[OpenAFS] Re: Openafs 1.3.78 and kernel 2.4.29 oopses ,same for 2.4.30 and openafs 1.3.82

[Dimitris Zilaskos]

         I am forwarding two messages from linux-kernel related to my
problem that dont seem to have made it to this list :


---------- Forwarded message ----------
Date: Fri, 6 May 2005 15:42:07 +0100
From: Christoph Hellwig <hch@infradead.org>
To: chas3@users.sourceforge.net
Cc: Dimitris Zilaskos <dzila@tassadar.physics.auth.gr>,
     Willy Tarreau <willy@w.ods.org>, openafs-info@openafs.org,
     linux-kernel@vger.kernel.org
Subject: Re: Openafs 1.3.78 and kernel 2.4.29 oopses ,
     same for 2.4.30 and openafs 1.3.82

On Fri, May 06, 2005 at 10:23:42AM -0400, chas williams - CONTRACTOR wrote:
> In message <Pine.LNX.4.62.0505061024070.450@tassadar.physics.auth.gr>,Dimitris
> Zilaskos writes:
>>May  6 04:55:29 system kernel: kernel BUG at inode.c:1204!
>
> looks like you might have one of those kernels with extra bits (in
> particular, i_notify).  please try a later version of afs like 1.3.81.

Btw, we get tons of bugsreports like that lately.  Is there a chance
that the OpenAFS folks could get their act together and stop abusing
internal interfaces all over.  In particular please stop shadowing struct
inode with your own version and use the proper export operations interfaces
if you're searching for an inode on another filesystem instead of the
utterly wrong blind iget().



---------- Forwarded message ----------
Date: Fri, 6 May 2005 13:50:33 -0300
From: Marcelo Tosatti <marcelo.tosatti@cyclades.com>
To: Willy Tarreau <willy@w.ods.org>
Cc: Dimitris Zilaskos <dzila@tassadar.physics.auth.gr>,
     openafs-info@openafs.org, linux-kernel@vger.kernel.org
Subject: Re: Openafs 1.3.78 and kernel 2.4.29 oopses ,
     same for 2.4.30 and openafs 1.3.82

On Fri, May 06, 2005 at 07:28:03AM +0200, Willy Tarreau wrote:
> Hi,
>
> On Fri, May 06, 2005 at 02:55:40AM +0300, Dimitris Zilaskos wrote:
>>
>> 	Hello ,
>>
>> [1.] One line summary of the problem:
>>
>> Oopses on an openafs client system using openafs 1.3.78 and kernel 2.4.29.
>> Oopses also occur afer moving to kernel 2.4.30 and openafs 1.3.82
>
> The problem you encounter on 2.4.30 is not the same as on 2.4.29. The problem
> in 2.4.29 is related to link_path_walk, which has been fixed in 2.4.30.
>
> You might want to try 1.3.78 (or other) with 2.4.29-hf7 to check if your
> 2.4.30 problem was brought by kernel 2.4.30 or openafs 1.3.82, as the
> link_path_walk bug is also fixed in hf7. The patch is available on :


Willy,

The link_path_walk fix in v2.4.30 is related to a reference counting
bug triggered by "umount"...

As Christoph noted OpenAFS seems to be doing nasty things...  it seems
to play with dentries inode i_state directly? If that is the case,
maybe it should define d_iput?

static inline void dentry_iput(struct dentry * dentry)
{
         struct inode *inode = dentry->d_inode;
         if (inode) {
                 dentry->d_inode = NULL;
                 list_del_init(&dentry->d_alias);
                 spin_unlock(&dcache_lock);
                 if (dentry->d_op && dentry->d_op->d_iput)
                         dentry->d_op->d_iput(dentry, inode);



May  6 04:55:29 system kernel: kernel BUG at inode.c:1204!



--
=============================================================================

Dimitris Zilaskos

Department of Physics @ Aristotle University of Thessaloniki , Greece
PGP key : http://tassadar.physics.auth.gr/~dzila/pgp_public_key.asc
 	  http://egnatia.ee.auth.gr/~dzila/pgp_public_key.asc
MD5sum  : de2bd8f73d545f0e4caf3096894ad83f  pgp_public_key.asc
=============================================================================