cvs commit: src/sys/kern vfs_subr.c

[Paul Richards]

On Fri, Aug 20, 2004 at 03:19:40PM -0700, Nate Lawson wrote:
> Peter Wemm wrote:
> >On Friday 20 August 2004 12:21 pm, Don Lewis wrote:
> >
> >> Modified files:
> >>   sys/kern             vfs_subr.c
> >> Log:
> >> Don't attempt to trigger the syncer thread final sync code in the
> >> shutdown_pre_sync state if the RB_NOSYNC flag is set.  This is the
> >> likely cause of hangs after a system panic that are keeping crash
> >> dumps from being done.
> >
> >
> >The hang that I saw was because interrupts are broken (they just stop 
> >for no apparent reason) and the ata_shutdown sync hooks are not in 
> >polled mode (which is correct - interrupts are supposed to be enabled 
> >at this point, and are for normal shutdowns). 
> 
> Did you got the "interrupt storm for xxx" message before this happened? 
>  I think the check for this is too sensitive -- it is triggered on 
> resume by my an0 card.  That's the only way I've seen interrupts stop 
> occuring during normal operation.

It's been happening on one of my boxes too. I can't boot it at all
with ACPI, I don't get any routing, without ACPI it runs for a while
then all of a sudden box starts getting ATA DMA timeouts because
interrupts stop working.

I was going to send some dmesg from the two boot scenarios but it
doesn't boot up at all anymore, just hangs I think just after the
timecounter/cpu throttle steps messages, I think that's the point
where it's probing for ATA devices - which might still be interrupt
issues.

It's not very useable considering it's -stable :-)

Paul.