[linux-cifs-client] [PATCH 0/4] cifs: convert oplock breaks to use slow_work facility

[Steve French]

 No objections to this.   Let me recheck the obvious today and merge.

On Fri, Sep 11, 2009 at 7:14 AM, Jeff Layton <jlayton at redhat.com> wrote:
> This patchset converts the oplock breaks to use the new slow_work
> facility that was added a couple of releases ago. In addition to fixing
> some races in the current scheme, this has the advantage that oplock
> breaks callbacks can potentially run in parallel.
>
> This patchset is identical to the one I sent the other day plus the
> patch to convert that to slow_work. The only difference is that I merged
> 4/4 and 5/4. I didn't see any reason to have a patch that converts the
> code to workqueues only to re-convert it to slow_work.
>
> This probably needs to go in ASAP since we have people hitting panics in
> the current code.
>
> Jeff Layton (4):
>  cifs: remove cifsInodeInfo.oplockPending flag
>  cifs: take read lock on GlobalSMBSes_lock in is_valid_oplock_break
>  cifs: have cifsFileInfo hold an extra inode reference
>  cifs: convert oplock breaks to use slow_work facility
>
>  fs/cifs/Kconfig     |    1 +
>  fs/cifs/cifsfs.c    |   95 +++------------------------------------------------
>  fs/cifs/cifsglob.h  |   17 ++++-----
>  fs/cifs/cifsproto.h |    4 --
>  fs/cifs/cifssmb.c   |    1 +
>  fs/cifs/connect.c   |    1 -
>  fs/cifs/dir.c       |    4 ++-
>  fs/cifs/file.c      |   73 +++++++++++++++++++++++++++++++++++++--
>  fs/cifs/misc.c      |   34 ++++++++++++------
>  fs/cifs/transport.c |   50 ---------------------------
>  10 files changed, 110 insertions(+), 170 deletions(-)
>
>



-- 
Thanks,

Steve