[PATCH] highres: warn on softirq handling an unsafe callback

[PATCH] highres: warn on softirq handling an unsafe callback

[Andres Salomon]

[-- Attachment #1: Type: text/plain, Size: 299 bytes --]

run_hrtimer_queue is called from a softirq context, and should never run
a callback that is classified as HRTIMER_CB_IRQSAFE_NO_SOFTIRQ.  A bug
in unrelated code had managed to trigger that bug, so let's add an
explicit warning to this function.

Signed-off-by: Andres Salomon <dilinger@debian.org>

[-- Attachment #2: hrtimer5.patch --]
[-- Type: text/x-patch, Size: 482 bytes --]

diff --git a/include/linux/hrtimer.h b/include/linux/hrtimer.h
diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c
index 476cb0c..167c296 100644
--- a/kernel/hrtimer.c
+++ b/kernel/hrtimer.c
@@ -1130,6 +1130,7 @@ static inline void run_hrtimer_queue(struct hrtimer_cpu_base *cpu_base,
 		if (base->softirq_time.tv64 <= timer->expires.tv64)
 			break;
 
+		WARN_ON_ONCE(timer->cb_mode == HRTIMER_CB_IRQSAFE_NO_SOFTIRQ);
 		timer_stats_account_hrtimer(timer);
 
 		fn = timer->function;

Re: [PATCH] highres: warn on softirq handling an unsafe callback

[Ingo Molnar]

* Andres Salomon <dilinger@debian.org> wrote:

> run_hrtimer_queue is called from a softirq context, and should never 
> run a callback that is classified as HRTIMER_CB_IRQSAFE_NO_SOFTIRQ.  A 
> bug in unrelated code had managed to trigger that bug, so let's add an 
> explicit warning to this function.
> 
> Signed-off-by: Andres Salomon <dilinger@debian.org>

thanks.

Acked-by: Ingo Molnar <mingo@elte.hu>

[suggestion for future submissions: please include the patch body as 
text, not as a MIME attachment - this makes it easier for upstream 
maintainers to process such mails.]

	Ingo