Messages in this thread Patch in this message |  | | From | kan.liang@linux ... | Subject | [PATCH 2/4] perf/x86/intel: fix event update for auto-reload | Date | Mon, 18 Dec 2017 03:34:49 -0800 |
| |
From: Kan Liang <kan.liang@linux.intel.com>
There is bug when mmap read event->count with large PEBS enabled. Here is an example. #./read_count 0x71f0 0x122c0 0x1000000001c54 0x100000001257d 0x200000000bdc5
There is auto-reload mechanism enabled for PEBS events in fixed period mode. But the calculation of event->count does not take the auto-reload values into account. Anyone who read the event->count will get wrong result, e.g x86_pmu_read. Also, the calculation of hwc->period_left is wrong either. It impacts the accuracy of period for the first record in PEBS multiple records.
The issue is introduced with the auto-reload mechanism enabled by commit 851559e35fd5 ("perf/x86/intel: Use the PEBS auto reload mechanism when possible")
For the auto-reload before the last time, it went through the whole period (reload value) every time. So period * times should be added into the event->count. For the last load, the elapsed delta (event-)time need to be corrected by adding the period (reload value). Because the start point is -period.
Signed-off-by: Kan Liang <kan.liang@linux.intel.com> --- arch/x86/events/core.c | 14 ++++++++++++++ arch/x86/events/intel/ds.c | 8 +++++++- 2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c index 35552ea..f74e21d 100644 --- a/arch/x86/events/core.c +++ b/arch/x86/events/core.c @@ -100,6 +100,20 @@ u64 x86_perf_event_update(struct perf_event *event, * of the count. */ delta = (new_raw_count << shift) - (prev_raw_count << shift); + + /* + * Take auto-reload into account + * For the auto-reload before the last time, it went through the + * whole period (reload_val) every time. + * Just simply add period * times to the event. + * + * For the last load, the elapsed delta (event-)time need to be + * corrected by adding the period. Because the start point is -period. + */ + if (reload_times > 0) { + delta += (reload_val << shift); + local64_add(reload_val * (reload_times - 1), &event->count); + } delta >>= shift; local64_add(delta, &event->count); diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c index 0b693b7..f0f6026 100644 --- a/arch/x86/events/intel/ds.c +++ b/arch/x86/events/intel/ds.c @@ -1256,11 +1256,17 @@ static void __intel_pmu_pebs_event(struct perf_event *event, void *base, void *top, int bit, int count) { + struct hw_perf_event *hwc = &event->hw; struct perf_sample_data data; struct pt_regs regs; void *at = get_next_pebs_record_by_bit(base, top, bit); - if (!intel_pmu_save_and_restart(event, 0, 0) && + /* + * Now, auto-reload is only enabled in fixed period mode. + * The reload value is always hwc->sample_period. + * May need to change it, if auto-reload is enabled in freq mode later. + */ + if (!intel_pmu_save_and_restart(event, hwc->sample_period, count - 1) && !(event->hw.flags & PERF_X86_EVENT_AUTO_RELOAD)) return; -- 2.7.4
|  |