On Mon, May 20, 2013 at 04:16:07PM -0500, Ben Myers wrote:
> On Mon, May 20, 2013 at 09:51:12AM +1000, Dave Chinner wrote:
> > From: Dave Chinner <dchinner@xxxxxxxxxx>
> >
> > There are several places where we use KM_SLEEP allocation contexts
> > and use the fact that there are called from transaction context to
> they
>
> (fixed)
>
> > add KM_NOFS where appropriate.
>
> I think you're referring to the usage of PF_FSTRANS and us clearing __GFP_FS
> in
> kmem_flags_convert?
Yes.
> > Unfortunately, there are several
> > places where the code makes this assumption but can be called from
> > outside transaction context but with filesystem locks held. These
> > places need explicit KM_NOFS annotations to avoid lockdep
> > complaining about reclaim contexts.
> >
> > Signed-off-by: Dave Chinner <dchinner@xxxxxxxxxx>
>
> Looks good. In each case you added KM_NOFS where there was no transaction and
> locks would have been held. Applied.
>
> Reviewed-by: Ben Myers <bpm@xxxxxxx>
Thanks.
-Dave
--
Dave Chinner
david@xxxxxxxxxxxxx
|