[Regression] cifs: sanity check length of data to send before sending

[Jeff Layton]

On Tue, 23 Sep 2014 15:09:23 -0400
Joseph Salisbury <joseph.salisbury at canonical.com> wrote:

> Hi Jeff,
> 
> A kernel bug report was opened against Ubuntu [0].  It was found that
> reverting the following commit resolved this bug:
> 
> 
> commit a26054d184763969a411e3939fe243516715ff59
> Author: Jeff Layton <jlayton at redhat.com>
> Date:   Fri Feb 14 07:21:00 2014 -0500
> 
>     cifs: sanity check length of data to send before sending
> 
>       
> The regression was introduced as of v3.14-rc7.
> 
> It appears the patch may be doing what it is supposed to, since the
> send_length != smb_buf_length:
> send_length=72
> and
> smb_buf_length=2164260932
> 
> However, without this check and WARN/return of -EIO the samba
> connections still work correctly.
> 
> I was hoping to get your feedback, since you are the patch author.  Do
> you think gathering any additional data will help diagnose this issue,
> or would it be best to submit a revert request?
>     
> Thanks,
>     
> Joe
> 
> [0] http://pad.lv/1372482
> 

(I'm no longer with Red Hat, so the jlayton at redhat.com address won't
reach me any longer...)

It looks like you probably just need to merge
dca1c8d17a2feae056f9e334ea75a462ae4cb52a into your kernel as well.

Cheers,
-- 
Jeff Layton <jlayton at samba.org>