sparc64/142102: FreeBSD 8.0 kernel panics on sparc64 when accessing NFS

[Marius Strobl]

The following reply was made to PR sparc64/142102; it has been noted by GNATS.

From: Marius Strobl <marius at alchemy.franken.de>
To: Manuel Tobias Schiller <mala at hinterbergen.de>
Cc: bug-followup at FreeBSD.org
Subject: Re: sparc64/142102: FreeBSD 8.0 kernel panics on sparc64 when accessing NFS
Date: Mon, 28 Dec 2009 19:26:47 +0100

 On Mon, Dec 28, 2009 at 06:07:36PM +0100, Manuel Tobias Schiller wrote:
 > On Mon, 28 Dec 2009 17:28:18 +0100
 > Marius Strobl <marius at alchemy.franken.de> wrote:
 > 
 > > On Mon, Dec 28, 2009 at 04:00:09PM +0000, Manuel Tobias Schiller wrote:
 > > > The following reply was made to PR sparc64/142102; it has been noted
 > > > by GNATS.
 > > > 
 > > > From: Manuel Tobias Schiller <mala at hinterbergen.de>
 > > > To: bug-followup at FreeBSD.org, mala at hinterbergen.de
 > > > Cc:  
 > > > Subject: Re: sparc64/142102: FreeBSD 8.0 kernel panics on sparc64 when
 > > >  accessing NFS
 > > > Date: Mon, 28 Dec 2009 16:40:04 +0100
 > > > 
 > > >  Dear all,
 > > >  
 > > >  the fix I gave in the original problem report appears to be
 > > > incomplete; the kernel still crashes sometimes during NFS access due
 > > > to improper alignment, this time in sys/nfsserver/nfs_srvsubs.c, line
 > > > 1355. I will try to continue debugging the issue and report back with
 > > > any fixes I manage to obtain.
 > > >  
 > > 
 > > Note that the original PR is a dupe of PR 140797. Could the second
 > > panic you're seeing be a result of using an incorrect size parameter
 > > in your change?
 > > 
 > > Marius
 > > 
 > > 
 > Dear Marius,
 > 
 > sorry, I searched, but I failed to spot PR 140797. I don't think that the
 > panic I'm seeing after my change has to do with the size parameter (if
 > it's incorrect though, please let me know). With the change I can read
 > files and do directory lookups just fine (tested with a 45 Gigabyte
 > file). I get the crash the minute I'm trying to write a file to the NFS
 > share.
 > 
 
 Yes, the size parameter in your patch is wrong. I also can't reproduce
 your remaining problem with the stock sources, I haven't checked whether
 it could be releated with using an incorrect size parameter though.
 
 Marius