[7067] in testers
Re: AFS cache coherency failure under 9.4
daemon@ATHENA.MIT.EDU (Bill Cattey)
Wed Jun 1 17:16:15 2005
From: Bill Cattey <wdc@MIT.EDU>
To: Robert A Basch <rbasch@mit.edu>
Cc: Greg Hudson <ghudson@mit.edu>, testers@mit.edu
In-Reply-To: <1117660359.11710.5.camel@anhedonia.mit.edu>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Message-Id: <1117660564.31147.122.camel@tokata.mit.edu>
Mime-Version: 1.0
Date: Wed, 01 Jun 2005 17:16:04 -0400
I saw this entry in the change log for 1.3.83 that might be relevant:
2005-05-23 21:09 shadow
* src/afs/LINUX/osi_vnodeops.c: DELTA
STABLE14-d-revalidate-redux-20050508 AUTHOR chas@cmf.nrl.navy.mil
FIXES 18588
when the callback for a directory is broken the children of this
vnode are not invalidated as well. so VerifyVCache() always
believes these vnodes to be up to date since they will still have
the CStatd flag.
perhaps breaking the callback on a directory should also clear
CStatd on children vnodes?
this isnt a problem for other operating system which go through the
dnlc which does purge the children entries for the vnode in
question during ClearCallBack().
since d_revalidate now avoids the dnlc (and afs_lookup) the
following patch tracks the parent vnode's DataVersion (just the
lower half really but this should be sufficient) in d_time. if the
parent has changed then the dentry is flagged as neededing updated.
On Wed, 2005-06-01 at 17:12, Robert A Basch wrote:
> On Wed, 2005-06-01 at 12:50, Greg Hudson wrote:
> > I observed another case of cache coherency failure
>
> I just encountered what seems to be the same problem, but on a file
> written from a 9.3 (Linux) client (and read from a 9.3 Solaris client).
> This file is also in the dev cell; a quick test in the athena cell
> did not exhibit a problem...
>
> Bob
>
