Re: [2.6.39] CIFS write failures where 2.6.38 works

[Jeff Layton <jlayton@redhat.com>]

On Fri, 03 Jun 2011 15:45:37 +0530
Suresh Jayaraman <sjayaraman@suse.de> wrote:

> [Cc linux-cifs@vger.kernel.org]
> 
> On 06/01/2011 03:41 PM, Helge Hafting wrote:
> > At work I use cifs for accessing a windows server. This has worked fine
> > for a long time, up to and including Debian's 2.6.38-2.
> > 
> > I just installed Debians's 2.6.39-1, and had to give up on it.
> > Mounting CIFS works, and I can see the files. But if I
> > try to make a new file (with cp), I get a long delay.
> 
> What is the security mechanism you are using? If you seeing the problem
> with ntlm, could you try using ntlmv2 and see whether the problem is
> reproducible?
> 
> 
> > In this time, anything accessing CIFS will pause, including "ls".
> > CIFS eventually becomes unstuck, and I will find that my "ls" succeeded,
> > but the file copy did not. The file was created, but
> > with 0 size.
> > 
> > Trying to copy the file again results in another stall, and so on.
> > 
> > 
> > When I mount CIFS, I get this message:
> > CIFS VFS: Unexpected SMB signature
> > 

I get this too on every mount using signing. It would be nice to fix
that but it's probably not related to the problem.

> > This happens with 2.6.38 as well as 2.6.39.
> > 
> > When writes fail in 2.6.39, I also get:
> > CIFS VFS: Send error in Close = -512
> > 

-512 is -ERESTARTSYS which probably means that it got a signal.

> > Followed by:
> > CIFS VFS: Server servername has not responded in 300 seconds.
> > Reconnecting...
> > 

So the server may not be responding to SMB echoes.

> > I have also seen
> > CIFS VFS: Send error in Close = -512
> > and
> > CIFS VFS: Send error in read = -13
> 

What sort of server is this? We'll probably need to see some wire
captures or debug logging to understand what's happening. It might be
good to open a bug at bugzilla.samba.org or bugzilla.kernel.org so we
can track progress on this.

-- 
Jeff Layton <jlayton@redhat.com>