]> pilppa.com Git - linux-2.6-omap-h63xx.git/commit
mm: don't mark_page_accessed in fault path
authorNick Piggin <npiggin@suse.de>
Tue, 6 Jan 2009 22:38:55 +0000 (14:38 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 6 Jan 2009 23:58:58 +0000 (15:58 -0800)
commitbf3f3bc5e734706730c12a323f9b2068052aa1f0
treed93fb6beb0916cc10aeb5674578bfa3ac40371c9
parent3340289ddf29ca75c3acfb3a6b72f234b2f74d5c
mm: don't mark_page_accessed in fault path

Doing a mark_page_accessed at fault-time, then doing SetPageReferenced at
unmap-time if the pte is young has a number of problems.

mark_page_accessed is supposed to be roughly the equivalent of a young pte
for unmapped references. Unfortunately it doesn't come with any context:
after being called, reclaim doesn't know who or why the page was touched.

So calling mark_page_accessed not only adds extra lru or PG_referenced
manipulations for pages that are already going to have pte_young ptes anyway,
but it also adds these references which are difficult to work with from the
context of vma specific references (eg. MADV_SEQUENTIAL pte_young may not
wish to contribute to the page being referenced).

Then, simply doing SetPageReferenced when zapping a pte and finding it is
young, is not a really good solution either. SetPageReferenced does not
correctly promote the page to the active list for example. So after removing
mark_page_accessed from the fault path, several mmap()+touch+munmap() would
have a very different result from several read(2) calls for example, which
is not really desirable.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Acked-by: Johannes Weiner <hannes@saeurebad.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/filemap.c
mm/memory.c