[PATCH v3] powerpc/ppc64: Use preempt_schedule_irq instead of preempt_schedule

Benjamin Herrenschmidt benh at kernel.crashing.org
Thu Oct 29 08:37:14 EST 2009


On Thu, 2009-10-29 at 00:28 +0300, Valentine wrote:
> Benjamin Herrenschmidt wrote:
> > On Wed, 2009-10-28 at 22:19 +0300, Valentine wrote:
> > 
> >> I'm just not sure that we need to clear HARDIRQEN here, since we don't 
> >> really hard-disable the the interrupts.
> > 
> > We do, or rather, we come in with the interrupts hard disabled, no ?
> 
> Yes, looks like the interrupts are disabled at this point (before 
> preempt_schedule_irq) most of the times, but we don't hard-disable them 
> here. We just soft-disable them to make preempt_schedule_irq happy. Even 
> if an interrupt fires, it will be hard-disabled and the hardirq flag 
> will be cleared by the exception handler right away. I just think that 
> there's no need to clear hardirq flag if we don't clear MSR_EE.

My point is that MSR_EE _is_ already clear... isn't it ? So either we
set it back, or we clear HARDIRQEN to reflect it. It will be re-enable
as soon as preempt_schedule_irq() calls local_irq_enable() which is soon
enough anyways.

Also that avoids perf interrupt sneaking in since those act as NMIs in
that regard and -will- get in even when soft disabled.

Cheers,
Ben.

> Thanks,
> Val.
> > 
> > Ben.
> > 
> >> Thanks,
> >> Val.
> >>
> >>> +	TRACE_DISABLE_INTS
> >>> +
> >>> +	/* Call the scheduler with soft IRQs off */
> >>> +1:	bl	.preempt_schedule_irq
> >>> +
> >>> +	/* Hard-disable interrupts again (and update PACA) */
> >>>  #ifdef CONFIG_PPC_BOOK3E
> >>> -	wrteei	1
> >>> -	bl	.preempt_schedule
> >>>  	wrteei	0
> >>>  #else
> >>> -	ori	r10,r10,MSR_EE
> >>> -	mtmsrd	r10,1		/* reenable interrupts */
> >>> -	bl	.preempt_schedule
> >>>  	mfmsr	r10
> >>> -	clrrdi	r9,r1,THREAD_SHIFT
> >>> -	rldicl	r10,r10,48,1	/* disable interrupts again */
> >>> +	rldicl	r10,r10,48,1
> >>>  	rotldi	r10,r10,16
> >>>  	mtmsrd	r10,1
> >>>  #endif /* CONFIG_PPC_BOOK3E */
> >>> +	li	r0,0
> >>> +	stb	r0,PACAHARDIRQEN(r13)
> >>> +
> >>> +	/* Re-test flags and eventually loop */
> >>> +	clrrdi	r9,r1,THREAD_SHIFT
> >>>  	ld	r4,TI_FLAGS(r9)
> >>>  	andi.	r0,r4,_TIF_NEED_RESCHED
> >>>  	bne	1b
> >>>  	b	restore
> >>>  
> >>>  user_work:
> >>> -#endif
> >>> +#endif /* CONFIG_PREEMPT */
> >>> +
> >>>  	/* Enable interrupts */
> >>>  #ifdef CONFIG_PPC_BOOK3E
> >>>  	wrteei	1
> > 
> > 




More information about the Linuxppc-dev mailing list