Re: [ 10/41] CIFS: Do not kmalloc under the flocks spinlock

[Greg KH <gregkh@linuxfoundation.org>]

On Mon, Mar 19, 2012 at 12:24:16PM -0700, Greg KH wrote:
> On Mon, Mar 19, 2012 at 11:11:35PM +0400, Pavel Shilovsky wrote:
> > 19 марта 2012 г. 19:50 пользователь Greg KH
> > <gregkh@linuxfoundation.org> написал:
> > > On Sat, Mar 17, 2012 at 11:52:24AM +0400, Pavel Shilovsky wrote:
> > >> 17 марта 2012 г. 11:32 пользователь Ben Hutchings <ben@decadent.org.uk> написал:
> > >> > On Sat, 2012-03-17 at 10:14 +0400, Pavel Shilovsky wrote:
> > >> >> 17 марта 2012 г. 6:37 пользователь Ben Hutchings <ben@decadent.org.uk> написал:
> > >> >> > On Fri, Mar 16, 2012 at 04:38:20PM -0700, Greg KH wrote:
> > >> >> >> 3.2-stable review patch.  If anyone has any objections, please let me know.
> > >> > [...]
> > >> >> > But we test this before flock->fl_flags & FL_POSIX, which means we
> > >> >> > don't know whether this lock actually needs to be assigned one of
> > >> >> > those structures.  So it appears that we might report a spurious error
> > >> >> > if the lock list ends with a mandatory lock.  If so, this is
> > >> >> > relatively harmless but does need to be fixed.
> > >> >> >
> > >> >>
> > >> >> You are right here, thanks for the catch! I will repost the patch asap.
> > >> >
> > >> > This has already been merged into Linus's tree, so you need to submit a
> > >> > patch to apply on top of it.
> > >> >
> > >>
> > >> I posted two patches:
> > >> 1) the whole fixed version for the stable tree [PATCH v2] CIFS: Do not
> > >> kmalloc under the flocks spinlock
> > >
> > > What do you mean by "fixed version"?
> > >
> > >> 2) fixup for mainline [PATCH] CIFS: Fix a spurious error in
> > >> cifs_push_posix_locks
> > >
> > > What do you mean by this?
> > 
> > Ok, seems I didn't understand this process correctly. I reposted the
> > new "fixed" version of this patch, because I thought it is more
> > suitable for stale to merge one correct patch rather than one
> > incorrect + follow-on fixup. Sorry if I was wrong.
> 
> I need to stay identical with Linus's tree as much as possible, it's
> easier for everyone in the end that way for tracking exactly what
> happens.
> 
> > > If there was a follow-on patch in Linus's tree that fixes a problem, I
> > > need that git commit id, not a "fixed" patch that does not match up with
> > > what is in Linus's tree right now.
> > >
> > > So, if that's the case, please let me know what the git commit id of
> > > that patch is please.
> > 
> > Steve has just merged the follow-on patch:
> > http://git.samba.org/?p=sfrench/cifs-2.6.git;a=commit;h=ce85852b90a214cf577fc1b4f49d99fd7e98784a
> > 
> > but seems hasn't sent a merge request to Linus yet - will let you know
> > when the patch comes to Linus's tree.
> 
> Ok, please let me know when this goes into Linus's tree and I will
> queue it up for the next releases.

It's there now, I've picked it up.

greg k-h