Re: [RFC][PATCH] lockdep: WARN about local_irq_{en,dis}able in NMIcontext

From: David Miller
Date: Tue Apr 06 2010 - 07:22:42 EST

From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Date: Tue, 06 Apr 2010 13:20:19 +0200

> On Tue, 2010-04-06 at 04:13 -0700, David Miller wrote:
>> From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
>> Date: Tue, 06 Apr 2010 13:12:01 +0200
>> > One thing we can do is make the code WARN about this, how about
>> > something like the below
>> It's going to warn every bootup in cpu_clock() on x86.
> *sigh*, yes, we could hack around that I suppose.. it would be nice to
> automate this check though, I bet you don't fancy tracking down more
> such splats than you have to.
> You could of course insert the debug code into your arch routines but
> that would limit the coverage checks to whatever you happen to run.

Yes, f.e. you could add local_irq_*_nmi() or similar that don't
complain when called inside of an NMI.

I would certainly welcome this debugging facility, for sure!

It would have saved me two days of work this time, in fact.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at