From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758076Ab0CKTMj (ORCPT ); Thu, 11 Mar 2010 14:12:39 -0500 Received: from mx2.mail.elte.hu ([157.181.151.9]:35548 "EHLO mx2.mail.elte.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755815Ab0CKTMg (ORCPT ); Thu, 11 Mar 2010 14:12:36 -0500 Date: Thu, 11 Mar 2010 20:12:04 +0100 From: Ingo Molnar To: Linus Torvalds Cc: linux-kernel@vger.kernel.org, Peter Zijlstra , Arnaldo Carvalho de Melo , Paul Mackerras , Andrew Morton Subject: [GIT PULL] perf fixes Message-ID: <20100311191204.GA368@elte.hu> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.20 (2009-08-17) X-ELTE-SpamScore: -2.0 X-ELTE-SpamLevel: X-ELTE-SpamCheck: no X-ELTE-SpamVersion: ELTE 2.0 X-ELTE-SpamCheck-Details: score=-2.0 required=5.9 tests=BAYES_00 autolearn=no SpamAssassin version=3.2.5 -2.0 BAYES_00 BODY: Bayesian spam probability is 0 to 1% [score: 0.0000] Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Linus, Please pull the latest perf-fixes-for-linus git tree from: git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git perf-fixes-for-linus Thanks, Ingo ------------------> Arnaldo Carvalho de Melo (1): perf archive: Don't try to collect files without a build-id Frederic Weisbecker (5): perf lock: Drop the buffers multiplexing dependency perf: Remove pointless breakpoint union x86/hw-breakpoints: Remove the name field hw-breakpoints: Remove stub unthrottle callback x86/stacktrace: Don't dereference bad frame pointers Hitoshi Mitake (1): perf lock: Fix and add misc documentally things Ingo Molnar (1): MAINTAINERS: Add Arnaldo as tools/perf/ co-maintainer Peter Zijlstra (3): perf, x86: Restrict the ANY flag perf_events, x86: Fixup fixed counter constraints perf: Provide generic perf_sample_data initialization Robert Richter (3): perf, x86: make IBS macros available in perf_event.h perf, x86: add some IBS macros to perf_event.h perf, x86: rename macro in ARCH_PERFMON_EVENTSEL_ENABLE Tejun Heo (1): percpu: Add __percpu sparse annotations to hw_breakpoint Tom Zanussi (2): perf trace/scripting: Remove extraneous header read perf trace: Don't use pager if scripting Will Deacon (1): perf, ARM: Modify kuser rmb() call to compile for Thumb-2 MAINTAINERS | 1 + arch/arm/kernel/perf_event.c | 4 +- arch/powerpc/kernel/perf_event.c | 8 +- arch/sparc/kernel/perf_event.c | 2 +- arch/x86/include/asm/hw_breakpoint.h | 1 - arch/x86/include/asm/perf_event.h | 16 +++- arch/x86/kernel/cpu/perf_event.c | 39 ++++++--- arch/x86/kernel/cpu/perf_event_intel.c | 37 +++++--- arch/x86/kernel/cpu/perf_event_p6.c | 8 +- arch/x86/kernel/cpu/perfctr-watchdog.c | 2 +- arch/x86/kernel/dumpstack_64.c | 10 ++- arch/x86/kernel/hw_breakpoint.c | 12 --- arch/x86/oprofile/op_model_amd.c | 23 ++---- arch/x86/oprofile/op_model_ppro.c | 6 +- include/linux/hw_breakpoint.h | 8 +- include/linux/perf_event.h | 27 +++++- kernel/hw_breakpoint.c | 11 +-- kernel/perf_event.c | 36 ++------ lib/Kconfig.debug | 8 ++ samples/hw_breakpoint/data_breakpoint.c | 6 +- tools/perf/Documentation/perf-lock.txt | 29 ++++++ tools/perf/builtin-lock.c | 148 ++++++++++++++++++++++++++++++- tools/perf/builtin-trace.c | 4 +- tools/perf/command-list.txt | 1 + tools/perf/perf-archive.sh | 3 +- tools/perf/perf.h | 4 +- tools/perf/util/probe-event.c | 2 +- 27 files changed, 327 insertions(+), 129 deletions(-) create mode 100644 tools/perf/Documentation/perf-lock.txt diff --git a/MAINTAINERS b/MAINTAINERS index c8a8b1f..14d5c4c 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4293,6 +4293,7 @@ PERFORMANCE EVENTS SUBSYSTEM M: Peter Zijlstra M: Paul Mackerras M: Ingo Molnar +M: Arnaldo Carvalho de Melo S: Supported F: kernel/perf_event.c F: include/linux/perf_event.h diff --git a/arch/arm/kernel/perf_event.c b/arch/arm/kernel/perf_event.c index c54ceb3..3875d99 100644 --- a/arch/arm/kernel/perf_event.c +++ b/arch/arm/kernel/perf_event.c @@ -965,7 +965,7 @@ armv6pmu_handle_irq(int irq_num, */ armv6_pmcr_write(pmcr); - data.addr = 0; + perf_sample_data_init(&data, 0); cpuc = &__get_cpu_var(cpu_hw_events); for (idx = 0; idx <= armpmu->num_events; ++idx) { @@ -1945,7 +1945,7 @@ static irqreturn_t armv7pmu_handle_irq(int irq_num, void *dev) */ regs = get_irq_regs(); - data.addr = 0; + perf_sample_data_init(&data, 0); cpuc = &__get_cpu_var(cpu_hw_events); for (idx = 0; idx <= armpmu->num_events; ++idx) { diff --git a/arch/powerpc/kernel/perf_event.c b/arch/powerpc/kernel/perf_event.c index b6cf8f1..5120bd4 100644 --- a/arch/powerpc/kernel/perf_event.c +++ b/arch/powerpc/kernel/perf_event.c @@ -1164,10 +1164,10 @@ static void record_and_restart(struct perf_event *event, unsigned long val, * Finally record data if requested. */ if (record) { - struct perf_sample_data data = { - .addr = ~0ULL, - .period = event->hw.last_period, - }; + struct perf_sample_data data; + + perf_sample_data_init(&data, ~0ULL); + data.period = event->hw.last_period; if (event->attr.sample_type & PERF_SAMPLE_ADDR) perf_get_data_addr(regs, &data.addr); diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c index 9f2b2ba..6504208 100644 --- a/arch/sparc/kernel/perf_event.c +++ b/arch/sparc/kernel/perf_event.c @@ -1189,7 +1189,7 @@ static int __kprobes perf_event_nmi_handler(struct notifier_block *self, regs = args->regs; - data.addr = 0; + perf_sample_data_init(&data, 0); cpuc = &__get_cpu_var(cpu_hw_events); diff --git a/arch/x86/include/asm/hw_breakpoint.h b/arch/x86/include/asm/hw_breakpoint.h index 0675a7c..2a1bd8f 100644 --- a/arch/x86/include/asm/hw_breakpoint.h +++ b/arch/x86/include/asm/hw_breakpoint.h @@ -10,7 +10,6 @@ * (display/resolving) */ struct arch_hw_breakpoint { - char *name; /* Contains name of the symbol to set bkpt */ unsigned long address; u8 len; u8 type; diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h index befd172..db6109a 100644 --- a/arch/x86/include/asm/perf_event.h +++ b/arch/x86/include/asm/perf_event.h @@ -18,7 +18,7 @@ #define MSR_ARCH_PERFMON_EVENTSEL0 0x186 #define MSR_ARCH_PERFMON_EVENTSEL1 0x187 -#define ARCH_PERFMON_EVENTSEL0_ENABLE (1 << 22) +#define ARCH_PERFMON_EVENTSEL_ENABLE (1 << 22) #define ARCH_PERFMON_EVENTSEL_ANY (1 << 21) #define ARCH_PERFMON_EVENTSEL_INT (1 << 20) #define ARCH_PERFMON_EVENTSEL_OS (1 << 17) @@ -50,7 +50,7 @@ INTEL_ARCH_INV_MASK| \ INTEL_ARCH_EDGE_MASK|\ INTEL_ARCH_UNIT_MASK|\ - INTEL_ARCH_EVTSEL_MASK) + INTEL_ARCH_EVENT_MASK) #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL 0x3c #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK (0x00 << 8) @@ -117,6 +117,18 @@ union cpuid10_edx { */ #define X86_PMC_IDX_FIXED_BTS (X86_PMC_IDX_FIXED + 16) +/* IbsFetchCtl bits/masks */ +#define IBS_FETCH_RAND_EN (1ULL<<57) +#define IBS_FETCH_VAL (1ULL<<49) +#define IBS_FETCH_ENABLE (1ULL<<48) +#define IBS_FETCH_CNT 0xFFFF0000ULL +#define IBS_FETCH_MAX_CNT 0x0000FFFFULL + +/* IbsOpCtl bits */ +#define IBS_OP_CNT_CTL (1ULL<<19) +#define IBS_OP_VAL (1ULL<<18) +#define IBS_OP_ENABLE (1ULL<<17) +#define IBS_OP_MAX_CNT 0x0000FFFFULL #ifdef CONFIG_PERF_EVENTS extern void init_hw_perf_events(void); diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index b1fbdee..42aafd1 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -73,10 +73,10 @@ struct debug_store { struct event_constraint { union { unsigned long idxmsk[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; - u64 idxmsk64[1]; + u64 idxmsk64; }; - int code; - int cmask; + u64 code; + u64 cmask; int weight; }; @@ -103,7 +103,7 @@ struct cpu_hw_events { }; #define __EVENT_CONSTRAINT(c, n, m, w) {\ - { .idxmsk64[0] = (n) }, \ + { .idxmsk64 = (n) }, \ .code = (c), \ .cmask = (m), \ .weight = (w), \ @@ -116,7 +116,7 @@ struct cpu_hw_events { EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVTSEL_MASK) #define FIXED_EVENT_CONSTRAINT(c, n) \ - EVENT_CONSTRAINT(c, n, INTEL_ARCH_FIXED_MASK) + EVENT_CONSTRAINT(c, (1ULL << (32+n)), INTEL_ARCH_FIXED_MASK) #define EVENT_CONSTRAINT_END \ EVENT_CONSTRAINT(0, 0, 0) @@ -503,6 +503,9 @@ static int __hw_perf_event_init(struct perf_event *event) */ if (attr->type == PERF_TYPE_RAW) { hwc->config |= x86_pmu.raw_event(attr->config); + if ((hwc->config & ARCH_PERFMON_EVENTSEL_ANY) && + perf_paranoid_cpu() && !capable(CAP_SYS_ADMIN)) + return -EACCES; return 0; } @@ -553,9 +556,9 @@ static void x86_pmu_disable_all(void) if (!test_bit(idx, cpuc->active_mask)) continue; rdmsrl(x86_pmu.eventsel + idx, val); - if (!(val & ARCH_PERFMON_EVENTSEL0_ENABLE)) + if (!(val & ARCH_PERFMON_EVENTSEL_ENABLE)) continue; - val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; + val &= ~ARCH_PERFMON_EVENTSEL_ENABLE; wrmsrl(x86_pmu.eventsel + idx, val); } } @@ -590,7 +593,7 @@ static void x86_pmu_enable_all(void) continue; val = event->hw.config; - val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + val |= ARCH_PERFMON_EVENTSEL_ENABLE; wrmsrl(x86_pmu.eventsel + idx, val); } } @@ -612,8 +615,8 @@ static int x86_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) bitmap_zero(used_mask, X86_PMC_IDX_MAX); for (i = 0; i < n; i++) { - constraints[i] = - x86_pmu.get_event_constraints(cpuc, cpuc->event_list[i]); + c = x86_pmu.get_event_constraints(cpuc, cpuc->event_list[i]); + constraints[i] = c; } /* @@ -853,7 +856,7 @@ void hw_perf_enable(void) static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc, int idx) { (void)checking_wrmsrl(hwc->config_base + idx, - hwc->config | ARCH_PERFMON_EVENTSEL0_ENABLE); + hwc->config | ARCH_PERFMON_EVENTSEL_ENABLE); } static inline void x86_pmu_disable_event(struct hw_perf_event *hwc, int idx) @@ -1094,8 +1097,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs) int idx, handled = 0; u64 val; - data.addr = 0; - data.raw = NULL; + perf_sample_data_init(&data, 0); cpuc = &__get_cpu_var(cpu_hw_events); @@ -1347,6 +1349,7 @@ static void __init pmu_check_apic(void) void __init init_hw_perf_events(void) { + struct event_constraint *c; int err; pr_info("Performance Events: "); @@ -1395,6 +1398,16 @@ void __init init_hw_perf_events(void) __EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_events) - 1, 0, x86_pmu.num_events); + if (x86_pmu.event_constraints) { + for_each_event_constraint(c, x86_pmu.event_constraints) { + if (c->cmask != INTEL_ARCH_FIXED_MASK) + continue; + + c->idxmsk64 |= (1ULL << x86_pmu.num_events) - 1; + c->weight += x86_pmu.num_events; + } + } + pr_info("... version: %d\n", x86_pmu.version); pr_info("... bit width: %d\n", x86_pmu.event_bits); pr_info("... generic registers: %d\n", x86_pmu.num_events); diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 977e754..44b60c8 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -1,7 +1,7 @@ #ifdef CONFIG_CPU_SUP_INTEL /* - * Intel PerfMon v3. Used on Core2 and later. + * Intel PerfMon, used on Core and later. */ static const u64 intel_perfmon_event_map[] = { @@ -27,8 +27,14 @@ static struct event_constraint intel_core_event_constraints[] = static struct event_constraint intel_core2_event_constraints[] = { - FIXED_EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ - FIXED_EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ + FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */ + FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */ + /* + * Core2 has Fixed Counter 2 listed as CPU_CLK_UNHALTED.REF and event + * 0x013c as CPU_CLK_UNHALTED.BUS and specifies there is a fixed + * ratio between these counters. + */ + /* FIXED_EVENT_CONSTRAINT(0x013c, 2), CPU_CLK_UNHALTED.REF */ INTEL_EVENT_CONSTRAINT(0x10, 0x1), /* FP_COMP_OPS_EXE */ INTEL_EVENT_CONSTRAINT(0x11, 0x2), /* FP_ASSIST */ INTEL_EVENT_CONSTRAINT(0x12, 0x2), /* MUL */ @@ -37,14 +43,16 @@ static struct event_constraint intel_core2_event_constraints[] = INTEL_EVENT_CONSTRAINT(0x18, 0x1), /* IDLE_DURING_DIV */ INTEL_EVENT_CONSTRAINT(0x19, 0x2), /* DELAYED_BYPASS */ INTEL_EVENT_CONSTRAINT(0xa1, 0x1), /* RS_UOPS_DISPATCH_CYCLES */ + INTEL_EVENT_CONSTRAINT(0xc9, 0x1), /* ITLB_MISS_RETIRED (T30-9) */ INTEL_EVENT_CONSTRAINT(0xcb, 0x1), /* MEM_LOAD_RETIRED */ EVENT_CONSTRAINT_END }; static struct event_constraint intel_nehalem_event_constraints[] = { - FIXED_EVENT_CONSTRAINT(0xc0, (0xf|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ - FIXED_EVENT_CONSTRAINT(0x3c, (0xf|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ + FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */ + FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */ + /* FIXED_EVENT_CONSTRAINT(0x013c, 2), CPU_CLK_UNHALTED.REF */ INTEL_EVENT_CONSTRAINT(0x40, 0x3), /* L1D_CACHE_LD */ INTEL_EVENT_CONSTRAINT(0x41, 0x3), /* L1D_CACHE_ST */ INTEL_EVENT_CONSTRAINT(0x42, 0x3), /* L1D_CACHE_LOCK */ @@ -58,8 +66,9 @@ static struct event_constraint intel_nehalem_event_constraints[] = static struct event_constraint intel_westmere_event_constraints[] = { - FIXED_EVENT_CONSTRAINT(0xc0, (0xf|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ - FIXED_EVENT_CONSTRAINT(0x3c, (0xf|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ + FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */ + FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */ + /* FIXED_EVENT_CONSTRAINT(0x013c, 2), CPU_CLK_UNHALTED.REF */ INTEL_EVENT_CONSTRAINT(0x51, 0x3), /* L1D */ INTEL_EVENT_CONSTRAINT(0x60, 0x1), /* OFFCORE_REQUESTS_OUTSTANDING */ INTEL_EVENT_CONSTRAINT(0x63, 0x3), /* CACHE_LOCK_CYCLES */ @@ -68,8 +77,9 @@ static struct event_constraint intel_westmere_event_constraints[] = static struct event_constraint intel_gen_event_constraints[] = { - FIXED_EVENT_CONSTRAINT(0xc0, (0x3|(1ULL<<32))), /* INSTRUCTIONS_RETIRED */ - FIXED_EVENT_CONSTRAINT(0x3c, (0x3|(1ULL<<33))), /* UNHALTED_CORE_CYCLES */ + FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */ + FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */ + /* FIXED_EVENT_CONSTRAINT(0x013c, 2), CPU_CLK_UNHALTED.REF */ EVENT_CONSTRAINT_END }; @@ -580,10 +590,9 @@ static void intel_pmu_drain_bts_buffer(void) ds->bts_index = ds->bts_buffer_base; + perf_sample_data_init(&data, 0); data.period = event->hw.last_period; - data.addr = 0; - data.raw = NULL; regs.ip = 0; /* @@ -732,8 +741,7 @@ static int intel_pmu_handle_irq(struct pt_regs *regs) int bit, loops; u64 ack, status; - data.addr = 0; - data.raw = NULL; + perf_sample_data_init(&data, 0); cpuc = &__get_cpu_var(cpu_hw_events); @@ -935,7 +943,7 @@ static __init int intel_pmu_init(void) x86_pmu.event_constraints = intel_nehalem_event_constraints; pr_cont("Nehalem/Corei7 events, "); break; - case 28: + case 28: /* Atom */ memcpy(hw_cache_event_ids, atom_hw_cache_event_ids, sizeof(hw_cache_event_ids)); @@ -951,6 +959,7 @@ static __init int intel_pmu_init(void) x86_pmu.event_constraints = intel_westmere_event_constraints; pr_cont("Westmere events, "); break; + default: /* * default constraints for v2 and up diff --git a/arch/x86/kernel/cpu/perf_event_p6.c b/arch/x86/kernel/cpu/perf_event_p6.c index 1ca5ba0..a4e67b9 100644 --- a/arch/x86/kernel/cpu/perf_event_p6.c +++ b/arch/x86/kernel/cpu/perf_event_p6.c @@ -62,7 +62,7 @@ static void p6_pmu_disable_all(void) /* p6 only has one enable register */ rdmsrl(MSR_P6_EVNTSEL0, val); - val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; + val &= ~ARCH_PERFMON_EVENTSEL_ENABLE; wrmsrl(MSR_P6_EVNTSEL0, val); } @@ -72,7 +72,7 @@ static void p6_pmu_enable_all(void) /* p6 only has one enable register */ rdmsrl(MSR_P6_EVNTSEL0, val); - val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + val |= ARCH_PERFMON_EVENTSEL_ENABLE; wrmsrl(MSR_P6_EVNTSEL0, val); } @@ -83,7 +83,7 @@ p6_pmu_disable_event(struct hw_perf_event *hwc, int idx) u64 val = P6_NOP_EVENT; if (cpuc->enabled) - val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + val |= ARCH_PERFMON_EVENTSEL_ENABLE; (void)checking_wrmsrl(hwc->config_base + idx, val); } @@ -95,7 +95,7 @@ static void p6_pmu_enable_event(struct hw_perf_event *hwc, int idx) val = hwc->config; if (cpuc->enabled) - val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + val |= ARCH_PERFMON_EVENTSEL_ENABLE; (void)checking_wrmsrl(hwc->config_base + idx, val); } diff --git a/arch/x86/kernel/cpu/perfctr-watchdog.c b/arch/x86/kernel/cpu/perfctr-watchdog.c index 74f4e85..fb329e9 100644 --- a/arch/x86/kernel/cpu/perfctr-watchdog.c +++ b/arch/x86/kernel/cpu/perfctr-watchdog.c @@ -680,7 +680,7 @@ static int setup_intel_arch_watchdog(unsigned nmi_hz) cpu_nmi_set_wd_enabled(); apic_write(APIC_LVTPC, APIC_DM_NMI); - evntsel |= ARCH_PERFMON_EVENTSEL0_ENABLE; + evntsel |= ARCH_PERFMON_EVENTSEL_ENABLE; wrmsr(evntsel_msr, evntsel, 0); intel_arch_wd_ops.checkbit = 1ULL << (eax.split.bit_width - 1); return 1; diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c index dce99ab..d5e2a2e 100644 --- a/arch/x86/kernel/dumpstack_64.c +++ b/arch/x86/kernel/dumpstack_64.c @@ -120,9 +120,15 @@ fixup_bp_irq_link(unsigned long bp, unsigned long *stack, { #ifdef CONFIG_FRAME_POINTER struct stack_frame *frame = (struct stack_frame *)bp; + unsigned long next; - if (!in_irq_stack(stack, irq_stack, irq_stack_end)) - return (unsigned long)frame->next_frame; + if (!in_irq_stack(stack, irq_stack, irq_stack_end)) { + if (!probe_kernel_address(&frame->next_frame, next)) + return next; + else + WARN_ONCE(1, "Perf: bad frame pointer = %p in " + "callchain\n", &frame->next_frame); + } #endif return bp; } diff --git a/arch/x86/kernel/hw_breakpoint.c b/arch/x86/kernel/hw_breakpoint.c index dca2802..d6cc065 100644 --- a/arch/x86/kernel/hw_breakpoint.c +++ b/arch/x86/kernel/hw_breakpoint.c @@ -344,13 +344,6 @@ int arch_validate_hwbkpt_settings(struct perf_event *bp, } /* - * For kernel-addresses, either the address or symbol name can be - * specified. - */ - if (info->name) - info->address = (unsigned long) - kallsyms_lookup_name(info->name); - /* * Check that the low-order bits of the address are appropriate * for the alignment implied by len. */ @@ -535,8 +528,3 @@ void hw_breakpoint_pmu_read(struct perf_event *bp) { /* TODO */ } - -void hw_breakpoint_pmu_unthrottle(struct perf_event *bp) -{ - /* TODO */ -} diff --git a/arch/x86/oprofile/op_model_amd.c b/arch/x86/oprofile/op_model_amd.c index 6a58256..090cbbe 100644 --- a/arch/x86/oprofile/op_model_amd.c +++ b/arch/x86/oprofile/op_model_amd.c @@ -46,17 +46,6 @@ static unsigned long reset_value[NUM_VIRT_COUNTERS]; -/* IbsFetchCtl bits/masks */ -#define IBS_FETCH_RAND_EN (1ULL<<57) -#define IBS_FETCH_VAL (1ULL<<49) -#define IBS_FETCH_ENABLE (1ULL<<48) -#define IBS_FETCH_CNT_MASK 0xFFFF0000ULL - -/* IbsOpCtl bits */ -#define IBS_OP_CNT_CTL (1ULL<<19) -#define IBS_OP_VAL (1ULL<<18) -#define IBS_OP_ENABLE (1ULL<<17) - #define IBS_FETCH_SIZE 6 #define IBS_OP_SIZE 12 @@ -182,7 +171,7 @@ static void op_amd_setup_ctrs(struct op_x86_model_spec const *model, continue; } rdmsrl(msrs->controls[i].addr, val); - if (val & ARCH_PERFMON_EVENTSEL0_ENABLE) + if (val & ARCH_PERFMON_EVENTSEL_ENABLE) op_x86_warn_in_use(i); val &= model->reserved; wrmsrl(msrs->controls[i].addr, val); @@ -290,7 +279,7 @@ op_amd_handle_ibs(struct pt_regs * const regs, oprofile_write_commit(&entry); /* reenable the IRQ */ - ctl &= ~(IBS_FETCH_VAL | IBS_FETCH_CNT_MASK); + ctl &= ~(IBS_FETCH_VAL | IBS_FETCH_CNT); ctl |= IBS_FETCH_ENABLE; wrmsrl(MSR_AMD64_IBSFETCHCTL, ctl); } @@ -330,7 +319,7 @@ static inline void op_amd_start_ibs(void) return; if (ibs_config.fetch_enabled) { - val = (ibs_config.max_cnt_fetch >> 4) & 0xFFFF; + val = (ibs_config.max_cnt_fetch >> 4) & IBS_FETCH_MAX_CNT; val |= ibs_config.rand_en ? IBS_FETCH_RAND_EN : 0; val |= IBS_FETCH_ENABLE; wrmsrl(MSR_AMD64_IBSFETCHCTL, val); @@ -352,7 +341,7 @@ static inline void op_amd_start_ibs(void) * avoid underflows. */ ibs_op_ctl = min(ibs_op_ctl + IBS_RANDOM_MAXCNT_OFFSET, - 0xFFFFULL); + IBS_OP_MAX_CNT); } if (ibs_caps & IBS_CAPS_OPCNT && ibs_config.dispatched_ops) ibs_op_ctl |= IBS_OP_CNT_CTL; @@ -409,7 +398,7 @@ static void op_amd_start(struct op_msrs const * const msrs) if (!reset_value[op_x86_phys_to_virt(i)]) continue; rdmsrl(msrs->controls[i].addr, val); - val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + val |= ARCH_PERFMON_EVENTSEL_ENABLE; wrmsrl(msrs->controls[i].addr, val); } @@ -429,7 +418,7 @@ static void op_amd_stop(struct op_msrs const * const msrs) if (!reset_value[op_x86_phys_to_virt(i)]) continue; rdmsrl(msrs->controls[i].addr, val); - val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; + val &= ~ARCH_PERFMON_EVENTSEL_ENABLE; wrmsrl(msrs->controls[i].addr, val); } diff --git a/arch/x86/oprofile/op_model_ppro.c b/arch/x86/oprofile/op_model_ppro.c index 5d1727b..2bf90fa 100644 --- a/arch/x86/oprofile/op_model_ppro.c +++ b/arch/x86/oprofile/op_model_ppro.c @@ -88,7 +88,7 @@ static void ppro_setup_ctrs(struct op_x86_model_spec const *model, continue; } rdmsrl(msrs->controls[i].addr, val); - if (val & ARCH_PERFMON_EVENTSEL0_ENABLE) + if (val & ARCH_PERFMON_EVENTSEL_ENABLE) op_x86_warn_in_use(i); val &= model->reserved; wrmsrl(msrs->controls[i].addr, val); @@ -166,7 +166,7 @@ static void ppro_start(struct op_msrs const * const msrs) for (i = 0; i < num_counters; ++i) { if (reset_value[i]) { rdmsrl(msrs->controls[i].addr, val); - val |= ARCH_PERFMON_EVENTSEL0_ENABLE; + val |= ARCH_PERFMON_EVENTSEL_ENABLE; wrmsrl(msrs->controls[i].addr, val); } } @@ -184,7 +184,7 @@ static void ppro_stop(struct op_msrs const * const msrs) if (!reset_value[i]) continue; rdmsrl(msrs->controls[i].addr, val); - val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE; + val &= ~ARCH_PERFMON_EVENTSEL_ENABLE; wrmsrl(msrs->controls[i].addr, val); } } diff --git a/include/linux/hw_breakpoint.h b/include/linux/hw_breakpoint.h index 5977b72..c70d27a 100644 --- a/include/linux/hw_breakpoint.h +++ b/include/linux/hw_breakpoint.h @@ -66,14 +66,14 @@ register_wide_hw_breakpoint_cpu(struct perf_event_attr *attr, perf_overflow_handler_t triggered, int cpu); -extern struct perf_event ** +extern struct perf_event * __percpu * register_wide_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered); extern int register_perf_hw_breakpoint(struct perf_event *bp); extern int __register_perf_hw_breakpoint(struct perf_event *bp); extern void unregister_hw_breakpoint(struct perf_event *bp); -extern void unregister_wide_hw_breakpoint(struct perf_event **cpu_events); +extern void unregister_wide_hw_breakpoint(struct perf_event * __percpu *cpu_events); extern int dbg_reserve_bp_slot(struct perf_event *bp); extern int dbg_release_bp_slot(struct perf_event *bp); @@ -100,7 +100,7 @@ static inline struct perf_event * register_wide_hw_breakpoint_cpu(struct perf_event_attr *attr, perf_overflow_handler_t triggered, int cpu) { return NULL; } -static inline struct perf_event ** +static inline struct perf_event * __percpu * register_wide_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered) { return NULL; } static inline int @@ -109,7 +109,7 @@ static inline int __register_perf_hw_breakpoint(struct perf_event *bp) { return -ENOSYS; } static inline void unregister_hw_breakpoint(struct perf_event *bp) { } static inline void -unregister_wide_hw_breakpoint(struct perf_event **cpu_events) { } +unregister_wide_hw_breakpoint(struct perf_event * __percpu *cpu_events) { } static inline int reserve_bp_slot(struct perf_event *bp) {return -ENOSYS; } static inline void release_bp_slot(struct perf_event *bp) { } diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 7b18b4f..6f8cd7d 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -487,9 +487,8 @@ struct hw_perf_event { struct hrtimer hrtimer; }; #ifdef CONFIG_HAVE_HW_BREAKPOINT - union { /* breakpoint */ - struct arch_hw_breakpoint info; - }; + /* breakpoint */ + struct arch_hw_breakpoint info; #endif }; atomic64_t prev_count; @@ -802,6 +801,13 @@ struct perf_sample_data { struct perf_raw_record *raw; }; +static inline +void perf_sample_data_init(struct perf_sample_data *data, u64 addr) +{ + data->addr = addr; + data->raw = NULL; +} + extern void perf_output_sample(struct perf_output_handle *handle, struct perf_event_header *header, struct perf_sample_data *data, @@ -858,6 +864,21 @@ extern int sysctl_perf_event_paranoid; extern int sysctl_perf_event_mlock; extern int sysctl_perf_event_sample_rate; +static inline bool perf_paranoid_tracepoint_raw(void) +{ + return sysctl_perf_event_paranoid > -1; +} + +static inline bool perf_paranoid_cpu(void) +{ + return sysctl_perf_event_paranoid > 0; +} + +static inline bool perf_paranoid_kernel(void) +{ + return sysctl_perf_event_paranoid > 1; +} + extern void perf_event_init(void); extern void perf_tp_event(int event_id, u64 addr, u64 count, void *record, int entry_size); extern void perf_bp_event(struct perf_event *event, void *data); diff --git a/kernel/hw_breakpoint.c b/kernel/hw_breakpoint.c index 967e661..03808ed 100644 --- a/kernel/hw_breakpoint.c +++ b/kernel/hw_breakpoint.c @@ -413,17 +413,17 @@ EXPORT_SYMBOL_GPL(unregister_hw_breakpoint); * * @return a set of per_cpu pointers to perf events */ -struct perf_event ** +struct perf_event * __percpu * register_wide_hw_breakpoint(struct perf_event_attr *attr, perf_overflow_handler_t triggered) { - struct perf_event **cpu_events, **pevent, *bp; + struct perf_event * __percpu *cpu_events, **pevent, *bp; long err; int cpu; cpu_events = alloc_percpu(typeof(*cpu_events)); if (!cpu_events) - return ERR_PTR(-ENOMEM); + return (void __percpu __force *)ERR_PTR(-ENOMEM); get_online_cpus(); for_each_online_cpu(cpu) { @@ -451,7 +451,7 @@ fail: put_online_cpus(); free_percpu(cpu_events); - return ERR_PTR(err); + return (void __percpu __force *)ERR_PTR(err); } EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); @@ -459,7 +459,7 @@ EXPORT_SYMBOL_GPL(register_wide_hw_breakpoint); * unregister_wide_hw_breakpoint - unregister a wide breakpoint in the kernel * @cpu_events: the per cpu set of events to unregister */ -void unregister_wide_hw_breakpoint(struct perf_event **cpu_events) +void unregister_wide_hw_breakpoint(struct perf_event * __percpu *cpu_events) { int cpu; struct perf_event **pevent; @@ -489,5 +489,4 @@ struct pmu perf_ops_bp = { .enable = arch_install_hw_breakpoint, .disable = arch_uninstall_hw_breakpoint, .read = hw_breakpoint_pmu_read, - .unthrottle = hw_breakpoint_pmu_unthrottle }; diff --git a/kernel/perf_event.c b/kernel/perf_event.c index f40560b..4393b9e 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -56,21 +56,6 @@ static atomic_t nr_task_events __read_mostly; */ int sysctl_perf_event_paranoid __read_mostly = 1; -static inline bool perf_paranoid_tracepoint_raw(void) -{ - return sysctl_perf_event_paranoid > -1; -} - -static inline bool perf_paranoid_cpu(void) -{ - return sysctl_perf_event_paranoid > 0; -} - -static inline bool perf_paranoid_kernel(void) -{ - return sysctl_perf_event_paranoid > 1; -} - int sysctl_perf_event_mlock __read_mostly = 512; /* 'free' kb per user */ /* @@ -4123,8 +4108,7 @@ void __perf_sw_event(u32 event_id, u64 nr, int nmi, if (rctx < 0) return; - data.addr = addr; - data.raw = NULL; + perf_sample_data_init(&data, addr); do_perf_sw_event(PERF_TYPE_SOFTWARE, event_id, nr, nmi, &data, regs); @@ -4169,11 +4153,10 @@ static enum hrtimer_restart perf_swevent_hrtimer(struct hrtimer *hrtimer) struct perf_event *event; u64 period; - event = container_of(hrtimer, struct perf_event, hw.hrtimer); + event = container_of(hrtimer, struct perf_event, hw.hrtimer); event->pmu->read(event); - data.addr = 0; - data.raw = NULL; + perf_sample_data_init(&data, 0); data.period = event->hw.last_period; regs = get_irq_regs(); /* @@ -4337,17 +4320,15 @@ static const struct pmu perf_ops_task_clock = { void perf_tp_event(int event_id, u64 addr, u64 count, void *record, int entry_size) { + struct pt_regs *regs = get_irq_regs(); + struct perf_sample_data data; struct perf_raw_record raw = { .size = entry_size, .data = record, }; - struct perf_sample_data data = { - .addr = addr, - .raw = &raw, - }; - - struct pt_regs *regs = get_irq_regs(); + perf_sample_data_init(&data, addr); + data.raw = &raw; if (!regs) regs = task_pt_regs(current); @@ -4463,8 +4444,7 @@ void perf_bp_event(struct perf_event *bp, void *data) struct perf_sample_data sample; struct pt_regs *regs = data; - sample.raw = NULL; - sample.addr = bp->attr.bp_addr; + perf_sample_data_init(&sample, bp->attr.bp_addr); if (!perf_exclude_event(bp, regs)) perf_swevent_add(bp, 1, 1, &sample, regs); diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index b520ec1..8e5ec5e 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -532,6 +532,14 @@ config LOCK_STAT For more details, see Documentation/lockstat.txt + This also enables lock events required by "perf lock", + subcommand of perf. + If you want to use "perf lock", you also need to turn on + CONFIG_EVENT_TRACING. + + CONFIG_LOCK_STAT defines "contended" and "acquired" lock events. + (CONFIG_LOCKDEP defines "acquire" and "release" events.) + config DEBUG_LOCKDEP bool "Lock dependency engine debugging" depends on DEBUG_KERNEL && LOCKDEP diff --git a/samples/hw_breakpoint/data_breakpoint.c b/samples/hw_breakpoint/data_breakpoint.c index c69cbe9..bd0f337 100644 --- a/samples/hw_breakpoint/data_breakpoint.c +++ b/samples/hw_breakpoint/data_breakpoint.c @@ -34,7 +34,7 @@ #include #include -struct perf_event **sample_hbp; +struct perf_event * __percpu *sample_hbp; static char ksym_name[KSYM_NAME_LEN] = "pid_max"; module_param_string(ksym, ksym_name, KSYM_NAME_LEN, S_IRUGO); @@ -61,8 +61,8 @@ static int __init hw_break_module_init(void) attr.bp_type = HW_BREAKPOINT_W | HW_BREAKPOINT_R; sample_hbp = register_wide_hw_breakpoint(&attr, sample_hbp_handler); - if (IS_ERR(sample_hbp)) { - ret = PTR_ERR(sample_hbp); + if (IS_ERR((void __force *)sample_hbp)) { + ret = PTR_ERR((void __force *)sample_hbp); goto fail; } diff --git a/tools/perf/Documentation/perf-lock.txt b/tools/perf/Documentation/perf-lock.txt new file mode 100644 index 0000000..b317102 --- /dev/null +++ b/tools/perf/Documentation/perf-lock.txt @@ -0,0 +1,29 @@ +perf-lock(1) +============ + +NAME +---- +perf-lock - Analyze lock events + +SYNOPSIS +-------- +[verse] +'perf lock' {record|report|trace} + +DESCRIPTION +----------- +You can analyze various lock behaviours +and statistics with this 'perf lock' command. + + 'perf lock record ' records lock events + between start and end . And this command + produces the file "perf.data" which contains tracing + results of lock events. + + 'perf lock trace' shows raw lock events. + + 'perf lock report' reports statistical data. + +SEE ALSO +-------- +linkperf:perf[1] diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c index fb9ab2a..e12c844 100644 --- a/tools/perf/builtin-lock.c +++ b/tools/perf/builtin-lock.c @@ -460,6 +460,150 @@ process_raw_event(void *data, int cpu, process_lock_release_event(data, event, cpu, timestamp, thread); } +struct raw_event_queue { + u64 timestamp; + int cpu; + void *data; + struct thread *thread; + struct list_head list; +}; + +static LIST_HEAD(raw_event_head); + +#define FLUSH_PERIOD (5 * NSEC_PER_SEC) + +static u64 flush_limit = ULLONG_MAX; +static u64 last_flush = 0; +struct raw_event_queue *last_inserted; + +static void flush_raw_event_queue(u64 limit) +{ + struct raw_event_queue *tmp, *iter; + + list_for_each_entry_safe(iter, tmp, &raw_event_head, list) { + if (iter->timestamp > limit) + return; + + if (iter == last_inserted) + last_inserted = NULL; + + process_raw_event(iter->data, iter->cpu, iter->timestamp, + iter->thread); + + last_flush = iter->timestamp; + list_del(&iter->list); + free(iter->data); + free(iter); + } +} + +static void __queue_raw_event_end(struct raw_event_queue *new) +{ + struct raw_event_queue *iter; + + list_for_each_entry_reverse(iter, &raw_event_head, list) { + if (iter->timestamp < new->timestamp) { + list_add(&new->list, &iter->list); + return; + } + } + + list_add(&new->list, &raw_event_head); +} + +static void __queue_raw_event_before(struct raw_event_queue *new, + struct raw_event_queue *iter) +{ + list_for_each_entry_continue_reverse(iter, &raw_event_head, list) { + if (iter->timestamp < new->timestamp) { + list_add(&new->list, &iter->list); + return; + } + } + + list_add(&new->list, &raw_event_head); +} + +static void __queue_raw_event_after(struct raw_event_queue *new, + struct raw_event_queue *iter) +{ + list_for_each_entry_continue(iter, &raw_event_head, list) { + if (iter->timestamp > new->timestamp) { + list_add_tail(&new->list, &iter->list); + return; + } + } + list_add_tail(&new->list, &raw_event_head); +} + +/* The queue is ordered by time */ +static void __queue_raw_event(struct raw_event_queue *new) +{ + if (!last_inserted) { + __queue_raw_event_end(new); + return; + } + + /* + * Most of the time the current event has a timestamp + * very close to the last event inserted, unless we just switched + * to another event buffer. Having a sorting based on a list and + * on the last inserted event that is close to the current one is + * probably more efficient than an rbtree based sorting. + */ + if (last_inserted->timestamp >= new->timestamp) + __queue_raw_event_before(new, last_inserted); + else + __queue_raw_event_after(new, last_inserted); +} + +static void queue_raw_event(void *data, int raw_size, int cpu, + u64 timestamp, struct thread *thread) +{ + struct raw_event_queue *new; + + if (flush_limit == ULLONG_MAX) + flush_limit = timestamp + FLUSH_PERIOD; + + if (timestamp < last_flush) { + printf("Warning: Timestamp below last timeslice flush\n"); + return; + } + + new = malloc(sizeof(*new)); + if (!new) + die("Not enough memory\n"); + + new->timestamp = timestamp; + new->cpu = cpu; + new->thread = thread; + + new->data = malloc(raw_size); + if (!new->data) + die("Not enough memory\n"); + + memcpy(new->data, data, raw_size); + + __queue_raw_event(new); + last_inserted = new; + + /* + * We want to have a slice of events covering 2 * FLUSH_PERIOD + * If FLUSH_PERIOD is big enough, it ensures every events that occured + * in the first half of the timeslice have all been buffered and there + * are none remaining (we need that because of the weakly ordered + * event recording we have). Then once we reach the 2 * FLUSH_PERIOD + * timeslice, we flush the first half to be gentle with the memory + * (the second half can still get new events in the middle, so wait + * another period to flush it) + */ + if (new->timestamp > flush_limit && + new->timestamp - flush_limit > FLUSH_PERIOD) { + flush_limit += FLUSH_PERIOD; + flush_raw_event_queue(flush_limit); + } +} + static int process_sample_event(event_t *event, struct perf_session *session) { struct thread *thread; @@ -480,7 +624,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) if (profile_cpu != -1 && profile_cpu != (int) data.cpu) return 0; - process_raw_event(data.raw_data, data.cpu, data.time, thread); + queue_raw_event(data.raw_data, data.raw_size, data.cpu, data.time, thread); return 0; } @@ -576,6 +720,7 @@ static void __cmd_report(void) setup_pager(); select_key(); read_events(); + flush_raw_event_queue(ULLONG_MAX); sort_result(); print_result(); } @@ -608,7 +753,6 @@ static const char *record_args[] = { "record", "-a", "-R", - "-M", "-f", "-m", "1024", "-c", "1", diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 5db687f..407041d 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -573,7 +573,8 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) if (symbol__init() < 0) return -1; - setup_pager(); + if (!script_name) + setup_pager(); session = perf_session__new(input_name, O_RDONLY, 0); if (session == NULL) @@ -608,7 +609,6 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) return -1; } - perf_header__read(&session->header, input); err = scripting_ops->generate_script("perf-trace"); goto out; } diff --git a/tools/perf/command-list.txt b/tools/perf/command-list.txt index 9afcff2..db6ee94 100644 --- a/tools/perf/command-list.txt +++ b/tools/perf/command-list.txt @@ -18,3 +18,4 @@ perf-top mainporcelain common perf-trace mainporcelain common perf-probe mainporcelain common perf-kmem mainporcelain common +perf-lock mainporcelain common diff --git a/tools/perf/perf-archive.sh b/tools/perf/perf-archive.sh index 45fbe2f..910468e 100644 --- a/tools/perf/perf-archive.sh +++ b/tools/perf/perf-archive.sh @@ -9,8 +9,9 @@ fi DEBUGDIR=~/.debug/ BUILDIDS=$(mktemp /tmp/perf-archive-buildids.XXXXXX) +NOBUILDID=0000000000000000000000000000000000000000 -perf buildid-list -i $PERF_DATA --with-hits > $BUILDIDS +perf buildid-list -i $PERF_DATA --with-hits | grep -v "^$NOBUILDID " > $BUILDIDS if [ ! -s $BUILDIDS ] ; then echo "perf archive: no build-ids found" rm -f $BUILDIDS diff --git a/tools/perf/perf.h b/tools/perf/perf.h index 75f941b..6fb379b 100644 --- a/tools/perf/perf.h +++ b/tools/perf/perf.h @@ -65,9 +65,7 @@ * Use the __kuser_memory_barrier helper in the CPU helper page. See * arch/arm/kernel/entry-armv.S in the kernel source for details. */ -#define rmb() asm volatile("mov r0, #0xffff0fff; mov lr, pc;" \ - "sub pc, r0, #95" ::: "r0", "lr", "cc", \ - "memory") +#define rmb() ((void(*)(void))0xffff0fa0)() #define cpu_relax() asm volatile("":::"memory") #endif diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index c971e81..53181db 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -508,8 +508,8 @@ void show_perf_probe_events(void) struct str_node *ent; setup_pager(); - memset(&pp, 0, sizeof(pp)); + fd = open_kprobe_events(O_RDONLY, 0); rawlist = get_trace_kprobe_event_rawlist(fd); close(fd);