In certain circumstances it may not be possible to schedule particular events due to constraints other than a lack of hardware counters (e.g. on big.LITTLE systems where CPUs support different events). The core perf event code does not distinguish these cases and pessimistically assumes that any failure to schedule an events is due to a lack of hardware counters, ending event group scheduling early despite hardware counters remaining available.
When such an unschedulable event exists in a ctx->flexible_groups list it can unnecessarily prevent event groups following it in the list from being scheduled until it is rotated to the end of the list. This can result in events being scheduled for only a portion of the time they would otherwise be eligible, and for short running programs unfortunate initial list ordering can result in no events being counted. This patch adds a new (optional) filter_match function pointer to struct pmu which backends can use to tell the perf core whether or not it is worth attempting to schedule an event. This plugs into the existing event_filter_match logic, and makes it possible to avoid the scheduling problem described above. Signed-off-by: Mark Rutland <[email protected]> Cc: Peter Zijlstra <[email protected]> Cc: Paul Mackerras <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Arnaldo Carvalho de Melo <[email protected]> --- include/linux/perf_event.h | 5 +++++ kernel/events/core.c | 8 +++++++- 2 files changed, 12 insertions(+), 1 deletion(-) diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 893a0d0..80c5f5f 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -263,6 +263,11 @@ struct pmu { * flush branch stack on context-switches (needed in cpu-wide mode) */ void (*flush_branch_stack) (void); + + /* + * Filter events for PMU-specific reasons. + */ + int (*filter_match) (struct perf_event *event); /* optional */ }; /** diff --git a/kernel/events/core.c b/kernel/events/core.c index 2b02c9f..770b276 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -1428,11 +1428,17 @@ static int __init perf_workqueue_init(void) core_initcall(perf_workqueue_init); +static inline int pmu_filter_match(struct perf_event *event) +{ + struct pmu *pmu = event->pmu; + return pmu->filter_match ? pmu->filter_match(event) : 1; +} + static inline int event_filter_match(struct perf_event *event) { return (event->cpu == -1 || event->cpu == smp_processor_id()) - && perf_cgroup_match(event); + && perf_cgroup_match(event) && pmu_filter_match(event); } static void -- 1.9.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
