RE: [PATCH V2] smp: Give WARN()ing when callingsmp_call_function_many()/single() in serving irq

From: Liu, Chuansheng
Date: Tue Feb 19 2013 - 20:22:16 EST




> -----Original Message-----
> From: Wu, Fengguang
> Sent: Wednesday, February 20, 2013 9:07 AM
> To: Andrew Morton
> Cc: Liu, Chuansheng; mingo@xxxxxxxxxx; peterz@xxxxxxxxxxxxx;
> jbeulich@xxxxxxxx; paulmck@xxxxxxxxxxxxxxxxxx; mina86@xxxxxxxxxx;
> srivatsa.bhat@xxxxxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; Zhang, Jun
> Subject: Re: [PATCH V2] smp: Give WARN()ing when calling
> smp_call_function_many()/single() in serving irq
>
> On Tue, Feb 19, 2013 at 03:01:23PM -0800, Andrew Morton wrote:
> > On Mon, 18 Feb 2013 09:38:52 +0800
> > Fengguang Wu <fengguang.wu@xxxxxxxxx> wrote:
> >
> > > Chuansheng,
> > >
> > > It works fine on tip/next. Thanks for the fix!
> > >
> > > Tested-by: Fengguang Wu <fengguang.wu@xxxxxxxxx>
> >
> > How did you test this? What did you do?
>
> I used the same kconfig to build a kernel from tip/next HEAD *plus*
> the v2 patch. The result is, no smp_call_function_single/many warnings
> are observed.
>
> > afacit the patch adds additional warnings, so you must have had some
> > buggy code and with the patch, runtime warnings are generated which
> > inform you of that bug?
>
> Andrew, the background is: Ingo has *reverted* the original patch that
> triggered the smp_call_function_single warning from his tip tree, so
> the test result means this patch's warning messages won't trigger in
> my test boots.
I also locally tested the below several cases, thanks.
1/ local_bh_disable(), then called smp_call_function_single(), no WARNING;
2/ In timer callback, then called smp_call_function_single(), WARNING;

>
> Thanks,
> Fengguang
--
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/