Date: Fri, 21 Aug 1998 00:32:54 +0200 (CEST) From: Geert Uytterhoeven To: LUTHER Sven cc: "'linux-apus@sunsite.auc.dk'" , Linux/m68k Subject: L68K: Re: TR: IDE shared interrupt patch In-Reply-To: Sender: owner-linux-m68k@phil.uni-sb.de On Thu, 20 Aug 1998, LUTHER Sven wrote: > > you will find here the patch for the ide stuff : > > > > Friendly, > > > > Sven > > > > --- drivers/block/ide.c.orig Thu Aug 20 14:58:56 1998 > > +++ drivers/block/ide.c Thu Aug 20 15:00:20 1998 > > @@ -1406,8 +1406,12 @@ > > * Probably not a shared PCI interrupt, > > * so we can safely try to do something about it: > > */ > > - (void)ide_ack_intr(hwif); > > - unexpected_intr(irq, hwgroup); > > + /* But don't forget that there are other > > architectures > > + * than i386, so better test the result of > > + * ide_ack_intr(hwif). > > + */ > > + if (ide_ack_intr(hwif)) > > + unexpected_intr(irq, hwgroup) ; > > } > > spin_unlock_irqrestore(&hwgroup->spinlock, flags); > > return; Actually Michael was right: we should always check the result of ide_ack_intr(). If we don't do that, hwgroup->handler() is still called if the IDE driver expected an interrupt but some other device generated one. Probably this explains the other IDE related messages we sometimes see on our consoles. The patch is untested, I already have more than enough troubles with the latest IDE integration patches (will we make it before 2.2? Let's hope so). --- jes-2.1.115/drivers/block/ide.c Sun Aug 9 21:47:47 1998 +++ m68k/drivers/block/ide.c Fri Aug 21 00:26:21 1998 @@ -1385,6 +1385,11 @@ __cli(); /* local CPU only */ spin_lock_irqsave(&hwgroup->spinlock, flags); hwif = hwgroup->hwif; + if (!ide_ack_intr(hwif)) { + spin_unlock_irqrestore(&hwgroup->spinlock, flags); + return; + } + if ((handler = hwgroup->handler) == NULL || hwgroup->poll_timeout != 0) { /* * Not expecting an interrupt from this drive. @@ -1406,7 +1411,6 @@ * Probably not a shared PCI interrupt, * so we can safely try to do something about it: */ - (void)ide_ack_intr(hwif); unexpected_intr(irq, hwgroup); } spin_unlock_irqrestore(&hwgroup->spinlock, flags); @@ -1418,7 +1422,6 @@ return; } hwgroup->handler = NULL; - (void)ide_ack_intr(hwif); del_timer(&(hwgroup->timer)); spin_unlock_irqrestore(&hwgroup->spinlock, flags); if (drive->unmask) Greetings, Geert -- Geert Uytterhoeven Geert.Uytterhoeven@cs.kuleuven.ac.be Wavelets, Linux/{m68k~Amiga,PPC~CHRP} http://www.cs.kuleuven.ac.be/~geert/ Department of Computer Science -- Katholieke Universiteit Leuven -- Belgium