Re: [PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-23 Thread Michal Hocko
It seems this patch hasn't reached the mmotm tree. Are there any unresolved concerns left? On Mon 16-11-15 14:22:19, mho...@kernel.org wrote: > From: Michal Hocko > > __alloc_pages_slowpath is looping over ALLOC_NO_WATERMARKS requests if > __GFP_NOFAIL is requested. This is fragile because we ar

Re: [PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-18 Thread Vlastimil Babka
On 11/18/2015 04:11 PM, Michal Hocko wrote: > On Wed 18-11-15 15:57:45, Vlastimil Babka wrote: > [...] >> > --- a/mm/page_alloc.c >> > +++ b/mm/page_alloc.c >> > @@ -3046,32 +3046,36 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned >> > int order, >> > * allocations are system rather

Re: [PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-18 Thread Michal Hocko
On Wed 18-11-15 15:57:45, Vlastimil Babka wrote: [...] > > --- a/mm/page_alloc.c > > +++ b/mm/page_alloc.c > > @@ -3046,32 +3046,36 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int > > order, > > * allocations are system rather than user orientated > > */ > >

Re: [PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-18 Thread Vlastimil Babka
On 11/16/2015 02:22 PM, mho...@kernel.org wrote: > From: Michal Hocko > > __alloc_pages_slowpath is looping over ALLOC_NO_WATERMARKS requests if > __GFP_NOFAIL is requested. This is fragile because we are basically > relying on somebody else to make the reclaim (be it the direct reclaim > or OOM

Re: [PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-18 Thread Michal Hocko
On Wed 18-11-15 10:11:01, Michal Hocko wrote: > Besides that I fail to see why a work item would ever > want to set PF_MEMALLOC for legitimate reasons. I have done a quick git > grep over the tree and there doesn't seem to be any user. OK, I have missed one case. xfs_btree_split_worker is really s

Re: [PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-18 Thread Michal Hocko
On Tue 17-11-15 19:58:09, Tetsuo Handa wrote: > Michal Hocko wrote: > > __alloc_pages_slowpath is looping over ALLOC_NO_WATERMARKS requests if > > __GFP_NOFAIL is requested. This is fragile because we are basically > > relying on somebody else to make the reclaim (be it the direct reclaim > > or OO

Re: [PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-17 Thread Tetsuo Handa
Michal Hocko wrote: > __alloc_pages_slowpath is looping over ALLOC_NO_WATERMARKS requests if > __GFP_NOFAIL is requested. This is fragile because we are basically > relying on somebody else to make the reclaim (be it the direct reclaim > or OOM killer) for us. The caller might be holding resources

Re: [PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-16 Thread David Rientjes
On Mon, 16 Nov 2015, mho...@kernel.org wrote: > From: Michal Hocko > > __alloc_pages_slowpath is looping over ALLOC_NO_WATERMARKS requests if > __GFP_NOFAIL is requested. This is fragile because we are basically > relying on somebody else to make the reclaim (be it the direct reclaim > or OOM ki

[PATCH 2/2] mm: do not loop over ALLOC_NO_WATERMARKS without triggering reclaim

2015-11-16 Thread mhocko
From: Michal Hocko __alloc_pages_slowpath is looping over ALLOC_NO_WATERMARKS requests if __GFP_NOFAIL is requested. This is fragile because we are basically relying on somebody else to make the reclaim (be it the direct reclaim or OOM killer) for us. The caller might be holding resources (e.g. l