perf_counter: rework context time
Since perf_counter_context is switched along with tasks, we can maintain the context time without using the task runtime clock. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Paul Mackerras <paulus@samba.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> LKML-Reference: <20090406094518.353552838@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
This commit is contained in:
		
					parent
					
						
							
								4c9e25428f
							
						
					
				
			
			
				commit
				
					
						4af4998b8a
					
				
			
		
					 2 changed files with 37 additions and 51 deletions
				
			
		|  | @ -477,14 +477,10 @@ struct perf_counter_context { | ||||||
| 	struct task_struct	*task; | 	struct task_struct	*task; | ||||||
| 
 | 
 | ||||||
| 	/*
 | 	/*
 | ||||||
| 	 * time_now is the current time in nanoseconds since an arbitrary | 	 * Context clock, runs when context enabled. | ||||||
| 	 * point in the past.  For per-task counters, this is based on the |  | ||||||
| 	 * task clock, and for per-cpu counters it is based on the cpu clock. |  | ||||||
| 	 * time_lost is an offset from the task/cpu clock, used to make it |  | ||||||
| 	 * appear that time only passes while the context is scheduled in. |  | ||||||
| 	 */ | 	 */ | ||||||
| 	u64			time_now; | 	u64			time; | ||||||
| 	u64			time_lost; | 	u64			timestamp; | ||||||
| #endif | #endif | ||||||
| }; | }; | ||||||
| 
 | 
 | ||||||
|  |  | ||||||
|  | @ -117,7 +117,7 @@ counter_sched_out(struct perf_counter *counter, | ||||||
| 		return; | 		return; | ||||||
| 
 | 
 | ||||||
| 	counter->state = PERF_COUNTER_STATE_INACTIVE; | 	counter->state = PERF_COUNTER_STATE_INACTIVE; | ||||||
| 	counter->tstamp_stopped = ctx->time_now; | 	counter->tstamp_stopped = ctx->time; | ||||||
| 	counter->hw_ops->disable(counter); | 	counter->hw_ops->disable(counter); | ||||||
| 	counter->oncpu = -1; | 	counter->oncpu = -1; | ||||||
| 
 | 
 | ||||||
|  | @ -253,27 +253,20 @@ retry: | ||||||
| 	spin_unlock_irq(&ctx->lock); | 	spin_unlock_irq(&ctx->lock); | ||||||
| } | } | ||||||
| 
 | 
 | ||||||
| /*
 | static inline u64 perf_clock(void) | ||||||
|  * Get the current time for this context. |  | ||||||
|  * If this is a task context, we use the task's task clock, |  | ||||||
|  * or for a per-cpu context, we use the cpu clock. |  | ||||||
|  */ |  | ||||||
| static u64 get_context_time(struct perf_counter_context *ctx, int update) |  | ||||||
| { | { | ||||||
| 	struct task_struct *curr = ctx->task; | 	return cpu_clock(smp_processor_id()); | ||||||
| 
 |  | ||||||
| 	if (!curr) |  | ||||||
| 		return cpu_clock(smp_processor_id()); |  | ||||||
| 
 |  | ||||||
| 	return __task_delta_exec(curr, update) + curr->se.sum_exec_runtime; |  | ||||||
| } | } | ||||||
| 
 | 
 | ||||||
| /*
 | /*
 | ||||||
|  * Update the record of the current time in a context. |  * Update the record of the current time in a context. | ||||||
|  */ |  */ | ||||||
| static void update_context_time(struct perf_counter_context *ctx, int update) | static void update_context_time(struct perf_counter_context *ctx) | ||||||
| { | { | ||||||
| 	ctx->time_now = get_context_time(ctx, update) - ctx->time_lost; | 	u64 now = perf_clock(); | ||||||
|  | 
 | ||||||
|  | 	ctx->time += now - ctx->timestamp; | ||||||
|  | 	ctx->timestamp = now; | ||||||
| } | } | ||||||
| 
 | 
 | ||||||
| /*
 | /*
 | ||||||
|  | @ -284,15 +277,17 @@ static void update_counter_times(struct perf_counter *counter) | ||||||
| 	struct perf_counter_context *ctx = counter->ctx; | 	struct perf_counter_context *ctx = counter->ctx; | ||||||
| 	u64 run_end; | 	u64 run_end; | ||||||
| 
 | 
 | ||||||
| 	if (counter->state >= PERF_COUNTER_STATE_INACTIVE) { | 	if (counter->state < PERF_COUNTER_STATE_INACTIVE) | ||||||
| 		counter->total_time_enabled = ctx->time_now - | 		return; | ||||||
| 			counter->tstamp_enabled; | 
 | ||||||
| 		if (counter->state == PERF_COUNTER_STATE_INACTIVE) | 	counter->total_time_enabled = ctx->time - counter->tstamp_enabled; | ||||||
| 			run_end = counter->tstamp_stopped; | 
 | ||||||
| 		else | 	if (counter->state == PERF_COUNTER_STATE_INACTIVE) | ||||||
| 			run_end = ctx->time_now; | 		run_end = counter->tstamp_stopped; | ||||||
| 		counter->total_time_running = run_end - counter->tstamp_running; | 	else | ||||||
| 	} | 		run_end = ctx->time; | ||||||
|  | 
 | ||||||
|  | 	counter->total_time_running = run_end - counter->tstamp_running; | ||||||
| } | } | ||||||
| 
 | 
 | ||||||
| /*
 | /*
 | ||||||
|  | @ -332,7 +327,7 @@ static void __perf_counter_disable(void *info) | ||||||
| 	 * If it is in error state, leave it in error state. | 	 * If it is in error state, leave it in error state. | ||||||
| 	 */ | 	 */ | ||||||
| 	if (counter->state >= PERF_COUNTER_STATE_INACTIVE) { | 	if (counter->state >= PERF_COUNTER_STATE_INACTIVE) { | ||||||
| 		update_context_time(ctx, 1); | 		update_context_time(ctx); | ||||||
| 		update_counter_times(counter); | 		update_counter_times(counter); | ||||||
| 		if (counter == counter->group_leader) | 		if (counter == counter->group_leader) | ||||||
| 			group_sched_out(counter, cpuctx, ctx); | 			group_sched_out(counter, cpuctx, ctx); | ||||||
|  | @ -426,7 +421,7 @@ counter_sched_in(struct perf_counter *counter, | ||||||
| 		return -EAGAIN; | 		return -EAGAIN; | ||||||
| 	} | 	} | ||||||
| 
 | 
 | ||||||
| 	counter->tstamp_running += ctx->time_now - counter->tstamp_stopped; | 	counter->tstamp_running += ctx->time - counter->tstamp_stopped; | ||||||
| 
 | 
 | ||||||
| 	if (!is_software_counter(counter)) | 	if (!is_software_counter(counter)) | ||||||
| 		cpuctx->active_oncpu++; | 		cpuctx->active_oncpu++; | ||||||
|  | @ -493,9 +488,9 @@ static void add_counter_to_ctx(struct perf_counter *counter, | ||||||
| 	list_add_counter(counter, ctx); | 	list_add_counter(counter, ctx); | ||||||
| 	ctx->nr_counters++; | 	ctx->nr_counters++; | ||||||
| 	counter->prev_state = PERF_COUNTER_STATE_OFF; | 	counter->prev_state = PERF_COUNTER_STATE_OFF; | ||||||
| 	counter->tstamp_enabled = ctx->time_now; | 	counter->tstamp_enabled = ctx->time; | ||||||
| 	counter->tstamp_running = ctx->time_now; | 	counter->tstamp_running = ctx->time; | ||||||
| 	counter->tstamp_stopped = ctx->time_now; | 	counter->tstamp_stopped = ctx->time; | ||||||
| } | } | ||||||
| 
 | 
 | ||||||
| /*
 | /*
 | ||||||
|  | @ -522,7 +517,7 @@ static void __perf_install_in_context(void *info) | ||||||
| 
 | 
 | ||||||
| 	curr_rq_lock_irq_save(&flags); | 	curr_rq_lock_irq_save(&flags); | ||||||
| 	spin_lock(&ctx->lock); | 	spin_lock(&ctx->lock); | ||||||
| 	update_context_time(ctx, 1); | 	update_context_time(ctx); | ||||||
| 
 | 
 | ||||||
| 	/*
 | 	/*
 | ||||||
| 	 * Protect the list operation against NMI by disabling the | 	 * Protect the list operation against NMI by disabling the | ||||||
|  | @ -648,13 +643,13 @@ static void __perf_counter_enable(void *info) | ||||||
| 
 | 
 | ||||||
| 	curr_rq_lock_irq_save(&flags); | 	curr_rq_lock_irq_save(&flags); | ||||||
| 	spin_lock(&ctx->lock); | 	spin_lock(&ctx->lock); | ||||||
| 	update_context_time(ctx, 1); | 	update_context_time(ctx); | ||||||
| 
 | 
 | ||||||
| 	counter->prev_state = counter->state; | 	counter->prev_state = counter->state; | ||||||
| 	if (counter->state >= PERF_COUNTER_STATE_INACTIVE) | 	if (counter->state >= PERF_COUNTER_STATE_INACTIVE) | ||||||
| 		goto unlock; | 		goto unlock; | ||||||
| 	counter->state = PERF_COUNTER_STATE_INACTIVE; | 	counter->state = PERF_COUNTER_STATE_INACTIVE; | ||||||
| 	counter->tstamp_enabled = ctx->time_now - counter->total_time_enabled; | 	counter->tstamp_enabled = ctx->time - counter->total_time_enabled; | ||||||
| 
 | 
 | ||||||
| 	/*
 | 	/*
 | ||||||
| 	 * If the counter is in a group and isn't the group leader, | 	 * If the counter is in a group and isn't the group leader, | ||||||
|  | @ -737,8 +732,8 @@ static void perf_counter_enable(struct perf_counter *counter) | ||||||
| 	 */ | 	 */ | ||||||
| 	if (counter->state == PERF_COUNTER_STATE_OFF) { | 	if (counter->state == PERF_COUNTER_STATE_OFF) { | ||||||
| 		counter->state = PERF_COUNTER_STATE_INACTIVE; | 		counter->state = PERF_COUNTER_STATE_INACTIVE; | ||||||
| 		counter->tstamp_enabled = ctx->time_now - | 		counter->tstamp_enabled = | ||||||
| 			counter->total_time_enabled; | 			ctx->time - counter->total_time_enabled; | ||||||
| 	} | 	} | ||||||
|  out: |  out: | ||||||
| 	spin_unlock_irq(&ctx->lock); | 	spin_unlock_irq(&ctx->lock); | ||||||
|  | @ -778,7 +773,7 @@ void __perf_counter_sched_out(struct perf_counter_context *ctx, | ||||||
| 	ctx->is_active = 0; | 	ctx->is_active = 0; | ||||||
| 	if (likely(!ctx->nr_counters)) | 	if (likely(!ctx->nr_counters)) | ||||||
| 		goto out; | 		goto out; | ||||||
| 	update_context_time(ctx, 0); | 	update_context_time(ctx); | ||||||
| 
 | 
 | ||||||
| 	flags = hw_perf_save_disable(); | 	flags = hw_perf_save_disable(); | ||||||
| 	if (ctx->nr_active) { | 	if (ctx->nr_active) { | ||||||
|  | @ -883,12 +878,7 @@ __perf_counter_sched_in(struct perf_counter_context *ctx, | ||||||
| 	if (likely(!ctx->nr_counters)) | 	if (likely(!ctx->nr_counters)) | ||||||
| 		goto out; | 		goto out; | ||||||
| 
 | 
 | ||||||
| 	/*
 | 	ctx->timestamp = perf_clock(); | ||||||
| 	 * Add any time since the last sched_out to the lost time |  | ||||||
| 	 * so it doesn't get included in the total_time_enabled and |  | ||||||
| 	 * total_time_running measures for counters in the context. |  | ||||||
| 	 */ |  | ||||||
| 	ctx->time_lost = get_context_time(ctx, 0) - ctx->time_now; |  | ||||||
| 
 | 
 | ||||||
| 	flags = hw_perf_save_disable(); | 	flags = hw_perf_save_disable(); | ||||||
| 
 | 
 | ||||||
|  | @ -1043,8 +1033,8 @@ int perf_counter_task_enable(void) | ||||||
| 		if (counter->state > PERF_COUNTER_STATE_OFF) | 		if (counter->state > PERF_COUNTER_STATE_OFF) | ||||||
| 			continue; | 			continue; | ||||||
| 		counter->state = PERF_COUNTER_STATE_INACTIVE; | 		counter->state = PERF_COUNTER_STATE_INACTIVE; | ||||||
| 		counter->tstamp_enabled = ctx->time_now - | 		counter->tstamp_enabled = | ||||||
| 			counter->total_time_enabled; | 			ctx->time - counter->total_time_enabled; | ||||||
| 		counter->hw_event.disabled = 0; | 		counter->hw_event.disabled = 0; | ||||||
| 	} | 	} | ||||||
| 	hw_perf_restore(perf_flags); | 	hw_perf_restore(perf_flags); | ||||||
|  | @ -1113,7 +1103,7 @@ static void __read(void *info) | ||||||
| 
 | 
 | ||||||
| 	curr_rq_lock_irq_save(&flags); | 	curr_rq_lock_irq_save(&flags); | ||||||
| 	if (ctx->is_active) | 	if (ctx->is_active) | ||||||
| 		update_context_time(ctx, 1); | 		update_context_time(ctx); | ||||||
| 	counter->hw_ops->read(counter); | 	counter->hw_ops->read(counter); | ||||||
| 	update_counter_times(counter); | 	update_counter_times(counter); | ||||||
| 	curr_rq_unlock_irq_restore(&flags); | 	curr_rq_unlock_irq_restore(&flags); | ||||||
|  |  | ||||||
		Loading…
	
	Add table
		Add a link
		
	
		Reference in a new issue
	
	 Peter Zijlstra
				Peter Zijlstra