[PATCH] preemptoff tracer: trace the preempt enable at preempt_schedule

[Todd Poynor]

preempt_schedule() uses the _notrace versions of the preemption
disable/enable routines for setting the PREEMPT_ACTIVE flag to avoid
a race condition per commit d1f74e2.  This causes the preemptoff
trace to not be properly terminated when the preempted thread is
rescheduled.  Since the original fix only converted the preempt
re-enable to use the _notrace version for symmetry, call the tracing
version instead.  The corresponding preemptoff start will be
traced by the preempt_disable() in schedule(), although this may
be relying on a fragile state of affairs, and it may be best to also
add a direct call to trace a preempt disable after preemption is
safely disabled and the race avoided.

Signed-off-by: Todd Poynor <toddpoynor@xxxxxxxxxx>
---
 kernel/sched.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/sched.c b/kernel/sched.c
index aa14a56..de83e30 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -4041,7 +4041,7 @@ asmlinkage void __sched notrace preempt_schedule(void)
 	do {
 		add_preempt_count_notrace(PREEMPT_ACTIVE);
 		schedule();
-		sub_preempt_count_notrace(PREEMPT_ACTIVE);
+		sub_preempt_count(PREEMPT_ACTIVE);
 
 		/*
 		 * Check again in case we missed a preemption opportunity
-- 
1.7.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/