Re: new IRQ scalability changes in 2.3.48

From: yodaiken@fsmlabs.com
Date: Fri Mar 03 2000 - 09:32:54 EST


On Fri, Mar 03, 2000 at 04:16:16PM +0100, Ingo Molnar wrote:
>
> On Fri, 3 Mar 2000 yodaiken@fsmlabs.com wrote:
>
> > Since we need a spin lock anyways in doIRQ, I still don't get how
> > this makes things any better.
>
> if different IRQs are delivered to different CPUs, then there is no global
> spinlock connection between them. Also see /proc/irq/*/smp_affinity. Eg.

OK. In any event, I figured out how to fix it in RTLinux. It would be
easier for me if there was a macro
               low_level_irq_controller_spin_lock(x);
which could be defined as
               spin_lock_irqsave(x)
in standard Linux.
and as
                rtl_hard_irqsave_spinlock(x)
in RTLinux.
Also, it's critical that those locks _never_ be exposed outside of the
lowest level routines.

So can you answer my questions about the effect of postponing the
ack until after the handler executes. Does that mask further ioapic
interrupts?

-- 
---------------------------------------------------------
Victor Yodaiken 
FSMLabs:  www.fsmlabs.com  www.rtlinux.com
FSMLabs is a servicemark and a service of 
VJY Associates L.L.C, New Mexico.

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Mar 07 2000 - 21:00:14 EST