aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorPeter Zijlstra <[email protected]>2016-03-18 17:31:27 +0100
committerIngo Molnar <[email protected]>2016-03-21 09:08:17 +0100
commite8d8a90fc5014e4cdcf1de2f320081d26c7f440f (patch)
treeadefe3bc51d02011ea1c6237d91bcadc0dec43f3
parent91a612eea9a316c464cc170ff8492ec09e7d1c69 (diff)
perf/x86/BTS: Fix RCU usage
This splat reminds us: [ 8166.045595] [ INFO: suspicious RCU usage. ] [ 8166.168972] [<ffffffff81127837>] lockdep_rcu_suspicious+0xe7/0x120 [ 8166.175966] [<ffffffff811e0bae>] perf_callchain+0x23e/0x250 [ 8166.182280] [<ffffffff811dda3d>] perf_prepare_sample+0x27d/0x350 [ 8166.189082] [<ffffffff8100f503>] intel_pmu_drain_bts_buffer+0x133/0x200 ... that as the core code does, one should hold rcu_read_lock() over that entire BTS event-output generation sequence as well. Signed-off-by: Peter Zijlstra (Intel) <[email protected]> Cc: Alexander Shishkin <[email protected]> Cc: Andy Lutomirski <[email protected]> Cc: Arnaldo Carvalho de Melo <[email protected]> Cc: Borislav Petkov <[email protected]> Cc: Brian Gerst <[email protected]> Cc: David Ahern <[email protected]> Cc: Denys Vlasenko <[email protected]> Cc: H. Peter Anvin <[email protected]> Cc: Jiri Olsa <[email protected]> Cc: Linus Torvalds <[email protected]> Cc: Namhyung Kim <[email protected]> Cc: Peter Zijlstra <[email protected]> Cc: Stephane Eranian <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Vince Weaver <[email protected]> Signed-off-by: Ingo Molnar <[email protected]>
-rw-r--r--arch/x86/events/intel/ds.c5
1 files changed, 4 insertions, 1 deletions
diff --git a/arch/x86/events/intel/ds.c b/arch/x86/events/intel/ds.c
index ce7211a07c0b..8584b90d8e0b 100644
--- a/arch/x86/events/intel/ds.c
+++ b/arch/x86/events/intel/ds.c
@@ -570,11 +570,12 @@ int intel_pmu_drain_bts_buffer(void)
* We will overwrite the from and to address before we output
* the sample.
*/
+ rcu_read_lock();
perf_prepare_sample(&header, &data, event, &regs);
if (perf_output_begin(&handle, event, header.size *
(top - base - skip)))
- return 1;
+ goto unlock;
for (at = base; at < top; at++) {
/* Filter out any records that contain kernel addresses. */
@@ -593,6 +594,8 @@ int intel_pmu_drain_bts_buffer(void)
/* There's new data available. */
event->hw.interrupts++;
event->pending_kill = POLL_IN;
+unlock:
+ rcu_read_unlock();
return 1;
}