]> git.karo-electronics.de Git - mv-sheeva.git/commit
thp: split_huge_page paging
authorAndrea Arcangeli <aarcange@redhat.com>
Thu, 13 Jan 2011 23:46:47 +0000 (15:46 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 14 Jan 2011 01:32:41 +0000 (17:32 -0800)
commit3f04f62f90d46a82dd73027c5fd7a15daed5c33d
treec267f1bdde3f3a7b2a8b52347fa5ad6053cff435
parentbae9c19bf12bb2a914a8e530270f41d36cc87c63
thp: split_huge_page paging

Paging logic that splits the page before it is unmapped and added to swap
to ensure backwards compatibility with the legacy swap code.  Eventually
swap should natively pageout the hugepages to increase performance and
decrease seeking and fragmentation of swap space.  swapoff can just skip
over huge pmd as they cannot be part of swap yet.  In add_to_swap be
careful to split the page only if we got a valid swap entry so we don't
split hugepages with a full swap.

In theory we could split pages before isolating them during the lru scan,
but for khugepaged to be safe, I'm relying on either mmap_sem write mode,
or PG_lock taken, so split_huge_page has to run either with mmap_sem
read/write mode or PG_lock taken.  Calling it from isolate_lru_page would
make locking more complicated, in addition to that split_huge_page would
deadlock if called by __isolate_lru_page because it has to take the lru
lock to add the tail pages.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/memory-failure.c
mm/rmap.c
mm/swap_state.c
mm/swapfile.c