Re: [PATCH] mm: Use GFP_KERNEL allocation for the page cache in page_cache_read

From: Mel Gorman
Date: Wed Mar 18 2015 - 10:39:03 EST

On Wed, Mar 18, 2015 at 03:09:26PM +0100, Michal Hocko wrote:
> page_cache_read has been historically using page_cache_alloc_cold to
> allocate a new page. This means that mapping_gfp_mask is used as the
> base for the gfp_mask. Many filesystems are setting this mask to
> GFP_NOFS to prevent from fs recursion issues. page_cache_read is,
> however, not called from the fs layer so it doesn't need this
> protection. Even ceph and ocfs2 which call filemap_fault from their
> fault handlers seem to be OK because they are not taking any fs lock
> before invoking generic implementation.
> The protection might be even harmful. There is a strong push to fail
> GFP_NOFS allocations rather than loop within allocator indefinitely with
> a very limited reclaim ability. Once we start failing those requests
> the OOM killer might be triggered prematurely because the page cache
> allocation failure is propagated up the page fault path and end up in
> pagefault_out_of_memory.
> Use GFP_KERNEL mask instead because it is safe from the reclaim
> recursion POV. We are already doing GFP_KERNEL allocations down
> add_to_page_cache_lru path.
> Reported-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
> Signed-off-by: Michal Hocko <mhocko@xxxxxxx>

I'm very far behind after LSF/MM so do not know where this came out of
but it loses addressing restriction hints from the driver such as

drivers/gpu/drm/gma500/gem.c: mapping_set_gfp_mask(r->gem.filp->f_mapping, GFP_KERNEL | __GFP_DMA32);

It also loses mobility hints for fragmentation avoidance.

fs/inode.c: mapping_set_gfp_mask(mapping, GFP_HIGHUSER_MOVABLE);

If users of mapping_set_gfp_mask are now being ignored then it should at
least trigger a once-off warning that the flags are being ignored so
it's obvious if a recursion does occur and cause problems.

Mel Gorman
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at