Re: New interrupt storm on ioapic0 pin 9 with -current 5.99.11

[Andrew Doran <ad%netbsd.org@localhost>]

On Wed, Apr 22, 2009 at 02:45:45PM +0200, Nicolas Joly wrote:

> On Wed, Apr 22, 2009 at 12:06:02PM +0000, Andrew Doran wrote:
> > On Wed, Apr 22, 2009 at 01:32:18PM +0200, Nicolas Joly wrote:
> > > On Mon, Apr 20, 2009 at 03:26:20PM +0200, Nicolas Joly wrote:
> > > > 
> > > > Hi,
> > > > 
> > > > I just updated my -current NetBSD/amd64 workstation, from sources
> > > > around April 10, and i not get an interrupt storm (about 60000 / sec)
> > > > on ioapic0 pin 9 / CPU0 ... which wasn't there previously.
> > > 
> > > It seems that the following commit is the culprit ...
> > > 
> > > http://mail-index.netbsd.org/source-changes/2009/04/19/msg220096.html
> > > 
> > > Reverting it make the storm disappear.
> > 
> > I have been offline. I'll sort it out later today along with the compile
> > error.
> > 
> > If you have ddb and a way to capture output, it would be useful to see the
> > output of:
> > 
> > call ioapic_dump
> > call intr_printconfig
> 
> [... lots of debug output ...]

Thanks, that is useful. I don't understand where the screwup is yet but
it has to be something fairly fundamental in our apic/spl code. It is
disappointing in one respect yet good because it highlights a bug.

Anyway, I have modified it so that it should direct all interrupts
back to the boot CPU again, for the time being (I have not been able
to test this today).