Re: -current dom0 crash

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Thu, Feb 02, 2012 at 06:48:23AM -0800, Jeff Rizzo wrote:
> On 2/2/12 2:35 AM, Manuel Bouyer wrote:
> >On Wed, Feb 01, 2012 at 01:22:24PM -0800, Jeff Rizzo wrote:
> >>I just updated one of my dom0s from 5.99.59 to 5.99.63, and got this
> >>crash when destroying a hung domU (an operation that worked fine
> >>under 5.99.59):
> >>
> >>Feb  1 13:20:02 xenserver1 xenbackendd[335]: Failed to read
> >>/local/domain/0/backend/console/2/0/script (No such file or
> >>directory)
> >>xbd backend: detach device mirror-wintermu for domain 2
> >>fatal protection fault in supervisor mode
> >>trap type 4 code 0 rip ffffffff804f7425 cs e030 rflags 10286 cr2
> >>ffffa000117db7dc cpl 6 rsp ffffa00000a5f880
> >>kernel: protection fault trap, code=0
> >>Stopped in pid 0.31 (system) at netbsd:nd6_purge+0xb5:  movl
> >>14(%r12),%eax
> >>nd6_purge() at netbsd:nd6_purge+0xb5
> >>in6_ifdetach() at netbsd:in6_ifdetach+0x21
> >>udp6_usrreq() at netbsd:udp6_usrreq+0x208
> >>if_detach() at netbsd:if_detach+0x112
> >>xennetback_xenbus_destroy() at netbsd:xennetback_xenbus_destroy+0x8b
> >>otherend_changed() at netbsd:otherend_changed+0xaf
> >>xenwatch_thread() at netbsd:xenwatch_thread+0xa2
> >It looks like an issue in the ipv6 code on interface destroy (it may be
> >a race condition), so it may not be xen-specific.
> >Anyone remembers inet6 or if_detach() related changes recently ?
> >
> 
> I have done more investigation, and this was triggered by the recent
> kmem changes;  I'm sure the problem was simply exposed by these
> changes, as have been some others.
> 
> rmind@ mentioned that it may even be in 5.1:
> http://gnats.netbsd.org/45764

I guess it's fixed with:
http://mail-index.netbsd.org/source-changes/2012/02/03/msg031457.html
http://mail-index.netbsd.org/source-changes/2012/02/03/msg031458.html

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--