X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=include%2Flinux%2Fperf_counter.h;h=5e970c7d3fd5b1c484ddc621dbbbc29921071926;hb=fa172f40068b50f0ad7ae352a2466d0acc579e00;hp=20cf5af27ade7ce20886a791ed3b3be5d778c32e;hpb=8be6e8f3c3a13900169f1141870562d0c723b010;p=safe%2Fjmp%2Flinux-2.6 diff --git a/include/linux/perf_counter.h b/include/linux/perf_counter.h index 20cf5af..5e970c7 100644 --- a/include/linux/perf_counter.h +++ b/include/linux/perf_counter.h @@ -1,12 +1,13 @@ /* * Performance counters: * - * Copyright(C) 2008, Thomas Gleixner - * Copyright(C) 2008, Red Hat, Inc., Ingo Molnar + * Copyright (C) 2008-2009, Thomas Gleixner + * Copyright (C) 2008-2009, Red Hat, Inc., Ingo Molnar + * Copyright (C) 2008-2009, Red Hat, Inc., Peter Zijlstra * * Data type definitions, declarations, prototypes. * - * Started by: Thomas Gleixner and Ingo Molnar + * Started by: Thomas Gleixner and Ingo Molnar * * For licencing details see kernel-base/COPYING */ @@ -25,18 +26,19 @@ * attr.type */ enum perf_type_id { - PERF_TYPE_HARDWARE = 0, - PERF_TYPE_SOFTWARE = 1, - PERF_TYPE_TRACEPOINT = 2, - PERF_TYPE_HW_CACHE = 3, - PERF_TYPE_RAW = 4, + PERF_TYPE_HARDWARE = 0, + PERF_TYPE_SOFTWARE = 1, + PERF_TYPE_TRACEPOINT = 2, + PERF_TYPE_HW_CACHE = 3, + PERF_TYPE_RAW = 4, - PERF_TYPE_MAX, /* non ABI */ + PERF_TYPE_MAX, /* non-ABI */ }; /* - * Generalized performance counter event types, used by the attr.event_id - * parameter of the sys_perf_counter_open() syscall: + * Generalized performance counter event types, used by the + * attr.event_id parameter of the sys_perf_counter_open() + * syscall: */ enum perf_hw_id { /* @@ -50,7 +52,7 @@ enum perf_hw_id { PERF_COUNT_HW_BRANCH_MISSES = 5, PERF_COUNT_HW_BUS_CYCLES = 6, - PERF_COUNT_HW_MAX, /* non ABI */ + PERF_COUNT_HW_MAX, /* non-ABI */ }; /* @@ -61,29 +63,29 @@ enum perf_hw_id { * { accesses, misses } */ enum perf_hw_cache_id { - PERF_COUNT_HW_CACHE_L1D = 0, - PERF_COUNT_HW_CACHE_L1I = 1, - PERF_COUNT_HW_CACHE_LL = 2, - PERF_COUNT_HW_CACHE_DTLB = 3, - PERF_COUNT_HW_CACHE_ITLB = 4, - PERF_COUNT_HW_CACHE_BPU = 5, - - PERF_COUNT_HW_CACHE_MAX, /* non ABI */ + PERF_COUNT_HW_CACHE_L1D = 0, + PERF_COUNT_HW_CACHE_L1I = 1, + PERF_COUNT_HW_CACHE_LL = 2, + PERF_COUNT_HW_CACHE_DTLB = 3, + PERF_COUNT_HW_CACHE_ITLB = 4, + PERF_COUNT_HW_CACHE_BPU = 5, + + PERF_COUNT_HW_CACHE_MAX, /* non-ABI */ }; enum perf_hw_cache_op_id { - PERF_COUNT_HW_CACHE_OP_READ = 0, - PERF_COUNT_HW_CACHE_OP_WRITE = 1, - PERF_COUNT_HW_CACHE_OP_PREFETCH = 2, + PERF_COUNT_HW_CACHE_OP_READ = 0, + PERF_COUNT_HW_CACHE_OP_WRITE = 1, + PERF_COUNT_HW_CACHE_OP_PREFETCH = 2, - PERF_COUNT_HW_CACHE_OP_MAX, /* non ABI */ + PERF_COUNT_HW_CACHE_OP_MAX, /* non-ABI */ }; enum perf_hw_cache_op_result_id { PERF_COUNT_HW_CACHE_RESULT_ACCESS = 0, PERF_COUNT_HW_CACHE_RESULT_MISS = 1, - PERF_COUNT_HW_CACHE_RESULT_MAX, /* non ABI */ + PERF_COUNT_HW_CACHE_RESULT_MAX, /* non-ABI */ }; /* @@ -93,15 +95,15 @@ enum perf_hw_cache_op_result_id { * well): */ enum perf_sw_ids { - PERF_COUNT_SW_CPU_CLOCK = 0, - PERF_COUNT_SW_TASK_CLOCK = 1, - PERF_COUNT_SW_PAGE_FAULTS = 2, - PERF_COUNT_SW_CONTEXT_SWITCHES = 3, - PERF_COUNT_SW_CPU_MIGRATIONS = 4, - PERF_COUNT_SW_PAGE_FAULTS_MIN = 5, - PERF_COUNT_SW_PAGE_FAULTS_MAJ = 6, - - PERF_COUNT_SW_MAX, /* non ABI */ + PERF_COUNT_SW_CPU_CLOCK = 0, + PERF_COUNT_SW_TASK_CLOCK = 1, + PERF_COUNT_SW_PAGE_FAULTS = 2, + PERF_COUNT_SW_CONTEXT_SWITCHES = 3, + PERF_COUNT_SW_CPU_MIGRATIONS = 4, + PERF_COUNT_SW_PAGE_FAULTS_MIN = 5, + PERF_COUNT_SW_PAGE_FAULTS_MAJ = 6, + + PERF_COUNT_SW_MAX, /* non-ABI */ }; /* @@ -109,15 +111,17 @@ enum perf_sw_ids { * in the overflow packets. */ enum perf_counter_sample_format { - PERF_SAMPLE_IP = 1U << 0, - PERF_SAMPLE_TID = 1U << 1, - PERF_SAMPLE_TIME = 1U << 2, - PERF_SAMPLE_ADDR = 1U << 3, - PERF_SAMPLE_GROUP = 1U << 4, - PERF_SAMPLE_CALLCHAIN = 1U << 5, - PERF_SAMPLE_ID = 1U << 6, - PERF_SAMPLE_CPU = 1U << 7, - PERF_SAMPLE_PERIOD = 1U << 8, + PERF_SAMPLE_IP = 1U << 0, + PERF_SAMPLE_TID = 1U << 1, + PERF_SAMPLE_TIME = 1U << 2, + PERF_SAMPLE_ADDR = 1U << 3, + PERF_SAMPLE_GROUP = 1U << 4, + PERF_SAMPLE_CALLCHAIN = 1U << 5, + PERF_SAMPLE_ID = 1U << 6, + PERF_SAMPLE_CPU = 1U << 7, + PERF_SAMPLE_PERIOD = 1U << 8, + + PERF_SAMPLE_MAX = 1U << 9, /* non-ABI */ }; /* @@ -126,20 +130,29 @@ enum perf_counter_sample_format { * in increasing order of bit value, after the counter value. */ enum perf_counter_read_format { - PERF_FORMAT_TOTAL_TIME_ENABLED = 1U << 0, - PERF_FORMAT_TOTAL_TIME_RUNNING = 1U << 1, - PERF_FORMAT_ID = 1U << 2, + PERF_FORMAT_TOTAL_TIME_ENABLED = 1U << 0, + PERF_FORMAT_TOTAL_TIME_RUNNING = 1U << 1, + PERF_FORMAT_ID = 1U << 2, + + PERF_FORMAT_MAX = 1U << 3, /* non-ABI */ }; +#define PERF_ATTR_SIZE_VER0 64 /* sizeof first published struct */ + /* * Hardware event to monitor via a performance monitoring counter: */ struct perf_counter_attr { + /* * Major type: hardware/software/tracepoint/etc. */ __u32 type; - __u32 __reserved_1; + + /* + * Size of the attr structure, for fwd/bwd compat. + */ + __u32 size; /* * Type specific configuration information. @@ -165,13 +178,15 @@ struct perf_counter_attr { mmap : 1, /* include mmap data */ comm : 1, /* include comm data */ freq : 1, /* use freq, not period */ + inherit_stat : 1, /* per task counts */ + enable_on_exec : 1, /* next exec enables */ - __reserved_2 : 53; + __reserved_1 : 51; __u32 wakeup_events; /* wakeup every n events */ - __u32 __reserved_3; + __u32 __reserved_2; - __u64 __reserved_4; + __u64 __reserved_3; }; /* @@ -219,22 +234,35 @@ struct perf_counter_mmap_page { __u32 lock; /* seqlock for synchronization */ __u32 index; /* hardware counter identifier */ __s64 offset; /* add to hardware counter value */ + __u64 time_enabled; /* time counter active */ + __u64 time_running; /* time counter on cpu */ + + /* + * Hole for extension of the self monitor capabilities + */ + + __u64 __reserved[123]; /* align to 1k */ /* * Control data for the mmap() data buffer. * - * User-space reading this value should issue an rmb(), on SMP capable - * platforms, after reading this value -- see perf_counter_wakeup(). + * User-space reading the @data_head value should issue an rmb(), on + * SMP capable platforms, after reading this value -- see + * perf_counter_wakeup(). + * + * When the mapping is PROT_WRITE the @data_tail value should be + * written by userspace to reflect the last read data. In this case + * the kernel will not over-write unread data. */ __u64 data_head; /* head in the data section */ + __u64 data_tail; /* user-space written tail */ }; -#define PERF_EVENT_MISC_CPUMODE_MASK (3 << 0) -#define PERF_EVENT_MISC_CPUMODE_UNKNOWN (0 << 0) -#define PERF_EVENT_MISC_KERNEL (1 << 0) -#define PERF_EVENT_MISC_USER (2 << 0) -#define PERF_EVENT_MISC_HYPERVISOR (3 << 0) -#define PERF_EVENT_MISC_OVERFLOW (1 << 2) +#define PERF_EVENT_MISC_CPUMODE_MASK (3 << 0) +#define PERF_EVENT_MISC_CPUMODE_UNKNOWN (0 << 0) +#define PERF_EVENT_MISC_KERNEL (1 << 0) +#define PERF_EVENT_MISC_USER (2 << 0) +#define PERF_EVENT_MISC_HYPERVISOR (3 << 0) struct perf_event_header { __u32 type; @@ -262,6 +290,15 @@ enum perf_event_type { /* * struct { + * struct perf_event_header header; + * u64 id; + * u64 lost; + * }; + */ + PERF_EVENT_LOST = 2, + + /* + * struct { * struct perf_event_header header; * * u32 pid, tid; @@ -284,6 +321,7 @@ enum perf_event_type { * struct { * struct perf_event_header header; * u64 time; + * u64 id; * }; */ PERF_EVENT_THROTTLE = 5, @@ -298,29 +336,51 @@ enum perf_event_type { PERF_EVENT_FORK = 7, /* - * When header.misc & PERF_EVENT_MISC_OVERFLOW the event_type field - * will be PERF_RECORD_* - * + * struct { + * struct perf_event_header header; + * u32 pid, tid; + * u64 value; + * { u64 time_enabled; } && PERF_FORMAT_ENABLED + * { u64 time_running; } && PERF_FORMAT_RUNNING + * { u64 parent_id; } && PERF_FORMAT_ID + * }; + */ + PERF_EVENT_READ = 8, + + /* * struct { * struct perf_event_header header; * - * { u64 ip; } && PERF_RECORD_IP - * { u32 pid, tid; } && PERF_RECORD_TID - * { u64 time; } && PERF_RECORD_TIME - * { u64 addr; } && PERF_RECORD_ADDR - * { u64 config; } && PERF_RECORD_CONFIG - * { u32 cpu, res; } && PERF_RECORD_CPU + * { u64 ip; } && PERF_SAMPLE_IP + * { u32 pid, tid; } && PERF_SAMPLE_TID + * { u64 time; } && PERF_SAMPLE_TIME + * { u64 addr; } && PERF_SAMPLE_ADDR + * { u64 id; } && PERF_SAMPLE_ID + * { u32 cpu, res; } && PERF_SAMPLE_CPU + * { u64 period; } && PERF_SAMPLE_PERIOD * * { u64 nr; - * { u64 id, val; } cnt[nr]; } && PERF_RECORD_GROUP + * { u64 id, val; } cnt[nr]; } && PERF_SAMPLE_GROUP * - * { u16 nr, - * hv, - * kernel, - * user; - * u64 ips[nr]; } && PERF_RECORD_CALLCHAIN + * { u64 nr, + * u64 ips[nr]; } && PERF_SAMPLE_CALLCHAIN * }; */ + PERF_EVENT_SAMPLE = 9, + + PERF_EVENT_MAX, /* non-ABI */ +}; + +enum perf_callchain_context { + PERF_CONTEXT_HV = (__u64)-32, + PERF_CONTEXT_KERNEL = (__u64)-128, + PERF_CONTEXT_USER = (__u64)-512, + + PERF_CONTEXT_GUEST = (__u64)-2048, + PERF_CONTEXT_GUEST_KERNEL = (__u64)-2176, + PERF_CONTEXT_GUEST_USER = (__u64)-2560, + + PERF_CONTEXT_MAX = (__u64)-4095, }; #ifdef __KERNEL__ @@ -342,6 +402,13 @@ enum perf_event_type { #include #include +#define PERF_MAX_STACK_DEPTH 255 + +struct perf_callchain_entry { + __u64 nr; + __u64 ip[PERF_MAX_STACK_DEPTH]; +}; + struct task_struct; /** @@ -351,14 +418,14 @@ struct hw_perf_counter { #ifdef CONFIG_PERF_COUNTERS union { struct { /* hardware */ - u64 config; - unsigned long config_base; - unsigned long counter_base; - int idx; + u64 config; + unsigned long config_base; + unsigned long counter_base; + int idx; }; union { /* software */ - atomic64_t count; - struct hrtimer hrtimer; + atomic64_t count; + struct hrtimer hrtimer; }; }; atomic64_t prev_count; @@ -400,6 +467,7 @@ struct file; struct perf_mmap_data { struct rcu_head rcu_head; int nr_pages; /* nr of data pages */ + int writable; /* are we writable */ int nr_locked; /* nr pages mlocked */ atomic_t poll; /* POLL_ for wakeups */ @@ -409,8 +477,8 @@ struct perf_mmap_data { atomic_long_t done_head; /* completed head */ atomic_t lock; /* concurrent writes */ - atomic_t wakeup; /* needs a wakeup */ + atomic_t lost; /* nr records lost */ struct perf_counter_mmap_page *user_page; void *data_pages[0]; @@ -523,37 +591,38 @@ struct perf_counter_context { * Protect the states of the counters in the list, * nr_active, and the list: */ - spinlock_t lock; + spinlock_t lock; /* * Protect the list of counters. Locking either mutex or lock * is sufficient to ensure the list doesn't change; to change * the list you need to lock both the mutex and the spinlock. */ - struct mutex mutex; + struct mutex mutex; - struct list_head counter_list; - struct list_head event_list; - int nr_counters; - int nr_active; - int is_active; - atomic_t refcount; - struct task_struct *task; + struct list_head counter_list; + struct list_head event_list; + int nr_counters; + int nr_active; + int is_active; + int nr_stat; + atomic_t refcount; + struct task_struct *task; /* * Context clock, runs when context enabled. */ - u64 time; - u64 timestamp; + u64 time; + u64 timestamp; /* * These fields let us detect when two contexts have both * been cloned (inherited) from a common ancestor. */ - struct perf_counter_context *parent_ctx; - u64 parent_gen; - u64 generation; - int pin_count; - struct rcu_head rcu_head; + struct perf_counter_context *parent_ctx; + u64 parent_gen; + u64 generation; + int pin_count; + struct rcu_head rcu_head; }; /** @@ -590,6 +659,7 @@ extern void perf_counter_task_tick(struct task_struct *task, int cpu); extern int perf_counter_init_task(struct task_struct *child); extern void perf_counter_exit_task(struct task_struct *child); extern void perf_counter_free_task(struct task_struct *task); +extern void set_perf_counter_pending(void); extern void perf_counter_do_pending(void); extern void perf_counter_print_debug(void); extern void __perf_disable(void); @@ -604,9 +674,9 @@ extern int hw_perf_group_sched_in(struct perf_counter *group_leader, extern void perf_counter_update_userpage(struct perf_counter *counter); struct perf_sample_data { - struct pt_regs *regs; - u64 addr; - u64 period; + struct pt_regs *regs; + u64 addr; + u64 period; }; extern int perf_counter_overflow(struct perf_counter *counter, int nmi, @@ -618,10 +688,20 @@ extern int perf_counter_overflow(struct perf_counter *counter, int nmi, static inline int is_software_counter(struct perf_counter *counter) { return (counter->attr.type != PERF_TYPE_RAW) && - (counter->attr.type != PERF_TYPE_HARDWARE); + (counter->attr.type != PERF_TYPE_HARDWARE) && + (counter->attr.type != PERF_TYPE_HW_CACHE); } -extern void perf_swcounter_event(u32, u64, int, struct pt_regs *, u64); +extern atomic_t perf_swcounter_enabled[PERF_COUNT_SW_MAX]; + +extern void __perf_swcounter_event(u32, u64, int, struct pt_regs *, u64); + +static inline void +perf_swcounter_event(u32 event, u64 nr, int nmi, struct pt_regs *regs, u64 addr) +{ + if (atomic_read(&perf_swcounter_enabled[event])) + __perf_swcounter_event(event, nr, nmi, regs, addr); +} extern void __perf_counter_mmap(struct vm_area_struct *vma); @@ -634,15 +714,6 @@ static inline void perf_counter_mmap(struct vm_area_struct *vma) extern void perf_counter_comm(struct task_struct *tsk); extern void perf_counter_fork(struct task_struct *tsk); -extern void perf_counter_task_migration(struct task_struct *task, int cpu); - -#define MAX_STACK_DEPTH 255 - -struct perf_callchain_entry { - u16 nr, hv, kernel, user; - u64 ip[MAX_STACK_DEPTH]; -}; - extern struct perf_callchain_entry *perf_callchain(struct pt_regs *regs); extern int sysctl_perf_counter_paranoid; @@ -683,8 +754,6 @@ static inline void perf_counter_mmap(struct vm_area_struct *vma) { } static inline void perf_counter_comm(struct task_struct *tsk) { } static inline void perf_counter_fork(struct task_struct *tsk) { } static inline void perf_counter_init(void) { } -static inline void perf_counter_task_migration(struct task_struct *task, - int cpu) { } #endif #endif /* __KERNEL__ */