From 61c32659b12c44e62de32fbf99f7e4ca783dc38b Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 18 Nov 2010 01:39:17 +0100 Subject: tracing: New flag to allow non privileged users to use a trace event This adds a new trace event internal flag that allows them to be used in perf by non privileged users in case of task bound tracing. This is desired for syscalls tracepoint because they don't leak global system informations, like some other tracepoints. Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Thomas Gleixner Cc: Steven Rostedt Cc: Li Zefan Cc: Jason Baron --- kernel/trace/trace_event_perf.c | 31 ++++++++++++++++++++++++++++++- 1 file changed, 30 insertions(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c index 39c059ca670e..19a359d5e6d5 100644 --- a/kernel/trace/trace_event_perf.c +++ b/kernel/trace/trace_event_perf.c @@ -21,17 +21,46 @@ typedef typeof(unsigned long [PERF_MAX_TRACE_SIZE / sizeof(unsigned long)]) /* Count the events in use (per event id, not per instance) */ static int total_ref_count; +static int perf_trace_event_perm(struct ftrace_event_call *tp_event, + struct perf_event *p_event) +{ + /* No tracing, just counting, so no obvious leak */ + if (!(p_event->attr.sample_type & PERF_SAMPLE_RAW)) + return 0; + + /* Some events are ok to be traced by non-root users... */ + if (p_event->attach_state == PERF_ATTACH_TASK) { + if (tp_event->flags & TRACE_EVENT_FL_CAP_ANY) + return 0; + } + + /* + * ...otherwise raw tracepoint data can be a severe data leak, + * only allow root to have these. + */ + if (perf_paranoid_tracepoint_raw() && !capable(CAP_SYS_ADMIN)) + return -EPERM; + + return 0; +} + static int perf_trace_event_init(struct ftrace_event_call *tp_event, struct perf_event *p_event) { struct hlist_head __percpu *list; - int ret = -ENOMEM; + int ret; int cpu; + ret = perf_trace_event_perm(tp_event, p_event); + if (ret) + return ret; + p_event->tp_event = tp_event; if (tp_event->perf_refcount++ > 0) return 0; + ret = -ENOMEM; + list = alloc_percpu(struct hlist_head); if (!list) goto fail; -- cgit v1.2.3 From 042957801626465492b9428860de39a3cb2a8219 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 12 Nov 2010 22:32:11 -0500 Subject: tracing/events: Show real number in array fields Currently we have in something like the sched_switch event: field:char prev_comm[TASK_COMM_LEN]; offset:12; size:16; signed:1; When a userspace tool such as perf tries to parse this, the TASK_COMM_LEN is meaningless. This is done because the TRACE_EVENT() macro simply uses a #len to show the string of the length. When the length is an enum, we get a string that means nothing for tools. By adding a static buffer and a mutex to protect it, we can store the string into that buffer with snprintf and show the actual number. Now we get: field:char prev_comm[16]; offset:12; size:16; signed:1; Something much more useful. Signed-off-by: Steven Rostedt --- include/linux/ftrace_event.h | 4 ++++ include/trace/ftrace.h | 14 ++++++++++---- kernel/trace/trace_events.c | 6 ++++++ kernel/trace/trace_export.c | 14 ++++++++++---- 4 files changed, 30 insertions(+), 8 deletions(-) (limited to 'kernel/trace') diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h index 725bf6bd39f7..47e3997f7b5c 100644 --- a/include/linux/ftrace_event.h +++ b/include/linux/ftrace_event.h @@ -225,6 +225,10 @@ enum { FILTER_PTR_STRING, }; +#define EVENT_STORAGE_SIZE 128 +extern struct mutex event_storage_mutex; +extern char event_storage[EVENT_STORAGE_SIZE]; + extern int trace_event_raw_init(struct ftrace_event_call *call); extern int trace_define_field(struct ftrace_event_call *call, const char *type, const char *name, int offset, int size, diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h index e718a917d897..e16610c208c9 100644 --- a/include/trace/ftrace.h +++ b/include/trace/ftrace.h @@ -296,13 +296,19 @@ static struct trace_event_functions ftrace_event_type_funcs_##call = { \ #undef __array #define __array(type, item, len) \ - BUILD_BUG_ON(len > MAX_FILTER_STR_VAL); \ - ret = trace_define_field(event_call, #type "[" #len "]", #item, \ + do { \ + mutex_lock(&event_storage_mutex); \ + BUILD_BUG_ON(len > MAX_FILTER_STR_VAL); \ + snprintf(event_storage, sizeof(event_storage), \ + "%s[%d]", #type, len); \ + ret = trace_define_field(event_call, event_storage, #item, \ offsetof(typeof(field), item), \ sizeof(field.item), \ is_signed_type(type), FILTER_OTHER); \ - if (ret) \ - return ret; + mutex_unlock(&event_storage_mutex); \ + if (ret) \ + return ret; \ + } while (0); #undef __dynamic_array #define __dynamic_array(type, item, len) \ diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 0725eeab1937..35fde09b81de 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -27,6 +27,12 @@ DEFINE_MUTEX(event_mutex); +DEFINE_MUTEX(event_storage_mutex); +EXPORT_SYMBOL_GPL(event_storage_mutex); + +char event_storage[EVENT_STORAGE_SIZE]; +EXPORT_SYMBOL_GPL(event_storage); + LIST_HEAD(ftrace_events); LIST_HEAD(ftrace_common_fields); diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 4ba44deaac25..4b74d71705c0 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -83,13 +83,19 @@ static void __always_unused ____ftrace_check_##name(void) \ #undef __array #define __array(type, item, len) \ - BUILD_BUG_ON(len > MAX_FILTER_STR_VAL); \ - ret = trace_define_field(event_call, #type "[" #len "]", #item, \ + do { \ + BUILD_BUG_ON(len > MAX_FILTER_STR_VAL); \ + mutex_lock(&event_storage_mutex); \ + snprintf(event_storage, sizeof(event_storage), \ + "%s[%d]", #type, len); \ + ret = trace_define_field(event_call, event_storage, #item, \ offsetof(typeof(field), item), \ sizeof(field.item), \ is_signed_type(type), FILTER_OTHER); \ - if (ret) \ - return ret; + mutex_unlock(&event_storage_mutex); \ + if (ret) \ + return ret; \ + } while (0); #undef __array_desc #define __array_desc(type, container, item, len) \ -- cgit v1.2.3 From 61a0d49c33c7fd57c14895e5b0760bd02b65ac1f Mon Sep 17 00:00:00 2001 From: Thomas Renninger Date: Mon, 3 Jan 2011 17:50:43 +0100 Subject: perf: Do not export power_frequency, but power_start event power_frequency moved to drivers/cpufreq/cpufreq.c which has to be compiled in, no need to export it. intel_idle can a be module though... Signed-off-by: Thomas Renninger Signed-off-by: Ingo Molnar Acked-by: Jean Pihet Cc: Jean Pihet Cc: Arjan van de Ven Cc: rjw@sisk.pl LKML-Reference: <1294073445-14812-2-git-send-email-trenn@suse.de> Signed-off-by: Ingo Molnar LKML-Reference: <1290072314-31155-2-git-send-email-trenn@suse.de> --- drivers/idle/intel_idle.c | 2 -- kernel/trace/power-traces.c | 2 +- 2 files changed, 1 insertion(+), 3 deletions(-) (limited to 'kernel/trace') diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c index c131d58bcb50..15783d5501a8 100644 --- a/drivers/idle/intel_idle.c +++ b/drivers/idle/intel_idle.c @@ -220,9 +220,7 @@ static int intel_idle(struct cpuidle_device *dev, struct cpuidle_state *state) kt_before = ktime_get_real(); stop_critical_timings(); -#ifndef MODULE trace_power_start(POWER_CSTATE, (eax >> 4) + 1, cpu); -#endif if (!need_resched()) { __monitor((void *)¤t_thread_info()->flags, 0, 0); diff --git a/kernel/trace/power-traces.c b/kernel/trace/power-traces.c index a22582a06161..0e0497d9fade 100644 --- a/kernel/trace/power-traces.c +++ b/kernel/trace/power-traces.c @@ -13,5 +13,5 @@ #define CREATE_TRACE_POINTS #include -EXPORT_TRACEPOINT_SYMBOL_GPL(power_frequency); +EXPORT_TRACEPOINT_SYMBOL_GPL(power_start); -- cgit v1.2.3 From 25e41933b58777f2d020c3b0186b430ea004ec28 Mon Sep 17 00:00:00 2001 From: Thomas Renninger Date: Mon, 3 Jan 2011 17:50:44 +0100 Subject: perf: Clean up power events by introducing new, more generic ones Add these new power trace events: power:cpu_idle power:cpu_frequency power:machine_suspend The old C-state/idle accounting events: power:power_start power:power_end Have now a replacement (but we are still keeping the old tracepoints for compatibility): power:cpu_idle and power:power_frequency is replaced with: power:cpu_frequency power:machine_suspend is newly introduced. Jean Pihet has a patch integrated into the generic layer (kernel/power/suspend.c) which will make use of it. the type= field got removed from both, it was never used and the type is differed by the event type itself. perf timechart userspace tool gets adjusted in a separate patch. Signed-off-by: Thomas Renninger Signed-off-by: Ingo Molnar Acked-by: Arjan van de Ven Acked-by: Jean Pihet Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Linus Torvalds Cc: rjw@sisk.pl LKML-Reference: <1294073445-14812-3-git-send-email-trenn@suse.de> Signed-off-by: Ingo Molnar LKML-Reference: <1290072314-31155-2-git-send-email-trenn@suse.de> --- arch/x86/kernel/process.c | 7 +++- arch/x86/kernel/process_32.c | 2 +- arch/x86/kernel/process_64.c | 2 + drivers/cpufreq/cpufreq.c | 1 + drivers/cpuidle/cpuidle.c | 1 + drivers/idle/intel_idle.c | 1 + include/trace/events/power.h | 98 ++++++++++++++++++++++++++++++++++++++++---- kernel/trace/Kconfig | 15 +++++++ kernel/trace/power-traces.c | 3 ++ 9 files changed, 119 insertions(+), 11 deletions(-) (limited to 'kernel/trace') diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index 96ed1aac543a..c852041bfc3d 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -373,6 +373,7 @@ void default_idle(void) { if (hlt_use_halt()) { trace_power_start(POWER_CSTATE, 1, smp_processor_id()); + trace_cpu_idle(1, smp_processor_id()); current_thread_info()->status &= ~TS_POLLING; /* * TS_POLLING-cleared state must be visible before we @@ -443,6 +444,7 @@ EXPORT_SYMBOL_GPL(cpu_idle_wait); void mwait_idle_with_hints(unsigned long ax, unsigned long cx) { trace_power_start(POWER_CSTATE, (ax>>4)+1, smp_processor_id()); + trace_cpu_idle((ax>>4)+1, smp_processor_id()); if (!need_resched()) { if (cpu_has(¤t_cpu_data, X86_FEATURE_CLFLUSH_MONITOR)) clflush((void *)¤t_thread_info()->flags); @@ -459,6 +461,7 @@ static void mwait_idle(void) { if (!need_resched()) { trace_power_start(POWER_CSTATE, 1, smp_processor_id()); + trace_cpu_idle(1, smp_processor_id()); if (cpu_has(¤t_cpu_data, X86_FEATURE_CLFLUSH_MONITOR)) clflush((void *)¤t_thread_info()->flags); @@ -480,10 +483,12 @@ static void mwait_idle(void) static void poll_idle(void) { trace_power_start(POWER_CSTATE, 0, smp_processor_id()); + trace_cpu_idle(0, smp_processor_id()); local_irq_enable(); while (!need_resched()) cpu_relax(); - trace_power_end(0); + trace_power_end(smp_processor_id()); + trace_cpu_idle(PWR_EVENT_EXIT, smp_processor_id()); } /* diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 96586c3cbbbf..4b9befa0e347 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -113,8 +113,8 @@ void cpu_idle(void) stop_critical_timings(); pm_idle(); start_critical_timings(); - trace_power_end(smp_processor_id()); + trace_cpu_idle(PWR_EVENT_EXIT, smp_processor_id()); } tick_nohz_restart_sched_tick(); preempt_enable_no_resched(); diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index b3d7a3a04f38..4c818a738396 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -142,6 +142,8 @@ void cpu_idle(void) start_critical_timings(); trace_power_end(smp_processor_id()); + trace_cpu_idle(PWR_EVENT_EXIT, + smp_processor_id()); /* In many cases the interrupt that ended idle has already called exit_idle. But some idle diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index c63a43823744..1109f6848a43 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -355,6 +355,7 @@ void cpufreq_notify_transition(struct cpufreq_freqs *freqs, unsigned int state) dprintk("FREQ: %lu - CPU: %lu", (unsigned long)freqs->new, (unsigned long)freqs->cpu); trace_power_frequency(POWER_PSTATE, freqs->new, freqs->cpu); + trace_cpu_frequency(freqs->new, freqs->cpu); srcu_notifier_call_chain(&cpufreq_transition_notifier_list, CPUFREQ_POSTCHANGE, freqs); if (likely(policy) && likely(policy->cpu == freqs->cpu)) diff --git a/drivers/cpuidle/cpuidle.c b/drivers/cpuidle/cpuidle.c index a50710843378..08d5f05378d9 100644 --- a/drivers/cpuidle/cpuidle.c +++ b/drivers/cpuidle/cpuidle.c @@ -107,6 +107,7 @@ static void cpuidle_idle_call(void) if (cpuidle_curr_governor->reflect) cpuidle_curr_governor->reflect(dev); trace_power_end(smp_processor_id()); + trace_cpu_idle(PWR_EVENT_EXIT, smp_processor_id()); } /** diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c index 15783d5501a8..56ac09d6c930 100644 --- a/drivers/idle/intel_idle.c +++ b/drivers/idle/intel_idle.c @@ -221,6 +221,7 @@ static int intel_idle(struct cpuidle_device *dev, struct cpuidle_state *state) stop_critical_timings(); trace_power_start(POWER_CSTATE, (eax >> 4) + 1, cpu); + trace_cpu_idle((eax >> 4) + 1, cpu); if (!need_resched()) { __monitor((void *)¤t_thread_info()->flags, 0, 0); diff --git a/include/trace/events/power.h b/include/trace/events/power.h index 286784d69b8f..1bcc2a8c00e2 100644 --- a/include/trace/events/power.h +++ b/include/trace/events/power.h @@ -7,16 +7,67 @@ #include #include -#ifndef _TRACE_POWER_ENUM_ -#define _TRACE_POWER_ENUM_ -enum { - POWER_NONE = 0, - POWER_CSTATE = 1, /* C-State */ - POWER_PSTATE = 2, /* Fequency change or DVFS */ - POWER_SSTATE = 3, /* Suspend */ -}; +DECLARE_EVENT_CLASS(cpu, + + TP_PROTO(unsigned int state, unsigned int cpu_id), + + TP_ARGS(state, cpu_id), + + TP_STRUCT__entry( + __field( u32, state ) + __field( u32, cpu_id ) + ), + + TP_fast_assign( + __entry->state = state; + __entry->cpu_id = cpu_id; + ), + + TP_printk("state=%lu cpu_id=%lu", (unsigned long)__entry->state, + (unsigned long)__entry->cpu_id) +); + +DEFINE_EVENT(cpu, cpu_idle, + + TP_PROTO(unsigned int state, unsigned int cpu_id), + + TP_ARGS(state, cpu_id) +); + +/* This file can get included multiple times, TRACE_HEADER_MULTI_READ at top */ +#ifndef _PWR_EVENT_AVOID_DOUBLE_DEFINING +#define _PWR_EVENT_AVOID_DOUBLE_DEFINING + +#define PWR_EVENT_EXIT -1 #endif +DEFINE_EVENT(cpu, cpu_frequency, + + TP_PROTO(unsigned int frequency, unsigned int cpu_id), + + TP_ARGS(frequency, cpu_id) +); + +TRACE_EVENT(machine_suspend, + + TP_PROTO(unsigned int state), + + TP_ARGS(state), + + TP_STRUCT__entry( + __field( u32, state ) + ), + + TP_fast_assign( + __entry->state = state; + ), + + TP_printk("state=%lu", (unsigned long)__entry->state) +); + +/* This code will be removed after deprecation time exceeded (2.6.41) */ +#ifdef CONFIG_EVENT_POWER_TRACING_DEPRECATED + /* * The power events are used for cpuidle & suspend (power_start, power_end) * and for cpufreq (power_frequency) @@ -75,6 +126,36 @@ TRACE_EVENT(power_end, ); +/* Deprecated dummy functions must be protected against multi-declartion */ +#ifndef _PWR_EVENT_AVOID_DOUBLE_DEFINING_DEPRECATED +#define _PWR_EVENT_AVOID_DOUBLE_DEFINING_DEPRECATED + +enum { + POWER_NONE = 0, + POWER_CSTATE = 1, + POWER_PSTATE = 2, +}; +#endif /* _PWR_EVENT_AVOID_DOUBLE_DEFINING_DEPRECATED */ + +#else /* CONFIG_EVENT_POWER_TRACING_DEPRECATED */ + +#ifndef _PWR_EVENT_AVOID_DOUBLE_DEFINING_DEPRECATED +#define _PWR_EVENT_AVOID_DOUBLE_DEFINING_DEPRECATED +enum { + POWER_NONE = 0, + POWER_CSTATE = 1, + POWER_PSTATE = 2, +}; + +/* These dummy declaration have to be ripped out when the deprecated + events get removed */ +static inline void trace_power_start(u64 type, u64 state, u64 cpuid) {}; +static inline void trace_power_end(u64 cpuid) {}; +static inline void trace_power_frequency(u64 type, u64 state, u64 cpuid) {}; +#endif /* _PWR_EVENT_AVOID_DOUBLE_DEFINING_DEPRECATED */ + +#endif /* CONFIG_EVENT_POWER_TRACING_DEPRECATED */ + /* * The clock events are used for clock enable/disable and for * clock rate change @@ -153,7 +234,6 @@ DEFINE_EVENT(power_domain, power_domain_target, TP_ARGS(name, state, cpu_id) ); - #endif /* _TRACE_POWER_H */ /* This part must be outside protection */ diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index ea37e2ff4164..14674dce77a6 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -69,6 +69,21 @@ config EVENT_TRACING select CONTEXT_SWITCH_TRACER bool +config EVENT_POWER_TRACING_DEPRECATED + depends on EVENT_TRACING + bool "Deprecated power event trace API, to be removed" + default y + help + Provides old power event types: + C-state/idle accounting events: + power:power_start + power:power_end + and old cpufreq accounting event: + power:power_frequency + This is for userspace compatibility + and will vanish after 5 kernel iterations, + namely 2.6.41. + config CONTEXT_SWITCH_TRACER bool diff --git a/kernel/trace/power-traces.c b/kernel/trace/power-traces.c index 0e0497d9fade..f55fcf61b223 100644 --- a/kernel/trace/power-traces.c +++ b/kernel/trace/power-traces.c @@ -13,5 +13,8 @@ #define CREATE_TRACE_POINTS #include +#ifdef EVENT_POWER_TRACING_DEPRECATED EXPORT_TRACEPOINT_SYMBOL_GPL(power_start); +#endif +EXPORT_TRACEPOINT_SYMBOL_GPL(cpu_idle); -- cgit v1.2.3