On Fri 16-12-16 16:05:58, Mike Galbraith wrote:
> On Thu, 2016-12-15 at 15:07 +0100, Michal Hocko wrote:
> > Hi,
> > I have posted the previous version here [1]. Since then I have added a
> > support to suppress reclaim lockdep warnings (__GFP_NOLOCKDEP) to allow
> > removing GFP_NOFS usage motivated by the lockdep false positives. On top
> > of that I've tried to convert few KM_NOFS usages to use the new flag in
> > the xfs code base. This would need a review from somebody familiar with
> > xfs of course.
>
> The wild ass guess below prevents the xfs explosion below when running
> ltp zram tests.
Yes this looks correct. Thanks for noticing. I will fold it to the
patch2. Thanks for testing Mike!
>
> ---
> fs/xfs/kmem.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> --- a/fs/xfs/kmem.h
> +++ b/fs/xfs/kmem.h
> @@ -45,7 +45,7 @@ kmem_flags_convert(xfs_km_flags_t flags)
> {
> gfp_t lflags;
>
> - BUG_ON(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL|KM_ZERO));
> + BUG_ON(flags &
> ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL|KM_ZERO|KM_NOLOCKDEP));
>
> if (flags & KM_NOSLEEP) {
> lflags = GFP_ATOMIC | __GFP_NOWARN;
--
Michal Hocko
SUSE Labs
