Re: [PATCH][RFC] per isr in_progress markers

[Stephan von Krawczynski <skraw@ithnet.com>]

On Mon, 9 Sep 2002 21:37:42 +0200 (SAST)
Zwane Mwaikambo <zwane@mwaikambo.name> wrote:

> > (Btw, if there is, that would also allow us to notice the "constantly
> > screaming PCI interrupt" without help from the low-level isrs)
> 
> As an aside, i just had an idea for another way to improve interrupt 
> handling latency. Instead of walking through all the isrs in the chain, 
> we can have an isr flag wether it was the source of the irq, and if so we 
> stop right there and not walk through the other isrs. Obviously taking 
> into account that some devices are dumb and have no real way of 
> determining.
> 
> 	Zwane

Hello,

a short note on that: this proved to be a particularly bad idea back in the
amiga-days. All that happened with this idea (Amiga which has basically only 2
usable interrupts has heavy interrupt sharing) is that every good programmer
told the system that it was not the source of the ongoing interrupt - even if
it was - because otherwise you lost interrupts in heavy-load environment.
Shared interrupts _can_ work well, but you have to do short interrupt-routines
and don't mess the thing by over-intelligent (in fact non-atomic) operation.

Regards,
Stephan


> 
> -- 
> function.linuxpower.ca
> 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>