[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <174601433215.22196.15691518087061204450.tip-bot2@tip-bot2>
Date: Wed, 30 Apr 2025 11:58:52 -0000
From: "tip-bot2 for Kan Liang" <tip-bot2@...utronix.de>
To: linux-tip-commits@...r.kernel.org
Cc: Kan Liang <kan.liang@...ux.intel.com>,
"Peter Zijlstra (Intel)" <peterz@...radead.org>, x86@...nel.org,
linux-kernel@...r.kernel.org
Subject: [tip: perf/core] perf/x86/intel/ds: Fix counter backwards of
non-precise events counters-snapshotting
The following commit has been merged into the perf/core branch of tip:
Commit-ID: 7da9960b59fb7e590eb8538c9428db55a4ea2d23
Gitweb: https://git.kernel.org/tip/7da9960b59fb7e590eb8538c9428db55a4ea2d23
Author: Kan Liang <kan.liang@...ux.intel.com>
AuthorDate: Thu, 24 Apr 2025 06:47:18 -07:00
Committer: Peter Zijlstra <peterz@...radead.org>
CommitterDate: Fri, 25 Apr 2025 14:55:19 +02:00
perf/x86/intel/ds: Fix counter backwards of non-precise events counters-snapshotting
The counter backwards may be observed in the PMI handler when
counters-snapshotting some non-precise events in the freq mode.
For the non-precise events, it's possible the counters-snapshotting
records a positive value for an overflowed PEBS event. Then the HW
auto-reload mechanism reset the counter to 0 immediately. Because the
pebs_event_reset is cleared in the freq mode, which doesn't set the
PERF_X86_EVENT_AUTO_RELOAD.
In the PMI handler, 0 will be read rather than the positive value
recorded in the counters-snapshotting record.
The counters-snapshotting case has to be specially handled. Since the
event value has been updated when processing the counters-snapshotting
record, only needs to set the new period for the counter via
x86_pmu_set_period().
Fixes: e02e9b0374c3 ("perf/x86/intel: Support PEBS counters snapshotting")
Signed-off-by: Kan Liang <kan.liang@...ux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
Link: https://lkml.kernel.org/r/20250424134718.311934-6-kan.liang@linux.intel.com
---
arch/x86/events/intel/ds.c | 21 +++++++++++++++++++--
1 file changed, 19 insertions(+), 2 deletions(-)
diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c
index 18c3ab5..9b20acc 100644
--- a/arch/x86/events/intel/ds.c
+++ b/arch/x86/events/intel/ds.c
@@ -2379,8 +2379,25 @@ __intel_pmu_pebs_last_event(struct perf_event *event,
*/
intel_pmu_save_and_restart_reload(event, count);
}
- } else
- intel_pmu_save_and_restart(event);
+ } else {
+ /*
+ * For a non-precise event, it's possible the
+ * counters-snapshotting records a positive value for the
+ * overflowed event. Then the HW auto-reload mechanism
+ * reset the counter to 0 immediately, because the
+ * pebs_event_reset is cleared if the PERF_X86_EVENT_AUTO_RELOAD
+ * is not set. The counter backwards may be observed in a
+ * PMI handler.
+ *
+ * Since the event value has been updated when processing the
+ * counters-snapshotting record, only needs to set the new
+ * period for the counter.
+ */
+ if (is_pebs_counter_event_group(event))
+ static_call(x86_pmu_set_period)(event);
+ else
+ intel_pmu_save_and_restart(event);
+ }
}
static __always_inline void
Powered by blists - more mailing lists