Re: NFSv4 poops itself

[Trond Myklebust <Trond.Myklebust@netapp.com>]

On Mon, 2007-08-06 at 09:40 -0400, J. Bruce Fields wrote:
> On Fri, Aug 03, 2007 at 03:13:36PM -0400, J. Bruce Fields wrote:
> > On Fri, Jul 27, 2007 at 01:29:00PM -0400, Trond Myklebust wrote:
> > > On Fri, 2007-07-27 at 13:16 -0400, Jeff Garzik wrote:
> > > > Trond Myklebust wrote:
> > > > > Jeff and Bruce, could you please try to reproduce the problem after
> > > > > either applying patches 001 to 004 or just the single NFS_ALL patch
> > > > > from 
> > > > > 
> > > > >    http://client.linux-nfs.org/Linux-2.6.x/2.6.23-rc1/
> > > > 
> > > > 
> > > > It's not easily reproducible here.  Any hints on triggering the behavior?
> > > 
> > > I suspect that the problem may involve mixing O_RDWR with O_RDONLY
> > > and/or O_WRONLY file access. Bruce might have a better suggestion,
> > > though.
> > 
> > I haven't done any more work to identify when exactly the problem is
> > triggered, but I did confirm that I could reproduce the problem reliably
> > with current git, but *not* with current git plus your 4-patch nfs-all.
> 
> And that result still holds after using the patched kernel for several
> more hours.

Thanks Bruce!
  Trond