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

[Dave Chinner <david@fromorbit.com>]

On Wed, Mar 18, 2015 at 03:55:28PM +0100, Michal Hocko wrote:
> On Wed 18-03-15 10:44:11, Rik van Riel wrote:
> > On 03/18/2015 10:09 AM, 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 
> > 
> > Is that true for filesystems that have directories in
> > the page cache?
> 
> I haven't found any explicit callers of filemap_fault except for ocfs2
> and ceph and those seem OK to me. Which filesystems you have in mind?

Just about every major filesystem calls filemap_fault through the
.fault callout.

C symbol: filemap_fault

  File           Function            Line
  0 9p/vfs_file.c  <global>             831 .fault = filemap_fault,
  1 9p/vfs_file.c  <global>             838 .fault = filemap_fault,
  2 btrfs/file.c   <global>            2081 .fault = filemap_fault,
  3 cifs/file.c    <global>            3242 .fault = filemap_fault,
  4 ext4/file.c    <global>             215 .fault = filemap_fault,
  5 f2fs/file.c    <global>              93 .fault = filemap_fault,
  6 fuse/file.c    <global>            2062 .fault = filemap_fault,
  7 gfs2/file.c    <global>             498 .fault = filemap_fault,
  8 nfs/file.c     <global>             653 .fault = filemap_fault,
  9 nilfs2/file.c  <global>             128 .fault = filemap_fault,
  a ubifs/file.c   <global>            1536 .fault = filemap_fault,
  b xfs/xfs_file.c <global>            1420 .fault = filemap_fault,


> Btw. how would that work as we already have GFP_KERNEL allocation few
> lines below?

GFP_KERNEL allocation for mappings is simply wrong. All mapping
allocations where the caller cannot pass a gfp_mask need to obey
the mapping_gfp_mask that is set by the mapping owner....

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>