[PATCH] ARM: perf: Don't sleep while atomic when enabling per-cpu interrupts

Will Deacon will.deacon at arm.com
Wed Sep 10 11:21:37 PDT 2014


On Tue, Sep 09, 2014 at 06:54:45PM +0100, Stephen Boyd wrote:
> On 09/09/14 04:39, Will Deacon wrote:
> > It's interesting that arm64 isn't affected by this problem, since we don't
> > update the active_irqs mask for PPIs there and consequently just pass the
> > irq instead of the cpu_pmu. I can't see why we actually need to update the
> > active_irqs mask for arch/arm/, so could we remove that and follow arm64's
> > lead instead? That would remove the need for a new struct definition too.
> >
> 
> I guess you're saying that we don't need the active_irqs mask in the
> percpu irq case? It looks like we still use it to determine when the
> last CPU PMU has been disabled in the non-percpu case.

Correct.

> Here's the interdiff. Is there a reason arm64 casts data to an unsigned
> int pointer when what's passed is an int pointer?

There has to be a cast to something because data is a void *.
enable_percpu_irq takes an unsigned int, so I guess that's why it was
chosen. I'm not fussed either way.

Feel free to submit the full patch with my ack:

  Acked-by: Will Deacon <will.deacon at arm.com>

Will



More information about the linux-arm-kernel mailing list