ksm: fix ksm swapin time optimization
authorAndrea Arcangeli <aarcange@redhat.com>
Tue, 10 Aug 2010 00:19:08 +0000 (17:19 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 10 Aug 2010 03:44:58 +0000 (20:44 -0700)
commitba6f0ff3981e6263ab81ac512f04cca55b85ec81
tree0052d4551de72731be503c8d78949a330ba865ed
parent26ba0cb63cb8df4e45394227f33c938920b11b88
ksm: fix ksm swapin time optimization

The new anon-vma code, was suboptimal and it lead to erratic invocation of
ksm_does_need_to_copy.  That leads to host hangs or guest vnc lockup, or
weird behavior.  It's unclear why ksm_does_need_to_copy is unstable but
the point is that when KSM is not in use, ksm_does_need_to_copy must never
run or we bounce pages for no good reason.  I suspect the same hangs will
happen with KVM swaps.  But this at least fixes the regression in the
new-anon-vma code and it only let KSM bugs triggers when KSM is in use.

The code in do_swap_page likely doesn't cope well with a not-swapcache,
especially the memcg code.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Rik van Riel <riel@redhat.com>
Cc: Izik Eidus <ieidus@yahoo.com>
Cc: Avi Kivity <avi@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
include/linux/ksm.h