Re: port-xen/39916: kernel panic with xen 3.1 on 5.0-Beta i386 (HP DL 380 G3) (200811090002Z)

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

The following reply was made to PR port-xen/39916; it has been noted by GNATS.

From: Manuel Bouyer <bouyer%antioche.eu.org@localhost>
To: Uwe Werler <uw%o3si.de@localhost>
Cc: gnats-bugs%NetBSD.org@localhost
Subject: Re: port-xen/39916: kernel panic with xen 3.1 on 5.0-Beta i386 (HP DL 
380 G3) (200811090002Z)
Date: Sat, 15 Nov 2008 16:53:47 +0100

 On Sat, Nov 15, 2008 at 04:49:45PM +0100, Uwe Werler wrote:
 > Am Thu, 13 Nov 2008 21:45:01 +0000 (UTC)
 > schrieb Manuel Bouyer <bouyer%antioche.eu.org@localhost>:
 > 
 > > The following reply was made to PR port-xen/39916; it has been noted
 > > by GNATS.
 > > 
 > > From: Manuel Bouyer <bouyer%antioche.eu.org@localhost>
 > > To: gnats-bugs%NetBSD.org@localhost
 > > Cc: port-xen-maintainer%NetBSD.org@localhost, 
 > > gnats-admin%NetBSD.org@localhost,
 > >         netbsd-bugs%NetBSD.org@localhost
 > > Subject: Re: port-xen/39916: kernel panic with xen 3.1 on 5.0-Beta
 > > i386 (HP DL 380 G3) (200811090002Z) Date: Thu, 13 Nov 2008 22:41:58
 > > +0100
 > > 
 > >  On Thu, Nov 13, 2008 at 09:20:00PM +0000, uw%o3si.de@localhost wrote:
 > >  > >Number:         39916
 > >  > >Category:       port-xen
 > >  > >Synopsis:       xen kernel fault on i389 with 5.0 Beta
 > >  > >Confidential:   no
 > >  > >Severity:       critical
 > >  > >Priority:       high
 > >  > >Responsible:    port-xen-maintainer
 > >  > >State:          open
 > >  > >Class:          sw-bug
 > >  > >Submitter-Id:   net
 > >  > >Arrival-Date:   Thu Nov 13 21:20:00 +0000 2008
 > >  > >Originator:     Charlie Root
 > >  > >Release:        NetBSD 5.0_BETA
 > >  > >Environment:
 > >  > System: NetBSD root03.o3si.de 5.0_BETA NetBSD 5.0_BETA (XEN3_DOM0)
 > >  > #1: Thu Nov 13 15:17:38 CET 2008
 > >  > root%root03.o3si.de@localhost:/usr/src/sys/arch/i386/compile/XEN3_DOM0 
 > > i386
 > >  > Architecture: i386 Machine: i386
 > >  > >Description:
 > >  > The system crashes after running ca. for two days with the
 > >  > following console output:
 > >  > 
 > >  > panic: kernel diagnostic assertion "iplbit !=3D 0" failed. file
 > >  > "../../../.= ./arch/
 > >  > xen/xen/evtchn.c", line 294
 > >  
 > >  This should already be partially fixed in current and netbsd-5
 > > (partially because it may still panic later on another KASSERT).
 > >  evtchn.c 1.41 should fix the panic (although the real issue, which is
 > >  that the IPL isn't lowered when it should, is still there it seems).
 > >  trap type 
 > >  evtchn.c 1.41 should print a debug printf with the address of the
 > > handler which didn't lower the IPL; I'm interested to know what this
 > > printf shows in dmesg after several days.
 > >  
 > >  -- 
 > >  Manuel Bouyer <bouyer%antioche.eu.org@localhost>
 > >       NetBSD: 26 ans d'experience feront toujours la difference
 > >  --
 > >  
 > 
 > Hello Manuel,
 > 
 > unfortunately the kernel paniced again. This is this time the output:
 > 
 > fatal page fault in supervisor mode
 > trap type 6 code 0 eip c054331c cs 9 eflags 10296 cr2 deaddead ilevel 0
 > kernel: supervisor trap page fault, code=0
 > Stopped in pid 0.3 (system) at netbsd:rn_match+0xc:  mov1    0(%eax),%eax
 > db>
 > 
 > What else can I do to help to fix this problem? Should I run the xen-debug 
 > kernel?
 
 The output of 'tr' would have helped here. It's probably a different
 problem.
 I don't think xen-debug would help here, as it doesn't seem to be an
 issue with hypervisor interractions (here it looks more like a use after
 free, or somthing like that).
 
 -- 
 Manuel Bouyer <bouyer%antioche.eu.org@localhost>
      NetBSD: 26 ans d'experience feront toujours la difference
 --