diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2010-10-21 12:54:49 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2010-10-21 12:54:49 -0700 |
commit | 5d70f79b5ef6ea2de4f72a37b2d96e2601e40a22 (patch) | |
tree | a0d6de0930ba83ecf4629c2e2e261f5eaa2d8f33 | |
parent | 888a6f77e0418b049f83d37547c209b904d30af4 (diff) | |
parent | 750ed158bf6c782d2813da1bca2c824365a0b777 (diff) | |
download | kernel_samsung_tuna-5d70f79b5ef6ea2de4f72a37b2d96e2601e40a22.zip kernel_samsung_tuna-5d70f79b5ef6ea2de4f72a37b2d96e2601e40a22.tar.gz kernel_samsung_tuna-5d70f79b5ef6ea2de4f72a37b2d96e2601e40a22.tar.bz2 |
Merge branch 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (163 commits)
tracing: Fix compile issue for trace_sched_wakeup.c
[S390] hardirq: remove pointless header file includes
[IA64] Move local_softirq_pending() definition
perf, powerpc: Fix power_pmu_event_init to not use event->ctx
ftrace: Remove recursion between recordmcount and scripts/mod/empty
jump_label: Add COND_STMT(), reducer wrappery
perf: Optimize sw events
perf: Use jump_labels to optimize the scheduler hooks
jump_label: Add atomic_t interface
jump_label: Use more consistent naming
perf, hw_breakpoint: Fix crash in hw_breakpoint creation
perf: Find task before event alloc
perf: Fix task refcount bugs
perf: Fix group moving
irq_work: Add generic hardirq context callbacks
perf_events: Fix transaction recovery in group_sched_in()
perf_events: Fix bogus AMD64 generic TLB events
perf_events: Fix bogus context time tracking
tracing: Remove parent recording in latency tracer graph options
tracing: Use one prologue for the preempt irqs off tracer function tracers
...
158 files changed, 7162 insertions, 3368 deletions
diff --git a/Documentation/kprobes.txt b/Documentation/kprobes.txt index 1762b81..741fe66 100644 --- a/Documentation/kprobes.txt +++ b/Documentation/kprobes.txt @@ -542,9 +542,11 @@ Kprobes does not use mutexes or allocate memory except during registration and unregistration. Probe handlers are run with preemption disabled. Depending on the -architecture, handlers may also run with interrupts disabled. In any -case, your handler should not yield the CPU (e.g., by attempting to -acquire a semaphore). +architecture and optimization state, handlers may also run with +interrupts disabled (e.g., kretprobe handlers and optimized kprobe +handlers run without interrupt disabled on x86/x86-64). In any case, +your handler should not yield the CPU (e.g., by attempting to acquire +a semaphore). Since a return probe is implemented by replacing the return address with the trampoline's address, stack backtraces and calls @@ -568,6 +568,12 @@ endif ifdef CONFIG_FUNCTION_TRACER KBUILD_CFLAGS += -pg +ifdef CONFIG_DYNAMIC_FTRACE + ifdef CONFIG_HAVE_C_RECORDMCOUNT + BUILD_C_RECORDMCOUNT := y + export BUILD_C_RECORDMCOUNT + endif +endif endif # We trigger additional mismatches with less inlining @@ -591,6 +597,11 @@ KBUILD_CFLAGS += $(call cc-option,-fno-strict-overflow) # conserve stack if available KBUILD_CFLAGS += $(call cc-option,-fconserve-stack) +# check for 'asm goto' +ifeq ($(shell $(CONFIG_SHELL) $(srctree)/scripts/gcc-goto.sh $(CC)), y) + KBUILD_CFLAGS += -DCC_HAVE_ASM_GOTO +endif + # Add user supplied CPPFLAGS, AFLAGS and CFLAGS as the last assignments # But warn user when we do so warn-assign = \ diff --git a/arch/Kconfig b/arch/Kconfig index fe48fc7..53d7f61 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -158,4 +158,7 @@ config HAVE_PERF_EVENTS_NMI subsystem. Also has support for calculating CPU cycle events to determine how many clock cycles in a given period. +config HAVE_ARCH_JUMP_LABEL + bool + source "kernel/gcov/Kconfig" diff --git a/arch/alpha/Kconfig b/arch/alpha/Kconfig index b9647bb..d04ccd7 100644 --- a/arch/alpha/Kconfig +++ b/arch/alpha/Kconfig @@ -9,6 +9,7 @@ config ALPHA select HAVE_IDE select HAVE_OPROFILE select HAVE_SYSCALL_WRAPPERS + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select HAVE_DMA_ATTRS help diff --git a/arch/alpha/include/asm/perf_event.h b/arch/alpha/include/asm/perf_event.h index 4157cd3..fe792ca 100644 --- a/arch/alpha/include/asm/perf_event.h +++ b/arch/alpha/include/asm/perf_event.h @@ -1,11 +1,6 @@ #ifndef __ASM_ALPHA_PERF_EVENT_H #define __ASM_ALPHA_PERF_EVENT_H -/* Alpha only supports software events through this interface. */ -extern void set_perf_event_pending(void); - -#define PERF_EVENT_INDEX_OFFSET 0 - #ifdef CONFIG_PERF_EVENTS extern void init_hw_perf_events(void); #else diff --git a/arch/alpha/kernel/perf_event.c b/arch/alpha/kernel/perf_event.c index 85d8e4f..1cc4968 100644 --- a/arch/alpha/kernel/perf_event.c +++ b/arch/alpha/kernel/perf_event.c @@ -307,7 +307,7 @@ again: new_raw_count) != prev_raw_count) goto again; - delta = (new_raw_count - (prev_raw_count & alpha_pmu->pmc_count_mask[idx])) + ovf; + delta = (new_raw_count - (prev_raw_count & alpha_pmu->pmc_count_mask[idx])) + ovf; /* It is possible on very rare occasions that the PMC has overflowed * but the interrupt is yet to come. Detect and fix this situation. @@ -402,14 +402,13 @@ static void maybe_change_configuration(struct cpu_hw_events *cpuc) struct hw_perf_event *hwc = &pe->hw; int idx = hwc->idx; - if (cpuc->current_idx[j] != PMC_NO_INDEX) { - cpuc->idx_mask |= (1<<cpuc->current_idx[j]); - continue; + if (cpuc->current_idx[j] == PMC_NO_INDEX) { + alpha_perf_event_set_period(pe, hwc, idx); + cpuc->current_idx[j] = idx; } - alpha_perf_event_set_period(pe, hwc, idx); - cpuc->current_idx[j] = idx; - cpuc->idx_mask |= (1<<cpuc->current_idx[j]); + if (!(hwc->state & PERF_HES_STOPPED)) + cpuc->idx_mask |= (1<<cpuc->current_idx[j]); } cpuc->config = cpuc->event[0]->hw.config_base; } @@ -420,12 +419,13 @@ static void maybe_change_configuration(struct cpu_hw_events *cpuc) * - this function is called from outside this module via the pmu struct * returned from perf event initialisation. */ -static int alpha_pmu_enable(struct perf_event *event) +static int alpha_pmu_add(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct hw_perf_event *hwc = &event->hw; int n0; int ret; - unsigned long flags; + unsigned long irq_flags; /* * The Sparc code has the IRQ disable first followed by the perf @@ -435,8 +435,8 @@ static int alpha_pmu_enable(struct perf_event *event) * nevertheless we disable the PMCs first to enable a potential * final PMI to occur before we disable interrupts. */ - perf_disable(); - local_irq_save(flags); + perf_pmu_disable(event->pmu); + local_irq_save(irq_flags); /* Default to error to be returned */ ret = -EAGAIN; @@ -455,8 +455,12 @@ static int alpha_pmu_enable(struct perf_event *event) } } - local_irq_restore(flags); - perf_enable(); + hwc->state = PERF_HES_UPTODATE; + if (!(flags & PERF_EF_START)) + hwc->state |= PERF_HES_STOPPED; + + local_irq_restore(irq_flags); + perf_pmu_enable(event->pmu); return ret; } @@ -467,15 +471,15 @@ static int alpha_pmu_enable(struct perf_event *event) * - this function is called from outside this module via the pmu struct * returned from perf event initialisation. */ -static void alpha_pmu_disable(struct perf_event *event) +static void alpha_pmu_del(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc = &event->hw; - unsigned long flags; + unsigned long irq_flags; int j; - perf_disable(); - local_irq_save(flags); + perf_pmu_disable(event->pmu); + local_irq_save(irq_flags); for (j = 0; j < cpuc->n_events; j++) { if (event == cpuc->event[j]) { @@ -501,8 +505,8 @@ static void alpha_pmu_disable(struct perf_event *event) } } - local_irq_restore(flags); - perf_enable(); + local_irq_restore(irq_flags); + perf_pmu_enable(event->pmu); } @@ -514,13 +518,44 @@ static void alpha_pmu_read(struct perf_event *event) } -static void alpha_pmu_unthrottle(struct perf_event *event) +static void alpha_pmu_stop(struct perf_event *event, int flags) +{ + struct hw_perf_event *hwc = &event->hw; + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + + if (!(hwc->state & PERF_HES_STOPPED)) { + cpuc->idx_mask &= ~(1UL<<hwc->idx); + hwc->state |= PERF_HES_STOPPED; + } + + if ((flags & PERF_EF_UPDATE) && !(hwc->state & PERF_HES_UPTODATE)) { + alpha_perf_event_update(event, hwc, hwc->idx, 0); + hwc->state |= PERF_HES_UPTODATE; + } + + if (cpuc->enabled) + wrperfmon(PERFMON_CMD_DISABLE, (1UL<<hwc->idx)); +} + + +static void alpha_pmu_start(struct perf_event *event, int flags) { struct hw_perf_event *hwc = &event->hw; struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + if (WARN_ON_ONCE(!(hwc->state & PERF_HES_STOPPED))) + return; + + if (flags & PERF_EF_RELOAD) { + WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE)); + alpha_perf_event_set_period(event, hwc, hwc->idx); + } + + hwc->state = 0; + cpuc->idx_mask |= 1UL<<hwc->idx; - wrperfmon(PERFMON_CMD_ENABLE, (1UL<<hwc->idx)); + if (cpuc->enabled) + wrperfmon(PERFMON_CMD_ENABLE, (1UL<<hwc->idx)); } @@ -642,39 +677,36 @@ static int __hw_perf_event_init(struct perf_event *event) return 0; } -static const struct pmu pmu = { - .enable = alpha_pmu_enable, - .disable = alpha_pmu_disable, - .read = alpha_pmu_read, - .unthrottle = alpha_pmu_unthrottle, -}; - - /* * Main entry point to initialise a HW performance event. */ -const struct pmu *hw_perf_event_init(struct perf_event *event) +static int alpha_pmu_event_init(struct perf_event *event) { int err; + switch (event->attr.type) { + case PERF_TYPE_RAW: + case PERF_TYPE_HARDWARE: + case PERF_TYPE_HW_CACHE: + break; + + default: + return -ENOENT; + } + if (!alpha_pmu) - return ERR_PTR(-ENODEV); + return -ENODEV; /* Do the real initialisation work. */ err = __hw_perf_event_init(event); - if (err) - return ERR_PTR(err); - - return &pmu; + return err; } - - /* * Main entry point - enable HW performance counters. */ -void hw_perf_enable(void) +static void alpha_pmu_enable(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); @@ -700,7 +732,7 @@ void hw_perf_enable(void) * Main entry point - disable HW performance counters. */ -void hw_perf_disable(void) +static void alpha_pmu_disable(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); @@ -713,6 +745,17 @@ void hw_perf_disable(void) wrperfmon(PERFMON_CMD_DISABLE, cpuc->idx_mask); } +static struct pmu pmu = { + .pmu_enable = alpha_pmu_enable, + .pmu_disable = alpha_pmu_disable, + .event_init = alpha_pmu_event_init, + .add = alpha_pmu_add, + .del = alpha_pmu_del, + .start = alpha_pmu_start, + .stop = alpha_pmu_stop, + .read = alpha_pmu_read, +}; + /* * Main entry point - don't know when this is called but it @@ -766,7 +809,7 @@ static void alpha_perf_event_irq_handler(unsigned long la_ptr, wrperfmon(PERFMON_CMD_DISABLE, cpuc->idx_mask); /* la_ptr is the counter that overflowed. */ - if (unlikely(la_ptr >= perf_max_events)) { + if (unlikely(la_ptr >= alpha_pmu->num_pmcs)) { /* This should never occur! */ irq_err_count++; pr_warning("PMI: silly index %ld\n", la_ptr); @@ -807,7 +850,7 @@ static void alpha_perf_event_irq_handler(unsigned long la_ptr, /* Interrupts coming too quickly; "throttle" the * counter, i.e., disable it for a little while. */ - cpuc->idx_mask &= ~(1UL<<idx); + alpha_pmu_stop(event, 0); } } wrperfmon(PERFMON_CMD_ENABLE, cpuc->idx_mask); @@ -837,6 +880,7 @@ void __init init_hw_perf_events(void) /* And set up PMU specification */ alpha_pmu = &ev67_pmu; - perf_max_events = alpha_pmu->num_pmcs; + + perf_pmu_register(&pmu); } diff --git a/arch/alpha/kernel/time.c b/arch/alpha/kernel/time.c index 396af17..0f1d849 100644 --- a/arch/alpha/kernel/time.c +++ b/arch/alpha/kernel/time.c @@ -41,7 +41,7 @@ #include <linux/init.h> #include <linux/bcd.h> #include <linux/profile.h> -#include <linux/perf_event.h> +#include <linux/irq_work.h> #include <asm/uaccess.h> #include <asm/io.h> @@ -83,25 +83,25 @@ static struct { unsigned long est_cycle_freq; -#ifdef CONFIG_PERF_EVENTS +#ifdef CONFIG_IRQ_WORK -DEFINE_PER_CPU(u8, perf_event_pending); +DEFINE_PER_CPU(u8, irq_work_pending); -#define set_perf_event_pending_flag() __get_cpu_var(perf_event_pending) = 1 -#define test_perf_event_pending() __get_cpu_var(perf_event_pending) -#define clear_perf_event_pending() __get_cpu_var(perf_event_pending) = 0 +#define set_irq_work_pending_flag() __get_cpu_var(irq_work_pending) = 1 +#define test_irq_work_pending() __get_cpu_var(irq_work_pending) +#define clear_irq_work_pending() __get_cpu_var(irq_work_pending) = 0 -void set_perf_event_pending(void) +void set_irq_work_pending(void) { - set_perf_event_pending_flag(); + set_irq_work_pending_flag(); } -#else /* CONFIG_PERF_EVENTS */ +#else /* CONFIG_IRQ_WORK */ -#define test_perf_event_pending() 0 -#define clear_perf_event_pending() +#define test_irq_work_pending() 0 +#define clear_irq_work_pending() -#endif /* CONFIG_PERF_EVENTS */ +#endif /* CONFIG_IRQ_WORK */ static inline __u32 rpcc(void) @@ -191,9 +191,9 @@ irqreturn_t timer_interrupt(int irq, void *dev) write_sequnlock(&xtime_lock); - if (test_perf_event_pending()) { - clear_perf_event_pending(); - perf_event_do_pending(); + if (test_irq_work_pending()) { + clear_irq_work_pending(); + irq_work_run(); } #ifndef CONFIG_SMP diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 9c26ba7..9103904 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -23,6 +23,7 @@ config ARM select HAVE_KERNEL_GZIP select HAVE_KERNEL_LZO select HAVE_KERNEL_LZMA + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC select HAVE_REGS_AND_STACK_ACCESS_API diff --git a/arch/arm/include/asm/perf_event.h b/arch/arm/include/asm/perf_event.h index b5799a3..c4aa4e8 100644 --- a/arch/arm/include/asm/perf_event.h +++ b/arch/arm/include/asm/perf_event.h @@ -12,18 +12,6 @@ #ifndef __ARM_PERF_EVENT_H__ #define __ARM_PERF_EVENT_H__ -/* - * NOP: on *most* (read: all supported) ARM platforms, the performance - * counter interrupts are regular interrupts and not an NMI. This - * means that when we receive the interrupt we can call - * perf_event_do_pending() that handles all of the work with - * interrupts disabled. - */ -static inline void -set_perf_event_pending(void) -{ -} - /* ARM performance counters start from 1 (in the cp15 accesses) so use the * same indexes here for consistency. */ #define PERF_EVENT_INDEX_OFFSET 1 diff --git a/arch/arm/kernel/perf_event.c b/arch/arm/kernel/perf_event.c index ecbb028..49643b1 100644 --- a/arch/arm/kernel/perf_event.c +++ b/arch/arm/kernel/perf_event.c @@ -123,6 +123,12 @@ armpmu_get_max_events(void) } EXPORT_SYMBOL_GPL(armpmu_get_max_events); +int perf_num_counters(void) +{ + return armpmu_get_max_events(); +} +EXPORT_SYMBOL_GPL(perf_num_counters); + #define HW_OP_UNSUPPORTED 0xFFFF #define C(_x) \ @@ -221,46 +227,56 @@ again: } static void -armpmu_disable(struct perf_event *event) +armpmu_read(struct perf_event *event) { - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc = &event->hw; - int idx = hwc->idx; - - WARN_ON(idx < 0); - - clear_bit(idx, cpuc->active_mask); - armpmu->disable(hwc, idx); - - barrier(); - armpmu_event_update(event, hwc, idx); - cpuc->events[idx] = NULL; - clear_bit(idx, cpuc->used_mask); + /* Don't read disabled counters! */ + if (hwc->idx < 0) + return; - perf_event_update_userpage(event); + armpmu_event_update(event, hwc, hwc->idx); } static void -armpmu_read(struct perf_event *event) +armpmu_stop(struct perf_event *event, int flags) { struct hw_perf_event *hwc = &event->hw; - /* Don't read disabled counters! */ - if (hwc->idx < 0) + if (!armpmu) return; - armpmu_event_update(event, hwc, hwc->idx); + /* + * ARM pmu always has to update the counter, so ignore + * PERF_EF_UPDATE, see comments in armpmu_start(). + */ + if (!(hwc->state & PERF_HES_STOPPED)) { + armpmu->disable(hwc, hwc->idx); + barrier(); /* why? */ + armpmu_event_update(event, hwc, hwc->idx); + hwc->state |= PERF_HES_STOPPED | PERF_HES_UPTODATE; + } } static void -armpmu_unthrottle(struct perf_event *event) +armpmu_start(struct perf_event *event, int flags) { struct hw_perf_event *hwc = &event->hw; + if (!armpmu) + return; + + /* + * ARM pmu always has to reprogram the period, so ignore + * PERF_EF_RELOAD, see the comment below. + */ + if (flags & PERF_EF_RELOAD) + WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE)); + + hwc->state = 0; /* * Set the period again. Some counters can't be stopped, so when we - * were throttled we simply disabled the IRQ source and the counter + * were stopped we simply disabled the IRQ source and the counter * may have been left counting. If we don't do this step then we may * get an interrupt too soon or *way* too late if the overflow has * happened since disabling. @@ -269,14 +285,33 @@ armpmu_unthrottle(struct perf_event *event) armpmu->enable(hwc, hwc->idx); } +static void +armpmu_del(struct perf_event *event, int flags) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct hw_perf_event *hwc = &event->hw; + int idx = hwc->idx; + + WARN_ON(idx < 0); + + clear_bit(idx, cpuc->active_mask); + armpmu_stop(event, PERF_EF_UPDATE); + cpuc->events[idx] = NULL; + clear_bit(idx, cpuc->used_mask); + + perf_event_update_userpage(event); +} + static int -armpmu_enable(struct perf_event *event) +armpmu_add(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx; int err = 0; + perf_pmu_disable(event->pmu); + /* If we don't have a space for the counter then finish early. */ idx = armpmu->get_event_idx(cpuc, hwc); if (idx < 0) { @@ -293,25 +328,19 @@ armpmu_enable(struct perf_event *event) cpuc->events[idx] = event; set_bit(idx, cpuc->active_mask); - /* Set the period for the event. */ - armpmu_event_set_period(event, hwc, idx); - - /* Enable the event. */ - armpmu->enable(hwc, idx); + hwc->state = PERF_HES_STOPPED | PERF_HES_UPTODATE; + if (flags & PERF_EF_START) + armpmu_start(event, PERF_EF_RELOAD); /* Propagate our changes to the userspace mapping. */ perf_event_update_userpage(event); out: + perf_pmu_enable(event->pmu); return err; } -static struct pmu pmu = { - .enable = armpmu_enable, - .disable = armpmu_disable, - .unthrottle = armpmu_unthrottle, - .read = armpmu_read, -}; +static struct pmu pmu; static int validate_event(struct cpu_hw_events *cpuc, @@ -491,20 +520,29 @@ __hw_perf_event_init(struct perf_event *event) return err; } -const struct pmu * -hw_perf_event_init(struct perf_event *event) +static int armpmu_event_init(struct perf_event *event) { int err = 0; + switch (event->attr.type) { + case PERF_TYPE_RAW: + case PERF_TYPE_HARDWARE: + case PERF_TYPE_HW_CACHE: + break; + + default: + return -ENOENT; + } + if (!armpmu) - return ERR_PTR(-ENODEV); + return -ENODEV; event->destroy = hw_perf_event_destroy; if (!atomic_inc_not_zero(&active_events)) { - if (atomic_read(&active_events) > perf_max_events) { + if (atomic_read(&active_events) > armpmu->num_events) { atomic_dec(&active_events); - return ERR_PTR(-ENOSPC); + return -ENOSPC; } mutex_lock(&pmu_reserve_mutex); @@ -518,17 +556,16 @@ hw_perf_event_init(struct perf_event *event) } if (err) - return ERR_PTR(err); + return err; err = __hw_perf_event_init(event); if (err) hw_perf_event_destroy(event); - return err ? ERR_PTR(err) : &pmu; + return err; } -void -hw_perf_enable(void) +static void armpmu_enable(struct pmu *pmu) { /* Enable all of the perf events on hardware. */ int idx; @@ -549,13 +586,23 @@ hw_perf_enable(void) armpmu->start(); } -void -hw_perf_disable(void) +static void armpmu_disable(struct pmu *pmu) { if (armpmu) armpmu->stop(); } +static struct pmu pmu = { + .pmu_enable = armpmu_enable, + .pmu_disable = armpmu_disable, + .event_init = armpmu_event_init, + .add = armpmu_add, + .del = armpmu_del, + .start = armpmu_start, + .stop = armpmu_stop, + .read = armpmu_read, +}; + /* * ARMv6 Performance counter handling code. * @@ -1045,7 +1092,7 @@ armv6pmu_handle_irq(int irq_num, * platforms that can have the PMU interrupts raised as an NMI, this * will not work. */ - perf_event_do_pending(); + irq_work_run(); return IRQ_HANDLED; } @@ -2021,7 +2068,7 @@ static irqreturn_t armv7pmu_handle_irq(int irq_num, void *dev) * platforms that can have the PMU interrupts raised as an NMI, this * will not work. */ - perf_event_do_pending(); + irq_work_run(); return IRQ_HANDLED; } @@ -2389,7 +2436,7 @@ xscale1pmu_handle_irq(int irq_num, void *dev) armpmu->disable(hwc, idx); } - perf_event_do_pending(); + irq_work_run(); /* * Re-enable the PMU. @@ -2716,7 +2763,7 @@ xscale2pmu_handle_irq(int irq_num, void *dev) armpmu->disable(hwc, idx); } - perf_event_do_pending(); + irq_work_run(); /* * Re-enable the PMU. @@ -2933,14 +2980,12 @@ init_hw_perf_events(void) armpmu = &armv6pmu; memcpy(armpmu_perf_cache_map, armv6_perf_cache_map, sizeof(armv6_perf_cache_map)); - perf_max_events = armv6pmu.num_events; break; case 0xB020: /* ARM11mpcore */ armpmu = &armv6mpcore_pmu; memcpy(armpmu_perf_cache_map, armv6mpcore_perf_cache_map, sizeof(armv6mpcore_perf_cache_map)); - perf_max_events = armv6mpcore_pmu.num_events; break; case 0xC080: /* Cortex-A8 */ armv7pmu.id = ARM_PERF_PMU_ID_CA8; @@ -2952,7 +2997,6 @@ init_hw_perf_events(void) /* Reset PMNC and read the nb of CNTx counters supported */ armv7pmu.num_events = armv7_reset_read_pmnc(); - perf_max_events = armv7pmu.num_events; break; case 0xC090: /* Cortex-A9 */ armv7pmu.id = ARM_PERF_PMU_ID_CA9; @@ -2964,7 +3008,6 @@ init_hw_perf_events(void) /* Reset PMNC and read the nb of CNTx counters supported */ armv7pmu.num_events = armv7_reset_read_pmnc(); - perf_max_events = armv7pmu.num_events; break; } /* Intel CPUs [xscale]. */ @@ -2975,13 +3018,11 @@ init_hw_perf_events(void) armpmu = &xscale1pmu; memcpy(armpmu_perf_cache_map, xscale_perf_cache_map, sizeof(xscale_perf_cache_map)); - perf_max_events = xscale1pmu.num_events; break; case 2: armpmu = &xscale2pmu; memcpy(armpmu_perf_cache_map, xscale_perf_cache_map, sizeof(xscale_perf_cache_map)); - perf_max_events = xscale2pmu.num_events; break; } } @@ -2991,9 +3032,10 @@ init_hw_perf_events(void) arm_pmu_names[armpmu->id], armpmu->num_events); } else { pr_info("no hardware support available\n"); - perf_max_events = -1; } + perf_pmu_register(&pmu); + return 0; } arch_initcall(init_hw_perf_events); @@ -3001,13 +3043,6 @@ arch_initcall(init_hw_perf_events); /* * Callchain handling code. */ -static inline void -callchain_store(struct perf_callchain_entry *entry, - u64 ip) -{ - if (entry->nr < PERF_MAX_STACK_DEPTH) - entry->ip[entry->nr++] = ip; -} /* * The registers we're interested in are at the end of the variable @@ -3039,7 +3074,7 @@ user_backtrace(struct frame_tail *tail, if (__copy_from_user_inatomic(&buftail, tail, sizeof(buftail))) return NULL; - callchain_store(entry, buftail.lr); + perf_callchain_store(entry, buftail.lr); /* * Frame pointers should strictly progress back up the stack @@ -3051,16 +3086,11 @@ user_backtrace(struct frame_tail *tail, return buftail.fp - 1; } -static void -perf_callchain_user(struct pt_regs *regs, - struct perf_callchain_entry *entry) +void +perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs) { struct frame_tail *tail; - callchain_store(entry, PERF_CONTEXT_USER); - - if (!user_mode(regs)) - regs = task_pt_regs(current); tail = (struct frame_tail *)regs->ARM_fp - 1; @@ -3078,56 +3108,18 @@ callchain_trace(struct stackframe *fr, void *data) { struct perf_callchain_entry *entry = data; - callchain_store(entry, fr->pc); + perf_callchain_store(entry, fr->pc); return 0; } -static void -perf_callchain_kernel(struct pt_regs *regs, - struct perf_callchain_entry *entry) +void +perf_callchain_kernel(struct perf_callchain_entry *entry, struct pt_regs *regs) { struct stackframe fr; - callchain_store(entry, PERF_CONTEXT_KERNEL); fr.fp = regs->ARM_fp; fr.sp = regs->ARM_sp; fr.lr = regs->ARM_lr; fr.pc = regs->ARM_pc; walk_stackframe(&fr, callchain_trace, entry); } - -static void -perf_do_callchain(struct pt_regs *regs, - struct perf_callchain_entry *entry) -{ - int is_user; - - if (!regs) - return; - - is_user = user_mode(regs); - - if (!current || !current->pid) - return; - - if (is_user && current->state != TASK_RUNNING) - return; - - if (!is_user) - perf_callchain_kernel(regs, entry); - - if (current->mm) - perf_callchain_user(regs, entry); -} - -static DEFINE_PER_CPU(struct perf_callchain_entry, pmc_irq_entry); - -struct perf_callchain_entry * -perf_callchain(struct pt_regs *regs) -{ - struct perf_callchain_entry *entry = &__get_cpu_var(pmc_irq_entry); - - entry->nr = 0; - perf_do_callchain(regs, entry); - return entry; -} diff --git a/arch/arm/oprofile/Makefile b/arch/arm/oprofile/Makefile index e666eaf..b2215c6 100644 --- a/arch/arm/oprofile/Makefile +++ b/arch/arm/oprofile/Makefile @@ -6,4 +6,8 @@ DRIVER_OBJS = $(addprefix ../../../drivers/oprofile/, \ oprofilefs.o oprofile_stats.o \ timer_int.o ) +ifeq ($(CONFIG_HW_PERF_EVENTS),y) +DRIVER_OBJS += $(addprefix ../../../drivers/oprofile/, oprofile_perf.o) +endif + oprofile-y := $(DRIVER_OBJS) common.o diff --git a/arch/arm/oprofile/common.c b/arch/arm/oprofile/common.c index 72e09eb..8aa9744 100644 --- a/arch/arm/oprofile/common.c +++ b/arch/arm/oprofile/common.c @@ -25,139 +25,10 @@ #include <asm/ptrace.h> #ifdef CONFIG_HW_PERF_EVENTS -/* - * Per performance monitor configuration as set via oprofilefs. - */ -struct op_counter_config { - unsigned long count; - unsigned long enabled; - unsigned long event; - unsigned long unit_mask; - unsigned long kernel; - unsigned long user; - struct perf_event_attr attr; -}; - -static int op_arm_enabled; -static DEFINE_MUTEX(op_arm_mutex); - -static struct op_counter_config *counter_config; -static struct perf_event **perf_events[nr_cpumask_bits]; -static int perf_num_counters; - -/* - * Overflow callback for oprofile. - */ -static void op_overflow_handler(struct perf_event *event, int unused, - struct perf_sample_data *data, struct pt_regs *regs) +char *op_name_from_perf_id(void) { - int id; - u32 cpu = smp_processor_id(); - - for (id = 0; id < perf_num_counters; ++id) - if (perf_events[cpu][id] == event) - break; - - if (id != perf_num_counters) - oprofile_add_sample(regs, id); - else - pr_warning("oprofile: ignoring spurious overflow " - "on cpu %u\n", cpu); -} - -/* - * Called by op_arm_setup to create perf attributes to mirror the oprofile - * settings in counter_config. Attributes are created as `pinned' events and - * so are permanently scheduled on the PMU. - */ -static void op_perf_setup(void) -{ - int i; - u32 size = sizeof(struct perf_event_attr); - struct perf_event_attr *attr; - - for (i = 0; i < perf_num_counters; ++i) { - attr = &counter_config[i].attr; - memset(attr, 0, size); - attr->type = PERF_TYPE_RAW; - attr->size = size; - attr->config = counter_config[i].event; - attr->sample_period = counter_config[i].count; - attr->pinned = 1; - } -} - -static int op_create_counter(int cpu, int event) -{ - int ret = 0; - struct perf_event *pevent; - - if (!counter_config[event].enabled || (perf_events[cpu][event] != NULL)) - return ret; - - pevent = perf_event_create_kernel_counter(&counter_config[event].attr, - cpu, -1, - op_overflow_handler); - - if (IS_ERR(pevent)) { - ret = PTR_ERR(pevent); - } else if (pevent->state != PERF_EVENT_STATE_ACTIVE) { - perf_event_release_kernel(pevent); - pr_warning("oprofile: failed to enable event %d " - "on CPU %d\n", event, cpu); - ret = -EBUSY; - } else { - perf_events[cpu][event] = pevent; - } - - return ret; -} + enum arm_perf_pmu_ids id = armpmu_get_pmu_id(); -static void op_destroy_counter(int cpu, int event) -{ - struct perf_event *pevent = perf_events[cpu][event]; - - if (pevent) { - perf_event_release_kernel(pevent); - perf_events[cpu][event] = NULL; - } -} - -/* - * Called by op_arm_start to create active perf events based on the - * perviously configured attributes. - */ -static int op_perf_start(void) -{ - int cpu, event, ret = 0; - - for_each_online_cpu(cpu) { - for (event = 0; event < perf_num_counters; ++event) { - ret = op_create_counter(cpu, event); - if (ret) - goto out; - } - } - -out: - return ret; -} - -/* - * Called by op_arm_stop at the end of a profiling run. - */ -static void op_perf_stop(void) -{ - int cpu, event; - - for_each_online_cpu(cpu) - for (event = 0; event < perf_num_counters; ++event) - op_destroy_counter(cpu, event); -} - - -static char *op_name_from_perf_id(enum arm_perf_pmu_ids id) -{ switch (id) { case ARM_PERF_PMU_ID_XSCALE1: return "arm/xscale1"; @@ -176,116 +47,6 @@ static char *op_name_from_perf_id(enum arm_perf_pmu_ids id) } } -static int op_arm_create_files(struct super_block *sb, struct dentry *root) -{ - unsigned int i; - - for (i = 0; i < perf_num_counters; i++) { - struct dentry *dir; - char buf[4]; - - snprintf(buf, sizeof buf, "%d", i); - dir = oprofilefs_mkdir(sb, root, buf); - oprofilefs_create_ulong(sb, dir, "enabled", &counter_config[i].enabled); - oprofilefs_create_ulong(sb, dir, "event", &counter_config[i].event); - oprofilefs_create_ulong(sb, dir, "count", &counter_config[i].count); - oprofilefs_create_ulong(sb, dir, "unit_mask", &counter_config[i].unit_mask); - oprofilefs_create_ulong(sb, dir, "kernel", &counter_config[i].kernel); - oprofilefs_create_ulong(sb, dir, "user", &counter_config[i].user); - } - - return 0; -} - -static int op_arm_setup(void) -{ - spin_lock(&oprofilefs_lock); - op_perf_setup(); - spin_unlock(&oprofilefs_lock); - return 0; -} - -static int op_arm_start(void) -{ - int ret = -EBUSY; - - mutex_lock(&op_arm_mutex); - if (!op_arm_enabled) { - ret = 0; - op_perf_start(); - op_arm_enabled = 1; - } - mutex_unlock(&op_arm_mutex); - return ret; -} - -static void op_arm_stop(void) -{ - mutex_lock(&op_arm_mutex); - if (op_arm_enabled) - op_perf_stop(); - op_arm_enabled = 0; - mutex_unlock(&op_arm_mutex); -} - -#ifdef CONFIG_PM -static int op_arm_suspend(struct platform_device *dev, pm_message_t state) -{ - mutex_lock(&op_arm_mutex); - if (op_arm_enabled) - op_perf_stop(); - mutex_unlock(&op_arm_mutex); - return 0; -} - -static int op_arm_resume(struct platform_device *dev) -{ - mutex_lock(&op_arm_mutex); - if (op_arm_enabled && op_perf_start()) - op_arm_enabled = 0; - mutex_unlock(&op_arm_mutex); - return 0; -} - -static struct platform_driver oprofile_driver = { - .driver = { - .name = "arm-oprofile", - }, - .resume = op_arm_resume, - .suspend = op_arm_suspend, -}; - -static struct platform_device *oprofile_pdev; - -static int __init init_driverfs(void) -{ - int ret; - - ret = platform_driver_register(&oprofile_driver); - if (ret) - goto out; - - oprofile_pdev = platform_device_register_simple( - oprofile_driver.driver.name, 0, NULL, 0); - if (IS_ERR(oprofile_pdev)) { - ret = PTR_ERR(oprofile_pdev); - platform_driver_unregister(&oprofile_driver); - } - -out: - return ret; -} - -static void exit_driverfs(void) -{ - platform_device_unregister(oprofile_pdev); - platform_driver_unregister(&oprofile_driver); -} -#else -static int __init init_driverfs(void) { return 0; } -#define exit_driverfs() do { } while (0) -#endif /* CONFIG_PM */ - static int report_trace(struct stackframe *frame, void *d) { unsigned int *depth = d; @@ -350,74 +111,14 @@ static void arm_backtrace(struct pt_regs * const regs, unsigned int depth) int __init oprofile_arch_init(struct oprofile_operations *ops) { - int cpu, ret = 0; - - perf_num_counters = armpmu_get_max_events(); - - counter_config = kcalloc(perf_num_counters, - sizeof(struct op_counter_config), GFP_KERNEL); - - if (!counter_config) { - pr_info("oprofile: failed to allocate %d " - "counters\n", perf_num_counters); - return -ENOMEM; - } - - ret = init_driverfs(); - if (ret) { - kfree(counter_config); - counter_config = NULL; - return ret; - } - - for_each_possible_cpu(cpu) { - perf_events[cpu] = kcalloc(perf_num_counters, - sizeof(struct perf_event *), GFP_KERNEL); - if (!perf_events[cpu]) { - pr_info("oprofile: failed to allocate %d perf events " - "for cpu %d\n", perf_num_counters, cpu); - while (--cpu >= 0) - kfree(perf_events[cpu]); - return -ENOMEM; - } - } - ops->backtrace = arm_backtrace; - ops->create_files = op_arm_create_files; - ops->setup = op_arm_setup; - ops->start = op_arm_start; - ops->stop = op_arm_stop; - ops->shutdown = op_arm_stop; - ops->cpu_type = op_name_from_perf_id(armpmu_get_pmu_id()); - - if (!ops->cpu_type) - ret = -ENODEV; - else - pr_info("oprofile: using %s\n", ops->cpu_type); - return ret; + return oprofile_perf_init(ops); } -void oprofile_arch_exit(void) +void __exit oprofile_arch_exit(void) { - int cpu, id; - struct perf_event *event; - - if (*perf_events) { - for_each_possible_cpu(cpu) { - for (id = 0; id < perf_num_counters; ++id) { - event = perf_events[cpu][id]; - if (event != NULL) - perf_event_release_kernel(event); - } - kfree(perf_events[cpu]); - } - } - - if (counter_config) { - kfree(counter_config); - exit_driverfs(); - } + oprofile_perf_exit(); } #else int __init oprofile_arch_init(struct oprofile_operations *ops) @@ -425,5 +126,5 @@ int __init oprofile_arch_init(struct oprofile_operations *ops) pr_info("oprofile: hardware counters not available\n"); return -ENODEV; } -void oprofile_arch_exit(void) {} +void __exit oprofile_arch_exit(void) {} #endif /* CONFIG_HW_PERF_EVENTS */ diff --git a/arch/frv/Kconfig b/arch/frv/Kconfig index 16399bd..0f2417d 100644 --- a/arch/frv/Kconfig +++ b/arch/frv/Kconfig @@ -7,6 +7,7 @@ config FRV default y select HAVE_IDE select HAVE_ARCH_TRACEHOOK + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS config ZONE_DMA diff --git a/arch/frv/lib/Makefile b/arch/frv/lib/Makefile index f470975..4ff2fb1 100644 --- a/arch/frv/lib/Makefile +++ b/arch/frv/lib/Makefile @@ -5,4 +5,4 @@ lib-y := \ __ashldi3.o __lshrdi3.o __muldi3.o __ashrdi3.o __negdi2.o __ucmpdi2.o \ checksum.o memcpy.o memset.o atomic-ops.o atomic64-ops.o \ - outsl_ns.o outsl_sw.o insl_ns.o insl_sw.o cache.o perf_event.o + outsl_ns.o outsl_sw.o insl_ns.o insl_sw.o cache.o diff --git a/arch/frv/lib/perf_event.c b/arch/frv/lib/perf_event.c deleted file mode 100644 index 9ac5acf..0000000 --- a/arch/frv/lib/perf_event.c +++ /dev/null @@ -1,19 +0,0 @@ -/* Performance event handling - * - * Copyright (C) 2009 Red Hat, Inc. All Rights Reserved. - * Written by David Howells (dhowells@redhat.com) - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public Licence - * as published by the Free Software Foundation; either version - * 2 of the Licence, or (at your option) any later version. - */ - -#include <linux/perf_event.h> - -/* - * mark the performance event as pending - */ -void set_perf_event_pending(void) -{ -} diff --git a/arch/ia64/include/asm/hardirq.h b/arch/ia64/include/asm/hardirq.h index d514cd9..8fb7d33 100644 --- a/arch/ia64/include/asm/hardirq.h +++ b/arch/ia64/include/asm/hardirq.h @@ -6,12 +6,6 @@ * David Mosberger-Tang <davidm@hpl.hp.com> */ - -#include <linux/threads.h> -#include <linux/irq.h> - -#include <asm/processor.h> - /* * No irq_cpustat_t for IA-64. The data is held in the per-CPU data structure. */ @@ -20,6 +14,11 @@ #define local_softirq_pending() (local_cpu_data->softirq_pending) +#include <linux/threads.h> +#include <linux/irq.h> + +#include <asm/processor.h> + extern void __iomem *ipi_base_addr; void ack_bad_irq(unsigned int irq); diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig index 907417d..79a04a9 100644 --- a/arch/parisc/Kconfig +++ b/arch/parisc/Kconfig @@ -16,6 +16,7 @@ config PARISC select RTC_DRV_GENERIC select INIT_ALL_POSSIBLE select BUG + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select GENERIC_ATOMIC64 if !64BIT help diff --git a/arch/parisc/include/asm/perf_event.h b/arch/parisc/include/asm/perf_event.h index cc14642..1e0fd8b 100644 --- a/arch/parisc/include/asm/perf_event.h +++ b/arch/parisc/include/asm/perf_event.h @@ -1,7 +1,6 @@ #ifndef __ASM_PARISC_PERF_EVENT_H #define __ASM_PARISC_PERF_EVENT_H -/* parisc only supports software events through this interface. */ -static inline void set_perf_event_pending(void) { } +/* Empty, just to avoid compiling error */ #endif /* __ASM_PARISC_PERF_EVENT_H */ diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 631e5a0..4b1e521 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -138,6 +138,7 @@ config PPC select HAVE_OPROFILE select HAVE_SYSCALL_WRAPPERS if PPC64 select GENERIC_ATOMIC64 if PPC32 + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select HAVE_REGS_AND_STACK_ACCESS_API select HAVE_HW_BREAKPOINT if PERF_EVENTS && PPC_BOOK3S_64 diff --git a/arch/powerpc/include/asm/paca.h b/arch/powerpc/include/asm/paca.h index 1ff6662..9b287fd 100644 --- a/arch/powerpc/include/asm/paca.h +++ b/arch/powerpc/include/asm/paca.h @@ -129,7 +129,7 @@ struct paca_struct { u8 soft_enabled; /* irq soft-enable flag */ u8 hard_enabled; /* set if irqs are enabled in MSR */ u8 io_sync; /* writel() needs spin_unlock sync */ - u8 perf_event_pending; /* PM interrupt while soft-disabled */ + u8 irq_work_pending; /* IRQ_WORK interrupt while soft-disable */ /* Stuff for accurate time accounting */ u64 user_time; /* accumulated usermode TB ticks */ diff --git a/arch/powerpc/kernel/perf_callchain.c b/arch/powerpc/kernel/perf_callchain.c index 95ad9da..d05ae42 100644 --- a/arch/powerpc/kernel/perf_callchain.c +++ b/arch/powerpc/kernel/perf_callchain.c @@ -23,18 +23,6 @@ #include "ppc32.h" #endif -/* - * Store another value in a callchain_entry. - */ -static inline void callchain_store(struct perf_callchain_entry *entry, u64 ip) -{ - unsigned int nr = entry->nr; - - if (nr < PERF_MAX_STACK_DEPTH) { - entry->ip[nr] = ip; - entry->nr = nr + 1; - } -} /* * Is sp valid as the address of the next kernel stack frame after prev_sp? @@ -58,8 +46,8 @@ static int valid_next_sp(unsigned long sp, unsigned long prev_sp) return 0; } -static void perf_callchain_kernel(struct pt_regs *regs, - struct perf_callchain_entry *entry) +void +perf_callchain_kernel(struct perf_callchain_entry *entry, struct pt_regs *regs) { unsigned long sp, next_sp; unsigned long next_ip; @@ -69,8 +57,7 @@ static void perf_callchain_kernel(struct pt_regs *regs, lr = regs->link; sp = regs->gpr[1]; - callchain_store(entry, PERF_CONTEXT_KERNEL); - callchain_store(entry, regs->nip); + perf_callchain_store(entry, regs->nip); if (!validate_sp(sp, current, STACK_FRAME_OVERHEAD)) return; @@ -89,7 +76,7 @@ static void perf_callchain_kernel(struct pt_regs *regs, next_ip = regs->nip; lr = regs->link; level = 0; - callchain_store(entry, PERF_CONTEXT_KERNEL); + perf_callchain_store(entry, PERF_CONTEXT_KERNEL); } else { if (level == 0) @@ -111,7 +98,7 @@ static void perf_callchain_kernel(struct pt_regs *regs, ++level; } - callchain_store(entry, next_ip); + perf_callchain_store(entry, next_ip); if (!valid_next_sp(next_sp, sp)) return; sp = next_sp; @@ -233,8 +220,8 @@ static int sane_signal_64_frame(unsigned long sp) puc == (unsigned long) &sf->uc; } -static void perf_callchain_user_64(struct pt_regs *regs, - struct perf_callchain_entry *entry) +static void perf_callchain_user_64(struct perf_callchain_entry *entry, + struct pt_regs *regs) { unsigned long sp, next_sp; unsigned long next_ip; @@ -246,8 +233,7 @@ static void perf_callchain_user_64(struct pt_regs *regs, next_ip = regs->nip; lr = regs->link; sp = regs->gpr[1]; - callchain_store(entry, PERF_CONTEXT_USER); - callchain_store(entry, next_ip); + perf_callchain_store(entry, next_ip); for (;;) { fp = (unsigned long __user *) sp; @@ -276,14 +262,14 @@ static void perf_callchain_user_64(struct pt_regs *regs, read_user_stack_64(&uregs[PT_R1], &sp)) return; level = 0; - callchain_store(entry, PERF_CONTEXT_USER); - callchain_store(entry, next_ip); + perf_callchain_store(entry, PERF_CONTEXT_USER); + perf_callchain_store(entry, next_ip); continue; } if (level == 0) next_ip = lr; - callchain_store(entry, next_ip); + perf_callchain_store(entry, next_ip); ++level; sp = next_sp; } @@ -315,8 +301,8 @@ static int read_user_stack_32(unsigned int __user *ptr, unsigned int *ret) return __get_user_inatomic(*ret, ptr); } -static inline void perf_callchain_user_64(struct pt_regs *regs, - struct perf_callchain_entry *entry) +static inline void perf_callchain_user_64(struct perf_callchain_entry *entry, + struct pt_regs *regs) { } @@ -435,8 +421,8 @@ static unsigned int __user *signal_frame_32_regs(unsigned int sp, return mctx->mc_gregs; } -static void perf_callchain_user_32(struct pt_regs *regs, - struct perf_callchain_entry *entry) +static void perf_callchain_user_32(struct perf_callchain_entry *entry, + struct pt_regs *regs) { unsigned int sp, next_sp; unsigned int next_ip; @@ -447,8 +433,7 @@ static void perf_callchain_user_32(struct pt_regs *regs, next_ip = regs->nip; lr = regs->link; sp = regs->gpr[1]; - callchain_store(entry, PERF_CONTEXT_USER); - callchain_store(entry, next_ip); + perf_callchain_store(entry, next_ip); while (entry->nr < PERF_MAX_STACK_DEPTH) { fp = (unsigned int __user *) (unsigned long) sp; @@ -470,45 +455,24 @@ static void perf_callchain_user_32(struct pt_regs *regs, read_user_stack_32(&uregs[PT_R1], &sp)) return; level = 0; - callchain_store(entry, PERF_CONTEXT_USER); - callchain_store(entry, next_ip); + perf_callchain_store(entry, PERF_CONTEXT_USER); + perf_callchain_store(entry, next_ip); continue; } if (level == 0) next_ip = lr; - callchain_store(entry, next_ip); + perf_callchain_store(entry, next_ip); ++level; sp = next_sp; } } -/* - * Since we can't get PMU interrupts inside a PMU interrupt handler, - * we don't need separate irq and nmi entries here. - */ -static DEFINE_PER_CPU(struct perf_callchain_entry, cpu_perf_callchain); - -struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) +void +perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs) { - struct perf_callchain_entry *entry = &__get_cpu_var(cpu_perf_callchain); - - entry->nr = 0; - - if (!user_mode(regs)) { - perf_callchain_kernel(regs, entry); - if (current->mm) - regs = task_pt_regs(current); - else - regs = NULL; - } - - if (regs) { - if (current_is_64bit()) - perf_callchain_user_64(regs, entry); - else - perf_callchain_user_32(regs, entry); - } - - return entry; + if (current_is_64bit()) + perf_callchain_user_64(entry, regs); + else + perf_callchain_user_32(entry, regs); } diff --git a/arch/powerpc/kernel/perf_event.c b/arch/powerpc/kernel/perf_event.c index d301a30..3129c85 100644 --- a/arch/powerpc/kernel/perf_event.c +++ b/arch/powerpc/kernel/perf_event.c @@ -402,6 +402,9 @@ static void power_pmu_read(struct perf_event *event) { s64 val, delta, prev; + if (event->hw.state & PERF_HES_STOPPED) + return; + if (!event->hw.idx) return; /* @@ -517,7 +520,7 @@ static void write_mmcr0(struct cpu_hw_events *cpuhw, unsigned long mmcr0) * Disable all events to prevent PMU interrupts and to allow * events to be added or removed. */ -void hw_perf_disable(void) +static void power_pmu_disable(struct pmu *pmu) { struct cpu_hw_events *cpuhw; unsigned long flags; @@ -565,7 +568,7 @@ void hw_perf_disable(void) * If we were previously disabled and events were added, then * put the new config on the PMU. */ -void hw_perf_enable(void) +static void power_pmu_enable(struct pmu *pmu) { struct perf_event *event; struct cpu_hw_events *cpuhw; @@ -672,6 +675,8 @@ void hw_perf_enable(void) } local64_set(&event->hw.prev_count, val); event->hw.idx = idx; + if (event->hw.state & PERF_HES_STOPPED) + val = 0; write_pmc(idx, val); perf_event_update_userpage(event); } @@ -727,7 +732,7 @@ static int collect_events(struct perf_event *group, int max_count, * re-enable the PMU in order to get hw_perf_enable to do the * actual work of reconfiguring the PMU. */ -static int power_pmu_enable(struct perf_event *event) +static int power_pmu_add(struct perf_event *event, int ef_flags) { struct cpu_hw_events *cpuhw; unsigned long flags; @@ -735,7 +740,7 @@ static int power_pmu_enable(struct perf_event *event) int ret = -EAGAIN; local_irq_save(flags); - perf_disable(); + perf_pmu_disable(event->pmu); /* * Add the event to the list (if there is room) @@ -749,6 +754,9 @@ static int power_pmu_enable(struct perf_event *event) cpuhw->events[n0] = event->hw.config; cpuhw->flags[n0] = event->hw.event_base; + if (!(ef_flags & PERF_EF_START)) + event->hw.state = PERF_HES_STOPPED | PERF_HES_UPTODATE; + /* * If group events scheduling transaction was started, * skip the schedulability test here, it will be peformed @@ -769,7 +777,7 @@ nocheck: ret = 0; out: - perf_enable(); + perf_pmu_enable(event->pmu); local_irq_restore(flags); return ret; } @@ -777,14 +785,14 @@ nocheck: /* * Remove a event from the PMU. */ -static void power_pmu_disable(struct perf_event *event) +static void power_pmu_del(struct perf_event *event, int ef_flags) { struct cpu_hw_events *cpuhw; long i; unsigned long flags; local_irq_save(flags); - perf_disable(); + perf_pmu_disable(event->pmu); power_pmu_read(event); @@ -821,34 +829,60 @@ static void power_pmu_disable(struct perf_event *event) cpuhw->mmcr[0] &= ~(MMCR0_PMXE | MMCR0_FCECE); } - perf_enable(); + perf_pmu_enable(event->pmu); local_irq_restore(flags); } /* - * Re-enable interrupts on a event after they were throttled - * because they were coming too fast. + * POWER-PMU does not support disabling individual counters, hence + * program their cycle counter to their max value and ignore the interrupts. */ -static void power_pmu_unthrottle(struct perf_event *event) + +static void power_pmu_start(struct perf_event *event, int ef_flags) +{ + unsigned long flags; + s64 left; + + if (!event->hw.idx || !event->hw.sample_period) + return; + + if (!(event->hw.state & PERF_HES_STOPPED)) + return; + + if (ef_flags & PERF_EF_RELOAD) + WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE)); + + local_irq_save(flags); + perf_pmu_disable(event->pmu); + + event->hw.state = 0; + left = local64_read(&event->hw.period_left); + write_pmc(event->hw.idx, left); + + perf_event_update_userpage(event); + perf_pmu_enable(event->pmu); + local_irq_restore(flags); +} + +static void power_pmu_stop(struct perf_event *event, int ef_flags) { - s64 val, left; unsigned long flags; if (!event->hw.idx || !event->hw.sample_period) return; + + if (event->hw.state & PERF_HES_STOPPED) + return; + local_irq_save(flags); - perf_disable(); + perf_pmu_disable(event->pmu); + power_pmu_read(event); - left = event->hw.sample_period; - event->hw.last_period = left; - val = 0; - if (left < 0x80000000L) - val = 0x80000000L - left; - write_pmc(event->hw.idx, val); - local64_set(&event->hw.prev_count, val); - local64_set(&event->hw.period_left, left); + event->hw.state |= PERF_HES_STOPPED | PERF_HES_UPTODATE; + write_pmc(event->hw.idx, 0); + perf_event_update_userpage(event); - perf_enable(); + perf_pmu_enable(event->pmu); local_irq_restore(flags); } @@ -857,10 +891,11 @@ static void power_pmu_unthrottle(struct perf_event *event) * Set the flag to make pmu::enable() not perform the * schedulability test, it will be performed at commit time */ -void power_pmu_start_txn(const struct pmu *pmu) +void power_pmu_start_txn(struct pmu *pmu) { struct cpu_hw_events *cpuhw = &__get_cpu_var(cpu_hw_events); + perf_pmu_disable(pmu); cpuhw->group_flag |= PERF_EVENT_TXN; cpuhw->n_txn_start = cpuhw->n_events; } @@ -870,11 +905,12 @@ void power_pmu_start_txn(const struct pmu *pmu) * Clear the flag and pmu::enable() will perform the * schedulability test. */ -void power_pmu_cancel_txn(const struct pmu *pmu) +void power_pmu_cancel_txn(struct pmu *pmu) { struct cpu_hw_events *cpuhw = &__get_cpu_var(cpu_hw_events); cpuhw->group_flag &= ~PERF_EVENT_TXN; + perf_pmu_enable(pmu); } /* @@ -882,7 +918,7 @@ void power_pmu_cancel_txn(const struct pmu *pmu) * Perform the group schedulability test as a whole * Return 0 if success */ -int power_pmu_commit_txn(const struct pmu *pmu) +int power_pmu_commit_txn(struct pmu *pmu) { struct cpu_hw_events *cpuhw; long i, n; @@ -901,19 +937,10 @@ int power_pmu_commit_txn(const struct pmu *pmu) cpuhw->event[i]->hw.config = cpuhw->events[i]; cpuhw->group_flag &= ~PERF_EVENT_TXN; + perf_pmu_enable(pmu); return 0; } -struct pmu power_pmu = { - .enable = power_pmu_enable, - .disable = power_pmu_disable, - .read = power_pmu_read, - .unthrottle = power_pmu_unthrottle, - .start_txn = power_pmu_start_txn, - .cancel_txn = power_pmu_cancel_txn, - .commit_txn = power_pmu_commit_txn, -}; - /* * Return 1 if we might be able to put event on a limited PMC, * or 0 if not. @@ -1014,7 +1041,7 @@ static int hw_perf_cache_event(u64 config, u64 *eventp) return 0; } -const struct pmu *hw_perf_event_init(struct perf_event *event) +static int power_pmu_event_init(struct perf_event *event) { u64 ev; unsigned long flags; @@ -1026,25 +1053,27 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) struct cpu_hw_events *cpuhw; if (!ppmu) - return ERR_PTR(-ENXIO); + return -ENOENT; + switch (event->attr.type) { case PERF_TYPE_HARDWARE: ev = event->attr.config; if (ev >= ppmu->n_generic || ppmu->generic_events[ev] == 0) - return ERR_PTR(-EOPNOTSUPP); + return -EOPNOTSUPP; ev = ppmu->generic_events[ev]; break; case PERF_TYPE_HW_CACHE: err = hw_perf_cache_event(event->attr.config, &ev); if (err) - return ERR_PTR(err); + return err; break; case PERF_TYPE_RAW: ev = event->attr.config; break; default: - return ERR_PTR(-EINVAL); + return -ENOENT; } + event->hw.config_base = ev; event->hw.idx = 0; @@ -1063,7 +1092,7 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) * XXX we should check if the task is an idle task. */ flags = 0; - if (event->ctx->task) + if (event->attach_state & PERF_ATTACH_TASK) flags |= PPMU_ONLY_COUNT_RUN; /* @@ -1081,7 +1110,7 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) */ ev = normal_pmc_alternative(ev, flags); if (!ev) - return ERR_PTR(-EINVAL); + return -EINVAL; } } @@ -1095,19 +1124,19 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) n = collect_events(event->group_leader, ppmu->n_counter - 1, ctrs, events, cflags); if (n < 0) - return ERR_PTR(-EINVAL); + return -EINVAL; } events[n] = ev; ctrs[n] = event; cflags[n] = flags; if (check_excludes(ctrs, cflags, n, 1)) - return ERR_PTR(-EINVAL); + return -EINVAL; cpuhw = &get_cpu_var(cpu_hw_events); err = power_check_constraints(cpuhw, events, cflags, n + 1); put_cpu_var(cpu_hw_events); if (err) - return ERR_PTR(-EINVAL); + return -EINVAL; event->hw.config = events[n]; event->hw.event_base = cflags[n]; @@ -1132,11 +1161,23 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) } event->destroy = hw_perf_event_destroy; - if (err) - return ERR_PTR(err); - return &power_pmu; + return err; } +struct pmu power_pmu = { + .pmu_enable = power_pmu_enable, + .pmu_disable = power_pmu_disable, + .event_init = power_pmu_event_init, + .add = power_pmu_add, + .del = power_pmu_del, + .start = power_pmu_start, + .stop = power_pmu_stop, + .read = power_pmu_read, + .start_txn = power_pmu_start_txn, + .cancel_txn = power_pmu_cancel_txn, + .commit_txn = power_pmu_commit_txn, +}; + /* * A counter has overflowed; update its count and record * things if requested. Note that interrupts are hard-disabled @@ -1149,6 +1190,11 @@ static void record_and_restart(struct perf_event *event, unsigned long val, s64 prev, delta, left; int record = 0; + if (event->hw.state & PERF_HES_STOPPED) { + write_pmc(event->hw.idx, 0); + return; + } + /* we don't have to worry about interrupts here */ prev = local64_read(&event->hw.prev_count); delta = (val - prev) & 0xfffffffful; @@ -1171,6 +1217,11 @@ static void record_and_restart(struct perf_event *event, unsigned long val, val = 0x80000000LL - left; } + write_pmc(event->hw.idx, val); + local64_set(&event->hw.prev_count, val); + local64_set(&event->hw.period_left, left); + perf_event_update_userpage(event); + /* * Finally record data if requested. */ @@ -1183,23 +1234,9 @@ static void record_and_restart(struct perf_event *event, unsigned long val, if (event->attr.sample_type & PERF_SAMPLE_ADDR) perf_get_data_addr(regs, &data.addr); - if (perf_event_overflow(event, nmi, &data, regs)) { - /* - * Interrupts are coming too fast - throttle them - * by setting the event to 0, so it will be - * at least 2^30 cycles until the next interrupt - * (assuming each event counts at most 2 counts - * per cycle). - */ - val = 0; - left = ~0ULL >> 1; - } + if (perf_event_overflow(event, nmi, &data, regs)) + power_pmu_stop(event, 0); } - - write_pmc(event->hw.idx, val); - local64_set(&event->hw.prev_count, val); - local64_set(&event->hw.period_left, left); - perf_event_update_userpage(event); } /* @@ -1342,6 +1379,7 @@ int register_power_pmu(struct power_pmu *pmu) freeze_events_kernel = MMCR0_FCHV; #endif /* CONFIG_PPC64 */ + perf_pmu_register(&power_pmu); perf_cpu_notifier(power_pmu_notifier); return 0; diff --git a/arch/powerpc/kernel/perf_event_fsl_emb.c b/arch/powerpc/kernel/perf_event_fsl_emb.c index 1ba4547..7ecca59 100644 --- a/arch/powerpc/kernel/perf_event_fsl_emb.c +++ b/arch/powerpc/kernel/perf_event_fsl_emb.c @@ -156,6 +156,9 @@ static void fsl_emb_pmu_read(struct perf_event *event) { s64 val, delta, prev; + if (event->hw.state & PERF_HES_STOPPED) + return; + /* * Performance monitor interrupts come even when interrupts * are soft-disabled, as long as interrupts are hard-enabled. @@ -177,7 +180,7 @@ static void fsl_emb_pmu_read(struct perf_event *event) * Disable all events to prevent PMU interrupts and to allow * events to be added or removed. */ -void hw_perf_disable(void) +static void fsl_emb_pmu_disable(struct pmu *pmu) { struct cpu_hw_events *cpuhw; unsigned long flags; @@ -216,7 +219,7 @@ void hw_perf_disable(void) * If we were previously disabled and events were added, then * put the new config on the PMU. */ -void hw_perf_enable(void) +static void fsl_emb_pmu_enable(struct pmu *pmu) { struct cpu_hw_events *cpuhw; unsigned long flags; @@ -262,8 +265,8 @@ static int collect_events(struct perf_event *group, int max_count, return n; } -/* perf must be disabled, context locked on entry */ -static int fsl_emb_pmu_enable(struct perf_event *event) +/* context locked on entry */ +static int fsl_emb_pmu_add(struct perf_event *event, int flags) { struct cpu_hw_events *cpuhw; int ret = -EAGAIN; @@ -271,6 +274,7 @@ static int fsl_emb_pmu_enable(struct perf_event *event) u64 val; int i; + perf_pmu_disable(event->pmu); cpuhw = &get_cpu_var(cpu_hw_events); if (event->hw.config & FSL_EMB_EVENT_RESTRICTED) @@ -301,6 +305,12 @@ static int fsl_emb_pmu_enable(struct perf_event *event) val = 0x80000000L - left; } local64_set(&event->hw.prev_count, val); + + if (!(flags & PERF_EF_START)) { + event->hw.state = PERF_HES_STOPPED | PERF_HES_UPTODATE; + val = 0; + } + write_pmc(i, val); perf_event_update_userpage(event); @@ -310,15 +320,17 @@ static int fsl_emb_pmu_enable(struct perf_event *event) ret = 0; out: put_cpu_var(cpu_hw_events); + perf_pmu_enable(event->pmu); return ret; } -/* perf must be disabled, context locked on entry */ -static void fsl_emb_pmu_disable(struct perf_event *event) +/* context locked on entry */ +static void fsl_emb_pmu_del(struct perf_event *event, int flags) { struct cpu_hw_events *cpuhw; int i = event->hw.idx; + perf_pmu_disable(event->pmu); if (i < 0) goto out; @@ -346,44 +358,57 @@ static void fsl_emb_pmu_disable(struct perf_event *event) cpuhw->n_events--; out: + perf_pmu_enable(event->pmu); put_cpu_var(cpu_hw_events); } -/* - * Re-enable interrupts on a event after they were throttled - * because they were coming too fast. - * - * Context is locked on entry, but perf is not disabled. - */ -static void fsl_emb_pmu_unthrottle(struct perf_event *event) +static void fsl_emb_pmu_start(struct perf_event *event, int ef_flags) { - s64 val, left; unsigned long flags; + s64 left; if (event->hw.idx < 0 || !event->hw.sample_period) return; + + if (!(event->hw.state & PERF_HES_STOPPED)) + return; + + if (ef_flags & PERF_EF_RELOAD) + WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE)); + local_irq_save(flags); - perf_disable(); - fsl_emb_pmu_read(event); - left = event->hw.sample_period; - event->hw.last_period = left; - val = 0; - if (left < 0x80000000L) - val = 0x80000000L - left; - write_pmc(event->hw.idx, val); - local64_set(&event->hw.prev_count, val); - local64_set(&event->hw.period_left, left); + perf_pmu_disable(event->pmu); + + event->hw.state = 0; + left = local64_read(&event->hw.period_left); + write_pmc(event->hw.idx, left); + perf_event_update_userpage(event); - perf_enable(); + perf_pmu_enable(event->pmu); local_irq_restore(flags); } -static struct pmu fsl_emb_pmu = { - .enable = fsl_emb_pmu_enable, - .disable = fsl_emb_pmu_disable, - .read = fsl_emb_pmu_read, - .unthrottle = fsl_emb_pmu_unthrottle, -}; +static void fsl_emb_pmu_stop(struct perf_event *event, int ef_flags) +{ + unsigned long flags; + + if (event->hw.idx < 0 || !event->hw.sample_period) + return; + + if (event->hw.state & PERF_HES_STOPPED) + return; + + local_irq_save(flags); + perf_pmu_disable(event->pmu); + + fsl_emb_pmu_read(event); + event->hw.state |= PERF_HES_STOPPED | PERF_HES_UPTODATE; + write_pmc(event->hw.idx, 0); + + perf_event_update_userpage(event); + perf_pmu_enable(event->pmu); + local_irq_restore(flags); +} /* * Release the PMU if this is the last perf_event. @@ -428,7 +453,7 @@ static int hw_perf_cache_event(u64 config, u64 *eventp) return 0; } -const struct pmu *hw_perf_event_init(struct perf_event *event) +static int fsl_emb_pmu_event_init(struct perf_event *event) { u64 ev; struct perf_event *events[MAX_HWEVENTS]; @@ -441,14 +466,14 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) case PERF_TYPE_HARDWARE: ev = event->attr.config; if (ev >= ppmu->n_generic || ppmu->generic_events[ev] == 0) - return ERR_PTR(-EOPNOTSUPP); + return -EOPNOTSUPP; ev = ppmu->generic_events[ev]; break; case PERF_TYPE_HW_CACHE: err = hw_perf_cache_event(event->attr.config, &ev); if (err) - return ERR_PTR(err); + return err; break; case PERF_TYPE_RAW: @@ -456,12 +481,12 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) break; default: - return ERR_PTR(-EINVAL); + return -ENOENT; } event->hw.config = ppmu->xlate_event(ev); if (!(event->hw.config & FSL_EMB_EVENT_VALID)) - return ERR_PTR(-EINVAL); + return -EINVAL; /* * If this is in a group, check if it can go on with all the @@ -473,7 +498,7 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) n = collect_events(event->group_leader, ppmu->n_counter - 1, events); if (n < 0) - return ERR_PTR(-EINVAL); + return -EINVAL; } if (event->hw.config & FSL_EMB_EVENT_RESTRICTED) { @@ -484,7 +509,7 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) } if (num_restricted >= ppmu->n_restricted) - return ERR_PTR(-EINVAL); + return -EINVAL; } event->hw.idx = -1; @@ -497,7 +522,7 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) if (event->attr.exclude_kernel) event->hw.config_base |= PMLCA_FCS; if (event->attr.exclude_idle) - return ERR_PTR(-ENOTSUPP); + return -ENOTSUPP; event->hw.last_period = event->hw.sample_period; local64_set(&event->hw.period_left, event->hw.last_period); @@ -523,11 +548,20 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) } event->destroy = hw_perf_event_destroy; - if (err) - return ERR_PTR(err); - return &fsl_emb_pmu; + return err; } +static struct pmu fsl_emb_pmu = { + .pmu_enable = fsl_emb_pmu_enable, + .pmu_disable = fsl_emb_pmu_disable, + .event_init = fsl_emb_pmu_event_init, + .add = fsl_emb_pmu_add, + .del = fsl_emb_pmu_del, + .start = fsl_emb_pmu_start, + .stop = fsl_emb_pmu_stop, + .read = fsl_emb_pmu_read, +}; + /* * A counter has overflowed; update its count and record * things if requested. Note that interrupts are hard-disabled @@ -540,6 +574,11 @@ static void record_and_restart(struct perf_event *event, unsigned long val, s64 prev, delta, left; int record = 0; + if (event->hw.state & PERF_HES_STOPPED) { + write_pmc(event->hw.idx, 0); + return; + } + /* we don't have to worry about interrupts here */ prev = local64_read(&event->hw.prev_count); delta = (val - prev) & 0xfffffffful; @@ -562,6 +601,11 @@ static void record_and_restart(struct perf_event *event, unsigned long val, val = 0x80000000LL - left; } + write_pmc(event->hw.idx, val); + local64_set(&event->hw.prev_count, val); + local64_set(&event->hw.period_left, left); + perf_event_update_userpage(event); + /* * Finally record data if requested. */ @@ -571,23 +615,9 @@ static void record_and_restart(struct perf_event *event, unsigned long val, perf_sample_data_init(&data, 0); data.period = event->hw.last_period; - if (perf_event_overflow(event, nmi, &data, regs)) { - /* - * Interrupts are coming too fast - throttle them - * by setting the event to 0, so it will be - * at least 2^30 cycles until the next interrupt - * (assuming each event counts at most 2 counts - * per cycle). - */ - val = 0; - left = ~0ULL >> 1; - } + if (perf_event_overflow(event, nmi, &data, regs)) + fsl_emb_pmu_stop(event, 0); } - - write_pmc(event->hw.idx, val); - local64_set(&event->hw.prev_count, val); - local64_set(&event->hw.period_left, left); - perf_event_update_userpage(event); } static void perf_event_interrupt(struct pt_regs *regs) @@ -651,5 +681,7 @@ int register_fsl_emb_pmu(struct fsl_emb_pmu *pmu) pr_info("%s performance monitor hardware support registered\n", pmu->name); + perf_pmu_register(&fsl_emb_pmu); + return 0; } diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index 8533b3b..54888eb 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -53,7 +53,7 @@ #include <linux/posix-timers.h> #include <linux/irq.h> #include <linux/delay.h> -#include <linux/perf_event.h> +#include <linux/irq_work.h> #include <asm/trace.h> #include <asm/io.h> @@ -493,60 +493,60 @@ void __init iSeries_time_init_early(void) } #endif /* CONFIG_PPC_ISERIES */ -#ifdef CONFIG_PERF_EVENTS +#ifdef CONFIG_IRQ_WORK /* * 64-bit uses a byte in the PACA, 32-bit uses a per-cpu variable... */ #ifdef CONFIG_PPC64 -static inline unsigned long test_perf_event_pending(void) +static inline unsigned long test_irq_work_pending(void) { unsigned long x; asm volatile("lbz %0,%1(13)" : "=r" (x) - : "i" (offsetof(struct paca_struct, perf_event_pending))); + : "i" (offsetof(struct paca_struct, irq_work_pending))); return x; } -static inline void set_perf_event_pending_flag(void) +static inline void set_irq_work_pending_flag(void) { asm volatile("stb %0,%1(13)" : : "r" (1), - "i" (offsetof(struct paca_struct, perf_event_pending))); + "i" (offsetof(struct paca_struct, irq_work_pending))); } -static inline void clear_perf_event_pending(void) +static inline void clear_irq_work_pending(void) { asm volatile("stb %0,%1(13)" : : "r" (0), - "i" (offsetof(struct paca_struct, perf_event_pending))); + "i" (offsetof(struct paca_struct, irq_work_pending))); } #else /* 32-bit */ -DEFINE_PER_CPU(u8, perf_event_pending); +DEFINE_PER_CPU(u8, irq_work_pending); -#define set_perf_event_pending_flag() __get_cpu_var(perf_event_pending) = 1 -#define test_perf_event_pending() __get_cpu_var(perf_event_pending) -#define clear_perf_event_pending() __get_cpu_var(perf_event_pending) = 0 +#define set_irq_work_pending_flag() __get_cpu_var(irq_work_pending) = 1 +#define test_irq_work_pending() __get_cpu_var(irq_work_pending) +#define clear_irq_work_pending() __get_cpu_var(irq_work_pending) = 0 #endif /* 32 vs 64 bit */ -void set_perf_event_pending(void) +void set_irq_work_pending(void) { preempt_disable(); - set_perf_event_pending_flag(); + set_irq_work_pending_flag(); set_dec(1); preempt_enable(); } -#else /* CONFIG_PERF_EVENTS */ +#else /* CONFIG_IRQ_WORK */ -#define test_perf_event_pending() 0 -#define clear_perf_event_pending() +#define test_irq_work_pending() 0 +#define clear_irq_work_pending() -#endif /* CONFIG_PERF_EVENTS */ +#endif /* CONFIG_IRQ_WORK */ /* * For iSeries shared processors, we have to let the hypervisor @@ -587,9 +587,9 @@ void timer_interrupt(struct pt_regs * regs) calculate_steal_time(); - if (test_perf_event_pending()) { - clear_perf_event_pending(); - perf_event_do_pending(); + if (test_irq_work_pending()) { + clear_irq_work_pending(); + irq_work_run(); } #ifdef CONFIG_PPC_ISERIES diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index f0777a4..958f0da 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -95,6 +95,7 @@ config S390 select HAVE_KVM if 64BIT select HAVE_ARCH_TRACEHOOK select INIT_ALL_POSSIBLE + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select HAVE_KERNEL_GZIP select HAVE_KERNEL_BZIP2 diff --git a/arch/s390/include/asm/hardirq.h b/arch/s390/include/asm/hardirq.h index 498bc38..881d945 100644 --- a/arch/s390/include/asm/hardirq.h +++ b/arch/s390/include/asm/hardirq.h @@ -12,10 +12,6 @@ #ifndef __ASM_HARDIRQ_H #define __ASM_HARDIRQ_H -#include <linux/threads.h> -#include <linux/sched.h> -#include <linux/cache.h> -#include <linux/interrupt.h> #include <asm/lowcore.h> #define local_softirq_pending() (S390_lowcore.softirq_pending) diff --git a/arch/s390/include/asm/perf_event.h b/arch/s390/include/asm/perf_event.h index 3840cbe..a75f168 100644 --- a/arch/s390/include/asm/perf_event.h +++ b/arch/s390/include/asm/perf_event.h @@ -4,7 +4,6 @@ * Copyright 2009 Martin Schwidefsky, IBM Corporation. */ -static inline void set_perf_event_pending(void) {} -static inline void clear_perf_event_pending(void) {} +/* Empty, just to avoid compiling error */ #define PERF_EVENT_INDEX_OFFSET 0 diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig index 33990fa..35b6879 100644 --- a/arch/sh/Kconfig +++ b/arch/sh/Kconfig @@ -16,6 +16,7 @@ config SUPERH select HAVE_ARCH_TRACEHOOK select HAVE_DMA_API_DEBUG select HAVE_DMA_ATTRS + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC select HAVE_KERNEL_GZIP @@ -249,6 +250,11 @@ config ARCH_SHMOBILE select PM select PM_RUNTIME +config CPU_HAS_PMU + depends on CPU_SH4 || CPU_SH4A + default y + bool + if SUPERH32 choice @@ -738,6 +744,14 @@ config GUSA_RB LLSC, this should be more efficient than the other alternative of disabling interrupts around the atomic sequence. +config HW_PERF_EVENTS + bool "Enable hardware performance counter support for perf events" + depends on PERF_EVENTS && CPU_HAS_PMU + default y + help + Enable hardware performance counter support for perf events. If + disabled, perf events will use software events only. + source "drivers/sh/Kconfig" endmenu diff --git a/arch/sh/include/asm/perf_event.h b/arch/sh/include/asm/perf_event.h index 3d0c9f3..14308be 100644 --- a/arch/sh/include/asm/perf_event.h +++ b/arch/sh/include/asm/perf_event.h @@ -26,11 +26,4 @@ extern int register_sh_pmu(struct sh_pmu *); extern int reserve_pmc_hardware(void); extern void release_pmc_hardware(void); -static inline void set_perf_event_pending(void) -{ - /* Nothing to see here, move along. */ -} - -#define PERF_EVENT_INDEX_OFFSET 0 - #endif /* __ASM_SH_PERF_EVENT_H */ diff --git a/arch/sh/kernel/perf_callchain.c b/arch/sh/kernel/perf_callchain.c index a9dd3ab..d5ca1ef 100644 --- a/arch/sh/kernel/perf_callchain.c +++ b/arch/sh/kernel/perf_callchain.c @@ -14,11 +14,6 @@ #include <asm/unwinder.h> #include <asm/ptrace.h> -static inline void callchain_store(struct perf_callchain_entry *entry, u64 ip) -{ - if (entry->nr < PERF_MAX_STACK_DEPTH) - entry->ip[entry->nr++] = ip; -} static void callchain_warning(void *data, char *msg) { @@ -39,7 +34,7 @@ static void callchain_address(void *data, unsigned long addr, int reliable) struct perf_callchain_entry *entry = data; if (reliable) - callchain_store(entry, addr); + perf_callchain_store(entry, addr); } static const struct stacktrace_ops callchain_ops = { @@ -49,47 +44,10 @@ static const struct stacktrace_ops callchain_ops = { .address = callchain_address, }; -static void -perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry) +void +perf_callchain_kernel(struct perf_callchain_entry *entry, struct pt_regs *regs) { - callchain_store(entry, PERF_CONTEXT_KERNEL); - callchain_store(entry, regs->pc); + perf_callchain_store(entry, regs->pc); unwind_stack(NULL, regs, NULL, &callchain_ops, entry); } - -static void -perf_do_callchain(struct pt_regs *regs, struct perf_callchain_entry *entry) -{ - int is_user; - - if (!regs) - return; - - is_user = user_mode(regs); - - if (is_user && current->state != TASK_RUNNING) - return; - - /* - * Only the kernel side is implemented for now. - */ - if (!is_user) - perf_callchain_kernel(regs, entry); -} - -/* - * No need for separate IRQ and NMI entries. - */ -static DEFINE_PER_CPU(struct perf_callchain_entry, callchain); - -struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) -{ - struct perf_callchain_entry *entry = &__get_cpu_var(callchain); - - entry->nr = 0; - - perf_do_callchain(regs, entry); - - return entry; -} diff --git a/arch/sh/kernel/perf_event.c b/arch/sh/kernel/perf_event.c index 7a3dc35..5a4b334 100644 --- a/arch/sh/kernel/perf_event.c +++ b/arch/sh/kernel/perf_event.c @@ -59,6 +59,24 @@ static inline int sh_pmu_initialized(void) return !!sh_pmu; } +const char *perf_pmu_name(void) +{ + if (!sh_pmu) + return NULL; + + return sh_pmu->name; +} +EXPORT_SYMBOL_GPL(perf_pmu_name); + +int perf_num_counters(void) +{ + if (!sh_pmu) + return 0; + + return sh_pmu->num_events; +} +EXPORT_SYMBOL_GPL(perf_num_counters); + /* * Release the PMU if this is the last perf_event. */ @@ -206,50 +224,80 @@ again: local64_add(delta, &event->count); } -static void sh_pmu_disable(struct perf_event *event) +static void sh_pmu_stop(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; - clear_bit(idx, cpuc->active_mask); - sh_pmu->disable(hwc, idx); + if (!(event->hw.state & PERF_HES_STOPPED)) { + sh_pmu->disable(hwc, idx); + cpuc->events[idx] = NULL; + event->hw.state |= PERF_HES_STOPPED; + } + + if ((flags & PERF_EF_UPDATE) && !(event->hw.state & PERF_HES_UPTODATE)) { + sh_perf_event_update(event, &event->hw, idx); + event->hw.state |= PERF_HES_UPTODATE; + } +} + +static void sh_pmu_start(struct perf_event *event, int flags) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct hw_perf_event *hwc = &event->hw; + int idx = hwc->idx; + + if (WARN_ON_ONCE(idx == -1)) + return; + + if (flags & PERF_EF_RELOAD) + WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE)); - barrier(); + cpuc->events[idx] = event; + event->hw.state = 0; + sh_pmu->enable(hwc, idx); +} - sh_perf_event_update(event, &event->hw, idx); +static void sh_pmu_del(struct perf_event *event, int flags) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - cpuc->events[idx] = NULL; - clear_bit(idx, cpuc->used_mask); + sh_pmu_stop(event, PERF_EF_UPDATE); + __clear_bit(event->hw.idx, cpuc->used_mask); perf_event_update_userpage(event); } -static int sh_pmu_enable(struct perf_event *event) +static int sh_pmu_add(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc = &event->hw; int idx = hwc->idx; + int ret = -EAGAIN; + + perf_pmu_disable(event->pmu); - if (test_and_set_bit(idx, cpuc->used_mask)) { + if (__test_and_set_bit(idx, cpuc->used_mask)) { idx = find_first_zero_bit(cpuc->used_mask, sh_pmu->num_events); if (idx == sh_pmu->num_events) - return -EAGAIN; + goto out; - set_bit(idx, cpuc->used_mask); + __set_bit(idx, cpuc->used_mask); hwc->idx = idx; } sh_pmu->disable(hwc, idx); - cpuc->events[idx] = event; - set_bit(idx, cpuc->active_mask); - - sh_pmu->enable(hwc, idx); + event->hw.state = PERF_HES_UPTODATE | PERF_HES_STOPPED; + if (flags & PERF_EF_START) + sh_pmu_start(event, PERF_EF_RELOAD); perf_event_update_userpage(event); - - return 0; + ret = 0; +out: + perf_pmu_enable(event->pmu); + return ret; } static void sh_pmu_read(struct perf_event *event) @@ -257,24 +305,56 @@ static void sh_pmu_read(struct perf_event *event) sh_perf_event_update(event, &event->hw, event->hw.idx); } -static const struct pmu pmu = { - .enable = sh_pmu_enable, - .disable = sh_pmu_disable, - .read = sh_pmu_read, -}; - -const struct pmu *hw_perf_event_init(struct perf_event *event) +static int sh_pmu_event_init(struct perf_event *event) { - int err = __hw_perf_event_init(event); + int err; + + switch (event->attr.type) { + case PERF_TYPE_RAW: + case PERF_TYPE_HW_CACHE: + case PERF_TYPE_HARDWARE: + err = __hw_perf_event_init(event); + break; + + default: + return -ENOENT; + } + if (unlikely(err)) { if (event->destroy) event->destroy(event); - return ERR_PTR(err); } - return &pmu; + return err; +} + +static void sh_pmu_enable(struct pmu *pmu) +{ + if (!sh_pmu_initialized()) + return; + + sh_pmu->enable_all(); +} + +static void sh_pmu_disable(struct pmu *pmu) +{ + if (!sh_pmu_initialized()) + return; + + sh_pmu->disable_all(); } +static struct pmu pmu = { + .pmu_enable = sh_pmu_enable, + .pmu_disable = sh_pmu_disable, + .event_init = sh_pmu_event_init, + .add = sh_pmu_add, + .del = sh_pmu_del, + .start = sh_pmu_start, + .stop = sh_pmu_stop, + .read = sh_pmu_read, +}; + static void sh_pmu_setup(int cpu) { struct cpu_hw_events *cpuhw = &per_cpu(cpu_hw_events, cpu); @@ -299,32 +379,17 @@ sh_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu) return NOTIFY_OK; } -void hw_perf_enable(void) -{ - if (!sh_pmu_initialized()) - return; - - sh_pmu->enable_all(); -} - -void hw_perf_disable(void) -{ - if (!sh_pmu_initialized()) - return; - - sh_pmu->disable_all(); -} - -int __cpuinit register_sh_pmu(struct sh_pmu *pmu) +int __cpuinit register_sh_pmu(struct sh_pmu *_pmu) { if (sh_pmu) return -EBUSY; - sh_pmu = pmu; + sh_pmu = _pmu; - pr_info("Performance Events: %s support registered\n", pmu->name); + pr_info("Performance Events: %s support registered\n", _pmu->name); - WARN_ON(pmu->num_events > MAX_HWEVENTS); + WARN_ON(_pmu->num_events > MAX_HWEVENTS); + perf_pmu_register(&pmu); perf_cpu_notifier(sh_pmu_notifier); return 0; } diff --git a/arch/sh/oprofile/Makefile b/arch/sh/oprofile/Makefile index 4886c5c..e85aae7 100644 --- a/arch/sh/oprofile/Makefile +++ b/arch/sh/oprofile/Makefile @@ -6,4 +6,8 @@ DRIVER_OBJS = $(addprefix ../../../drivers/oprofile/, \ oprofilefs.o oprofile_stats.o \ timer_int.o ) +ifeq ($(CONFIG_HW_PERF_EVENTS),y) +DRIVER_OBJS += $(addprefix ../../../drivers/oprofile/, oprofile_perf.o) +endif + oprofile-y := $(DRIVER_OBJS) common.o backtrace.o diff --git a/arch/sh/oprofile/common.c b/arch/sh/oprofile/common.c index ac60493..e10d893 100644 --- a/arch/sh/oprofile/common.c +++ b/arch/sh/oprofile/common.c @@ -17,114 +17,45 @@ #include <linux/init.h> #include <linux/errno.h> #include <linux/smp.h> +#include <linux/perf_event.h> #include <asm/processor.h> -#include "op_impl.h" - -static struct op_sh_model *model; - -static struct op_counter_config ctr[20]; +#ifdef CONFIG_HW_PERF_EVENTS extern void sh_backtrace(struct pt_regs * const regs, unsigned int depth); -static int op_sh_setup(void) -{ - /* Pre-compute the values to stuff in the hardware registers. */ - model->reg_setup(ctr); - - /* Configure the registers on all cpus. */ - on_each_cpu(model->cpu_setup, NULL, 1); - - return 0; -} - -static int op_sh_create_files(struct super_block *sb, struct dentry *root) +char *op_name_from_perf_id(void) { - int i, ret = 0; + const char *pmu; + char buf[20]; + int size; - for (i = 0; i < model->num_counters; i++) { - struct dentry *dir; - char buf[4]; + pmu = perf_pmu_name(); + if (!pmu) + return NULL; - snprintf(buf, sizeof(buf), "%d", i); - dir = oprofilefs_mkdir(sb, root, buf); + size = snprintf(buf, sizeof(buf), "sh/%s", pmu); + if (size > -1 && size < sizeof(buf)) + return buf; - ret |= oprofilefs_create_ulong(sb, dir, "enabled", &ctr[i].enabled); - ret |= oprofilefs_create_ulong(sb, dir, "event", &ctr[i].event); - ret |= oprofilefs_create_ulong(sb, dir, "kernel", &ctr[i].kernel); - ret |= oprofilefs_create_ulong(sb, dir, "user", &ctr[i].user); - - if (model->create_files) - ret |= model->create_files(sb, dir); - else - ret |= oprofilefs_create_ulong(sb, dir, "count", &ctr[i].count); - - /* Dummy entries */ - ret |= oprofilefs_create_ulong(sb, dir, "unit_mask", &ctr[i].unit_mask); - } - - return ret; + return NULL; } -static int op_sh_start(void) +int __init oprofile_arch_init(struct oprofile_operations *ops) { - /* Enable performance monitoring for all counters. */ - on_each_cpu(model->cpu_start, NULL, 1); + ops->backtrace = sh_backtrace; - return 0; + return oprofile_perf_init(ops); } -static void op_sh_stop(void) +void __exit oprofile_arch_exit(void) { - /* Disable performance monitoring for all counters. */ - on_each_cpu(model->cpu_stop, NULL, 1); + oprofile_perf_exit(); } - +#else int __init oprofile_arch_init(struct oprofile_operations *ops) { - struct op_sh_model *lmodel = NULL; - int ret; - - /* - * Always assign the backtrace op. If the counter initialization - * fails, we fall back to the timer which will still make use of - * this. - */ - ops->backtrace = sh_backtrace; - - /* - * XXX - * - * All of the SH7750/SH-4A counters have been converted to perf, - * this infrastructure hook is left for other users until they've - * had a chance to convert over, at which point all of this - * will be deleted. - */ - - if (!lmodel) - return -ENODEV; - if (!(current_cpu_data.flags & CPU_HAS_PERF_COUNTER)) - return -ENODEV; - - ret = lmodel->init(); - if (unlikely(ret != 0)) - return ret; - - model = lmodel; - - ops->setup = op_sh_setup; - ops->create_files = op_sh_create_files; - ops->start = op_sh_start; - ops->stop = op_sh_stop; - ops->cpu_type = lmodel->cpu_type; - - printk(KERN_INFO "oprofile: using %s performance monitoring.\n", - lmodel->cpu_type); - - return 0; -} - -void oprofile_arch_exit(void) -{ - if (model && model->exit) - model->exit(); + pr_info("oprofile: hardware counters not available\n"); + return -ENODEV; } +void __exit oprofile_arch_exit(void) {} +#endif /* CONFIG_HW_PERF_EVENTS */ diff --git a/arch/sh/oprofile/op_impl.h b/arch/sh/oprofile/op_impl.h deleted file mode 100644 index 1244479..0000000 --- a/arch/sh/oprofile/op_impl.h +++ /dev/null @@ -1,33 +0,0 @@ -#ifndef __OP_IMPL_H -#define __OP_IMPL_H - -/* Per-counter configuration as set via oprofilefs. */ -struct op_counter_config { - unsigned long enabled; - unsigned long event; - - unsigned long count; - - /* Dummy values for userspace tool compliance */ - unsigned long kernel; - unsigned long user; - unsigned long unit_mask; -}; - -/* Per-architecture configury and hooks. */ -struct op_sh_model { - void (*reg_setup)(struct op_counter_config *); - int (*create_files)(struct super_block *sb, struct dentry *dir); - void (*cpu_setup)(void *dummy); - int (*init)(void); - void (*exit)(void); - void (*cpu_start)(void *args); - void (*cpu_stop)(void *args); - char *cpu_type; - unsigned char num_counters; -}; - -/* arch/sh/oprofile/common.c */ -extern void sh_backtrace(struct pt_regs * const regs, unsigned int depth); - -#endif /* __OP_IMPL_H */ diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig index 491e9d6..3e9d314 100644 --- a/arch/sparc/Kconfig +++ b/arch/sparc/Kconfig @@ -26,10 +26,12 @@ config SPARC select ARCH_WANT_OPTIONAL_GPIOLIB select RTC_CLASS select RTC_DRV_M48T59 + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC select HAVE_DMA_ATTRS select HAVE_DMA_API_DEBUG + select HAVE_ARCH_JUMP_LABEL config SPARC32 def_bool !64BIT @@ -53,6 +55,7 @@ config SPARC64 select RTC_DRV_BQ4802 select RTC_DRV_SUN4V select RTC_DRV_STARFIRE + select HAVE_IRQ_WORK select HAVE_PERF_EVENTS select PERF_USE_VMALLOC diff --git a/arch/sparc/include/asm/jump_label.h b/arch/sparc/include/asm/jump_label.h new file mode 100644 index 0000000..62e66d7 --- /dev/null +++ b/arch/sparc/include/asm/jump_label.h @@ -0,0 +1,32 @@ +#ifndef _ASM_SPARC_JUMP_LABEL_H +#define _ASM_SPARC_JUMP_LABEL_H + +#ifdef __KERNEL__ + +#include <linux/types.h> +#include <asm/system.h> + +#define JUMP_LABEL_NOP_SIZE 4 + +#define JUMP_LABEL(key, label) \ + do { \ + asm goto("1:\n\t" \ + "nop\n\t" \ + "nop\n\t" \ + ".pushsection __jump_table, \"a\"\n\t"\ + ".word 1b, %l[" #label "], %c0\n\t" \ + ".popsection \n\t" \ + : : "i" (key) : : label);\ + } while (0) + +#endif /* __KERNEL__ */ + +typedef u32 jump_label_t; + +struct jump_entry { + jump_label_t code; + jump_label_t target; + jump_label_t key; +}; + +#endif diff --git a/arch/sparc/include/asm/perf_event.h b/arch/sparc/include/asm/perf_event.h index 727af70..6e8bfa1 100644 --- a/arch/sparc/include/asm/perf_event.h +++ b/arch/sparc/include/asm/perf_event.h @@ -1,10 +1,6 @@ #ifndef __ASM_SPARC_PERF_EVENT_H #define __ASM_SPARC_PERF_EVENT_H -extern void set_perf_event_pending(void); - -#define PERF_EVENT_INDEX_OFFSET 0 - #ifdef CONFIG_PERF_EVENTS #include <asm/ptrace.h> diff --git a/arch/sparc/kernel/Makefile b/arch/sparc/kernel/Makefile index 0c2dc1f..599398f 100644 --- a/arch/sparc/kernel/Makefile +++ b/arch/sparc/kernel/Makefile @@ -119,3 +119,5 @@ obj-$(CONFIG_COMPAT) += $(audit--y) pc--$(CONFIG_PERF_EVENTS) := perf_event.o obj-$(CONFIG_SPARC64) += $(pc--y) + +obj-$(CONFIG_SPARC64) += jump_label.o diff --git a/arch/sparc/kernel/jump_label.c b/arch/sparc/kernel/jump_label.c new file mode 100644 index 0000000..ea2dafc --- /dev/null +++ b/arch/sparc/kernel/jump_label.c @@ -0,0 +1,47 @@ +#include <linux/kernel.h> +#include <linux/types.h> +#include <linux/mutex.h> +#include <linux/cpu.h> + +#include <linux/jump_label.h> +#include <linux/memory.h> + +#ifdef HAVE_JUMP_LABEL + +void arch_jump_label_transform(struct jump_entry *entry, + enum jump_label_type type) +{ + u32 val; + u32 *insn = (u32 *) (unsigned long) entry->code; + + if (type == JUMP_LABEL_ENABLE) { + s32 off = (s32)entry->target - (s32)entry->code; + +#ifdef CONFIG_SPARC64 + /* ba,pt %xcc, . + (off << 2) */ + val = 0x10680000 | ((u32) off >> 2); +#else + /* ba . + (off << 2) */ + val = 0x10800000 | ((u32) off >> 2); +#endif + } else { + val = 0x01000000; + } + + get_online_cpus(); + mutex_lock(&text_mutex); + *insn = val; + flushi(insn); + mutex_unlock(&text_mutex); + put_online_cpus(); +} + +void arch_jump_label_text_poke_early(jump_label_t addr) +{ + u32 *insn_p = (u32 *) (unsigned long) addr; + + *insn_p = 0x01000000; + flushi(insn_p); +} + +#endif diff --git a/arch/sparc/kernel/module.c b/arch/sparc/kernel/module.c index f848aad..ee3c7dd 100644 --- a/arch/sparc/kernel/module.c +++ b/arch/sparc/kernel/module.c @@ -18,6 +18,9 @@ #include <asm/spitfire.h> #ifdef CONFIG_SPARC64 + +#include <linux/jump_label.h> + static void *module_map(unsigned long size) { struct vm_struct *area; @@ -227,6 +230,9 @@ int module_finalize(const Elf_Ehdr *hdr, const Elf_Shdr *sechdrs, struct module *me) { + /* make jump label nops */ + jump_label_apply_nops(me); + /* Cheetah's I-cache is fully coherent. */ if (tlb_type == spitfire) { unsigned long va; diff --git a/arch/sparc/kernel/pcr.c b/arch/sparc/kernel/pcr.c index c4a6a50..b87873c 100644 --- a/arch/sparc/kernel/pcr.c +++ b/arch/sparc/kernel/pcr.c @@ -7,7 +7,7 @@ #include <linux/init.h> #include <linux/irq.h> -#include <linux/perf_event.h> +#include <linux/irq_work.h> #include <linux/ftrace.h> #include <asm/pil.h> @@ -43,14 +43,14 @@ void __irq_entry deferred_pcr_work_irq(int irq, struct pt_regs *regs) old_regs = set_irq_regs(regs); irq_enter(); -#ifdef CONFIG_PERF_EVENTS - perf_event_do_pending(); +#ifdef CONFIG_IRQ_WORK + irq_work_run(); #endif irq_exit(); set_irq_regs(old_regs); } -void set_perf_event_pending(void) +void arch_irq_work_raise(void) { set_softint(1 << PIL_DEFERRED_PCR_WORK); } diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c index 6318e62..0d6deb5 100644 --- a/arch/sparc/kernel/perf_event.c +++ b/arch/sparc/kernel/perf_event.c @@ -658,13 +658,16 @@ static u64 maybe_change_configuration(struct cpu_hw_events *cpuc, u64 pcr) enc = perf_event_get_enc(cpuc->events[i]); pcr &= ~mask_for_index(idx); - pcr |= event_encoding(enc, idx); + if (hwc->state & PERF_HES_STOPPED) + pcr |= nop_for_index(idx); + else + pcr |= event_encoding(enc, idx); } out: return pcr; } -void hw_perf_enable(void) +static void sparc_pmu_enable(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); u64 pcr; @@ -691,7 +694,7 @@ void hw_perf_enable(void) pcr_ops->write(cpuc->pcr); } -void hw_perf_disable(void) +static void sparc_pmu_disable(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); u64 val; @@ -710,19 +713,65 @@ void hw_perf_disable(void) pcr_ops->write(cpuc->pcr); } -static void sparc_pmu_disable(struct perf_event *event) +static int active_event_index(struct cpu_hw_events *cpuc, + struct perf_event *event) +{ + int i; + + for (i = 0; i < cpuc->n_events; i++) { + if (cpuc->event[i] == event) + break; + } + BUG_ON(i == cpuc->n_events); + return cpuc->current_idx[i]; +} + +static void sparc_pmu_start(struct perf_event *event, int flags) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + int idx = active_event_index(cpuc, event); + + if (flags & PERF_EF_RELOAD) { + WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE)); + sparc_perf_event_set_period(event, &event->hw, idx); + } + + event->hw.state = 0; + + sparc_pmu_enable_event(cpuc, &event->hw, idx); +} + +static void sparc_pmu_stop(struct perf_event *event, int flags) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + int idx = active_event_index(cpuc, event); + + if (!(event->hw.state & PERF_HES_STOPPED)) { + sparc_pmu_disable_event(cpuc, &event->hw, idx); + event->hw.state |= PERF_HES_STOPPED; + } + + if (!(event->hw.state & PERF_HES_UPTODATE) && (flags & PERF_EF_UPDATE)) { + sparc_perf_event_update(event, &event->hw, idx); + event->hw.state |= PERF_HES_UPTODATE; + } +} + +static void sparc_pmu_del(struct perf_event *event, int _flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - struct hw_perf_event *hwc = &event->hw; unsigned long flags; int i; local_irq_save(flags); - perf_disable(); + perf_pmu_disable(event->pmu); for (i = 0; i < cpuc->n_events; i++) { if (event == cpuc->event[i]) { - int idx = cpuc->current_idx[i]; + /* Absorb the final count and turn off the + * event. + */ + sparc_pmu_stop(event, PERF_EF_UPDATE); /* Shift remaining entries down into * the existing slot. @@ -734,13 +783,6 @@ static void sparc_pmu_disable(struct perf_event *event) cpuc->current_idx[i]; } - /* Absorb the final count and turn off the - * event. - */ - sparc_pmu_disable_event(cpuc, hwc, idx); - barrier(); - sparc_perf_event_update(event, hwc, idx); - perf_event_update_userpage(event); cpuc->n_events--; @@ -748,23 +790,10 @@ static void sparc_pmu_disable(struct perf_event *event) } } - perf_enable(); + perf_pmu_enable(event->pmu); local_irq_restore(flags); } -static int active_event_index(struct cpu_hw_events *cpuc, - struct perf_event *event) -{ - int i; - - for (i = 0; i < cpuc->n_events; i++) { - if (cpuc->event[i] == event) - break; - } - BUG_ON(i == cpuc->n_events); - return cpuc->current_idx[i]; -} - static void sparc_pmu_read(struct perf_event *event) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); @@ -774,15 +803,6 @@ static void sparc_pmu_read(struct perf_event *event) sparc_perf_event_update(event, hwc, idx); } -static void sparc_pmu_unthrottle(struct perf_event *event) -{ - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - int idx = active_event_index(cpuc, event); - struct hw_perf_event *hwc = &event->hw; - - sparc_pmu_enable_event(cpuc, hwc, idx); -} - static atomic_t active_events = ATOMIC_INIT(0); static DEFINE_MUTEX(pmc_grab_mutex); @@ -877,7 +897,7 @@ static int sparc_check_constraints(struct perf_event **evts, if (!n_ev) return 0; - if (n_ev > perf_max_events) + if (n_ev > MAX_HWEVENTS) return -1; msk0 = perf_event_get_msk(events[0]); @@ -984,23 +1004,27 @@ static int collect_events(struct perf_event *group, int max_count, return n; } -static int sparc_pmu_enable(struct perf_event *event) +static int sparc_pmu_add(struct perf_event *event, int ef_flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int n0, ret = -EAGAIN; unsigned long flags; local_irq_save(flags); - perf_disable(); + perf_pmu_disable(event->pmu); n0 = cpuc->n_events; - if (n0 >= perf_max_events) + if (n0 >= MAX_HWEVENTS) goto out; cpuc->event[n0] = event; cpuc->events[n0] = event->hw.event_base; cpuc->current_idx[n0] = PIC_NO_INDEX; + event->hw.state = PERF_HES_UPTODATE; + if (!(ef_flags & PERF_EF_START)) + event->hw.state |= PERF_HES_STOPPED; + /* * If group events scheduling transaction was started, * skip the schedulability test here, it will be peformed @@ -1020,12 +1044,12 @@ nocheck: ret = 0; out: - perf_enable(); + perf_pmu_enable(event->pmu); local_irq_restore(flags); return ret; } -static int __hw_perf_event_init(struct perf_event *event) +static int sparc_pmu_event_init(struct perf_event *event) { struct perf_event_attr *attr = &event->attr; struct perf_event *evts[MAX_HWEVENTS]; @@ -1038,22 +1062,33 @@ static int __hw_perf_event_init(struct perf_event *event) if (atomic_read(&nmi_active) < 0) return -ENODEV; - pmap = NULL; - if (attr->type == PERF_TYPE_HARDWARE) { + switch (attr->type) { + case PERF_TYPE_HARDWARE: if (attr->config >= sparc_pmu->max_events) return -EINVAL; pmap = sparc_pmu->event_map(attr->config); - } else if (attr->type == PERF_TYPE_HW_CACHE) { + break; + + case PERF_TYPE_HW_CACHE: pmap = sparc_map_cache_event(attr->config); if (IS_ERR(pmap)) return PTR_ERR(pmap); - } else if (attr->type != PERF_TYPE_RAW) - return -EOPNOTSUPP; + break; + + case PERF_TYPE_RAW: + pmap = NULL; + break; + + default: + return -ENOENT; + + } if (pmap) { hwc->event_base = perf_event_encode(pmap); } else { - /* User gives us "(encoding << 16) | pic_mask" for + /* + * User gives us "(encoding << 16) | pic_mask" for * PERF_TYPE_RAW events. */ hwc->event_base = attr->config; @@ -1071,7 +1106,7 @@ static int __hw_perf_event_init(struct perf_event *event) n = 0; if (event->group_leader != event) { n = collect_events(event->group_leader, - perf_max_events - 1, + MAX_HWEVENTS - 1, evts, events, current_idx_dmy); if (n < 0) return -EINVAL; @@ -1107,10 +1142,11 @@ static int __hw_perf_event_init(struct perf_event *event) * Set the flag to make pmu::enable() not perform the * schedulability test, it will be performed at commit time */ -static void sparc_pmu_start_txn(const struct pmu *pmu) +static void sparc_pmu_start_txn(struct pmu *pmu) { struct cpu_hw_events *cpuhw = &__get_cpu_var(cpu_hw_events); + perf_pmu_disable(pmu); cpuhw->group_flag |= PERF_EVENT_TXN; } @@ -1119,11 +1155,12 @@ static void sparc_pmu_start_txn(const struct pmu *pmu) * Clear the flag and pmu::enable() will perform the * schedulability test. */ -static void sparc_pmu_cancel_txn(const struct pmu *pmu) +static void sparc_pmu_cancel_txn(struct pmu *pmu) { struct cpu_hw_events *cpuhw = &__get_cpu_var(cpu_hw_events); cpuhw->group_flag &= ~PERF_EVENT_TXN; + perf_pmu_enable(pmu); } /* @@ -1131,7 +1168,7 @@ static void sparc_pmu_cancel_txn(const struct pmu *pmu) * Perform the group schedulability test as a whole * Return 0 if success */ -static int sparc_pmu_commit_txn(const struct pmu *pmu) +static int sparc_pmu_commit_txn(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int n; @@ -1147,28 +1184,24 @@ static int sparc_pmu_commit_txn(const struct pmu *pmu) return -EAGAIN; cpuc->group_flag &= ~PERF_EVENT_TXN; + perf_pmu_enable(pmu); return 0; } -static const struct pmu pmu = { - .enable = sparc_pmu_enable, - .disable = sparc_pmu_disable, +static struct pmu pmu = { + .pmu_enable = sparc_pmu_enable, + .pmu_disable = sparc_pmu_disable, + .event_init = sparc_pmu_event_init, + .add = sparc_pmu_add, + .del = sparc_pmu_del, + .start = sparc_pmu_start, + .stop = sparc_pmu_stop, .read = sparc_pmu_read, - .unthrottle = sparc_pmu_unthrottle, .start_txn = sparc_pmu_start_txn, .cancel_txn = sparc_pmu_cancel_txn, .commit_txn = sparc_pmu_commit_txn, }; -const struct pmu *hw_perf_event_init(struct perf_event *event) -{ - int err = __hw_perf_event_init(event); - - if (err) - return ERR_PTR(err); - return &pmu; -} - void perf_event_print_debug(void) { unsigned long flags; @@ -1244,7 +1277,7 @@ static int __kprobes perf_event_nmi_handler(struct notifier_block *self, continue; if (perf_event_overflow(event, 1, &data, regs)) - sparc_pmu_disable_event(cpuc, hwc, idx); + sparc_pmu_stop(event, 0); } return NOTIFY_STOP; @@ -1285,28 +1318,21 @@ void __init init_hw_perf_events(void) pr_cont("Supported PMU type is '%s'\n", sparc_pmu_type); - /* All sparc64 PMUs currently have 2 events. */ - perf_max_events = 2; - + perf_pmu_register(&pmu); register_die_notifier(&perf_event_nmi_notifier); } -static inline void callchain_store(struct perf_callchain_entry *entry, u64 ip) -{ - if (entry->nr < PERF_MAX_STACK_DEPTH) - entry->ip[entry->nr++] = ip; -} - -static void perf_callchain_kernel(struct pt_regs *regs, - struct perf_callchain_entry *entry) +void perf_callchain_kernel(struct perf_callchain_entry *entry, + struct pt_regs *regs) { unsigned long ksp, fp; #ifdef CONFIG_FUNCTION_GRAPH_TRACER int graph = 0; #endif - callchain_store(entry, PERF_CONTEXT_KERNEL); - callchain_store(entry, regs->tpc); + stack_trace_flush(); + + perf_callchain_store(entry, regs->tpc); ksp = regs->u_regs[UREG_I6]; fp = ksp + STACK_BIAS; @@ -1330,13 +1356,13 @@ static void perf_callchain_kernel(struct pt_regs *regs, pc = sf->callers_pc; fp = (unsigned long)sf->fp + STACK_BIAS; } - callchain_store(entry, pc); + perf_callchain_store(entry, pc); #ifdef CONFIG_FUNCTION_GRAPH_TRACER if ((pc + 8UL) == (unsigned long) &return_to_handler) { int index = current->curr_ret_stack; if (current->ret_stack && index >= graph) { pc = current->ret_stack[index - graph].ret; - callchain_store(entry, pc); + perf_callchain_store(entry, pc); graph++; } } @@ -1344,13 +1370,12 @@ static void perf_callchain_kernel(struct pt_regs *regs, } while (entry->nr < PERF_MAX_STACK_DEPTH); } -static void perf_callchain_user_64(struct pt_regs *regs, - struct perf_callchain_entry *entry) +static void perf_callchain_user_64(struct perf_callchain_entry *entry, + struct pt_regs *regs) { unsigned long ufp; - callchain_store(entry, PERF_CONTEXT_USER); - callchain_store(entry, regs->tpc); + perf_callchain_store(entry, regs->tpc); ufp = regs->u_regs[UREG_I6] + STACK_BIAS; do { @@ -1363,17 +1388,16 @@ static void perf_callchain_user_64(struct pt_regs *regs, pc = sf.callers_pc; ufp = (unsigned long)sf.fp + STACK_BIAS; - callchain_store(entry, pc); + perf_callchain_store(entry, pc); } while (entry->nr < PERF_MAX_STACK_DEPTH); } -static void perf_callchain_user_32(struct pt_regs *regs, - struct perf_callchain_entry *entry) +static void perf_callchain_user_32(struct perf_callchain_entry *entry, + struct pt_regs *regs) { unsigned long ufp; - callchain_store(entry, PERF_CONTEXT_USER); - callchain_store(entry, regs->tpc); + perf_callchain_store(entry, regs->tpc); ufp = regs->u_regs[UREG_I6] & 0xffffffffUL; do { @@ -1386,34 +1410,16 @@ static void perf_callchain_user_32(struct pt_regs *regs, pc = sf.callers_pc; ufp = (unsigned long)sf.fp; - callchain_store(entry, pc); + perf_callchain_store(entry, pc); } while (entry->nr < PERF_MAX_STACK_DEPTH); } -/* Like powerpc we can't get PMU interrupts within the PMU handler, - * so no need for separate NMI and IRQ chains as on x86. - */ -static DEFINE_PER_CPU(struct perf_callchain_entry, callchain); - -struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) +void +perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs) { - struct perf_callchain_entry *entry = &__get_cpu_var(callchain); - - entry->nr = 0; - if (!user_mode(regs)) { - stack_trace_flush(); - perf_callchain_kernel(regs, entry); - if (current->mm) - regs = task_pt_regs(current); - else - regs = NULL; - } - if (regs) { - flushw_user(); - if (test_thread_flag(TIF_32BIT)) - perf_callchain_user_32(regs, entry); - else - perf_callchain_user_64(regs, entry); - } - return entry; + flushw_user(); + if (test_thread_flag(TIF_32BIT)) + perf_callchain_user_32(entry, regs); + else + perf_callchain_user_64(entry, regs); } diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index cea0cd9..fd227d6 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -25,6 +25,7 @@ config X86 select HAVE_IDE select HAVE_OPROFILE select HAVE_PERF_EVENTS if (!M386 && !M486) + select HAVE_IRQ_WORK select HAVE_IOREMAP_PROT select HAVE_KPROBES select ARCH_WANT_OPTIONAL_GPIOLIB @@ -33,6 +34,7 @@ config X86 select HAVE_KRETPROBES select HAVE_OPTPROBES select HAVE_FTRACE_MCOUNT_RECORD + select HAVE_C_RECORDMCOUNT select HAVE_DYNAMIC_FTRACE select HAVE_FUNCTION_TRACER select HAVE_FUNCTION_GRAPH_TRACER @@ -59,6 +61,8 @@ config X86 select ANON_INODES select HAVE_ARCH_KMEMCHECK select HAVE_USER_RETURN_NOTIFIER + select HAVE_ARCH_JUMP_LABEL + select HAVE_TEXT_POKE_SMP config INSTRUCTION_DECODER def_bool (KPROBES || PERF_EVENTS) @@ -2125,6 +2129,10 @@ config HAVE_ATOMIC_IOMAP def_bool y depends on X86_32 +config HAVE_TEXT_POKE_SMP + bool + select STOP_MACHINE if SMP + source "net/Kconfig" source "drivers/Kconfig" diff --git a/arch/x86/include/asm/alternative.h b/arch/x86/include/asm/alternative.h index bc6abb7..76561d2 100644 --- a/arch/x86/include/asm/alternative.h +++ b/arch/x86/include/asm/alternative.h @@ -4,6 +4,7 @@ #include <linux/types.h> #include <linux/stddef.h> #include <linux/stringify.h> +#include <linux/jump_label.h> #include <asm/asm.h> /* @@ -160,6 +161,8 @@ static inline void apply_paravirt(struct paravirt_patch_site *start, #define __parainstructions_end NULL #endif +extern void *text_poke_early(void *addr, const void *opcode, size_t len); + /* * Clear and restore the kernel write-protection flag on the local CPU. * Allows the kernel to edit read-only pages. @@ -180,4 +183,12 @@ static inline void apply_paravirt(struct paravirt_patch_site *start, extern void *text_poke(void *addr, const void *opcode, size_t len); extern void *text_poke_smp(void *addr, const void *opcode, size_t len); +#if defined(CONFIG_DYNAMIC_FTRACE) || defined(HAVE_JUMP_LABEL) +#define IDEAL_NOP_SIZE_5 5 +extern unsigned char ideal_nop5[IDEAL_NOP_SIZE_5]; +extern void arch_init_ideal_nop5(void); +#else +static inline void arch_init_ideal_nop5(void) {} +#endif + #endif /* _ASM_X86_ALTERNATIVE_H */ diff --git a/arch/x86/include/asm/entry_arch.h b/arch/x86/include/asm/entry_arch.h index 8e8ec66..b8e96a1 100644 --- a/arch/x86/include/asm/entry_arch.h +++ b/arch/x86/include/asm/entry_arch.h @@ -49,8 +49,8 @@ BUILD_INTERRUPT(apic_timer_interrupt,LOCAL_TIMER_VECTOR) BUILD_INTERRUPT(error_interrupt,ERROR_APIC_VECTOR) BUILD_INTERRUPT(spurious_interrupt,SPURIOUS_APIC_VECTOR) -#ifdef CONFIG_PERF_EVENTS -BUILD_INTERRUPT(perf_pending_interrupt, LOCAL_PENDING_VECTOR) +#ifdef CONFIG_IRQ_WORK +BUILD_INTERRUPT(irq_work_interrupt, IRQ_WORK_VECTOR) #endif #ifdef CONFIG_X86_THERMAL_VECTOR diff --git a/arch/x86/include/asm/hardirq.h b/arch/x86/include/asm/hardirq.h index aeab29a..55e4de6 100644 --- a/arch/x86/include/asm/hardirq.h +++ b/arch/x86/include/asm/hardirq.h @@ -14,7 +14,7 @@ typedef struct { #endif unsigned int x86_platform_ipis; /* arch dependent */ unsigned int apic_perf_irqs; - unsigned int apic_pending_irqs; + unsigned int apic_irq_work_irqs; #ifdef CONFIG_SMP unsigned int irq_resched_count; unsigned int irq_call_count; diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index 46c0fe0..3a54a1c 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -29,7 +29,7 @@ extern void apic_timer_interrupt(void); extern void x86_platform_ipi(void); extern void error_interrupt(void); -extern void perf_pending_interrupt(void); +extern void irq_work_interrupt(void); extern void spurious_interrupt(void); extern void thermal_interrupt(void); diff --git a/arch/x86/include/asm/irq_vectors.h b/arch/x86/include/asm/irq_vectors.h index e2ca300..6af0894 100644 --- a/arch/x86/include/asm/irq_vectors.h +++ b/arch/x86/include/asm/irq_vectors.h @@ -114,9 +114,9 @@ #define X86_PLATFORM_IPI_VECTOR 0xed /* - * Performance monitoring pending work vector: + * IRQ work vector: */ -#define LOCAL_PENDING_VECTOR 0xec +#define IRQ_WORK_VECTOR 0xec #define UV_BAU_MESSAGE 0xea diff --git a/arch/x86/include/asm/jump_label.h b/arch/x86/include/asm/jump_label.h new file mode 100644 index 0000000..f52d42e --- /dev/null +++ b/arch/x86/include/asm/jump_label.h @@ -0,0 +1,37 @@ +#ifndef _ASM_X86_JUMP_LABEL_H +#define _ASM_X86_JUMP_LABEL_H + +#ifdef __KERNEL__ + +#include <linux/types.h> +#include <asm/nops.h> + +#define JUMP_LABEL_NOP_SIZE 5 + +# define JUMP_LABEL_INITIAL_NOP ".byte 0xe9 \n\t .long 0\n\t" + +# define JUMP_LABEL(key, label) \ + do { \ + asm goto("1:" \ + JUMP_LABEL_INITIAL_NOP \ + ".pushsection __jump_table, \"a\" \n\t"\ + _ASM_PTR "1b, %l[" #label "], %c0 \n\t" \ + ".popsection \n\t" \ + : : "i" (key) : : label); \ + } while (0) + +#endif /* __KERNEL__ */ + +#ifdef CONFIG_X86_64 +typedef u64 jump_label_t; +#else +typedef u32 jump_label_t; +#endif + +struct jump_entry { + jump_label_t code; + jump_label_t target; + jump_label_t key; +}; + +#endif diff --git a/arch/x86/include/asm/perf_event_p4.h b/arch/x86/include/asm/perf_event_p4.h index def5007..a70cd21 100644 --- a/arch/x86/include/asm/perf_event_p4.h +++ b/arch/x86/include/asm/perf_event_p4.h @@ -36,19 +36,6 @@ #define P4_ESCR_EMASK(v) ((v) << P4_ESCR_EVENTMASK_SHIFT) #define P4_ESCR_TAG(v) ((v) << P4_ESCR_TAG_SHIFT) -/* Non HT mask */ -#define P4_ESCR_MASK \ - (P4_ESCR_EVENT_MASK | \ - P4_ESCR_EVENTMASK_MASK | \ - P4_ESCR_TAG_MASK | \ - P4_ESCR_TAG_ENABLE | \ - P4_ESCR_T0_OS | \ - P4_ESCR_T0_USR) - -/* HT mask */ -#define P4_ESCR_MASK_HT \ - (P4_ESCR_MASK | P4_ESCR_T1_OS | P4_ESCR_T1_USR) - #define P4_CCCR_OVF 0x80000000U #define P4_CCCR_CASCADE 0x40000000U #define P4_CCCR_OVF_PMI_T0 0x04000000U @@ -70,23 +57,6 @@ #define P4_CCCR_THRESHOLD(v) ((v) << P4_CCCR_THRESHOLD_SHIFT) #define P4_CCCR_ESEL(v) ((v) << P4_CCCR_ESCR_SELECT_SHIFT) -/* Non HT mask */ -#define P4_CCCR_MASK \ - (P4_CCCR_OVF | \ - P4_CCCR_CASCADE | \ - P4_CCCR_OVF_PMI_T0 | \ - P4_CCCR_FORCE_OVF | \ - P4_CCCR_EDGE | \ - P4_CCCR_THRESHOLD_MASK | \ - P4_CCCR_COMPLEMENT | \ - P4_CCCR_COMPARE | \ - P4_CCCR_ESCR_SELECT_MASK | \ - P4_CCCR_ENABLE) - -/* HT mask */ -#define P4_CCCR_MASK_HT \ - (P4_CCCR_MASK | P4_CCCR_OVF_PMI_T1 | P4_CCCR_THREAD_ANY) - #define P4_GEN_ESCR_EMASK(class, name, bit) \ class##__##name = ((1 << bit) << P4_ESCR_EVENTMASK_SHIFT) #define P4_ESCR_EMASK_BIT(class, name) class##__##name @@ -127,6 +97,28 @@ #define P4_CONFIG_HT_SHIFT 63 #define P4_CONFIG_HT (1ULL << P4_CONFIG_HT_SHIFT) +/* + * The bits we allow to pass for RAW events + */ +#define P4_CONFIG_MASK_ESCR \ + P4_ESCR_EVENT_MASK | \ + P4_ESCR_EVENTMASK_MASK | \ + P4_ESCR_TAG_MASK | \ + P4_ESCR_TAG_ENABLE + +#define P4_CONFIG_MASK_CCCR \ + P4_CCCR_EDGE | \ + P4_CCCR_THRESHOLD_MASK | \ + P4_CCCR_COMPLEMENT | \ + P4_CCCR_COMPARE | \ + P4_CCCR_THREAD_ANY | \ + P4_CCCR_RESERVED + +/* some dangerous bits are reserved for kernel internals */ +#define P4_CONFIG_MASK \ + (p4_config_pack_escr(P4_CONFIG_MASK_ESCR)) | \ + (p4_config_pack_cccr(P4_CONFIG_MASK_CCCR)) + static inline bool p4_is_event_cascaded(u64 config) { u32 cccr = p4_config_unpack_cccr(config); diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index fedf32a..7490bf8 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -34,7 +34,8 @@ GCOV_PROFILE_paravirt.o := n obj-y := process_$(BITS).o signal.o entry_$(BITS).o obj-y += traps.o irq.o irq_$(BITS).o dumpstack_$(BITS).o obj-y += time.o ioport.o ldt.o dumpstack.o -obj-y += setup.o x86_init.o i8259.o irqinit.o +obj-y += setup.o x86_init.o i8259.o irqinit.o jump_label.o +obj-$(CONFIG_IRQ_WORK) += irq_work.o obj-$(CONFIG_X86_VISWS) += visws_quirks.o obj-$(CONFIG_X86_32) += probe_roms_32.o obj-$(CONFIG_X86_32) += sys_i386_32.o i386_ksyms_32.o diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index f65ab8b..a36bb90 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -195,7 +195,7 @@ static void __init_or_module add_nops(void *insns, unsigned int len) extern struct alt_instr __alt_instructions[], __alt_instructions_end[]; extern s32 __smp_locks[], __smp_locks_end[]; -static void *text_poke_early(void *addr, const void *opcode, size_t len); +void *text_poke_early(void *addr, const void *opcode, size_t len); /* Replace instructions with better alternatives for this CPU type. This runs before SMP is initialized to avoid SMP problems with @@ -522,7 +522,7 @@ void __init alternative_instructions(void) * instructions. And on the local CPU you need to be protected again NMI or MCE * handlers seeing an inconsistent instruction while you patch. */ -static void *__init_or_module text_poke_early(void *addr, const void *opcode, +void *__init_or_module text_poke_early(void *addr, const void *opcode, size_t len) { unsigned long flags; @@ -637,7 +637,72 @@ void *__kprobes text_poke_smp(void *addr, const void *opcode, size_t len) tpp.len = len; atomic_set(&stop_machine_first, 1); wrote_text = 0; - stop_machine(stop_machine_text_poke, (void *)&tpp, NULL); + /* Use __stop_machine() because the caller already got online_cpus. */ + __stop_machine(stop_machine_text_poke, (void *)&tpp, NULL); return addr; } +#if defined(CONFIG_DYNAMIC_FTRACE) || defined(HAVE_JUMP_LABEL) + +unsigned char ideal_nop5[IDEAL_NOP_SIZE_5]; + +void __init arch_init_ideal_nop5(void) +{ + extern const unsigned char ftrace_test_p6nop[]; + extern const unsigned char ftrace_test_nop5[]; + extern const unsigned char ftrace_test_jmp[]; + int faulted = 0; + + /* + * There is no good nop for all x86 archs. + * We will default to using the P6_NOP5, but first we + * will test to make sure that the nop will actually + * work on this CPU. If it faults, we will then + * go to a lesser efficient 5 byte nop. If that fails + * we then just use a jmp as our nop. This isn't the most + * efficient nop, but we can not use a multi part nop + * since we would then risk being preempted in the middle + * of that nop, and if we enabled tracing then, it might + * cause a system crash. + * + * TODO: check the cpuid to determine the best nop. + */ + asm volatile ( + "ftrace_test_jmp:" + "jmp ftrace_test_p6nop\n" + "nop\n" + "nop\n" + "nop\n" /* 2 byte jmp + 3 bytes */ + "ftrace_test_p6nop:" + P6_NOP5 + "jmp 1f\n" + "ftrace_test_nop5:" + ".byte 0x66,0x66,0x66,0x66,0x90\n" + "1:" + ".section .fixup, \"ax\"\n" + "2: movl $1, %0\n" + " jmp ftrace_test_nop5\n" + "3: movl $2, %0\n" + " jmp 1b\n" + ".previous\n" + _ASM_EXTABLE(ftrace_test_p6nop, 2b) + _ASM_EXTABLE(ftrace_test_nop5, 3b) + : "=r"(faulted) : "0" (faulted)); + + switch (faulted) { + case 0: + pr_info("converting mcount calls to 0f 1f 44 00 00\n"); + memcpy(ideal_nop5, ftrace_test_p6nop, IDEAL_NOP_SIZE_5); + break; + case 1: + pr_info("converting mcount calls to 66 66 66 66 90\n"); + memcpy(ideal_nop5, ftrace_test_nop5, IDEAL_NOP_SIZE_5); + break; + case 2: + pr_info("converting mcount calls to jmp . + 5\n"); + memcpy(ideal_nop5, ftrace_test_jmp, IDEAL_NOP_SIZE_5); + break; + } + +} +#endif diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 03a5b03..fe73c18 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -531,7 +531,7 @@ static int x86_pmu_hw_config(struct perf_event *event) /* * Setup the hardware configuration for a given attr_type */ -static int __hw_perf_event_init(struct perf_event *event) +static int __x86_pmu_event_init(struct perf_event *event) { int err; @@ -584,7 +584,7 @@ static void x86_pmu_disable_all(void) } } -void hw_perf_disable(void) +static void x86_pmu_disable(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); @@ -619,7 +619,7 @@ static void x86_pmu_enable_all(int added) } } -static const struct pmu pmu; +static struct pmu pmu; static inline int is_x86_event(struct perf_event *event) { @@ -801,10 +801,10 @@ static inline int match_prev_assignment(struct hw_perf_event *hwc, hwc->last_tag == cpuc->tags[i]; } -static int x86_pmu_start(struct perf_event *event); -static void x86_pmu_stop(struct perf_event *event); +static void x86_pmu_start(struct perf_event *event, int flags); +static void x86_pmu_stop(struct perf_event *event, int flags); -void hw_perf_enable(void) +static void x86_pmu_enable(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct perf_event *event; @@ -840,7 +840,14 @@ void hw_perf_enable(void) match_prev_assignment(hwc, cpuc, i)) continue; - x86_pmu_stop(event); + /* + * Ensure we don't accidentally enable a stopped + * counter simply because we rescheduled. + */ + if (hwc->state & PERF_HES_STOPPED) + hwc->state |= PERF_HES_ARCH; + + x86_pmu_stop(event, PERF_EF_UPDATE); } for (i = 0; i < cpuc->n_events; i++) { @@ -852,7 +859,10 @@ void hw_perf_enable(void) else if (i < n_running) continue; - x86_pmu_start(event); + if (hwc->state & PERF_HES_ARCH) + continue; + + x86_pmu_start(event, PERF_EF_RELOAD); } cpuc->n_added = 0; perf_events_lapic_init(); @@ -953,15 +963,12 @@ static void x86_pmu_enable_event(struct perf_event *event) } /* - * activate a single event + * Add a single event to the PMU. * * The event is added to the group of enabled events * but only if it can be scehduled with existing events. - * - * Called with PMU disabled. If successful and return value 1, - * then guaranteed to call perf_enable() and hw_perf_enable() */ -static int x86_pmu_enable(struct perf_event *event) +static int x86_pmu_add(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc; @@ -970,58 +977,67 @@ static int x86_pmu_enable(struct perf_event *event) hwc = &event->hw; + perf_pmu_disable(event->pmu); n0 = cpuc->n_events; - n = collect_events(cpuc, event, false); - if (n < 0) - return n; + ret = n = collect_events(cpuc, event, false); + if (ret < 0) + goto out; + + hwc->state = PERF_HES_UPTODATE | PERF_HES_STOPPED; + if (!(flags & PERF_EF_START)) + hwc->state |= PERF_HES_ARCH; /* * If group events scheduling transaction was started, * skip the schedulability test here, it will be peformed - * at commit time(->commit_txn) as a whole + * at commit time (->commit_txn) as a whole */ if (cpuc->group_flag & PERF_EVENT_TXN) - goto out; + goto done_collect; ret = x86_pmu.schedule_events(cpuc, n, assign); if (ret) - return ret; + goto out; /* * copy new assignment, now we know it is possible * will be used by hw_perf_enable() */ memcpy(cpuc->assign, assign, n*sizeof(int)); -out: +done_collect: cpuc->n_events = n; cpuc->n_added += n - n0; cpuc->n_txn += n - n0; - return 0; + ret = 0; +out: + perf_pmu_enable(event->pmu); + return ret; } -static int x86_pmu_start(struct perf_event *event) +static void x86_pmu_start(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int idx = event->hw.idx; - if (idx == -1) - return -EAGAIN; + if (WARN_ON_ONCE(!(event->hw.state & PERF_HES_STOPPED))) + return; + + if (WARN_ON_ONCE(idx == -1)) + return; + + if (flags & PERF_EF_RELOAD) { + WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE)); + x86_perf_event_set_period(event); + } + + event->hw.state = 0; - x86_perf_event_set_period(event); cpuc->events[idx] = event; __set_bit(idx, cpuc->active_mask); __set_bit(idx, cpuc->running); x86_pmu.enable(event); perf_event_update_userpage(event); - - return 0; -} - -static void x86_pmu_unthrottle(struct perf_event *event) -{ - int ret = x86_pmu_start(event); - WARN_ON_ONCE(ret); } void perf_event_print_debug(void) @@ -1078,27 +1094,29 @@ void perf_event_print_debug(void) local_irq_restore(flags); } -static void x86_pmu_stop(struct perf_event *event) +static void x86_pmu_stop(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct hw_perf_event *hwc = &event->hw; - int idx = hwc->idx; - if (!__test_and_clear_bit(idx, cpuc->active_mask)) - return; - - x86_pmu.disable(event); - - /* - * Drain the remaining delta count out of a event - * that we are disabling: - */ - x86_perf_event_update(event); + if (__test_and_clear_bit(hwc->idx, cpuc->active_mask)) { + x86_pmu.disable(event); + cpuc->events[hwc->idx] = NULL; + WARN_ON_ONCE(hwc->state & PERF_HES_STOPPED); + hwc->state |= PERF_HES_STOPPED; + } - cpuc->events[idx] = NULL; + if ((flags & PERF_EF_UPDATE) && !(hwc->state & PERF_HES_UPTODATE)) { + /* + * Drain the remaining delta count out of a event + * that we are disabling: + */ + x86_perf_event_update(event); + hwc->state |= PERF_HES_UPTODATE; + } } -static void x86_pmu_disable(struct perf_event *event) +static void x86_pmu_del(struct perf_event *event, int flags) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int i; @@ -1111,7 +1129,7 @@ static void x86_pmu_disable(struct perf_event *event) if (cpuc->group_flag & PERF_EVENT_TXN) return; - x86_pmu_stop(event); + x86_pmu_stop(event, PERF_EF_UPDATE); for (i = 0; i < cpuc->n_events; i++) { if (event == cpuc->event_list[i]) { @@ -1134,7 +1152,6 @@ static int x86_pmu_handle_irq(struct pt_regs *regs) struct perf_sample_data data; struct cpu_hw_events *cpuc; struct perf_event *event; - struct hw_perf_event *hwc; int idx, handled = 0; u64 val; @@ -1155,7 +1172,6 @@ static int x86_pmu_handle_irq(struct pt_regs *regs) } event = cpuc->events[idx]; - hwc = &event->hw; val = x86_perf_event_update(event); if (val & (1ULL << (x86_pmu.cntval_bits - 1))) @@ -1171,7 +1187,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs) continue; if (perf_event_overflow(event, 1, &data, regs)) - x86_pmu_stop(event); + x86_pmu_stop(event, 0); } if (handled) @@ -1180,25 +1196,6 @@ static int x86_pmu_handle_irq(struct pt_regs *regs) return handled; } -void smp_perf_pending_interrupt(struct pt_regs *regs) -{ - irq_enter(); - ack_APIC_irq(); - inc_irq_stat(apic_pending_irqs); - perf_event_do_pending(); - irq_exit(); -} - -void set_perf_event_pending(void) -{ -#ifdef CONFIG_X86_LOCAL_APIC - if (!x86_pmu.apic || !x86_pmu_initialized()) - return; - - apic->send_IPI_self(LOCAL_PENDING_VECTOR); -#endif -} - void perf_events_lapic_init(void) { if (!x86_pmu.apic || !x86_pmu_initialized()) @@ -1388,7 +1385,6 @@ void __init init_hw_perf_events(void) x86_pmu.num_counters = X86_PMC_MAX_GENERIC; } x86_pmu.intel_ctrl = (1 << x86_pmu.num_counters) - 1; - perf_max_events = x86_pmu.num_counters; if (x86_pmu.num_counters_fixed > X86_PMC_MAX_FIXED) { WARN(1, KERN_ERR "hw perf events fixed %d > max(%d), clipping!", @@ -1424,6 +1420,7 @@ void __init init_hw_perf_events(void) pr_info("... fixed-purpose events: %d\n", x86_pmu.num_counters_fixed); pr_info("... event mask: %016Lx\n", x86_pmu.intel_ctrl); + perf_pmu_register(&pmu); perf_cpu_notifier(x86_pmu_notifier); } @@ -1437,10 +1434,11 @@ static inline void x86_pmu_read(struct perf_event *event) * Set the flag to make pmu::enable() not perform the * schedulability test, it will be performed at commit time */ -static void x86_pmu_start_txn(const struct pmu *pmu) +static void x86_pmu_start_txn(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + perf_pmu_disable(pmu); cpuc->group_flag |= PERF_EVENT_TXN; cpuc->n_txn = 0; } @@ -1450,7 +1448,7 @@ static void x86_pmu_start_txn(const struct pmu *pmu) * Clear the flag and pmu::enable() will perform the * schedulability test. */ -static void x86_pmu_cancel_txn(const struct pmu *pmu) +static void x86_pmu_cancel_txn(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); @@ -1460,6 +1458,7 @@ static void x86_pmu_cancel_txn(const struct pmu *pmu) */ cpuc->n_added -= cpuc->n_txn; cpuc->n_events -= cpuc->n_txn; + perf_pmu_enable(pmu); } /* @@ -1467,7 +1466,7 @@ static void x86_pmu_cancel_txn(const struct pmu *pmu) * Perform the group schedulability test as a whole * Return 0 if success */ -static int x86_pmu_commit_txn(const struct pmu *pmu) +static int x86_pmu_commit_txn(struct pmu *pmu) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int assign[X86_PMC_IDX_MAX]; @@ -1489,22 +1488,10 @@ static int x86_pmu_commit_txn(const struct pmu *pmu) memcpy(cpuc->assign, assign, n*sizeof(int)); cpuc->group_flag &= ~PERF_EVENT_TXN; - + perf_pmu_enable(pmu); return 0; } -static const struct pmu pmu = { - .enable = x86_pmu_enable, - .disable = x86_pmu_disable, - .start = x86_pmu_start, - .stop = x86_pmu_stop, - .read = x86_pmu_read, - .unthrottle = x86_pmu_unthrottle, - .start_txn = x86_pmu_start_txn, - .cancel_txn = x86_pmu_cancel_txn, - .commit_txn = x86_pmu_commit_txn, -}; - /* * validate that we can schedule this event */ @@ -1579,12 +1566,22 @@ out: return ret; } -const struct pmu *hw_perf_event_init(struct perf_event *event) +int x86_pmu_event_init(struct perf_event *event) { - const struct pmu *tmp; + struct pmu *tmp; int err; - err = __hw_perf_event_init(event); + switch (event->attr.type) { + case PERF_TYPE_RAW: + case PERF_TYPE_HARDWARE: + case PERF_TYPE_HW_CACHE: + break; + + default: + return -ENOENT; + } + + err = __x86_pmu_event_init(event); if (!err) { /* * we temporarily connect event to its pmu @@ -1604,26 +1601,31 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) if (err) { if (event->destroy) event->destroy(event); - return ERR_PTR(err); } - return &pmu; + return err; } -/* - * callchain support - */ +static struct pmu pmu = { + .pmu_enable = x86_pmu_enable, + .pmu_disable = x86_pmu_disable, -static inline -void callchain_store(struct perf_callchain_entry *entry, u64 ip) -{ - if (entry->nr < PERF_MAX_STACK_DEPTH) - entry->ip[entry->nr++] = ip; -} + .event_init = x86_pmu_event_init, -static DEFINE_PER_CPU(struct perf_callchain_entry, pmc_irq_entry); -static DEFINE_PER_CPU(struct perf_callchain_entry, pmc_nmi_entry); + .add = x86_pmu_add, + .del = x86_pmu_del, + .start = x86_pmu_start, + .stop = x86_pmu_stop, + .read = x86_pmu_read, + .start_txn = x86_pmu_start_txn, + .cancel_txn = x86_pmu_cancel_txn, + .commit_txn = x86_pmu_commit_txn, +}; + +/* + * callchain support + */ static void backtrace_warning_symbol(void *data, char *msg, unsigned long symbol) @@ -1645,7 +1647,7 @@ static void backtrace_address(void *data, unsigned long addr, int reliable) { struct perf_callchain_entry *entry = data; - callchain_store(entry, addr); + perf_callchain_store(entry, addr); } static const struct stacktrace_ops backtrace_ops = { @@ -1656,11 +1658,15 @@ static const struct stacktrace_ops backtrace_ops = { .walk_stack = print_context_stack_bp, }; -static void -perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry) +void +perf_callchain_kernel(struct perf_callchain_entry *entry, struct pt_regs *regs) { - callchain_store(entry, PERF_CONTEXT_KERNEL); - callchain_store(entry, regs->ip); + if (perf_guest_cbs && perf_guest_cbs->is_in_guest()) { + /* TODO: We don't support guest os callchain now */ + return; + } + + perf_callchain_store(entry, regs->ip); dump_trace(NULL, regs, NULL, regs->bp, &backtrace_ops, entry); } @@ -1689,7 +1695,7 @@ perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry) if (fp < compat_ptr(regs->sp)) break; - callchain_store(entry, frame.return_address); + perf_callchain_store(entry, frame.return_address); fp = compat_ptr(frame.next_frame); } return 1; @@ -1702,19 +1708,20 @@ perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry) } #endif -static void -perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry) +void +perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs) { struct stack_frame frame; const void __user *fp; - if (!user_mode(regs)) - regs = task_pt_regs(current); + if (perf_guest_cbs && perf_guest_cbs->is_in_guest()) { + /* TODO: We don't support guest os callchain now */ + return; + } fp = (void __user *)regs->bp; - callchain_store(entry, PERF_CONTEXT_USER); - callchain_store(entry, regs->ip); + perf_callchain_store(entry, regs->ip); if (perf_callchain_user32(regs, entry)) return; @@ -1731,52 +1738,11 @@ perf_callchain_user(struct pt_regs *regs, struct perf_callchain_entry *entry) if ((unsigned long)fp < regs->sp) break; - callchain_store(entry, frame.return_address); + perf_callchain_store(entry, frame.return_address); fp = frame.next_frame; } } -static void -perf_do_callchain(struct pt_regs *regs, struct perf_callchain_entry *entry) -{ - int is_user; - - if (!regs) - return; - - is_user = user_mode(regs); - - if (is_user && current->state != TASK_RUNNING) - return; - - if (!is_user) - perf_callchain_kernel(regs, entry); - - if (current->mm) - perf_callchain_user(regs, entry); -} - -struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) -{ - struct perf_callchain_entry *entry; - - if (perf_guest_cbs && perf_guest_cbs->is_in_guest()) { - /* TODO: We don't support guest os callchain now */ - return NULL; - } - - if (in_nmi()) - entry = &__get_cpu_var(pmc_nmi_entry); - else - entry = &__get_cpu_var(pmc_irq_entry); - - entry->nr = 0; - - perf_do_callchain(regs, entry); - - return entry; -} - unsigned long perf_instruction_pointer(struct pt_regs *regs) { unsigned long ip; diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c index c2897b7..46d5844 100644 --- a/arch/x86/kernel/cpu/perf_event_amd.c +++ b/arch/x86/kernel/cpu/perf_event_amd.c @@ -52,7 +52,7 @@ static __initconst const u64 amd_hw_cache_event_ids [ C(DTLB) ] = { [ C(OP_READ) ] = { [ C(RESULT_ACCESS) ] = 0x0040, /* Data Cache Accesses */ - [ C(RESULT_MISS) ] = 0x0046, /* L1 DTLB and L2 DLTB Miss */ + [ C(RESULT_MISS) ] = 0x0746, /* L1_DTLB_AND_L2_DLTB_MISS.ALL */ }, [ C(OP_WRITE) ] = { [ C(RESULT_ACCESS) ] = 0, @@ -66,7 +66,7 @@ static __initconst const u64 amd_hw_cache_event_ids [ C(ITLB) ] = { [ C(OP_READ) ] = { [ C(RESULT_ACCESS) ] = 0x0080, /* Instruction fecthes */ - [ C(RESULT_MISS) ] = 0x0085, /* Instr. fetch ITLB misses */ + [ C(RESULT_MISS) ] = 0x0385, /* L1_ITLB_AND_L2_ITLB_MISS.ALL */ }, [ C(OP_WRITE) ] = { [ C(RESULT_ACCESS) ] = -1, diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index ee05c90..c8f5c08 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -713,18 +713,18 @@ static int intel_pmu_handle_irq(struct pt_regs *regs) struct cpu_hw_events *cpuc; int bit, loops; u64 status; - int handled = 0; + int handled; perf_sample_data_init(&data, 0); cpuc = &__get_cpu_var(cpu_hw_events); intel_pmu_disable_all(); - intel_pmu_drain_bts_buffer(); + handled = intel_pmu_drain_bts_buffer(); status = intel_pmu_get_status(); if (!status) { intel_pmu_enable_all(0); - return 0; + return handled; } loops = 0; @@ -763,7 +763,7 @@ again: data.period = event->hw.last_period; if (perf_event_overflow(event, 1, &data, regs)) - x86_pmu_stop(event); + x86_pmu_stop(event, 0); } /* diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c index 18018d1..4977f9c 100644 --- a/arch/x86/kernel/cpu/perf_event_intel_ds.c +++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c @@ -214,7 +214,7 @@ static void intel_pmu_disable_bts(void) update_debugctlmsr(debugctlmsr); } -static void intel_pmu_drain_bts_buffer(void) +static int intel_pmu_drain_bts_buffer(void) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct debug_store *ds = cpuc->ds; @@ -231,16 +231,16 @@ static void intel_pmu_drain_bts_buffer(void) struct pt_regs regs; if (!event) - return; + return 0; if (!ds) - return; + return 0; at = (struct bts_record *)(unsigned long)ds->bts_buffer_base; top = (struct bts_record *)(unsigned long)ds->bts_index; if (top <= at) - return; + return 0; ds->bts_index = ds->bts_buffer_base; @@ -256,7 +256,7 @@ static void intel_pmu_drain_bts_buffer(void) perf_prepare_sample(&header, &data, event, ®s); if (perf_output_begin(&handle, event, header.size * (top - at), 1, 1)) - return; + return 1; for (; at < top; at++) { data.ip = at->from; @@ -270,6 +270,7 @@ static void intel_pmu_drain_bts_buffer(void) /* There's new data available. */ event->hw.interrupts++; event->pending_kill = POLL_IN; + return 1; } /* @@ -491,7 +492,7 @@ static void __intel_pmu_pebs_event(struct perf_event *event, regs.flags &= ~PERF_EFLAGS_EXACT; if (perf_event_overflow(event, 1, &data, ®s)) - x86_pmu_stop(event); + x86_pmu_stop(event, 0); } static void intel_pmu_drain_pebs_core(struct pt_regs *iregs) diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c index 2490151..81400b9 100644 --- a/arch/x86/kernel/cpu/perf_event_p4.c +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -18,6 +18,8 @@ struct p4_event_bind { unsigned int opcode; /* Event code and ESCR selector */ unsigned int escr_msr[2]; /* ESCR MSR for this event */ + unsigned int escr_emask; /* valid ESCR EventMask bits */ + unsigned int shared; /* event is shared across threads */ char cntr[2][P4_CNTR_LIMIT]; /* counter index (offset), -1 on abscence */ }; @@ -66,231 +68,435 @@ static struct p4_event_bind p4_event_bind_map[] = { [P4_EVENT_TC_DELIVER_MODE] = { .opcode = P4_OPCODE(P4_EVENT_TC_DELIVER_MODE), .escr_msr = { MSR_P4_TC_ESCR0, MSR_P4_TC_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_TC_DELIVER_MODE, DD) | + P4_ESCR_EMASK_BIT(P4_EVENT_TC_DELIVER_MODE, DB) | + P4_ESCR_EMASK_BIT(P4_EVENT_TC_DELIVER_MODE, DI) | + P4_ESCR_EMASK_BIT(P4_EVENT_TC_DELIVER_MODE, BD) | + P4_ESCR_EMASK_BIT(P4_EVENT_TC_DELIVER_MODE, BB) | + P4_ESCR_EMASK_BIT(P4_EVENT_TC_DELIVER_MODE, BI) | + P4_ESCR_EMASK_BIT(P4_EVENT_TC_DELIVER_MODE, ID), + .shared = 1, .cntr = { {4, 5, -1}, {6, 7, -1} }, }, [P4_EVENT_BPU_FETCH_REQUEST] = { .opcode = P4_OPCODE(P4_EVENT_BPU_FETCH_REQUEST), .escr_msr = { MSR_P4_BPU_ESCR0, MSR_P4_BPU_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_BPU_FETCH_REQUEST, TCMISS), .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_ITLB_REFERENCE] = { .opcode = P4_OPCODE(P4_EVENT_ITLB_REFERENCE), .escr_msr = { MSR_P4_ITLB_ESCR0, MSR_P4_ITLB_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_ITLB_REFERENCE, HIT) | + P4_ESCR_EMASK_BIT(P4_EVENT_ITLB_REFERENCE, MISS) | + P4_ESCR_EMASK_BIT(P4_EVENT_ITLB_REFERENCE, HIT_UK), .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_MEMORY_CANCEL] = { .opcode = P4_OPCODE(P4_EVENT_MEMORY_CANCEL), .escr_msr = { MSR_P4_DAC_ESCR0, MSR_P4_DAC_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_MEMORY_CANCEL, ST_RB_FULL) | + P4_ESCR_EMASK_BIT(P4_EVENT_MEMORY_CANCEL, 64K_CONF), .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_MEMORY_COMPLETE] = { .opcode = P4_OPCODE(P4_EVENT_MEMORY_COMPLETE), .escr_msr = { MSR_P4_SAAT_ESCR0 , MSR_P4_SAAT_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_MEMORY_COMPLETE, LSC) | + P4_ESCR_EMASK_BIT(P4_EVENT_MEMORY_COMPLETE, SSC), .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_LOAD_PORT_REPLAY] = { .opcode = P4_OPCODE(P4_EVENT_LOAD_PORT_REPLAY), .escr_msr = { MSR_P4_SAAT_ESCR0, MSR_P4_SAAT_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_LOAD_PORT_REPLAY, SPLIT_LD), .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_STORE_PORT_REPLAY] = { .opcode = P4_OPCODE(P4_EVENT_STORE_PORT_REPLAY), .escr_msr = { MSR_P4_SAAT_ESCR0 , MSR_P4_SAAT_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_STORE_PORT_REPLAY, SPLIT_ST), .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_MOB_LOAD_REPLAY] = { .opcode = P4_OPCODE(P4_EVENT_MOB_LOAD_REPLAY), .escr_msr = { MSR_P4_MOB_ESCR0, MSR_P4_MOB_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_MOB_LOAD_REPLAY, NO_STA) | + P4_ESCR_EMASK_BIT(P4_EVENT_MOB_LOAD_REPLAY, NO_STD) | + P4_ESCR_EMASK_BIT(P4_EVENT_MOB_LOAD_REPLAY, PARTIAL_DATA) | + P4_ESCR_EMASK_BIT(P4_EVENT_MOB_LOAD_REPLAY, UNALGN_ADDR), .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_PAGE_WALK_TYPE] = { .opcode = P4_OPCODE(P4_EVENT_PAGE_WALK_TYPE), .escr_msr = { MSR_P4_PMH_ESCR0, MSR_P4_PMH_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_PAGE_WALK_TYPE, DTMISS) | + P4_ESCR_EMASK_BIT(P4_EVENT_PAGE_WALK_TYPE, ITMISS), + .shared = 1, .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_BSQ_CACHE_REFERENCE] = { .opcode = P4_OPCODE(P4_EVENT_BSQ_CACHE_REFERENCE), .escr_msr = { MSR_P4_BSU_ESCR0, MSR_P4_BSU_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITS) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITM) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITS) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITM) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_MISS) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_MISS) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, WR_2ndL_MISS), .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_IOQ_ALLOCATION] = { .opcode = P4_OPCODE(P4_EVENT_IOQ_ALLOCATION), .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, DEFAULT) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, ALL_READ) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, ALL_WRITE) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, MEM_UC) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, MEM_WC) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, MEM_WT) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, MEM_WP) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, MEM_WB) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, OWN) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, OTHER) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ALLOCATION, PREFETCH), .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_IOQ_ACTIVE_ENTRIES] = { /* shared ESCR */ .opcode = P4_OPCODE(P4_EVENT_IOQ_ACTIVE_ENTRIES), .escr_msr = { MSR_P4_FSB_ESCR1, MSR_P4_FSB_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, DEFAULT) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, ALL_READ) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, ALL_WRITE) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_UC) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_WC) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_WT) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_WP) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_WB) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, OWN) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, OTHER) | + P4_ESCR_EMASK_BIT(P4_EVENT_IOQ_ACTIVE_ENTRIES, PREFETCH), .cntr = { {2, -1, -1}, {3, -1, -1} }, }, [P4_EVENT_FSB_DATA_ACTIVITY] = { .opcode = P4_OPCODE(P4_EVENT_FSB_DATA_ACTIVITY), .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_DRV) | + P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_OWN) | + P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_OTHER) | + P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DBSY_DRV) | + P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DBSY_OWN) | + P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DBSY_OTHER), + .shared = 1, .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_BSQ_ALLOCATION] = { /* shared ESCR, broken CCCR1 */ .opcode = P4_OPCODE(P4_EVENT_BSQ_ALLOCATION), .escr_msr = { MSR_P4_BSU_ESCR0, MSR_P4_BSU_ESCR0 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_TYPE0) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_TYPE1) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_LEN0) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_LEN1) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_IO_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_LOCK_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_CACHE_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_SPLIT_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_DEM_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, REQ_ORD_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, MEM_TYPE0) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, MEM_TYPE1) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ALLOCATION, MEM_TYPE2), .cntr = { {0, -1, -1}, {1, -1, -1} }, }, [P4_EVENT_BSQ_ACTIVE_ENTRIES] = { /* shared ESCR */ .opcode = P4_OPCODE(P4_EVENT_BSQ_ACTIVE_ENTRIES), .escr_msr = { MSR_P4_BSU_ESCR1 , MSR_P4_BSU_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_TYPE0) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_TYPE1) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_LEN0) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_LEN1) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_IO_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_LOCK_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_CACHE_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_SPLIT_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_DEM_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_ORD_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, MEM_TYPE0) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, MEM_TYPE1) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_ACTIVE_ENTRIES, MEM_TYPE2), .cntr = { {2, -1, -1}, {3, -1, -1} }, }, [P4_EVENT_SSE_INPUT_ASSIST] = { .opcode = P4_OPCODE(P4_EVENT_SSE_INPUT_ASSIST), .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_SSE_INPUT_ASSIST, ALL), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_PACKED_SP_UOP] = { .opcode = P4_OPCODE(P4_EVENT_PACKED_SP_UOP), .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_PACKED_SP_UOP, ALL), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_PACKED_DP_UOP] = { .opcode = P4_OPCODE(P4_EVENT_PACKED_DP_UOP), .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_PACKED_DP_UOP, ALL), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_SCALAR_SP_UOP] = { .opcode = P4_OPCODE(P4_EVENT_SCALAR_SP_UOP), .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_SCALAR_SP_UOP, ALL), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_SCALAR_DP_UOP] = { .opcode = P4_OPCODE(P4_EVENT_SCALAR_DP_UOP), .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_SCALAR_DP_UOP, ALL), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_64BIT_MMX_UOP] = { .opcode = P4_OPCODE(P4_EVENT_64BIT_MMX_UOP), .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_64BIT_MMX_UOP, ALL), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_128BIT_MMX_UOP] = { .opcode = P4_OPCODE(P4_EVENT_128BIT_MMX_UOP), .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_128BIT_MMX_UOP, ALL), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_X87_FP_UOP] = { .opcode = P4_OPCODE(P4_EVENT_X87_FP_UOP), .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_X87_FP_UOP, ALL), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_TC_MISC] = { .opcode = P4_OPCODE(P4_EVENT_TC_MISC), .escr_msr = { MSR_P4_TC_ESCR0, MSR_P4_TC_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_TC_MISC, FLUSH), .cntr = { {4, 5, -1}, {6, 7, -1} }, }, [P4_EVENT_GLOBAL_POWER_EVENTS] = { .opcode = P4_OPCODE(P4_EVENT_GLOBAL_POWER_EVENTS), .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING), .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_TC_MS_XFER] = { .opcode = P4_OPCODE(P4_EVENT_TC_MS_XFER), .escr_msr = { MSR_P4_MS_ESCR0, MSR_P4_MS_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_TC_MS_XFER, CISC), .cntr = { {4, 5, -1}, {6, 7, -1} }, }, [P4_EVENT_UOP_QUEUE_WRITES] = { .opcode = P4_OPCODE(P4_EVENT_UOP_QUEUE_WRITES), .escr_msr = { MSR_P4_MS_ESCR0, MSR_P4_MS_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_UOP_QUEUE_WRITES, FROM_TC_BUILD) | + P4_ESCR_EMASK_BIT(P4_EVENT_UOP_QUEUE_WRITES, FROM_TC_DELIVER) | + P4_ESCR_EMASK_BIT(P4_EVENT_UOP_QUEUE_WRITES, FROM_ROM), .cntr = { {4, 5, -1}, {6, 7, -1} }, }, [P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE] = { .opcode = P4_OPCODE(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE), .escr_msr = { MSR_P4_TBPU_ESCR0 , MSR_P4_TBPU_ESCR0 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, CONDITIONAL) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, CALL) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, RETURN) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, INDIRECT), .cntr = { {4, 5, -1}, {6, 7, -1} }, }, [P4_EVENT_RETIRED_BRANCH_TYPE] = { .opcode = P4_OPCODE(P4_EVENT_RETIRED_BRANCH_TYPE), .escr_msr = { MSR_P4_TBPU_ESCR0 , MSR_P4_TBPU_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, CONDITIONAL) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, CALL) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, RETURN) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, INDIRECT), .cntr = { {4, 5, -1}, {6, 7, -1} }, }, [P4_EVENT_RESOURCE_STALL] = { .opcode = P4_OPCODE(P4_EVENT_RESOURCE_STALL), .escr_msr = { MSR_P4_ALF_ESCR0, MSR_P4_ALF_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_RESOURCE_STALL, SBFULL), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_WC_BUFFER] = { .opcode = P4_OPCODE(P4_EVENT_WC_BUFFER), .escr_msr = { MSR_P4_DAC_ESCR0, MSR_P4_DAC_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_WC_BUFFER, WCB_EVICTS) | + P4_ESCR_EMASK_BIT(P4_EVENT_WC_BUFFER, WCB_FULL_EVICTS), + .shared = 1, .cntr = { {8, 9, -1}, {10, 11, -1} }, }, [P4_EVENT_B2B_CYCLES] = { .opcode = P4_OPCODE(P4_EVENT_B2B_CYCLES), .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .escr_emask = 0, .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_BNR] = { .opcode = P4_OPCODE(P4_EVENT_BNR), .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .escr_emask = 0, .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_SNOOP] = { .opcode = P4_OPCODE(P4_EVENT_SNOOP), .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .escr_emask = 0, .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_RESPONSE] = { .opcode = P4_OPCODE(P4_EVENT_RESPONSE), .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .escr_emask = 0, .cntr = { {0, -1, -1}, {2, -1, -1} }, }, [P4_EVENT_FRONT_END_EVENT] = { .opcode = P4_OPCODE(P4_EVENT_FRONT_END_EVENT), .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_FRONT_END_EVENT, NBOGUS) | + P4_ESCR_EMASK_BIT(P4_EVENT_FRONT_END_EVENT, BOGUS), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_EXECUTION_EVENT] = { .opcode = P4_OPCODE(P4_EVENT_EXECUTION_EVENT), .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS0) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS1) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS2) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, NBOGUS3) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS0) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS1) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS2) | + P4_ESCR_EMASK_BIT(P4_EVENT_EXECUTION_EVENT, BOGUS3), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_REPLAY_EVENT] = { .opcode = P4_OPCODE(P4_EVENT_REPLAY_EVENT), .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_REPLAY_EVENT, NBOGUS) | + P4_ESCR_EMASK_BIT(P4_EVENT_REPLAY_EVENT, BOGUS), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_INSTR_RETIRED] = { .opcode = P4_OPCODE(P4_EVENT_INSTR_RETIRED), .escr_msr = { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_RETIRED, NBOGUSNTAG) | + P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_RETIRED, NBOGUSTAG) | + P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_RETIRED, BOGUSNTAG) | + P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_RETIRED, BOGUSTAG), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_UOPS_RETIRED] = { .opcode = P4_OPCODE(P4_EVENT_UOPS_RETIRED), .escr_msr = { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_UOPS_RETIRED, NBOGUS) | + P4_ESCR_EMASK_BIT(P4_EVENT_UOPS_RETIRED, BOGUS), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_UOP_TYPE] = { .opcode = P4_OPCODE(P4_EVENT_UOP_TYPE), .escr_msr = { MSR_P4_RAT_ESCR0, MSR_P4_RAT_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_UOP_TYPE, TAGLOADS) | + P4_ESCR_EMASK_BIT(P4_EVENT_UOP_TYPE, TAGSTORES), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_BRANCH_RETIRED] = { .opcode = P4_OPCODE(P4_EVENT_BRANCH_RETIRED), .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_BRANCH_RETIRED, MMNP) | + P4_ESCR_EMASK_BIT(P4_EVENT_BRANCH_RETIRED, MMNM) | + P4_ESCR_EMASK_BIT(P4_EVENT_BRANCH_RETIRED, MMTP) | + P4_ESCR_EMASK_BIT(P4_EVENT_BRANCH_RETIRED, MMTM), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_MISPRED_BRANCH_RETIRED] = { .opcode = P4_OPCODE(P4_EVENT_MISPRED_BRANCH_RETIRED), .escr_msr = { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_MISPRED_BRANCH_RETIRED, NBOGUS), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_X87_ASSIST] = { .opcode = P4_OPCODE(P4_EVENT_X87_ASSIST), .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_X87_ASSIST, FPSU) | + P4_ESCR_EMASK_BIT(P4_EVENT_X87_ASSIST, FPSO) | + P4_ESCR_EMASK_BIT(P4_EVENT_X87_ASSIST, POAO) | + P4_ESCR_EMASK_BIT(P4_EVENT_X87_ASSIST, POAU) | + P4_ESCR_EMASK_BIT(P4_EVENT_X87_ASSIST, PREA), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_MACHINE_CLEAR] = { .opcode = P4_OPCODE(P4_EVENT_MACHINE_CLEAR), .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_MACHINE_CLEAR, CLEAR) | + P4_ESCR_EMASK_BIT(P4_EVENT_MACHINE_CLEAR, MOCLEAR) | + P4_ESCR_EMASK_BIT(P4_EVENT_MACHINE_CLEAR, SMCLEAR), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, [P4_EVENT_INSTR_COMPLETED] = { .opcode = P4_OPCODE(P4_EVENT_INSTR_COMPLETED), .escr_msr = { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 }, + .escr_emask = + P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_COMPLETED, NBOGUS) | + P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_COMPLETED, BOGUS), .cntr = { {12, 13, 16}, {14, 15, 17} }, }, }; @@ -428,29 +634,73 @@ static u64 p4_pmu_event_map(int hw_event) return config; } +/* check cpu model specifics */ +static bool p4_event_match_cpu_model(unsigned int event_idx) +{ + /* INSTR_COMPLETED event only exist for model 3, 4, 6 (Prescott) */ + if (event_idx == P4_EVENT_INSTR_COMPLETED) { + if (boot_cpu_data.x86_model != 3 && + boot_cpu_data.x86_model != 4 && + boot_cpu_data.x86_model != 6) + return false; + } + + /* + * For info + * - IQ_ESCR0, IQ_ESCR1 only for models 1 and 2 + */ + + return true; +} + static int p4_validate_raw_event(struct perf_event *event) { - unsigned int v; + unsigned int v, emask; - /* user data may have out-of-bound event index */ + /* User data may have out-of-bound event index */ v = p4_config_unpack_event(event->attr.config); - if (v >= ARRAY_SIZE(p4_event_bind_map)) { - pr_warning("P4 PMU: Unknown event code: %d\n", v); + if (v >= ARRAY_SIZE(p4_event_bind_map)) + return -EINVAL; + + /* It may be unsupported: */ + if (!p4_event_match_cpu_model(v)) return -EINVAL; + + /* + * NOTE: P4_CCCR_THREAD_ANY has not the same meaning as + * in Architectural Performance Monitoring, it means not + * on _which_ logical cpu to count but rather _when_, ie it + * depends on logical cpu state -- count event if one cpu active, + * none, both or any, so we just allow user to pass any value + * desired. + * + * In turn we always set Tx_OS/Tx_USR bits bound to logical + * cpu without their propagation to another cpu + */ + + /* + * if an event is shared accross the logical threads + * the user needs special permissions to be able to use it + */ + if (p4_event_bind_map[v].shared) { + if (perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN)) + return -EACCES; } + /* ESCR EventMask bits may be invalid */ + emask = p4_config_unpack_escr(event->attr.config) & P4_ESCR_EVENTMASK_MASK; + if (emask & ~p4_event_bind_map[v].escr_emask) + return -EINVAL; + /* - * it may have some screwed PEBS bits + * it may have some invalid PEBS bits */ - if (p4_config_pebs_has(event->attr.config, P4_PEBS_CONFIG_ENABLE)) { - pr_warning("P4 PMU: PEBS are not supported yet\n"); + if (p4_config_pebs_has(event->attr.config, P4_PEBS_CONFIG_ENABLE)) return -EINVAL; - } + v = p4_config_unpack_metric(event->attr.config); - if (v >= ARRAY_SIZE(p4_pebs_bind_map)) { - pr_warning("P4 PMU: Unknown metric code: %d\n", v); + if (v >= ARRAY_SIZE(p4_pebs_bind_map)) return -EINVAL; - } return 0; } @@ -478,27 +728,21 @@ static int p4_hw_config(struct perf_event *event) if (event->attr.type == PERF_TYPE_RAW) { + /* + * Clear bits we reserve to be managed by kernel itself + * and never allowed from a user space + */ + event->attr.config &= P4_CONFIG_MASK; + rc = p4_validate_raw_event(event); if (rc) goto out; /* - * We don't control raw events so it's up to the caller - * to pass sane values (and we don't count the thread number - * on HT machine but allow HT-compatible specifics to be - * passed on) - * * Note that for RAW events we allow user to use P4_CCCR_RESERVED * bits since we keep additional info here (for cache events and etc) - * - * XXX: HT wide things should check perf_paranoid_cpu() && - * CAP_SYS_ADMIN */ - event->hw.config |= event->attr.config & - (p4_config_pack_escr(P4_ESCR_MASK_HT) | - p4_config_pack_cccr(P4_CCCR_MASK_HT | P4_CCCR_RESERVED)); - - event->hw.config &= ~P4_CCCR_FORCE_OVF; + event->hw.config |= event->attr.config; } rc = x86_setup_perfctr(event); diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 17be5ec..c375c79 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -1023,9 +1023,9 @@ apicinterrupt ERROR_APIC_VECTOR \ apicinterrupt SPURIOUS_APIC_VECTOR \ spurious_interrupt smp_spurious_interrupt -#ifdef CONFIG_PERF_EVENTS -apicinterrupt LOCAL_PENDING_VECTOR \ - perf_pending_interrupt smp_perf_pending_interrupt +#ifdef CONFIG_IRQ_WORK +apicinterrupt IRQ_WORK_VECTOR \ + irq_work_interrupt smp_irq_work_interrupt #endif /* diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index cd37469..3afb33f 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -257,14 +257,9 @@ do_ftrace_mod_code(unsigned long ip, void *new_code) return mod_code_status; } - - - -static unsigned char ftrace_nop[MCOUNT_INSN_SIZE]; - static unsigned char *ftrace_nop_replace(void) { - return ftrace_nop; + return ideal_nop5; } static int @@ -338,62 +333,6 @@ int ftrace_update_ftrace_func(ftrace_func_t func) int __init ftrace_dyn_arch_init(void *data) { - extern const unsigned char ftrace_test_p6nop[]; - extern const unsigned char ftrace_test_nop5[]; - extern const unsigned char ftrace_test_jmp[]; - int faulted = 0; - - /* - * There is no good nop for all x86 archs. - * We will default to using the P6_NOP5, but first we - * will test to make sure that the nop will actually - * work on this CPU. If it faults, we will then - * go to a lesser efficient 5 byte nop. If that fails - * we then just use a jmp as our nop. This isn't the most - * efficient nop, but we can not use a multi part nop - * since we would then risk being preempted in the middle - * of that nop, and if we enabled tracing then, it might - * cause a system crash. - * - * TODO: check the cpuid to determine the best nop. - */ - asm volatile ( - "ftrace_test_jmp:" - "jmp ftrace_test_p6nop\n" - "nop\n" - "nop\n" - "nop\n" /* 2 byte jmp + 3 bytes */ - "ftrace_test_p6nop:" - P6_NOP5 - "jmp 1f\n" - "ftrace_test_nop5:" - ".byte 0x66,0x66,0x66,0x66,0x90\n" - "1:" - ".section .fixup, \"ax\"\n" - "2: movl $1, %0\n" - " jmp ftrace_test_nop5\n" - "3: movl $2, %0\n" - " jmp 1b\n" - ".previous\n" - _ASM_EXTABLE(ftrace_test_p6nop, 2b) - _ASM_EXTABLE(ftrace_test_nop5, 3b) - : "=r"(faulted) : "0" (faulted)); - - switch (faulted) { - case 0: - pr_info("converting mcount calls to 0f 1f 44 00 00\n"); - memcpy(ftrace_nop, ftrace_test_p6nop, MCOUNT_INSN_SIZE); - break; - case 1: - pr_info("converting mcount calls to 66 66 66 66 90\n"); - memcpy(ftrace_nop, ftrace_test_nop5, MCOUNT_INSN_SIZE); - break; - case 2: - pr_info("converting mcount calls to jmp . + 5\n"); - memcpy(ftrace_nop, ftrace_test_jmp, MCOUNT_INSN_SIZE); - break; - } - /* The return code is retured via data */ *(unsigned long *)data = 0; diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index 91fd0c7..44edb03 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -67,10 +67,10 @@ static int show_other_interrupts(struct seq_file *p, int prec) for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->apic_perf_irqs); seq_printf(p, " Performance monitoring interrupts\n"); - seq_printf(p, "%*s: ", prec, "PND"); + seq_printf(p, "%*s: ", prec, "IWI"); for_each_online_cpu(j) - seq_printf(p, "%10u ", irq_stats(j)->apic_pending_irqs); - seq_printf(p, " Performance pending work\n"); + seq_printf(p, "%10u ", irq_stats(j)->apic_irq_work_irqs); + seq_printf(p, " IRQ work interrupts\n"); #endif if (x86_platform_ipi_callback) { seq_printf(p, "%*s: ", prec, "PLT"); @@ -185,7 +185,7 @@ u64 arch_irq_stat_cpu(unsigned int cpu) sum += irq_stats(cpu)->apic_timer_irqs; sum += irq_stats(cpu)->irq_spurious_count; sum += irq_stats(cpu)->apic_perf_irqs; - sum += irq_stats(cpu)->apic_pending_irqs; + sum += irq_stats(cpu)->apic_irq_work_irqs; #endif if (x86_platform_ipi_callback) sum += irq_stats(cpu)->x86_platform_ipis; diff --git a/arch/x86/kernel/irq_work.c b/arch/x86/kernel/irq_work.c new file mode 100644 index 0000000..ca8f703 --- /dev/null +++ b/arch/x86/kernel/irq_work.c @@ -0,0 +1,30 @@ +/* + * x86 specific code for irq_work + * + * Copyright (C) 2010 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + */ + +#include <linux/kernel.h> +#include <linux/irq_work.h> +#include <linux/hardirq.h> +#include <asm/apic.h> + +void smp_irq_work_interrupt(struct pt_regs *regs) +{ + irq_enter(); + ack_APIC_irq(); + inc_irq_stat(apic_irq_work_irqs); + irq_work_run(); + irq_exit(); +} + +void arch_irq_work_raise(void) +{ +#ifdef CONFIG_X86_LOCAL_APIC + if (!cpu_has_apic) + return; + + apic->send_IPI_self(IRQ_WORK_VECTOR); + apic_wait_icr_idle(); +#endif +} diff --git a/arch/x86/kernel/irqinit.c b/arch/x86/kernel/irqinit.c index 990ae7c..713969b 100644 --- a/arch/x86/kernel/irqinit.c +++ b/arch/x86/kernel/irqinit.c @@ -224,9 +224,9 @@ static void __init apic_intr_init(void) alloc_intr_gate(SPURIOUS_APIC_VECTOR, spurious_interrupt); alloc_intr_gate(ERROR_APIC_VECTOR, error_interrupt); - /* Performance monitoring interrupts: */ -# ifdef CONFIG_PERF_EVENTS - alloc_intr_gate(LOCAL_PENDING_VECTOR, perf_pending_interrupt); + /* IRQ work interrupts: */ +# ifdef CONFIG_IRQ_WORK + alloc_intr_gate(IRQ_WORK_VECTOR, irq_work_interrupt); # endif #endif diff --git a/arch/x86/kernel/jump_label.c b/arch/x86/kernel/jump_label.c new file mode 100644 index 0000000..961b6b3 --- /dev/null +++ b/arch/x86/kernel/jump_label.c @@ -0,0 +1,50 @@ +/* + * jump label x86 support + * + * Copyright (C) 2009 Jason Baron <jbaron@redhat.com> + * + */ +#include <linux/jump_label.h> +#include <linux/memory.h> +#include <linux/uaccess.h> +#include <linux/module.h> +#include <linux/list.h> +#include <linux/jhash.h> +#include <linux/cpu.h> +#include <asm/kprobes.h> +#include <asm/alternative.h> + +#ifdef HAVE_JUMP_LABEL + +union jump_code_union { + char code[JUMP_LABEL_NOP_SIZE]; + struct { + char jump; + int offset; + } __attribute__((packed)); +}; + +void arch_jump_label_transform(struct jump_entry *entry, + enum jump_label_type type) +{ + union jump_code_union code; + + if (type == JUMP_LABEL_ENABLE) { + code.jump = 0xe9; + code.offset = entry->target - + (entry->code + JUMP_LABEL_NOP_SIZE); + } else + memcpy(&code, ideal_nop5, JUMP_LABEL_NOP_SIZE); + get_online_cpus(); + mutex_lock(&text_mutex); + text_poke_smp((void *)entry->code, &code, JUMP_LABEL_NOP_SIZE); + mutex_unlock(&text_mutex); + put_online_cpus(); +} + +void arch_jump_label_text_poke_early(jump_label_t addr) +{ + text_poke_early((void *)addr, ideal_nop5, JUMP_LABEL_NOP_SIZE); +} + +#endif diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c index 770ebfb..1cbd54c 100644 --- a/arch/x86/kernel/kprobes.c +++ b/arch/x86/kernel/kprobes.c @@ -230,9 +230,6 @@ static int recover_probed_instruction(kprobe_opcode_t *buf, unsigned long addr) return 0; } -/* Dummy buffers for kallsyms_lookup */ -static char __dummy_buf[KSYM_NAME_LEN]; - /* Check if paddr is at an instruction boundary */ static int __kprobes can_probe(unsigned long paddr) { @@ -241,7 +238,7 @@ static int __kprobes can_probe(unsigned long paddr) struct insn insn; kprobe_opcode_t buf[MAX_INSN_SIZE]; - if (!kallsyms_lookup(paddr, NULL, &offset, NULL, __dummy_buf)) + if (!kallsyms_lookup_size_offset(paddr, NULL, &offset)) return 0; /* Decode instructions */ @@ -1129,7 +1126,7 @@ static void __kprobes synthesize_set_arg1(kprobe_opcode_t *addr, *(unsigned long *)addr = val; } -void __kprobes kprobes_optinsn_template_holder(void) +static void __used __kprobes kprobes_optinsn_template_holder(void) { asm volatile ( ".global optprobe_template_entry\n" @@ -1221,7 +1218,8 @@ static int __kprobes copy_optimized_instructions(u8 *dest, u8 *src) } /* Check whether the address range is reserved */ if (ftrace_text_reserved(src, src + len - 1) || - alternatives_text_reserved(src, src + len - 1)) + alternatives_text_reserved(src, src + len - 1) || + jump_label_text_reserved(src, src + len - 1)) return -EBUSY; return len; @@ -1269,11 +1267,9 @@ static int __kprobes can_optimize(unsigned long paddr) unsigned long addr, size = 0, offset = 0; struct insn insn; kprobe_opcode_t buf[MAX_INSN_SIZE]; - /* Dummy buffers for lookup_symbol_attrs */ - static char __dummy_buf[KSYM_NAME_LEN]; /* Lookup symbol including addr */ - if (!kallsyms_lookup(paddr, &size, &offset, NULL, __dummy_buf)) + if (!kallsyms_lookup_size_offset(paddr, &size, &offset)) return 0; /* Check there is enough space for a relative jump. */ diff --git a/arch/x86/kernel/module.c b/arch/x86/kernel/module.c index 1c355c5..8f29560 100644 --- a/arch/x86/kernel/module.c +++ b/arch/x86/kernel/module.c @@ -239,6 +239,9 @@ int module_finalize(const Elf_Ehdr *hdr, apply_paravirt(pseg, pseg + para->sh_size); } + /* make jump label nops */ + jump_label_apply_nops(me); + return 0; } diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index c3a4fbb..00e1678 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -112,6 +112,7 @@ #include <asm/numa_64.h> #endif #include <asm/mce.h> +#include <asm/alternative.h> /* * end_pfn only includes RAM, while max_pfn_mapped includes all e820 entries. @@ -726,6 +727,7 @@ void __init setup_arch(char **cmdline_p) { int acpi = 0; int k8 = 0; + unsigned long flags; #ifdef CONFIG_X86_32 memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data)); @@ -1071,6 +1073,10 @@ void __init setup_arch(char **cmdline_p) x86_init.oem.banner(); mcheck_init(); + + local_irq_save(flags); + arch_init_ideal_nop5(); + local_irq_restore(flags); } #ifdef CONFIG_X86_32 diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c index 4c4508e..a24c6cf 100644 --- a/arch/x86/mm/fault.c +++ b/arch/x86/mm/fault.c @@ -251,6 +251,8 @@ static noinline __kprobes int vmalloc_fault(unsigned long address) if (!(address >= VMALLOC_START && address < VMALLOC_END)) return -1; + WARN_ON_ONCE(in_nmi()); + /* * Synchronize this task's top level page-table * with the 'reference' page table. @@ -369,6 +371,8 @@ static noinline __kprobes int vmalloc_fault(unsigned long address) if (!(address >= VMALLOC_START && address < VMALLOC_END)) return -1; + WARN_ON_ONCE(in_nmi()); + /* * Copy kernel mappings over when needed. This can also * happen within a race in page table update. In the later diff --git a/arch/x86/mm/kmemcheck/kmemcheck.c b/arch/x86/mm/kmemcheck/kmemcheck.c index b3b531a..d87dd6d 100644 --- a/arch/x86/mm/kmemcheck/kmemcheck.c +++ b/arch/x86/mm/kmemcheck/kmemcheck.c @@ -631,6 +631,8 @@ bool kmemcheck_fault(struct pt_regs *regs, unsigned long address, if (!pte) return false; + WARN_ON_ONCE(in_nmi()); + if (error_code & 2) kmemcheck_access(regs, address, KMEMCHECK_WRITE); else diff --git a/arch/x86/oprofile/backtrace.c b/arch/x86/oprofile/backtrace.c index 3855096..2d49d4e 100644 --- a/arch/x86/oprofile/backtrace.c +++ b/arch/x86/oprofile/backtrace.c @@ -14,6 +14,7 @@ #include <asm/ptrace.h> #include <asm/uaccess.h> #include <asm/stacktrace.h> +#include <linux/compat.h> static void backtrace_warning_symbol(void *data, char *msg, unsigned long symbol) @@ -48,14 +49,12 @@ static struct stacktrace_ops backtrace_ops = { .walk_stack = print_context_stack, }; -struct frame_head { - struct frame_head *bp; - unsigned long ret; -} __attribute__((packed)); - -static struct frame_head *dump_user_backtrace(struct frame_head *head) +#ifdef CONFIG_COMPAT +static struct stack_frame_ia32 * +dump_user_backtrace_32(struct stack_frame_ia32 *head) { - struct frame_head bufhead[2]; + struct stack_frame_ia32 bufhead[2]; + struct stack_frame_ia32 *fp; /* Also check accessibility of one struct frame_head beyond */ if (!access_ok(VERIFY_READ, head, sizeof(bufhead))) @@ -63,20 +62,66 @@ static struct frame_head *dump_user_backtrace(struct frame_head *head) if (__copy_from_user_inatomic(bufhead, head, sizeof(bufhead))) return NULL; - oprofile_add_trace(bufhead[0].ret); + fp = (struct stack_frame_ia32 *) compat_ptr(bufhead[0].next_frame); + + oprofile_add_trace(bufhead[0].return_address); + + /* frame pointers should strictly progress back up the stack + * (towards higher addresses) */ + if (head >= fp) + return NULL; + + return fp; +} + +static inline int +x86_backtrace_32(struct pt_regs * const regs, unsigned int depth) +{ + struct stack_frame_ia32 *head; + + /* User process is 32-bit */ + if (!current || !test_thread_flag(TIF_IA32)) + return 0; + + head = (struct stack_frame_ia32 *) regs->bp; + while (depth-- && head) + head = dump_user_backtrace_32(head); + + return 1; +} + +#else +static inline int +x86_backtrace_32(struct pt_regs * const regs, unsigned int depth) +{ + return 0; +} +#endif /* CONFIG_COMPAT */ + +static struct stack_frame *dump_user_backtrace(struct stack_frame *head) +{ + struct stack_frame bufhead[2]; + + /* Also check accessibility of one struct stack_frame beyond */ + if (!access_ok(VERIFY_READ, head, sizeof(bufhead))) + return NULL; + if (__copy_from_user_inatomic(bufhead, head, sizeof(bufhead))) + return NULL; + + oprofile_add_trace(bufhead[0].return_address); /* frame pointers should strictly progress back up the stack * (towards higher addresses) */ - if (head >= bufhead[0].bp) + if (head >= bufhead[0].next_frame) return NULL; - return bufhead[0].bp; + return bufhead[0].next_frame; } void x86_backtrace(struct pt_regs * const regs, unsigned int depth) { - struct frame_head *head = (struct frame_head *)frame_pointer(regs); + struct stack_frame *head = (struct stack_frame *)frame_pointer(regs); if (!user_mode_vm(regs)) { unsigned long stack = kernel_stack_pointer(regs); @@ -86,6 +131,9 @@ x86_backtrace(struct pt_regs * const regs, unsigned int depth) return; } + if (x86_backtrace_32(regs, depth)) + return; + while (depth-- && head) head = dump_user_backtrace(head); } diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index f1575c9..bd1489c 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -695,9 +695,6 @@ static int __init ppro_init(char **cpu_type) return 1; } -/* in order to get sysfs right */ -static int using_nmi; - int __init op_nmi_init(struct oprofile_operations *ops) { __u8 vendor = boot_cpu_data.x86_vendor; @@ -705,8 +702,6 @@ int __init op_nmi_init(struct oprofile_operations *ops) char *cpu_type = NULL; int ret = 0; - using_nmi = 0; - if (!cpu_has_apic) return -ENODEV; @@ -790,13 +785,11 @@ int __init op_nmi_init(struct oprofile_operations *ops) if (ret) return ret; - using_nmi = 1; printk(KERN_INFO "oprofile: using NMI interrupt.\n"); return 0; } void op_nmi_exit(void) { - if (using_nmi) - exit_sysfs(); + exit_sysfs(); } diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c index b336cd9..f9bda64 100644 --- a/drivers/oprofile/oprof.c +++ b/drivers/oprofile/oprof.c @@ -225,26 +225,17 @@ post_sync: mutex_unlock(&start_mutex); } -int oprofile_set_backtrace(unsigned long val) +int oprofile_set_ulong(unsigned long *addr, unsigned long val) { - int err = 0; + int err = -EBUSY; mutex_lock(&start_mutex); - - if (oprofile_started) { - err = -EBUSY; - goto out; - } - - if (!oprofile_ops.backtrace) { - err = -EINVAL; - goto out; + if (!oprofile_started) { + *addr = val; + err = 0; } - - oprofile_backtrace_depth = val; - -out: mutex_unlock(&start_mutex); + return err; } @@ -257,16 +248,9 @@ static int __init oprofile_init(void) printk(KERN_INFO "oprofile: using timer interrupt.\n"); err = oprofile_timer_init(&oprofile_ops); if (err) - goto out_arch; + return err; } - err = oprofilefs_register(); - if (err) - goto out_arch; - return 0; - -out_arch: - oprofile_arch_exit(); - return err; + return oprofilefs_register(); } diff --git a/drivers/oprofile/oprof.h b/drivers/oprofile/oprof.h index 47e12cb..177b73d 100644 --- a/drivers/oprofile/oprof.h +++ b/drivers/oprofile/oprof.h @@ -37,7 +37,7 @@ void oprofile_create_files(struct super_block *sb, struct dentry *root); int oprofile_timer_init(struct oprofile_operations *ops); void oprofile_timer_exit(void); -int oprofile_set_backtrace(unsigned long depth); +int oprofile_set_ulong(unsigned long *addr, unsigned long val); int oprofile_set_timeout(unsigned long time); #endif /* OPROF_H */ diff --git a/drivers/oprofile/oprofile_files.c b/drivers/oprofile/oprofile_files.c index bbd7516..ccf099e 100644 --- a/drivers/oprofile/oprofile_files.c +++ b/drivers/oprofile/oprofile_files.c @@ -79,14 +79,17 @@ static ssize_t depth_write(struct file *file, char const __user *buf, size_t cou if (*offset) return -EINVAL; + if (!oprofile_ops.backtrace) + return -EINVAL; + retval = oprofilefs_ulong_from_user(&val, buf, count); if (retval) return retval; - retval = oprofile_set_backtrace(val); - + retval = oprofile_set_ulong(&oprofile_backtrace_depth, val); if (retval) return retval; + return count; } diff --git a/drivers/oprofile/oprofile_perf.c b/drivers/oprofile/oprofile_perf.c new file mode 100644 index 0000000..9046f7b --- /dev/null +++ b/drivers/oprofile/oprofile_perf.c @@ -0,0 +1,328 @@ +/* + * Copyright 2010 ARM Ltd. + * + * Perf-events backend for OProfile. + */ +#include <linux/perf_event.h> +#include <linux/platform_device.h> +#include <linux/oprofile.h> +#include <linux/slab.h> + +/* + * Per performance monitor configuration as set via oprofilefs. + */ +struct op_counter_config { + unsigned long count; + unsigned long enabled; + unsigned long event; + unsigned long unit_mask; + unsigned long kernel; + unsigned long user; + struct perf_event_attr attr; +}; + +static int oprofile_perf_enabled; +static DEFINE_MUTEX(oprofile_perf_mutex); + +static struct op_counter_config *counter_config; +static struct perf_event **perf_events[nr_cpumask_bits]; +static int num_counters; + +/* + * Overflow callback for oprofile. + */ +static void op_overflow_handler(struct perf_event *event, int unused, + struct perf_sample_data *data, struct pt_regs *regs) +{ + int id; + u32 cpu = smp_processor_id(); + + for (id = 0; id < num_counters; ++id) + if (perf_events[cpu][id] == event) + break; + + if (id != num_counters) + oprofile_add_sample(regs, id); + else + pr_warning("oprofile: ignoring spurious overflow " + "on cpu %u\n", cpu); +} + +/* + * Called by oprofile_perf_setup to create perf attributes to mirror the oprofile + * settings in counter_config. Attributes are created as `pinned' events and + * so are permanently scheduled on the PMU. + */ +static void op_perf_setup(void) +{ + int i; + u32 size = sizeof(struct perf_event_attr); + struct perf_event_attr *attr; + + for (i = 0; i < num_counters; ++i) { + attr = &counter_config[i].attr; + memset(attr, 0, size); + attr->type = PERF_TYPE_RAW; + attr->size = size; + attr->config = counter_config[i].event; + attr->sample_period = counter_config[i].count; + attr->pinned = 1; + } +} + +static int op_create_counter(int cpu, int event) +{ + struct perf_event *pevent; + + if (!counter_config[event].enabled || perf_events[cpu][event]) + return 0; + + pevent = perf_event_create_kernel_counter(&counter_config[event].attr, + cpu, NULL, + op_overflow_handler); + + if (IS_ERR(pevent)) + return PTR_ERR(pevent); + + if (pevent->state != PERF_EVENT_STATE_ACTIVE) { + perf_event_release_kernel(pevent); + pr_warning("oprofile: failed to enable event %d " + "on CPU %d\n", event, cpu); + return -EBUSY; + } + + perf_events[cpu][event] = pevent; + + return 0; +} + +static void op_destroy_counter(int cpu, int event) +{ + struct perf_event *pevent = perf_events[cpu][event]; + + if (pevent) { + perf_event_release_kernel(pevent); + perf_events[cpu][event] = NULL; + } +} + +/* + * Called by oprofile_perf_start to create active perf events based on the + * perviously configured attributes. + */ +static int op_perf_start(void) +{ + int cpu, event, ret = 0; + + for_each_online_cpu(cpu) { + for (event = 0; event < num_counters; ++event) { + ret = op_create_counter(cpu, event); + if (ret) + return ret; + } + } + + return ret; +} + +/* + * Called by oprofile_perf_stop at the end of a profiling run. + */ +static void op_perf_stop(void) +{ + int cpu, event; + + for_each_online_cpu(cpu) + for (event = 0; event < num_counters; ++event) + op_destroy_counter(cpu, event); +} + +static int oprofile_perf_create_files(struct super_block *sb, struct dentry *root) +{ + unsigned int i; + + for (i = 0; i < num_counters; i++) { + struct dentry *dir; + char buf[4]; + + snprintf(buf, sizeof buf, "%d", i); + dir = oprofilefs_mkdir(sb, root, buf); + oprofilefs_create_ulong(sb, dir, "enabled", &counter_config[i].enabled); + oprofilefs_create_ulong(sb, dir, "event", &counter_config[i].event); + oprofilefs_create_ulong(sb, dir, "count", &counter_config[i].count); + oprofilefs_create_ulong(sb, dir, "unit_mask", &counter_config[i].unit_mask); + oprofilefs_create_ulong(sb, dir, "kernel", &counter_config[i].kernel); + oprofilefs_create_ulong(sb, dir, "user", &counter_config[i].user); + } + + return 0; +} + +static int oprofile_perf_setup(void) +{ + spin_lock(&oprofilefs_lock); + op_perf_setup(); + spin_unlock(&oprofilefs_lock); + return 0; +} + +static int oprofile_perf_start(void) +{ + int ret = -EBUSY; + + mutex_lock(&oprofile_perf_mutex); + if (!oprofile_perf_enabled) { + ret = 0; + op_perf_start(); + oprofile_perf_enabled = 1; + } + mutex_unlock(&oprofile_perf_mutex); + return ret; +} + +static void oprofile_perf_stop(void) +{ + mutex_lock(&oprofile_perf_mutex); + if (oprofile_perf_enabled) + op_perf_stop(); + oprofile_perf_enabled = 0; + mutex_unlock(&oprofile_perf_mutex); +} + +#ifdef CONFIG_PM + +static int oprofile_perf_suspend(struct platform_device *dev, pm_message_t state) +{ + mutex_lock(&oprofile_perf_mutex); + if (oprofile_perf_enabled) + op_perf_stop(); + mutex_unlock(&oprofile_perf_mutex); + return 0; +} + +static int oprofile_perf_resume(struct platform_device *dev) +{ + mutex_lock(&oprofile_perf_mutex); + if (oprofile_perf_enabled && op_perf_start()) + oprofile_perf_enabled = 0; + mutex_unlock(&oprofile_perf_mutex); + return 0; +} + +static struct platform_driver oprofile_driver = { + .driver = { + .name = "oprofile-perf", + }, + .resume = oprofile_perf_resume, + .suspend = oprofile_perf_suspend, +}; + +static struct platform_device *oprofile_pdev; + +static int __init init_driverfs(void) +{ + int ret; + + ret = platform_driver_register(&oprofile_driver); + if (ret) + return ret; + + oprofile_pdev = platform_device_register_simple( + oprofile_driver.driver.name, 0, NULL, 0); + if (IS_ERR(oprofile_pdev)) { + ret = PTR_ERR(oprofile_pdev); + platform_driver_unregister(&oprofile_driver); + } + + return ret; +} + +static void exit_driverfs(void) +{ + platform_device_unregister(oprofile_pdev); + platform_driver_unregister(&oprofile_driver); +} + +#else + +static inline int init_driverfs(void) { return 0; } +static inline void exit_driverfs(void) { } + +#endif /* CONFIG_PM */ + +void oprofile_perf_exit(void) +{ + int cpu, id; + struct perf_event *event; + + for_each_possible_cpu(cpu) { + for (id = 0; id < num_counters; ++id) { + event = perf_events[cpu][id]; + if (event) + perf_event_release_kernel(event); + } + + kfree(perf_events[cpu]); + } + + kfree(counter_config); + exit_driverfs(); +} + +int __init oprofile_perf_init(struct oprofile_operations *ops) +{ + int cpu, ret = 0; + + ret = init_driverfs(); + if (ret) + return ret; + + memset(&perf_events, 0, sizeof(perf_events)); + + num_counters = perf_num_counters(); + if (num_counters <= 0) { + pr_info("oprofile: no performance counters\n"); + ret = -ENODEV; + goto out; + } + + counter_config = kcalloc(num_counters, + sizeof(struct op_counter_config), GFP_KERNEL); + + if (!counter_config) { + pr_info("oprofile: failed to allocate %d " + "counters\n", num_counters); + ret = -ENOMEM; + num_counters = 0; + goto out; + } + + for_each_possible_cpu(cpu) { + perf_events[cpu] = kcalloc(num_counters, + sizeof(struct perf_event *), GFP_KERNEL); + if (!perf_events[cpu]) { + pr_info("oprofile: failed to allocate %d perf events " + "for cpu %d\n", num_counters, cpu); + ret = -ENOMEM; + goto out; + } + } + + ops->create_files = oprofile_perf_create_files; + ops->setup = oprofile_perf_setup; + ops->start = oprofile_perf_start; + ops->stop = oprofile_perf_stop; + ops->shutdown = oprofile_perf_stop; + ops->cpu_type = op_name_from_perf_id(); + + if (!ops->cpu_type) + ret = -ENODEV; + else + pr_info("oprofile: using %s\n", ops->cpu_type); + +out: + if (ret) + oprofile_perf_exit(); + + return ret; +} diff --git a/drivers/oprofile/oprofilefs.c b/drivers/oprofile/oprofilefs.c index 2766a6d..1944621 100644 --- a/drivers/oprofile/oprofilefs.c +++ b/drivers/oprofile/oprofilefs.c @@ -91,16 +91,20 @@ static ssize_t ulong_read_file(struct file *file, char __user *buf, size_t count static ssize_t ulong_write_file(struct file *file, char const __user *buf, size_t count, loff_t *offset) { - unsigned long *value = file->private_data; + unsigned long value; int retval; if (*offset) return -EINVAL; - retval = oprofilefs_ulong_from_user(value, buf, count); + retval = oprofilefs_ulong_from_user(&value, buf, count); + if (retval) + return retval; + retval = oprofile_set_ulong(file->private_data, value); if (retval) return retval; + return count; } @@ -126,50 +130,41 @@ static const struct file_operations ulong_ro_fops = { }; -static struct dentry *__oprofilefs_create_file(struct super_block *sb, +static int __oprofilefs_create_file(struct super_block *sb, struct dentry *root, char const *name, const struct file_operations *fops, - int perm) + int perm, void *priv) { struct dentry *dentry; struct inode *inode; dentry = d_alloc_name(root, name); if (!dentry) - return NULL; + return -ENOMEM; inode = oprofilefs_get_inode(sb, S_IFREG | perm); if (!inode) { dput(dentry); - return NULL; + return -ENOMEM; } inode->i_fop = fops; d_add(dentry, inode); - return dentry; + dentry->d_inode->i_private = priv; + return 0; } int oprofilefs_create_ulong(struct super_block *sb, struct dentry *root, char const *name, unsigned long *val) { - struct dentry *d = __oprofilefs_create_file(sb, root, name, - &ulong_fops, 0644); - if (!d) - return -EFAULT; - - d->d_inode->i_private = val; - return 0; + return __oprofilefs_create_file(sb, root, name, + &ulong_fops, 0644, val); } int oprofilefs_create_ro_ulong(struct super_block *sb, struct dentry *root, char const *name, unsigned long *val) { - struct dentry *d = __oprofilefs_create_file(sb, root, name, - &ulong_ro_fops, 0444); - if (!d) - return -EFAULT; - - d->d_inode->i_private = val; - return 0; + return __oprofilefs_create_file(sb, root, name, + &ulong_ro_fops, 0444, val); } @@ -189,31 +184,22 @@ static const struct file_operations atomic_ro_fops = { int oprofilefs_create_ro_atomic(struct super_block *sb, struct dentry *root, char const *name, atomic_t *val) { - struct dentry *d = __oprofilefs_create_file(sb, root, name, - &atomic_ro_fops, 0444); - if (!d) - return -EFAULT; - - d->d_inode->i_private = val; - return 0; + return __oprofilefs_create_file(sb, root, name, + &atomic_ro_fops, 0444, val); } int oprofilefs_create_file(struct super_block *sb, struct dentry *root, char const *name, const struct file_operations *fops) { - if (!__oprofilefs_create_file(sb, root, name, fops, 0644)) - return -EFAULT; - return 0; + return __oprofilefs_create_file(sb, root, name, fops, 0644, NULL); } int oprofilefs_create_file_perm(struct super_block *sb, struct dentry *root, char const *name, const struct file_operations *fops, int perm) { - if (!__oprofilefs_create_file(sb, root, name, fops, perm)) - return -EFAULT; - return 0; + return __oprofilefs_create_file(sb, root, name, fops, perm, NULL); } diff --git a/include/asm-generic/hardirq.h b/include/asm-generic/hardirq.h index 62f5908..04d0a97 100644 --- a/include/asm-generic/hardirq.h +++ b/include/asm-generic/hardirq.h @@ -3,13 +3,13 @@ #include <linux/cache.h> #include <linux/threads.h> -#include <linux/irq.h> typedef struct { unsigned int __softirq_pending; } ____cacheline_aligned irq_cpustat_t; #include <linux/irq_cpustat.h> /* Standard mappings for irq_cpustat_t above */ +#include <linux/irq.h> #ifndef ack_bad_irq static inline void ack_bad_irq(unsigned int irq) diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 8a92a17..ef2af99 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -220,6 +220,8 @@ \ BUG_TABLE \ \ + JUMP_TABLE \ + \ /* PCI quirks */ \ .pci_fixup : AT(ADDR(.pci_fixup) - LOAD_OFFSET) { \ VMLINUX_SYMBOL(__start_pci_fixups_early) = .; \ @@ -563,6 +565,14 @@ #define BUG_TABLE #endif +#define JUMP_TABLE \ + . = ALIGN(8); \ + __jump_table : AT(ADDR(__jump_table) - LOAD_OFFSET) { \ + VMLINUX_SYMBOL(__start___jump_table) = .; \ + *(__jump_table) \ + VMLINUX_SYMBOL(__stop___jump_table) = .; \ + } + #ifdef CONFIG_PM_TRACE #define TRACEDATA \ . = ALIGN(4); \ diff --git a/include/linux/dynamic_debug.h b/include/linux/dynamic_debug.h index 52c0da4..bef3cda 100644 --- a/include/linux/dynamic_debug.h +++ b/include/linux/dynamic_debug.h @@ -1,6 +1,8 @@ #ifndef _DYNAMIC_DEBUG_H #define _DYNAMIC_DEBUG_H +#include <linux/jump_label.h> + /* dynamic_printk_enabled, and dynamic_printk_enabled2 are bitmasks in which * bit n is set to 1 if any modname hashes into the bucket n, 0 otherwise. They * use independent hash functions, to reduce the chance of false positives. @@ -22,8 +24,6 @@ struct _ddebug { const char *function; const char *filename; const char *format; - char primary_hash; - char secondary_hash; unsigned int lineno:24; /* * The flags field controls the behaviour at the callsite. @@ -33,6 +33,7 @@ struct _ddebug { #define _DPRINTK_FLAGS_PRINT (1<<0) /* printk() a message using the format */ #define _DPRINTK_FLAGS_DEFAULT 0 unsigned int flags:8; + char enabled; } __attribute__((aligned(8))); @@ -42,33 +43,35 @@ int ddebug_add_module(struct _ddebug *tab, unsigned int n, #if defined(CONFIG_DYNAMIC_DEBUG) extern int ddebug_remove_module(const char *mod_name); -#define __dynamic_dbg_enabled(dd) ({ \ - int __ret = 0; \ - if (unlikely((dynamic_debug_enabled & (1LL << DEBUG_HASH)) && \ - (dynamic_debug_enabled2 & (1LL << DEBUG_HASH2)))) \ - if (unlikely(dd.flags)) \ - __ret = 1; \ - __ret; }) - #define dynamic_pr_debug(fmt, ...) do { \ + __label__ do_printk; \ + __label__ out; \ static struct _ddebug descriptor \ __used \ __attribute__((section("__verbose"), aligned(8))) = \ - { KBUILD_MODNAME, __func__, __FILE__, fmt, DEBUG_HASH, \ - DEBUG_HASH2, __LINE__, _DPRINTK_FLAGS_DEFAULT }; \ - if (__dynamic_dbg_enabled(descriptor)) \ - printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__); \ + { KBUILD_MODNAME, __func__, __FILE__, fmt, __LINE__, \ + _DPRINTK_FLAGS_DEFAULT }; \ + JUMP_LABEL(&descriptor.enabled, do_printk); \ + goto out; \ +do_printk: \ + printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__); \ +out: ; \ } while (0) #define dynamic_dev_dbg(dev, fmt, ...) do { \ + __label__ do_printk; \ + __label__ out; \ static struct _ddebug descriptor \ __used \ __attribute__((section("__verbose"), aligned(8))) = \ - { KBUILD_MODNAME, __func__, __FILE__, fmt, DEBUG_HASH, \ - DEBUG_HASH2, __LINE__, _DPRINTK_FLAGS_DEFAULT }; \ - if (__dynamic_dbg_enabled(descriptor)) \ - dev_printk(KERN_DEBUG, dev, fmt, ##__VA_ARGS__); \ + { KBUILD_MODNAME, __func__, __FILE__, fmt, __LINE__, \ + _DPRINTK_FLAGS_DEFAULT }; \ + JUMP_LABEL(&descriptor.enabled, do_printk); \ + goto out; \ +do_printk: \ + dev_printk(KERN_DEBUG, dev, fmt, ##__VA_ARGS__); \ +out: ; \ } while (0) #else diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h index 02b8b24..8beabb9 100644 --- a/include/linux/ftrace_event.h +++ b/include/linux/ftrace_event.h @@ -191,8 +191,8 @@ struct ftrace_event_call { unsigned int flags; #ifdef CONFIG_PERF_EVENTS - int perf_refcount; - struct hlist_head *perf_events; + int perf_refcount; + struct hlist_head __percpu *perf_events; #endif }; @@ -252,8 +252,8 @@ DECLARE_PER_CPU(struct pt_regs, perf_trace_regs); extern int perf_trace_init(struct perf_event *event); extern void perf_trace_destroy(struct perf_event *event); -extern int perf_trace_enable(struct perf_event *event); -extern void perf_trace_disable(struct perf_event *event); +extern int perf_trace_add(struct perf_event *event, int flags); +extern void perf_trace_del(struct perf_event *event, int flags); extern int ftrace_profile_set_filter(struct perf_event *event, int event_id, char *filter_str); extern void ftrace_profile_free_filter(struct perf_event *event); diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index a0384a4..531495d 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -18,6 +18,7 @@ #include <asm/atomic.h> #include <asm/ptrace.h> #include <asm/system.h> +#include <trace/events/irq.h> /* * These correspond to the IORESOURCE_IRQ_* defines in @@ -407,7 +408,12 @@ asmlinkage void do_softirq(void); asmlinkage void __do_softirq(void); extern void open_softirq(int nr, void (*action)(struct softirq_action *)); extern void softirq_init(void); -#define __raise_softirq_irqoff(nr) do { or_softirq_pending(1UL << (nr)); } while (0) +static inline void __raise_softirq_irqoff(unsigned int nr) +{ + trace_softirq_raise((struct softirq_action *)(unsigned long)nr, NULL); + or_softirq_pending(1UL << nr); +} + extern void raise_softirq_irqoff(unsigned int nr); extern void raise_softirq(unsigned int nr); extern void wakeup_softirqd(void); diff --git a/include/linux/irq_work.h b/include/linux/irq_work.h new file mode 100644 index 0000000..4fa09d4 --- /dev/null +++ b/include/linux/irq_work.h @@ -0,0 +1,20 @@ +#ifndef _LINUX_IRQ_WORK_H +#define _LINUX_IRQ_WORK_H + +struct irq_work { + struct irq_work *next; + void (*func)(struct irq_work *); +}; + +static inline +void init_irq_work(struct irq_work *entry, void (*func)(struct irq_work *)) +{ + entry->next = NULL; + entry->func = func; +} + +bool irq_work_queue(struct irq_work *entry); +void irq_work_run(void); +void irq_work_sync(struct irq_work *entry); + +#endif /* _LINUX_IRQ_WORK_H */ diff --git a/include/linux/jump_label.h b/include/linux/jump_label.h new file mode 100644 index 0000000..b67cb18 --- /dev/null +++ b/include/linux/jump_label.h @@ -0,0 +1,74 @@ +#ifndef _LINUX_JUMP_LABEL_H +#define _LINUX_JUMP_LABEL_H + +#if defined(CC_HAVE_ASM_GOTO) && defined(CONFIG_HAVE_ARCH_JUMP_LABEL) +# include <asm/jump_label.h> +# define HAVE_JUMP_LABEL +#endif + +enum jump_label_type { + JUMP_LABEL_ENABLE, + JUMP_LABEL_DISABLE +}; + +struct module; + +#ifdef HAVE_JUMP_LABEL + +extern struct jump_entry __start___jump_table[]; +extern struct jump_entry __stop___jump_table[]; + +extern void arch_jump_label_transform(struct jump_entry *entry, + enum jump_label_type type); +extern void arch_jump_label_text_poke_early(jump_label_t addr); +extern void jump_label_update(unsigned long key, enum jump_label_type type); +extern void jump_label_apply_nops(struct module *mod); +extern int jump_label_text_reserved(void *start, void *end); + +#define jump_label_enable(key) \ + jump_label_update((unsigned long)key, JUMP_LABEL_ENABLE); + +#define jump_label_disable(key) \ + jump_label_update((unsigned long)key, JUMP_LABEL_DISABLE); + +#else + +#define JUMP_LABEL(key, label) \ +do { \ + if (unlikely(*key)) \ + goto label; \ +} while (0) + +#define jump_label_enable(cond_var) \ +do { \ + *(cond_var) = 1; \ +} while (0) + +#define jump_label_disable(cond_var) \ +do { \ + *(cond_var) = 0; \ +} while (0) + +static inline int jump_label_apply_nops(struct module *mod) +{ + return 0; +} + +static inline int jump_label_text_reserved(void *start, void *end) +{ + return 0; +} + +#endif + +#define COND_STMT(key, stmt) \ +do { \ + __label__ jl_enabled; \ + JUMP_LABEL(key, jl_enabled); \ + if (0) { \ +jl_enabled: \ + stmt; \ + } \ +} while (0) + +#endif diff --git a/include/linux/jump_label_ref.h b/include/linux/jump_label_ref.h new file mode 100644 index 0000000..e5d012a --- /dev/null +++ b/include/linux/jump_label_ref.h @@ -0,0 +1,44 @@ +#ifndef _LINUX_JUMP_LABEL_REF_H +#define _LINUX_JUMP_LABEL_REF_H + +#include <linux/jump_label.h> +#include <asm/atomic.h> + +#ifdef HAVE_JUMP_LABEL + +static inline void jump_label_inc(atomic_t *key) +{ + if (atomic_add_return(1, key) == 1) + jump_label_enable(key); +} + +static inline void jump_label_dec(atomic_t *key) +{ + if (atomic_dec_and_test(key)) + jump_label_disable(key); +} + +#else /* !HAVE_JUMP_LABEL */ + +static inline void jump_label_inc(atomic_t *key) +{ + atomic_inc(key); +} + +static inline void jump_label_dec(atomic_t *key) +{ + atomic_dec(key); +} + +#undef JUMP_LABEL +#define JUMP_LABEL(key, label) \ +do { \ + if (unlikely(__builtin_choose_expr( \ + __builtin_types_compatible_p(typeof(key), atomic_t *), \ + atomic_read((atomic_t *)(key)), *(key)))) \ + goto label; \ +} while (0) + +#endif /* HAVE_JUMP_LABEL */ + +#endif /* _LINUX_JUMP_LABEL_REF_H */ diff --git a/include/linux/module.h b/include/linux/module.h index aace066..b29e745 100644 --- a/include/linux/module.h +++ b/include/linux/module.h @@ -350,7 +350,10 @@ struct module struct tracepoint *tracepoints; unsigned int num_tracepoints; #endif - +#ifdef HAVE_JUMP_LABEL + struct jump_entry *jump_entries; + unsigned int num_jump_entries; +#endif #ifdef CONFIG_TRACING const char **trace_bprintk_fmt_start; unsigned int num_trace_bprintk_fmt; diff --git a/include/linux/oprofile.h b/include/linux/oprofile.h index 5171639..32fb812 100644 --- a/include/linux/oprofile.h +++ b/include/linux/oprofile.h @@ -15,6 +15,7 @@ #include <linux/types.h> #include <linux/spinlock.h> +#include <linux/init.h> #include <asm/atomic.h> /* Each escaped entry is prefixed by ESCAPE_CODE @@ -185,4 +186,10 @@ int oprofile_add_data(struct op_entry *entry, unsigned long val); int oprofile_add_data64(struct op_entry *entry, u64 val); int oprofile_write_commit(struct op_entry *entry); +#ifdef CONFIG_PERF_EVENTS +int __init oprofile_perf_init(struct oprofile_operations *ops); +void oprofile_perf_exit(void); +char *op_name_from_perf_id(void); +#endif /* CONFIG_PERF_EVENTS */ + #endif /* OPROFILE_H */ diff --git a/include/linux/percpu.h b/include/linux/percpu.h index 49466b1..0eb5083 100644 --- a/include/linux/percpu.h +++ b/include/linux/percpu.h @@ -39,6 +39,15 @@ preempt_enable(); \ } while (0) +#define get_cpu_ptr(var) ({ \ + preempt_disable(); \ + this_cpu_ptr(var); }) + +#define put_cpu_ptr(var) do { \ + (void)(var); \ + preempt_enable(); \ +} while (0) + #ifdef CONFIG_SMP /* minimum unit size, also is the maximum supported allocation size */ diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 716f99b..057bf22 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -486,6 +486,8 @@ struct perf_guest_info_callbacks { #include <linux/workqueue.h> #include <linux/ftrace.h> #include <linux/cpu.h> +#include <linux/irq_work.h> +#include <linux/jump_label_ref.h> #include <asm/atomic.h> #include <asm/local.h> @@ -529,16 +531,22 @@ struct hw_perf_event { int last_cpu; }; struct { /* software */ - s64 remaining; struct hrtimer hrtimer; }; #ifdef CONFIG_HAVE_HW_BREAKPOINT struct { /* breakpoint */ struct arch_hw_breakpoint info; struct list_head bp_list; + /* + * Crufty hack to avoid the chicken and egg + * problem hw_breakpoint has with context + * creation and event initalization. + */ + struct task_struct *bp_target; }; #endif }; + int state; local64_t prev_count; u64 sample_period; u64 last_period; @@ -550,6 +558,13 @@ struct hw_perf_event { #endif }; +/* + * hw_perf_event::state flags + */ +#define PERF_HES_STOPPED 0x01 /* the counter is stopped */ +#define PERF_HES_UPTODATE 0x02 /* event->count up-to-date */ +#define PERF_HES_ARCH 0x04 + struct perf_event; /* @@ -561,36 +576,70 @@ struct perf_event; * struct pmu - generic performance monitoring unit */ struct pmu { - int (*enable) (struct perf_event *event); - void (*disable) (struct perf_event *event); - int (*start) (struct perf_event *event); - void (*stop) (struct perf_event *event); - void (*read) (struct perf_event *event); - void (*unthrottle) (struct perf_event *event); + struct list_head entry; + + int * __percpu pmu_disable_count; + struct perf_cpu_context * __percpu pmu_cpu_context; + int task_ctx_nr; + + /* + * Fully disable/enable this PMU, can be used to protect from the PMI + * as well as for lazy/batch writing of the MSRs. + */ + void (*pmu_enable) (struct pmu *pmu); /* optional */ + void (*pmu_disable) (struct pmu *pmu); /* optional */ /* - * Group events scheduling is treated as a transaction, add group - * events as a whole and perform one schedulability test. If the test - * fails, roll back the whole group + * Try and initialize the event for this PMU. + * Should return -ENOENT when the @event doesn't match this PMU. */ + int (*event_init) (struct perf_event *event); + +#define PERF_EF_START 0x01 /* start the counter when adding */ +#define PERF_EF_RELOAD 0x02 /* reload the counter when starting */ +#define PERF_EF_UPDATE 0x04 /* update the counter when stopping */ /* - * Start the transaction, after this ->enable() doesn't need - * to do schedulability tests. + * Adds/Removes a counter to/from the PMU, can be done inside + * a transaction, see the ->*_txn() methods. */ - void (*start_txn) (const struct pmu *pmu); + int (*add) (struct perf_event *event, int flags); + void (*del) (struct perf_event *event, int flags); + /* - * If ->start_txn() disabled the ->enable() schedulability test + * Starts/Stops a counter present on the PMU. The PMI handler + * should stop the counter when perf_event_overflow() returns + * !0. ->start() will be used to continue. + */ + void (*start) (struct perf_event *event, int flags); + void (*stop) (struct perf_event *event, int flags); + + /* + * Updates the counter value of the event. + */ + void (*read) (struct perf_event *event); + + /* + * Group events scheduling is treated as a transaction, add + * group events as a whole and perform one schedulability test. + * If the test fails, roll back the whole group + * + * Start the transaction, after this ->add() doesn't need to + * do schedulability tests. + */ + void (*start_txn) (struct pmu *pmu); /* optional */ + /* + * If ->start_txn() disabled the ->add() schedulability test * then ->commit_txn() is required to perform one. On success * the transaction is closed. On error the transaction is kept * open until ->cancel_txn() is called. */ - int (*commit_txn) (const struct pmu *pmu); + int (*commit_txn) (struct pmu *pmu); /* optional */ /* - * Will cancel the transaction, assumes ->disable() is called for - * each successfull ->enable() during the transaction. + * Will cancel the transaction, assumes ->del() is called + * for each successfull ->add() during the transaction. */ - void (*cancel_txn) (const struct pmu *pmu); + void (*cancel_txn) (struct pmu *pmu); /* optional */ }; /** @@ -631,11 +680,6 @@ struct perf_buffer { void *data_pages[0]; }; -struct perf_pending_entry { - struct perf_pending_entry *next; - void (*func)(struct perf_pending_entry *); -}; - struct perf_sample_data; typedef void (*perf_overflow_handler_t)(struct perf_event *, int, @@ -656,6 +700,7 @@ struct swevent_hlist { #define PERF_ATTACH_CONTEXT 0x01 #define PERF_ATTACH_GROUP 0x02 +#define PERF_ATTACH_TASK 0x04 /** * struct perf_event - performance event kernel representation: @@ -669,7 +714,7 @@ struct perf_event { int nr_siblings; int group_flags; struct perf_event *group_leader; - const struct pmu *pmu; + struct pmu *pmu; enum perf_event_active_state state; unsigned int attach_state; @@ -743,7 +788,7 @@ struct perf_event { int pending_wakeup; int pending_kill; int pending_disable; - struct perf_pending_entry pending; + struct irq_work pending; atomic_t event_limit; @@ -763,12 +808,19 @@ struct perf_event { #endif /* CONFIG_PERF_EVENTS */ }; +enum perf_event_context_type { + task_context, + cpu_context, +}; + /** * struct perf_event_context - event context structure * * Used as a container for task events and CPU events as well: */ struct perf_event_context { + enum perf_event_context_type type; + struct pmu *pmu; /* * Protect the states of the events in the list, * nr_active, and the list: @@ -808,6 +860,12 @@ struct perf_event_context { struct rcu_head rcu_head; }; +/* + * Number of contexts where an event can trigger: + * task, softirq, hardirq, nmi. + */ +#define PERF_NR_CONTEXTS 4 + /** * struct perf_event_cpu_context - per cpu event context structure */ @@ -815,18 +873,9 @@ struct perf_cpu_context { struct perf_event_context ctx; struct perf_event_context *task_ctx; int active_oncpu; - int max_pertask; int exclusive; - struct swevent_hlist *swevent_hlist; - struct mutex hlist_mutex; - int hlist_refcount; - - /* - * Recursion avoidance: - * - * task, softirq, irq, nmi context - */ - int recursion[4]; + struct list_head rotation_list; + int jiffies_interval; }; struct perf_output_handle { @@ -842,26 +891,34 @@ struct perf_output_handle { #ifdef CONFIG_PERF_EVENTS -/* - * Set by architecture code: - */ -extern int perf_max_events; +extern int perf_pmu_register(struct pmu *pmu); +extern void perf_pmu_unregister(struct pmu *pmu); + +extern int perf_num_counters(void); +extern const char *perf_pmu_name(void); +extern void __perf_event_task_sched_in(struct task_struct *task); +extern void __perf_event_task_sched_out(struct task_struct *task, struct task_struct *next); -extern const struct pmu *hw_perf_event_init(struct perf_event *event); +extern atomic_t perf_task_events; + +static inline void perf_event_task_sched_in(struct task_struct *task) +{ + COND_STMT(&perf_task_events, __perf_event_task_sched_in(task)); +} + +static inline +void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next) +{ + COND_STMT(&perf_task_events, __perf_event_task_sched_out(task, next)); +} -extern void perf_event_task_sched_in(struct task_struct *task); -extern void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next); -extern void perf_event_task_tick(struct task_struct *task); extern int perf_event_init_task(struct task_struct *child); extern void perf_event_exit_task(struct task_struct *child); extern void perf_event_free_task(struct task_struct *task); -extern void set_perf_event_pending(void); -extern void perf_event_do_pending(void); +extern void perf_event_delayed_put(struct task_struct *task); extern void perf_event_print_debug(void); -extern void __perf_disable(void); -extern bool __perf_enable(void); -extern void perf_disable(void); -extern void perf_enable(void); +extern void perf_pmu_disable(struct pmu *pmu); +extern void perf_pmu_enable(struct pmu *pmu); extern int perf_event_task_disable(void); extern int perf_event_task_enable(void); extern void perf_event_update_userpage(struct perf_event *event); @@ -869,7 +926,7 @@ 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, + struct task_struct *task, perf_overflow_handler_t callback); extern u64 perf_event_read_value(struct perf_event *event, u64 *enabled, u64 *running); @@ -920,14 +977,7 @@ extern int perf_event_overflow(struct perf_event *event, int nmi, */ static inline int is_software_event(struct perf_event *event) { - switch (event->attr.type) { - case PERF_TYPE_SOFTWARE: - case PERF_TYPE_TRACEPOINT: - /* for now the breakpoint stuff also works as software event */ - case PERF_TYPE_BREAKPOINT: - return 1; - } - return 0; + return event->pmu->task_ctx_nr == perf_sw_context; } extern atomic_t perf_swevent_enabled[PERF_COUNT_SW_MAX]; @@ -954,18 +1004,20 @@ static inline void perf_fetch_caller_regs(struct pt_regs *regs) perf_arch_fetch_caller_regs(regs, CALLER_ADDR0); } -static inline void +static __always_inline void perf_sw_event(u32 event_id, u64 nr, int nmi, struct pt_regs *regs, u64 addr) { - if (atomic_read(&perf_swevent_enabled[event_id])) { - struct pt_regs hot_regs; - - if (!regs) { - perf_fetch_caller_regs(&hot_regs); - regs = &hot_regs; - } - __perf_sw_event(event_id, nr, nmi, regs, addr); + struct pt_regs hot_regs; + + JUMP_LABEL(&perf_swevent_enabled[event_id], have_event); + return; + +have_event: + if (!regs) { + perf_fetch_caller_regs(&hot_regs); + regs = &hot_regs; } + __perf_sw_event(event_id, nr, nmi, regs, addr); } extern void perf_event_mmap(struct vm_area_struct *vma); @@ -976,7 +1028,21 @@ extern int perf_unregister_guest_info_callbacks(struct perf_guest_info_callbacks extern void perf_event_comm(struct task_struct *tsk); extern void perf_event_fork(struct task_struct *tsk); -extern struct perf_callchain_entry *perf_callchain(struct pt_regs *regs); +/* Callchains */ +DECLARE_PER_CPU(struct perf_callchain_entry, perf_callchain_entry); + +extern void perf_callchain_user(struct perf_callchain_entry *entry, + struct pt_regs *regs); +extern void perf_callchain_kernel(struct perf_callchain_entry *entry, + struct pt_regs *regs); + + +static inline void +perf_callchain_store(struct perf_callchain_entry *entry, u64 ip) +{ + if (entry->nr < PERF_MAX_STACK_DEPTH) + entry->ip[entry->nr++] = ip; +} extern int sysctl_perf_event_paranoid; extern int sysctl_perf_event_mlock; @@ -1019,21 +1085,18 @@ extern int perf_swevent_get_recursion_context(void); extern void perf_swevent_put_recursion_context(int rctx); extern void perf_event_enable(struct perf_event *event); extern void perf_event_disable(struct perf_event *event); +extern void perf_event_task_tick(void); #else static inline void perf_event_task_sched_in(struct task_struct *task) { } static inline void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next) { } -static inline void -perf_event_task_tick(struct task_struct *task) { } static inline int perf_event_init_task(struct task_struct *child) { return 0; } static inline void perf_event_exit_task(struct task_struct *child) { } static inline void perf_event_free_task(struct task_struct *task) { } -static inline void perf_event_do_pending(void) { } +static inline void perf_event_delayed_put(struct task_struct *task) { } static inline void perf_event_print_debug(void) { } -static inline void perf_disable(void) { } -static inline void perf_enable(void) { } static inline int perf_event_task_disable(void) { return -EINVAL; } static inline int perf_event_task_enable(void) { return -EINVAL; } @@ -1056,6 +1119,7 @@ static inline int perf_swevent_get_recursion_context(void) { return -1; } static inline void perf_swevent_put_recursion_context(int rctx) { } static inline void perf_event_enable(struct perf_event *event) { } static inline void perf_event_disable(struct perf_event *event) { } +static inline void perf_event_task_tick(void) { } #endif #define perf_output_put(handle, x) \ diff --git a/include/linux/sched.h b/include/linux/sched.h index e18473f..61b4ecf 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1160,6 +1160,13 @@ struct sched_rt_entity { struct rcu_node; +enum perf_event_task_context { + perf_invalid_context = -1, + perf_hw_context = 0, + perf_sw_context, + perf_nr_task_contexts, +}; + struct task_struct { volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */ void *stack; @@ -1433,7 +1440,7 @@ struct task_struct { struct futex_pi_state *pi_state_cache; #endif #ifdef CONFIG_PERF_EVENTS - struct perf_event_context *perf_event_ctxp; + struct perf_event_context *perf_event_ctxp[perf_nr_task_contexts]; struct mutex perf_event_mutex; struct list_head perf_event_list; #endif diff --git a/include/linux/stop_machine.h b/include/linux/stop_machine.h index 6b524a0..1808960 100644 --- a/include/linux/stop_machine.h +++ b/include/linux/stop_machine.h @@ -126,8 +126,8 @@ int __stop_machine(int (*fn)(void *), void *data, const struct cpumask *cpus); #else /* CONFIG_STOP_MACHINE && CONFIG_SMP */ -static inline int stop_machine(int (*fn)(void *), void *data, - const struct cpumask *cpus) +static inline int __stop_machine(int (*fn)(void *), void *data, + const struct cpumask *cpus) { int ret; local_irq_disable(); @@ -136,5 +136,11 @@ static inline int stop_machine(int (*fn)(void *), void *data, return ret; } +static inline int stop_machine(int (*fn)(void *), void *data, + const struct cpumask *cpus) +{ + return __stop_machine(fn, data, cpus); +} + #endif /* CONFIG_STOP_MACHINE && CONFIG_SMP */ #endif /* _LINUX_STOP_MACHINE */ diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 103d1b6..a4a90b6 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -17,6 +17,7 @@ #include <linux/errno.h> #include <linux/types.h> #include <linux/rcupdate.h> +#include <linux/jump_label.h> struct module; struct tracepoint; @@ -145,7 +146,9 @@ static inline void tracepoint_update_probe_range(struct tracepoint *begin, extern struct tracepoint __tracepoint_##name; \ static inline void trace_##name(proto) \ { \ - if (unlikely(__tracepoint_##name.state)) \ + JUMP_LABEL(&__tracepoint_##name.state, do_trace); \ + return; \ +do_trace: \ __DO_TRACE(&__tracepoint_##name, \ TP_PROTO(data_proto), \ TP_ARGS(data_args)); \ diff --git a/include/trace/events/irq.h b/include/trace/events/irq.h index 0e4cfb6..6fa7cba 100644 --- a/include/trace/events/irq.h +++ b/include/trace/events/irq.h @@ -5,7 +5,9 @@ #define _TRACE_IRQ_H #include <linux/tracepoint.h> -#include <linux/interrupt.h> + +struct irqaction; +struct softirq_action; #define softirq_name(sirq) { sirq##_SOFTIRQ, #sirq } #define show_softirq_name(val) \ @@ -93,7 +95,10 @@ DECLARE_EVENT_CLASS(softirq, ), TP_fast_assign( - __entry->vec = (int)(h - vec); + if (vec) + __entry->vec = (int)(h - vec); + else + __entry->vec = (int)(long)h; ), TP_printk("vec=%d [action=%s]", __entry->vec, @@ -136,6 +141,23 @@ DEFINE_EVENT(softirq, softirq_exit, TP_ARGS(h, vec) ); +/** + * softirq_raise - called immediately when a softirq is raised + * @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 softirq vector number which is + * raised. @vec is NULL and it means @h includes vector number not + * softirq_action. When used in combination with the softirq_entry tracepoint + * we can determine the softirq raise latency. + */ +DEFINE_EVENT(softirq, softirq_raise, + + TP_PROTO(struct softirq_action *h, struct softirq_action *vec), + + TP_ARGS(h, vec) +); + #endif /* _TRACE_IRQ_H */ /* This part must be outside protection */ diff --git a/include/trace/events/napi.h b/include/trace/events/napi.h index 188deca..8fe1e93 100644 --- a/include/trace/events/napi.h +++ b/include/trace/events/napi.h @@ -6,10 +6,31 @@ #include <linux/netdevice.h> #include <linux/tracepoint.h> +#include <linux/ftrace.h> + +#define NO_DEV "(no_device)" + +TRACE_EVENT(napi_poll, -DECLARE_TRACE(napi_poll, TP_PROTO(struct napi_struct *napi), - TP_ARGS(napi)); + + TP_ARGS(napi), + + TP_STRUCT__entry( + __field( struct napi_struct *, napi) + __string( dev_name, napi->dev ? napi->dev->name : NO_DEV) + ), + + TP_fast_assign( + __entry->napi = napi; + __assign_str(dev_name, napi->dev ? napi->dev->name : NO_DEV); + ), + + TP_printk("napi poll on napi struct %p for device %s", + __entry->napi, __get_str(dev_name)) +); + +#undef NO_DEV #endif /* _TRACE_NAPI_H_ */ diff --git a/include/trace/events/net.h b/include/trace/events/net.h new file mode 100644 index 0000000..5f247f5 --- /dev/null +++ b/include/trace/events/net.h @@ -0,0 +1,82 @@ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM net + +#if !defined(_TRACE_NET_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_NET_H + +#include <linux/skbuff.h> +#include <linux/netdevice.h> +#include <linux/ip.h> +#include <linux/tracepoint.h> + +TRACE_EVENT(net_dev_xmit, + + TP_PROTO(struct sk_buff *skb, + int rc), + + TP_ARGS(skb, rc), + + TP_STRUCT__entry( + __field( void *, skbaddr ) + __field( unsigned int, len ) + __field( int, rc ) + __string( name, skb->dev->name ) + ), + + TP_fast_assign( + __entry->skbaddr = skb; + __entry->len = skb->len; + __entry->rc = rc; + __assign_str(name, skb->dev->name); + ), + + TP_printk("dev=%s skbaddr=%p len=%u rc=%d", + __get_str(name), __entry->skbaddr, __entry->len, __entry->rc) +); + +DECLARE_EVENT_CLASS(net_dev_template, + + TP_PROTO(struct sk_buff *skb), + + TP_ARGS(skb), + + TP_STRUCT__entry( + __field( void *, skbaddr ) + __field( unsigned int, len ) + __string( name, skb->dev->name ) + ), + + TP_fast_assign( + __entry->skbaddr = skb; + __entry->len = skb->len; + __assign_str(name, skb->dev->name); + ), + + TP_printk("dev=%s skbaddr=%p len=%u", + __get_str(name), __entry->skbaddr, __entry->len) +) + +DEFINE_EVENT(net_dev_template, net_dev_queue, + + TP_PROTO(struct sk_buff *skb), + + TP_ARGS(skb) +); + +DEFINE_EVENT(net_dev_template, netif_receive_skb, + + TP_PROTO(struct sk_buff *skb), + + TP_ARGS(skb) +); + +DEFINE_EVENT(net_dev_template, netif_rx, + + TP_PROTO(struct sk_buff *skb), + + TP_ARGS(skb) +); +#endif /* _TRACE_NET_H */ + +/* This part must be outside protection */ +#include <trace/define_trace.h> diff --git a/include/trace/events/power.h b/include/trace/events/power.h index 35a2a6e..286784d 100644 --- a/include/trace/events/power.h +++ b/include/trace/events/power.h @@ -10,12 +10,17 @@ #ifndef _TRACE_POWER_ENUM_ #define _TRACE_POWER_ENUM_ enum { - POWER_NONE = 0, - POWER_CSTATE = 1, - POWER_PSTATE = 2, + POWER_NONE = 0, + POWER_CSTATE = 1, /* C-State */ + POWER_PSTATE = 2, /* Fequency change or DVFS */ + POWER_SSTATE = 3, /* Suspend */ }; #endif +/* + * The power events are used for cpuidle & suspend (power_start, power_end) + * and for cpufreq (power_frequency) + */ DECLARE_EVENT_CLASS(power, TP_PROTO(unsigned int type, unsigned int state, unsigned int cpu_id), @@ -70,6 +75,85 @@ TRACE_EVENT(power_end, ); +/* + * The clock events are used for clock enable/disable and for + * clock rate change + */ +DECLARE_EVENT_CLASS(clock, + + TP_PROTO(const char *name, unsigned int state, unsigned int cpu_id), + + TP_ARGS(name, state, cpu_id), + + TP_STRUCT__entry( + __string( name, name ) + __field( u64, state ) + __field( u64, cpu_id ) + ), + + TP_fast_assign( + __assign_str(name, name); + __entry->state = state; + __entry->cpu_id = cpu_id; + ), + + TP_printk("%s state=%lu cpu_id=%lu", __get_str(name), + (unsigned long)__entry->state, (unsigned long)__entry->cpu_id) +); + +DEFINE_EVENT(clock, clock_enable, + + TP_PROTO(const char *name, unsigned int state, unsigned int cpu_id), + + TP_ARGS(name, state, cpu_id) +); + +DEFINE_EVENT(clock, clock_disable, + + TP_PROTO(const char *name, unsigned int state, unsigned int cpu_id), + + TP_ARGS(name, state, cpu_id) +); + +DEFINE_EVENT(clock, clock_set_rate, + + TP_PROTO(const char *name, unsigned int state, unsigned int cpu_id), + + TP_ARGS(name, state, cpu_id) +); + +/* + * The power domain events are used for power domains transitions + */ +DECLARE_EVENT_CLASS(power_domain, + + TP_PROTO(const char *name, unsigned int state, unsigned int cpu_id), + + TP_ARGS(name, state, cpu_id), + + TP_STRUCT__entry( + __string( name, name ) + __field( u64, state ) + __field( u64, cpu_id ) + ), + + TP_fast_assign( + __assign_str(name, name); + __entry->state = state; + __entry->cpu_id = cpu_id; +), + + TP_printk("%s state=%lu cpu_id=%lu", __get_str(name), + (unsigned long)__entry->state, (unsigned long)__entry->cpu_id) +); + +DEFINE_EVENT(power_domain, power_domain_target, + + TP_PROTO(const char *name, unsigned int state, unsigned int cpu_id), + + TP_ARGS(name, state, cpu_id) +); + #endif /* _TRACE_POWER_H */ /* This part must be outside protection */ diff --git a/include/trace/events/skb.h b/include/trace/events/skb.h index 4b2be6d..75ce9d5 100644 --- a/include/trace/events/skb.h +++ b/include/trace/events/skb.h @@ -35,6 +35,23 @@ TRACE_EVENT(kfree_skb, __entry->skbaddr, __entry->protocol, __entry->location) ); +TRACE_EVENT(consume_skb, + + TP_PROTO(struct sk_buff *skb), + + TP_ARGS(skb), + + TP_STRUCT__entry( + __field( void *, skbaddr ) + ), + + TP_fast_assign( + __entry->skbaddr = skb; + ), + + TP_printk("skbaddr=%p", __entry->skbaddr) +); + TRACE_EVENT(skb_copy_datagram_iovec, TP_PROTO(const struct sk_buff *skb, int len), diff --git a/init/Kconfig b/init/Kconfig index a619a1a..7b920aa 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -21,6 +21,13 @@ config CONSTRUCTORS depends on !UML default y +config HAVE_IRQ_WORK + bool + +config IRQ_WORK + bool + depends on HAVE_IRQ_WORK + menu "General setup" config EXPERIMENTAL @@ -1005,6 +1012,7 @@ config PERF_EVENTS default y if (PROFILING || PERF_COUNTERS) depends on HAVE_PERF_EVENTS select ANON_INODES + select IRQ_WORK help Enable kernel support for various performance events provided by software and hardware. diff --git a/kernel/Makefile b/kernel/Makefile index 17046b6..e2c9d52 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -10,7 +10,7 @@ obj-y = sched.o fork.o exec_domain.o panic.o printk.o \ kthread.o wait.o kfifo.o sys_ni.o posix-cpu-timers.o mutex.o \ hrtimer.o rwsem.o nsproxy.o srcu.o semaphore.o \ notifier.o ksysfs.o pm_qos_params.o sched_clock.o cred.o \ - async.o range.o + async.o range.o jump_label.o obj-$(CONFIG_HAVE_EARLY_RES) += early_res.o obj-y += groups.o @@ -23,6 +23,7 @@ CFLAGS_REMOVE_rtmutex-debug.o = -pg CFLAGS_REMOVE_cgroup-debug.o = -pg CFLAGS_REMOVE_sched_clock.o = -pg CFLAGS_REMOVE_perf_event.o = -pg +CFLAGS_REMOVE_irq_work.o = -pg endif obj-$(CONFIG_FREEZER) += freezer.o @@ -101,6 +102,7 @@ obj-$(CONFIG_TRACING) += trace/ obj-$(CONFIG_X86_DS) += trace/ obj-$(CONFIG_RING_BUFFER) += trace/ obj-$(CONFIG_SMP) += sched_cpupri.o +obj-$(CONFIG_IRQ_WORK) += irq_work.o obj-$(CONFIG_PERF_EVENTS) += perf_event.o obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o obj-$(CONFIG_USER_RETURN_NOTIFIER) += user-return-notifier.o diff --git a/kernel/exit.c b/kernel/exit.c index 0312022..e2bdf37 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -149,9 +149,7 @@ static void delayed_put_task_struct(struct rcu_head *rhp) { struct task_struct *tsk = container_of(rhp, struct task_struct, rcu); -#ifdef CONFIG_PERF_EVENTS - WARN_ON_ONCE(tsk->perf_event_ctxp); -#endif + perf_event_delayed_put(tsk); trace_sched_process_free(tsk); put_task_struct(tsk); } diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index c7c2aed..2c9120f 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -113,12 +113,12 @@ static unsigned int max_task_bp_pinned(int cpu, enum bp_type_idx type) */ static int task_bp_pinned(struct perf_event *bp, enum bp_type_idx type) { - struct perf_event_context *ctx = bp->ctx; + struct task_struct *tsk = bp->hw.bp_target; struct perf_event *iter; int count = 0; list_for_each_entry(iter, &bp_task_head, hw.bp_list) { - if (iter->ctx == ctx && find_slot_idx(iter) == type) + if (iter->hw.bp_target == tsk && find_slot_idx(iter) == type) count += hw_breakpoint_weight(iter); } @@ -134,7 +134,7 @@ fetch_bp_busy_slots(struct bp_busy_slots *slots, struct perf_event *bp, enum bp_type_idx type) { int cpu = bp->cpu; - struct task_struct *tsk = bp->ctx->task; + struct task_struct *tsk = bp->hw.bp_target; if (cpu >= 0) { slots->pinned = per_cpu(nr_cpu_bp_pinned[type], cpu); @@ -213,7 +213,7 @@ toggle_bp_slot(struct perf_event *bp, bool enable, enum bp_type_idx type, int weight) { int cpu = bp->cpu; - struct task_struct *tsk = bp->ctx->task; + struct task_struct *tsk = bp->hw.bp_target; /* Pinned counter cpu profiling */ if (!tsk) { @@ -433,8 +433,7 @@ register_user_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered, struct task_struct *tsk) { - return perf_event_create_kernel_counter(attr, -1, task_pid_vnr(tsk), - triggered); + return perf_event_create_kernel_counter(attr, -1, tsk, triggered); } EXPORT_SYMBOL_GPL(register_user_hw_breakpoint); @@ -516,7 +515,7 @@ register_wide_hw_breakpoint(struct perf_event_attr *attr, get_online_cpus(); for_each_online_cpu(cpu) { pevent = per_cpu_ptr(cpu_events, cpu); - bp = perf_event_create_kernel_counter(attr, cpu, -1, triggered); + bp = perf_event_create_kernel_counter(attr, cpu, NULL, triggered); *pevent = bp; @@ -566,6 +565,61 @@ static struct notifier_block hw_breakpoint_exceptions_nb = { .priority = 0x7fffffff }; +static void bp_perf_event_destroy(struct perf_event *event) +{ + release_bp_slot(event); +} + +static int hw_breakpoint_event_init(struct perf_event *bp) +{ + int err; + + if (bp->attr.type != PERF_TYPE_BREAKPOINT) + return -ENOENT; + + err = register_perf_hw_breakpoint(bp); + if (err) + return err; + + bp->destroy = bp_perf_event_destroy; + + return 0; +} + +static int hw_breakpoint_add(struct perf_event *bp, int flags) +{ + if (!(flags & PERF_EF_START)) + bp->hw.state = PERF_HES_STOPPED; + + return arch_install_hw_breakpoint(bp); +} + +static void hw_breakpoint_del(struct perf_event *bp, int flags) +{ + arch_uninstall_hw_breakpoint(bp); +} + +static void hw_breakpoint_start(struct perf_event *bp, int flags) +{ + bp->hw.state = 0; +} + +static void hw_breakpoint_stop(struct perf_event *bp, int flags) +{ + bp->hw.state = PERF_HES_STOPPED; +} + +static struct pmu perf_breakpoint = { + .task_ctx_nr = perf_sw_context, /* could eventually get its own */ + + .event_init = hw_breakpoint_event_init, + .add = hw_breakpoint_add, + .del = hw_breakpoint_del, + .start = hw_breakpoint_start, + .stop = hw_breakpoint_stop, + .read = hw_breakpoint_pmu_read, +}; + static int __init init_hw_breakpoint(void) { unsigned int **task_bp_pinned; @@ -587,6 +641,8 @@ static int __init init_hw_breakpoint(void) constraints_initialized = 1; + perf_pmu_register(&perf_breakpoint); + return register_die_notifier(&hw_breakpoint_exceptions_nb); err_alloc: @@ -602,8 +658,3 @@ static int __init init_hw_breakpoint(void) 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, -}; diff --git a/kernel/irq_work.c b/kernel/irq_work.c new file mode 100644 index 0000000..f16763f --- /dev/null +++ b/kernel/irq_work.c @@ -0,0 +1,164 @@ +/* + * Copyright (C) 2010 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * + * Provides a framework for enqueueing and running callbacks from hardirq + * context. The enqueueing is NMI-safe. + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/irq_work.h> +#include <linux/hardirq.h> + +/* + * An entry can be in one of four states: + * + * free NULL, 0 -> {claimed} : free to be used + * claimed NULL, 3 -> {pending} : claimed to be enqueued + * pending next, 3 -> {busy} : queued, pending callback + * busy NULL, 2 -> {free, claimed} : callback in progress, can be claimed + * + * We use the lower two bits of the next pointer to keep PENDING and BUSY + * flags. + */ + +#define IRQ_WORK_PENDING 1UL +#define IRQ_WORK_BUSY 2UL +#define IRQ_WORK_FLAGS 3UL + +static inline bool irq_work_is_set(struct irq_work *entry, int flags) +{ + return (unsigned long)entry->next & flags; +} + +static inline struct irq_work *irq_work_next(struct irq_work *entry) +{ + unsigned long next = (unsigned long)entry->next; + next &= ~IRQ_WORK_FLAGS; + return (struct irq_work *)next; +} + +static inline struct irq_work *next_flags(struct irq_work *entry, int flags) +{ + unsigned long next = (unsigned long)entry; + next |= flags; + return (struct irq_work *)next; +} + +static DEFINE_PER_CPU(struct irq_work *, irq_work_list); + +/* + * Claim the entry so that no one else will poke at it. + */ +static bool irq_work_claim(struct irq_work *entry) +{ + struct irq_work *next, *nflags; + + do { + next = entry->next; + if ((unsigned long)next & IRQ_WORK_PENDING) + return false; + nflags = next_flags(next, IRQ_WORK_FLAGS); + } while (cmpxchg(&entry->next, next, nflags) != next); + + return true; +} + + +void __weak arch_irq_work_raise(void) +{ + /* + * Lame architectures will get the timer tick callback + */ +} + +/* + * Queue the entry and raise the IPI if needed. + */ +static void __irq_work_queue(struct irq_work *entry) +{ + struct irq_work **head, *next; + + head = &get_cpu_var(irq_work_list); + + do { + next = *head; + /* Can assign non-atomic because we keep the flags set. */ + entry->next = next_flags(next, IRQ_WORK_FLAGS); + } while (cmpxchg(head, next, entry) != next); + + /* The list was empty, raise self-interrupt to start processing. */ + if (!irq_work_next(entry)) + arch_irq_work_raise(); + + put_cpu_var(irq_work_list); +} + +/* + * Enqueue the irq_work @entry, returns true on success, failure when the + * @entry was already enqueued by someone else. + * + * Can be re-enqueued while the callback is still in progress. + */ +bool irq_work_queue(struct irq_work *entry) +{ + if (!irq_work_claim(entry)) { + /* + * Already enqueued, can't do! + */ + return false; + } + + __irq_work_queue(entry); + return true; +} +EXPORT_SYMBOL_GPL(irq_work_queue); + +/* + * Run the irq_work entries on this cpu. Requires to be ran from hardirq + * context with local IRQs disabled. + */ +void irq_work_run(void) +{ + struct irq_work *list, **head; + + head = &__get_cpu_var(irq_work_list); + if (*head == NULL) + return; + + BUG_ON(!in_irq()); + BUG_ON(!irqs_disabled()); + + list = xchg(head, NULL); + while (list != NULL) { + struct irq_work *entry = list; + + list = irq_work_next(list); + + /* + * Clear the PENDING bit, after this point the @entry + * can be re-used. + */ + entry->next = next_flags(NULL, IRQ_WORK_BUSY); + entry->func(entry); + /* + * Clear the BUSY bit and return to the free state if + * no-one else claimed it meanwhile. + */ + cmpxchg(&entry->next, next_flags(NULL, IRQ_WORK_BUSY), NULL); + } +} +EXPORT_SYMBOL_GPL(irq_work_run); + +/* + * Synchronize against the irq_work @entry, ensures the entry is not + * currently in use. + */ +void irq_work_sync(struct irq_work *entry) +{ + WARN_ON_ONCE(irqs_disabled()); + + while (irq_work_is_set(entry, IRQ_WORK_BUSY)) + cpu_relax(); +} +EXPORT_SYMBOL_GPL(irq_work_sync); diff --git a/kernel/jump_label.c b/kernel/jump_label.c new file mode 100644 index 0000000..7be868b --- /dev/null +++ b/kernel/jump_label.c @@ -0,0 +1,429 @@ +/* + * jump label support + * + * Copyright (C) 2009 Jason Baron <jbaron@redhat.com> + * + */ +#include <linux/jump_label.h> +#include <linux/memory.h> +#include <linux/uaccess.h> +#include <linux/module.h> +#include <linux/list.h> +#include <linux/jhash.h> +#include <linux/slab.h> +#include <linux/sort.h> +#include <linux/err.h> + +#ifdef HAVE_JUMP_LABEL + +#define JUMP_LABEL_HASH_BITS 6 +#define JUMP_LABEL_TABLE_SIZE (1 << JUMP_LABEL_HASH_BITS) +static struct hlist_head jump_label_table[JUMP_LABEL_TABLE_SIZE]; + +/* mutex to protect coming/going of the the jump_label table */ +static DEFINE_MUTEX(jump_label_mutex); + +struct jump_label_entry { + struct hlist_node hlist; + struct jump_entry *table; + int nr_entries; + /* hang modules off here */ + struct hlist_head modules; + unsigned long key; +}; + +struct jump_label_module_entry { + struct hlist_node hlist; + struct jump_entry *table; + int nr_entries; + struct module *mod; +}; + +static int jump_label_cmp(const void *a, const void *b) +{ + const struct jump_entry *jea = a; + const struct jump_entry *jeb = b; + + if (jea->key < jeb->key) + return -1; + + if (jea->key > jeb->key) + return 1; + + return 0; +} + +static void +sort_jump_label_entries(struct jump_entry *start, struct jump_entry *stop) +{ + unsigned long size; + + size = (((unsigned long)stop - (unsigned long)start) + / sizeof(struct jump_entry)); + sort(start, size, sizeof(struct jump_entry), jump_label_cmp, NULL); +} + +static struct jump_label_entry *get_jump_label_entry(jump_label_t key) +{ + struct hlist_head *head; + struct hlist_node *node; + struct jump_label_entry *e; + u32 hash = jhash((void *)&key, sizeof(jump_label_t), 0); + + head = &jump_label_table[hash & (JUMP_LABEL_TABLE_SIZE - 1)]; + hlist_for_each_entry(e, node, head, hlist) { + if (key == e->key) + return e; + } + return NULL; +} + +static struct jump_label_entry * +add_jump_label_entry(jump_label_t key, int nr_entries, struct jump_entry *table) +{ + struct hlist_head *head; + struct jump_label_entry *e; + u32 hash; + + e = get_jump_label_entry(key); + if (e) + return ERR_PTR(-EEXIST); + + e = kmalloc(sizeof(struct jump_label_entry), GFP_KERNEL); + if (!e) + return ERR_PTR(-ENOMEM); + + hash = jhash((void *)&key, sizeof(jump_label_t), 0); + head = &jump_label_table[hash & (JUMP_LABEL_TABLE_SIZE - 1)]; + e->key = key; + e->table = table; + e->nr_entries = nr_entries; + INIT_HLIST_HEAD(&(e->modules)); + hlist_add_head(&e->hlist, head); + return e; +} + +static int +build_jump_label_hashtable(struct jump_entry *start, struct jump_entry *stop) +{ + struct jump_entry *iter, *iter_begin; + struct jump_label_entry *entry; + int count; + + sort_jump_label_entries(start, stop); + iter = start; + while (iter < stop) { + entry = get_jump_label_entry(iter->key); + if (!entry) { + iter_begin = iter; + count = 0; + while ((iter < stop) && + (iter->key == iter_begin->key)) { + iter++; + count++; + } + entry = add_jump_label_entry(iter_begin->key, + count, iter_begin); + if (IS_ERR(entry)) + return PTR_ERR(entry); + } else { + WARN_ONCE(1, KERN_ERR "build_jump_hashtable: unexpected entry!\n"); + return -1; + } + } + return 0; +} + +/*** + * jump_label_update - update jump label text + * @key - key value associated with a a jump label + * @type - enum set to JUMP_LABEL_ENABLE or JUMP_LABEL_DISABLE + * + * Will enable/disable the jump for jump label @key, depending on the + * value of @type. + * + */ + +void jump_label_update(unsigned long key, enum jump_label_type type) +{ + struct jump_entry *iter; + struct jump_label_entry *entry; + struct hlist_node *module_node; + struct jump_label_module_entry *e_module; + int count; + + mutex_lock(&jump_label_mutex); + entry = get_jump_label_entry((jump_label_t)key); + if (entry) { + count = entry->nr_entries; + iter = entry->table; + while (count--) { + if (kernel_text_address(iter->code)) + arch_jump_label_transform(iter, type); + iter++; + } + /* eanble/disable jump labels in modules */ + hlist_for_each_entry(e_module, module_node, &(entry->modules), + hlist) { + count = e_module->nr_entries; + iter = e_module->table; + while (count--) { + if (kernel_text_address(iter->code)) + arch_jump_label_transform(iter, type); + iter++; + } + } + } + mutex_unlock(&jump_label_mutex); +} + +static int addr_conflict(struct jump_entry *entry, void *start, void *end) +{ + if (entry->code <= (unsigned long)end && + entry->code + JUMP_LABEL_NOP_SIZE > (unsigned long)start) + return 1; + + return 0; +} + +#ifdef CONFIG_MODULES + +static int module_conflict(void *start, void *end) +{ + struct hlist_head *head; + struct hlist_node *node, *node_next, *module_node, *module_node_next; + struct jump_label_entry *e; + struct jump_label_module_entry *e_module; + struct jump_entry *iter; + int i, count; + int conflict = 0; + + for (i = 0; i < JUMP_LABEL_TABLE_SIZE; i++) { + head = &jump_label_table[i]; + hlist_for_each_entry_safe(e, node, node_next, head, hlist) { + hlist_for_each_entry_safe(e_module, module_node, + module_node_next, + &(e->modules), hlist) { + count = e_module->nr_entries; + iter = e_module->table; + while (count--) { + if (addr_conflict(iter, start, end)) { + conflict = 1; + goto out; + } + iter++; + } + } + } + } +out: + return conflict; +} + +#endif + +/*** + * jump_label_text_reserved - check if addr range is reserved + * @start: start text addr + * @end: end text addr + * + * checks if the text addr located between @start and @end + * overlaps with any of the jump label patch addresses. Code + * that wants to modify kernel text should first verify that + * it does not overlap with any of the jump label addresses. + * + * returns 1 if there is an overlap, 0 otherwise + */ +int jump_label_text_reserved(void *start, void *end) +{ + struct jump_entry *iter; + struct jump_entry *iter_start = __start___jump_table; + struct jump_entry *iter_stop = __start___jump_table; + int conflict = 0; + + mutex_lock(&jump_label_mutex); + iter = iter_start; + while (iter < iter_stop) { + if (addr_conflict(iter, start, end)) { + conflict = 1; + goto out; + } + iter++; + } + + /* now check modules */ +#ifdef CONFIG_MODULES + conflict = module_conflict(start, end); +#endif +out: + mutex_unlock(&jump_label_mutex); + return conflict; +} + +static __init int init_jump_label(void) +{ + int ret; + struct jump_entry *iter_start = __start___jump_table; + struct jump_entry *iter_stop = __stop___jump_table; + struct jump_entry *iter; + + mutex_lock(&jump_label_mutex); + ret = build_jump_label_hashtable(__start___jump_table, + __stop___jump_table); + iter = iter_start; + while (iter < iter_stop) { + arch_jump_label_text_poke_early(iter->code); + iter++; + } + mutex_unlock(&jump_label_mutex); + return ret; +} +early_initcall(init_jump_label); + +#ifdef CONFIG_MODULES + +static struct jump_label_module_entry * +add_jump_label_module_entry(struct jump_label_entry *entry, + struct jump_entry *iter_begin, + int count, struct module *mod) +{ + struct jump_label_module_entry *e; + + e = kmalloc(sizeof(struct jump_label_module_entry), GFP_KERNEL); + if (!e) + return ERR_PTR(-ENOMEM); + e->mod = mod; + e->nr_entries = count; + e->table = iter_begin; + hlist_add_head(&e->hlist, &entry->modules); + return e; +} + +static int add_jump_label_module(struct module *mod) +{ + struct jump_entry *iter, *iter_begin; + struct jump_label_entry *entry; + struct jump_label_module_entry *module_entry; + int count; + + /* if the module doesn't have jump label entries, just return */ + if (!mod->num_jump_entries) + return 0; + + sort_jump_label_entries(mod->jump_entries, + mod->jump_entries + mod->num_jump_entries); + iter = mod->jump_entries; + while (iter < mod->jump_entries + mod->num_jump_entries) { + entry = get_jump_label_entry(iter->key); + iter_begin = iter; + count = 0; + while ((iter < mod->jump_entries + mod->num_jump_entries) && + (iter->key == iter_begin->key)) { + iter++; + count++; + } + if (!entry) { + entry = add_jump_label_entry(iter_begin->key, 0, NULL); + if (IS_ERR(entry)) + return PTR_ERR(entry); + } + module_entry = add_jump_label_module_entry(entry, iter_begin, + count, mod); + if (IS_ERR(module_entry)) + return PTR_ERR(module_entry); + } + return 0; +} + +static void remove_jump_label_module(struct module *mod) +{ + struct hlist_head *head; + struct hlist_node *node, *node_next, *module_node, *module_node_next; + struct jump_label_entry *e; + struct jump_label_module_entry *e_module; + int i; + + /* if the module doesn't have jump label entries, just return */ + if (!mod->num_jump_entries) + return; + + for (i = 0; i < JUMP_LABEL_TABLE_SIZE; i++) { + head = &jump_label_table[i]; + hlist_for_each_entry_safe(e, node, node_next, head, hlist) { + hlist_for_each_entry_safe(e_module, module_node, + module_node_next, + &(e->modules), hlist) { + if (e_module->mod == mod) { + hlist_del(&e_module->hlist); + kfree(e_module); + } + } + if (hlist_empty(&e->modules) && (e->nr_entries == 0)) { + hlist_del(&e->hlist); + kfree(e); + } + } + } +} + +static int +jump_label_module_notify(struct notifier_block *self, unsigned long val, + void *data) +{ + struct module *mod = data; + int ret = 0; + + switch (val) { + case MODULE_STATE_COMING: + mutex_lock(&jump_label_mutex); + ret = add_jump_label_module(mod); + if (ret) + remove_jump_label_module(mod); + mutex_unlock(&jump_label_mutex); + break; + case MODULE_STATE_GOING: + mutex_lock(&jump_label_mutex); + remove_jump_label_module(mod); + mutex_unlock(&jump_label_mutex); + break; + } + return ret; +} + +/*** + * apply_jump_label_nops - patch module jump labels with arch_get_jump_label_nop() + * @mod: module to patch + * + * Allow for run-time selection of the optimal nops. Before the module + * loads patch these with arch_get_jump_label_nop(), which is specified by + * the arch specific jump label code. + */ +void jump_label_apply_nops(struct module *mod) +{ + struct jump_entry *iter; + + /* if the module doesn't have jump label entries, just return */ + if (!mod->num_jump_entries) + return; + + iter = mod->jump_entries; + while (iter < mod->jump_entries + mod->num_jump_entries) { + arch_jump_label_text_poke_early(iter->code); + iter++; + } +} + +struct notifier_block jump_label_module_nb = { + .notifier_call = jump_label_module_notify, + .priority = 0, +}; + +static __init int init_jump_label_module(void) +{ + return register_module_notifier(&jump_label_module_nb); +} +early_initcall(init_jump_label_module); + +#endif /* CONFIG_MODULES */ + +#endif diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 282035f..ec4210c 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -47,6 +47,7 @@ #include <linux/memory.h> #include <linux/ftrace.h> #include <linux/cpu.h> +#include <linux/jump_label.h> #include <asm-generic/sections.h> #include <asm/cacheflush.h> @@ -399,7 +400,7 @@ static inline int kprobe_optready(struct kprobe *p) * Return an optimized kprobe whose optimizing code replaces * instructions including addr (exclude breakpoint). */ -struct kprobe *__kprobes get_optimized_kprobe(unsigned long addr) +static struct kprobe *__kprobes get_optimized_kprobe(unsigned long addr) { int i; struct kprobe *p = NULL; @@ -831,6 +832,7 @@ void __kprobes recycle_rp_inst(struct kretprobe_instance *ri, void __kprobes kretprobe_hash_lock(struct task_struct *tsk, struct hlist_head **head, unsigned long *flags) +__acquires(hlist_lock) { unsigned long hash = hash_ptr(tsk, KPROBE_HASH_BITS); spinlock_t *hlist_lock; @@ -842,6 +844,7 @@ void __kprobes kretprobe_hash_lock(struct task_struct *tsk, static void __kprobes kretprobe_table_lock(unsigned long hash, unsigned long *flags) +__acquires(hlist_lock) { spinlock_t *hlist_lock = kretprobe_table_lock_ptr(hash); spin_lock_irqsave(hlist_lock, *flags); @@ -849,6 +852,7 @@ static void __kprobes kretprobe_table_lock(unsigned long hash, void __kprobes kretprobe_hash_unlock(struct task_struct *tsk, unsigned long *flags) +__releases(hlist_lock) { unsigned long hash = hash_ptr(tsk, KPROBE_HASH_BITS); spinlock_t *hlist_lock; @@ -857,7 +861,9 @@ void __kprobes kretprobe_hash_unlock(struct task_struct *tsk, spin_unlock_irqrestore(hlist_lock, *flags); } -void __kprobes kretprobe_table_unlock(unsigned long hash, unsigned long *flags) +static void __kprobes kretprobe_table_unlock(unsigned long hash, + unsigned long *flags) +__releases(hlist_lock) { spinlock_t *hlist_lock = kretprobe_table_lock_ptr(hash); spin_unlock_irqrestore(hlist_lock, *flags); @@ -1141,7 +1147,8 @@ int __kprobes register_kprobe(struct kprobe *p) preempt_disable(); if (!kernel_text_address((unsigned long) p->addr) || in_kprobes_functions((unsigned long) p->addr) || - ftrace_text_reserved(p->addr, p->addr)) { + ftrace_text_reserved(p->addr, p->addr) || + jump_label_text_reserved(p->addr, p->addr)) { preempt_enable(); return -EINVAL; } @@ -1339,18 +1346,19 @@ int __kprobes register_jprobes(struct jprobe **jps, int num) if (num <= 0) return -EINVAL; for (i = 0; i < num; i++) { - unsigned long addr; + unsigned long addr, offset; jp = jps[i]; addr = arch_deref_entry_point(jp->entry); - if (!kernel_text_address(addr)) - ret = -EINVAL; - else { - /* Todo: Verify probepoint is a function entry point */ + /* Verify probepoint is a function entry point */ + if (kallsyms_lookup_size_offset(addr, NULL, &offset) && + offset == 0) { jp->kp.pre_handler = setjmp_pre_handler; jp->kp.break_handler = longjmp_break_handler; ret = register_kprobe(&jp->kp); - } + } else + ret = -EINVAL; + if (ret < 0) { if (i > 0) unregister_jprobes(jps, i); diff --git a/kernel/module.c b/kernel/module.c index ccd6419..2df4630 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -55,6 +55,7 @@ #include <linux/async.h> #include <linux/percpu.h> #include <linux/kmemleak.h> +#include <linux/jump_label.h> #define CREATE_TRACE_POINTS #include <trace/events/module.h> @@ -2309,6 +2310,11 @@ static void find_module_sections(struct module *mod, struct load_info *info) sizeof(*mod->tracepoints), &mod->num_tracepoints); #endif +#ifdef HAVE_JUMP_LABEL + mod->jump_entries = section_objs(info, "__jump_table", + sizeof(*mod->jump_entries), + &mod->num_jump_entries); +#endif #ifdef CONFIG_EVENT_TRACING mod->trace_events = section_objs(info, "_ftrace_events", sizeof(*mod->trace_events), diff --git a/kernel/perf_event.c b/kernel/perf_event.c index b98bed3..f309e80 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -31,24 +31,18 @@ #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> -/* - * Each CPU has a list of per CPU events: - */ -static DEFINE_PER_CPU(struct perf_cpu_context, perf_cpu_context); - -int perf_max_events __read_mostly = 1; -static int perf_reserved_percpu __read_mostly; -static int perf_overcommit __read_mostly = 1; - -static atomic_t nr_events __read_mostly; +atomic_t perf_task_events __read_mostly; static atomic_t nr_mmap_events __read_mostly; static atomic_t nr_comm_events __read_mostly; static atomic_t nr_task_events __read_mostly; +static LIST_HEAD(pmus); +static DEFINE_MUTEX(pmus_lock); +static struct srcu_struct pmus_srcu; + /* * perf event paranoia level: * -1 - not paranoid at all @@ -67,36 +61,43 @@ int sysctl_perf_event_sample_rate __read_mostly = 100000; static atomic64_t perf_event_id; -/* - * Lock for (sysadmin-configurable) event reservations: - */ -static DEFINE_SPINLOCK(perf_resource_lock); +void __weak perf_event_print_debug(void) { } -/* - * Architecture provided APIs - weak aliases: - */ -extern __weak const struct pmu *hw_perf_event_init(struct perf_event *event) +extern __weak const char *perf_pmu_name(void) { - return NULL; + return "pmu"; } -void __weak hw_perf_disable(void) { barrier(); } -void __weak hw_perf_enable(void) { barrier(); } - -void __weak perf_event_print_debug(void) { } - -static DEFINE_PER_CPU(int, perf_disable_count); +void perf_pmu_disable(struct pmu *pmu) +{ + int *count = this_cpu_ptr(pmu->pmu_disable_count); + if (!(*count)++) + pmu->pmu_disable(pmu); +} -void perf_disable(void) +void perf_pmu_enable(struct pmu *pmu) { - if (!__get_cpu_var(perf_disable_count)++) - hw_perf_disable(); + int *count = this_cpu_ptr(pmu->pmu_disable_count); + if (!--(*count)) + pmu->pmu_enable(pmu); } -void perf_enable(void) +static DEFINE_PER_CPU(struct list_head, rotation_list); + +/* + * perf_pmu_rotate_start() and perf_rotate_context() are fully serialized + * because they're strictly cpu affine and rotate_start is called with IRQs + * disabled, while rotate_context is called from IRQ context. + */ +static void perf_pmu_rotate_start(struct pmu *pmu) { - if (!--__get_cpu_var(perf_disable_count)) - hw_perf_enable(); + struct perf_cpu_context *cpuctx = this_cpu_ptr(pmu->pmu_cpu_context); + struct list_head *head = &__get_cpu_var(rotation_list); + + WARN_ON(!irqs_disabled()); + + if (list_empty(&cpuctx->rotation_list)) + list_add(&cpuctx->rotation_list, head); } static void get_ctx(struct perf_event_context *ctx) @@ -151,13 +152,13 @@ static u64 primary_event_id(struct perf_event *event) * the context could get moved to another task. */ static struct perf_event_context * -perf_lock_task_context(struct task_struct *task, unsigned long *flags) +perf_lock_task_context(struct task_struct *task, int ctxn, unsigned long *flags) { struct perf_event_context *ctx; rcu_read_lock(); - retry: - ctx = rcu_dereference(task->perf_event_ctxp); +retry: + ctx = rcu_dereference(task->perf_event_ctxp[ctxn]); if (ctx) { /* * If this context is a clone of another, it might @@ -170,7 +171,7 @@ perf_lock_task_context(struct task_struct *task, unsigned long *flags) * can't get swapped on us any more. */ raw_spin_lock_irqsave(&ctx->lock, *flags); - if (ctx != rcu_dereference(task->perf_event_ctxp)) { + if (ctx != rcu_dereference(task->perf_event_ctxp[ctxn])) { raw_spin_unlock_irqrestore(&ctx->lock, *flags); goto retry; } @@ -189,12 +190,13 @@ perf_lock_task_context(struct task_struct *task, unsigned long *flags) * can't get swapped to another task. This also increments its * reference count so that the context can't get freed. */ -static struct perf_event_context *perf_pin_task_context(struct task_struct *task) +static struct perf_event_context * +perf_pin_task_context(struct task_struct *task, int ctxn) { struct perf_event_context *ctx; unsigned long flags; - ctx = perf_lock_task_context(task, &flags); + ctx = perf_lock_task_context(task, ctxn, &flags); if (ctx) { ++ctx->pin_count; raw_spin_unlock_irqrestore(&ctx->lock, flags); @@ -302,6 +304,8 @@ list_add_event(struct perf_event *event, struct perf_event_context *ctx) } list_add_rcu(&event->event_entry, &ctx->event_list); + if (!ctx->nr_events) + perf_pmu_rotate_start(ctx->pmu); ctx->nr_events++; if (event->attr.inherit_stat) ctx->nr_stat++; @@ -311,7 +315,12 @@ static void perf_group_attach(struct perf_event *event) { struct perf_event *group_leader = event->group_leader; - WARN_ON_ONCE(event->attach_state & PERF_ATTACH_GROUP); + /* + * We can have double attach due to group movement in perf_event_open. + */ + if (event->attach_state & PERF_ATTACH_GROUP) + return; + event->attach_state |= PERF_ATTACH_GROUP; if (group_leader == event) @@ -408,8 +417,8 @@ event_filter_match(struct perf_event *event) return event->cpu == -1 || event->cpu == smp_processor_id(); } -static void -event_sched_out(struct perf_event *event, +static int +__event_sched_out(struct perf_event *event, struct perf_cpu_context *cpuctx, struct perf_event_context *ctx) { @@ -428,15 +437,14 @@ event_sched_out(struct perf_event *event, } if (event->state != PERF_EVENT_STATE_ACTIVE) - return; + return 0; event->state = PERF_EVENT_STATE_INACTIVE; if (event->pending_disable) { event->pending_disable = 0; event->state = PERF_EVENT_STATE_OFF; } - event->tstamp_stopped = ctx->time; - event->pmu->disable(event); + event->pmu->del(event, 0); event->oncpu = -1; if (!is_software_event(event)) @@ -444,6 +452,19 @@ event_sched_out(struct perf_event *event, ctx->nr_active--; if (event->attr.exclusive || !cpuctx->active_oncpu) cpuctx->exclusive = 0; + return 1; +} + +static void +event_sched_out(struct perf_event *event, + struct perf_cpu_context *cpuctx, + struct perf_event_context *ctx) +{ + int ret; + + ret = __event_sched_out(event, cpuctx, ctx); + if (ret) + event->tstamp_stopped = ctx->time; } static void @@ -466,6 +487,12 @@ group_sched_out(struct perf_event *group_event, cpuctx->exclusive = 0; } +static inline struct perf_cpu_context * +__get_cpu_context(struct perf_event_context *ctx) +{ + return this_cpu_ptr(ctx->pmu->pmu_cpu_context); +} + /* * Cross CPU call to remove a performance event * @@ -474,9 +501,9 @@ group_sched_out(struct perf_event *group_event, */ static void __perf_event_remove_from_context(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; + struct perf_cpu_context *cpuctx = __get_cpu_context(ctx); /* * If this is a task context, we need to check whether it is @@ -487,27 +514,11 @@ static void __perf_event_remove_from_context(void *info) return; raw_spin_lock(&ctx->lock); - /* - * Protect the list operation against NMI by disabling the - * events on a global level. - */ - perf_disable(); event_sched_out(event, cpuctx, ctx); list_del_event(event, ctx); - if (!ctx->task) { - /* - * Allow more per task events with respect to the - * reservation: - */ - cpuctx->max_pertask = - min(perf_max_events - ctx->nr_events, - perf_max_events - perf_reserved_percpu); - } - - perf_enable(); raw_spin_unlock(&ctx->lock); } @@ -572,8 +583,8 @@ retry: static void __perf_event_disable(void *info) { struct perf_event *event = info; - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_event_context *ctx = event->ctx; + struct perf_cpu_context *cpuctx = __get_cpu_context(ctx); /* * If this is a per-task event, need to check whether this @@ -628,7 +639,7 @@ void perf_event_disable(struct perf_event *event) return; } - retry: +retry: task_oncpu_function_call(task, __perf_event_disable, event); raw_spin_lock_irq(&ctx->lock); @@ -653,7 +664,7 @@ void perf_event_disable(struct perf_event *event) } static int -event_sched_in(struct perf_event *event, +__event_sched_in(struct perf_event *event, struct perf_cpu_context *cpuctx, struct perf_event_context *ctx) { @@ -667,14 +678,12 @@ event_sched_in(struct perf_event *event, */ smp_wmb(); - if (event->pmu->enable(event)) { + if (event->pmu->add(event, PERF_EF_START)) { event->state = PERF_EVENT_STATE_INACTIVE; event->oncpu = -1; return -EAGAIN; } - event->tstamp_running += ctx->time - event->tstamp_stopped; - if (!is_software_event(event)) cpuctx->active_oncpu++; ctx->nr_active++; @@ -685,28 +694,56 @@ event_sched_in(struct perf_event *event, return 0; } +static inline int +event_sched_in(struct perf_event *event, + struct perf_cpu_context *cpuctx, + struct perf_event_context *ctx) +{ + int ret = __event_sched_in(event, cpuctx, ctx); + if (ret) + return ret; + event->tstamp_running += ctx->time - event->tstamp_stopped; + return 0; +} + +static void +group_commit_event_sched_in(struct perf_event *group_event, + struct perf_cpu_context *cpuctx, + struct perf_event_context *ctx) +{ + struct perf_event *event; + u64 now = ctx->time; + + group_event->tstamp_running += now - group_event->tstamp_stopped; + /* + * Schedule in siblings as one group (if any): + */ + list_for_each_entry(event, &group_event->sibling_list, group_entry) { + event->tstamp_running += now - event->tstamp_stopped; + } +} + static int group_sched_in(struct perf_event *group_event, struct perf_cpu_context *cpuctx, struct perf_event_context *ctx) { struct perf_event *event, *partial_group = NULL; - const struct pmu *pmu = group_event->pmu; - bool txn = false; + struct pmu *pmu = group_event->pmu; if (group_event->state == PERF_EVENT_STATE_OFF) return 0; - /* Check if group transaction availabe */ - if (pmu->start_txn) - txn = true; + pmu->start_txn(pmu); - if (txn) - pmu->start_txn(pmu); - - if (event_sched_in(group_event, cpuctx, ctx)) { - if (txn) - pmu->cancel_txn(pmu); + /* + * use __event_sched_in() to delay updating tstamp_running + * until the transaction is committed. In case of failure + * we will keep an unmodified tstamp_running which is a + * requirement to get correct timing information + */ + if (__event_sched_in(group_event, cpuctx, ctx)) { + pmu->cancel_txn(pmu); return -EAGAIN; } @@ -714,29 +751,33 @@ group_sched_in(struct perf_event *group_event, * Schedule in siblings as one group (if any): */ list_for_each_entry(event, &group_event->sibling_list, group_entry) { - if (event_sched_in(event, cpuctx, ctx)) { + if (__event_sched_in(event, cpuctx, ctx)) { partial_group = event; goto group_error; } } - if (!txn || !pmu->commit_txn(pmu)) + if (!pmu->commit_txn(pmu)) { + /* commit tstamp_running */ + group_commit_event_sched_in(group_event, cpuctx, ctx); return 0; - + } group_error: /* * Groups can be scheduled in as one unit only, so undo any * partial group before returning: + * + * use __event_sched_out() to avoid updating tstamp_stopped + * because the event never actually ran */ list_for_each_entry(event, &group_event->sibling_list, group_entry) { if (event == partial_group) break; - event_sched_out(event, cpuctx, ctx); + __event_sched_out(event, cpuctx, ctx); } - event_sched_out(group_event, cpuctx, ctx); + __event_sched_out(group_event, cpuctx, ctx); - if (txn) - pmu->cancel_txn(pmu); + pmu->cancel_txn(pmu); return -EAGAIN; } @@ -789,10 +830,10 @@ static void add_event_to_ctx(struct perf_event *event, */ static void __perf_install_in_context(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; struct perf_event *leader = event->group_leader; + struct perf_cpu_context *cpuctx = __get_cpu_context(ctx); int err; /* @@ -812,12 +853,6 @@ static void __perf_install_in_context(void *info) ctx->is_active = 1; update_context_time(ctx); - /* - * Protect the list operation against NMI by disabling the - * events on a global level. NOP for non NMI based events. - */ - perf_disable(); - add_event_to_ctx(event, ctx); if (event->cpu != -1 && event->cpu != smp_processor_id()) @@ -855,12 +890,7 @@ static void __perf_install_in_context(void *info) } } - if (!err && !ctx->task && cpuctx->max_pertask) - cpuctx->max_pertask--; - - unlock: - perf_enable(); - +unlock: raw_spin_unlock(&ctx->lock); } @@ -883,6 +913,8 @@ perf_install_in_context(struct perf_event_context *ctx, { struct task_struct *task = ctx->task; + event->ctx = ctx; + if (!task) { /* * Per cpu events are installed via an smp call and @@ -931,10 +963,12 @@ static void __perf_event_mark_enabled(struct perf_event *event, event->state = PERF_EVENT_STATE_INACTIVE; event->tstamp_enabled = ctx->time - event->total_time_enabled; - list_for_each_entry(sub, &event->sibling_list, group_entry) - if (sub->state >= PERF_EVENT_STATE_INACTIVE) + list_for_each_entry(sub, &event->sibling_list, group_entry) { + if (sub->state >= PERF_EVENT_STATE_INACTIVE) { sub->tstamp_enabled = ctx->time - sub->total_time_enabled; + } + } } /* @@ -943,9 +977,9 @@ static void __perf_event_mark_enabled(struct perf_event *event, static void __perf_event_enable(void *info) { struct perf_event *event = info; - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); struct perf_event_context *ctx = event->ctx; struct perf_event *leader = event->group_leader; + struct perf_cpu_context *cpuctx = __get_cpu_context(ctx); int err; /* @@ -979,12 +1013,10 @@ static void __perf_event_enable(void *info) if (!group_can_go_on(event, cpuctx, 1)) { err = -EEXIST; } else { - perf_disable(); if (event == leader) err = group_sched_in(event, cpuctx, ctx); else err = event_sched_in(event, cpuctx, ctx); - perf_enable(); } if (err) { @@ -1000,7 +1032,7 @@ static void __perf_event_enable(void *info) } } - unlock: +unlock: raw_spin_unlock(&ctx->lock); } @@ -1041,7 +1073,7 @@ void perf_event_enable(struct perf_event *event) if (event->state == PERF_EVENT_STATE_ERROR) event->state = PERF_EVENT_STATE_OFF; - retry: +retry: raw_spin_unlock_irq(&ctx->lock); task_oncpu_function_call(task, __perf_event_enable, event); @@ -1061,7 +1093,7 @@ void perf_event_enable(struct perf_event *event) if (event->state == PERF_EVENT_STATE_OFF) __perf_event_mark_enabled(event, ctx); - out: +out: raw_spin_unlock_irq(&ctx->lock); } @@ -1092,26 +1124,26 @@ static void ctx_sched_out(struct perf_event_context *ctx, struct perf_event *event; raw_spin_lock(&ctx->lock); + perf_pmu_disable(ctx->pmu); ctx->is_active = 0; if (likely(!ctx->nr_events)) goto out; update_context_time(ctx); - perf_disable(); if (!ctx->nr_active) - goto out_enable; + goto out; - if (event_type & EVENT_PINNED) + if (event_type & EVENT_PINNED) { list_for_each_entry(event, &ctx->pinned_groups, group_entry) group_sched_out(event, cpuctx, ctx); + } - if (event_type & EVENT_FLEXIBLE) + if (event_type & EVENT_FLEXIBLE) { list_for_each_entry(event, &ctx->flexible_groups, group_entry) group_sched_out(event, cpuctx, ctx); - - out_enable: - perf_enable(); - out: + } +out: + perf_pmu_enable(ctx->pmu); raw_spin_unlock(&ctx->lock); } @@ -1209,34 +1241,25 @@ static void perf_event_sync_stat(struct perf_event_context *ctx, } } -/* - * Called from scheduler to remove the events of the current task, - * with interrupts disabled. - * - * We stop each event and update the event value in event->count. - * - * This does not protect us against NMI, but disable() - * sets the disabled bit in the control field of event _before_ - * accessing the event control register. If a NMI hits, then it will - * not restart the event. - */ -void perf_event_task_sched_out(struct task_struct *task, - struct task_struct *next) +void perf_event_context_sched_out(struct task_struct *task, int ctxn, + struct task_struct *next) { - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); - struct perf_event_context *ctx = task->perf_event_ctxp; + struct perf_event_context *ctx = task->perf_event_ctxp[ctxn]; struct perf_event_context *next_ctx; struct perf_event_context *parent; + struct perf_cpu_context *cpuctx; int do_switch = 1; - perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, 1, NULL, 0); + if (likely(!ctx)) + return; - if (likely(!ctx || !cpuctx->task_ctx)) + cpuctx = __get_cpu_context(ctx); + if (!cpuctx->task_ctx) return; rcu_read_lock(); parent = rcu_dereference(ctx->parent_ctx); - next_ctx = next->perf_event_ctxp; + next_ctx = next->perf_event_ctxp[ctxn]; if (parent && next_ctx && rcu_dereference(next_ctx->parent_ctx) == parent) { /* @@ -1255,8 +1278,8 @@ void perf_event_task_sched_out(struct task_struct *task, * XXX do we need a memory barrier of sorts * wrt to rcu_dereference() of perf_event_ctxp */ - task->perf_event_ctxp = next_ctx; - next->perf_event_ctxp = ctx; + task->perf_event_ctxp[ctxn] = next_ctx; + next->perf_event_ctxp[ctxn] = ctx; ctx->task = next; next_ctx->task = task; do_switch = 0; @@ -1274,10 +1297,35 @@ void perf_event_task_sched_out(struct task_struct *task, } } +#define for_each_task_context_nr(ctxn) \ + for ((ctxn) = 0; (ctxn) < perf_nr_task_contexts; (ctxn)++) + +/* + * Called from scheduler to remove the events of the current task, + * with interrupts disabled. + * + * We stop each event and update the event value in event->count. + * + * This does not protect us against NMI, but disable() + * sets the disabled bit in the control field of event _before_ + * accessing the event control register. If a NMI hits, then it will + * not restart the event. + */ +void __perf_event_task_sched_out(struct task_struct *task, + struct task_struct *next) +{ + int ctxn; + + perf_sw_event(PERF_COUNT_SW_CONTEXT_SWITCHES, 1, 1, NULL, 0); + + for_each_task_context_nr(ctxn) + perf_event_context_sched_out(task, ctxn, next); +} + static void task_ctx_sched_out(struct perf_event_context *ctx, enum event_type_t event_type) { - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + struct perf_cpu_context *cpuctx = __get_cpu_context(ctx); if (!cpuctx->task_ctx) return; @@ -1292,14 +1340,6 @@ static void task_ctx_sched_out(struct perf_event_context *ctx, /* * Called with IRQs disabled */ -static void __perf_event_task_sched_out(struct perf_event_context *ctx) -{ - task_ctx_sched_out(ctx, EVENT_ALL); -} - -/* - * Called with IRQs disabled - */ static void cpu_ctx_sched_out(struct perf_cpu_context *cpuctx, enum event_type_t event_type) { @@ -1350,9 +1390,10 @@ ctx_flexible_sched_in(struct perf_event_context *ctx, if (event->cpu != -1 && event->cpu != smp_processor_id()) continue; - if (group_can_go_on(event, cpuctx, can_add_hw)) + if (group_can_go_on(event, cpuctx, can_add_hw)) { if (group_sched_in(event, cpuctx, ctx)) can_add_hw = 0; + } } } @@ -1368,8 +1409,6 @@ ctx_sched_in(struct perf_event_context *ctx, ctx->timestamp = perf_clock(); - perf_disable(); - /* * First go through the list and put on any pinned groups * in order to give them the best chance of going on. @@ -1381,8 +1420,7 @@ ctx_sched_in(struct perf_event_context *ctx, if (event_type & EVENT_FLEXIBLE) ctx_flexible_sched_in(ctx, cpuctx); - perf_enable(); - out: +out: raw_spin_unlock(&ctx->lock); } @@ -1394,43 +1432,28 @@ static void cpu_ctx_sched_in(struct perf_cpu_context *cpuctx, ctx_sched_in(ctx, cpuctx, event_type); } -static void task_ctx_sched_in(struct task_struct *task, +static void task_ctx_sched_in(struct perf_event_context *ctx, enum event_type_t event_type) { - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); - struct perf_event_context *ctx = task->perf_event_ctxp; + struct perf_cpu_context *cpuctx; - if (likely(!ctx)) - return; + cpuctx = __get_cpu_context(ctx); if (cpuctx->task_ctx == ctx) return; + ctx_sched_in(ctx, cpuctx, event_type); cpuctx->task_ctx = ctx; } -/* - * Called from scheduler to add the events of the current task - * with interrupts disabled. - * - * We restore the event value and then enable it. - * - * This does not protect us against NMI, but enable() - * sets the enabled bit in the control field of event _before_ - * accessing the event control register. If a NMI hits, then it will - * keep the event running. - */ -void perf_event_task_sched_in(struct task_struct *task) -{ - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); - struct perf_event_context *ctx = task->perf_event_ctxp; - if (likely(!ctx)) - return; +void perf_event_context_sched_in(struct perf_event_context *ctx) +{ + struct perf_cpu_context *cpuctx; + cpuctx = __get_cpu_context(ctx); if (cpuctx->task_ctx == ctx) return; - perf_disable(); - + perf_pmu_disable(ctx->pmu); /* * We want to keep the following priority order: * cpu pinned (that don't need to move), task pinned, @@ -1444,7 +1467,37 @@ void perf_event_task_sched_in(struct task_struct *task) cpuctx->task_ctx = ctx; - perf_enable(); + /* + * Since these rotations are per-cpu, we need to ensure the + * cpu-context we got scheduled on is actually rotating. + */ + perf_pmu_rotate_start(ctx->pmu); + perf_pmu_enable(ctx->pmu); +} + +/* + * Called from scheduler to add the events of the current task + * with interrupts disabled. + * + * We restore the event value and then enable it. + * + * This does not protect us against NMI, but enable() + * sets the enabled bit in the control field of event _before_ + * accessing the event control register. If a NMI hits, then it will + * keep the event running. + */ +void __perf_event_task_sched_in(struct task_struct *task) +{ + struct perf_event_context *ctx; + int ctxn; + + for_each_task_context_nr(ctxn) { + ctx = task->perf_event_ctxp[ctxn]; + if (likely(!ctx)) + continue; + + perf_event_context_sched_in(ctx); + } } #define MAX_INTERRUPTS (~0ULL) @@ -1524,22 +1577,6 @@ do { \ return div64_u64(dividend, divisor); } -static void perf_event_stop(struct perf_event *event) -{ - if (!event->pmu->stop) - return event->pmu->disable(event); - - return event->pmu->stop(event); -} - -static int perf_event_start(struct perf_event *event) -{ - if (!event->pmu->start) - return event->pmu->enable(event); - - return event->pmu->start(event); -} - static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count) { struct hw_perf_event *hwc = &event->hw; @@ -1559,15 +1596,13 @@ static void perf_adjust_period(struct perf_event *event, u64 nsec, u64 count) hwc->sample_period = sample_period; if (local64_read(&hwc->period_left) > 8*sample_period) { - perf_disable(); - perf_event_stop(event); + event->pmu->stop(event, PERF_EF_UPDATE); local64_set(&hwc->period_left, 0); - perf_event_start(event); - perf_enable(); + event->pmu->start(event, PERF_EF_RELOAD); } } -static void perf_ctx_adjust_freq(struct perf_event_context *ctx) +static void perf_ctx_adjust_freq(struct perf_event_context *ctx, u64 period) { struct perf_event *event; struct hw_perf_event *hwc; @@ -1592,23 +1627,19 @@ static void perf_ctx_adjust_freq(struct perf_event_context *ctx) */ if (interrupts == MAX_INTERRUPTS) { perf_log_throttle(event, 1); - perf_disable(); - event->pmu->unthrottle(event); - perf_enable(); + event->pmu->start(event, 0); } if (!event->attr.freq || !event->attr.sample_freq) continue; - perf_disable(); event->pmu->read(event); now = local64_read(&event->count); delta = now - hwc->freq_count_stamp; hwc->freq_count_stamp = now; if (delta > 0) - perf_adjust_period(event, TICK_NSEC, delta); - perf_enable(); + perf_adjust_period(event, period, delta); } raw_spin_unlock(&ctx->lock); } @@ -1626,32 +1657,38 @@ static void rotate_ctx(struct perf_event_context *ctx) raw_spin_unlock(&ctx->lock); } -void perf_event_task_tick(struct task_struct *curr) +/* + * perf_pmu_rotate_start() and perf_rotate_context() are fully serialized + * because they're strictly cpu affine and rotate_start is called with IRQs + * disabled, while rotate_context is called from IRQ context. + */ +static void perf_rotate_context(struct perf_cpu_context *cpuctx) { - struct perf_cpu_context *cpuctx; - struct perf_event_context *ctx; - int rotate = 0; - - if (!atomic_read(&nr_events)) - return; + u64 interval = (u64)cpuctx->jiffies_interval * TICK_NSEC; + struct perf_event_context *ctx = NULL; + int rotate = 0, remove = 1; - cpuctx = &__get_cpu_var(perf_cpu_context); - if (cpuctx->ctx.nr_events && - cpuctx->ctx.nr_events != cpuctx->ctx.nr_active) - rotate = 1; + if (cpuctx->ctx.nr_events) { + remove = 0; + if (cpuctx->ctx.nr_events != cpuctx->ctx.nr_active) + rotate = 1; + } - ctx = curr->perf_event_ctxp; - if (ctx && ctx->nr_events && ctx->nr_events != ctx->nr_active) - rotate = 1; + ctx = cpuctx->task_ctx; + if (ctx && ctx->nr_events) { + remove = 0; + if (ctx->nr_events != ctx->nr_active) + rotate = 1; + } - perf_ctx_adjust_freq(&cpuctx->ctx); + perf_pmu_disable(cpuctx->ctx.pmu); + perf_ctx_adjust_freq(&cpuctx->ctx, interval); if (ctx) - perf_ctx_adjust_freq(ctx); + perf_ctx_adjust_freq(ctx, interval); if (!rotate) - return; + goto done; - perf_disable(); cpu_ctx_sched_out(cpuctx, EVENT_FLEXIBLE); if (ctx) task_ctx_sched_out(ctx, EVENT_FLEXIBLE); @@ -1662,8 +1699,27 @@ void perf_event_task_tick(struct task_struct *curr) cpu_ctx_sched_in(cpuctx, EVENT_FLEXIBLE); if (ctx) - task_ctx_sched_in(curr, EVENT_FLEXIBLE); - perf_enable(); + task_ctx_sched_in(ctx, EVENT_FLEXIBLE); + +done: + if (remove) + list_del_init(&cpuctx->rotation_list); + + perf_pmu_enable(cpuctx->ctx.pmu); +} + +void perf_event_task_tick(void) +{ + struct list_head *head = &__get_cpu_var(rotation_list); + struct perf_cpu_context *cpuctx, *tmp; + + WARN_ON(!irqs_disabled()); + + list_for_each_entry_safe(cpuctx, tmp, head, rotation_list) { + if (cpuctx->jiffies_interval == 1 || + !(jiffies % cpuctx->jiffies_interval)) + perf_rotate_context(cpuctx); + } } static int event_enable_on_exec(struct perf_event *event, @@ -1685,20 +1741,18 @@ static int event_enable_on_exec(struct perf_event *event, * Enable all of a task's events that have been marked enable-on-exec. * This expects task == current. */ -static void perf_event_enable_on_exec(struct task_struct *task) +static void perf_event_enable_on_exec(struct perf_event_context *ctx) { - struct perf_event_context *ctx; struct perf_event *event; unsigned long flags; int enabled = 0; int ret; local_irq_save(flags); - ctx = task->perf_event_ctxp; if (!ctx || !ctx->nr_events) goto out; - __perf_event_task_sched_out(ctx); + task_ctx_sched_out(ctx, EVENT_ALL); raw_spin_lock(&ctx->lock); @@ -1722,8 +1776,8 @@ static void perf_event_enable_on_exec(struct task_struct *task) raw_spin_unlock(&ctx->lock); - perf_event_task_sched_in(task); - out: + perf_event_context_sched_in(ctx); +out: local_irq_restore(flags); } @@ -1732,9 +1786,9 @@ static void perf_event_enable_on_exec(struct task_struct *task) */ 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; + struct perf_cpu_context *cpuctx = __get_cpu_context(ctx); /* * If this is a task context, we need to check whether it is @@ -1773,7 +1827,13 @@ static u64 perf_event_read(struct perf_event *event) unsigned long flags; raw_spin_lock_irqsave(&ctx->lock, flags); - update_context_time(ctx); + /* + * may read while context is not active + * (e.g., thread is blocked), in that case + * we cannot update context time + */ + if (ctx->is_active) + update_context_time(ctx); update_event_times(event); raw_spin_unlock_irqrestore(&ctx->lock, flags); } @@ -1782,11 +1842,219 @@ static u64 perf_event_read(struct perf_event *event) } /* - * Initialize the perf_event context in a task_struct: + * Callchain support */ + +struct callchain_cpus_entries { + struct rcu_head rcu_head; + struct perf_callchain_entry *cpu_entries[0]; +}; + +static DEFINE_PER_CPU(int, callchain_recursion[PERF_NR_CONTEXTS]); +static atomic_t nr_callchain_events; +static DEFINE_MUTEX(callchain_mutex); +struct callchain_cpus_entries *callchain_cpus_entries; + + +__weak void perf_callchain_kernel(struct perf_callchain_entry *entry, + struct pt_regs *regs) +{ +} + +__weak void perf_callchain_user(struct perf_callchain_entry *entry, + struct pt_regs *regs) +{ +} + +static void release_callchain_buffers_rcu(struct rcu_head *head) +{ + struct callchain_cpus_entries *entries; + int cpu; + + entries = container_of(head, struct callchain_cpus_entries, rcu_head); + + for_each_possible_cpu(cpu) + kfree(entries->cpu_entries[cpu]); + + kfree(entries); +} + +static void release_callchain_buffers(void) +{ + struct callchain_cpus_entries *entries; + + entries = callchain_cpus_entries; + rcu_assign_pointer(callchain_cpus_entries, NULL); + call_rcu(&entries->rcu_head, release_callchain_buffers_rcu); +} + +static int alloc_callchain_buffers(void) +{ + int cpu; + int size; + struct callchain_cpus_entries *entries; + + /* + * We can't use the percpu allocation API for data that can be + * accessed from NMI. Use a temporary manual per cpu allocation + * until that gets sorted out. + */ + size = sizeof(*entries) + sizeof(struct perf_callchain_entry *) * + num_possible_cpus(); + + entries = kzalloc(size, GFP_KERNEL); + if (!entries) + return -ENOMEM; + + size = sizeof(struct perf_callchain_entry) * PERF_NR_CONTEXTS; + + for_each_possible_cpu(cpu) { + entries->cpu_entries[cpu] = kmalloc_node(size, GFP_KERNEL, + cpu_to_node(cpu)); + if (!entries->cpu_entries[cpu]) + goto fail; + } + + rcu_assign_pointer(callchain_cpus_entries, entries); + + return 0; + +fail: + for_each_possible_cpu(cpu) + kfree(entries->cpu_entries[cpu]); + kfree(entries); + + return -ENOMEM; +} + +static int get_callchain_buffers(void) +{ + int err = 0; + int count; + + mutex_lock(&callchain_mutex); + + count = atomic_inc_return(&nr_callchain_events); + if (WARN_ON_ONCE(count < 1)) { + err = -EINVAL; + goto exit; + } + + if (count > 1) { + /* If the allocation failed, give up */ + if (!callchain_cpus_entries) + err = -ENOMEM; + goto exit; + } + + err = alloc_callchain_buffers(); + if (err) + release_callchain_buffers(); +exit: + mutex_unlock(&callchain_mutex); + + return err; +} + +static void put_callchain_buffers(void) +{ + if (atomic_dec_and_mutex_lock(&nr_callchain_events, &callchain_mutex)) { + release_callchain_buffers(); + mutex_unlock(&callchain_mutex); + } +} + +static int get_recursion_context(int *recursion) +{ + int rctx; + + if (in_nmi()) + rctx = 3; + else if (in_irq()) + rctx = 2; + else if (in_softirq()) + rctx = 1; + else + rctx = 0; + + if (recursion[rctx]) + return -1; + + recursion[rctx]++; + barrier(); + + return rctx; +} + +static inline void put_recursion_context(int *recursion, int rctx) +{ + barrier(); + recursion[rctx]--; +} + +static struct perf_callchain_entry *get_callchain_entry(int *rctx) +{ + int cpu; + struct callchain_cpus_entries *entries; + + *rctx = get_recursion_context(__get_cpu_var(callchain_recursion)); + if (*rctx == -1) + return NULL; + + entries = rcu_dereference(callchain_cpus_entries); + if (!entries) + return NULL; + + cpu = smp_processor_id(); + + return &entries->cpu_entries[cpu][*rctx]; +} + static void -__perf_event_init_context(struct perf_event_context *ctx, - struct task_struct *task) +put_callchain_entry(int rctx) +{ + put_recursion_context(__get_cpu_var(callchain_recursion), rctx); +} + +static struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) +{ + int rctx; + struct perf_callchain_entry *entry; + + + entry = get_callchain_entry(&rctx); + if (rctx == -1) + return NULL; + + if (!entry) + goto exit_put; + + entry->nr = 0; + + if (!user_mode(regs)) { + perf_callchain_store(entry, PERF_CONTEXT_KERNEL); + perf_callchain_kernel(entry, regs); + if (current->mm) + regs = task_pt_regs(current); + else + regs = NULL; + } + + if (regs) { + perf_callchain_store(entry, PERF_CONTEXT_USER); + perf_callchain_user(entry, regs); + } + +exit_put: + put_callchain_entry(rctx); + + return entry; +} + +/* + * Initialize the perf_event context in a task_struct: + */ +static void __perf_event_init_context(struct perf_event_context *ctx) { raw_spin_lock_init(&ctx->lock); mutex_init(&ctx->mutex); @@ -1794,45 +2062,38 @@ __perf_event_init_context(struct perf_event_context *ctx, INIT_LIST_HEAD(&ctx->flexible_groups); INIT_LIST_HEAD(&ctx->event_list); atomic_set(&ctx->refcount, 1); - ctx->task = task; } -static struct perf_event_context *find_get_context(pid_t pid, int cpu) +static struct perf_event_context * +alloc_perf_context(struct pmu *pmu, struct task_struct *task) { struct perf_event_context *ctx; - struct perf_cpu_context *cpuctx; - struct task_struct *task; - unsigned long flags; - int err; - - if (pid == -1 && cpu != -1) { - /* Must be root to operate on a CPU event: */ - if (perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN)) - return ERR_PTR(-EACCES); - if (cpu < 0 || cpu >= nr_cpumask_bits) - return ERR_PTR(-EINVAL); + ctx = kzalloc(sizeof(struct perf_event_context), GFP_KERNEL); + if (!ctx) + return NULL; - /* - * We could be clever and allow to attach a event to an - * offline CPU and activate it when the CPU comes up, but - * that's for later. - */ - if (!cpu_online(cpu)) - return ERR_PTR(-ENODEV); + __perf_event_init_context(ctx); + if (task) { + ctx->task = task; + get_task_struct(task); + } + ctx->pmu = pmu; - cpuctx = &per_cpu(perf_cpu_context, cpu); - ctx = &cpuctx->ctx; - get_ctx(ctx); + return ctx; +} - return ctx; - } +static struct task_struct * +find_lively_task_by_vpid(pid_t vpid) +{ + struct task_struct *task; + int err; rcu_read_lock(); - if (!pid) + if (!vpid) task = current; else - task = find_task_by_vpid(pid); + task = find_task_by_vpid(vpid); if (task) get_task_struct(task); rcu_read_unlock(); @@ -1852,36 +2113,78 @@ static struct perf_event_context *find_get_context(pid_t pid, int cpu) if (!ptrace_may_access(task, PTRACE_MODE_READ)) goto errout; - retry: - ctx = perf_lock_task_context(task, &flags); + return task; +errout: + put_task_struct(task); + return ERR_PTR(err); + +} + +static struct perf_event_context * +find_get_context(struct pmu *pmu, struct task_struct *task, int cpu) +{ + struct perf_event_context *ctx; + struct perf_cpu_context *cpuctx; + unsigned long flags; + int ctxn, err; + + if (!task && cpu != -1) { + /* Must be root to operate on a CPU event: */ + if (perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN)) + return ERR_PTR(-EACCES); + + if (cpu < 0 || cpu >= nr_cpumask_bits) + return ERR_PTR(-EINVAL); + + /* + * We could be clever and allow to attach a event to an + * offline CPU and activate it when the CPU comes up, but + * that's for later. + */ + if (!cpu_online(cpu)) + return ERR_PTR(-ENODEV); + + cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu); + ctx = &cpuctx->ctx; + get_ctx(ctx); + + return ctx; + } + + err = -EINVAL; + ctxn = pmu->task_ctx_nr; + if (ctxn < 0) + goto errout; + +retry: + ctx = perf_lock_task_context(task, ctxn, &flags); if (ctx) { unclone_ctx(ctx); raw_spin_unlock_irqrestore(&ctx->lock, flags); } if (!ctx) { - ctx = kzalloc(sizeof(struct perf_event_context), GFP_KERNEL); + ctx = alloc_perf_context(pmu, task); err = -ENOMEM; if (!ctx) goto errout; - __perf_event_init_context(ctx, task); + get_ctx(ctx); - if (cmpxchg(&task->perf_event_ctxp, NULL, ctx)) { + + if (cmpxchg(&task->perf_event_ctxp[ctxn], NULL, ctx)) { /* * We raced with some other task; use * the context they set. */ + put_task_struct(task); kfree(ctx); goto retry; } - get_task_struct(task); } - put_task_struct(task); return ctx; - errout: - put_task_struct(task); +errout: return ERR_PTR(err); } @@ -1898,21 +2201,23 @@ static void free_event_rcu(struct rcu_head *head) kfree(event); } -static void perf_pending_sync(struct perf_event *event); static void perf_buffer_put(struct perf_buffer *buffer); static void free_event(struct perf_event *event) { - perf_pending_sync(event); + irq_work_sync(&event->pending); if (!event->parent) { - atomic_dec(&nr_events); + if (event->attach_state & PERF_ATTACH_TASK) + jump_label_dec(&perf_task_events); if (event->attr.mmap || event->attr.mmap_data) atomic_dec(&nr_mmap_events); if (event->attr.comm) atomic_dec(&nr_comm_events); if (event->attr.task) atomic_dec(&nr_task_events); + if (event->attr.sample_type & PERF_SAMPLE_CALLCHAIN) + put_callchain_buffers(); } if (event->buffer) { @@ -1923,7 +2228,9 @@ static void free_event(struct perf_event *event) if (event->destroy) event->destroy(event); - put_ctx(event->ctx); + if (event->ctx) + put_ctx(event->ctx); + call_rcu(&event->rcu_head, free_event_rcu); } @@ -2342,6 +2649,9 @@ int perf_event_task_disable(void) static int perf_event_index(struct perf_event *event) { + if (event->hw.state & PERF_HES_STOPPED) + return 0; + if (event->state != PERF_EVENT_STATE_ACTIVE) return 0; @@ -2845,16 +3155,7 @@ void perf_event_wakeup(struct perf_event *event) } } -/* - * Pending wakeups - * - * Handle the case where we need to wakeup up from NMI (or rq->lock) context. - * - * The NMI bit means we cannot possibly take locks. Therefore, maintain a - * single linked list and use cmpxchg() to add entries lockless. - */ - -static void perf_pending_event(struct perf_pending_entry *entry) +static void perf_pending_event(struct irq_work *entry) { struct perf_event *event = container_of(entry, struct perf_event, pending); @@ -2870,99 +3171,6 @@ static void perf_pending_event(struct perf_pending_entry *entry) } } -#define PENDING_TAIL ((struct perf_pending_entry *)-1UL) - -static DEFINE_PER_CPU(struct perf_pending_entry *, perf_pending_head) = { - PENDING_TAIL, -}; - -static void perf_pending_queue(struct perf_pending_entry *entry, - void (*func)(struct perf_pending_entry *)) -{ - struct perf_pending_entry **head; - - if (cmpxchg(&entry->next, NULL, PENDING_TAIL) != NULL) - return; - - entry->func = func; - - head = &get_cpu_var(perf_pending_head); - - do { - entry->next = *head; - } while (cmpxchg(head, entry->next, entry) != entry->next); - - set_perf_event_pending(); - - put_cpu_var(perf_pending_head); -} - -static int __perf_pending_run(void) -{ - struct perf_pending_entry *list; - int nr = 0; - - list = xchg(&__get_cpu_var(perf_pending_head), PENDING_TAIL); - while (list != PENDING_TAIL) { - void (*func)(struct perf_pending_entry *); - struct perf_pending_entry *entry = list; - - list = list->next; - - func = entry->func; - entry->next = NULL; - /* - * Ensure we observe the unqueue before we issue the wakeup, - * so that we won't be waiting forever. - * -- see perf_not_pending(). - */ - smp_wmb(); - - func(entry); - nr++; - } - - return nr; -} - -static inline int perf_not_pending(struct perf_event *event) -{ - /* - * If we flush on whatever cpu we run, there is a chance we don't - * need to wait. - */ - get_cpu(); - __perf_pending_run(); - put_cpu(); - - /* - * Ensure we see the proper queue state before going to sleep - * so that we do not miss the wakeup. -- see perf_pending_handle() - */ - smp_rmb(); - return event->pending.next == NULL; -} - -static void perf_pending_sync(struct perf_event *event) -{ - wait_event(event->waitq, perf_not_pending(event)); -} - -void perf_event_do_pending(void) -{ - __perf_pending_run(); -} - -/* - * Callchain support -- arch specific - */ - -__weak struct perf_callchain_entry *perf_callchain(struct pt_regs *regs) -{ - return NULL; -} - - /* * We assume there is only KVM supporting the callbacks. * Later on, we might change it to a list if there is @@ -3012,8 +3220,7 @@ static void perf_output_wakeup(struct perf_output_handle *handle) if (handle->nmi) { handle->event->pending_wakeup = 1; - perf_pending_queue(&handle->event->pending, - perf_pending_event); + irq_work_queue(&handle->event->pending); } else perf_event_wakeup(handle->event); } @@ -3069,7 +3276,7 @@ again: if (handle->wakeup != local_read(&buffer->wakeup)) perf_output_wakeup(handle); - out: +out: preempt_enable(); } @@ -3457,14 +3664,20 @@ static void perf_event_output(struct perf_event *event, int nmi, struct perf_output_handle handle; struct perf_event_header header; + /* protect the callchain buffers */ + rcu_read_lock(); + perf_prepare_sample(&header, data, event, regs); if (perf_output_begin(&handle, event, header.size, nmi, 1)) - return; + goto exit; perf_output_sample(&handle, &header, data, event); perf_output_end(&handle); + +exit: + rcu_read_unlock(); } /* @@ -3578,16 +3791,27 @@ static void perf_event_task_ctx(struct perf_event_context *ctx, 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; + struct perf_event_context *ctx; + struct pmu *pmu; + int ctxn; rcu_read_lock(); - cpuctx = &get_cpu_var(perf_cpu_context); - perf_event_task_ctx(&cpuctx->ctx, task_event); - if (!ctx) - ctx = rcu_dereference(current->perf_event_ctxp); - if (ctx) - perf_event_task_ctx(ctx, task_event); - put_cpu_var(perf_cpu_context); + list_for_each_entry_rcu(pmu, &pmus, entry) { + cpuctx = get_cpu_ptr(pmu->pmu_cpu_context); + perf_event_task_ctx(&cpuctx->ctx, task_event); + + ctx = task_event->task_ctx; + if (!ctx) { + ctxn = pmu->task_ctx_nr; + if (ctxn < 0) + goto next; + ctx = rcu_dereference(current->perf_event_ctxp[ctxn]); + } + if (ctx) + perf_event_task_ctx(ctx, task_event); +next: + put_cpu_ptr(pmu->pmu_cpu_context); + } rcu_read_unlock(); } @@ -3692,8 +3916,10 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event) { struct perf_cpu_context *cpuctx; struct perf_event_context *ctx; - unsigned int size; char comm[TASK_COMM_LEN]; + unsigned int size; + struct pmu *pmu; + int ctxn; memset(comm, 0, sizeof(comm)); strlcpy(comm, comm_event->task->comm, sizeof(comm)); @@ -3705,21 +3931,36 @@ 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); - ctx = rcu_dereference(current->perf_event_ctxp); - if (ctx) - perf_event_comm_ctx(ctx, comm_event); - put_cpu_var(perf_cpu_context); + list_for_each_entry_rcu(pmu, &pmus, entry) { + cpuctx = get_cpu_ptr(pmu->pmu_cpu_context); + perf_event_comm_ctx(&cpuctx->ctx, comm_event); + + ctxn = pmu->task_ctx_nr; + if (ctxn < 0) + goto next; + + ctx = rcu_dereference(current->perf_event_ctxp[ctxn]); + if (ctx) + perf_event_comm_ctx(ctx, comm_event); +next: + put_cpu_ptr(pmu->pmu_cpu_context); + } rcu_read_unlock(); } void perf_event_comm(struct task_struct *task) { struct perf_comm_event comm_event; + struct perf_event_context *ctx; + int ctxn; + + for_each_task_context_nr(ctxn) { + ctx = task->perf_event_ctxp[ctxn]; + if (!ctx) + continue; - if (task->perf_event_ctxp) - perf_event_enable_on_exec(task); + perf_event_enable_on_exec(ctx); + } if (!atomic_read(&nr_comm_events)) return; @@ -3821,6 +4062,8 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) char tmp[16]; char *buf = NULL; const char *name; + struct pmu *pmu; + int ctxn; memset(tmp, 0, sizeof(tmp)); @@ -3873,12 +4116,23 @@ 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, vma->vm_flags & VM_EXEC); - ctx = rcu_dereference(current->perf_event_ctxp); - if (ctx) - perf_event_mmap_ctx(ctx, mmap_event, vma->vm_flags & VM_EXEC); - put_cpu_var(perf_cpu_context); + list_for_each_entry_rcu(pmu, &pmus, entry) { + cpuctx = get_cpu_ptr(pmu->pmu_cpu_context); + perf_event_mmap_ctx(&cpuctx->ctx, mmap_event, + vma->vm_flags & VM_EXEC); + + ctxn = pmu->task_ctx_nr; + if (ctxn < 0) + goto next; + + ctx = rcu_dereference(current->perf_event_ctxp[ctxn]); + if (ctx) { + perf_event_mmap_ctx(ctx, mmap_event, + vma->vm_flags & VM_EXEC); + } +next: + put_cpu_ptr(pmu->pmu_cpu_context); + } rcu_read_unlock(); kfree(buf); @@ -3960,8 +4214,6 @@ static int __perf_event_overflow(struct perf_event *event, int nmi, struct hw_perf_event *hwc = &event->hw; int ret = 0; - throttle = (throttle && event->pmu->unthrottle != NULL); - if (!throttle) { hwc->interrupts++; } else { @@ -4004,8 +4256,7 @@ static int __perf_event_overflow(struct perf_event *event, int nmi, event->pending_kill = POLL_HUP; if (nmi) { event->pending_disable = 1; - perf_pending_queue(&event->pending, - perf_pending_event); + irq_work_queue(&event->pending); } else perf_event_disable(event); } @@ -4029,6 +4280,17 @@ int perf_event_overflow(struct perf_event *event, int nmi, * Generic software event infrastructure */ +struct swevent_htable { + struct swevent_hlist *swevent_hlist; + struct mutex hlist_mutex; + int hlist_refcount; + + /* Recursion avoidance in each contexts */ + int recursion[PERF_NR_CONTEXTS]; +}; + +static DEFINE_PER_CPU(struct swevent_htable, swevent_htable); + /* * We directly increment event->count and keep a second value in * event->hw.period_left to count intervals. This period event @@ -4086,7 +4348,7 @@ static void perf_swevent_overflow(struct perf_event *event, u64 overflow, } } -static void perf_swevent_add(struct perf_event *event, u64 nr, +static void perf_swevent_event(struct perf_event *event, u64 nr, int nmi, struct perf_sample_data *data, struct pt_regs *regs) { @@ -4112,6 +4374,9 @@ static void perf_swevent_add(struct perf_event *event, u64 nr, static int perf_exclude_event(struct perf_event *event, struct pt_regs *regs) { + if (event->hw.state & PERF_HES_STOPPED) + return 0; + if (regs) { if (event->attr.exclude_user && user_mode(regs)) return 1; @@ -4158,11 +4423,11 @@ __find_swevent_head(struct swevent_hlist *hlist, u64 type, u32 event_id) /* For the read side: events when they trigger */ static inline struct hlist_head * -find_swevent_head_rcu(struct perf_cpu_context *ctx, u64 type, u32 event_id) +find_swevent_head_rcu(struct swevent_htable *swhash, u64 type, u32 event_id) { struct swevent_hlist *hlist; - hlist = rcu_dereference(ctx->swevent_hlist); + hlist = rcu_dereference(swhash->swevent_hlist); if (!hlist) return NULL; @@ -4171,7 +4436,7 @@ find_swevent_head_rcu(struct perf_cpu_context *ctx, u64 type, u32 event_id) /* For the event head insertion and removal in the hlist */ static inline struct hlist_head * -find_swevent_head(struct perf_cpu_context *ctx, struct perf_event *event) +find_swevent_head(struct swevent_htable *swhash, struct perf_event *event) { struct swevent_hlist *hlist; u32 event_id = event->attr.config; @@ -4182,7 +4447,7 @@ find_swevent_head(struct perf_cpu_context *ctx, struct perf_event *event) * and release. Which makes the protected version suitable here. * The context lock guarantees that. */ - hlist = rcu_dereference_protected(ctx->swevent_hlist, + hlist = rcu_dereference_protected(swhash->swevent_hlist, lockdep_is_held(&event->ctx->lock)); if (!hlist) return NULL; @@ -4195,23 +4460,19 @@ static void do_perf_sw_event(enum perf_type_id type, u32 event_id, struct perf_sample_data *data, struct pt_regs *regs) { - struct perf_cpu_context *cpuctx; + struct swevent_htable *swhash = &__get_cpu_var(swevent_htable); struct perf_event *event; struct hlist_node *node; struct hlist_head *head; - cpuctx = &__get_cpu_var(perf_cpu_context); - rcu_read_lock(); - - head = find_swevent_head_rcu(cpuctx, type, event_id); - + head = find_swevent_head_rcu(swhash, type, event_id); if (!head) goto end; hlist_for_each_entry_rcu(event, node, head, hlist_entry) { if (perf_swevent_match(event, type, event_id, data, regs)) - perf_swevent_add(event, nr, nmi, data, regs); + perf_swevent_event(event, nr, nmi, data, regs); } end: rcu_read_unlock(); @@ -4219,33 +4480,17 @@ end: int perf_swevent_get_recursion_context(void) { - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); - int rctx; - - if (in_nmi()) - rctx = 3; - else if (in_irq()) - rctx = 2; - else if (in_softirq()) - rctx = 1; - else - rctx = 0; - - if (cpuctx->recursion[rctx]) - return -1; + struct swevent_htable *swhash = &__get_cpu_var(swevent_htable); - cpuctx->recursion[rctx]++; - barrier(); - - return rctx; + return get_recursion_context(swhash->recursion); } EXPORT_SYMBOL_GPL(perf_swevent_get_recursion_context); void inline perf_swevent_put_recursion_context(int rctx) { - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); - barrier(); - cpuctx->recursion[rctx]--; + struct swevent_htable *swhash = &__get_cpu_var(swevent_htable); + + put_recursion_context(swhash->recursion, rctx); } void __perf_sw_event(u32 event_id, u64 nr, int nmi, @@ -4271,20 +4516,20 @@ static void perf_swevent_read(struct perf_event *event) { } -static int perf_swevent_enable(struct perf_event *event) +static int perf_swevent_add(struct perf_event *event, int flags) { + struct swevent_htable *swhash = &__get_cpu_var(swevent_htable); struct hw_perf_event *hwc = &event->hw; - struct perf_cpu_context *cpuctx; struct hlist_head *head; - cpuctx = &__get_cpu_var(perf_cpu_context); - if (hwc->sample_period) { hwc->last_period = hwc->sample_period; perf_swevent_set_period(event); } - head = find_swevent_head(cpuctx, event); + hwc->state = !(flags & PERF_EF_START); + + head = find_swevent_head(swhash, event); if (WARN_ON_ONCE(!head)) return -EINVAL; @@ -4293,202 +4538,27 @@ static int perf_swevent_enable(struct perf_event *event) return 0; } -static void perf_swevent_disable(struct perf_event *event) +static void perf_swevent_del(struct perf_event *event, int flags) { hlist_del_rcu(&event->hlist_entry); } -static void perf_swevent_void(struct perf_event *event) -{ -} - -static int perf_swevent_int(struct perf_event *event) -{ - return 0; -} - -static const struct pmu perf_ops_generic = { - .enable = perf_swevent_enable, - .disable = perf_swevent_disable, - .start = perf_swevent_int, - .stop = perf_swevent_void, - .read = perf_swevent_read, - .unthrottle = perf_swevent_void, /* hwc->interrupts already reset */ -}; - -/* - * hrtimer based swevent callback - */ - -static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer) +static void perf_swevent_start(struct perf_event *event, int flags) { - enum hrtimer_restart ret = HRTIMER_RESTART; - struct perf_sample_data data; - struct pt_regs *regs; - struct perf_event *event; - u64 period; - - event = container_of(hrtimer, struct perf_event, hw.hrtimer); - event->pmu->read(event); - - perf_sample_data_init(&data, 0); - data.period = event->hw.last_period; - regs = get_irq_regs(); - - if (regs && !perf_exclude_event(event, regs)) { - if (!(event->attr.exclude_idle && current->pid == 0)) - if (perf_event_overflow(event, 0, &data, regs)) - ret = HRTIMER_NORESTART; - } - - period = max_t(u64, 10000, event->hw.sample_period); - hrtimer_forward_now(hrtimer, ns_to_ktime(period)); - - return ret; + event->hw.state = 0; } -static void perf_swevent_start_hrtimer(struct perf_event *event) +static void perf_swevent_stop(struct perf_event *event, int flags) { - struct hw_perf_event *hwc = &event->hw; - - hrtimer_init(&hwc->hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); - hwc->hrtimer.function = perf_swevent_hrtimer; - if (hwc->sample_period) { - u64 period; - - if (hwc->remaining) { - if (hwc->remaining < 0) - period = 10000; - else - period = hwc->remaining; - hwc->remaining = 0; - } else { - period = max_t(u64, 10000, hwc->sample_period); - } - __hrtimer_start_range_ns(&hwc->hrtimer, - ns_to_ktime(period), 0, - HRTIMER_MODE_REL, 0); - } -} - -static void perf_swevent_cancel_hrtimer(struct perf_event *event) -{ - struct hw_perf_event *hwc = &event->hw; - - if (hwc->sample_period) { - ktime_t remaining = hrtimer_get_remaining(&hwc->hrtimer); - hwc->remaining = ktime_to_ns(remaining); - - hrtimer_cancel(&hwc->hrtimer); - } -} - -/* - * Software event: cpu wall time clock - */ - -static void cpu_clock_perf_event_update(struct perf_event *event) -{ - int cpu = raw_smp_processor_id(); - s64 prev; - u64 now; - - now = cpu_clock(cpu); - prev = local64_xchg(&event->hw.prev_count, now); - local64_add(now - prev, &event->count); -} - -static int cpu_clock_perf_event_enable(struct perf_event *event) -{ - struct hw_perf_event *hwc = &event->hw; - int cpu = raw_smp_processor_id(); - - local64_set(&hwc->prev_count, cpu_clock(cpu)); - perf_swevent_start_hrtimer(event); - - return 0; -} - -static void cpu_clock_perf_event_disable(struct perf_event *event) -{ - perf_swevent_cancel_hrtimer(event); - cpu_clock_perf_event_update(event); -} - -static void cpu_clock_perf_event_read(struct perf_event *event) -{ - cpu_clock_perf_event_update(event); -} - -static const struct pmu perf_ops_cpu_clock = { - .enable = cpu_clock_perf_event_enable, - .disable = cpu_clock_perf_event_disable, - .read = cpu_clock_perf_event_read, -}; - -/* - * Software event: task time clock - */ - -static void task_clock_perf_event_update(struct perf_event *event, u64 now) -{ - u64 prev; - s64 delta; - - prev = local64_xchg(&event->hw.prev_count, now); - delta = now - prev; - local64_add(delta, &event->count); -} - -static int task_clock_perf_event_enable(struct perf_event *event) -{ - struct hw_perf_event *hwc = &event->hw; - u64 now; - - now = event->ctx->time; - - local64_set(&hwc->prev_count, now); - - perf_swevent_start_hrtimer(event); - - return 0; -} - -static void task_clock_perf_event_disable(struct perf_event *event) -{ - perf_swevent_cancel_hrtimer(event); - task_clock_perf_event_update(event, event->ctx->time); - -} - -static void task_clock_perf_event_read(struct perf_event *event) -{ - u64 time; - - if (!in_nmi()) { - update_context_time(event->ctx); - time = event->ctx->time; - } else { - u64 now = perf_clock(); - u64 delta = now - event->ctx->timestamp; - time = event->ctx->time + delta; - } - - task_clock_perf_event_update(event, time); + event->hw.state = PERF_HES_STOPPED; } -static const struct pmu perf_ops_task_clock = { - .enable = task_clock_perf_event_enable, - .disable = task_clock_perf_event_disable, - .read = task_clock_perf_event_read, -}; - /* Deref the hlist from the update side */ static inline struct swevent_hlist * -swevent_hlist_deref(struct perf_cpu_context *cpuctx) +swevent_hlist_deref(struct swevent_htable *swhash) { - return rcu_dereference_protected(cpuctx->swevent_hlist, - lockdep_is_held(&cpuctx->hlist_mutex)); + return rcu_dereference_protected(swhash->swevent_hlist, + lockdep_is_held(&swhash->hlist_mutex)); } static void swevent_hlist_release_rcu(struct rcu_head *rcu_head) @@ -4499,27 +4569,27 @@ static void swevent_hlist_release_rcu(struct rcu_head *rcu_head) kfree(hlist); } -static void swevent_hlist_release(struct perf_cpu_context *cpuctx) +static void swevent_hlist_release(struct swevent_htable *swhash) { - struct swevent_hlist *hlist = swevent_hlist_deref(cpuctx); + struct swevent_hlist *hlist = swevent_hlist_deref(swhash); if (!hlist) return; - rcu_assign_pointer(cpuctx->swevent_hlist, NULL); + rcu_assign_pointer(swhash->swevent_hlist, NULL); call_rcu(&hlist->rcu_head, swevent_hlist_release_rcu); } static void swevent_hlist_put_cpu(struct perf_event *event, int cpu) { - struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct swevent_htable *swhash = &per_cpu(swevent_htable, cpu); - mutex_lock(&cpuctx->hlist_mutex); + mutex_lock(&swhash->hlist_mutex); - if (!--cpuctx->hlist_refcount) - swevent_hlist_release(cpuctx); + if (!--swhash->hlist_refcount) + swevent_hlist_release(swhash); - mutex_unlock(&cpuctx->hlist_mutex); + mutex_unlock(&swhash->hlist_mutex); } static void swevent_hlist_put(struct perf_event *event) @@ -4537,12 +4607,12 @@ static void swevent_hlist_put(struct perf_event *event) static int swevent_hlist_get_cpu(struct perf_event *event, int cpu) { - struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); + struct swevent_htable *swhash = &per_cpu(swevent_htable, cpu); int err = 0; - mutex_lock(&cpuctx->hlist_mutex); + mutex_lock(&swhash->hlist_mutex); - if (!swevent_hlist_deref(cpuctx) && cpu_online(cpu)) { + if (!swevent_hlist_deref(swhash) && cpu_online(cpu)) { struct swevent_hlist *hlist; hlist = kzalloc(sizeof(*hlist), GFP_KERNEL); @@ -4550,11 +4620,11 @@ static int swevent_hlist_get_cpu(struct perf_event *event, int cpu) err = -ENOMEM; goto exit; } - rcu_assign_pointer(cpuctx->swevent_hlist, hlist); + rcu_assign_pointer(swhash->swevent_hlist, hlist); } - cpuctx->hlist_refcount++; - exit: - mutex_unlock(&cpuctx->hlist_mutex); + swhash->hlist_refcount++; +exit: + mutex_unlock(&swhash->hlist_mutex); return err; } @@ -4578,7 +4648,7 @@ static int swevent_hlist_get(struct perf_event *event) put_online_cpus(); return 0; - fail: +fail: for_each_possible_cpu(cpu) { if (cpu == failed_cpu) break; @@ -4589,17 +4659,64 @@ static int swevent_hlist_get(struct perf_event *event) return err; } -#ifdef CONFIG_EVENT_TRACING +atomic_t perf_swevent_enabled[PERF_COUNT_SW_MAX]; + +static void sw_perf_event_destroy(struct perf_event *event) +{ + u64 event_id = event->attr.config; + + WARN_ON(event->parent); + + jump_label_dec(&perf_swevent_enabled[event_id]); + swevent_hlist_put(event); +} + +static int perf_swevent_init(struct perf_event *event) +{ + int event_id = event->attr.config; + + if (event->attr.type != PERF_TYPE_SOFTWARE) + return -ENOENT; + + switch (event_id) { + case PERF_COUNT_SW_CPU_CLOCK: + case PERF_COUNT_SW_TASK_CLOCK: + return -ENOENT; -static const struct pmu perf_ops_tracepoint = { - .enable = perf_trace_enable, - .disable = perf_trace_disable, - .start = perf_swevent_int, - .stop = perf_swevent_void, + default: + break; + } + + if (event_id > PERF_COUNT_SW_MAX) + return -ENOENT; + + if (!event->parent) { + int err; + + err = swevent_hlist_get(event); + if (err) + return err; + + jump_label_inc(&perf_swevent_enabled[event_id]); + event->destroy = sw_perf_event_destroy; + } + + return 0; +} + +static struct pmu perf_swevent = { + .task_ctx_nr = perf_sw_context, + + .event_init = perf_swevent_init, + .add = perf_swevent_add, + .del = perf_swevent_del, + .start = perf_swevent_start, + .stop = perf_swevent_stop, .read = perf_swevent_read, - .unthrottle = perf_swevent_void, }; +#ifdef CONFIG_EVENT_TRACING + static int perf_tp_filter_match(struct perf_event *event, struct perf_sample_data *data) { @@ -4643,7 +4760,7 @@ void perf_tp_event(u64 addr, u64 count, void *record, int entry_size, hlist_for_each_entry_rcu(event, node, head, hlist_entry) { if (perf_tp_event_match(event, &data, regs)) - perf_swevent_add(event, count, 1, &data, regs); + perf_swevent_event(event, count, 1, &data, regs); } perf_swevent_put_recursion_context(rctx); @@ -4655,10 +4772,13 @@ static void tp_perf_event_destroy(struct perf_event *event) perf_trace_destroy(event); } -static const struct pmu *tp_perf_event_init(struct perf_event *event) +static int perf_tp_event_init(struct perf_event *event) { int err; + if (event->attr.type != PERF_TYPE_TRACEPOINT) + return -ENOENT; + /* * Raw tracepoint data is a severe data leak, only allow root to * have these. @@ -4666,15 +4786,31 @@ static const struct pmu *tp_perf_event_init(struct perf_event *event) if ((event->attr.sample_type & PERF_SAMPLE_RAW) && perf_paranoid_tracepoint_raw() && !capable(CAP_SYS_ADMIN)) - return ERR_PTR(-EPERM); + return -EPERM; err = perf_trace_init(event); if (err) - return NULL; + return err; event->destroy = tp_perf_event_destroy; - return &perf_ops_tracepoint; + return 0; +} + +static struct pmu perf_tracepoint = { + .task_ctx_nr = perf_sw_context, + + .event_init = perf_tp_event_init, + .add = perf_trace_add, + .del = perf_trace_del, + .start = perf_swevent_start, + .stop = perf_swevent_stop, + .read = perf_swevent_read, +}; + +static inline void perf_tp_register(void) +{ + perf_pmu_register(&perf_tracepoint); } static int perf_event_set_filter(struct perf_event *event, void __user *arg) @@ -4702,9 +4838,8 @@ static void perf_event_free_filter(struct perf_event *event) #else -static const struct pmu *tp_perf_event_init(struct perf_event *event) +static inline void perf_tp_register(void) { - return NULL; } static int perf_event_set_filter(struct perf_event *event, void __user *arg) @@ -4719,105 +4854,389 @@ static void perf_event_free_filter(struct perf_event *event) #endif /* CONFIG_EVENT_TRACING */ #ifdef CONFIG_HAVE_HW_BREAKPOINT -static void bp_perf_event_destroy(struct perf_event *event) +void perf_bp_event(struct perf_event *bp, void *data) { - release_bp_slot(event); + struct perf_sample_data sample; + struct pt_regs *regs = data; + + perf_sample_data_init(&sample, bp->attr.bp_addr); + + if (!bp->hw.state && !perf_exclude_event(bp, regs)) + perf_swevent_event(bp, 1, 1, &sample, regs); } +#endif -static const struct pmu *bp_perf_event_init(struct perf_event *bp) +/* + * hrtimer based swevent callback + */ + +static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer) { - int err; + enum hrtimer_restart ret = HRTIMER_RESTART; + struct perf_sample_data data; + struct pt_regs *regs; + struct perf_event *event; + u64 period; - err = register_perf_hw_breakpoint(bp); - if (err) - return ERR_PTR(err); + event = container_of(hrtimer, struct perf_event, hw.hrtimer); + event->pmu->read(event); + + perf_sample_data_init(&data, 0); + data.period = event->hw.last_period; + regs = get_irq_regs(); + + if (regs && !perf_exclude_event(event, regs)) { + if (!(event->attr.exclude_idle && current->pid == 0)) + if (perf_event_overflow(event, 0, &data, regs)) + ret = HRTIMER_NORESTART; + } + + period = max_t(u64, 10000, event->hw.sample_period); + hrtimer_forward_now(hrtimer, ns_to_ktime(period)); - bp->destroy = bp_perf_event_destroy; + return ret; +} - return &perf_ops_bp; +static void perf_swevent_start_hrtimer(struct perf_event *event) +{ + struct hw_perf_event *hwc = &event->hw; + + hrtimer_init(&hwc->hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + hwc->hrtimer.function = perf_swevent_hrtimer; + if (hwc->sample_period) { + s64 period = local64_read(&hwc->period_left); + + if (period) { + if (period < 0) + period = 10000; + + local64_set(&hwc->period_left, 0); + } else { + period = max_t(u64, 10000, hwc->sample_period); + } + __hrtimer_start_range_ns(&hwc->hrtimer, + ns_to_ktime(period), 0, + HRTIMER_MODE_REL_PINNED, 0); + } } -void perf_bp_event(struct perf_event *bp, void *data) +static void perf_swevent_cancel_hrtimer(struct perf_event *event) { - struct perf_sample_data sample; - struct pt_regs *regs = data; + struct hw_perf_event *hwc = &event->hw; - perf_sample_data_init(&sample, bp->attr.bp_addr); + if (hwc->sample_period) { + ktime_t remaining = hrtimer_get_remaining(&hwc->hrtimer); + local64_set(&hwc->period_left, ktime_to_ns(remaining)); - if (!perf_exclude_event(bp, regs)) - perf_swevent_add(bp, 1, 1, &sample, regs); + hrtimer_cancel(&hwc->hrtimer); + } } -#else -static const struct pmu *bp_perf_event_init(struct perf_event *bp) + +/* + * Software event: cpu wall time clock + */ + +static void cpu_clock_event_update(struct perf_event *event) { - return NULL; + s64 prev; + u64 now; + + now = local_clock(); + prev = local64_xchg(&event->hw.prev_count, now); + local64_add(now - prev, &event->count); } -void perf_bp_event(struct perf_event *bp, void *regs) +static void cpu_clock_event_start(struct perf_event *event, int flags) { + local64_set(&event->hw.prev_count, local_clock()); + perf_swevent_start_hrtimer(event); } -#endif -atomic_t perf_swevent_enabled[PERF_COUNT_SW_MAX]; +static void cpu_clock_event_stop(struct perf_event *event, int flags) +{ + perf_swevent_cancel_hrtimer(event); + cpu_clock_event_update(event); +} -static void sw_perf_event_destroy(struct perf_event *event) +static int cpu_clock_event_add(struct perf_event *event, int flags) { - u64 event_id = event->attr.config; + if (flags & PERF_EF_START) + cpu_clock_event_start(event, flags); - WARN_ON(event->parent); + return 0; +} - atomic_dec(&perf_swevent_enabled[event_id]); - swevent_hlist_put(event); +static void cpu_clock_event_del(struct perf_event *event, int flags) +{ + cpu_clock_event_stop(event, flags); } -static const struct pmu *sw_perf_event_init(struct perf_event *event) +static void cpu_clock_event_read(struct perf_event *event) { - const struct pmu *pmu = NULL; - u64 event_id = event->attr.config; + cpu_clock_event_update(event); +} + +static int cpu_clock_event_init(struct perf_event *event) +{ + if (event->attr.type != PERF_TYPE_SOFTWARE) + return -ENOENT; + + if (event->attr.config != PERF_COUNT_SW_CPU_CLOCK) + return -ENOENT; + + return 0; +} +static struct pmu perf_cpu_clock = { + .task_ctx_nr = perf_sw_context, + + .event_init = cpu_clock_event_init, + .add = cpu_clock_event_add, + .del = cpu_clock_event_del, + .start = cpu_clock_event_start, + .stop = cpu_clock_event_stop, + .read = cpu_clock_event_read, +}; + +/* + * Software event: task time clock + */ + +static void task_clock_event_update(struct perf_event *event, u64 now) +{ + u64 prev; + s64 delta; + + prev = local64_xchg(&event->hw.prev_count, now); + delta = now - prev; + local64_add(delta, &event->count); +} + +static void task_clock_event_start(struct perf_event *event, int flags) +{ + local64_set(&event->hw.prev_count, event->ctx->time); + perf_swevent_start_hrtimer(event); +} + +static void task_clock_event_stop(struct perf_event *event, int flags) +{ + perf_swevent_cancel_hrtimer(event); + task_clock_event_update(event, event->ctx->time); +} + +static int task_clock_event_add(struct perf_event *event, int flags) +{ + if (flags & PERF_EF_START) + task_clock_event_start(event, flags); + + return 0; +} + +static void task_clock_event_del(struct perf_event *event, int flags) +{ + task_clock_event_stop(event, PERF_EF_UPDATE); +} + +static void task_clock_event_read(struct perf_event *event) +{ + u64 time; + + if (!in_nmi()) { + update_context_time(event->ctx); + time = event->ctx->time; + } else { + u64 now = perf_clock(); + u64 delta = now - event->ctx->timestamp; + time = event->ctx->time + delta; + } + + task_clock_event_update(event, time); +} + +static int task_clock_event_init(struct perf_event *event) +{ + if (event->attr.type != PERF_TYPE_SOFTWARE) + return -ENOENT; + + if (event->attr.config != PERF_COUNT_SW_TASK_CLOCK) + return -ENOENT; + + return 0; +} + +static struct pmu perf_task_clock = { + .task_ctx_nr = perf_sw_context, + + .event_init = task_clock_event_init, + .add = task_clock_event_add, + .del = task_clock_event_del, + .start = task_clock_event_start, + .stop = task_clock_event_stop, + .read = task_clock_event_read, +}; + +static void perf_pmu_nop_void(struct pmu *pmu) +{ +} + +static int perf_pmu_nop_int(struct pmu *pmu) +{ + return 0; +} + +static void perf_pmu_start_txn(struct pmu *pmu) +{ + perf_pmu_disable(pmu); +} + +static int perf_pmu_commit_txn(struct pmu *pmu) +{ + perf_pmu_enable(pmu); + return 0; +} + +static void perf_pmu_cancel_txn(struct pmu *pmu) +{ + perf_pmu_enable(pmu); +} + +/* + * Ensures all contexts with the same task_ctx_nr have the same + * pmu_cpu_context too. + */ +static void *find_pmu_context(int ctxn) +{ + struct pmu *pmu; + + if (ctxn < 0) + return NULL; + + list_for_each_entry(pmu, &pmus, entry) { + if (pmu->task_ctx_nr == ctxn) + return pmu->pmu_cpu_context; + } + + return NULL; +} + +static void free_pmu_context(void * __percpu cpu_context) +{ + struct pmu *pmu; + + mutex_lock(&pmus_lock); /* - * Software events (currently) can't in general distinguish - * between user, kernel and hypervisor events. - * However, context switches and cpu migrations are considered - * to be kernel events, and page faults are never hypervisor - * events. + * Like a real lame refcount. */ - switch (event_id) { - case PERF_COUNT_SW_CPU_CLOCK: - pmu = &perf_ops_cpu_clock; + list_for_each_entry(pmu, &pmus, entry) { + if (pmu->pmu_cpu_context == cpu_context) + goto out; + } - break; - case PERF_COUNT_SW_TASK_CLOCK: - /* - * If the user instantiates this as a per-cpu event, - * use the cpu_clock event instead. - */ - if (event->ctx->task) - pmu = &perf_ops_task_clock; - else - pmu = &perf_ops_cpu_clock; + free_percpu(cpu_context); +out: + mutex_unlock(&pmus_lock); +} - break; - case PERF_COUNT_SW_PAGE_FAULTS: - case PERF_COUNT_SW_PAGE_FAULTS_MIN: - case PERF_COUNT_SW_PAGE_FAULTS_MAJ: - case PERF_COUNT_SW_CONTEXT_SWITCHES: - case PERF_COUNT_SW_CPU_MIGRATIONS: - case PERF_COUNT_SW_ALIGNMENT_FAULTS: - case PERF_COUNT_SW_EMULATION_FAULTS: - if (!event->parent) { - int err; - - err = swevent_hlist_get(event); - if (err) - return ERR_PTR(err); +int perf_pmu_register(struct pmu *pmu) +{ + int cpu, ret; + + mutex_lock(&pmus_lock); + ret = -ENOMEM; + pmu->pmu_disable_count = alloc_percpu(int); + if (!pmu->pmu_disable_count) + goto unlock; - atomic_inc(&perf_swevent_enabled[event_id]); - event->destroy = sw_perf_event_destroy; + pmu->pmu_cpu_context = find_pmu_context(pmu->task_ctx_nr); + if (pmu->pmu_cpu_context) + goto got_cpu_context; + + pmu->pmu_cpu_context = alloc_percpu(struct perf_cpu_context); + if (!pmu->pmu_cpu_context) + goto free_pdc; + + for_each_possible_cpu(cpu) { + struct perf_cpu_context *cpuctx; + + cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu); + __perf_event_init_context(&cpuctx->ctx); + cpuctx->ctx.type = cpu_context; + cpuctx->ctx.pmu = pmu; + cpuctx->jiffies_interval = 1; + INIT_LIST_HEAD(&cpuctx->rotation_list); + } + +got_cpu_context: + if (!pmu->start_txn) { + if (pmu->pmu_enable) { + /* + * If we have pmu_enable/pmu_disable calls, install + * transaction stubs that use that to try and batch + * hardware accesses. + */ + pmu->start_txn = perf_pmu_start_txn; + pmu->commit_txn = perf_pmu_commit_txn; + pmu->cancel_txn = perf_pmu_cancel_txn; + } else { + pmu->start_txn = perf_pmu_nop_void; + pmu->commit_txn = perf_pmu_nop_int; + pmu->cancel_txn = perf_pmu_nop_void; + } + } + + if (!pmu->pmu_enable) { + pmu->pmu_enable = perf_pmu_nop_void; + pmu->pmu_disable = perf_pmu_nop_void; + } + + list_add_rcu(&pmu->entry, &pmus); + ret = 0; +unlock: + mutex_unlock(&pmus_lock); + + return ret; + +free_pdc: + free_percpu(pmu->pmu_disable_count); + goto unlock; +} + +void perf_pmu_unregister(struct pmu *pmu) +{ + mutex_lock(&pmus_lock); + list_del_rcu(&pmu->entry); + mutex_unlock(&pmus_lock); + + /* + * We dereference the pmu list under both SRCU and regular RCU, so + * synchronize against both of those. + */ + synchronize_srcu(&pmus_srcu); + synchronize_rcu(); + + free_percpu(pmu->pmu_disable_count); + free_pmu_context(pmu->pmu_cpu_context); +} + +struct pmu *perf_init_event(struct perf_event *event) +{ + struct pmu *pmu = NULL; + int idx; + + idx = srcu_read_lock(&pmus_srcu); + list_for_each_entry_rcu(pmu, &pmus, entry) { + int ret = pmu->event_init(event); + if (!ret) + goto unlock; + + if (ret != -ENOENT) { + pmu = ERR_PTR(ret); + goto unlock; } - pmu = &perf_ops_generic; - break; } + pmu = ERR_PTR(-ENOENT); +unlock: + srcu_read_unlock(&pmus_srcu, idx); return pmu; } @@ -4826,20 +5245,18 @@ static const struct pmu *sw_perf_event_init(struct perf_event *event) * Allocate and initialize a event structure */ static struct perf_event * -perf_event_alloc(struct perf_event_attr *attr, - int cpu, - struct perf_event_context *ctx, - struct perf_event *group_leader, - struct perf_event *parent_event, - perf_overflow_handler_t overflow_handler, - gfp_t gfpflags) -{ - const struct pmu *pmu; +perf_event_alloc(struct perf_event_attr *attr, int cpu, + struct task_struct *task, + struct perf_event *group_leader, + struct perf_event *parent_event, + perf_overflow_handler_t overflow_handler) +{ + struct pmu *pmu; struct perf_event *event; struct hw_perf_event *hwc; long err; - event = kzalloc(sizeof(*event), gfpflags); + event = kzalloc(sizeof(*event), GFP_KERNEL); if (!event) return ERR_PTR(-ENOMEM); @@ -4857,6 +5274,7 @@ perf_event_alloc(struct perf_event_attr *attr, INIT_LIST_HEAD(&event->event_entry); INIT_LIST_HEAD(&event->sibling_list); init_waitqueue_head(&event->waitq); + init_irq_work(&event->pending, perf_pending_event); mutex_init(&event->mmap_mutex); @@ -4864,7 +5282,6 @@ perf_event_alloc(struct perf_event_attr *attr, event->attr = *attr; event->group_leader = group_leader; event->pmu = NULL; - event->ctx = ctx; event->oncpu = -1; event->parent = parent_event; @@ -4874,6 +5291,17 @@ perf_event_alloc(struct perf_event_attr *attr, event->state = PERF_EVENT_STATE_INACTIVE; + if (task) { + event->attach_state = PERF_ATTACH_TASK; +#ifdef CONFIG_HAVE_HW_BREAKPOINT + /* + * hw_breakpoint is a bit difficult here.. + */ + if (attr->type == PERF_TYPE_BREAKPOINT) + event->hw.bp_target = task; +#endif + } + if (!overflow_handler && parent_event) overflow_handler = parent_event->overflow_handler; @@ -4898,29 +5326,8 @@ perf_event_alloc(struct perf_event_attr *attr, if (attr->inherit && (attr->read_format & PERF_FORMAT_GROUP)) goto done; - switch (attr->type) { - case PERF_TYPE_RAW: - case PERF_TYPE_HARDWARE: - case PERF_TYPE_HW_CACHE: - pmu = hw_perf_event_init(event); - break; - - case PERF_TYPE_SOFTWARE: - pmu = sw_perf_event_init(event); - break; - - case PERF_TYPE_TRACEPOINT: - pmu = tp_perf_event_init(event); - break; + pmu = perf_init_event(event); - case PERF_TYPE_BREAKPOINT: - pmu = bp_perf_event_init(event); - break; - - - default: - break; - } done: err = 0; if (!pmu) @@ -4938,13 +5345,21 @@ done: event->pmu = pmu; if (!event->parent) { - atomic_inc(&nr_events); + if (event->attach_state & PERF_ATTACH_TASK) + jump_label_inc(&perf_task_events); if (event->attr.mmap || event->attr.mmap_data) atomic_inc(&nr_mmap_events); if (event->attr.comm) atomic_inc(&nr_comm_events); if (event->attr.task) atomic_inc(&nr_task_events); + if (event->attr.sample_type & PERF_SAMPLE_CALLCHAIN) { + err = get_callchain_buffers(); + if (err) { + free_event(event); + return ERR_PTR(err); + } + } } return event; @@ -5092,12 +5507,16 @@ SYSCALL_DEFINE5(perf_event_open, struct perf_event_attr __user *, attr_uptr, pid_t, pid, int, cpu, int, group_fd, unsigned long, flags) { - struct perf_event *event, *group_leader = NULL, *output_event = NULL; + struct perf_event *group_leader = NULL, *output_event = NULL; + struct perf_event *event, *sibling; struct perf_event_attr attr; struct perf_event_context *ctx; struct file *event_file = NULL; struct file *group_file = NULL; + struct task_struct *task = NULL; + struct pmu *pmu; int event_fd; + int move_group = 0; int fput_needed = 0; int err; @@ -5123,20 +5542,11 @@ SYSCALL_DEFINE5(perf_event_open, if (event_fd < 0) return event_fd; - /* - * Get the target context (task or percpu): - */ - ctx = find_get_context(pid, cpu); - if (IS_ERR(ctx)) { - err = PTR_ERR(ctx); - goto err_fd; - } - if (group_fd != -1) { group_leader = perf_fget_light(group_fd, &fput_needed); if (IS_ERR(group_leader)) { err = PTR_ERR(group_leader); - goto err_put_context; + goto err_fd; } group_file = group_leader->filp; if (flags & PERF_FLAG_FD_OUTPUT) @@ -5145,6 +5555,58 @@ SYSCALL_DEFINE5(perf_event_open, group_leader = NULL; } + if (pid != -1) { + task = find_lively_task_by_vpid(pid); + if (IS_ERR(task)) { + err = PTR_ERR(task); + goto err_group_fd; + } + } + + event = perf_event_alloc(&attr, cpu, task, group_leader, NULL, NULL); + if (IS_ERR(event)) { + err = PTR_ERR(event); + goto err_task; + } + + /* + * Special case software events and allow them to be part of + * any hardware group. + */ + pmu = event->pmu; + + if (group_leader && + (is_software_event(event) != is_software_event(group_leader))) { + if (is_software_event(event)) { + /* + * If event and group_leader are not both a software + * event, and event is, then group leader is not. + * + * Allow the addition of software events to !software + * groups, this is safe because software events never + * fail to schedule. + */ + pmu = group_leader->pmu; + } else if (is_software_event(group_leader) && + (group_leader->group_flags & PERF_GROUP_SOFTWARE)) { + /* + * In case the group is a pure software group, and we + * try to add a hardware event, move the whole group to + * the hardware context. + */ + move_group = 1; + } + } + + /* + * Get the target context (task or percpu): + */ + ctx = find_get_context(pmu, task, cpu); + if (IS_ERR(ctx)) { + err = PTR_ERR(ctx); + goto err_alloc; + } + /* * Look up the group leader (we will attach this event to it): */ @@ -5156,42 +5618,66 @@ SYSCALL_DEFINE5(perf_event_open, * becoming part of another group-sibling): */ if (group_leader->group_leader != group_leader) - goto err_put_context; + goto err_context; /* * Do not allow to attach to a group in a different * task or CPU context: */ - if (group_leader->ctx != ctx) - goto err_put_context; + if (move_group) { + if (group_leader->ctx->type != ctx->type) + goto err_context; + } else { + if (group_leader->ctx != ctx) + goto err_context; + } + /* * Only a group leader can be exclusive or pinned */ if (attr.exclusive || attr.pinned) - goto err_put_context; - } - - event = perf_event_alloc(&attr, cpu, ctx, group_leader, - NULL, NULL, GFP_KERNEL); - if (IS_ERR(event)) { - err = PTR_ERR(event); - goto err_put_context; + goto err_context; } if (output_event) { err = perf_event_set_output(event, output_event); if (err) - goto err_free_put_context; + goto err_context; } event_file = anon_inode_getfile("[perf_event]", &perf_fops, event, O_RDWR); if (IS_ERR(event_file)) { err = PTR_ERR(event_file); - goto err_free_put_context; + goto err_context; + } + + if (move_group) { + struct perf_event_context *gctx = group_leader->ctx; + + mutex_lock(&gctx->mutex); + perf_event_remove_from_context(group_leader); + list_for_each_entry(sibling, &group_leader->sibling_list, + group_entry) { + perf_event_remove_from_context(sibling); + put_ctx(gctx); + } + mutex_unlock(&gctx->mutex); + put_ctx(gctx); } event->filp = event_file; WARN_ON_ONCE(ctx->parent_ctx); mutex_lock(&ctx->mutex); + + if (move_group) { + perf_install_in_context(ctx, group_leader, cpu); + get_ctx(ctx); + list_for_each_entry(sibling, &group_leader->sibling_list, + group_entry) { + perf_install_in_context(ctx, sibling, cpu); + get_ctx(ctx); + } + } + perf_install_in_context(ctx, event, cpu); ++ctx->generation; mutex_unlock(&ctx->mutex); @@ -5212,11 +5698,15 @@ SYSCALL_DEFINE5(perf_event_open, fd_install(event_fd, event_file); return event_fd; -err_free_put_context: +err_context: + put_ctx(ctx); +err_alloc: free_event(event); -err_put_context: +err_task: + if (task) + put_task_struct(task); +err_group_fd: fput_light(group_file, fput_needed); - put_ctx(ctx); err_fd: put_unused_fd(event_fd); return err; @@ -5227,32 +5717,31 @@ err_fd: * * @attr: attributes of the counter to create * @cpu: cpu in which the counter is bound - * @pid: task to profile + * @task: task to profile (NULL for percpu) */ struct perf_event * perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, - pid_t pid, + struct task_struct *task, perf_overflow_handler_t overflow_handler) { - struct perf_event *event; struct perf_event_context *ctx; + struct perf_event *event; 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, overflow_handler, GFP_KERNEL); + event = perf_event_alloc(attr, cpu, task, NULL, NULL, overflow_handler); if (IS_ERR(event)) { err = PTR_ERR(event); - goto err_put_context; + goto err; + } + + ctx = find_get_context(event->pmu, task, cpu); + if (IS_ERR(ctx)) { + err = PTR_ERR(ctx); + goto err_free; } event->filp = NULL; @@ -5270,112 +5759,13 @@ perf_event_create_kernel_counter(struct perf_event_attr *attr, int cpu, return event; - err_put_context: - put_ctx(ctx); - err_exit: +err_free: + free_event(event); +err: return ERR_PTR(err); } EXPORT_SYMBOL_GPL(perf_event_create_kernel_counter); -/* - * inherit a event from parent task to child task: - */ -static struct perf_event * -inherit_event(struct perf_event *parent_event, - struct task_struct *parent, - struct perf_event_context *parent_ctx, - struct task_struct *child, - struct perf_event *group_leader, - struct perf_event_context *child_ctx) -{ - struct perf_event *child_event; - - /* - * Instead of creating recursive hierarchies of events, - * we link inherited events back to the original parent, - * which has a filp for sure, which we use as the reference - * count: - */ - if (parent_event->parent) - parent_event = parent_event->parent; - - child_event = perf_event_alloc(&parent_event->attr, - parent_event->cpu, child_ctx, - group_leader, parent_event, - NULL, GFP_KERNEL); - if (IS_ERR(child_event)) - return child_event; - get_ctx(child_ctx); - - /* - * Make the child state follow the state of the parent event, - * not its attr.disabled bit. We hold the parent's mutex, - * so we won't race with perf_event_{en, dis}able_family. - */ - if (parent_event->state >= PERF_EVENT_STATE_INACTIVE) - child_event->state = PERF_EVENT_STATE_INACTIVE; - else - child_event->state = PERF_EVENT_STATE_OFF; - - if (parent_event->attr.freq) { - u64 sample_period = parent_event->hw.sample_period; - struct hw_perf_event *hwc = &child_event->hw; - - hwc->sample_period = sample_period; - hwc->last_period = sample_period; - - local64_set(&hwc->period_left, sample_period); - } - - child_event->overflow_handler = parent_event->overflow_handler; - - /* - * Link it up in the child's context: - */ - add_event_to_ctx(child_event, child_ctx); - - /* - * Get a reference to the parent filp - we will fput it - * when the child event exits. This is safe to do because - * we are in the parent and we know that the filp still - * exists and has a nonzero count: - */ - atomic_long_inc(&parent_event->filp->f_count); - - /* - * Link this into the parent event's child list - */ - WARN_ON_ONCE(parent_event->ctx->parent_ctx); - mutex_lock(&parent_event->child_mutex); - list_add_tail(&child_event->child_list, &parent_event->child_list); - mutex_unlock(&parent_event->child_mutex); - - return child_event; -} - -static int inherit_group(struct perf_event *parent_event, - struct task_struct *parent, - struct perf_event_context *parent_ctx, - struct task_struct *child, - struct perf_event_context *child_ctx) -{ - struct perf_event *leader; - struct perf_event *sub; - struct perf_event *child_ctr; - - leader = inherit_event(parent_event, parent, parent_ctx, - child, NULL, child_ctx); - if (IS_ERR(leader)) - return PTR_ERR(leader); - list_for_each_entry(sub, &parent_event->sibling_list, group_entry) { - child_ctr = inherit_event(sub, parent, parent_ctx, - child, leader, child_ctx); - if (IS_ERR(child_ctr)) - return PTR_ERR(child_ctr); - } - return 0; -} - static void sync_child_event(struct perf_event *child_event, struct task_struct *child) { @@ -5432,16 +5822,13 @@ __perf_event_exit_task(struct perf_event *child_event, } } -/* - * When a child task exits, feed back event values to parent events. - */ -void perf_event_exit_task(struct task_struct *child) +static void perf_event_exit_task_context(struct task_struct *child, int ctxn) { struct perf_event *child_event, *tmp; struct perf_event_context *child_ctx; unsigned long flags; - if (likely(!child->perf_event_ctxp)) { + if (likely(!child->perf_event_ctxp[ctxn])) { perf_event_task(child, NULL, 0); return; } @@ -5453,8 +5840,8 @@ void perf_event_exit_task(struct task_struct *child) * scheduled, so we are now safe from rescheduling changing * our context. */ - child_ctx = child->perf_event_ctxp; - __perf_event_task_sched_out(child_ctx); + child_ctx = child->perf_event_ctxp[ctxn]; + task_ctx_sched_out(child_ctx, EVENT_ALL); /* * Take the context lock here so that if find_get_context is @@ -5462,7 +5849,7 @@ void perf_event_exit_task(struct task_struct *child) * incremented the context's refcount before we do put_ctx below. */ raw_spin_lock(&child_ctx->lock); - child->perf_event_ctxp = NULL; + child->perf_event_ctxp[ctxn] = NULL; /* * If this context is a clone; unclone it so it can't get * swapped to another process while we're removing all @@ -5515,6 +5902,17 @@ again: put_ctx(child_ctx); } +/* + * When a child task exits, feed back event values to parent events. + */ +void perf_event_exit_task(struct task_struct *child) +{ + int ctxn; + + for_each_task_context_nr(ctxn) + perf_event_exit_task_context(child, ctxn); +} + static void perf_free_event(struct perf_event *event, struct perf_event_context *ctx) { @@ -5536,48 +5934,166 @@ static void perf_free_event(struct perf_event *event, /* * free an unexposed, unused context as created by inheritance by - * init_task below, used by fork() in case of fail. + * perf_event_init_task below, used by fork() in case of fail. */ void perf_event_free_task(struct task_struct *task) { - struct perf_event_context *ctx = task->perf_event_ctxp; + struct perf_event_context *ctx; struct perf_event *event, *tmp; + int ctxn; - if (!ctx) - return; + for_each_task_context_nr(ctxn) { + ctx = task->perf_event_ctxp[ctxn]; + if (!ctx) + continue; - mutex_lock(&ctx->mutex); + mutex_lock(&ctx->mutex); again: - list_for_each_entry_safe(event, tmp, &ctx->pinned_groups, group_entry) - perf_free_event(event, ctx); + list_for_each_entry_safe(event, tmp, &ctx->pinned_groups, + group_entry) + perf_free_event(event, ctx); - list_for_each_entry_safe(event, tmp, &ctx->flexible_groups, - group_entry) - perf_free_event(event, ctx); + list_for_each_entry_safe(event, tmp, &ctx->flexible_groups, + group_entry) + perf_free_event(event, ctx); - if (!list_empty(&ctx->pinned_groups) || - !list_empty(&ctx->flexible_groups)) - goto again; + if (!list_empty(&ctx->pinned_groups) || + !list_empty(&ctx->flexible_groups)) + goto again; - mutex_unlock(&ctx->mutex); + mutex_unlock(&ctx->mutex); - put_ctx(ctx); + put_ctx(ctx); + } +} + +void perf_event_delayed_put(struct task_struct *task) +{ + int ctxn; + + for_each_task_context_nr(ctxn) + WARN_ON_ONCE(task->perf_event_ctxp[ctxn]); +} + +/* + * inherit a event from parent task to child task: + */ +static struct perf_event * +inherit_event(struct perf_event *parent_event, + struct task_struct *parent, + struct perf_event_context *parent_ctx, + struct task_struct *child, + struct perf_event *group_leader, + struct perf_event_context *child_ctx) +{ + struct perf_event *child_event; + unsigned long flags; + + /* + * Instead of creating recursive hierarchies of events, + * we link inherited events back to the original parent, + * which has a filp for sure, which we use as the reference + * count: + */ + if (parent_event->parent) + parent_event = parent_event->parent; + + child_event = perf_event_alloc(&parent_event->attr, + parent_event->cpu, + child, + group_leader, parent_event, + NULL); + if (IS_ERR(child_event)) + return child_event; + get_ctx(child_ctx); + + /* + * Make the child state follow the state of the parent event, + * not its attr.disabled bit. We hold the parent's mutex, + * so we won't race with perf_event_{en, dis}able_family. + */ + if (parent_event->state >= PERF_EVENT_STATE_INACTIVE) + child_event->state = PERF_EVENT_STATE_INACTIVE; + else + child_event->state = PERF_EVENT_STATE_OFF; + + if (parent_event->attr.freq) { + u64 sample_period = parent_event->hw.sample_period; + struct hw_perf_event *hwc = &child_event->hw; + + hwc->sample_period = sample_period; + hwc->last_period = sample_period; + + local64_set(&hwc->period_left, sample_period); + } + + child_event->ctx = child_ctx; + child_event->overflow_handler = parent_event->overflow_handler; + + /* + * Link it up in the child's context: + */ + raw_spin_lock_irqsave(&child_ctx->lock, flags); + add_event_to_ctx(child_event, child_ctx); + raw_spin_unlock_irqrestore(&child_ctx->lock, flags); + + /* + * Get a reference to the parent filp - we will fput it + * when the child event exits. This is safe to do because + * we are in the parent and we know that the filp still + * exists and has a nonzero count: + */ + atomic_long_inc(&parent_event->filp->f_count); + + /* + * Link this into the parent event's child list + */ + WARN_ON_ONCE(parent_event->ctx->parent_ctx); + mutex_lock(&parent_event->child_mutex); + list_add_tail(&child_event->child_list, &parent_event->child_list); + mutex_unlock(&parent_event->child_mutex); + + return child_event; +} + +static int inherit_group(struct perf_event *parent_event, + struct task_struct *parent, + struct perf_event_context *parent_ctx, + struct task_struct *child, + struct perf_event_context *child_ctx) +{ + struct perf_event *leader; + struct perf_event *sub; + struct perf_event *child_ctr; + + leader = inherit_event(parent_event, parent, parent_ctx, + child, NULL, child_ctx); + if (IS_ERR(leader)) + return PTR_ERR(leader); + list_for_each_entry(sub, &parent_event->sibling_list, group_entry) { + child_ctr = inherit_event(sub, parent, parent_ctx, + child, leader, child_ctx); + if (IS_ERR(child_ctr)) + return PTR_ERR(child_ctr); + } + return 0; } static int inherit_task_group(struct perf_event *event, struct task_struct *parent, struct perf_event_context *parent_ctx, - struct task_struct *child, + struct task_struct *child, int ctxn, int *inherited_all) { int ret; - struct perf_event_context *child_ctx = child->perf_event_ctxp; + struct perf_event_context *child_ctx; if (!event->attr.inherit) { *inherited_all = 0; return 0; } + child_ctx = child->perf_event_ctxp[ctxn]; if (!child_ctx) { /* * This is executed from the parent task context, so @@ -5586,14 +6102,11 @@ inherit_task_group(struct perf_event *event, struct task_struct *parent, * child. */ - child_ctx = kzalloc(sizeof(struct perf_event_context), - GFP_KERNEL); + child_ctx = alloc_perf_context(event->pmu, child); if (!child_ctx) return -ENOMEM; - __perf_event_init_context(child_ctx, child); - child->perf_event_ctxp = child_ctx; - get_task_struct(child); + child->perf_event_ctxp[ctxn] = child_ctx; } ret = inherit_group(event, parent, parent_ctx, @@ -5605,11 +6118,10 @@ inherit_task_group(struct perf_event *event, struct task_struct *parent, return ret; } - /* * Initialize the perf_event context in task_struct */ -int perf_event_init_task(struct task_struct *child) +int perf_event_init_context(struct task_struct *child, int ctxn) { struct perf_event_context *child_ctx, *parent_ctx; struct perf_event_context *cloned_ctx; @@ -5618,19 +6130,19 @@ int perf_event_init_task(struct task_struct *child) int inherited_all = 1; int ret = 0; - child->perf_event_ctxp = NULL; + child->perf_event_ctxp[ctxn] = NULL; mutex_init(&child->perf_event_mutex); INIT_LIST_HEAD(&child->perf_event_list); - if (likely(!parent->perf_event_ctxp)) + if (likely(!parent->perf_event_ctxp[ctxn])) return 0; /* * If the parent's context is a clone, pin it so it won't get * swapped under us. */ - parent_ctx = perf_pin_task_context(parent); + parent_ctx = perf_pin_task_context(parent, ctxn); /* * No need to check if parent_ctx != NULL here; since we saw @@ -5650,20 +6162,20 @@ int perf_event_init_task(struct task_struct *child) * the list, not manipulating it: */ list_for_each_entry(event, &parent_ctx->pinned_groups, group_entry) { - ret = inherit_task_group(event, parent, parent_ctx, child, - &inherited_all); + ret = inherit_task_group(event, parent, parent_ctx, + child, ctxn, &inherited_all); if (ret) break; } list_for_each_entry(event, &parent_ctx->flexible_groups, group_entry) { - ret = inherit_task_group(event, parent, parent_ctx, child, - &inherited_all); + ret = inherit_task_group(event, parent, parent_ctx, + child, ctxn, &inherited_all); if (ret) break; } - child_ctx = child->perf_event_ctxp; + child_ctx = child->perf_event_ctxp[ctxn]; if (child_ctx && inherited_all) { /* @@ -5692,63 +6204,98 @@ int perf_event_init_task(struct task_struct *child) return ret; } +/* + * Initialize the perf_event context in task_struct + */ +int perf_event_init_task(struct task_struct *child) +{ + int ctxn, ret; + + for_each_task_context_nr(ctxn) { + ret = perf_event_init_context(child, ctxn); + if (ret) + return ret; + } + + return 0; +} + static void __init perf_event_init_all_cpus(void) { + struct swevent_htable *swhash; int cpu; - struct perf_cpu_context *cpuctx; for_each_possible_cpu(cpu) { - cpuctx = &per_cpu(perf_cpu_context, cpu); - mutex_init(&cpuctx->hlist_mutex); - __perf_event_init_context(&cpuctx->ctx, NULL); + swhash = &per_cpu(swevent_htable, cpu); + mutex_init(&swhash->hlist_mutex); + INIT_LIST_HEAD(&per_cpu(rotation_list, cpu)); } } static void __cpuinit perf_event_init_cpu(int cpu) { - struct perf_cpu_context *cpuctx; - - cpuctx = &per_cpu(perf_cpu_context, cpu); + struct swevent_htable *swhash = &per_cpu(swevent_htable, cpu); - spin_lock(&perf_resource_lock); - cpuctx->max_pertask = perf_max_events - perf_reserved_percpu; - spin_unlock(&perf_resource_lock); - - mutex_lock(&cpuctx->hlist_mutex); - if (cpuctx->hlist_refcount > 0) { + mutex_lock(&swhash->hlist_mutex); + if (swhash->hlist_refcount > 0) { struct swevent_hlist *hlist; - hlist = kzalloc(sizeof(*hlist), GFP_KERNEL); - WARN_ON_ONCE(!hlist); - rcu_assign_pointer(cpuctx->swevent_hlist, hlist); + hlist = kzalloc_node(sizeof(*hlist), GFP_KERNEL, cpu_to_node(cpu)); + WARN_ON(!hlist); + rcu_assign_pointer(swhash->swevent_hlist, hlist); } - mutex_unlock(&cpuctx->hlist_mutex); + mutex_unlock(&swhash->hlist_mutex); } #ifdef CONFIG_HOTPLUG_CPU -static void __perf_event_exit_cpu(void *info) +static void perf_pmu_rotate_stop(struct pmu *pmu) { - struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); - struct perf_event_context *ctx = &cpuctx->ctx; + struct perf_cpu_context *cpuctx = this_cpu_ptr(pmu->pmu_cpu_context); + + WARN_ON(!irqs_disabled()); + + list_del_init(&cpuctx->rotation_list); +} + +static void __perf_event_exit_context(void *__info) +{ + struct perf_event_context *ctx = __info; struct perf_event *event, *tmp; + perf_pmu_rotate_stop(ctx->pmu); + list_for_each_entry_safe(event, tmp, &ctx->pinned_groups, group_entry) __perf_event_remove_from_context(event); list_for_each_entry_safe(event, tmp, &ctx->flexible_groups, group_entry) __perf_event_remove_from_context(event); } + +static void perf_event_exit_cpu_context(int cpu) +{ + struct perf_event_context *ctx; + struct pmu *pmu; + int idx; + + idx = srcu_read_lock(&pmus_srcu); + list_for_each_entry_rcu(pmu, &pmus, entry) { + ctx = &per_cpu_ptr(pmu->pmu_cpu_context, cpu)->ctx; + + mutex_lock(&ctx->mutex); + smp_call_function_single(cpu, __perf_event_exit_context, ctx, 1); + mutex_unlock(&ctx->mutex); + } + srcu_read_unlock(&pmus_srcu, idx); +} + static void perf_event_exit_cpu(int cpu) { - struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); - struct perf_event_context *ctx = &cpuctx->ctx; + struct swevent_htable *swhash = &per_cpu(swevent_htable, cpu); - mutex_lock(&cpuctx->hlist_mutex); - swevent_hlist_release(cpuctx); - mutex_unlock(&cpuctx->hlist_mutex); + mutex_lock(&swhash->hlist_mutex); + swevent_hlist_release(swhash); + mutex_unlock(&swhash->hlist_mutex); - mutex_lock(&ctx->mutex); - smp_call_function_single(cpu, __perf_event_exit_cpu, NULL, 1); - mutex_unlock(&ctx->mutex); + perf_event_exit_cpu_context(cpu); } #else static inline void perf_event_exit_cpu(int cpu) { } @@ -5778,118 +6325,13 @@ perf_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) return NOTIFY_OK; } -/* - * This has to have a higher priority than migration_notifier in sched.c. - */ -static struct notifier_block __cpuinitdata perf_cpu_nb = { - .notifier_call = perf_cpu_notify, - .priority = 20, -}; - void __init perf_event_init(void) { perf_event_init_all_cpus(); - perf_cpu_notify(&perf_cpu_nb, (unsigned long)CPU_UP_PREPARE, - (void *)(long)smp_processor_id()); - perf_cpu_notify(&perf_cpu_nb, (unsigned long)CPU_ONLINE, - (void *)(long)smp_processor_id()); - register_cpu_notifier(&perf_cpu_nb); -} - -static ssize_t perf_show_reserve_percpu(struct sysdev_class *class, - struct sysdev_class_attribute *attr, - char *buf) -{ - return sprintf(buf, "%d\n", perf_reserved_percpu); -} - -static ssize_t -perf_set_reserve_percpu(struct sysdev_class *class, - struct sysdev_class_attribute *attr, - const char *buf, - size_t count) -{ - struct perf_cpu_context *cpuctx; - unsigned long val; - int err, cpu, mpt; - - err = strict_strtoul(buf, 10, &val); - if (err) - return err; - if (val > perf_max_events) - return -EINVAL; - - spin_lock(&perf_resource_lock); - perf_reserved_percpu = val; - for_each_online_cpu(cpu) { - cpuctx = &per_cpu(perf_cpu_context, cpu); - raw_spin_lock_irq(&cpuctx->ctx.lock); - mpt = min(perf_max_events - cpuctx->ctx.nr_events, - perf_max_events - perf_reserved_percpu); - cpuctx->max_pertask = mpt; - raw_spin_unlock_irq(&cpuctx->ctx.lock); - } - spin_unlock(&perf_resource_lock); - - return count; -} - -static ssize_t perf_show_overcommit(struct sysdev_class *class, - struct sysdev_class_attribute *attr, - char *buf) -{ - return sprintf(buf, "%d\n", perf_overcommit); -} - -static ssize_t -perf_set_overcommit(struct sysdev_class *class, - struct sysdev_class_attribute *attr, - const char *buf, size_t count) -{ - unsigned long val; - int err; - - err = strict_strtoul(buf, 10, &val); - if (err) - return err; - if (val > 1) - return -EINVAL; - - spin_lock(&perf_resource_lock); - perf_overcommit = val; - spin_unlock(&perf_resource_lock); - - return count; -} - -static SYSDEV_CLASS_ATTR( - reserve_percpu, - 0644, - perf_show_reserve_percpu, - perf_set_reserve_percpu - ); - -static SYSDEV_CLASS_ATTR( - overcommit, - 0644, - perf_show_overcommit, - perf_set_overcommit - ); - -static struct attribute *perfclass_attrs[] = { - &attr_reserve_percpu.attr, - &attr_overcommit.attr, - NULL -}; - -static struct attribute_group perfclass_attr_group = { - .attrs = perfclass_attrs, - .name = "perf_events", -}; - -static int __init perf_event_sysfs_init(void) -{ - return sysfs_create_group(&cpu_sysdev_class.kset.kobj, - &perfclass_attr_group); + init_srcu_struct(&pmus_srcu); + perf_pmu_register(&perf_swevent); + perf_pmu_register(&perf_cpu_clock); + perf_pmu_register(&perf_task_clock); + perf_tp_register(); + perf_cpu_notifier(perf_cpu_notify); } -device_initcall(perf_event_sysfs_init); diff --git a/kernel/sched.c b/kernel/sched.c index 0bba34a..5a5cc33 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -3584,7 +3584,7 @@ void scheduler_tick(void) curr->sched_class->task_tick(rq, curr, 0); raw_spin_unlock(&rq->lock); - perf_event_task_tick(curr); + perf_event_task_tick(); #ifdef CONFIG_SMP rq->idle_at_tick = idle_cpu(cpu); diff --git a/kernel/test_kprobes.c b/kernel/test_kprobes.c index 4f10451..f8b11a2 100644 --- a/kernel/test_kprobes.c +++ b/kernel/test_kprobes.c @@ -115,7 +115,9 @@ static int test_kprobes(void) int ret; struct kprobe *kps[2] = {&kp, &kp2}; - kp.addr = 0; /* addr should be cleard for reusing kprobe. */ + /* addr and flags should be cleard for reusing kprobe. */ + kp.addr = NULL; + kp.flags = 0; ret = register_kprobes(kps, 2); if (ret < 0) { printk(KERN_ERR "Kprobe smoke test failed: " @@ -210,7 +212,9 @@ static int test_jprobes(void) int ret; struct jprobe *jps[2] = {&jp, &jp2}; - jp.kp.addr = 0; /* addr should be cleard for reusing kprobe. */ + /* addr and flags should be cleard for reusing kprobe. */ + jp.kp.addr = NULL; + jp.kp.flags = 0; ret = register_jprobes(jps, 2); if (ret < 0) { printk(KERN_ERR "Kprobe smoke test failed: " @@ -323,7 +327,9 @@ static int test_kretprobes(void) int ret; struct kretprobe *rps[2] = {&rp, &rp2}; - rp.kp.addr = 0; /* addr should be cleard for reusing kprobe. */ + /* addr and flags should be cleard for reusing kprobe. */ + rp.kp.addr = NULL; + rp.kp.flags = 0; ret = register_kretprobes(rps, 2); if (ret < 0) { printk(KERN_ERR "Kprobe smoke test failed: " diff --git a/kernel/timer.c b/kernel/timer.c index 97bf05b..68a9ae7 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -37,7 +37,7 @@ #include <linux/delay.h> #include <linux/tick.h> #include <linux/kallsyms.h> -#include <linux/perf_event.h> +#include <linux/irq_work.h> #include <linux/sched.h> #include <linux/slab.h> @@ -1279,7 +1279,10 @@ void update_process_times(int user_tick) run_local_timers(); rcu_check_callbacks(cpu, user_tick); printk_tick(); - perf_event_do_pending(); +#ifdef CONFIG_IRQ_WORK + if (in_irq()) + irq_work_run(); +#endif scheduler_tick(); run_posix_cpu_timers(p); } diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 538501c..e550d2ed 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -49,6 +49,11 @@ config HAVE_SYSCALL_TRACEPOINTS help See Documentation/trace/ftrace-design.txt +config HAVE_C_RECORDMCOUNT + bool + help + C version of recordmcount available? + config TRACER_MAX_TRACE bool diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index fa7ece6..ebd80d5 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -884,10 +884,8 @@ enum { FTRACE_ENABLE_CALLS = (1 << 0), FTRACE_DISABLE_CALLS = (1 << 1), FTRACE_UPDATE_TRACE_FUNC = (1 << 2), - FTRACE_ENABLE_MCOUNT = (1 << 3), - FTRACE_DISABLE_MCOUNT = (1 << 4), - FTRACE_START_FUNC_RET = (1 << 5), - FTRACE_STOP_FUNC_RET = (1 << 6), + FTRACE_START_FUNC_RET = (1 << 3), + FTRACE_STOP_FUNC_RET = (1 << 4), }; static int ftrace_filtered; @@ -1226,8 +1224,6 @@ static void ftrace_shutdown(int command) static void ftrace_startup_sysctl(void) { - int command = FTRACE_ENABLE_MCOUNT; - if (unlikely(ftrace_disabled)) return; @@ -1235,23 +1231,17 @@ static void ftrace_startup_sysctl(void) saved_ftrace_func = NULL; /* ftrace_start_up is true if we want ftrace running */ if (ftrace_start_up) - command |= FTRACE_ENABLE_CALLS; - - ftrace_run_update_code(command); + ftrace_run_update_code(FTRACE_ENABLE_CALLS); } static void ftrace_shutdown_sysctl(void) { - int command = FTRACE_DISABLE_MCOUNT; - if (unlikely(ftrace_disabled)) return; /* ftrace_start_up is true if ftrace is running */ if (ftrace_start_up) - command |= FTRACE_DISABLE_CALLS; - - ftrace_run_update_code(command); + ftrace_run_update_code(FTRACE_DISABLE_CALLS); } static cycle_t ftrace_update_time; @@ -1368,24 +1358,29 @@ enum { #define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */ struct ftrace_iterator { - struct ftrace_page *pg; - int hidx; - int idx; - unsigned flags; - struct trace_parser parser; + loff_t pos; + loff_t func_pos; + struct ftrace_page *pg; + struct dyn_ftrace *func; + struct ftrace_func_probe *probe; + struct trace_parser parser; + int hidx; + int idx; + unsigned flags; }; static void * -t_hash_next(struct seq_file *m, void *v, loff_t *pos) +t_hash_next(struct seq_file *m, loff_t *pos) { struct ftrace_iterator *iter = m->private; - struct hlist_node *hnd = v; + struct hlist_node *hnd = NULL; struct hlist_head *hhd; - WARN_ON(!(iter->flags & FTRACE_ITER_HASH)); - (*pos)++; + iter->pos = *pos; + if (iter->probe) + hnd = &iter->probe->node; retry: if (iter->hidx >= FTRACE_FUNC_HASHSIZE) return NULL; @@ -1408,7 +1403,12 @@ t_hash_next(struct seq_file *m, void *v, loff_t *pos) } } - return hnd; + if (WARN_ON_ONCE(!hnd)) + return NULL; + + iter->probe = hlist_entry(hnd, struct ftrace_func_probe, node); + + return iter; } static void *t_hash_start(struct seq_file *m, loff_t *pos) @@ -1417,26 +1417,32 @@ static void *t_hash_start(struct seq_file *m, loff_t *pos) void *p = NULL; loff_t l; - if (!(iter->flags & FTRACE_ITER_HASH)) - *pos = 0; - - iter->flags |= FTRACE_ITER_HASH; + if (iter->func_pos > *pos) + return NULL; iter->hidx = 0; - for (l = 0; l <= *pos; ) { - p = t_hash_next(m, p, &l); + for (l = 0; l <= (*pos - iter->func_pos); ) { + p = t_hash_next(m, &l); if (!p) break; } - return p; + if (!p) + return NULL; + + /* Only set this if we have an item */ + iter->flags |= FTRACE_ITER_HASH; + + return iter; } -static int t_hash_show(struct seq_file *m, void *v) +static int +t_hash_show(struct seq_file *m, struct ftrace_iterator *iter) { struct ftrace_func_probe *rec; - struct hlist_node *hnd = v; - rec = hlist_entry(hnd, struct ftrace_func_probe, node); + rec = iter->probe; + if (WARN_ON_ONCE(!rec)) + return -EIO; if (rec->ops->print) return rec->ops->print(m, rec->ip, rec->ops, rec->data); @@ -1457,12 +1463,13 @@ t_next(struct seq_file *m, void *v, loff_t *pos) struct dyn_ftrace *rec = NULL; if (iter->flags & FTRACE_ITER_HASH) - return t_hash_next(m, v, pos); + return t_hash_next(m, pos); (*pos)++; + iter->pos = *pos; if (iter->flags & FTRACE_ITER_PRINTALL) - return NULL; + return t_hash_start(m, pos); retry: if (iter->idx >= iter->pg->index) { @@ -1491,7 +1498,20 @@ t_next(struct seq_file *m, void *v, loff_t *pos) } } - return rec; + if (!rec) + return t_hash_start(m, pos); + + iter->func_pos = *pos; + iter->func = rec; + + return iter; +} + +static void reset_iter_read(struct ftrace_iterator *iter) +{ + iter->pos = 0; + iter->func_pos = 0; + iter->flags &= ~(FTRACE_ITER_PRINTALL & FTRACE_ITER_HASH); } static void *t_start(struct seq_file *m, loff_t *pos) @@ -1502,6 +1522,12 @@ static void *t_start(struct seq_file *m, loff_t *pos) mutex_lock(&ftrace_lock); /* + * If an lseek was done, then reset and start from beginning. + */ + if (*pos < iter->pos) + reset_iter_read(iter); + + /* * For set_ftrace_filter reading, if we have the filter * off, we can short cut and just print out that all * functions are enabled. @@ -1518,6 +1544,11 @@ static void *t_start(struct seq_file *m, loff_t *pos) if (iter->flags & FTRACE_ITER_HASH) return t_hash_start(m, pos); + /* + * Unfortunately, we need to restart at ftrace_pages_start + * every time we let go of the ftrace_mutex. This is because + * those pointers can change without the lock. + */ iter->pg = ftrace_pages_start; iter->idx = 0; for (l = 0; l <= *pos; ) { @@ -1526,10 +1557,14 @@ static void *t_start(struct seq_file *m, loff_t *pos) break; } - if (!p && iter->flags & FTRACE_ITER_FILTER) - return t_hash_start(m, pos); + if (!p) { + if (iter->flags & FTRACE_ITER_FILTER) + return t_hash_start(m, pos); - return p; + return NULL; + } + + return iter; } static void t_stop(struct seq_file *m, void *p) @@ -1540,16 +1575,18 @@ static void t_stop(struct seq_file *m, void *p) static int t_show(struct seq_file *m, void *v) { struct ftrace_iterator *iter = m->private; - struct dyn_ftrace *rec = v; + struct dyn_ftrace *rec; if (iter->flags & FTRACE_ITER_HASH) - return t_hash_show(m, v); + return t_hash_show(m, iter); if (iter->flags & FTRACE_ITER_PRINTALL) { seq_printf(m, "#### all functions enabled ####\n"); return 0; } + rec = iter->func; + if (!rec) return 0; @@ -1601,8 +1638,8 @@ ftrace_failures_open(struct inode *inode, struct file *file) ret = ftrace_avail_open(inode, file); if (!ret) { - m = (struct seq_file *)file->private_data; - iter = (struct ftrace_iterator *)m->private; + m = file->private_data; + iter = m->private; iter->flags = FTRACE_ITER_FAILURES; } @@ -2418,7 +2455,7 @@ static const struct file_operations ftrace_filter_fops = { .open = ftrace_filter_open, .read = seq_read, .write = ftrace_filter_write, - .llseek = no_llseek, + .llseek = ftrace_regex_lseek, .release = ftrace_filter_release, }; diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index bca9637..c5a632a 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2606,6 +2606,19 @@ void ring_buffer_record_enable_cpu(struct ring_buffer *buffer, int cpu) } EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu); +/* + * The total entries in the ring buffer is the running counter + * of entries entered into the ring buffer, minus the sum of + * the entries read from the ring buffer and the number of + * entries that were overwritten. + */ +static inline unsigned long +rb_num_of_entries(struct ring_buffer_per_cpu *cpu_buffer) +{ + return local_read(&cpu_buffer->entries) - + (local_read(&cpu_buffer->overrun) + cpu_buffer->read); +} + /** * ring_buffer_entries_cpu - get the number of entries in a cpu buffer * @buffer: The ring buffer @@ -2614,16 +2627,13 @@ EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu); unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; - unsigned long ret; if (!cpumask_test_cpu(cpu, buffer->cpumask)) return 0; cpu_buffer = buffer->buffers[cpu]; - ret = (local_read(&cpu_buffer->entries) - local_read(&cpu_buffer->overrun)) - - cpu_buffer->read; - return ret; + return rb_num_of_entries(cpu_buffer); } EXPORT_SYMBOL_GPL(ring_buffer_entries_cpu); @@ -2684,8 +2694,7 @@ unsigned long ring_buffer_entries(struct ring_buffer *buffer) /* if you care about this being correct, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; - entries += (local_read(&cpu_buffer->entries) - - local_read(&cpu_buffer->overrun)) - cpu_buffer->read; + entries += rb_num_of_entries(cpu_buffer); } return entries; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 9ec59f5..001bcd2 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2196,7 +2196,7 @@ int tracing_open_generic(struct inode *inode, struct file *filp) static int tracing_release(struct inode *inode, struct file *file) { - struct seq_file *m = (struct seq_file *)file->private_data; + struct seq_file *m = file->private_data; struct trace_iterator *iter; int cpu; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index d39b3c5..9021f8c 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -343,6 +343,10 @@ void trace_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc); +void trace_graph_function(struct trace_array *tr, + unsigned long ip, + unsigned long parent_ip, + unsigned long flags, int pc); void trace_default_header(struct seq_file *m); void print_trace_header(struct seq_file *m, struct trace_iterator *iter); int trace_empty(struct trace_iterator *iter); diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c index 31cc4cb..39c059c 100644 --- a/kernel/trace/trace_event_perf.c +++ b/kernel/trace/trace_event_perf.c @@ -9,7 +9,7 @@ #include <linux/kprobes.h> #include "trace.h" -static char *perf_trace_buf[4]; +static char __percpu *perf_trace_buf[PERF_NR_CONTEXTS]; /* * Force it to be aligned to unsigned long to avoid misaligned accesses @@ -24,7 +24,7 @@ static int total_ref_count; static int perf_trace_event_init(struct ftrace_event_call *tp_event, struct perf_event *p_event) { - struct hlist_head *list; + struct hlist_head __percpu *list; int ret = -ENOMEM; int cpu; @@ -42,11 +42,11 @@ static int perf_trace_event_init(struct ftrace_event_call *tp_event, tp_event->perf_events = list; if (!total_ref_count) { - char *buf; + char __percpu *buf; int i; - for (i = 0; i < 4; i++) { - buf = (char *)alloc_percpu(perf_trace_t); + for (i = 0; i < PERF_NR_CONTEXTS; i++) { + buf = (char __percpu *)alloc_percpu(perf_trace_t); if (!buf) goto fail; @@ -65,7 +65,7 @@ fail: if (!total_ref_count) { int i; - for (i = 0; i < 4; i++) { + for (i = 0; i < PERF_NR_CONTEXTS; i++) { free_percpu(perf_trace_buf[i]); perf_trace_buf[i] = NULL; } @@ -101,22 +101,26 @@ int perf_trace_init(struct perf_event *p_event) return ret; } -int perf_trace_enable(struct perf_event *p_event) +int perf_trace_add(struct perf_event *p_event, int flags) { struct ftrace_event_call *tp_event = p_event->tp_event; + struct hlist_head __percpu *pcpu_list; struct hlist_head *list; - list = tp_event->perf_events; - if (WARN_ON_ONCE(!list)) + pcpu_list = tp_event->perf_events; + if (WARN_ON_ONCE(!pcpu_list)) return -EINVAL; - list = this_cpu_ptr(list); + if (!(flags & PERF_EF_START)) + p_event->hw.state = PERF_HES_STOPPED; + + list = this_cpu_ptr(pcpu_list); hlist_add_head_rcu(&p_event->hlist_entry, list); return 0; } -void perf_trace_disable(struct perf_event *p_event) +void perf_trace_del(struct perf_event *p_event, int flags) { hlist_del_rcu(&p_event->hlist_entry); } @@ -142,7 +146,7 @@ void perf_trace_destroy(struct perf_event *p_event) tp_event->perf_events = NULL; if (!--total_ref_count) { - for (i = 0; i < 4; i++) { + for (i = 0; i < PERF_NR_CONTEXTS; i++) { free_percpu(perf_trace_buf[i]); perf_trace_buf[i] = NULL; } diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 4c758f1..398c0e8 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -600,21 +600,29 @@ out: enum { FORMAT_HEADER = 1, - FORMAT_PRINTFMT = 2, + FORMAT_FIELD_SEPERATOR = 2, + FORMAT_PRINTFMT = 3, }; static void *f_next(struct seq_file *m, void *v, loff_t *pos) { struct ftrace_event_call *call = m->private; struct ftrace_event_field *field; - struct list_head *head; + struct list_head *common_head = &ftrace_common_fields; + struct list_head *head = trace_get_fields(call); (*pos)++; switch ((unsigned long)v) { case FORMAT_HEADER: - head = &ftrace_common_fields; + if (unlikely(list_empty(common_head))) + return NULL; + + field = list_entry(common_head->prev, + struct ftrace_event_field, link); + return field; + case FORMAT_FIELD_SEPERATOR: if (unlikely(list_empty(head))) return NULL; @@ -626,31 +634,10 @@ static void *f_next(struct seq_file *m, void *v, loff_t *pos) return NULL; } - head = trace_get_fields(call); - - /* - * To separate common fields from event fields, the - * LSB is set on the first event field. Clear it in case. - */ - v = (void *)((unsigned long)v & ~1L); - field = v; - /* - * If this is a common field, and at the end of the list, then - * continue with main list. - */ - if (field->link.prev == &ftrace_common_fields) { - if (unlikely(list_empty(head))) - return NULL; - field = list_entry(head->prev, struct ftrace_event_field, link); - /* Set the LSB to notify f_show to print an extra newline */ - field = (struct ftrace_event_field *) - ((unsigned long)field | 1); - return field; - } - - /* If we are done tell f_show to print the format */ - if (field->link.prev == head) + if (field->link.prev == common_head) + return (void *)FORMAT_FIELD_SEPERATOR; + else if (field->link.prev == head) return (void *)FORMAT_PRINTFMT; field = list_entry(field->link.prev, struct ftrace_event_field, link); @@ -688,22 +675,16 @@ static int f_show(struct seq_file *m, void *v) seq_printf(m, "format:\n"); return 0; + case FORMAT_FIELD_SEPERATOR: + seq_putc(m, '\n'); + return 0; + case FORMAT_PRINTFMT: seq_printf(m, "\nprint fmt: %s\n", call->print_fmt); return 0; } - /* - * To separate common fields from event fields, the - * LSB is set on the first event field. Clear it and - * print a newline if it is set. - */ - if ((unsigned long)v & 1) { - seq_putc(m, '\n'); - v = (void *)((unsigned long)v & ~1L); - } - field = v; /* diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 6f23369..76b0598 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -15,15 +15,19 @@ #include "trace.h" #include "trace_output.h" +/* When set, irq functions will be ignored */ +static int ftrace_graph_skip_irqs; + struct fgraph_cpu_data { pid_t last_pid; int depth; + int depth_irq; int ignore; unsigned long enter_funcs[FTRACE_RETFUNC_DEPTH]; }; struct fgraph_data { - struct fgraph_cpu_data *cpu_data; + struct fgraph_cpu_data __percpu *cpu_data; /* Place to preserve last processed entry. */ struct ftrace_graph_ent_entry ent; @@ -41,6 +45,7 @@ struct fgraph_data { #define TRACE_GRAPH_PRINT_PROC 0x8 #define TRACE_GRAPH_PRINT_DURATION 0x10 #define TRACE_GRAPH_PRINT_ABS_TIME 0x20 +#define TRACE_GRAPH_PRINT_IRQS 0x40 static struct tracer_opt trace_opts[] = { /* Display overruns? (for self-debug purpose) */ @@ -55,13 +60,15 @@ static struct tracer_opt trace_opts[] = { { TRACER_OPT(funcgraph-duration, TRACE_GRAPH_PRINT_DURATION) }, /* Display absolute time of an entry */ { TRACER_OPT(funcgraph-abstime, TRACE_GRAPH_PRINT_ABS_TIME) }, + /* Display interrupts */ + { TRACER_OPT(funcgraph-irqs, TRACE_GRAPH_PRINT_IRQS) }, { } /* Empty entry */ }; static struct tracer_flags tracer_flags = { /* Don't display overruns and proc by default */ .val = TRACE_GRAPH_PRINT_CPU | TRACE_GRAPH_PRINT_OVERHEAD | - TRACE_GRAPH_PRINT_DURATION, + TRACE_GRAPH_PRINT_DURATION | TRACE_GRAPH_PRINT_IRQS, .opts = trace_opts }; @@ -204,6 +211,14 @@ int __trace_graph_entry(struct trace_array *tr, return 1; } +static inline int ftrace_graph_ignore_irqs(void) +{ + if (!ftrace_graph_skip_irqs) + return 0; + + return in_irq(); +} + int trace_graph_entry(struct ftrace_graph_ent *trace) { struct trace_array *tr = graph_array; @@ -218,7 +233,8 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) return 0; /* trace it when it is-nested-in or is a function enabled. */ - if (!(trace->depth || ftrace_graph_addr(trace->func))) + if (!(trace->depth || ftrace_graph_addr(trace->func)) || + ftrace_graph_ignore_irqs()) return 0; local_irq_save(flags); @@ -246,6 +262,34 @@ int trace_graph_thresh_entry(struct ftrace_graph_ent *trace) return trace_graph_entry(trace); } +static void +__trace_graph_function(struct trace_array *tr, + unsigned long ip, unsigned long flags, int pc) +{ + u64 time = trace_clock_local(); + struct ftrace_graph_ent ent = { + .func = ip, + .depth = 0, + }; + struct ftrace_graph_ret ret = { + .func = ip, + .depth = 0, + .calltime = time, + .rettime = time, + }; + + __trace_graph_entry(tr, &ent, flags, pc); + __trace_graph_return(tr, &ret, flags, pc); +} + +void +trace_graph_function(struct trace_array *tr, + unsigned long ip, unsigned long parent_ip, + unsigned long flags, int pc) +{ + __trace_graph_function(tr, ip, flags, pc); +} + void __trace_graph_return(struct trace_array *tr, struct ftrace_graph_ret *trace, unsigned long flags, @@ -649,8 +693,9 @@ trace_print_graph_duration(unsigned long long duration, struct trace_seq *s) /* Print nsecs (we don't want to exceed 7 numbers) */ if (len < 7) { - snprintf(nsecs_str, min(sizeof(nsecs_str), 8UL - len), "%03lu", - nsecs_rem); + size_t slen = min_t(size_t, sizeof(nsecs_str), 8UL - len); + + snprintf(nsecs_str, slen, "%03lu", nsecs_rem); ret = trace_seq_printf(s, ".%s", nsecs_str); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -855,6 +900,108 @@ print_graph_prologue(struct trace_iterator *iter, struct trace_seq *s, return 0; } +/* + * Entry check for irq code + * + * returns 1 if + * - we are inside irq code + * - we just extered irq code + * + * retunns 0 if + * - funcgraph-interrupts option is set + * - we are not inside irq code + */ +static int +check_irq_entry(struct trace_iterator *iter, u32 flags, + unsigned long addr, int depth) +{ + int cpu = iter->cpu; + int *depth_irq; + struct fgraph_data *data = iter->private; + + /* + * If we are either displaying irqs, or we got called as + * a graph event and private data does not exist, + * then we bypass the irq check. + */ + if ((flags & TRACE_GRAPH_PRINT_IRQS) || + (!data)) + return 0; + + depth_irq = &(per_cpu_ptr(data->cpu_data, cpu)->depth_irq); + + /* + * We are inside the irq code + */ + if (*depth_irq >= 0) + return 1; + + if ((addr < (unsigned long)__irqentry_text_start) || + (addr >= (unsigned long)__irqentry_text_end)) + return 0; + + /* + * We are entering irq code. + */ + *depth_irq = depth; + return 1; +} + +/* + * Return check for irq code + * + * returns 1 if + * - we are inside irq code + * - we just left irq code + * + * returns 0 if + * - funcgraph-interrupts option is set + * - we are not inside irq code + */ +static int +check_irq_return(struct trace_iterator *iter, u32 flags, int depth) +{ + int cpu = iter->cpu; + int *depth_irq; + struct fgraph_data *data = iter->private; + + /* + * If we are either displaying irqs, or we got called as + * a graph event and private data does not exist, + * then we bypass the irq check. + */ + if ((flags & TRACE_GRAPH_PRINT_IRQS) || + (!data)) + return 0; + + depth_irq = &(per_cpu_ptr(data->cpu_data, cpu)->depth_irq); + + /* + * We are not inside the irq code. + */ + if (*depth_irq == -1) + return 0; + + /* + * We are inside the irq code, and this is returning entry. + * Let's not trace it and clear the entry depth, since + * we are out of irq code. + * + * This condition ensures that we 'leave the irq code' once + * we are out of the entry depth. Thus protecting us from + * the RETURN entry loss. + */ + if (*depth_irq >= depth) { + *depth_irq = -1; + return 1; + } + + /* + * We are inside the irq code, and this is not the entry. + */ + return 1; +} + static enum print_line_t print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, struct trace_iterator *iter, u32 flags) @@ -865,6 +1012,9 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, static enum print_line_t ret; int cpu = iter->cpu; + if (check_irq_entry(iter, flags, call->func, call->depth)) + return TRACE_TYPE_HANDLED; + if (print_graph_prologue(iter, s, TRACE_GRAPH_ENT, call->func, flags)) return TRACE_TYPE_PARTIAL_LINE; @@ -902,6 +1052,9 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, int ret; int i; + if (check_irq_return(iter, flags, trace->depth)) + return TRACE_TYPE_HANDLED; + if (data) { struct fgraph_cpu_data *cpu_data; int cpu = iter->cpu; @@ -1054,7 +1207,7 @@ print_graph_comment(struct trace_seq *s, struct trace_entry *ent, enum print_line_t -print_graph_function_flags(struct trace_iterator *iter, u32 flags) +__print_graph_function_flags(struct trace_iterator *iter, u32 flags) { struct ftrace_graph_ent_entry *field; struct fgraph_data *data = iter->private; @@ -1117,7 +1270,18 @@ print_graph_function_flags(struct trace_iterator *iter, u32 flags) static enum print_line_t print_graph_function(struct trace_iterator *iter) { - return print_graph_function_flags(iter, tracer_flags.val); + return __print_graph_function_flags(iter, tracer_flags.val); +} + +enum print_line_t print_graph_function_flags(struct trace_iterator *iter, + u32 flags) +{ + if (trace_flags & TRACE_ITER_LATENCY_FMT) + flags |= TRACE_GRAPH_PRINT_DURATION; + else + flags |= TRACE_GRAPH_PRINT_ABS_TIME; + + return __print_graph_function_flags(iter, flags); } static enum print_line_t @@ -1149,7 +1313,7 @@ static void print_lat_header(struct seq_file *s, u32 flags) seq_printf(s, "#%.*s|||| / \n", size, spaces); } -void print_graph_headers_flags(struct seq_file *s, u32 flags) +static void __print_graph_headers_flags(struct seq_file *s, u32 flags) { int lat = trace_flags & TRACE_ITER_LATENCY_FMT; @@ -1190,6 +1354,23 @@ void print_graph_headers(struct seq_file *s) print_graph_headers_flags(s, tracer_flags.val); } +void print_graph_headers_flags(struct seq_file *s, u32 flags) +{ + struct trace_iterator *iter = s->private; + + if (trace_flags & TRACE_ITER_LATENCY_FMT) { + /* print nothing if the buffers are empty */ + if (trace_empty(iter)) + return; + + print_trace_header(s, iter); + flags |= TRACE_GRAPH_PRINT_DURATION; + } else + flags |= TRACE_GRAPH_PRINT_ABS_TIME; + + __print_graph_headers_flags(s, flags); +} + void graph_trace_open(struct trace_iterator *iter) { /* pid and depth on the last trace processed */ @@ -1210,9 +1391,12 @@ void graph_trace_open(struct trace_iterator *iter) pid_t *pid = &(per_cpu_ptr(data->cpu_data, cpu)->last_pid); int *depth = &(per_cpu_ptr(data->cpu_data, cpu)->depth); int *ignore = &(per_cpu_ptr(data->cpu_data, cpu)->ignore); + int *depth_irq = &(per_cpu_ptr(data->cpu_data, cpu)->depth_irq); + *pid = -1; *depth = 0; *ignore = 0; + *depth_irq = -1; } iter->private = data; @@ -1235,6 +1419,14 @@ void graph_trace_close(struct trace_iterator *iter) } } +static int func_graph_set_flag(u32 old_flags, u32 bit, int set) +{ + if (bit == TRACE_GRAPH_PRINT_IRQS) + ftrace_graph_skip_irqs = !set; + + return 0; +} + static struct trace_event_functions graph_functions = { .trace = print_graph_function_event, }; @@ -1261,6 +1453,7 @@ static struct tracer graph_trace __read_mostly = { .print_line = print_graph_function, .print_header = print_graph_headers, .flags = &tracer_flags, + .set_flag = func_graph_set_flag, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_function_graph, #endif diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index 73a6b06..5cf8c60 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -87,14 +87,22 @@ static __cacheline_aligned_in_smp unsigned long max_sequence; #ifdef CONFIG_FUNCTION_TRACER /* - * irqsoff uses its own tracer function to keep the overhead down: + * Prologue for the preempt and irqs off function tracers. + * + * Returns 1 if it is OK to continue, and data->disabled is + * incremented. + * 0 if the trace is to be ignored, and data->disabled + * is kept the same. + * + * Note, this function is also used outside this ifdef but + * inside the #ifdef of the function graph tracer below. + * This is OK, since the function graph tracer is + * dependent on the function tracer. */ -static void -irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip) +static int func_prolog_dec(struct trace_array *tr, + struct trace_array_cpu **data, + unsigned long *flags) { - struct trace_array *tr = irqsoff_trace; - struct trace_array_cpu *data; - unsigned long flags; long disabled; int cpu; @@ -106,18 +114,38 @@ irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip) */ cpu = raw_smp_processor_id(); if (likely(!per_cpu(tracing_cpu, cpu))) - return; + return 0; - local_save_flags(flags); + local_save_flags(*flags); /* slight chance to get a false positive on tracing_cpu */ - if (!irqs_disabled_flags(flags)) - return; + if (!irqs_disabled_flags(*flags)) + return 0; - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); + *data = tr->data[cpu]; + disabled = atomic_inc_return(&(*data)->disabled); if (likely(disabled == 1)) - trace_function(tr, ip, parent_ip, flags, preempt_count()); + return 1; + + atomic_dec(&(*data)->disabled); + + return 0; +} + +/* + * irqsoff uses its own tracer function to keep the overhead down: + */ +static void +irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip) +{ + struct trace_array *tr = irqsoff_trace; + struct trace_array_cpu *data; + unsigned long flags; + + if (!func_prolog_dec(tr, &data, &flags)) + return; + + trace_function(tr, ip, parent_ip, flags, preempt_count()); atomic_dec(&data->disabled); } @@ -155,30 +183,16 @@ static int irqsoff_graph_entry(struct ftrace_graph_ent *trace) struct trace_array *tr = irqsoff_trace; struct trace_array_cpu *data; unsigned long flags; - long disabled; int ret; - int cpu; int pc; - cpu = raw_smp_processor_id(); - if (likely(!per_cpu(tracing_cpu, cpu))) + if (!func_prolog_dec(tr, &data, &flags)) return 0; - local_save_flags(flags); - /* slight chance to get a false positive on tracing_cpu */ - if (!irqs_disabled_flags(flags)) - return 0; - - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - - if (likely(disabled == 1)) { - pc = preempt_count(); - ret = __trace_graph_entry(tr, trace, flags, pc); - } else - ret = 0; - + pc = preempt_count(); + ret = __trace_graph_entry(tr, trace, flags, pc); atomic_dec(&data->disabled); + return ret; } @@ -187,27 +201,13 @@ static void irqsoff_graph_return(struct ftrace_graph_ret *trace) struct trace_array *tr = irqsoff_trace; struct trace_array_cpu *data; unsigned long flags; - long disabled; - int cpu; int pc; - cpu = raw_smp_processor_id(); - if (likely(!per_cpu(tracing_cpu, cpu))) + if (!func_prolog_dec(tr, &data, &flags)) return; - local_save_flags(flags); - /* slight chance to get a false positive on tracing_cpu */ - if (!irqs_disabled_flags(flags)) - return; - - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - - if (likely(disabled == 1)) { - pc = preempt_count(); - __trace_graph_return(tr, trace, flags, pc); - } - + pc = preempt_count(); + __trace_graph_return(tr, trace, flags, pc); atomic_dec(&data->disabled); } @@ -229,75 +229,33 @@ static void irqsoff_trace_close(struct trace_iterator *iter) static enum print_line_t irqsoff_print_line(struct trace_iterator *iter) { - u32 flags = GRAPH_TRACER_FLAGS; - - if (trace_flags & TRACE_ITER_LATENCY_FMT) - flags |= TRACE_GRAPH_PRINT_DURATION; - else - flags |= TRACE_GRAPH_PRINT_ABS_TIME; - /* * In graph mode call the graph tracer output function, * otherwise go with the TRACE_FN event handler */ if (is_graph()) - return print_graph_function_flags(iter, flags); + return print_graph_function_flags(iter, GRAPH_TRACER_FLAGS); return TRACE_TYPE_UNHANDLED; } static void irqsoff_print_header(struct seq_file *s) { - if (is_graph()) { - struct trace_iterator *iter = s->private; - u32 flags = GRAPH_TRACER_FLAGS; - - if (trace_flags & TRACE_ITER_LATENCY_FMT) { - /* print nothing if the buffers are empty */ - if (trace_empty(iter)) - return; - - print_trace_header(s, iter); - flags |= TRACE_GRAPH_PRINT_DURATION; - } else - flags |= TRACE_GRAPH_PRINT_ABS_TIME; - - print_graph_headers_flags(s, flags); - } else + if (is_graph()) + print_graph_headers_flags(s, GRAPH_TRACER_FLAGS); + else trace_default_header(s); } static void -trace_graph_function(struct trace_array *tr, - unsigned long ip, unsigned long flags, int pc) -{ - u64 time = trace_clock_local(); - struct ftrace_graph_ent ent = { - .func = ip, - .depth = 0, - }; - struct ftrace_graph_ret ret = { - .func = ip, - .depth = 0, - .calltime = time, - .rettime = time, - }; - - __trace_graph_entry(tr, &ent, flags, pc); - __trace_graph_return(tr, &ret, flags, pc); -} - -static void __trace_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc) { - if (!is_graph()) + if (is_graph()) + trace_graph_function(tr, ip, parent_ip, flags, pc); + else trace_function(tr, ip, parent_ip, flags, pc); - else { - trace_graph_function(tr, parent_ip, flags, pc); - trace_graph_function(tr, ip, flags, pc); - } } #else diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 4086eae..7319559 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -31,48 +31,98 @@ static int wakeup_rt; static arch_spinlock_t wakeup_lock = (arch_spinlock_t)__ARCH_SPIN_LOCK_UNLOCKED; +static void wakeup_reset(struct trace_array *tr); static void __wakeup_reset(struct trace_array *tr); +static int wakeup_graph_entry(struct ftrace_graph_ent *trace); +static void wakeup_graph_return(struct ftrace_graph_ret *trace); static int save_lat_flag; +#define TRACE_DISPLAY_GRAPH 1 + +static struct tracer_opt trace_opts[] = { +#ifdef CONFIG_FUNCTION_GRAPH_TRACER + /* display latency trace as call graph */ + { TRACER_OPT(display-graph, TRACE_DISPLAY_GRAPH) }, +#endif + { } /* Empty entry */ +}; + +static struct tracer_flags tracer_flags = { + .val = 0, + .opts = trace_opts, +}; + +#define is_graph() (tracer_flags.val & TRACE_DISPLAY_GRAPH) + #ifdef CONFIG_FUNCTION_TRACER + /* - * irqsoff uses its own tracer function to keep the overhead down: + * Prologue for the wakeup function tracers. + * + * Returns 1 if it is OK to continue, and preemption + * is disabled and data->disabled is incremented. + * 0 if the trace is to be ignored, and preemption + * is not disabled and data->disabled is + * kept the same. + * + * Note, this function is also used outside this ifdef but + * inside the #ifdef of the function graph tracer below. + * This is OK, since the function graph tracer is + * dependent on the function tracer. */ -static void -wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) +static int +func_prolog_preempt_disable(struct trace_array *tr, + struct trace_array_cpu **data, + int *pc) { - struct trace_array *tr = wakeup_trace; - struct trace_array_cpu *data; - unsigned long flags; long disabled; int cpu; - int pc; if (likely(!wakeup_task)) - return; + return 0; - pc = preempt_count(); + *pc = preempt_count(); preempt_disable_notrace(); cpu = raw_smp_processor_id(); if (cpu != wakeup_current_cpu) goto out_enable; - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); + *data = tr->data[cpu]; + disabled = atomic_inc_return(&(*data)->disabled); if (unlikely(disabled != 1)) goto out; - local_irq_save(flags); + return 1; - trace_function(tr, ip, parent_ip, flags, pc); +out: + atomic_dec(&(*data)->disabled); + +out_enable: + preempt_enable_notrace(); + return 0; +} +/* + * wakeup uses its own tracer function to keep the overhead down: + */ +static void +wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) +{ + struct trace_array *tr = wakeup_trace; + struct trace_array_cpu *data; + unsigned long flags; + int pc; + + if (!func_prolog_preempt_disable(tr, &data, &pc)) + return; + + local_irq_save(flags); + trace_function(tr, ip, parent_ip, flags, pc); local_irq_restore(flags); - out: atomic_dec(&data->disabled); - out_enable: preempt_enable_notrace(); } @@ -82,6 +132,156 @@ static struct ftrace_ops trace_ops __read_mostly = }; #endif /* CONFIG_FUNCTION_TRACER */ +static int start_func_tracer(int graph) +{ + int ret; + + if (!graph) + ret = register_ftrace_function(&trace_ops); + else + ret = register_ftrace_graph(&wakeup_graph_return, + &wakeup_graph_entry); + + if (!ret && tracing_is_enabled()) + tracer_enabled = 1; + else + tracer_enabled = 0; + + return ret; +} + +static void stop_func_tracer(int graph) +{ + tracer_enabled = 0; + + if (!graph) + unregister_ftrace_function(&trace_ops); + else + unregister_ftrace_graph(); +} + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +static int wakeup_set_flag(u32 old_flags, u32 bit, int set) +{ + + if (!(bit & TRACE_DISPLAY_GRAPH)) + return -EINVAL; + + if (!(is_graph() ^ set)) + return 0; + + stop_func_tracer(!set); + + wakeup_reset(wakeup_trace); + tracing_max_latency = 0; + + return start_func_tracer(set); +} + +static int wakeup_graph_entry(struct ftrace_graph_ent *trace) +{ + struct trace_array *tr = wakeup_trace; + struct trace_array_cpu *data; + unsigned long flags; + int pc, ret = 0; + + if (!func_prolog_preempt_disable(tr, &data, &pc)) + return 0; + + local_save_flags(flags); + ret = __trace_graph_entry(tr, trace, flags, pc); + atomic_dec(&data->disabled); + preempt_enable_notrace(); + + return ret; +} + +static void wakeup_graph_return(struct ftrace_graph_ret *trace) +{ + struct trace_array *tr = wakeup_trace; + struct trace_array_cpu *data; + unsigned long flags; + int pc; + + if (!func_prolog_preempt_disable(tr, &data, &pc)) + return; + + local_save_flags(flags); + __trace_graph_return(tr, trace, flags, pc); + atomic_dec(&data->disabled); + + preempt_enable_notrace(); + return; +} + +static void wakeup_trace_open(struct trace_iterator *iter) +{ + if (is_graph()) + graph_trace_open(iter); +} + +static void wakeup_trace_close(struct trace_iterator *iter) +{ + if (iter->private) + graph_trace_close(iter); +} + +#define GRAPH_TRACER_FLAGS (TRACE_GRAPH_PRINT_PROC) + +static enum print_line_t wakeup_print_line(struct trace_iterator *iter) +{ + /* + * In graph mode call the graph tracer output function, + * otherwise go with the TRACE_FN event handler + */ + if (is_graph()) + return print_graph_function_flags(iter, GRAPH_TRACER_FLAGS); + + return TRACE_TYPE_UNHANDLED; +} + +static void wakeup_print_header(struct seq_file *s) +{ + if (is_graph()) + print_graph_headers_flags(s, GRAPH_TRACER_FLAGS); + else + trace_default_header(s); +} + +static void +__trace_function(struct trace_array *tr, + unsigned long ip, unsigned long parent_ip, + unsigned long flags, int pc) +{ + if (is_graph()) + trace_graph_function(tr, ip, parent_ip, flags, pc); + else + trace_function(tr, ip, parent_ip, flags, pc); +} +#else +#define __trace_function trace_function + +static int wakeup_set_flag(u32 old_flags, u32 bit, int set) +{ + return -EINVAL; +} + +static int wakeup_graph_entry(struct ftrace_graph_ent *trace) +{ + return -1; +} + +static enum print_line_t wakeup_print_line(struct trace_iterator *iter) +{ + return TRACE_TYPE_UNHANDLED; +} + +static void wakeup_graph_return(struct ftrace_graph_ret *trace) { } +static void wakeup_print_header(struct seq_file *s) { } +static void wakeup_trace_open(struct trace_iterator *iter) { } +static void wakeup_trace_close(struct trace_iterator *iter) { } +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ + /* * Should this new latency be reported/recorded? */ @@ -152,7 +352,7 @@ probe_wakeup_sched_switch(void *ignore, /* The task we are waiting for is waking up */ data = wakeup_trace->data[wakeup_cpu]; - trace_function(wakeup_trace, CALLER_ADDR0, CALLER_ADDR1, flags, pc); + __trace_function(wakeup_trace, CALLER_ADDR0, CALLER_ADDR1, flags, pc); tracing_sched_switch_trace(wakeup_trace, prev, next, flags, pc); T0 = data->preempt_timestamp; @@ -252,7 +452,7 @@ probe_wakeup(void *ignore, struct task_struct *p, int success) * is not called by an assembly function (where as schedule is) * it should be safe to use it here. */ - trace_function(wakeup_trace, CALLER_ADDR1, CALLER_ADDR2, flags, pc); + __trace_function(wakeup_trace, CALLER_ADDR1, CALLER_ADDR2, flags, pc); out_locked: arch_spin_unlock(&wakeup_lock); @@ -303,12 +503,8 @@ static void start_wakeup_tracer(struct trace_array *tr) */ smp_wmb(); - register_ftrace_function(&trace_ops); - - if (tracing_is_enabled()) - tracer_enabled = 1; - else - tracer_enabled = 0; + if (start_func_tracer(is_graph())) + printk(KERN_ERR "failed to start wakeup tracer\n"); return; fail_deprobe_wake_new: @@ -320,7 +516,7 @@ fail_deprobe: static void stop_wakeup_tracer(struct trace_array *tr) { tracer_enabled = 0; - unregister_ftrace_function(&trace_ops); + stop_func_tracer(is_graph()); unregister_trace_sched_switch(probe_wakeup_sched_switch, NULL); unregister_trace_sched_wakeup_new(probe_wakeup, NULL); unregister_trace_sched_wakeup(probe_wakeup, NULL); @@ -379,9 +575,15 @@ static struct tracer wakeup_tracer __read_mostly = .start = wakeup_tracer_start, .stop = wakeup_tracer_stop, .print_max = 1, + .print_header = wakeup_print_header, + .print_line = wakeup_print_line, + .flags = &tracer_flags, + .set_flag = wakeup_set_flag, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_wakeup, #endif + .open = wakeup_trace_open, + .close = wakeup_trace_close, .use_max_tr = 1, }; @@ -394,9 +596,15 @@ static struct tracer wakeup_rt_tracer __read_mostly = .stop = wakeup_tracer_stop, .wait_pipe = poll_wait_pipe, .print_max = 1, + .print_header = wakeup_print_header, + .print_line = wakeup_print_line, + .flags = &tracer_flags, + .set_flag = wakeup_set_flag, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_wakeup, #endif + .open = wakeup_trace_open, + .close = wakeup_trace_close, .use_max_tr = 1, }; diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c index a7cc379..209b379 100644 --- a/kernel/trace/trace_workqueue.c +++ b/kernel/trace/trace_workqueue.c @@ -263,6 +263,11 @@ int __init trace_workqueue_early_init(void) { int ret, cpu; + for_each_possible_cpu(cpu) { + spin_lock_init(&workqueue_cpu_stat(cpu)->lock); + INIT_LIST_HEAD(&workqueue_cpu_stat(cpu)->list); + } + ret = register_trace_workqueue_insertion(probe_workqueue_insertion, NULL); if (ret) goto out; @@ -279,11 +284,6 @@ int __init trace_workqueue_early_init(void) if (ret) goto no_creation; - for_each_possible_cpu(cpu) { - spin_lock_init(&workqueue_cpu_stat(cpu)->lock); - INIT_LIST_HEAD(&workqueue_cpu_stat(cpu)->list); - } - return 0; no_creation: diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c index c77f3ec..e95ee7f 100644 --- a/kernel/tracepoint.c +++ b/kernel/tracepoint.c @@ -25,6 +25,7 @@ #include <linux/err.h> #include <linux/slab.h> #include <linux/sched.h> +#include <linux/jump_label.h> extern struct tracepoint __start___tracepoints[]; extern struct tracepoint __stop___tracepoints[]; @@ -263,7 +264,13 @@ static void set_tracepoint(struct tracepoint_entry **entry, * is used. */ rcu_assign_pointer(elem->funcs, (*entry)->funcs); - elem->state = active; + if (!elem->state && active) { + jump_label_enable(&elem->state); + elem->state = active; + } else if (elem->state && !active) { + jump_label_disable(&elem->state); + elem->state = active; + } } /* @@ -277,7 +284,10 @@ static void disable_tracepoint(struct tracepoint *elem) if (elem->unregfunc && elem->state) elem->unregfunc(); - elem->state = 0; + if (elem->state) { + jump_label_disable(&elem->state); + elem->state = 0; + } rcu_assign_pointer(elem->funcs, NULL); } diff --git a/kernel/watchdog.c b/kernel/watchdog.c index 2feb287..bafba68 100644 --- a/kernel/watchdog.c +++ b/kernel/watchdog.c @@ -43,7 +43,6 @@ static DEFINE_PER_CPU(unsigned long, hrtimer_interrupts_saved); static DEFINE_PER_CPU(struct perf_event *, watchdog_ev); #endif -static int __read_mostly did_panic; static int __initdata no_watchdog; @@ -187,18 +186,6 @@ static int is_softlockup(unsigned long touch_ts) return 0; } -static int -watchdog_panic(struct notifier_block *this, unsigned long event, void *ptr) -{ - did_panic = 1; - - return NOTIFY_DONE; -} - -static struct notifier_block panic_block = { - .notifier_call = watchdog_panic, -}; - #ifdef CONFIG_HARDLOCKUP_DETECTOR static struct perf_event_attr wd_hw_attr = { .type = PERF_TYPE_HARDWARE, @@ -371,14 +358,14 @@ static int watchdog_nmi_enable(int cpu) /* Try to register using hardware perf events */ wd_attr = &wd_hw_attr; wd_attr->sample_period = hw_nmi_get_sample_period(); - event = perf_event_create_kernel_counter(wd_attr, cpu, -1, watchdog_overflow_callback); + event = perf_event_create_kernel_counter(wd_attr, cpu, NULL, watchdog_overflow_callback); if (!IS_ERR(event)) { printk(KERN_INFO "NMI watchdog enabled, takes one hw-pmu counter.\n"); goto out_save; } printk(KERN_ERR "NMI watchdog failed to create perf event on cpu%i: %p\n", cpu, event); - return -1; + return PTR_ERR(event); /* success path */ out_save: @@ -422,17 +409,19 @@ static int watchdog_prepare_cpu(int cpu) static int watchdog_enable(int cpu) { struct task_struct *p = per_cpu(softlockup_watchdog, cpu); + int err; /* enable the perf event */ - if (watchdog_nmi_enable(cpu) != 0) - return -1; + err = watchdog_nmi_enable(cpu); + if (err) + return err; /* create the watchdog thread */ if (!p) { p = kthread_create(watchdog, (void *)(unsigned long)cpu, "watchdog/%d", cpu); if (IS_ERR(p)) { printk(KERN_ERR "softlockup watchdog for %i failed\n", cpu); - return -1; + return PTR_ERR(p); } kthread_bind(p, cpu); per_cpu(watchdog_touch_ts, cpu) = 0; @@ -484,6 +473,9 @@ static void watchdog_disable_all_cpus(void) { int cpu; + if (no_watchdog) + return; + for_each_online_cpu(cpu) watchdog_disable(cpu); @@ -526,17 +518,16 @@ static int __cpuinit cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) { int hotcpu = (unsigned long)hcpu; + int err = 0; switch (action) { case CPU_UP_PREPARE: case CPU_UP_PREPARE_FROZEN: - if (watchdog_prepare_cpu(hotcpu)) - return NOTIFY_BAD; + err = watchdog_prepare_cpu(hotcpu); break; case CPU_ONLINE: case CPU_ONLINE_FROZEN: - if (watchdog_enable(hotcpu)) - return NOTIFY_BAD; + err = watchdog_enable(hotcpu); break; #ifdef CONFIG_HOTPLUG_CPU case CPU_UP_CANCELED: @@ -549,7 +540,7 @@ cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) break; #endif /* CONFIG_HOTPLUG_CPU */ } - return NOTIFY_OK; + return notifier_from_errno(err); } static struct notifier_block __cpuinitdata cpu_nfb = { @@ -565,13 +556,11 @@ static int __init spawn_watchdog_task(void) return 0; err = cpu_callback(&cpu_nfb, CPU_UP_PREPARE, cpu); - WARN_ON(err == NOTIFY_BAD); + WARN_ON(notifier_to_errno(err)); cpu_callback(&cpu_nfb, CPU_ONLINE, cpu); register_cpu_notifier(&cpu_nfb); - atomic_notifier_chain_register(&panic_notifier_list, &panic_block); - return 0; } early_initcall(spawn_watchdog_task); diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 9886cf5..21ac830 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -482,6 +482,7 @@ config PROVE_LOCKING select DEBUG_SPINLOCK select DEBUG_MUTEXES select DEBUG_LOCK_ALLOC + select TRACE_IRQFLAGS default n help This feature enables the kernel to prove that all locking @@ -596,11 +597,10 @@ config DEBUG_LOCKDEP of more runtime overhead. config TRACE_IRQFLAGS - depends on DEBUG_KERNEL bool - default y - depends on TRACE_IRQFLAGS_SUPPORT - depends on PROVE_LOCKING + help + Enables hooks to interrupt enabling and disabling for + either tracing or lock debugging. config DEBUG_SPINLOCK_SLEEP bool "Spinlock debugging: sleep-inside-spinlock checking" diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c index 02afc25..7bd6df7 100644 --- a/lib/dynamic_debug.c +++ b/lib/dynamic_debug.c @@ -26,19 +26,11 @@ #include <linux/dynamic_debug.h> #include <linux/debugfs.h> #include <linux/slab.h> +#include <linux/jump_label.h> extern struct _ddebug __start___verbose[]; extern struct _ddebug __stop___verbose[]; -/* dynamic_debug_enabled, and dynamic_debug_enabled2 are bitmasks in which - * bit n is set to 1 if any modname hashes into the bucket n, 0 otherwise. They - * use independent hash functions, to reduce the chance of false positives. - */ -long long dynamic_debug_enabled; -EXPORT_SYMBOL_GPL(dynamic_debug_enabled); -long long dynamic_debug_enabled2; -EXPORT_SYMBOL_GPL(dynamic_debug_enabled2); - struct ddebug_table { struct list_head link; char *mod_name; @@ -88,26 +80,6 @@ static char *ddebug_describe_flags(struct _ddebug *dp, char *buf, } /* - * must be called with ddebug_lock held - */ - -static int disabled_hash(char hash, bool first_table) -{ - struct ddebug_table *dt; - char table_hash_value; - - list_for_each_entry(dt, &ddebug_tables, link) { - if (first_table) - table_hash_value = dt->ddebugs->primary_hash; - else - table_hash_value = dt->ddebugs->secondary_hash; - if (dt->num_enabled && (hash == table_hash_value)) - return 0; - } - return 1; -} - -/* * Search the tables for _ddebug's which match the given * `query' and apply the `flags' and `mask' to them. Tells * the user which ddebug's were changed, or whether none @@ -170,17 +142,9 @@ static void ddebug_change(const struct ddebug_query *query, dt->num_enabled++; dp->flags = newflags; if (newflags) { - dynamic_debug_enabled |= - (1LL << dp->primary_hash); - dynamic_debug_enabled2 |= - (1LL << dp->secondary_hash); + jump_label_enable(&dp->enabled); } else { - if (disabled_hash(dp->primary_hash, true)) - dynamic_debug_enabled &= - ~(1LL << dp->primary_hash); - if (disabled_hash(dp->secondary_hash, false)) - dynamic_debug_enabled2 &= - ~(1LL << dp->secondary_hash); + jump_label_disable(&dp->enabled); } if (verbose) printk(KERN_INFO diff --git a/net/core/datagram.c b/net/core/datagram.c index 251997a..282806b 100644 --- a/net/core/datagram.c +++ b/net/core/datagram.c @@ -243,6 +243,7 @@ void skb_free_datagram_locked(struct sock *sk, struct sk_buff *skb) unlock_sock_fast(sk, slow); /* skb is now orphaned, can be freed outside of locked section */ + trace_kfree_skb(skb, skb_free_datagram_locked); __kfree_skb(skb); } EXPORT_SYMBOL(skb_free_datagram_locked); diff --git a/net/core/dev.c b/net/core/dev.c index 660dd41..7ec85e2 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -128,6 +128,8 @@ #include <linux/jhash.h> #include <linux/random.h> #include <trace/events/napi.h> +#include <trace/events/net.h> +#include <trace/events/skb.h> #include <linux/pci.h> #include "net-sysfs.h" @@ -1978,6 +1980,7 @@ int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev, } rc = ops->ndo_start_xmit(skb, dev); + trace_net_dev_xmit(skb, rc); if (rc == NETDEV_TX_OK) txq_trans_update(txq); return rc; @@ -1998,6 +2001,7 @@ gso: skb_dst_drop(nskb); rc = ops->ndo_start_xmit(nskb, dev); + trace_net_dev_xmit(nskb, rc); if (unlikely(rc != NETDEV_TX_OK)) { if (rc & ~NETDEV_TX_MASK) goto out_kfree_gso_skb; @@ -2186,6 +2190,7 @@ int dev_queue_xmit(struct sk_buff *skb) #ifdef CONFIG_NET_CLS_ACT skb->tc_verd = SET_TC_AT(skb->tc_verd, AT_EGRESS); #endif + trace_net_dev_queue(skb); if (q->enqueue) { rc = __dev_xmit_skb(skb, q, dev, txq); goto out; @@ -2512,6 +2517,7 @@ int netif_rx(struct sk_buff *skb) if (netdev_tstamp_prequeue) net_timestamp_check(skb); + trace_netif_rx(skb); #ifdef CONFIG_RPS { struct rps_dev_flow voidflow, *rflow = &voidflow; @@ -2571,6 +2577,7 @@ static void net_tx_action(struct softirq_action *h) clist = clist->next; WARN_ON(atomic_read(&skb->users)); + trace_kfree_skb(skb, net_tx_action); __kfree_skb(skb); } } @@ -2828,6 +2835,7 @@ static int __netif_receive_skb(struct sk_buff *skb) if (!netdev_tstamp_prequeue) net_timestamp_check(skb); + trace_netif_receive_skb(skb); if (vlan_tx_tag_present(skb) && vlan_hwaccel_do_receive(skb)) return NET_RX_SUCCESS; diff --git a/net/core/net-traces.c b/net/core/net-traces.c index afa6380..7f1bb2a 100644 --- a/net/core/net-traces.c +++ b/net/core/net-traces.c @@ -26,6 +26,7 @@ #define CREATE_TRACE_POINTS #include <trace/events/skb.h> +#include <trace/events/net.h> #include <trace/events/napi.h> EXPORT_TRACEPOINT_SYMBOL_GPL(kfree_skb); diff --git a/net/core/skbuff.c b/net/core/skbuff.c index c83b421..56ba3c4 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -466,6 +466,7 @@ void consume_skb(struct sk_buff *skb) smp_rmb(); else if (likely(!atomic_dec_and_test(&skb->users))) return; + trace_consume_skb(skb); __kfree_skb(skb); } EXPORT_SYMBOL(consume_skb); diff --git a/scripts/Makefile b/scripts/Makefile index 842dbc2..2e08810 100644 --- a/scripts/Makefile +++ b/scripts/Makefile @@ -11,6 +11,7 @@ hostprogs-$(CONFIG_KALLSYMS) += kallsyms hostprogs-$(CONFIG_LOGO) += pnmtologo hostprogs-$(CONFIG_VT) += conmakehash hostprogs-$(CONFIG_IKCONFIG) += bin2c +hostprogs-$(BUILD_C_RECORDMCOUNT) += recordmcount always := $(hostprogs-y) $(hostprogs-m) diff --git a/scripts/Makefile.build b/scripts/Makefile.build index a1a5cf9..843bd4f 100644 --- a/scripts/Makefile.build +++ b/scripts/Makefile.build @@ -209,12 +209,22 @@ cmd_modversions = \ endif ifdef CONFIG_FTRACE_MCOUNT_RECORD +ifdef BUILD_C_RECORDMCOUNT +# Due to recursion, we must skip empty.o. +# The empty.o file is created in the make process in order to determine +# the target endianness and word size. It is made before all other C +# files, including recordmcount. +cmd_record_mcount = if [ $(@) != "scripts/mod/empty.o" ]; then \ + $(objtree)/scripts/recordmcount "$(@)"; \ + fi; +else cmd_record_mcount = set -e ; perl $(srctree)/scripts/recordmcount.pl "$(ARCH)" \ "$(if $(CONFIG_CPU_BIG_ENDIAN),big,little)" \ "$(if $(CONFIG_64BIT),64,32)" \ "$(OBJDUMP)" "$(OBJCOPY)" "$(CC)" "$(LD)" "$(NM)" "$(RM)" "$(MV)" \ "$(if $(part-of-module),1,0)" "$(@)"; endif +endif define rule_cc_o_c $(call echo-cmd,checksrc) $(cmd_checksrc) \ diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib index 54fd1b7..7bfcf1a 100644 --- a/scripts/Makefile.lib +++ b/scripts/Makefile.lib @@ -101,14 +101,6 @@ basename_flags = -D"KBUILD_BASENAME=KBUILD_STR($(call name-fix,$(basetarget)))" modname_flags = $(if $(filter 1,$(words $(modname))),\ -D"KBUILD_MODNAME=KBUILD_STR($(call name-fix,$(modname)))") -#hash values -ifdef CONFIG_DYNAMIC_DEBUG -debug_flags = -D"DEBUG_HASH=$(shell ./scripts/basic/hash djb2 $(@D)$(modname))"\ - -D"DEBUG_HASH2=$(shell ./scripts/basic/hash r5 $(@D)$(modname))" -else -debug_flags = -endif - orig_c_flags = $(KBUILD_CPPFLAGS) $(KBUILD_CFLAGS) $(KBUILD_SUBDIR_CCFLAGS) \ $(ccflags-y) $(CFLAGS_$(basetarget).o) _c_flags = $(filter-out $(CFLAGS_REMOVE_$(basetarget).o), $(orig_c_flags)) @@ -152,8 +144,7 @@ endif c_flags = -Wp,-MD,$(depfile) $(NOSTDINC_FLAGS) $(LINUXINCLUDE) \ $(__c_flags) $(modkern_cflags) \ - -D"KBUILD_STR(s)=\#s" $(basename_flags) $(modname_flags) \ - $(debug_flags) + -D"KBUILD_STR(s)=\#s" $(basename_flags) $(modname_flags) a_flags = -Wp,-MD,$(depfile) $(NOSTDINC_FLAGS) $(LINUXINCLUDE) \ $(__a_flags) $(modkern_aflags) diff --git a/scripts/basic/Makefile b/scripts/basic/Makefile index 0955995..4c324a1 100644 --- a/scripts/basic/Makefile +++ b/scripts/basic/Makefile @@ -9,7 +9,7 @@ # fixdep: Used to generate dependency information during build process # docproc: Used in Documentation/DocBook -hostprogs-y := fixdep docproc hash +hostprogs-y := fixdep docproc always := $(hostprogs-y) # fixdep is needed to compile other host programs diff --git a/scripts/basic/hash.c b/scripts/basic/hash.c deleted file mode 100644 index 2ef5d3f..0000000 --- a/scripts/basic/hash.c +++ /dev/null @@ -1,64 +0,0 @@ -/* - * Copyright (C) 2008 Red Hat, Inc., Jason Baron <jbaron@redhat.com> - * - */ - -#include <stdio.h> -#include <stdlib.h> -#include <string.h> - -#define DYNAMIC_DEBUG_HASH_BITS 6 - -static const char *program; - -static void usage(void) -{ - printf("Usage: %s <djb2|r5> <modname>\n", program); - exit(1); -} - -/* djb2 hashing algorithm by Dan Bernstein. From: - * http://www.cse.yorku.ca/~oz/hash.html - */ - -static unsigned int djb2_hash(char *str) -{ - unsigned long hash = 5381; - int c; - - c = *str; - while (c) { - hash = ((hash << 5) + hash) + c; - c = *++str; - } - return (unsigned int)(hash & ((1 << DYNAMIC_DEBUG_HASH_BITS) - 1)); -} - -static unsigned int r5_hash(char *str) -{ - unsigned long hash = 0; - int c; - - c = *str; - while (c) { - hash = (hash + (c << 4) + (c >> 4)) * 11; - c = *++str; - } - return (unsigned int)(hash & ((1 << DYNAMIC_DEBUG_HASH_BITS) - 1)); -} - -int main(int argc, char *argv[]) -{ - program = argv[0]; - - if (argc != 3) - usage(); - if (!strcmp(argv[1], "djb2")) - printf("%d\n", djb2_hash(argv[2])); - else if (!strcmp(argv[1], "r5")) - printf("%d\n", r5_hash(argv[2])); - else - usage(); - exit(0); -} - diff --git a/scripts/gcc-goto.sh b/scripts/gcc-goto.sh new file mode 100644 index 0000000..520d16b --- /dev/null +++ b/scripts/gcc-goto.sh @@ -0,0 +1,5 @@ +#!/bin/sh +# Test for gcc 'asm goto' suport +# Copyright (C) 2010, Jason Baron <jbaron@redhat.com> + +echo "int main(void) { entry: asm goto (\"\"::::entry); return 0; }" | $@ -x c - -c -o /dev/null >/dev/null 2>&1 && echo "y" diff --git a/scripts/recordmcount.c b/scripts/recordmcount.c new file mode 100644 index 0000000..26e1271 --- /dev/null +++ b/scripts/recordmcount.c @@ -0,0 +1,363 @@ +/* + * recordmcount.c: construct a table of the locations of calls to 'mcount' + * so that ftrace can find them quickly. + * Copyright 2009 John F. Reiser <jreiser@BitWagon.com>. All rights reserved. + * Licensed under the GNU General Public License, version 2 (GPLv2). + * + * Restructured to fit Linux format, as well as other updates: + * Copyright 2010 Steven Rostedt <srostedt@redhat.com>, Red Hat Inc. + */ + +/* + * Strategy: alter the .o file in-place. + * + * Append a new STRTAB that has the new section names, followed by a new array + * ElfXX_Shdr[] that has the new section headers, followed by the section + * contents for __mcount_loc and its relocations. The old shstrtab strings, + * and the old ElfXX_Shdr[] array, remain as "garbage" (commonly, a couple + * kilobytes.) Subsequent processing by /bin/ld (or the kernel module loader) + * will ignore the garbage regions, because they are not designated by the + * new .e_shoff nor the new ElfXX_Shdr[]. [In order to remove the garbage, + * then use "ld -r" to create a new file that omits the garbage.] + */ + +#include <sys/types.h> +#include <sys/mman.h> +#include <sys/stat.h> +#include <elf.h> +#include <fcntl.h> +#include <setjmp.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <unistd.h> + +static int fd_map; /* File descriptor for file being modified. */ +static int mmap_failed; /* Boolean flag. */ +static void *ehdr_curr; /* current ElfXX_Ehdr * for resource cleanup */ +static char gpfx; /* prefix for global symbol name (sometimes '_') */ +static struct stat sb; /* Remember .st_size, etc. */ +static jmp_buf jmpenv; /* setjmp/longjmp per-file error escape */ + +/* setjmp() return values */ +enum { + SJ_SETJMP = 0, /* hardwired first return */ + SJ_FAIL, + SJ_SUCCEED +}; + +/* Per-file resource cleanup when multiple files. */ +static void +cleanup(void) +{ + if (!mmap_failed) + munmap(ehdr_curr, sb.st_size); + else + free(ehdr_curr); + close(fd_map); +} + +static void __attribute__((noreturn)) +fail_file(void) +{ + cleanup(); + longjmp(jmpenv, SJ_FAIL); +} + +static void __attribute__((noreturn)) +succeed_file(void) +{ + cleanup(); + longjmp(jmpenv, SJ_SUCCEED); +} + +/* ulseek, uread, ...: Check return value for errors. */ + +static off_t +ulseek(int const fd, off_t const offset, int const whence) +{ + off_t const w = lseek(fd, offset, whence); + if ((off_t)-1 == w) { + perror("lseek"); + fail_file(); + } + return w; +} + +static size_t +uread(int const fd, void *const buf, size_t const count) +{ + size_t const n = read(fd, buf, count); + if (n != count) { + perror("read"); + fail_file(); + } + return n; +} + +static size_t +uwrite(int const fd, void const *const buf, size_t const count) +{ + size_t const n = write(fd, buf, count); + if (n != count) { + perror("write"); + fail_file(); + } + return n; +} + +static void * +umalloc(size_t size) +{ + void *const addr = malloc(size); + if (0 == addr) { + fprintf(stderr, "malloc failed: %zu bytes\n", size); + fail_file(); + } + return addr; +} + +/* + * Get the whole file as a programming convenience in order to avoid + * malloc+lseek+read+free of many pieces. If successful, then mmap + * avoids copying unused pieces; else just read the whole file. + * Open for both read and write; new info will be appended to the file. + * Use MAP_PRIVATE so that a few changes to the in-memory ElfXX_Ehdr + * do not propagate to the file until an explicit overwrite at the last. + * This preserves most aspects of consistency (all except .st_size) + * for simultaneous readers of the file while we are appending to it. + * However, multiple writers still are bad. We choose not to use + * locking because it is expensive and the use case of kernel build + * makes multiple writers unlikely. + */ +static void *mmap_file(char const *fname) +{ + void *addr; + + fd_map = open(fname, O_RDWR); + if (0 > fd_map || 0 > fstat(fd_map, &sb)) { + perror(fname); + fail_file(); + } + if (!S_ISREG(sb.st_mode)) { + fprintf(stderr, "not a regular file: %s\n", fname); + fail_file(); + } + addr = mmap(0, sb.st_size, PROT_READ|PROT_WRITE, MAP_PRIVATE, + fd_map, 0); + mmap_failed = 0; + if (MAP_FAILED == addr) { + mmap_failed = 1; + addr = umalloc(sb.st_size); + uread(fd_map, addr, sb.st_size); + } + return addr; +} + +/* w8rev, w8nat, ...: Handle endianness. */ + +static uint64_t w8rev(uint64_t const x) +{ + return ((0xff & (x >> (0 * 8))) << (7 * 8)) + | ((0xff & (x >> (1 * 8))) << (6 * 8)) + | ((0xff & (x >> (2 * 8))) << (5 * 8)) + | ((0xff & (x >> (3 * 8))) << (4 * 8)) + | ((0xff & (x >> (4 * 8))) << (3 * 8)) + | ((0xff & (x >> (5 * 8))) << (2 * 8)) + | ((0xff & (x >> (6 * 8))) << (1 * 8)) + | ((0xff & (x >> (7 * 8))) << (0 * 8)); +} + +static uint32_t w4rev(uint32_t const x) +{ + return ((0xff & (x >> (0 * 8))) << (3 * 8)) + | ((0xff & (x >> (1 * 8))) << (2 * 8)) + | ((0xff & (x >> (2 * 8))) << (1 * 8)) + | ((0xff & (x >> (3 * 8))) << (0 * 8)); +} + +static uint32_t w2rev(uint16_t const x) +{ + return ((0xff & (x >> (0 * 8))) << (1 * 8)) + | ((0xff & (x >> (1 * 8))) << (0 * 8)); +} + +static uint64_t w8nat(uint64_t const x) +{ + return x; +} + +static uint32_t w4nat(uint32_t const x) +{ + return x; +} + +static uint32_t w2nat(uint16_t const x) +{ + return x; +} + +static uint64_t (*w8)(uint64_t); +static uint32_t (*w)(uint32_t); +static uint32_t (*w2)(uint16_t); + +/* Names of the sections that could contain calls to mcount. */ +static int +is_mcounted_section_name(char const *const txtname) +{ + return 0 == strcmp(".text", txtname) || + 0 == strcmp(".sched.text", txtname) || + 0 == strcmp(".spinlock.text", txtname) || + 0 == strcmp(".irqentry.text", txtname) || + 0 == strcmp(".text.unlikely", txtname); +} + +/* 32 bit and 64 bit are very similar */ +#include "recordmcount.h" +#define RECORD_MCOUNT_64 +#include "recordmcount.h" + +static void +do_file(char const *const fname) +{ + Elf32_Ehdr *const ehdr = mmap_file(fname); + unsigned int reltype = 0; + + ehdr_curr = ehdr; + w = w4nat; + w2 = w2nat; + w8 = w8nat; + switch (ehdr->e_ident[EI_DATA]) { + static unsigned int const endian = 1; + default: { + fprintf(stderr, "unrecognized ELF data encoding %d: %s\n", + ehdr->e_ident[EI_DATA], fname); + fail_file(); + } break; + case ELFDATA2LSB: { + if (1 != *(unsigned char const *)&endian) { + /* main() is big endian, file.o is little endian. */ + w = w4rev; + w2 = w2rev; + w8 = w8rev; + } + } break; + case ELFDATA2MSB: { + if (0 != *(unsigned char const *)&endian) { + /* main() is little endian, file.o is big endian. */ + w = w4rev; + w2 = w2rev; + w8 = w8rev; + } + } break; + } /* end switch */ + if (0 != memcmp(ELFMAG, ehdr->e_ident, SELFMAG) + || ET_REL != w2(ehdr->e_type) + || EV_CURRENT != ehdr->e_ident[EI_VERSION]) { + fprintf(stderr, "unrecognized ET_REL file %s\n", fname); + fail_file(); + } + + gpfx = 0; + switch (w2(ehdr->e_machine)) { + default: { + fprintf(stderr, "unrecognized e_machine %d %s\n", + w2(ehdr->e_machine), fname); + fail_file(); + } break; + case EM_386: reltype = R_386_32; break; + case EM_ARM: reltype = R_ARM_ABS32; break; + case EM_IA_64: reltype = R_IA64_IMM64; gpfx = '_'; break; + case EM_PPC: reltype = R_PPC_ADDR32; gpfx = '_'; break; + case EM_PPC64: reltype = R_PPC64_ADDR64; gpfx = '_'; break; + case EM_S390: /* reltype: e_class */ gpfx = '_'; break; + case EM_SH: reltype = R_SH_DIR32; break; + case EM_SPARCV9: reltype = R_SPARC_64; gpfx = '_'; break; + case EM_X86_64: reltype = R_X86_64_64; break; + } /* end switch */ + + switch (ehdr->e_ident[EI_CLASS]) { + default: { + fprintf(stderr, "unrecognized ELF class %d %s\n", + ehdr->e_ident[EI_CLASS], fname); + fail_file(); + } break; + case ELFCLASS32: { + if (sizeof(Elf32_Ehdr) != w2(ehdr->e_ehsize) + || sizeof(Elf32_Shdr) != w2(ehdr->e_shentsize)) { + fprintf(stderr, + "unrecognized ET_REL file: %s\n", fname); + fail_file(); + } + if (EM_S390 == w2(ehdr->e_machine)) + reltype = R_390_32; + do32(ehdr, fname, reltype); + } break; + case ELFCLASS64: { + Elf64_Ehdr *const ghdr = (Elf64_Ehdr *)ehdr; + if (sizeof(Elf64_Ehdr) != w2(ghdr->e_ehsize) + || sizeof(Elf64_Shdr) != w2(ghdr->e_shentsize)) { + fprintf(stderr, + "unrecognized ET_REL file: %s\n", fname); + fail_file(); + } + if (EM_S390 == w2(ghdr->e_machine)) + reltype = R_390_64; + do64(ghdr, fname, reltype); + } break; + } /* end switch */ + + cleanup(); +} + +int +main(int argc, char const *argv[]) +{ + const char ftrace[] = "kernel/trace/ftrace.o"; + int ftrace_size = sizeof(ftrace) - 1; + int n_error = 0; /* gcc-4.3.0 false positive complaint */ + + if (argc <= 1) { + fprintf(stderr, "usage: recordmcount file.o...\n"); + return 0; + } + + /* Process each file in turn, allowing deep failure. */ + for (--argc, ++argv; 0 < argc; --argc, ++argv) { + int const sjval = setjmp(jmpenv); + int len; + + /* + * The file kernel/trace/ftrace.o references the mcount + * function but does not call it. Since ftrace.o should + * not be traced anyway, we just skip it. + */ + len = strlen(argv[0]); + if (len >= ftrace_size && + strcmp(argv[0] + (len - ftrace_size), ftrace) == 0) + continue; + + switch (sjval) { + default: { + fprintf(stderr, "internal error: %s\n", argv[0]); + exit(1); + } break; + case SJ_SETJMP: { /* normal sequence */ + /* Avoid problems if early cleanup() */ + fd_map = -1; + ehdr_curr = NULL; + mmap_failed = 1; + do_file(argv[0]); + } break; + case SJ_FAIL: { /* error in do_file or below */ + ++n_error; + } break; + case SJ_SUCCEED: { /* premature success */ + /* do nothing */ + } break; + } /* end switch */ + } + return !!n_error; +} + + diff --git a/scripts/recordmcount.h b/scripts/recordmcount.h new file mode 100644 index 0000000..7f39d09 --- /dev/null +++ b/scripts/recordmcount.h @@ -0,0 +1,366 @@ +/* + * recordmcount.h + * + * This code was taken out of recordmcount.c written by + * Copyright 2009 John F. Reiser <jreiser@BitWagon.com>. All rights reserved. + * + * The original code had the same algorithms for both 32bit + * and 64bit ELF files, but the code was duplicated to support + * the difference in structures that were used. This + * file creates a macro of everything that is different between + * the 64 and 32 bit code, such that by including this header + * twice we can create both sets of functions by including this + * header once with RECORD_MCOUNT_64 undefined, and again with + * it defined. + * + * This conversion to macros was done by: + * Copyright 2010 Steven Rostedt <srostedt@redhat.com>, Red Hat Inc. + * + * Licensed under the GNU General Public License, version 2 (GPLv2). + */ +#undef append_func +#undef sift_rel_mcount +#undef find_secsym_ndx +#undef __has_rel_mcount +#undef has_rel_mcount +#undef tot_relsize +#undef do_func +#undef Elf_Ehdr +#undef Elf_Shdr +#undef Elf_Rel +#undef Elf_Rela +#undef Elf_Sym +#undef ELF_R_SYM +#undef ELF_R_INFO +#undef ELF_ST_BIND +#undef uint_t +#undef _w +#undef _align +#undef _size + +#ifdef RECORD_MCOUNT_64 +# define append_func append64 +# define sift_rel_mcount sift64_rel_mcount +# define find_secsym_ndx find64_secsym_ndx +# define __has_rel_mcount __has64_rel_mcount +# define has_rel_mcount has64_rel_mcount +# define tot_relsize tot64_relsize +# define do_func do64 +# define Elf_Ehdr Elf64_Ehdr +# define Elf_Shdr Elf64_Shdr +# define Elf_Rel Elf64_Rel +# define Elf_Rela Elf64_Rela +# define Elf_Sym Elf64_Sym +# define ELF_R_SYM ELF64_R_SYM +# define ELF_R_INFO ELF64_R_INFO +# define ELF_ST_BIND ELF64_ST_BIND +# define uint_t uint64_t +# define _w w8 +# define _align 7u +# define _size 8 +#else +# define append_func append32 +# define sift_rel_mcount sift32_rel_mcount +# define find_secsym_ndx find32_secsym_ndx +# define __has_rel_mcount __has32_rel_mcount +# define has_rel_mcount has32_rel_mcount +# define tot_relsize tot32_relsize +# define do_func do32 +# define Elf_Ehdr Elf32_Ehdr +# define Elf_Shdr Elf32_Shdr +# define Elf_Rel Elf32_Rel +# define Elf_Rela Elf32_Rela +# define Elf_Sym Elf32_Sym +# define ELF_R_SYM ELF32_R_SYM +# define ELF_R_INFO ELF32_R_INFO +# define ELF_ST_BIND ELF32_ST_BIND +# define uint_t uint32_t +# define _w w +# define _align 3u +# define _size 4 +#endif + +/* Append the new shstrtab, Elf_Shdr[], __mcount_loc and its relocations. */ +static void append_func(Elf_Ehdr *const ehdr, + Elf_Shdr *const shstr, + uint_t const *const mloc0, + uint_t const *const mlocp, + Elf_Rel const *const mrel0, + Elf_Rel const *const mrelp, + unsigned int const rel_entsize, + unsigned int const symsec_sh_link) +{ + /* Begin constructing output file */ + Elf_Shdr mcsec; + char const *mc_name = (sizeof(Elf_Rela) == rel_entsize) + ? ".rela__mcount_loc" + : ".rel__mcount_loc"; + unsigned const old_shnum = w2(ehdr->e_shnum); + uint_t const old_shoff = _w(ehdr->e_shoff); + uint_t const old_shstr_sh_size = _w(shstr->sh_size); + uint_t const old_shstr_sh_offset = _w(shstr->sh_offset); + uint_t t = 1 + strlen(mc_name) + _w(shstr->sh_size); + uint_t new_e_shoff; + + shstr->sh_size = _w(t); + shstr->sh_offset = _w(sb.st_size); + t += sb.st_size; + t += (_align & -t); /* word-byte align */ + new_e_shoff = t; + + /* body for new shstrtab */ + ulseek(fd_map, sb.st_size, SEEK_SET); + uwrite(fd_map, old_shstr_sh_offset + (void *)ehdr, old_shstr_sh_size); + uwrite(fd_map, mc_name, 1 + strlen(mc_name)); + + /* old(modified) Elf_Shdr table, word-byte aligned */ + ulseek(fd_map, t, SEEK_SET); + t += sizeof(Elf_Shdr) * old_shnum; + uwrite(fd_map, old_shoff + (void *)ehdr, + sizeof(Elf_Shdr) * old_shnum); + + /* new sections __mcount_loc and .rel__mcount_loc */ + t += 2*sizeof(mcsec); + mcsec.sh_name = w((sizeof(Elf_Rela) == rel_entsize) + strlen(".rel") + + old_shstr_sh_size); + mcsec.sh_type = w(SHT_PROGBITS); + mcsec.sh_flags = _w(SHF_ALLOC); + mcsec.sh_addr = 0; + mcsec.sh_offset = _w(t); + mcsec.sh_size = _w((void *)mlocp - (void *)mloc0); + mcsec.sh_link = 0; + mcsec.sh_info = 0; + mcsec.sh_addralign = _w(_size); + mcsec.sh_entsize = _w(_size); + uwrite(fd_map, &mcsec, sizeof(mcsec)); + + mcsec.sh_name = w(old_shstr_sh_size); + mcsec.sh_type = (sizeof(Elf_Rela) == rel_entsize) + ? w(SHT_RELA) + : w(SHT_REL); + mcsec.sh_flags = 0; + mcsec.sh_addr = 0; + mcsec.sh_offset = _w((void *)mlocp - (void *)mloc0 + t); + mcsec.sh_size = _w((void *)mrelp - (void *)mrel0); + mcsec.sh_link = w(symsec_sh_link); + mcsec.sh_info = w(old_shnum); + mcsec.sh_addralign = _w(_size); + mcsec.sh_entsize = _w(rel_entsize); + uwrite(fd_map, &mcsec, sizeof(mcsec)); + + uwrite(fd_map, mloc0, (void *)mlocp - (void *)mloc0); + uwrite(fd_map, mrel0, (void *)mrelp - (void *)mrel0); + + ehdr->e_shoff = _w(new_e_shoff); + ehdr->e_shnum = w2(2 + w2(ehdr->e_shnum)); /* {.rel,}__mcount_loc */ + ulseek(fd_map, 0, SEEK_SET); + uwrite(fd_map, ehdr, sizeof(*ehdr)); +} + + +/* + * Look at the relocations in order to find the calls to mcount. + * Accumulate the section offsets that are found, and their relocation info, + * onto the end of the existing arrays. + */ +static uint_t *sift_rel_mcount(uint_t *mlocp, + unsigned const offbase, + Elf_Rel **const mrelpp, + Elf_Shdr const *const relhdr, + Elf_Ehdr const *const ehdr, + unsigned const recsym, + uint_t const recval, + unsigned const reltype) +{ + uint_t *const mloc0 = mlocp; + Elf_Rel *mrelp = *mrelpp; + Elf_Shdr *const shdr0 = (Elf_Shdr *)(_w(ehdr->e_shoff) + + (void *)ehdr); + unsigned const symsec_sh_link = w(relhdr->sh_link); + Elf_Shdr const *const symsec = &shdr0[symsec_sh_link]; + Elf_Sym const *const sym0 = (Elf_Sym const *)(_w(symsec->sh_offset) + + (void *)ehdr); + + Elf_Shdr const *const strsec = &shdr0[w(symsec->sh_link)]; + char const *const str0 = (char const *)(_w(strsec->sh_offset) + + (void *)ehdr); + + Elf_Rel const *const rel0 = (Elf_Rel const *)(_w(relhdr->sh_offset) + + (void *)ehdr); + unsigned rel_entsize = _w(relhdr->sh_entsize); + unsigned const nrel = _w(relhdr->sh_size) / rel_entsize; + Elf_Rel const *relp = rel0; + + unsigned mcountsym = 0; + unsigned t; + + for (t = nrel; t; --t) { + if (!mcountsym) { + Elf_Sym const *const symp = + &sym0[ELF_R_SYM(_w(relp->r_info))]; + char const *symname = &str0[w(symp->st_name)]; + + if ('.' == symname[0]) + ++symname; /* ppc64 hack */ + if (0 == strcmp((('_' == gpfx) ? "_mcount" : "mcount"), + symname)) + mcountsym = ELF_R_SYM(_w(relp->r_info)); + } + + if (mcountsym == ELF_R_SYM(_w(relp->r_info))) { + uint_t const addend = _w(_w(relp->r_offset) - recval); + + mrelp->r_offset = _w(offbase + + ((void *)mlocp - (void *)mloc0)); + mrelp->r_info = _w(ELF_R_INFO(recsym, reltype)); + if (sizeof(Elf_Rela) == rel_entsize) { + ((Elf_Rela *)mrelp)->r_addend = addend; + *mlocp++ = 0; + } else + *mlocp++ = addend; + + mrelp = (Elf_Rel *)(rel_entsize + (void *)mrelp); + } + relp = (Elf_Rel const *)(rel_entsize + (void *)relp); + } + *mrelpp = mrelp; + return mlocp; +} + + +/* + * Find a symbol in the given section, to be used as the base for relocating + * the table of offsets of calls to mcount. A local or global symbol suffices, + * but avoid a Weak symbol because it may be overridden; the change in value + * would invalidate the relocations of the offsets of the calls to mcount. + * Often the found symbol will be the unnamed local symbol generated by + * GNU 'as' for the start of each section. For example: + * Num: Value Size Type Bind Vis Ndx Name + * 2: 00000000 0 SECTION LOCAL DEFAULT 1 + */ +static unsigned find_secsym_ndx(unsigned const txtndx, + char const *const txtname, + uint_t *const recvalp, + Elf_Shdr const *const symhdr, + Elf_Ehdr const *const ehdr) +{ + Elf_Sym const *const sym0 = (Elf_Sym const *)(_w(symhdr->sh_offset) + + (void *)ehdr); + unsigned const nsym = _w(symhdr->sh_size) / _w(symhdr->sh_entsize); + Elf_Sym const *symp; + unsigned t; + + for (symp = sym0, t = nsym; t; --t, ++symp) { + unsigned int const st_bind = ELF_ST_BIND(symp->st_info); + + if (txtndx == w2(symp->st_shndx) + /* avoid STB_WEAK */ + && (STB_LOCAL == st_bind || STB_GLOBAL == st_bind)) { + *recvalp = _w(symp->st_value); + return symp - sym0; + } + } + fprintf(stderr, "Cannot find symbol for section %d: %s.\n", + txtndx, txtname); + fail_file(); +} + + +/* Evade ISO C restriction: no declaration after statement in has_rel_mcount. */ +static char const * +__has_rel_mcount(Elf_Shdr const *const relhdr, /* is SHT_REL or SHT_RELA */ + Elf_Shdr const *const shdr0, + char const *const shstrtab, + char const *const fname) +{ + /* .sh_info depends on .sh_type == SHT_REL[,A] */ + Elf_Shdr const *const txthdr = &shdr0[w(relhdr->sh_info)]; + char const *const txtname = &shstrtab[w(txthdr->sh_name)]; + + if (0 == strcmp("__mcount_loc", txtname)) { + fprintf(stderr, "warning: __mcount_loc already exists: %s\n", + fname); + succeed_file(); + } + if (SHT_PROGBITS != w(txthdr->sh_type) || + !is_mcounted_section_name(txtname)) + return NULL; + return txtname; +} + +static char const *has_rel_mcount(Elf_Shdr const *const relhdr, + Elf_Shdr const *const shdr0, + char const *const shstrtab, + char const *const fname) +{ + if (SHT_REL != w(relhdr->sh_type) && SHT_RELA != w(relhdr->sh_type)) + return NULL; + return __has_rel_mcount(relhdr, shdr0, shstrtab, fname); +} + + +static unsigned tot_relsize(Elf_Shdr const *const shdr0, + unsigned nhdr, + const char *const shstrtab, + const char *const fname) +{ + unsigned totrelsz = 0; + Elf_Shdr const *shdrp = shdr0; + + for (; nhdr; --nhdr, ++shdrp) { + if (has_rel_mcount(shdrp, shdr0, shstrtab, fname)) + totrelsz += _w(shdrp->sh_size); + } + return totrelsz; +} + + +/* Overall supervision for Elf32 ET_REL file. */ +static void +do_func(Elf_Ehdr *const ehdr, char const *const fname, unsigned const reltype) +{ + Elf_Shdr *const shdr0 = (Elf_Shdr *)(_w(ehdr->e_shoff) + + (void *)ehdr); + unsigned const nhdr = w2(ehdr->e_shnum); + Elf_Shdr *const shstr = &shdr0[w2(ehdr->e_shstrndx)]; + char const *const shstrtab = (char const *)(_w(shstr->sh_offset) + + (void *)ehdr); + + Elf_Shdr const *relhdr; + unsigned k; + + /* Upper bound on space: assume all relevant relocs are for mcount. */ + unsigned const totrelsz = tot_relsize(shdr0, nhdr, shstrtab, fname); + Elf_Rel *const mrel0 = umalloc(totrelsz); + Elf_Rel * mrelp = mrel0; + + /* 2*sizeof(address) <= sizeof(Elf_Rel) */ + uint_t *const mloc0 = umalloc(totrelsz>>1); + uint_t * mlocp = mloc0; + + unsigned rel_entsize = 0; + unsigned symsec_sh_link = 0; + + for (relhdr = shdr0, k = nhdr; k; --k, ++relhdr) { + char const *const txtname = has_rel_mcount(relhdr, shdr0, + shstrtab, fname); + if (txtname) { + uint_t recval = 0; + unsigned const recsym = find_secsym_ndx( + w(relhdr->sh_info), txtname, &recval, + &shdr0[symsec_sh_link = w(relhdr->sh_link)], + ehdr); + + rel_entsize = _w(relhdr->sh_entsize); + mlocp = sift_rel_mcount(mlocp, + (void *)mlocp - (void *)mloc0, &mrelp, + relhdr, ehdr, recsym, recval, reltype); + } + } + if (mloc0 != mlocp) { + append_func(ehdr, shstr, mloc0, mlocp, mrel0, mrelp, + rel_entsize, symsec_sh_link); + } + free(mrel0); + free(mloc0); +} diff --git a/tools/perf/Documentation/perf-annotate.txt b/tools/perf/Documentation/perf-annotate.txt index 5164a65..b2c6330 100644 --- a/tools/perf/Documentation/perf-annotate.txt +++ b/tools/perf/Documentation/perf-annotate.txt @@ -8,7 +8,7 @@ perf-annotate - Read perf.data (created by perf record) and display annotated co SYNOPSIS -------- [verse] -'perf annotate' [-i <file> | --input=file] symbol_name +'perf annotate' [-i <file> | --input=file] [symbol_name] DESCRIPTION ----------- @@ -24,6 +24,13 @@ OPTIONS --input=:: Input file name. (default: perf.data) +--stdio:: Use the stdio interface. + +--tui:: Use the TUI interface Use of --tui requires a tty, if one is not + present, as when piping to other commands, the stdio interface is + used. This interfaces starts by centering on the line with more + samples, TAB/UNTAB cycles thru the lines with more samples. + SEE ALSO -------- -linkperf:perf-record[1] +linkperf:perf-record[1], linkperf:perf-report[1] diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt index abfabe9..12052c9 100644 --- a/tools/perf/Documentation/perf-report.txt +++ b/tools/perf/Documentation/perf-report.txt @@ -65,6 +65,13 @@ OPTIONS the tree is considered as a new profiled object. + Default: fractal,0.5. +--stdio:: Use the stdio interface. + +--tui:: Use the TUI interface, that is integrated with annotate and allows + zooming into DSOs or threads, among other features. Use of --tui + requires a tty, if one is not present, as when piping to other + commands, the stdio interface is used. + SEE ALSO -------- linkperf:perf-stat[1] diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 1950e19..d1db0f6 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -313,6 +313,9 @@ TEST_PROGRAMS = SCRIPT_SH += perf-archive.sh +grep-libs = $(filter -l%,$(1)) +strip-libs = $(filter-out -l%,$(1)) + # # No Perl scripts right now: # @@ -588,14 +591,17 @@ endif ifdef NO_LIBPERL BASIC_CFLAGS += -DNO_LIBPERL else - PERL_EMBED_LDOPTS = `perl -MExtUtils::Embed -e ldopts 2>/dev/null` + PERL_EMBED_LDOPTS = $(shell perl -MExtUtils::Embed -e ldopts 2>/dev/null) + PERL_EMBED_LDFLAGS = $(call strip-libs,$(PERL_EMBED_LDOPTS)) + PERL_EMBED_LIBADD = $(call grep-libs,$(PERL_EMBED_LDOPTS)) PERL_EMBED_CCOPTS = `perl -MExtUtils::Embed -e ccopts 2>/dev/null` FLAGS_PERL_EMBED=$(PERL_EMBED_CCOPTS) $(PERL_EMBED_LDOPTS) ifneq ($(call try-cc,$(SOURCE_PERL_EMBED),$(FLAGS_PERL_EMBED)),y) BASIC_CFLAGS += -DNO_LIBPERL else - ALL_LDFLAGS += $(PERL_EMBED_LDOPTS) + ALL_LDFLAGS += $(PERL_EMBED_LDFLAGS) + EXTLIBS += $(PERL_EMBED_LIBADD) LIB_OBJS += $(OUTPUT)util/scripting-engines/trace-event-perl.o LIB_OBJS += $(OUTPUT)scripts/perl/Perf-Trace-Util/Context.o endif @@ -604,13 +610,16 @@ endif ifdef NO_LIBPYTHON BASIC_CFLAGS += -DNO_LIBPYTHON else - PYTHON_EMBED_LDOPTS = `python-config --ldflags 2>/dev/null` + PYTHON_EMBED_LDOPTS = $(shell python-config --ldflags 2>/dev/null) + PYTHON_EMBED_LDFLAGS = $(call strip-libs,$(PYTHON_EMBED_LDOPTS)) + PYTHON_EMBED_LIBADD = $(call grep-libs,$(PYTHON_EMBED_LDOPTS)) PYTHON_EMBED_CCOPTS = `python-config --cflags 2>/dev/null` FLAGS_PYTHON_EMBED=$(PYTHON_EMBED_CCOPTS) $(PYTHON_EMBED_LDOPTS) ifneq ($(call try-cc,$(SOURCE_PYTHON_EMBED),$(FLAGS_PYTHON_EMBED)),y) BASIC_CFLAGS += -DNO_LIBPYTHON else - ALL_LDFLAGS += $(PYTHON_EMBED_LDOPTS) + ALL_LDFLAGS += $(PYTHON_EMBED_LDFLAGS) + EXTLIBS += $(PYTHON_EMBED_LIBADD) LIB_OBJS += $(OUTPUT)util/scripting-engines/trace-event-python.o LIB_OBJS += $(OUTPUT)scripts/python/Perf-Trace-Util/Context.o endif @@ -653,6 +662,15 @@ else endif endif + +ifdef NO_STRLCPY + BASIC_CFLAGS += -DNO_STRLCPY +else + ifneq ($(call try-cc,$(SOURCE_STRLCPY),),y) + BASIC_CFLAGS += -DNO_STRLCPY + endif +endif + ifndef CC_LD_DYNPATH ifdef NO_R_TO_GCC_LINKER # Some gcc does not accept and pass -R to the linker to specify @@ -910,8 +928,8 @@ $(OUTPUT)perf.o: perf.c $(OUTPUT)common-cmds.h $(OUTPUT)PERF-CFLAGS $(ALL_CFLAGS) -c $(filter %.c,$^) -o $@ $(OUTPUT)perf$X: $(OUTPUT)perf.o $(BUILTIN_OBJS) $(PERFLIBS) - $(QUIET_LINK)$(CC) $(ALL_CFLAGS) -o $@ $(OUTPUT)perf.o \ - $(BUILTIN_OBJS) $(ALL_LDFLAGS) $(LIBS) + $(QUIET_LINK)$(CC) $(ALL_CFLAGS) $(ALL_LDFLAGS) $(OUTPUT)perf.o \ + $(BUILTIN_OBJS) $(LIBS) -o $@ $(OUTPUT)builtin-help.o: builtin-help.c $(OUTPUT)common-cmds.h $(OUTPUT)PERF-CFLAGS $(QUIET_CC)$(CC) -o $@ -c $(ALL_CFLAGS) \ diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 1478dc6..6d5604d8d 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -28,7 +28,7 @@ static char const *input_name = "perf.data"; -static bool force; +static bool force, use_tui, use_stdio; static bool full_paths; @@ -321,7 +321,7 @@ static int hist_entry__tty_annotate(struct hist_entry *he) static void hists__find_annotations(struct hists *self) { - struct rb_node *first = rb_first(&self->entries), *nd = first; + struct rb_node *nd = rb_first(&self->entries), *next; int key = KEY_RIGHT; while (nd) { @@ -343,20 +343,19 @@ find_next: if (use_browser > 0) { key = hist_entry__tui_annotate(he); - if (is_exit_key(key)) - break; switch (key) { case KEY_RIGHT: - case '\t': - nd = rb_next(nd); + next = rb_next(nd); break; case KEY_LEFT: - if (nd == first) - continue; - nd = rb_prev(nd); - default: + next = rb_prev(nd); break; + default: + return; } + + if (next != NULL) + nd = next; } else { hist_entry__tty_annotate(he); nd = rb_next(nd); @@ -428,6 +427,8 @@ 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_BOOLEAN(0, "tui", &use_tui, "Use the TUI interface"), + OPT_BOOLEAN(0, "stdio", &use_stdio, "Use the stdio interface"), OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, "file", "vmlinux pathname"), OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules, @@ -443,6 +444,11 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __used) { argc = parse_options(argc, argv, options, annotate_usage, 0); + if (use_stdio) + use_browser = 0; + else if (use_tui) + use_browser = 1; + setup_browser(); symbol_conf.priv_size = sizeof(struct sym_priv); diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 55fc1f4..5de405d 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -32,7 +32,7 @@ static char const *input_name = "perf.data"; -static bool force; +static bool force, use_tui, use_stdio; static bool hide_unresolved; static bool dont_use_callchains; @@ -107,7 +107,8 @@ static int perf_session__add_hist_entry(struct perf_session *self, goto out_free_syms; err = 0; if (symbol_conf.use_callchain) { - err = append_chain(he->callchain, data->callchain, syms, data->period); + err = callchain_append(he->callchain, data->callchain, syms, + data->period); if (err) goto out_free_syms; } @@ -450,6 +451,8 @@ static const struct option options[] = { "Show per-thread event counters"), OPT_STRING(0, "pretty", &pretty_printing_style, "key", "pretty printing style key: normal raw"), + OPT_BOOLEAN(0, "tui", &use_tui, "Use the TUI interface"), + OPT_BOOLEAN(0, "stdio", &use_stdio, "Use the stdio interface"), OPT_STRING('s', "sort", &sort_order, "key[,key2...]", "sort by key(s): pid, comm, dso, symbol, parent"), OPT_BOOLEAN(0, "showcpuutilization", &symbol_conf.show_cpu_utilization, @@ -482,8 +485,15 @@ int cmd_report(int argc, const char **argv, const char *prefix __used) { argc = parse_options(argc, argv, options, report_usage, 0); + if (use_stdio) + use_browser = 0; + else if (use_tui) + use_browser = 1; + if (strcmp(input_name, "-") != 0) setup_browser(); + else + use_browser = 0; /* * Only in the newt browser we are doing integrated annotation, * so don't allocate extra space that won't be used in the stdio diff --git a/tools/perf/feature-tests.mak b/tools/perf/feature-tests.mak index 7a7b608..b253db6 100644 --- a/tools/perf/feature-tests.mak +++ b/tools/perf/feature-tests.mak @@ -110,6 +110,17 @@ int main(void) } endef +define SOURCE_STRLCPY +#include <stdlib.h> +extern size_t strlcpy(char *dest, const char *src, size_t size); + +int main(void) +{ + strlcpy(NULL, NULL, 0); + return 0; +} +endef + # try-cc # Usage: option = $(call try-cc, source-to-build, cc-options) try-cc = $(shell sh -c \ diff --git a/tools/perf/scripts/python/bin/netdev-times-record b/tools/perf/scripts/python/bin/netdev-times-record new file mode 100644 index 0000000..d931a82 --- /dev/null +++ b/tools/perf/scripts/python/bin/netdev-times-record @@ -0,0 +1,8 @@ +#!/bin/bash +perf record -a -e net:net_dev_xmit -e net:net_dev_queue \ + -e net:netif_receive_skb -e net:netif_rx \ + -e skb:consume_skb -e skb:kfree_skb \ + -e skb:skb_copy_datagram_iovec -e napi:napi_poll \ + -e irq:irq_handler_entry -e irq:irq_handler_exit \ + -e irq:softirq_entry -e irq:softirq_exit \ + -e irq:softirq_raise $@ diff --git a/tools/perf/scripts/python/bin/netdev-times-report b/tools/perf/scripts/python/bin/netdev-times-report new file mode 100644 index 0000000..c3d0a63 --- /dev/null +++ b/tools/perf/scripts/python/bin/netdev-times-report @@ -0,0 +1,5 @@ +#!/bin/bash +# description: display a process of packet and processing time +# args: [tx] [rx] [dev=] [debug] + +perf trace -s ~/libexec/perf-core/scripts/python/netdev-times.py $@ diff --git a/tools/perf/scripts/python/netdev-times.py b/tools/perf/scripts/python/netdev-times.py new file mode 100644 index 0000000..9aa0a32 --- /dev/null +++ b/tools/perf/scripts/python/netdev-times.py @@ -0,0 +1,464 @@ +# Display a process of packets and processed time. +# It helps us to investigate networking or network device. +# +# options +# tx: show only tx chart +# rx: show only rx chart +# dev=: show only thing related to specified device +# debug: work with debug mode. It shows buffer status. + +import os +import sys + +sys.path.append(os.environ['PERF_EXEC_PATH'] + \ + '/scripts/python/Perf-Trace-Util/lib/Perf/Trace') + +from perf_trace_context import * +from Core import * +from Util import * + +all_event_list = []; # insert all tracepoint event related with this script +irq_dic = {}; # key is cpu and value is a list which stacks irqs + # which raise NET_RX softirq +net_rx_dic = {}; # key is cpu and value include time of NET_RX softirq-entry + # and a list which stacks receive +receive_hunk_list = []; # a list which include a sequence of receive events +rx_skb_list = []; # received packet list for matching + # skb_copy_datagram_iovec + +buffer_budget = 65536; # the budget of rx_skb_list, tx_queue_list and + # tx_xmit_list +of_count_rx_skb_list = 0; # overflow count + +tx_queue_list = []; # list of packets which pass through dev_queue_xmit +of_count_tx_queue_list = 0; # overflow count + +tx_xmit_list = []; # list of packets which pass through dev_hard_start_xmit +of_count_tx_xmit_list = 0; # overflow count + +tx_free_list = []; # list of packets which is freed + +# options +show_tx = 0; +show_rx = 0; +dev = 0; # store a name of device specified by option "dev=" +debug = 0; + +# indices of event_info tuple +EINFO_IDX_NAME= 0 +EINFO_IDX_CONTEXT=1 +EINFO_IDX_CPU= 2 +EINFO_IDX_TIME= 3 +EINFO_IDX_PID= 4 +EINFO_IDX_COMM= 5 + +# Calculate a time interval(msec) from src(nsec) to dst(nsec) +def diff_msec(src, dst): + return (dst - src) / 1000000.0 + +# Display a process of transmitting a packet +def print_transmit(hunk): + if dev != 0 and hunk['dev'].find(dev) < 0: + return + print "%7s %5d %6d.%06dsec %12.3fmsec %12.3fmsec" % \ + (hunk['dev'], hunk['len'], + nsecs_secs(hunk['queue_t']), + nsecs_nsecs(hunk['queue_t'])/1000, + diff_msec(hunk['queue_t'], hunk['xmit_t']), + diff_msec(hunk['xmit_t'], hunk['free_t'])) + +# Format for displaying rx packet processing +PF_IRQ_ENTRY= " irq_entry(+%.3fmsec irq=%d:%s)" +PF_SOFT_ENTRY=" softirq_entry(+%.3fmsec)" +PF_NAPI_POLL= " napi_poll_exit(+%.3fmsec %s)" +PF_JOINT= " |" +PF_WJOINT= " | |" +PF_NET_RECV= " |---netif_receive_skb(+%.3fmsec skb=%x len=%d)" +PF_NET_RX= " |---netif_rx(+%.3fmsec skb=%x)" +PF_CPY_DGRAM= " | skb_copy_datagram_iovec(+%.3fmsec %d:%s)" +PF_KFREE_SKB= " | kfree_skb(+%.3fmsec location=%x)" +PF_CONS_SKB= " | consume_skb(+%.3fmsec)" + +# Display a process of received packets and interrputs associated with +# a NET_RX softirq +def print_receive(hunk): + show_hunk = 0 + irq_list = hunk['irq_list'] + cpu = irq_list[0]['cpu'] + base_t = irq_list[0]['irq_ent_t'] + # check if this hunk should be showed + if dev != 0: + for i in range(len(irq_list)): + if irq_list[i]['name'].find(dev) >= 0: + show_hunk = 1 + break + else: + show_hunk = 1 + if show_hunk == 0: + return + + print "%d.%06dsec cpu=%d" % \ + (nsecs_secs(base_t), nsecs_nsecs(base_t)/1000, cpu) + for i in range(len(irq_list)): + print PF_IRQ_ENTRY % \ + (diff_msec(base_t, irq_list[i]['irq_ent_t']), + irq_list[i]['irq'], irq_list[i]['name']) + print PF_JOINT + irq_event_list = irq_list[i]['event_list'] + for j in range(len(irq_event_list)): + irq_event = irq_event_list[j] + if irq_event['event'] == 'netif_rx': + print PF_NET_RX % \ + (diff_msec(base_t, irq_event['time']), + irq_event['skbaddr']) + print PF_JOINT + print PF_SOFT_ENTRY % \ + diff_msec(base_t, hunk['sirq_ent_t']) + print PF_JOINT + event_list = hunk['event_list'] + for i in range(len(event_list)): + event = event_list[i] + if event['event_name'] == 'napi_poll': + print PF_NAPI_POLL % \ + (diff_msec(base_t, event['event_t']), event['dev']) + if i == len(event_list) - 1: + print "" + else: + print PF_JOINT + else: + print PF_NET_RECV % \ + (diff_msec(base_t, event['event_t']), event['skbaddr'], + event['len']) + if 'comm' in event.keys(): + print PF_WJOINT + print PF_CPY_DGRAM % \ + (diff_msec(base_t, event['comm_t']), + event['pid'], event['comm']) + elif 'handle' in event.keys(): + print PF_WJOINT + if event['handle'] == "kfree_skb": + print PF_KFREE_SKB % \ + (diff_msec(base_t, + event['comm_t']), + event['location']) + elif event['handle'] == "consume_skb": + print PF_CONS_SKB % \ + diff_msec(base_t, + event['comm_t']) + print PF_JOINT + +def trace_begin(): + global show_tx + global show_rx + global dev + global debug + + for i in range(len(sys.argv)): + if i == 0: + continue + arg = sys.argv[i] + if arg == 'tx': + show_tx = 1 + elif arg =='rx': + show_rx = 1 + elif arg.find('dev=',0, 4) >= 0: + dev = arg[4:] + elif arg == 'debug': + debug = 1 + if show_tx == 0 and show_rx == 0: + show_tx = 1 + show_rx = 1 + +def trace_end(): + # order all events in time + all_event_list.sort(lambda a,b :cmp(a[EINFO_IDX_TIME], + b[EINFO_IDX_TIME])) + # process all events + for i in range(len(all_event_list)): + event_info = all_event_list[i] + name = event_info[EINFO_IDX_NAME] + if name == 'irq__softirq_exit': + handle_irq_softirq_exit(event_info) + elif name == 'irq__softirq_entry': + handle_irq_softirq_entry(event_info) + elif name == 'irq__softirq_raise': + handle_irq_softirq_raise(event_info) + elif name == 'irq__irq_handler_entry': + handle_irq_handler_entry(event_info) + elif name == 'irq__irq_handler_exit': + handle_irq_handler_exit(event_info) + elif name == 'napi__napi_poll': + handle_napi_poll(event_info) + elif name == 'net__netif_receive_skb': + handle_netif_receive_skb(event_info) + elif name == 'net__netif_rx': + handle_netif_rx(event_info) + elif name == 'skb__skb_copy_datagram_iovec': + handle_skb_copy_datagram_iovec(event_info) + elif name == 'net__net_dev_queue': + handle_net_dev_queue(event_info) + elif name == 'net__net_dev_xmit': + handle_net_dev_xmit(event_info) + elif name == 'skb__kfree_skb': + handle_kfree_skb(event_info) + elif name == 'skb__consume_skb': + handle_consume_skb(event_info) + # display receive hunks + if show_rx: + for i in range(len(receive_hunk_list)): + print_receive(receive_hunk_list[i]) + # display transmit hunks + if show_tx: + print " dev len Qdisc " \ + " netdevice free" + for i in range(len(tx_free_list)): + print_transmit(tx_free_list[i]) + if debug: + print "debug buffer status" + print "----------------------------" + print "xmit Qdisc:remain:%d overflow:%d" % \ + (len(tx_queue_list), of_count_tx_queue_list) + print "xmit netdevice:remain:%d overflow:%d" % \ + (len(tx_xmit_list), of_count_tx_xmit_list) + print "receive:remain:%d overflow:%d" % \ + (len(rx_skb_list), of_count_rx_skb_list) + +# called from perf, when it finds a correspoinding event +def irq__softirq_entry(name, context, cpu, sec, nsec, pid, comm, vec): + if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX": + return + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec) + all_event_list.append(event_info) + +def irq__softirq_exit(name, context, cpu, sec, nsec, pid, comm, vec): + if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX": + return + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec) + all_event_list.append(event_info) + +def irq__softirq_raise(name, context, cpu, sec, nsec, pid, comm, vec): + if symbol_str("irq__softirq_entry", "vec", vec) != "NET_RX": + return + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, vec) + all_event_list.append(event_info) + +def irq__irq_handler_entry(name, context, cpu, sec, nsec, pid, comm, + irq, irq_name): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + irq, irq_name) + all_event_list.append(event_info) + +def irq__irq_handler_exit(name, context, cpu, sec, nsec, pid, comm, irq, ret): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, irq, ret) + all_event_list.append(event_info) + +def napi__napi_poll(name, context, cpu, sec, nsec, pid, comm, napi, dev_name): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + napi, dev_name) + all_event_list.append(event_info) + +def net__netif_receive_skb(name, context, cpu, sec, nsec, pid, comm, skbaddr, + skblen, dev_name): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + skbaddr, skblen, dev_name) + all_event_list.append(event_info) + +def net__netif_rx(name, context, cpu, sec, nsec, pid, comm, skbaddr, + skblen, dev_name): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + skbaddr, skblen, dev_name) + all_event_list.append(event_info) + +def net__net_dev_queue(name, context, cpu, sec, nsec, pid, comm, + skbaddr, skblen, dev_name): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + skbaddr, skblen, dev_name) + all_event_list.append(event_info) + +def net__net_dev_xmit(name, context, cpu, sec, nsec, pid, comm, + skbaddr, skblen, rc, dev_name): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + skbaddr, skblen, rc ,dev_name) + all_event_list.append(event_info) + +def skb__kfree_skb(name, context, cpu, sec, nsec, pid, comm, + skbaddr, protocol, location): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + skbaddr, protocol, location) + all_event_list.append(event_info) + +def skb__consume_skb(name, context, cpu, sec, nsec, pid, comm, skbaddr): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + skbaddr) + all_event_list.append(event_info) + +def skb__skb_copy_datagram_iovec(name, context, cpu, sec, nsec, pid, comm, + skbaddr, skblen): + event_info = (name, context, cpu, nsecs(sec, nsec), pid, comm, + skbaddr, skblen) + all_event_list.append(event_info) + +def handle_irq_handler_entry(event_info): + (name, context, cpu, time, pid, comm, irq, irq_name) = event_info + if cpu not in irq_dic.keys(): + irq_dic[cpu] = [] + irq_record = {'irq':irq, 'name':irq_name, 'cpu':cpu, 'irq_ent_t':time} + irq_dic[cpu].append(irq_record) + +def handle_irq_handler_exit(event_info): + (name, context, cpu, time, pid, comm, irq, ret) = event_info + if cpu not in irq_dic.keys(): + return + irq_record = irq_dic[cpu].pop() + if irq != irq_record['irq']: + return + irq_record.update({'irq_ext_t':time}) + # if an irq doesn't include NET_RX softirq, drop. + if 'event_list' in irq_record.keys(): + irq_dic[cpu].append(irq_record) + +def handle_irq_softirq_raise(event_info): + (name, context, cpu, time, pid, comm, vec) = event_info + if cpu not in irq_dic.keys() \ + or len(irq_dic[cpu]) == 0: + return + irq_record = irq_dic[cpu].pop() + if 'event_list' in irq_record.keys(): + irq_event_list = irq_record['event_list'] + else: + irq_event_list = [] + irq_event_list.append({'time':time, 'event':'sirq_raise'}) + irq_record.update({'event_list':irq_event_list}) + irq_dic[cpu].append(irq_record) + +def handle_irq_softirq_entry(event_info): + (name, context, cpu, time, pid, comm, vec) = event_info + net_rx_dic[cpu] = {'sirq_ent_t':time, 'event_list':[]} + +def handle_irq_softirq_exit(event_info): + (name, context, cpu, time, pid, comm, vec) = event_info + irq_list = [] + event_list = 0 + if cpu in irq_dic.keys(): + irq_list = irq_dic[cpu] + del irq_dic[cpu] + if cpu in net_rx_dic.keys(): + sirq_ent_t = net_rx_dic[cpu]['sirq_ent_t'] + event_list = net_rx_dic[cpu]['event_list'] + del net_rx_dic[cpu] + if irq_list == [] or event_list == 0: + return + rec_data = {'sirq_ent_t':sirq_ent_t, 'sirq_ext_t':time, + 'irq_list':irq_list, 'event_list':event_list} + # merge information realted to a NET_RX softirq + receive_hunk_list.append(rec_data) + +def handle_napi_poll(event_info): + (name, context, cpu, time, pid, comm, napi, dev_name) = event_info + if cpu in net_rx_dic.keys(): + event_list = net_rx_dic[cpu]['event_list'] + rec_data = {'event_name':'napi_poll', + 'dev':dev_name, 'event_t':time} + event_list.append(rec_data) + +def handle_netif_rx(event_info): + (name, context, cpu, time, pid, comm, + skbaddr, skblen, dev_name) = event_info + if cpu not in irq_dic.keys() \ + or len(irq_dic[cpu]) == 0: + return + irq_record = irq_dic[cpu].pop() + if 'event_list' in irq_record.keys(): + irq_event_list = irq_record['event_list'] + else: + irq_event_list = [] + irq_event_list.append({'time':time, 'event':'netif_rx', + 'skbaddr':skbaddr, 'skblen':skblen, 'dev_name':dev_name}) + irq_record.update({'event_list':irq_event_list}) + irq_dic[cpu].append(irq_record) + +def handle_netif_receive_skb(event_info): + global of_count_rx_skb_list + + (name, context, cpu, time, pid, comm, + skbaddr, skblen, dev_name) = event_info + if cpu in net_rx_dic.keys(): + rec_data = {'event_name':'netif_receive_skb', + 'event_t':time, 'skbaddr':skbaddr, 'len':skblen} + event_list = net_rx_dic[cpu]['event_list'] + event_list.append(rec_data) + rx_skb_list.insert(0, rec_data) + if len(rx_skb_list) > buffer_budget: + rx_skb_list.pop() + of_count_rx_skb_list += 1 + +def handle_net_dev_queue(event_info): + global of_count_tx_queue_list + + (name, context, cpu, time, pid, comm, + skbaddr, skblen, dev_name) = event_info + skb = {'dev':dev_name, 'skbaddr':skbaddr, 'len':skblen, 'queue_t':time} + tx_queue_list.insert(0, skb) + if len(tx_queue_list) > buffer_budget: + tx_queue_list.pop() + of_count_tx_queue_list += 1 + +def handle_net_dev_xmit(event_info): + global of_count_tx_xmit_list + + (name, context, cpu, time, pid, comm, + skbaddr, skblen, rc, dev_name) = event_info + if rc == 0: # NETDEV_TX_OK + for i in range(len(tx_queue_list)): + skb = tx_queue_list[i] + if skb['skbaddr'] == skbaddr: + skb['xmit_t'] = time + tx_xmit_list.insert(0, skb) + del tx_queue_list[i] + if len(tx_xmit_list) > buffer_budget: + tx_xmit_list.pop() + of_count_tx_xmit_list += 1 + return + +def handle_kfree_skb(event_info): + (name, context, cpu, time, pid, comm, + skbaddr, protocol, location) = event_info + for i in range(len(tx_queue_list)): + skb = tx_queue_list[i] + if skb['skbaddr'] == skbaddr: + del tx_queue_list[i] + return + for i in range(len(tx_xmit_list)): + skb = tx_xmit_list[i] + if skb['skbaddr'] == skbaddr: + skb['free_t'] = time + tx_free_list.append(skb) + del tx_xmit_list[i] + return + for i in range(len(rx_skb_list)): + rec_data = rx_skb_list[i] + if rec_data['skbaddr'] == skbaddr: + rec_data.update({'handle':"kfree_skb", + 'comm':comm, 'pid':pid, 'comm_t':time}) + del rx_skb_list[i] + return + +def handle_consume_skb(event_info): + (name, context, cpu, time, pid, comm, skbaddr) = event_info + for i in range(len(tx_xmit_list)): + skb = tx_xmit_list[i] + if skb['skbaddr'] == skbaddr: + skb['free_t'] = time + tx_free_list.append(skb) + del tx_xmit_list[i] + return + +def handle_skb_copy_datagram_iovec(event_info): + (name, context, cpu, time, pid, comm, skbaddr, skblen) = event_info + for i in range(len(rx_skb_list)): + rec_data = rx_skb_list[i] + if skbaddr == rec_data['skbaddr']: + rec_data.update({'handle':"skb_copy_datagram_iovec", + 'comm':comm, 'pid':pid, 'comm_t':time}) + del rx_skb_list[i] + return diff --git a/tools/perf/util/cache.h b/tools/perf/util/cache.h index 27e9ebe..a772979 100644 --- a/tools/perf/util/cache.h +++ b/tools/perf/util/cache.h @@ -82,6 +82,8 @@ extern char *perf_path(const char *fmt, ...) __attribute__((format (printf, 1, 2 extern char *perf_pathdup(const char *fmt, ...) __attribute__((format (printf, 1, 2))); +#ifdef NO_STRLCPY extern size_t strlcpy(char *dest, const char *src, size_t size); +#endif #endif /* __PERF_CACHE_H */ diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index f231f43..e12d539 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -28,6 +28,9 @@ bool ip_callchain__valid(struct ip_callchain *chain, const event_t *event) #define chain_for_each_child(child, parent) \ list_for_each_entry(child, &parent->children, brothers) +#define chain_for_each_child_safe(child, next, parent) \ + list_for_each_entry_safe(child, next, &parent->children, brothers) + static void rb_insert_callchain(struct rb_root *root, struct callchain_node *chain, enum chain_mode mode) @@ -86,10 +89,10 @@ __sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node, * sort them by hit */ static void -sort_chain_flat(struct rb_root *rb_root, struct callchain_node *node, +sort_chain_flat(struct rb_root *rb_root, struct callchain_root *root, u64 min_hit, struct callchain_param *param __used) { - __sort_chain_flat(rb_root, node, min_hit); + __sort_chain_flat(rb_root, &root->node, min_hit); } static void __sort_chain_graph_abs(struct callchain_node *node, @@ -108,11 +111,11 @@ static void __sort_chain_graph_abs(struct callchain_node *node, } static void -sort_chain_graph_abs(struct rb_root *rb_root, struct callchain_node *chain_root, +sort_chain_graph_abs(struct rb_root *rb_root, struct callchain_root *chain_root, u64 min_hit, struct callchain_param *param __used) { - __sort_chain_graph_abs(chain_root, min_hit); - rb_root->rb_node = chain_root->rb_root.rb_node; + __sort_chain_graph_abs(&chain_root->node, min_hit); + rb_root->rb_node = chain_root->node.rb_root.rb_node; } static void __sort_chain_graph_rel(struct callchain_node *node, @@ -133,11 +136,11 @@ static void __sort_chain_graph_rel(struct callchain_node *node, } static void -sort_chain_graph_rel(struct rb_root *rb_root, struct callchain_node *chain_root, +sort_chain_graph_rel(struct rb_root *rb_root, struct callchain_root *chain_root, u64 min_hit __used, struct callchain_param *param) { - __sort_chain_graph_rel(chain_root, param->min_percent / 100.0); - rb_root->rb_node = chain_root->rb_root.rb_node; + __sort_chain_graph_rel(&chain_root->node, param->min_percent / 100.0); + rb_root->rb_node = chain_root->node.rb_root.rb_node; } int register_callchain_param(struct callchain_param *param) @@ -284,19 +287,18 @@ split_add_child(struct callchain_node *parent, struct resolved_chain *chain, } static int -__append_chain(struct callchain_node *root, struct resolved_chain *chain, - unsigned int start, u64 period); +append_chain(struct callchain_node *root, struct resolved_chain *chain, + unsigned int start, u64 period); static void -__append_chain_children(struct callchain_node *root, - struct resolved_chain *chain, - unsigned int start, u64 period) +append_chain_children(struct callchain_node *root, struct resolved_chain *chain, + unsigned int start, u64 period) { struct callchain_node *rnode; /* lookup in childrens */ chain_for_each_child(rnode, root) { - unsigned int ret = __append_chain(rnode, chain, start, period); + unsigned int ret = append_chain(rnode, chain, start, period); if (!ret) goto inc_children_hit; @@ -309,8 +311,8 @@ inc_children_hit: } static int -__append_chain(struct callchain_node *root, struct resolved_chain *chain, - unsigned int start, u64 period) +append_chain(struct callchain_node *root, struct resolved_chain *chain, + unsigned int start, u64 period) { struct callchain_list *cnode; unsigned int i = start; @@ -357,7 +359,7 @@ __append_chain(struct callchain_node *root, struct resolved_chain *chain, } /* We match the node and still have a part remaining */ - __append_chain_children(root, chain, i, period); + append_chain_children(root, chain, i, period); return 0; } @@ -380,8 +382,8 @@ static void filter_context(struct ip_callchain *old, struct resolved_chain *new, } -int append_chain(struct callchain_node *root, struct ip_callchain *chain, - struct map_symbol *syms, u64 period) +int callchain_append(struct callchain_root *root, struct ip_callchain *chain, + struct map_symbol *syms, u64 period) { struct resolved_chain *filtered; @@ -398,9 +400,65 @@ int append_chain(struct callchain_node *root, struct ip_callchain *chain, if (!filtered->nr) goto end; - __append_chain_children(root, filtered, 0, period); + append_chain_children(&root->node, filtered, 0, period); + + if (filtered->nr > root->max_depth) + root->max_depth = filtered->nr; end: free(filtered); return 0; } + +static int +merge_chain_branch(struct callchain_node *dst, struct callchain_node *src, + struct resolved_chain *chain) +{ + struct callchain_node *child, *next_child; + struct callchain_list *list, *next_list; + int old_pos = chain->nr; + int err = 0; + + list_for_each_entry_safe(list, next_list, &src->val, list) { + chain->ips[chain->nr].ip = list->ip; + chain->ips[chain->nr].ms = list->ms; + chain->nr++; + list_del(&list->list); + free(list); + } + + if (src->hit) + append_chain_children(dst, chain, 0, src->hit); + + chain_for_each_child_safe(child, next_child, src) { + err = merge_chain_branch(dst, child, chain); + if (err) + break; + + list_del(&child->brothers); + free(child); + } + + chain->nr = old_pos; + + return err; +} + +int callchain_merge(struct callchain_root *dst, struct callchain_root *src) +{ + struct resolved_chain *chain; + int err; + + chain = malloc(sizeof(*chain) + + src->max_depth * sizeof(struct resolved_ip)); + if (!chain) + return -ENOMEM; + + chain->nr = 0; + + err = merge_chain_branch(&dst->node, &src->node, chain); + + free(chain); + + return err; +} diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index 6de4313..c15fb8c 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -26,9 +26,14 @@ struct callchain_node { u64 children_hit; }; +struct callchain_root { + u64 max_depth; + struct callchain_node node; +}; + struct callchain_param; -typedef void (*sort_chain_func_t)(struct rb_root *, struct callchain_node *, +typedef void (*sort_chain_func_t)(struct rb_root *, struct callchain_root *, u64, struct callchain_param *); struct callchain_param { @@ -44,15 +49,16 @@ struct callchain_list { struct list_head list; }; -static inline void callchain_init(struct callchain_node *node) +static inline void callchain_init(struct callchain_root *root) { - INIT_LIST_HEAD(&node->brothers); - INIT_LIST_HEAD(&node->children); - INIT_LIST_HEAD(&node->val); + INIT_LIST_HEAD(&root->node.brothers); + INIT_LIST_HEAD(&root->node.children); + INIT_LIST_HEAD(&root->node.val); - node->children_hit = 0; - node->parent = NULL; - node->hit = 0; + root->node.parent = NULL; + root->node.hit = 0; + root->node.children_hit = 0; + root->max_depth = 0; } static inline u64 cumul_hits(struct callchain_node *node) @@ -61,8 +67,9 @@ static inline u64 cumul_hits(struct callchain_node *node) } int register_callchain_param(struct callchain_param *param); -int append_chain(struct callchain_node *root, struct ip_callchain *chain, - struct map_symbol *syms, u64 period); +int callchain_append(struct callchain_root *root, struct ip_callchain *chain, + struct map_symbol *syms, u64 period); +int callchain_merge(struct callchain_root *dst, struct callchain_root *src); bool ip_callchain__valid(struct ip_callchain *chain, const event_t *event); #endif /* __PERF_CALLCHAIN_H */ diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index be22ae6..2022e87 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -87,7 +87,7 @@ static void hist_entry__add_cpumode_period(struct hist_entry *self, static struct hist_entry *hist_entry__new(struct hist_entry *template) { - size_t callchain_size = symbol_conf.use_callchain ? sizeof(struct callchain_node) : 0; + size_t callchain_size = symbol_conf.use_callchain ? sizeof(struct callchain_root) : 0; struct hist_entry *self = malloc(sizeof(*self) + callchain_size); if (self != NULL) { @@ -226,6 +226,8 @@ static bool collapse__insert_entry(struct rb_root *root, struct hist_entry *he) if (!cmp) { iter->period += he->period; + if (symbol_conf.use_callchain) + callchain_merge(iter->callchain, he->callchain); hist_entry__free(he); return false; } diff --git a/tools/perf/util/path.c b/tools/perf/util/path.c index 58a470d..bd74977 100644 --- a/tools/perf/util/path.c +++ b/tools/perf/util/path.c @@ -22,6 +22,7 @@ static const char *get_perf_dir(void) return "."; } +#ifdef NO_STRLCPY size_t strlcpy(char *dest, const char *src, size_t size) { size_t ret = strlen(src); @@ -33,7 +34,7 @@ size_t strlcpy(char *dest, const char *src, size_t size) } return ret; } - +#endif static char *get_pathname(void) { diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h index 46e531d..0b91053 100644 --- a/tools/perf/util/sort.h +++ b/tools/perf/util/sort.h @@ -70,7 +70,7 @@ struct hist_entry { struct hist_entry *pair; struct rb_root sorted_chain; }; - struct callchain_node callchain[0]; + struct callchain_root callchain[0]; }; enum sort_type { diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index b2f5ae9..b39f499 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -388,6 +388,20 @@ size_t dso__fprintf_buildid(struct dso *self, FILE *fp) return fprintf(fp, "%s", sbuild_id); } +size_t dso__fprintf_symbols_by_name(struct dso *self, enum map_type type, FILE *fp) +{ + size_t ret = 0; + struct rb_node *nd; + struct symbol_name_rb_node *pos; + + for (nd = rb_first(&self->symbol_names[type]); nd; nd = rb_next(nd)) { + pos = rb_entry(nd, struct symbol_name_rb_node, rb_node); + fprintf(fp, "%s\n", pos->sym.name); + } + + return ret; +} + size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp) { struct rb_node *nd; diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index ea95c27..038f220 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -182,6 +182,7 @@ size_t machines__fprintf_dsos(struct rb_root *self, FILE *fp); size_t machines__fprintf_dsos_buildid(struct rb_root *self, FILE *fp, bool with_hits); size_t dso__fprintf_buildid(struct dso *self, FILE *fp); +size_t dso__fprintf_symbols_by_name(struct dso *self, enum map_type type, FILE *fp); size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp); enum dso_origin { diff --git a/tools/perf/util/ui/browser.c b/tools/perf/util/ui/browser.c index 66f2d58..6d0df80 100644 --- a/tools/perf/util/ui/browser.c +++ b/tools/perf/util/ui/browser.c @@ -1,16 +1,6 @@ -#define _GNU_SOURCE -#include <stdio.h> -#undef _GNU_SOURCE -/* - * slang versions <= 2.0.6 have a "#if HAVE_LONG_LONG" that breaks - * the build if it isn't defined. Use the equivalent one that glibc - * has on features.h. - */ -#include <features.h> -#ifndef HAVE_LONG_LONG -#define HAVE_LONG_LONG __GLIBC_HAVE_LONG_LONG -#endif #include <slang.h> +#include "libslang.h" +#include <linux/compiler.h> #include <linux/list.h> #include <linux/rbtree.h> #include <stdlib.h> @@ -19,17 +9,9 @@ #include "helpline.h" #include "../color.h" #include "../util.h" +#include <stdio.h> -#if SLANG_VERSION < 20104 -#define sltt_set_color(obj, name, fg, bg) \ - SLtt_set_color(obj,(char *)name, (char *)fg, (char *)bg) -#else -#define sltt_set_color SLtt_set_color -#endif - -newtComponent newt_form__new(void); - -int ui_browser__percent_color(double percent, bool current) +static int ui_browser__percent_color(double percent, bool current) { if (current) return HE_COLORSET_SELECTED; @@ -40,6 +22,23 @@ int ui_browser__percent_color(double percent, bool current) return HE_COLORSET_NORMAL; } +void ui_browser__set_color(struct ui_browser *self __used, int color) +{ + SLsmg_set_color(color); +} + +void ui_browser__set_percent_color(struct ui_browser *self, + double percent, bool current) +{ + int color = ui_browser__percent_color(percent, current); + ui_browser__set_color(self, color); +} + +void ui_browser__gotorc(struct ui_browser *self, int y, int x) +{ + SLsmg_gotorc(self->y + y, self->x + x); +} + void ui_browser__list_head_seek(struct ui_browser *self, off_t offset, int whence) { struct list_head *head = self->entries; @@ -111,7 +110,7 @@ unsigned int ui_browser__rb_tree_refresh(struct ui_browser *self) nd = self->top; while (nd != NULL) { - SLsmg_gotorc(self->y + row, self->x); + ui_browser__gotorc(self, row, 0); self->write(self, nd, row); if (++row == self->height) break; @@ -131,13 +130,10 @@ void ui_browser__refresh_dimensions(struct ui_browser *self) int cols, rows; newtGetScreenSize(&cols, &rows); - if (self->width > cols - 4) - self->width = cols - 4; - self->height = rows - 5; - if (self->height > self->nr_entries) - self->height = self->nr_entries; - self->y = (rows - self->height) / 2; - self->x = (cols - self->width) / 2; + self->width = cols - 1; + self->height = rows - 2; + self->y = 1; + self->x = 0; } void ui_browser__reset_index(struct ui_browser *self) @@ -146,34 +142,48 @@ void ui_browser__reset_index(struct ui_browser *self) self->seek(self, 0, SEEK_SET); } +void ui_browser__add_exit_key(struct ui_browser *self, int key) +{ + newtFormAddHotKey(self->form, key); +} + +void ui_browser__add_exit_keys(struct ui_browser *self, int keys[]) +{ + int i = 0; + + while (keys[i] && i < 64) { + ui_browser__add_exit_key(self, keys[i]); + ++i; + } +} + int ui_browser__show(struct ui_browser *self, const char *title, const char *helpline, ...) { va_list ap; + int keys[] = { NEWT_KEY_UP, NEWT_KEY_DOWN, NEWT_KEY_PGUP, + NEWT_KEY_PGDN, NEWT_KEY_HOME, NEWT_KEY_END, ' ', + NEWT_KEY_LEFT, NEWT_KEY_ESCAPE, 'q', CTRL('c'), 0 }; - if (self->form != NULL) { + if (self->form != NULL) newtFormDestroy(self->form); - newtPopWindow(); - } + ui_browser__refresh_dimensions(self); - newtCenteredWindow(self->width, self->height, title); - self->form = newt_form__new(); + self->form = newtForm(NULL, NULL, 0); if (self->form == NULL) return -1; - self->sb = newtVerticalScrollbar(self->width, 0, self->height, + self->sb = newtVerticalScrollbar(self->width, 1, self->height, HE_COLORSET_NORMAL, HE_COLORSET_SELECTED); if (self->sb == NULL) return -1; - newtFormAddHotKey(self->form, NEWT_KEY_UP); - newtFormAddHotKey(self->form, NEWT_KEY_DOWN); - newtFormAddHotKey(self->form, NEWT_KEY_PGUP); - newtFormAddHotKey(self->form, NEWT_KEY_PGDN); - newtFormAddHotKey(self->form, NEWT_KEY_HOME); - newtFormAddHotKey(self->form, NEWT_KEY_END); - newtFormAddHotKey(self->form, ' '); + SLsmg_gotorc(0, 0); + ui_browser__set_color(self, NEWT_COLORSET_ROOT); + slsmg_write_nstring(title, self->width); + + ui_browser__add_exit_keys(self, keys); newtFormAddComponent(self->form, self->sb); va_start(ap, helpline); @@ -185,7 +195,6 @@ int ui_browser__show(struct ui_browser *self, const char *title, void ui_browser__hide(struct ui_browser *self) { newtFormDestroy(self->form); - newtPopWindow(); self->form = NULL; ui_helpline__pop(); } @@ -196,28 +205,28 @@ int ui_browser__refresh(struct ui_browser *self) newtScrollbarSet(self->sb, self->index, self->nr_entries - 1); row = self->refresh(self); - SLsmg_set_color(HE_COLORSET_NORMAL); + ui_browser__set_color(self, HE_COLORSET_NORMAL); SLsmg_fill_region(self->y + row, self->x, self->height - row, self->width, ' '); return 0; } -int ui_browser__run(struct ui_browser *self, struct newtExitStruct *es) +int ui_browser__run(struct ui_browser *self) { + struct newtExitStruct es; + if (ui_browser__refresh(self) < 0) return -1; while (1) { off_t offset; - newtFormRun(self->form, es); + newtFormRun(self->form, &es); - if (es->reason != NEWT_EXIT_HOTKEY) + if (es.reason != NEWT_EXIT_HOTKEY) break; - if (is_exit_key(es->u.key)) - return es->u.key; - switch (es->u.key) { + switch (es.u.key) { case NEWT_KEY_DOWN: if (self->index == self->nr_entries - 1) break; @@ -274,12 +283,12 @@ int ui_browser__run(struct ui_browser *self, struct newtExitStruct *es) self->seek(self, -offset, SEEK_END); break; default: - return es->u.key; + return es.u.key; } if (ui_browser__refresh(self) < 0) return -1; } - return 0; + return -1; } unsigned int ui_browser__list_head_refresh(struct ui_browser *self) @@ -294,7 +303,7 @@ unsigned int ui_browser__list_head_refresh(struct ui_browser *self) pos = self->top; list_for_each_from(pos, head) { - SLsmg_gotorc(self->y + row, self->x); + ui_browser__gotorc(self, row, 0); self->write(self, pos, row); if (++row == self->height) break; diff --git a/tools/perf/util/ui/browser.h b/tools/perf/util/ui/browser.h index 0b9f829..0dc7e4d 100644 --- a/tools/perf/util/ui/browser.h +++ b/tools/perf/util/ui/browser.h @@ -25,16 +25,21 @@ struct ui_browser { }; -int ui_browser__percent_color(double percent, bool current); +void ui_browser__set_color(struct ui_browser *self, int color); +void ui_browser__set_percent_color(struct ui_browser *self, + double percent, bool current); bool ui_browser__is_current_entry(struct ui_browser *self, unsigned row); void ui_browser__refresh_dimensions(struct ui_browser *self); void ui_browser__reset_index(struct ui_browser *self); +void ui_browser__gotorc(struct ui_browser *self, int y, int x); +void ui_browser__add_exit_key(struct ui_browser *self, int key); +void ui_browser__add_exit_keys(struct ui_browser *self, int keys[]); int ui_browser__show(struct ui_browser *self, const char *title, const char *helpline, ...); void ui_browser__hide(struct ui_browser *self); int ui_browser__refresh(struct ui_browser *self); -int ui_browser__run(struct ui_browser *self, struct newtExitStruct *es); +int ui_browser__run(struct ui_browser *self); void ui_browser__rb_tree_seek(struct ui_browser *self, off_t offset, int whence); unsigned int ui_browser__rb_tree_refresh(struct ui_browser *self); diff --git a/tools/perf/util/ui/browsers/annotate.c b/tools/perf/util/ui/browsers/annotate.c index a90273e..82b78f9 100644 --- a/tools/perf/util/ui/browsers/annotate.c +++ b/tools/perf/util/ui/browsers/annotate.c @@ -40,14 +40,12 @@ static void annotate_browser__write(struct ui_browser *self, void *entry, int ro if (ol->offset != -1) { struct objdump_line_rb_node *olrb = objdump_line__rb(ol); - int color = ui_browser__percent_color(olrb->percent, current_entry); - SLsmg_set_color(color); + ui_browser__set_percent_color(self, olrb->percent, current_entry); slsmg_printf(" %7.2f ", olrb->percent); if (!current_entry) - SLsmg_set_color(HE_COLORSET_CODE); + ui_browser__set_color(self, HE_COLORSET_CODE); } else { - int color = ui_browser__percent_color(0, current_entry); - SLsmg_set_color(color); + ui_browser__set_percent_color(self, 0, current_entry); slsmg_write_nstring(" ", 9); } @@ -135,32 +133,31 @@ static void annotate_browser__set_top(struct annotate_browser *self, self->curr_hot = nd; } -static int annotate_browser__run(struct annotate_browser *self, - struct newtExitStruct *es) +static int annotate_browser__run(struct annotate_browser *self) { struct rb_node *nd; struct hist_entry *he = self->b.priv; + int key; if (ui_browser__show(&self->b, he->ms.sym->name, - "<- or ESC: exit, TAB/shift+TAB: cycle thru samples") < 0) + "<-, -> or ESC: exit, TAB/shift+TAB: cycle thru samples") < 0) return -1; - - newtFormAddHotKey(self->b.form, NEWT_KEY_LEFT); - newtFormAddHotKey(self->b.form, NEWT_KEY_RIGHT); + /* + * To allow builtin-annotate to cycle thru multiple symbols by + * examining the exit key for this function. + */ + ui_browser__add_exit_key(&self->b, NEWT_KEY_RIGHT); nd = self->curr_hot; if (nd) { - newtFormAddHotKey(self->b.form, NEWT_KEY_TAB); - newtFormAddHotKey(self->b.form, NEWT_KEY_UNTAB); + int tabs[] = { NEWT_KEY_TAB, NEWT_KEY_UNTAB, 0 }; + ui_browser__add_exit_keys(&self->b, tabs); } while (1) { - ui_browser__run(&self->b, es); - - if (es->reason != NEWT_EXIT_HOTKEY) - break; + key = ui_browser__run(&self->b); - switch (es->u.key) { + switch (key) { case NEWT_KEY_TAB: nd = rb_prev(nd); if (nd == NULL) @@ -179,12 +176,11 @@ static int annotate_browser__run(struct annotate_browser *self, } out: ui_browser__hide(&self->b); - return es->u.key; + return key; } int hist_entry__tui_annotate(struct hist_entry *self) { - struct newtExitStruct es; struct objdump_line *pos, *n; struct objdump_line_rb_node *rbpos; LIST_HEAD(head); @@ -232,7 +228,7 @@ int hist_entry__tui_annotate(struct hist_entry *self) annotate_browser__set_top(&browser, browser.curr_hot); browser.b.width += 18; /* Percentage */ - ret = annotate_browser__run(&browser, &es); + ret = annotate_browser__run(&browser); list_for_each_entry_safe(pos, n, &head, node) { list_del(&pos->node); objdump_line__free(pos); diff --git a/tools/perf/util/ui/browsers/hists.c b/tools/perf/util/ui/browsers/hists.c index 6866aa4..ebda8c3 100644 --- a/tools/perf/util/ui/browsers/hists.c +++ b/tools/perf/util/ui/browsers/hists.c @@ -58,6 +58,11 @@ static char callchain_list__folded(const struct callchain_list *self) return map_symbol__folded(&self->ms); } +static void map_symbol__set_folding(struct map_symbol *self, bool unfold) +{ + self->unfolded = unfold ? self->has_children : false; +} + static int callchain_node__count_rows_rb_tree(struct callchain_node *self) { int n = 0; @@ -129,16 +134,16 @@ static void callchain_node__init_have_children_rb_tree(struct callchain_node *se for (nd = rb_first(&self->rb_root); nd; nd = rb_next(nd)) { struct callchain_node *child = rb_entry(nd, struct callchain_node, rb_node); struct callchain_list *chain; - int first = true; + bool first = true; list_for_each_entry(chain, &child->val, list) { if (first) { first = false; chain->ms.has_children = chain->list.next != &child->val || - rb_first(&child->rb_root) != NULL; + !RB_EMPTY_ROOT(&child->rb_root); } else chain->ms.has_children = chain->list.next == &child->val && - rb_first(&child->rb_root) != NULL; + !RB_EMPTY_ROOT(&child->rb_root); } callchain_node__init_have_children_rb_tree(child); @@ -150,7 +155,7 @@ static void callchain_node__init_have_children(struct callchain_node *self) struct callchain_list *chain; list_for_each_entry(chain, &self->val, list) - chain->ms.has_children = rb_first(&self->rb_root) != NULL; + chain->ms.has_children = !RB_EMPTY_ROOT(&self->rb_root); callchain_node__init_have_children_rb_tree(self); } @@ -168,6 +173,7 @@ static void callchain__init_have_children(struct rb_root *self) static void hist_entry__init_have_children(struct hist_entry *self) { if (!self->init_have_children) { + self->ms.has_children = !RB_EMPTY_ROOT(&self->sorted_chain); callchain__init_have_children(&self->sorted_chain); self->init_have_children = true; } @@ -195,43 +201,114 @@ static bool hist_browser__toggle_fold(struct hist_browser *self) return false; } -static int hist_browser__run(struct hist_browser *self, const char *title, - struct newtExitStruct *es) +static int callchain_node__set_folding_rb_tree(struct callchain_node *self, bool unfold) +{ + int n = 0; + struct rb_node *nd; + + for (nd = rb_first(&self->rb_root); nd; nd = rb_next(nd)) { + struct callchain_node *child = rb_entry(nd, struct callchain_node, rb_node); + struct callchain_list *chain; + bool has_children = false; + + list_for_each_entry(chain, &child->val, list) { + ++n; + map_symbol__set_folding(&chain->ms, unfold); + has_children = chain->ms.has_children; + } + + if (has_children) + n += callchain_node__set_folding_rb_tree(child, unfold); + } + + return n; +} + +static int callchain_node__set_folding(struct callchain_node *node, bool unfold) +{ + struct callchain_list *chain; + bool has_children = false; + int n = 0; + + list_for_each_entry(chain, &node->val, list) { + ++n; + map_symbol__set_folding(&chain->ms, unfold); + has_children = chain->ms.has_children; + } + + if (has_children) + n += callchain_node__set_folding_rb_tree(node, unfold); + + return n; +} + +static int callchain__set_folding(struct rb_root *chain, bool unfold) +{ + struct rb_node *nd; + int n = 0; + + for (nd = rb_first(chain); nd; nd = rb_next(nd)) { + struct callchain_node *node = rb_entry(nd, struct callchain_node, rb_node); + n += callchain_node__set_folding(node, unfold); + } + + return n; +} + +static void hist_entry__set_folding(struct hist_entry *self, bool unfold) +{ + hist_entry__init_have_children(self); + map_symbol__set_folding(&self->ms, unfold); + + if (self->ms.has_children) { + int n = callchain__set_folding(&self->sorted_chain, unfold); + self->nr_rows = unfold ? n : 0; + } else + self->nr_rows = 0; +} + +static void hists__set_folding(struct hists *self, bool unfold) +{ + struct rb_node *nd; + + self->nr_entries = 0; + + for (nd = rb_first(&self->entries); nd; nd = rb_next(nd)) { + struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node); + hist_entry__set_folding(he, unfold); + self->nr_entries += 1 + he->nr_rows; + } +} + +static void hist_browser__set_folding(struct hist_browser *self, bool unfold) +{ + hists__set_folding(self->hists, unfold); + self->b.nr_entries = self->hists->nr_entries; + /* Go to the start, we may be way after valid entries after a collapse */ + ui_browser__reset_index(&self->b); +} + +static int hist_browser__run(struct hist_browser *self, const char *title) { - char str[256], unit; - unsigned long nr_events = self->hists->stats.nr_events[PERF_RECORD_SAMPLE]; + int key; + int exit_keys[] = { 'a', '?', 'h', 'C', 'd', 'D', 'E', 't', + NEWT_KEY_ENTER, NEWT_KEY_RIGHT, NEWT_KEY_LEFT, 0, }; self->b.entries = &self->hists->entries; self->b.nr_entries = self->hists->nr_entries; hist_browser__refresh_dimensions(self); - nr_events = convert_unit(nr_events, &unit); - snprintf(str, sizeof(str), "Events: %lu%c ", - nr_events, unit); - newtDrawRootText(0, 0, str); - if (ui_browser__show(&self->b, title, "Press '?' for help on key bindings") < 0) return -1; - newtFormAddHotKey(self->b.form, 'a'); - newtFormAddHotKey(self->b.form, '?'); - newtFormAddHotKey(self->b.form, 'h'); - newtFormAddHotKey(self->b.form, 'd'); - newtFormAddHotKey(self->b.form, 'D'); - newtFormAddHotKey(self->b.form, 't'); - - newtFormAddHotKey(self->b.form, NEWT_KEY_LEFT); - newtFormAddHotKey(self->b.form, NEWT_KEY_RIGHT); - newtFormAddHotKey(self->b.form, NEWT_KEY_ENTER); + ui_browser__add_exit_keys(&self->b, exit_keys); while (1) { - ui_browser__run(&self->b, es); + key = ui_browser__run(&self->b); - if (es->reason != NEWT_EXIT_HOTKEY) - break; - switch (es->u.key) { + switch (key) { case 'D': { /* Debug */ static int seq; struct hist_entry *h = rb_entry(self->b.top, @@ -245,18 +322,26 @@ static int hist_browser__run(struct hist_browser *self, const char *title, self->b.top_idx, h->row_offset, h->nr_rows); } - continue; + break; + case 'C': + /* Collapse the whole world. */ + hist_browser__set_folding(self, false); + break; + case 'E': + /* Expand the whole world. */ + hist_browser__set_folding(self, true); + break; case NEWT_KEY_ENTER: if (hist_browser__toggle_fold(self)) break; /* fall thru */ default: - return 0; + goto out; } } - +out: ui_browser__hide(&self->b); - return 0; + return key; } static char *callchain_list__sym_name(struct callchain_list *self, @@ -306,15 +391,10 @@ static int hist_browser__show_callchain_node_rb_tree(struct hist_browser *self, int color; bool was_first = first; - if (first) { + if (first) first = false; - chain->ms.has_children = chain->list.next != &child->val || - rb_first(&child->rb_root) != NULL; - } else { + else extra_offset = LEVEL_OFFSET_STEP; - chain->ms.has_children = chain->list.next == &child->val && - rb_first(&child->rb_root) != NULL; - } folded_sign = callchain_list__folded(chain); if (*row_offset != 0) { @@ -341,8 +421,8 @@ static int hist_browser__show_callchain_node_rb_tree(struct hist_browser *self, *is_current_entry = true; } - SLsmg_set_color(color); - SLsmg_gotorc(self->b.y + row, self->b.x); + ui_browser__set_color(&self->b, color); + ui_browser__gotorc(&self->b, row, 0); slsmg_write_nstring(" ", offset + extra_offset); slsmg_printf("%c ", folded_sign); slsmg_write_nstring(str, width); @@ -384,12 +464,7 @@ static int hist_browser__show_callchain_node(struct hist_browser *self, list_for_each_entry(chain, &node->val, list) { char ipstr[BITS_PER_LONG / 4 + 1], *s; int color; - /* - * FIXME: This should be moved to somewhere else, - * probably when the callchain is created, so as not to - * traverse it all over again - */ - chain->ms.has_children = rb_first(&node->rb_root) != NULL; + folded_sign = callchain_list__folded(chain); if (*row_offset != 0) { @@ -405,8 +480,8 @@ static int hist_browser__show_callchain_node(struct hist_browser *self, } s = callchain_list__sym_name(chain, ipstr, sizeof(ipstr)); - SLsmg_gotorc(self->b.y + row, self->b.x); - SLsmg_set_color(color); + ui_browser__gotorc(&self->b, row, 0); + ui_browser__set_color(&self->b, color); slsmg_write_nstring(" ", offset); slsmg_printf("%c ", folded_sign); slsmg_write_nstring(s, width - 2); @@ -465,7 +540,7 @@ static int hist_browser__show_entry(struct hist_browser *self, } if (symbol_conf.use_callchain) { - entry->ms.has_children = !RB_EMPTY_ROOT(&entry->sorted_chain); + hist_entry__init_have_children(entry); folded_sign = hist_entry__folded(entry); } @@ -484,8 +559,8 @@ static int hist_browser__show_entry(struct hist_browser *self, color = HE_COLORSET_NORMAL; } - SLsmg_set_color(color); - SLsmg_gotorc(self->b.y + row, self->b.x); + ui_browser__set_color(&self->b, color); + ui_browser__gotorc(&self->b, row, 0); if (symbol_conf.use_callchain) { slsmg_printf("%c ", folded_sign); width -= 2; @@ -687,8 +762,6 @@ static struct hist_browser *hist_browser__new(struct hists *hists) static void hist_browser__delete(struct hist_browser *self) { - newtFormDestroy(self->b.form); - newtPopWindow(); free(self); } @@ -702,21 +775,26 @@ static struct thread *hist_browser__selected_thread(struct hist_browser *self) return self->he_selection->thread; } -static int hist_browser__title(char *bf, size_t size, const char *ev_name, - const struct dso *dso, const struct thread *thread) +static int hists__browser_title(struct hists *self, char *bf, size_t size, + const char *ev_name, const struct dso *dso, + const struct thread *thread) { - int printed = 0; + char unit; + int printed; + unsigned long nr_events = self->stats.nr_events[PERF_RECORD_SAMPLE]; + + nr_events = convert_unit(nr_events, &unit); + printed = snprintf(bf, size, "Events: %lu%c %s", nr_events, unit, ev_name); if (thread) printed += snprintf(bf + printed, size - printed, - "Thread: %s(%d)", - (thread->comm_set ? thread->comm : ""), + ", Thread: %s(%d)", + (thread->comm_set ? thread->comm : ""), thread->pid); if (dso) printed += snprintf(bf + printed, size - printed, - "%sDSO: %s", thread ? " " : "", - dso->short_name); - return printed ?: snprintf(bf, size, "Event: %s", ev_name); + ", DSO: %s", dso->short_name); + return printed; } int hists__browse(struct hists *self, const char *helpline, const char *ev_name) @@ -725,7 +803,6 @@ int hists__browse(struct hists *self, const char *helpline, const char *ev_name) struct pstack *fstack; const struct thread *thread_filter = NULL; const struct dso *dso_filter = NULL; - struct newtExitStruct es; char msg[160]; int key = -1; @@ -738,9 +815,8 @@ int hists__browse(struct hists *self, const char *helpline, const char *ev_name) ui_helpline__push(helpline); - hist_browser__title(msg, sizeof(msg), ev_name, - dso_filter, thread_filter); - + hists__browser_title(self, msg, sizeof(msg), ev_name, + dso_filter, thread_filter); while (1) { const struct thread *thread; const struct dso *dso; @@ -749,70 +825,63 @@ int hists__browse(struct hists *self, const char *helpline, const char *ev_name) annotate = -2, zoom_dso = -2, zoom_thread = -2, browse_map = -2; - if (hist_browser__run(browser, msg, &es)) - break; + key = hist_browser__run(browser, msg); thread = hist_browser__selected_thread(browser); dso = browser->selection->map ? browser->selection->map->dso : NULL; - if (es.reason == NEWT_EXIT_HOTKEY) { - key = es.u.key; - - switch (key) { - case NEWT_KEY_F1: - goto do_help; - case NEWT_KEY_TAB: - case NEWT_KEY_UNTAB: - /* - * Exit the browser, let hists__browser_tree - * go to the next or previous - */ - goto out_free_stack; - default:; - } - - switch (key) { - case 'a': - if (browser->selection->map == NULL || - browser->selection->map->dso->annotate_warned) - continue; - goto do_annotate; - case 'd': - goto zoom_dso; - case 't': - goto zoom_thread; - case 'h': - case '?': -do_help: - ui__help_window("-> Zoom into DSO/Threads & Annotate current symbol\n" - "<- Zoom out\n" - "a Annotate current symbol\n" - "h/?/F1 Show this window\n" - "d Zoom into current DSO\n" - "t Zoom into current Thread\n" - "q/CTRL+C Exit browser"); + switch (key) { + case NEWT_KEY_TAB: + case NEWT_KEY_UNTAB: + /* + * Exit the browser, let hists__browser_tree + * go to the next or previous + */ + goto out_free_stack; + case 'a': + if (browser->selection->map == NULL && + browser->selection->map->dso->annotate_warned) continue; - default:; - } - if (is_exit_key(key)) { - if (key == NEWT_KEY_ESCAPE && - !ui__dialog_yesno("Do you really want to exit?")) - continue; - break; - } - - if (es.u.key == NEWT_KEY_LEFT) { - const void *top; + goto do_annotate; + case 'd': + goto zoom_dso; + case 't': + goto zoom_thread; + case NEWT_KEY_F1: + case 'h': + case '?': + ui__help_window("-> Zoom into DSO/Threads & Annotate current symbol\n" + "<- Zoom out\n" + "a Annotate current symbol\n" + "h/?/F1 Show this window\n" + "C Collapse all callchains\n" + "E Expand all callchains\n" + "d Zoom into current DSO\n" + "t Zoom into current Thread\n" + "q/CTRL+C Exit browser"); + continue; + case NEWT_KEY_ENTER: + case NEWT_KEY_RIGHT: + /* menu */ + break; + case NEWT_KEY_LEFT: { + const void *top; - if (pstack__empty(fstack)) - continue; - top = pstack__pop(fstack); - if (top == &dso_filter) - goto zoom_out_dso; - if (top == &thread_filter) - goto zoom_out_thread; + if (pstack__empty(fstack)) continue; - } + top = pstack__pop(fstack); + if (top == &dso_filter) + goto zoom_out_dso; + if (top == &thread_filter) + goto zoom_out_thread; + continue; + } + case NEWT_KEY_ESCAPE: + if (!ui__dialog_yesno("Do you really want to exit?")) + continue; + /* Fall thru */ + default: + goto out_free_stack; } if (browser->selection->sym != NULL && @@ -885,8 +954,8 @@ zoom_out_dso: pstack__push(fstack, &dso_filter); } hists__filter_by_dso(self, dso_filter); - hist_browser__title(msg, sizeof(msg), ev_name, - dso_filter, thread_filter); + hists__browser_title(self, msg, sizeof(msg), ev_name, + dso_filter, thread_filter); hist_browser__reset(browser); } else if (choice == zoom_thread) { zoom_thread: @@ -903,8 +972,8 @@ zoom_out_thread: pstack__push(fstack, &thread_filter); } hists__filter_by_thread(self, thread_filter); - hist_browser__title(msg, sizeof(msg), ev_name, - dso_filter, thread_filter); + hists__browser_title(self, msg, sizeof(msg), ev_name, + dso_filter, thread_filter); hist_browser__reset(browser); } } @@ -925,10 +994,6 @@ int hists__tui_browse_tree(struct rb_root *self, const char *help) const char *ev_name = __event_name(hists->type, hists->config); key = hists__browse(hists, help, ev_name); - - if (is_exit_key(key)) - break; - switch (key) { case NEWT_KEY_TAB: next = rb_next(nd); @@ -940,7 +1005,7 @@ int hists__tui_browse_tree(struct rb_root *self, const char *help) continue; nd = rb_prev(nd); default: - break; + return key; } } diff --git a/tools/perf/util/ui/browsers/map.c b/tools/perf/util/ui/browsers/map.c index 142b825..e35437d 100644 --- a/tools/perf/util/ui/browsers/map.c +++ b/tools/perf/util/ui/browsers/map.c @@ -1,6 +1,5 @@ #include "../libslang.h" #include <elf.h> -#include <newt.h> #include <sys/ttydefaults.h> #include <ctype.h> #include <string.h> @@ -47,7 +46,6 @@ out_free_form: struct map_browser { struct ui_browser b; struct map *map; - u16 namelen; u8 addrlen; }; @@ -56,14 +54,16 @@ static void map_browser__write(struct ui_browser *self, void *nd, int row) struct symbol *sym = rb_entry(nd, struct symbol, rb_node); struct map_browser *mb = container_of(self, struct map_browser, b); bool current_entry = ui_browser__is_current_entry(self, row); - int color = ui_browser__percent_color(0, current_entry); + int width; - SLsmg_set_color(color); + ui_browser__set_percent_color(self, 0, current_entry); slsmg_printf("%*llx %*llx %c ", mb->addrlen, sym->start, mb->addrlen, sym->end, sym->binding == STB_GLOBAL ? 'g' : sym->binding == STB_LOCAL ? 'l' : 'w'); - slsmg_write_nstring(sym->name, mb->namelen); + width = self->width - ((mb->addrlen * 2) + 4); + if (width > 0) + slsmg_write_nstring(sym->name, width); } /* FIXME uber-kludgy, see comment on cmd_report... */ @@ -98,31 +98,29 @@ static int map_browser__search(struct map_browser *self) return 0; } -static int map_browser__run(struct map_browser *self, struct newtExitStruct *es) +static int map_browser__run(struct map_browser *self) { + int key; + if (ui_browser__show(&self->b, self->map->dso->long_name, "Press <- or ESC to exit, %s / to search", verbose ? "" : "restart with -v to use") < 0) return -1; - newtFormAddHotKey(self->b.form, NEWT_KEY_LEFT); - newtFormAddHotKey(self->b.form, NEWT_KEY_ENTER); if (verbose) - newtFormAddHotKey(self->b.form, '/'); + ui_browser__add_exit_key(&self->b, '/'); while (1) { - ui_browser__run(&self->b, es); + key = ui_browser__run(&self->b); - if (es->reason != NEWT_EXIT_HOTKEY) - break; - if (verbose && es->u.key == '/') + if (verbose && key == '/') map_browser__search(self); else break; } ui_browser__hide(&self->b); - return 0; + return key; } int map__browse(struct map *self) @@ -136,7 +134,6 @@ int map__browse(struct map *self) }, .map = self, }; - struct newtExitStruct es; struct rb_node *nd; char tmp[BITS_PER_LONG / 4]; u64 maxaddr = 0; @@ -144,8 +141,6 @@ int map__browse(struct map *self) for (nd = rb_first(mb.b.entries); nd; nd = rb_next(nd)) { struct symbol *pos = rb_entry(nd, struct symbol, rb_node); - if (mb.namelen < pos->namelen) - mb.namelen = pos->namelen; if (maxaddr < pos->end) maxaddr = pos->end; if (verbose) { @@ -156,6 +151,5 @@ int map__browse(struct map *self) } mb.addrlen = snprintf(tmp, sizeof(tmp), "%llx", maxaddr); - mb.b.width += mb.addrlen * 2 + 4 + mb.namelen; - return map_browser__run(&mb, &es); + return map_browser__run(&mb); } diff --git a/tools/perf/util/ui/util.c b/tools/perf/util/ui/util.c index 04600e2..9706d9d 100644 --- a/tools/perf/util/ui/util.c +++ b/tools/perf/util/ui/util.c @@ -11,8 +11,6 @@ #include "helpline.h" #include "util.h" -newtComponent newt_form__new(void); - static void newt_form__set_exit_keys(newtComponent self) { newtFormAddHotKey(self, NEWT_KEY_LEFT); @@ -22,7 +20,7 @@ static void newt_form__set_exit_keys(newtComponent self) newtFormAddHotKey(self, CTRL('c')); } -newtComponent newt_form__new(void) +static newtComponent newt_form__new(void) { newtComponent self = newtForm(NULL, NULL, 0); if (self) diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h index f380fed..7562707 100644 --- a/tools/perf/util/util.h +++ b/tools/perf/util/util.h @@ -266,19 +266,6 @@ bool strglobmatch(const char *str, const char *pat); bool strlazymatch(const char *str, const char *pat); unsigned long convert_unit(unsigned long value, char *unit); -#ifndef ESC -#define ESC 27 -#endif - -static inline bool is_exit_key(int key) -{ - char up; - if (key == CTRL('c') || key == ESC) - return true; - up = toupper(key); - return up == 'Q'; -} - #define _STR(x) #x #define STR(x) _STR(x) |