[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-arm-kernel
Subject:    Re: SA_INTERRUPT note
From:       Phil Blundell <pb () nexus ! co ! uk>
Date:       2003-10-31 13:59:16
[Download RAW message or body]

on Fri, Oct 31, 2003 at 08:44:07AM -0500, Zwane Mwaikambo wrote:
> On Fri, 31 Oct 2003, Phil Blundell wrote:
> 
> > on Wed, Oct 29, 2003 at 10:26:57AM -0500, Zwane Mwaikambo wrote:
> > > In which case shouldn't we be doing the following for the shared interrupt 
> > > case?
> > 
> > All the handlers for a particular interrupt have to agree on whether
> > it's SA_INTERRUPT or not, so this ought not to be necessary.
> 
> In setup_irq that appears only to be true for SA_SHIRQ, perhaps i'm 
> missing something else?

Mmm, seems you're right.  I'm sure that request_irq used to check for
mismatches at one time, but it seems not to any more.  SA_INTERRUPT has
been through a few changes in semantics over the years, including if I
recall correctly a period where it had no effect at all.

A quick survey of some other architectures in 2.4 suggests that i386
behaves the same as arm, alpha behaves as you were suggesting, and sparc
behaves in the way I was thinking of.

p.

-------------------------------------------------------------------
Subscription options: http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm-kernel
FAQ/Etiquette:       http://www.arm.linux.org.uk/armlinux/mailinglists.php

[prev in list] [next in list] [prev in thread] [next in thread]