On Thu, 15 Sep 2016 18:32:02 +0530 Madhavan Srinivasan <ma...@linux.vnet.ibm.com> wrote:
> diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c > index 9e5e9a6d4147..ae31b1e85fdb 100644 > --- a/arch/powerpc/kernel/irq.c > +++ b/arch/powerpc/kernel/irq.c > @@ -209,6 +209,10 @@ notrace void arch_local_irq_restore(unsigned long en) > unsigned char irq_happened; > unsigned int replay; > > +#ifdef CONFIG_IRQ_DEBUG_SUPPORT > + WARN_ON(en & local_paca->soft_enabled & ~IRQ_DISABLE_MASK_LINUX); > +#endif > + > /* Write the new soft-enabled value */ > soft_enabled_set(en); > Oh one other quick thing I just noticed: you could put this debug check into your soft_enabled accessors. We did decide it's okay for your masking level to go both ways, didn't we? I.e., local_irq_disable(); local_irq_pmu_save(flags); local_irq_pmu_restore(flags); local_irq_enable(); -> LINUX -> LINUX|PMU -> LINUX -> This means PMU interrupts would not get replayed despite being enabled here. In practice I think that doesn't matter/can't happen because a PMU interrupt while masked would hard disable anyway. A comment explaining it might be nice though. Thanks, Nick