Re: [PATCH] powerpc/irq: Run softirqs off the top of the irq stack

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Mon, 2013-09-23 at 17:56 +1000, Stephen Rothwell wrote:
> Hi Ben,
> 
> On Mon, 23 Sep 2013 14:35:58 +1000 Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:
> >
> > diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
> > index c69440c..0c9646f 100644
> > --- a/arch/powerpc/kernel/irq.c
> > +++ b/arch/powerpc/kernel/irq.c
> > @@ -443,46 +443,7 @@ void migrate_irqs(void)
> >  
> >  static inline void handle_one_irq(unsigned int irq)
> >  {
> > -	struct thread_info *curtp, *irqtp;
> > -	unsigned long saved_sp_limit;
> > -	struct irq_desc *desc;
> > -
> > -	desc = irq_to_desc(irq);
> > -	if (!desc)
> > -		return;
> > -
> > -	/* Switch to the irq stack to handle this */
> > -	curtp = current_thread_info();
> > -	irqtp = hardirq_ctx[smp_processor_id()];
> > -
> > -	if (curtp == irqtp) {
> > -		/* We're already on the irq stack, just handle it */
> > -		desc->handle_irq(irq, desc);
> > -		return;
> > -	}
> > -
> > -	saved_sp_limit = current->thread.ksp_limit;
> > -
> > -	irqtp->task = curtp->task;
> > -	irqtp->flags = 0;
> > -
> > -	/* Copy the softirq bits in preempt_count so that the
> > -	 * softirq checks work in the hardirq context. */
> > -	irqtp->preempt_count = (irqtp->preempt_count & ~SOFTIRQ_MASK) |
> > -			       (curtp->preempt_count & SOFTIRQ_MASK);
> >  
> > -	current->thread.ksp_limit = (unsigned long)irqtp +
> > -		_ALIGN_UP(sizeof(struct thread_info), 16);
> > -
> > -	call_handle_irq(irq, desc, irqtp, desc->handle_irq);
> > -	current->thread.ksp_limit = saved_sp_limit;
> > -	irqtp->task = NULL;
> > -
> > -	/* Set any flag that may have been set on the
> > -	 * alternate stack
> > -	 */
> > -	if (irqtp->flags)
> > -		set_bits(irqtp->flags, &curtp->flags);
> >  }
> 
> This function ends up as a single blank line ...
> 
> > @@ -519,18 +480,64 @@ void do_IRQ(struct pt_regs *regs)
> >  	 */
> >  	irq = ppc_md.get_irq();
> >  
> > -	/* We can hard enable interrupts now */
> > +	/* We can hard enable interrupts now to allow perf interrupts */
> >  	may_hard_irq_enable();
> >  
> >  	/* And finally process it */
> > -	if (irq != NO_IRQ)
> > -		handle_one_irq(irq);
> 
> then you remove the only call, so why not just remove the function
> completely?

Because I'm an idiot ? :-)

I moved bits and pieces to do_IRQ and forgot to remove the remainder
(and gcc didn't warn :-)

Cheers,
Ben.