diff options
author | Peter Zijlstra <peterz@infradead.org> | 2020-10-05 11:09:06 +0300 |
---|---|---|
committer | Peter Zijlstra <peterz@infradead.org> | 2020-10-06 16:18:17 +0300 |
commit | 871a93b0aad65a7f44ee25f2d17932ef6d559850 (patch) | |
tree | f335936c9c50598f143aebb2436185dddf02b9de | |
parent | 59d5396a4666195f89a67e118e9e627ddd6f53a1 (diff) | |
download | linux-871a93b0aad65a7f44ee25f2d17932ef6d559850.tar.xz |
perf/x86: Fix n_pair for cancelled txn
Kan reported that n_metric gets corrupted for cancelled transactions;
a similar issue exists for n_pair for AMD's Large Increment thing.
The problem was confirmed and confirmed fixed by Kim using:
sudo perf stat -e "{cycles,cycles,cycles,cycles}:D" -a sleep 10 &
# should succeed:
sudo perf stat -e "{fp_ret_sse_avx_ops.all}:D" -a workload
# should fail:
sudo perf stat -e "{fp_ret_sse_avx_ops.all,fp_ret_sse_avx_ops.all,cycles}:D" -a workload
# previously failed, now succeeds with this patch:
sudo perf stat -e "{fp_ret_sse_avx_ops.all}:D" -a workload
Fixes: 5738891229a2 ("perf/x86/amd: Add support for Large Increment per Cycle Events")
Reported-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Kim Phillips <kim.phillips@amd.com>
Link: https://lkml.kernel.org/r/20201005082516.GG2628@hirez.programming.kicks-ass.net
-rw-r--r-- | arch/x86/events/core.c | 6 | ||||
-rw-r--r-- | arch/x86/events/perf_event.h | 1 |
2 files changed, 6 insertions, 1 deletions
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c index cb5cfef3a938..a7248a3c4b2f 100644 --- a/arch/x86/events/core.c +++ b/arch/x86/events/core.c @@ -1064,8 +1064,10 @@ static int collect_event(struct cpu_hw_events *cpuc, struct perf_event *event, return -EINVAL; cpuc->event_list[n] = event; - if (is_counter_pair(&event->hw)) + if (is_counter_pair(&event->hw)) { cpuc->n_pair++; + cpuc->n_txn_pair++; + } return 0; } @@ -2006,6 +2008,7 @@ static void x86_pmu_start_txn(struct pmu *pmu, unsigned int txn_flags) perf_pmu_disable(pmu); __this_cpu_write(cpu_hw_events.n_txn, 0); + __this_cpu_write(cpu_hw_events.n_txn_pair, 0); } /* @@ -2031,6 +2034,7 @@ static void x86_pmu_cancel_txn(struct pmu *pmu) */ __this_cpu_sub(cpu_hw_events.n_added, __this_cpu_read(cpu_hw_events.n_txn)); __this_cpu_sub(cpu_hw_events.n_events, __this_cpu_read(cpu_hw_events.n_txn)); + __this_cpu_sub(cpu_hw_events.n_pair, __this_cpu_read(cpu_hw_events.n_txn_pair)); perf_pmu_enable(pmu); } diff --git a/arch/x86/events/perf_event.h b/arch/x86/events/perf_event.h index 345442410a4d..93e56d76980f 100644 --- a/arch/x86/events/perf_event.h +++ b/arch/x86/events/perf_event.h @@ -235,6 +235,7 @@ struct cpu_hw_events { they've never been enabled yet */ int n_txn; /* the # last events in the below arrays; added in the current transaction */ + int n_txn_pair; int assign[X86_PMC_IDX_MAX]; /* event to counter assignment */ u64 tags[X86_PMC_IDX_MAX]; |