&kernel_maps, ip); - - if (mapp) - *mapp = map; - - if (map) { - ip = map->map_ip(map, ip); - return map->dso->find_symbol(map->dso, ip); - } - - return NULL; -} - -struct map *kernel_maps__find_by_dso_name(const char *name) +static struct map *thread__find_map_by_name(struct thread *self, char *name) { struct rb_node *nd; - for (nd = rb_first(&kernel_maps); nd; nd = rb_next(nd)) { + for (nd = rb_first(&self->maps[MAP__FUNCTION]); nd; nd = rb_next(nd)) { struct map *map = rb_entry(nd, struct map, rb_node); if (map->dso && strcmp(map->dso->name, name) == 0) @@ -1118,32 +1193,13 @@ struct map *kernel_maps__find_by_dso_name(const char *name) return NULL; } -static int dso__load_module_sym(struct dso *self, struct map *map, - symbol_filter_t filter) -{ - int err = 0, fd = open(self->long_name, O_RDONLY); - - self->loaded = 1; - - if (fd < 0) { - pr_err("%s: cannot open %s\n", __func__, self->long_name); - return err; - } - - err = dso__load_sym(self, map, self->long_name, fd, filter, 0, 1); - close(fd); - - return err; -} - -static int dsos__load_modules_sym_dir(char *dirname, symbol_filter_t filter) +static int dsos__set_modules_path_dir(char *dirname) { struct dirent *dent; - int nr_symbols = 0, err; DIR *dir = opendir(dirname); if (!dir) { - pr_err("%s: cannot open %s dir\n", __func__, dirname); + pr_debug("%s: cannot open %s dir\n", __func__, dirname); return -1; } @@ -1157,14 +1213,13 @@ static int dsos__load_modules_sym_dir(char *dirname, symbol_filter_t filter) snprintf(path, sizeof(path), "%s/%s", dirname, dent->d_name); - err = dsos__load_modules_sym_dir(path, filter); - if (err < 0) + if (dsos__set_modules_path_dir(path) < 0) goto failure; } else { char *dot = strrchr(dent->d_name, '.'), dso_name[PATH_MAX]; struct map *map; - struct rb_node *last; + char *long_name; if (dot == NULL || strcmp(dot, ".ko")) continue; @@ -1172,45 +1227,27 @@ static int dsos__load_modules_sym_dir(char *dirname, symbol_filter_t filter) (int)(dot - dent->d_name), dent->d_name); strxfrchar(dso_name, '-', '_'); - map = kernel_maps__find_by_dso_name(dso_name); + map = thread__find_map_by_name(kthread, dso_name); if (map == NULL) continue; snprintf(path, sizeof(path), "%s/%s", dirname, dent->d_name); - map->dso->long_name = strdup(path); - if (map->dso->long_name == NULL) - goto failure; - - err = dso__load_module_sym(map->dso, map, filter); - if (err < 0) + long_name = strdup(path); + if (long_name == NULL) goto failure; - last = rb_last(&map->dso->syms); - if (last) { - struct symbol *sym; - /* - * We do this here as well, even having the - * symbol size found in the symtab because - * misannotated ASM symbols may have the size - * set to zero. - */ - dso__fixup_sym_end(map->dso); - - sym = rb_entry(last, struct symbol, rb_node); - map->end = map->start + sym->end; - } + dso__set_long_name(map->dso, long_name); } - nr_symbols += err; } - return nr_symbols; + return 0; failure: closedir(dir); return -1; } -static int dsos__load_modules_sym(symbol_filter_t filter) +static int dsos__set_modules_path(void) { struct utsname uts; char modules_path[PATH_MAX]; @@ -1221,7 +1258,7 @@ static int dsos__load_modules_sym(symbol_filter_t filter) snprintf(modules_path, sizeof(modules_path), "/lib/modules/%s/kernel", uts.release); - return dsos__load_modules_sym_dir(modules_path, filter); + return dsos__set_modules_path_dir(modules_path); } /* @@ -1229,7 +1266,7 @@ static int dsos__load_modules_sym(symbol_filter_t filter) * they are loaded) and for vmlinux, where only after we load all the * symbols we'll know where it starts and ends. */ -static struct map *map__new2(u64 start, struct dso *dso) +static struct map *map__new2(u64 start, struct dso *dso, enum map_type type) { struct map *self = malloc(sizeof(*self)); @@ -1237,13 +1274,13 @@ static struct map *map__new2(u64 start, struct dso *dso) /* * ->end will be filled after we load all the symbols */ - map__init(self, start, 0, 0, dso); + map__init(self, type, start, 0, 0, dso); } return self; } -static int dsos__load_modules(void) +static int thread__create_module_maps(struct thread *self) { char *line = NULL; size_t n; @@ -1287,21 +1324,27 @@ static int dsos__load_modules(void) if (dso == NULL) goto out_delete_line; - map = map__new2(start, dso); + map = map__new2(start, dso, MAP__FUNCTION); if (map == NULL) { dso__delete(dso); goto out_delete_line; } + snprintf(name, sizeof(name), + "/sys/module/%s/notes/.note.gnu.build-id", line); + if (sysfs__read_build_id(name, dso->build_id, + sizeof(dso->build_id)) == 0) + dso->has_build_id = true; + dso->origin = DSO__ORIG_KMODULE; - kernel_maps__insert(map); - dsos__add(dso); + __thread__insert_map(self, map); + dsos__add(&dsos__kernel, dso); } free(line); fclose(file); - return 0; + return dsos__set_modules_path(); out_delete_line: free(line); @@ -1309,106 +1352,106 @@ out_failure: return -1; } -static int dso__load_vmlinux(struct dso *self, struct map *map, +static int dso__load_vmlinux(struct dso *self, struct map *map, struct thread *thread, const char *vmlinux, symbol_filter_t filter) { - int err, fd = open(vmlinux, O_RDONLY); + int err = -1, fd; + + if (self->has_build_id) { + u8 build_id[BUILD_ID_SIZE]; - self->loaded = 1; + if (filename__read_build_id(vmlinux, build_id, + sizeof(build_id)) < 0) { + pr_debug("No build_id in %s, ignoring it\n", vmlinux); + return -1; + } + if (!dso__build_id_equal(self, build_id)) { + char expected_build_id[BUILD_ID_SIZE * 2 + 1], + vmlinux_build_id[BUILD_ID_SIZE * 2 + 1]; + + build_id__sprintf(self->build_id, + sizeof(self->build_id), + expected_build_id); + build_id__sprintf(build_id, sizeof(build_id), + vmlinux_build_id); + pr_debug("build_id in %s is %s while expected is %s, " + "ignoring it\n", vmlinux, vmlinux_build_id, + expected_build_id); + return -1; + } + } + fd = open(vmlinux, O_RDONLY); if (fd < 0) return -1; - err = dso__load_sym(self, map, self->long_name, fd, filter, 1, 0); - + dso__set_loaded(self, map->type); + err = dso__load_sym(self, map, thread, self->long_name, fd, filter, 1, 0); close(fd); return err; } -int dsos__load_kernel(const char *vmlinux, symbol_filter_t filter, - int use_modules) +static int dso__load_kernel_sym(struct dso *self, struct map *map, + struct thread *thread, symbol_filter_t filter) { - int err = -1; - struct dso *dso = dso__new(vmlinux); - - if (dso == NULL) - return -1; - - dso->short_name = "[kernel]"; - kernel_map = map__new2(0, dso); - if (kernel_map == NULL) - goto out_delete_dso; - - kernel_map->map_ip = kernel_map->unmap_ip = identity__map_ip; - - if (use_modules && dsos__load_modules() < 0) { - pr_warning("Failed to load list of modules in use! " - "Continuing...\n"); - use_modules = 0; - } - - if (vmlinux) { - err = dso__load_vmlinux(dso, kernel_map, vmlinux, filter); - if (err > 0 && use_modules) { - int syms = dsos__load_modules_sym(filter); - - if (syms < 0) - pr_warning("Failed to read module symbols!" - " Continuing...\n"); - else - err += syms; + int err; + bool is_kallsyms; + + if (vmlinux_path != NULL) { + int i; + pr_debug("Looking at the vmlinux_path (%d entries long)\n", + vmlinux_path__nr_entries); + for (i = 0; i < vmlinux_path__nr_entries; ++i) { + err = dso__load_vmlinux(self, map, thread, + vmlinux_path[i], filter); + if (err > 0) { + pr_debug("Using %s for symbols\n", + vmlinux_path[i]); + dso__set_long_name(self, + strdup(vmlinux_path[i])); + goto out_fixup; + } } } - if (err <= 0) - err = kernel_maps__load_kallsyms(filter, use_modules); + is_kallsyms = self->long_name[0] == '['; + if (is_kallsyms) + goto do_kallsyms; + + err = dso__load_vmlinux(self, map, thread, self->long_name, filter); + if (err <= 0) { + pr_info("The file %s cannot be used, " + "trying to use /proc/kallsyms...", self->long_name); +do_kallsyms: + err = dso__load_kallsyms(self, map, thread, filter); + if (err > 0 && !is_kallsyms) + dso__set_long_name(self, strdup("[kernel.kallsyms]")); + } if (err > 0) { - struct rb_node *node = rb_first(&dso->syms); - struct symbol *sym = rb_entry(node, struct symbol, rb_node); - - kernel_map->start = sym->start; - node = rb_last(&dso->syms); - sym = rb_entry(node, struct symbol, rb_node); - kernel_map->end = sym->end; - - dso->origin = DSO__ORIG_KERNEL; - kernel_maps__insert(kernel_map); - /* - * Now that we have all sorted out, just set the ->end of all - * maps: - */ - kernel_maps__fixup_end(); - dsos__add(dso); - - if (verbose) - kernel_maps__fprintf(stderr); +out_fixup: + map__fixup_start(map); + map__fixup_end(map); } return err; - -out_delete_dso: - dso__delete(dso); - return -1; } -LIST_HEAD(dsos); -struct dso *vdso; - -const char *vmlinux_name = "vmlinux"; -int modules; +LIST_HEAD(dsos__user); +LIST_HEAD(dsos__kernel); +struct dso *vdso; -static void dsos__add(struct dso *dso) +static void dsos__add(struct list_head *head, struct dso *dso) { - list_add_tail(&dso->node, &dsos); + list_add_tail(&dso->node, head); } -static struct dso *dsos__find(const char *name) +static struct dso *dsos__find(struct list_head *head, const char *name) { struct dso *pos; - list_for_each_entry(pos, &dsos, node) + list_for_each_entry(pos, head, node) if (strcmp(pos->name, name) == 0) return pos; return NULL; @@ -1416,53 +1459,170 @@ static struct dso *dsos__find(const char *name) struct dso *dsos__findnew(const char *name) { - struct dso *dso = dsos__find(name); + struct dso *dso = dsos__find(&dsos__user, name); if (!dso) { dso = dso__new(name); - if (dso != NULL) - dsos__add(dso); + if (dso != NULL) { + dsos__add(&dsos__user, dso); + dso__set_basename(dso); + } } return dso; } -void dsos__fprintf(FILE *fp) +static void __dsos__fprintf(struct list_head *head, FILE *fp) { struct dso *pos; - list_for_each_entry(pos, &dsos, node) - dso__fprintf(pos, fp); + list_for_each_entry(pos, head, node) { + int i; + for (i = 0; i < MAP__NR_TYPES; ++i) + dso__fprintf(pos, i, fp); + } } -size_t dsos__fprintf_buildid(FILE *fp) +void dsos__fprintf(FILE *fp) +{ + __dsos__fprintf(&dsos__kernel, fp); + __dsos__fprintf(&dsos__user, fp); +} + +static size_t __dsos__fprintf_buildid(struct list_head *head, FILE *fp) { struct dso *pos; size_t ret = 0; - list_for_each_entry(pos, &dsos, node) { + list_for_each_entry(pos, head, node) { ret += dso__fprintf_buildid(pos, fp); ret += fprintf(fp, " %s\n", pos->long_name); } return ret; } -int load_kernel(symbol_filter_t filter) +size_t dsos__fprintf_buildid(FILE *fp) +{ + return (__dsos__fprintf_buildid(&dsos__kernel, fp) + + __dsos__fprintf_buildid(&dsos__user, fp)); +} + +static int thread__create_kernel_map(struct thread *self, const char *vmlinux) { - if (dsos__load_kernel(vmlinux_name, filter, modules) <= 0) + struct map *kmap; + struct dso *kernel = dso__new(vmlinux ?: "[kernel.kallsyms]"); + + if (kernel == NULL) return -1; + kmap = map__new2(0, kernel, MAP__FUNCTION); + if (kmap == NULL) + goto out_delete_kernel_dso; + + kmap->map_ip = kmap->unmap_ip = identity__map_ip; + kernel->short_name = "[kernel]"; + kernel->kernel = 1; + vdso = dso__new("[vdso]"); - if (!vdso) + if (vdso == NULL) + goto out_delete_kernel_map; + dso__set_loaded(vdso, MAP__FUNCTION); + + if (sysfs__read_build_id("/sys/kernel/notes", kernel->build_id, + sizeof(kernel->build_id)) == 0) + kernel->has_build_id = true; + + __thread__insert_map(self, kmap); + dsos__add(&dsos__kernel, kernel); + dsos__add(&dsos__user, vdso); + + return 0; + +out_delete_kernel_map: + map__delete(kmap); +out_delete_kernel_dso: + dso__delete(kernel); + return -1; +} + +static void vmlinux_path__exit(void) +{ + while (--vmlinux_path__nr_entries >= 0) { + free(vmlinux_path[vmlinux_path__nr_entries]); + vmlinux_path[vmlinux_path__nr_entries] = NULL; + } + + free(vmlinux_path); + vmlinux_path = NULL; +} + +static int vmlinux_path__init(void) +{ + struct utsname uts; + char bf[PATH_MAX]; + + if (uname(&uts) < 0) return -1; - dsos__add(vdso); + vmlinux_path = malloc(sizeof(char *) * 5); + if (vmlinux_path == NULL) + return -1; + + vmlinux_path[vmlinux_path__nr_entries] = strdup("vmlinux"); + if (vmlinux_path[vmlinux_path__nr_entries] == NULL) + goto out_fail; + ++vmlinux_path__nr_entries; + vmlinux_path[vmlinux_path__nr_entries] = strdup("/boot/vmlinux"); + if (vmlinux_path[vmlinux_path__nr_entries] == NULL) + goto out_fail; + ++vmlinux_path__nr_entries; + snprintf(bf, sizeof(bf), "/boot/vmlinux-%s", uts.release); + vmlinux_path[vmlinux_path__nr_entries] = strdup(bf); + if (vmlinux_path[vmlinux_path__nr_entries] == NULL) + goto out_fail; + ++vmlinux_path__nr_entries; + snprintf(bf, sizeof(bf), "/lib/modules/%s/build/vmlinux", uts.release); + vmlinux_path[vmlinux_path__nr_entries] = strdup(bf); + if (vmlinux_path[vmlinux_path__nr_entries] == NULL) + goto out_fail; + ++vmlinux_path__nr_entries; + snprintf(bf, sizeof(bf), "/usr/lib/debug/lib/modules/%s/vmlinux", + uts.release); + vmlinux_path[vmlinux_path__nr_entries] = strdup(bf); + if (vmlinux_path[vmlinux_path__nr_entries] == NULL) + goto out_fail; + ++vmlinux_path__nr_entries; return 0; + +out_fail: + vmlinux_path__exit(); + return -1; } -void symbol__init(unsigned int priv_size) +int symbol__init(struct symbol_conf *conf) { + const struct symbol_conf *pconf = conf ?: &symbol_conf__defaults; + elf_version(EV_CURRENT); - symbol__priv_size = priv_size; + symbol__priv_size = pconf->priv_size; + thread__init(kthread, 0); + + if (pconf->try_vmlinux_path && vmlinux_path__init() < 0) + return -1; + + if (thread__create_kernel_map(kthread, pconf->vmlinux_name) < 0) { + vmlinux_path__exit(); + return -1; + } + + kthread->use_modules = pconf->use_modules; + if (pconf->use_modules && thread__create_module_maps(kthread) < 0) + pr_debug("Failed to load list of modules in use, " + "continuing...\n"); + /* + * Now that we have all the maps created, just set the ->end of them: + */ + thread__fixup_maps_end(kthread); + return 0; } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 51c5a4a0813..17003efa0b3 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -49,6 +49,13 @@ struct symbol { char name[0]; }; +struct symbol_conf { + unsigned short priv_size; + bool try_vmlinux_path, + use_modules; + const char *vmlinux_name; +}; + extern unsigned int symbol__priv_size; static inline void *symbol__priv(struct symbol *self) @@ -56,16 +63,27 @@ static inline void *symbol__priv(struct symbol *self) return ((void *)self) - symbol__priv_size; } +struct addr_location { + struct thread *thread; + struct map *map; + struct symbol *sym; + u64 addr; + char level; +}; + struct dso { struct list_head node; - struct rb_root syms; - struct symbol *(*find_symbol)(struct dso *, u64 ip); + struct rb_root symbols[MAP__NR_TYPES]; + struct symbol *(*find_symbol)(struct dso *self, + enum map_type type, u64 addr); u8 adjust_symbols:1; u8 slen_calculated:1; - u8 loaded:1; u8 has_build_id:1; + u8 kernel:1; unsigned char origin; + u8 loaded; u8 build_id[BUILD_ID_SIZE]; + u16 long_name_len; const char *short_name; char *long_name; char name[0]; @@ -74,30 +92,29 @@ struct dso { struct dso *dso__new(const char *name); void dso__delete(struct dso *self); -struct symbol *dso__find_symbol(struct dso *self, u64 ip); +bool dso__loaded(const struct dso *self, enum map_type type); -int dsos__load_kernel(const char *vmlinux, symbol_filter_t filter, int modules); struct dso *dsos__findnew(const char *name); int dso__load(struct dso *self, struct map *map, symbol_filter_t filter); void dsos__fprintf(FILE *fp); size_t dsos__fprintf_buildid(FILE *fp); size_t dso__fprintf_buildid(struct dso *self, FILE *fp); -size_t dso__fprintf(struct dso *self, FILE *fp); +size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp); char dso__symtab_origin(const struct dso *self); void dso__set_build_id(struct dso *self, void *build_id); int filename__read_build_id(const char *filename, void *bf, size_t size); -bool fetch_build_id_table(struct list_head *head); +int sysfs__read_build_id(const char *filename, void *bf, size_t size); +bool dsos__read_build_ids(void); int build_id__sprintf(u8 *self, int len, char *bf); -int load_kernel(symbol_filter_t filter); +size_t kernel_maps__fprintf(FILE *fp); -void symbol__init(unsigned int priv_size); +int symbol__init(struct symbol_conf *conf); -extern struct list_head dsos; -extern struct map *kernel_map; +struct thread; +struct thread *kthread; +extern struct list_head dsos__user, dsos__kernel; extern struct dso *vdso; -extern const char *vmlinux_name; -extern int modules; #endif /* __PERF_SYMBOL */ diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 0f6d78c9863..603f5610861 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -9,17 +9,26 @@ static struct rb_root threads; static struct thread *last_match; +void thread__init(struct thread *self, pid_t pid) +{ + int i; + self->pid = pid; + self->comm = NULL; + for (i = 0; i < MAP__NR_TYPES; ++i) { + self->maps[i] = RB_ROOT; + INIT_LIST_HEAD(&self->removed_maps[i]); + } +} + static struct thread *thread__new(pid_t pid) { - struct thread *self = calloc(1, sizeof(*self)); + struct thread *self = zalloc(sizeof(*self)); if (self != NULL) { - self->pid = pid; + thread__init(self, pid); self->comm = malloc(32); if (self->comm) snprintf(self->comm, 32, ":%d", self->pid); - self->maps = RB_ROOT; - INIT_LIST_HEAD(&self->removed_maps); } return self; @@ -44,24 +53,68 @@ int thread__comm_len(struct thread *self) return self->comm_len; } -static size_t thread__fprintf(struct thread *self, FILE *fp) +static const char *map_type__name[MAP__NR_TYPES] = { + [MAP__FUNCTION] = "Functions", +}; + +static size_t __thread__fprintf_maps(struct thread *self, + enum map_type type, FILE *fp) { + size_t printed = fprintf(fp, "%s:\n", map_type__name[type]); struct rb_node *nd; - struct map *pos; - size_t ret = fprintf(fp, "Thread %d %s\nCurrent maps:\n", - self->pid, self->comm); - for (nd = rb_first(&self->maps); nd; nd = rb_next(nd)) { - pos = rb_entry(nd, struct map, rb_node); - ret += map__fprintf(pos, fp); + for (nd = rb_first(&self->maps[type]); nd; nd = rb_next(nd)) { + struct map *pos = rb_entry(nd, struct map, rb_node); + printed += fprintf(fp, "Map:"); + printed += map__fprintf(pos, fp); + if (verbose > 1) { + printed += dso__fprintf(pos->dso, type, fp); + printed += fprintf(fp, "--\n"); + } } - ret = fprintf(fp, "Removed maps:\n"); + return printed; +} - list_for_each_entry(pos, &self->removed_maps, node) - ret += map__fprintf(pos, fp); +size_t thread__fprintf_maps(struct thread *self, FILE *fp) +{ + size_t printed = 0, i; + for (i = 0; i < MAP__NR_TYPES; ++i) + printed += __thread__fprintf_maps(self, i, fp); + return printed; +} - return ret; +static size_t __thread__fprintf_removed_maps(struct thread *self, + enum map_type type, FILE *fp) +{ + struct map *pos; + size_t printed = 0; + + list_for_each_entry(pos, &self->removed_maps[type], node) { + printed += fprintf(fp, "Map:"); + printed += map__fprintf(pos, fp); + if (verbose > 1) { + printed += dso__fprintf(pos->dso, type, fp); + printed += fprintf(fp, "--\n"); + } + } + return printed; +} + +static size_t thread__fprintf_removed_maps(struct thread *self, FILE *fp) +{ + size_t printed = 0, i; + for (i = 0; i < MAP__NR_TYPES; ++i) + printed += __thread__fprintf_removed_maps(self, i, fp); + return printed; +} + +static size_t thread__fprintf(struct thread *self, FILE *fp) +{ + size_t printed = fprintf(fp, "Thread %d %s\n", self->pid, self->comm); + printed += thread__fprintf_removed_maps(self, fp); + printed += fprintf(fp, "Removed maps:\n"); + return printed + thread__fprintf_removed_maps(self, fp); } struct thread *threads__findnew(pid_t pid) @@ -117,7 +170,8 @@ struct thread *register_idle_thread(void) static void thread__remove_overlappings(struct thread *self, struct map *map) { - struct rb_node *next = rb_first(&self->maps); + struct rb_root *root = &self->maps[map->type]; + struct rb_node *next = rb_first(root); while (next) { struct map *pos = rb_entry(next, struct map, rb_node); @@ -132,13 +186,13 @@ static void thread__remove_overlappings(struct thread *self, struct map *map) map__fprintf(pos, stderr); } - rb_erase(&pos->rb_node, &self->maps); + rb_erase(&pos->rb_node, root); /* * We may have references to this map, for instance in some * hist_entry instances, so just move them to a separate * list. */ - list_add_tail(&pos->node, &self->removed_maps); + list_add_tail(&pos->node, &self->removed_maps[map->type]); } } @@ -185,12 +239,26 @@ struct map *maps__find(struct rb_root *maps, u64 ip) void thread__insert_map(struct thread *self, struct map *map) { thread__remove_overlappings(self, map); - maps__insert(&self->maps, map); + maps__insert(&self->maps[map->type], map); } -int thread__fork(struct thread *self, struct thread *parent) +static int thread__clone_maps(struct thread *self, struct thread *parent, + enum map_type type) { struct rb_node *nd; + for (nd = rb_first(&parent->maps[type]); nd; nd = rb_next(nd)) { + struct map *map = rb_entry(nd, struct map, rb_node); + struct map *new = map__clone(map); + if (new == NULL) + return -ENOMEM; + thread__insert_map(self, new); + } + return 0; +} + +int thread__fork(struct thread *self, struct thread *parent) +{ + int i; if (self->comm) free(self->comm); @@ -198,14 +266,9 @@ int thread__fork(struct thread *self, struct thread *parent) if (!self->comm) return -ENOMEM; - for (nd = rb_first(&parent->maps); nd; nd = rb_next(nd)) { - struct map *map = rb_entry(nd, struct map, rb_node); - struct map *new = map__clone(map); - if (!new) + for (i = 0; i < MAP__NR_TYPES; ++i) + if (thread__clone_maps(self, parent, i) < 0) return -ENOMEM; - thread__insert_map(self, new); - } - return 0; } @@ -222,3 +285,15 @@ size_t threads__fprintf(FILE *fp) return ret; } + +struct symbol *thread__find_symbol(struct thread *self, + enum map_type type, u64 addr, + symbol_filter_t filter) +{ + struct map *map = thread__find_map(self, type, addr); + + if (map != NULL) + return map__find_symbol(map, map->map_ip(map, addr), filter); + + return NULL; +} diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index 53addd77ce8..686d6e914d9 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -7,31 +7,50 @@ struct thread { struct rb_node rb_node; - struct rb_root maps; - struct list_head removed_maps; + struct rb_root maps[MAP__NR_TYPES]; + struct list_head removed_maps[MAP__NR_TYPES]; pid_t pid; + bool use_modules; char shortname[3]; char *comm; int comm_len; }; +void thread__init(struct thread *self, pid_t pid); int thread__set_comm(struct thread *self, const char *comm); int thread__comm_len(struct thread *self); struct thread *threads__findnew(pid_t pid); struct thread *register_idle_thread(void); void thread__insert_map(struct thread *self, struct map *map); int thread__fork(struct thread *self, struct thread *parent); +size_t thread__fprintf_maps(struct thread *self, FILE *fp); size_t threads__fprintf(FILE *fp); void maps__insert(struct rb_root *maps, struct map *map); -struct map *maps__find(struct rb_root *maps, u64 ip); +struct map *maps__find(struct rb_root *maps, u64 addr); -struct symbol *kernel_maps__find_symbol(const u64 ip, struct map **mapp); -struct map *kernel_maps__find_by_dso_name(const char *name); +static inline struct map *thread__find_map(struct thread *self, + enum map_type type, u64 addr) +{ + return self ? maps__find(&self->maps[type], addr) : NULL; +} -static inline struct map *thread__find_map(struct thread *self, u64 ip) +static inline void __thread__insert_map(struct thread *self, struct map *map) { - return self ? maps__find(&self->maps, ip) : NULL; + maps__insert(&self->maps[map->type], map); } +void thread__find_addr_location(struct thread *self, u8 cpumode, + enum map_type type, u64 addr, + struct addr_location *al, + symbol_filter_t filter); +struct symbol *thread__find_symbol(struct thread *self, + enum map_type type, u64 addr, + symbol_filter_t filter); + +static inline struct symbol * +thread__find_function(struct thread *self, u64 addr, symbol_filter_t filter) +{ + return thread__find_symbol(self, MAP__FUNCTION, addr, filter); +} #endif /* __PERF_THREAD_H */ diff --git a/tools/perf/util/trace-event-info.c b/tools/perf/util/trace-event-info.c index 831052d4b4f..cace3559553 100644 --- a/tools/perf/util/trace-event-info.c +++ b/tools/perf/util/trace-event-info.c @@ -33,11 +33,11 @@ #include <ctype.h> #include <errno.h> #include <stdbool.h> +#include <linux/kernel.h> #include "../perf.h" #include "trace-event.h" - #define VERSION "0.5" #define _STR(x) #x @@ -483,23 +483,31 @@ static struct tracepoint_path * get_tracepoints_path(struct perf_event_attr *pattrs, int nb_events) { struct tracepoint_path path, *ppath = &path; - int i; + int i, nr_tracepoints = 0; for (i = 0; i < nb_events; i++) { if (pattrs[i].type != PERF_TYPE_TRACEPOINT) continue; + ++nr_tracepoints; ppath->next = tracepoint_id_to_path(pattrs[i].config); if (!ppath->next) die("%s\n", "No memory to alloc tracepoints list"); ppath = ppath->next; } - return path.next; + return nr_tracepoints > 0 ? path.next : NULL; } -void read_tracing_data(int fd, struct perf_event_attr *pattrs, int nb_events) + +int read_tracing_data(int fd, struct perf_event_attr *pattrs, int nb_events) { char buf[BUFSIZ]; - struct tracepoint_path *tps; + struct tracepoint_path *tps = get_tracepoints_path(pattrs, nb_events); + + /* + * What? No tracepoints? No sense writing anything here, bail out. + */ + if (tps == NULL) + return -1; output_fd = fd; @@ -528,11 +536,11 @@ void read_tracing_data(int fd, struct perf_event_attr *pattrs, int nb_events) page_size = getpagesize(); write_or_die(&page_size, 4); - tps = get_tracepoints_path(pattrs, nb_events); - read_header_files(); read_ftrace_files(tps); read_event_files(tps); read_proc_kallsyms(); read_ftrace_printk(); + + return 0; } diff --git a/tools/perf/util/trace-event-parse.c b/tools/perf/util/trace-event-parse.c index eae56050308..0302405aa2c 100644 --- a/tools/perf/util/trace-event-parse.c +++ b/tools/perf/util/trace-event-parse.c @@ -48,6 +48,11 @@ static unsigned long long input_buf_siz; static int cpus; static int long_size; +static int is_flag_field; +static int is_symbolic_field; + +static struct format_field * +find_any_field(struct event *event, const char *name); static void init_input_buf(char *buf, unsigned long long size) { @@ -941,7 +946,8 @@ static int event_read_fields(struct event *event, struct format_field **fields) if (read_expect_type(EVENT_ITEM, &token)) goto fail; - /* add signed type */ + if (strtoul(token, NULL, 0)) + field->flags |= FIELD_IS_SIGNED; free_token(token); if (read_expected(EVENT_OP, ";") < 0) @@ -1300,6 +1306,16 @@ process_entry(struct event *event __unused, struct print_arg *arg, arg->type = PRINT_FIELD; arg->field.name = field; + if (is_flag_field) { + arg->field.field = find_any_field(event, arg->field.name); + arg->field.field->flags |= FIELD_IS_FLAG; + is_flag_field = 0; + } else if (is_symbolic_field) { + arg->field.field = find_any_field(event, arg->field.name); + arg->field.field->flags |= FIELD_IS_SYMBOLIC; + is_symbolic_field = 0; + } + type = read_token(&token); *tok = token; @@ -1667,9 +1683,11 @@ process_arg_token(struct event *event, struct print_arg *arg, type = process_entry(event, arg, &token); } else if (strcmp(token, "__print_flags") == 0) { free_token(token); + is_flag_field = 1; type = process_flags(event, arg, &token); } else if (strcmp(token, "__print_symbolic") == 0) { free_token(token); + is_symbolic_field = 1; type = process_symbols(event, arg, &token); } else if (strcmp(token, "__get_str") == 0) { free_token(token); @@ -1870,7 +1888,7 @@ find_any_field(struct event *event, const char *name) return find_field(event, name); } -static unsigned long long read_size(void *ptr, int size) +unsigned long long read_size(void *ptr, int size) { switch (size) { case 1: @@ -1955,7 +1973,7 @@ int trace_parse_common_type(void *data) "common_type"); } -static int parse_common_pid(void *data) +int trace_parse_common_pid(void *data) { static int pid_offset; static int pid_size; @@ -1964,7 +1982,7 @@ static int parse_common_pid(void *data) "common_pid"); } -static int parse_common_pc(void *data) +int parse_common_pc(void *data) { static int pc_offset; static int pc_size; @@ -1973,7 +1991,7 @@ static int parse_common_pc(void *data) "common_preempt_count"); } -static int parse_common_flags(void *data) +int parse_common_flags(void *data) { static int flags_offset; static int flags_size; @@ -1982,7 +2000,7 @@ static int parse_common_flags(void *data) "common_flags"); } -static int parse_common_lock_depth(void *data) +int parse_common_lock_depth(void *data) { static int ld_offset; static int ld_size; @@ -2007,6 +2025,14 @@ struct event *trace_find_event(int id) return event; } +struct event *trace_find_next_event(struct event *event) +{ + if (!event) + return event_list; + + return event->next; +} + static unsigned long long eval_num_arg(void *data, int size, struct event *event, struct print_arg *arg) { @@ -2146,7 +2172,7 @@ static const struct flag flags[] = { { "HRTIMER_RESTART", 1 }, }; -static unsigned long long eval_flag(const char *flag) +unsigned long long eval_flag(const char *flag) { int i; @@ -2676,7 +2702,7 @@ get_return_for_leaf(int cpu, int cur_pid, unsigned long long cur_func, if (!(event->flags & EVENT_FL_ISFUNCRET)) return NULL; - pid = parse_common_pid(next->data); + pid = trace_parse_common_pid(next->data); field = find_field(event, "func"); if (!field) die("function return does not have field func"); @@ -2962,7 +2988,7 @@ void print_event(int cpu, void *data, int size, unsigned long long nsecs, return; } - pid = parse_common_pid(data); + pid = trace_parse_common_pid(data); if (event->flags & (EVENT_FL_ISFUNCENT | EVENT_FL_ISFUNCRET)) return pretty_print_func_graph(data, size, event, cpu, diff --git a/tools/perf/util/trace-event-perl.c b/tools/perf/util/trace-event-perl.c new file mode 100644 index 00000000000..51e833fd58c --- /dev/null +++ b/tools/perf/util/trace-event-perl.c @@ -0,0 +1,598 @@ +/* + * trace-event-perl. Feed perf trace events to an embedded Perl interpreter. + * + * Copyright (C) 2009 Tom Zanussi <tzanussi@gmail.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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + */ + +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <ctype.h> +#include <errno.h> + +#include "../perf.h" +#include "util.h" +#include "trace-event.h" +#include "trace-event-perl.h" + +void xs_init(pTHX); + +void boot_Perf__Trace__Context(pTHX_ CV *cv); +void boot_DynaLoader(pTHX_ CV *cv); + +void xs_init(pTHX) +{ + const char *file = __FILE__; + dXSUB_SYS; + + newXS("Perf::Trace::Context::bootstrap", boot_Perf__Trace__Context, + file); + newXS("DynaLoader::boot_DynaLoader", boot_DynaLoader, file); +} + +INTERP my_perl; + +#define FTRACE_MAX_EVENT \ + ((1 << (sizeof(unsigned short) * 8)) - 1) + +struct event *events[FTRACE_MAX_EVENT]; + +static struct scripting_context *scripting_context; + +static char *cur_field_name; +static int zero_flag_atom; + +static void define_symbolic_value(const char *ev_name, + const char *field_name, + const char *field_value, + const char *field_str) +{ + unsigned long long value; + dSP; + + value = eval_flag(field_value); + + ENTER; + SAVETMPS; + PUSHMARK(SP); + + XPUSHs(sv_2mortal(newSVpv(ev_name, 0))); + XPUSHs(sv_2mortal(newSVpv(field_name, 0))); + XPUSHs(sv_2mortal(newSVuv(value))); + XPUSHs(sv_2mortal(newSVpv(field_str, 0))); + + PUTBACK; + if (get_cv("main::define_symbolic_value", 0)) + call_pv("main::define_symbolic_value", G_SCALAR); + SPAGAIN; + PUTBACK; + FREETMPS; + LEAVE; +} + +static void define_symbolic_values(struct print_flag_sym *field, + const char *ev_name, + const char *field_name) +{ + define_symbolic_value(ev_name, field_name, field->value, field->str); + if (field->next) + define_symbolic_values(field->next, ev_name, field_name); +} + +static void define_symbolic_field(const char *ev_name, + const char *field_name) +{ + dSP; + + ENTER; + SAVETMPS; + PUSHMARK(SP); + + XPUSHs(sv_2mortal(newSVpv(ev_name, 0))); + XPUSHs(sv_2mortal(newSVpv(field_name, 0))); + + PUTBACK; + if (get_cv("main::define_symbolic_field", 0)) + call_pv("main::define_symbolic_field", G_SCALAR); + SPAGAIN; + PUTBACK; + FREETMPS; + LEAVE; +} + +static void define_flag_value(const char *ev_name, + const char *field_name, + const char *field_value, + const char *field_str) +{ + unsigned long long value; + dSP; + + value = eval_flag(field_value); + + ENTER; + SAVETMPS; + PUSHMARK(SP); + + XPUSHs(sv_2mortal(newSVpv(ev_name, 0))); + XPUSHs(sv_2mortal(newSVpv(field_name, 0))); + XPUSHs(sv_2mortal(newSVuv(value))); + XPUSHs(sv_2mortal(newSVpv(field_str, 0))); + + PUTBACK; + if (get_cv("main::define_flag_value", 0)) + call_pv("main::define_flag_value", G_SCALAR); + SPAGAIN; + PUTBACK; + FREETMPS; + LEAVE; +} + +static void define_flag_values(struct print_flag_sym *field, + const char *ev_name, + const char *field_name) +{ + define_flag_value(ev_name, field_name, field->value, field->str); + if (field->next) + define_flag_values(field->next, ev_name, field_name); +} + +static void define_flag_field(const char *ev_name, + const char *field_name, + const char *delim) +{ + dSP; + + ENTER; + SAVETMPS; + PUSHMARK(SP); + + XPUSHs(sv_2mortal(newSVpv(ev_name, 0))); + XPUSHs(sv_2mortal(newSVpv(field_name, 0))); + XPUSHs(sv_2mortal(newSVpv(delim, 0))); + + PUTBACK; + if (get_cv("main::define_flag_field", 0)) + call_pv("main::define_flag_field", G_SCALAR); + SPAGAIN; + PUTBACK; + FREETMPS; + LEAVE; +} + +static void define_event_symbols(struct event *event, + const char *ev_name, + struct print_arg *args) +{ + switch (args->type) { + case PRINT_NULL: + break; + case PRINT_ATOM: + define_flag_value(ev_name, cur_field_name, "0", + args->atom.atom); + zero_flag_atom = 0; + break; + case PRINT_FIELD: + if (cur_field_name) + free(cur_field_name); + cur_field_name = strdup(args->field.name); + break; + case PRINT_FLAGS: + define_event_symbols(event, ev_name, args->flags.field); + define_flag_field(ev_name, cur_field_name, args->flags.delim); + define_flag_values(args->flags.flags, ev_name, cur_field_name); + break; + case PRINT_SYMBOL: + define_event_symbols(event, ev_name, args->symbol.field); + define_symbolic_field(ev_name, cur_field_name); + define_symbolic_values(args->symbol.symbols, ev_name, + cur_field_name); + break; + case PRINT_STRING: + break; + case PRINT_TYPE: + define_event_symbols(event, ev_name, args->typecast.item); + break; + case PRINT_OP: + if (strcmp(args->op.op, ":") == 0) + zero_flag_atom = 1; + define_event_symbols(event, ev_name, args->op.left); + define_event_symbols(event, ev_name, args->op.right); + break; + default: + /* we should warn... */ + return; + } + + if (args->next) + define_event_symbols(event, ev_name, args->next); +} + +static inline struct event *find_cache_event(int type) +{ + static char ev_name[256]; + struct event *event; + + if (events[type]) + return events[type]; + + events[type] = event = trace_find_event(type); + if (!event) + return NULL; + + sprintf(ev_name, "%s::%s", event->system, event->name); + + define_event_symbols(event, ev_name, event->print_fmt.args); + + return event; +} + +int common_pc(struct scripting_context *context) +{ + int pc; + + pc = parse_common_pc(context->event_data); + + return pc; +} + +int common_flags(struct scripting_context *context) +{ + int flags; + + flags = parse_common_flags(context->event_data); + + return flags; +} + +int common_lock_depth(struct scripting_context *context) +{ + int lock_depth; + + lock_depth = parse_common_lock_depth(context->event_data); + + return lock_depth; +} + +static void perl_process_event(int cpu, void *data, + int size __attribute((unused)), + unsigned long long nsecs, char *comm) +{ + struct format_field *field; + static char handler[256]; + unsigned long long val; + unsigned long s, ns; + struct event *event; + int type; + int pid; + + dSP; + + type = trace_parse_common_type(data); + + event = find_cache_event(type); + if (!event) + die("ug! no event found for type %d", type); + + pid = trace_parse_common_pid(data); + + sprintf(handler, "%s::%s", event->system, event->name); + + s = nsecs / NSECS_PER_SEC; + ns = nsecs - s * NSECS_PER_SEC; + + scripting_context->event_data = data; + + ENTER; + SAVETMPS; + PUSHMARK(SP); + + XPUSHs(sv_2mortal(newSVpv(handler, 0))); + XPUSHs(sv_2mortal(newSViv(PTR2IV(scripting_context)))); + XPUSHs(sv_2mortal(newSVuv(cpu))); + XPUSHs(sv_2mortal(newSVuv(s))); + XPUSHs(sv_2mortal(newSVuv(ns))); + XPUSHs(sv_2mortal(newSViv(pid))); + XPUSHs(sv_2mortal(newSVpv(comm, 0))); + + /* common fields other than pid can be accessed via xsub fns */ + + for (field = event->format.fields; field; field = field->next) { + if (field->flags & FIELD_IS_STRING) { + int offset; + if (field->flags & FIELD_IS_DYNAMIC) { + offset = *(int *)(data + field->offset); + offset &= 0xffff; + } else + offset = field->offset; + XPUSHs(sv_2mortal(newSVpv((char *)data + offset, 0))); + } else { /* FIELD_IS_NUMERIC */ + val = read_size(data + field->offset, field->size); + if (field->flags & FIELD_IS_SIGNED) { + XPUSHs(sv_2mortal(newSViv(val))); + } else { + XPUSHs(sv_2mortal(newSVuv(val))); + } + } + } + + PUTBACK; + + if (get_cv(handler, 0)) + call_pv(handler, G_SCALAR); + else if (get_cv("main::trace_unhandled", 0)) { + XPUSHs(sv_2mortal(newSVpv(handler, 0))); + XPUSHs(sv_2mortal(newSViv(PTR2IV(scripting_context)))); + XPUSHs(sv_2mortal(newSVuv(cpu))); + XPUSHs(sv_2mortal(newSVuv(nsecs))); + XPUSHs(sv_2mortal(newSViv(pid))); + XPUSHs(sv_2mortal(newSVpv(comm, 0))); + call_pv("main::trace_unhandled", G_SCALAR); + } + SPAGAIN; + PUTBACK; + FREETMPS; + LEAVE; +} + +static void run_start_sub(void) +{ + dSP; /* access to Perl stack */ + PUSHMARK(SP); + + if (get_cv("main::trace_begin", 0)) + call_pv("main::trace_begin", G_DISCARD | G_NOARGS); +} + +/* + * Start trace script + */ +static int perl_start_script(const char *script) +{ + const char *command_line[2] = { "", NULL }; + + command_line[1] = script; + + my_perl = perl_alloc(); + perl_construct(my_perl); + + if (perl_parse(my_perl, xs_init, 2, (char **)command_line, + (char **)NULL)) + return -1; + + perl_run(my_perl); + if (SvTRUE(ERRSV)) + return -1; + + run_start_sub(); + + fprintf(stderr, "perf trace started with Perl script %s\n\n", script); + + return 0; +} + +/* + * Stop trace script + */ +static int perl_stop_script(void) +{ + dSP; /* access to Perl stack */ + PUSHMARK(SP); + + if (get_cv("main::trace_end", 0)) + call_pv("main::trace_end", G_DISCARD | G_NOARGS); + + perl_destruct(my_perl); + perl_free(my_perl); + + fprintf(stderr, "\nperf trace Perl script stopped\n"); + + return 0; +} + +static int perl_generate_script(const char *outfile) +{ + struct event *event = NULL; + struct format_field *f; + char fname[PATH_MAX]; + int not_first, count; + FILE *ofp; + + sprintf(fname, "%s.pl", outfile); + ofp = fopen(fname, "w"); + if (ofp == NULL) { + fprintf(stderr, "couldn't open %s\n", fname); + return -1; + } + + fprintf(ofp, "# perf trace event handlers, " + "generated by perf trace -g perl\n"); + + fprintf(ofp, "# Licensed under the terms of the GNU GPL" + " License version 2\n\n"); + + fprintf(ofp, "# The common_* event handler fields are the most useful " + "fields common to\n"); + + fprintf(ofp, "# all events. They don't necessarily correspond to " + "the 'common_*' fields\n"); + + fprintf(ofp, "# in the format files. Those fields not available as " + "handler params can\n"); + + fprintf(ofp, "# be retrieved using Perl functions of the form " + "common_*($context).\n"); + + fprintf(ofp, "# See Context.pm for the list of available " + "functions.\n\n"); + + fprintf(ofp, "use lib \"$ENV{'PERF_EXEC_PATH'}/scripts/perl/" + "Perf-Trace-Util/lib\";\n"); + + fprintf(ofp, "use lib \"./Perf-Trace-Util/lib\";\n"); + fprintf(ofp, "use Perf::Trace::Core;\n"); + fprintf(ofp, "use Perf::Trace::Context;\n"); + fprintf(ofp, "use Perf::Trace::Util;\n\n"); + + fprintf(ofp, "sub trace_begin\n{\n\t# optional\n}\n\n"); + fprintf(ofp, "sub trace_end\n{\n\t# optional\n}\n\n"); + + while ((event = trace_find_next_event(event))) { + fprintf(ofp, "sub %s::%s\n{\n", event->system, event->name); + fprintf(ofp, "\tmy ("); + + fprintf(ofp, "$event_name, "); + fprintf(ofp, "$context, "); + fprintf(ofp, "$common_cpu, "); + fprintf(ofp, "$common_secs, "); + fprintf(ofp, "$common_nsecs,\n"); + fprintf(ofp, "\t $common_pid, "); + fprintf(ofp, "$common_comm,\n\t "); + + not_first = 0; + count = 0; + + for (f = event->format.fields; f; f = f->next) { + if (not_first++) + fprintf(ofp, ", "); + if (++count % 5 == 0) + fprintf(ofp, "\n\t "); + + fprintf(ofp, "$%s", f->name); + } + fprintf(ofp, ") = @_;\n\n"); + + fprintf(ofp, "\tprint_header($event_name, $common_cpu, " + "$common_secs, $common_nsecs,\n\t " + "$common_pid, $common_comm);\n\n"); + + fprintf(ofp, "\tprintf(\""); + + not_first = 0; + count = 0; + + for (f = event->format.fields; f; f = f->next) { + if (not_first++) + fprintf(ofp, ", "); + if (count && count % 4 == 0) { + fprintf(ofp, "\".\n\t \""); + } + count++; + + fprintf(ofp, "%s=", f->name); + if (f->flags & FIELD_IS_STRING || + f->flags & FIELD_IS_FLAG || + f->flags & FIELD_IS_SYMBOLIC) + fprintf(ofp, "%%s"); + else if (f->flags & FIELD_IS_SIGNED) + fprintf(ofp, "%%d"); + else + fprintf(ofp, "%%u"); + } + + fprintf(ofp, "\\n\",\n\t "); + + not_first = 0; + count = 0; + + for (f = event->format.fields; f; f = f->next) { + if (not_first++) + fprintf(ofp, ", "); + + if (++count % 5 == 0) + fprintf(ofp, "\n\t "); + + if (f->flags & FIELD_IS_FLAG) { + if ((count - 1) % 5 != 0) { + fprintf(ofp, "\n\t "); + count = 4; + } + fprintf(ofp, "flag_str(\""); + fprintf(ofp, "%s::%s\", ", event->system, + event->name); + fprintf(ofp, "\"%s\", $%s)", f->name, + f->name); + } else if (f->flags & FIELD_IS_SYMBOLIC) { + if ((count - 1) % 5 != 0) { + fprintf(ofp, "\n\t "); + count = 4; + } + fprintf(ofp, "symbol_str(\""); + fprintf(ofp, "%s::%s\", ", event->system, + event->name); + fprintf(ofp, "\"%s\", $%s)", f->name, + f->name); + } else + fprintf(ofp, "$%s", f->name); + } + + fprintf(ofp, ");\n"); + fprintf(ofp, "}\n\n"); + } + + fprintf(ofp, "sub trace_unhandled\n{\n\tmy ($event_name, $context, " + "$common_cpu, $common_secs, $common_nsecs,\n\t " + "$common_pid, $common_comm) = @_;\n\n"); + + fprintf(ofp, "\tprint_header($event_name, $common_cpu, " + "$common_secs, $common_nsecs,\n\t $common_pid, " + "$common_comm);\n}\n\n"); + + fprintf(ofp, "sub print_header\n{\n" + "\tmy ($event_name, $cpu, $secs, $nsecs, $pid, $comm) = @_;\n\n" + "\tprintf(\"%%-20s %%5u %%05u.%%09u %%8u %%-20s \",\n\t " + "$event_name, $cpu, $secs, $nsecs, $pid, $comm);\n}"); + + fclose(ofp); + + fprintf(stderr, "generated Perl script: %s\n", fname); + + return 0; +} + +struct scripting_ops perl_scripting_ops = { + .name = "Perl", + .start_script = perl_start_script, + .stop_script = perl_stop_script, + .process_event = perl_process_event, + .generate_script = perl_generate_script, +}; + +#ifdef NO_LIBPERL +void setup_perl_scripting(void) +{ + fprintf(stderr, "Perl scripting not supported." + " Install libperl and rebuild perf to enable it. e.g. " + "apt-get install libperl-dev (ubuntu), yum install " + "perl-ExtUtils-Embed (Fedora), etc.\n"); +} +#else +void setup_perl_scripting(void) +{ + int err; + err = script_spec_register("Perl", &perl_scripting_ops); + if (err) + die("error registering Perl script extension"); + + err = script_spec_register("pl", &perl_scripting_ops); + if (err) + die("error registering pl script extension"); + + scripting_context = malloc(sizeof(struct scripting_context)); +} +#endif diff --git a/tools/perf/util/trace-event-perl.h b/tools/perf/util/trace-event-perl.h new file mode 100644 index 00000000000..8fe0d866fe1 --- /dev/null +++ b/tools/perf/util/trace-event-perl.h @@ -0,0 +1,51 @@ +#ifndef __PERF_TRACE_EVENT_PERL_H +#define __PERF_TRACE_EVENT_PERL_H +#ifdef NO_LIBPERL +typedef int INTERP; +#define dSP +#define ENTER +#define SAVETMPS +#define PUTBACK +#define SPAGAIN +#define FREETMPS +#define LEAVE +#define SP +#define ERRSV +#define G_SCALAR (0) +#define G_DISCARD (0) +#define G_NOARGS (0) +#define PUSHMARK(a) +#define SvTRUE(a) (0) +#define XPUSHs(s) +#define sv_2mortal(a) +#define newSVpv(a,b) +#define newSVuv(a) +#define newSViv(a) +#define get_cv(a,b) (0) +#define call_pv(a,b) (0) +#define perl_alloc() (0) +#define perl_construct(a) (0) +#define perl_parse(a,b,c,d,e) (0) +#define perl_run(a) (0) +#define perl_destruct(a) (0) +#define perl_free(a) (0) +#define pTHX void +#define CV void +#define dXSUB_SYS +#define pTHX_ +static inline void newXS(const char *a, void *b, const char *c) {} +#else +#include <EXTERN.h> +#include <perl.h> +typedef PerlInterpreter * INTERP; +#endif + +struct scripting_context { + void *event_data; +}; + +int common_pc(struct scripting_context *context); +int common_flags(struct scripting_context *context); +int common_lock_depth(struct scripting_context *context); + +#endif /* __PERF_TRACE_EVENT_PERL_H */ diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c index 44292e06cca..342dfdd43f8 100644 --- a/tools/perf/util/trace-event-read.c +++ b/tools/perf/util/trace-event-read.c @@ -471,11 +471,11 @@ void trace_report(int fd) read_or_die(buf, 3); if (memcmp(buf, test, 3) != 0) - die("not an trace data file"); + die("no trace data in the file"); read_or_die(buf, 7); if (memcmp(buf, "tracing", 7) != 0) - die("not a trace file (missing tracing)"); + die("not a trace file (missing 'tracing' tag)"); version = read_string(); if (show_version) diff --git a/tools/perf/util/trace-event.h b/tools/perf/util/trace-event.h index f6637c2fa1f..81698d5e650 100644 --- a/tools/perf/util/trace-event.h +++ b/tools/perf/util/trace-event.h @@ -29,6 +29,8 @@ enum format_flags { FIELD_IS_SIGNED = 4, FIELD_IS_STRING = 8, FIELD_IS_DYNAMIC = 16, + FIELD_IS_FLAG = 32, + FIELD_IS_SYMBOLIC = 64, }; struct format_field { @@ -243,12 +245,19 @@ extern int latency_format; int parse_header_page(char *buf, unsigned long size); int trace_parse_common_type(void *data); +int trace_parse_common_pid(void *data); +int parse_common_pc(void *data); +int parse_common_flags(void *data); +int parse_common_lock_depth(void *data); struct event *trace_find_event(int id); +struct event *trace_find_next_event(struct event *event); +unsigned long long read_size(void *ptr, int size); unsigned long long raw_field_value(struct event *event, const char *name, void *data); void *raw_field_ptr(struct event *event, const char *name, void *data); +unsigned long long eval_flag(const char *flag); -void read_tracing_data(int fd, struct perf_event_attr *pattrs, int nb_events); +int read_tracing_data(int fd, struct perf_event_attr *pattrs, int nb_events); /* taken from kernel/trace/trace.h */ enum trace_flag_type { @@ -259,4 +268,18 @@ enum trace_flag_type { TRACE_FLAG_SOFTIRQ = 0x10, }; +struct scripting_ops { + const char *name; + int (*start_script) (const char *); + int (*stop_script) (void); + void (*process_event) (int cpu, void *data, int size, + unsigned long long nsecs, char *comm); + int (*generate_script) (const char *outfile); +}; + +int script_spec_register(const char *spec, struct scripting_ops *ops); + +extern struct scripting_ops perl_scripting_ops; +void setup_perl_scripting(void); + #endif /* __PERF_TRACE_EVENTS_H */ diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h index f2203a0946b..c673d882588 100644 --- a/tools/perf/util/util.h +++ b/tools/perf/util/util.h @@ -84,6 +84,9 @@ #include <iconv.h> #endif +extern const char *graph_line; +extern const char *graph_dotted_line; + /* On most systems <limits.h> would have given us this, but * not on some systems (e.g. GNU/Hurd). */ @@ -287,17 +290,15 @@ static inline char *gitstrchrnul(const char *s, int c) * Wrappers: */ extern char *xstrdup(const char *str); -extern void *xmalloc(size_t size); +extern void *xmalloc(size_t size) __attribute__((weak)); extern void *xmemdupz(const void *data, size_t len); extern char *xstrndup(const char *str, size_t len); -extern void *xrealloc(void *ptr, size_t size); -extern void *xcalloc(size_t nmemb, size_t size); -extern void *xmmap(void *start, size_t length, int prot, int flags, int fd, off_t offset); -extern ssize_t xread(int fd, void *buf, size_t len); -extern ssize_t xwrite(int fd, const void *buf, size_t len); -extern int xdup(int fd); -extern FILE *xfdopen(int fd, const char *mode); -extern int xmkstemp(char *template); +extern void *xrealloc(void *ptr, size_t size) __attribute__((weak)); + +static inline void *zalloc(size_t size) +{ + return calloc(1, size); +} static inline size_t xsize_t(off_t len) { diff --git a/tools/perf/util/wrapper.c b/tools/perf/util/wrapper.c index 4574ac28396..bf44ca85d23 100644 --- a/tools/perf/util/wrapper.c +++ b/tools/perf/util/wrapper.c @@ -79,43 +79,12 @@ void *xrealloc(void *ptr, size_t size) return ret; } -void *xcalloc(size_t nmemb, size_t size) -{ - void *ret = calloc(nmemb, size); - if (!ret && (!nmemb || !size)) - ret = calloc(1, 1); - if (!ret) { - release_pack_memory(nmemb * size, -1); - ret = calloc(nmemb, size); - if (!ret && (!nmemb || !size)) - ret = calloc(1, 1); - if (!ret) - die("Out of memory, calloc failed"); - } - return ret; -} - -void *xmmap(void *start, size_t length, - int prot, int flags, int fd, off_t offset) -{ - void *ret = mmap(start, length, prot, flags, fd, offset); - if (ret == MAP_FAILED) { - if (!length) - return NULL; - release_pack_memory(length, fd); - ret = mmap(start, length, prot, flags, fd, offset); - if (ret == MAP_FAILED) - die("Out of memory? mmap failed: %s", strerror(errno)); - } - return ret; -} - /* * xread() is the same a read(), but it automatically restarts read() * operations with a recoverable error (EAGAIN and EINTR). xread() * DOES NOT GUARANTEE that "len" bytes is read even if the data is available. */ -ssize_t xread(int fd, void *buf, size_t len) +static ssize_t xread(int fd, void *buf, size_t len) { ssize_t nr; while (1) { @@ -131,7 +100,7 @@ ssize_t xread(int fd, void *buf, size_t len) * operations with a recoverable error (EAGAIN and EINTR). xwrite() DOES NOT * GUARANTEE that "len" bytes is written even if the operation is successful. */ -ssize_t xwrite(int fd, const void *buf, size_t len) +static ssize_t xwrite(int fd, const void *buf, size_t len) { ssize_t nr; while (1) { @@ -179,29 +148,3 @@ ssize_t write_in_full(int fd, const void *buf, size_t count) return total; } - -int xdup(int fd) -{ - int ret = dup(fd); - if (ret < 0) - die("dup failed: %s", strerror(errno)); - return ret; -} - -FILE *xfdopen(int fd, const char *mode) -{ - FILE *stream = fdopen(fd, mode); - if (stream == NULL) - die("Out of memory? fdopen failed: %s", strerror(errno)); - return stream; -} - -int xmkstemp(char *template) -{ - int fd; - - fd = mkstemp(template); - if (fd < 0) - die("Unable to create temporary file: %s", strerror(errno)); - return fd; -} |