mm: mark page accessed before we write_end()
commitc8236db9cd7aa492dcfcdcca702638e704abed49
authorJosef Bacik <josef@redhat.com>
Sun, 5 Jul 2009 19:08:18 +0000 (5 12:08 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Mon, 6 Jul 2009 20:57:03 +0000 (6 13:57 -0700)
tree75eb83051298fd1f9df426cafaaa6d3c422d1b68
parenta65e7bfcd74e4c0939f235d2bf9f48ddb3a57991
mm: mark page accessed before we write_end()

In testing a backport of the write_begin/write_end AOPs, a 10% re-read
regression was noticed when running iozone.  This regression was
introduced because the old AOPs would always do a mark_page_accessed(page)
after the commit_write, but when the new AOPs where introduced, the only
place this was kept was in pagecache_write_end().

This patch does the same thing in the generic case as what is done in
pagecache_write_end(), which is just to mark the page accessed before we
do write_end().

Signed-off-by: Josef Bacik <jbacik@redhat.com>
Acked-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/filemap.c