]> git.baikalelectronics.ru Git - kernel.git/commit
revert "mm: vmscan: do not swap anon pages just because free+file is low"
authorJohannes Weiner <hannes@cmpxchg.org>
Tue, 6 May 2014 19:50:07 +0000 (12:50 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 6 May 2014 20:04:59 +0000 (13:04 -0700)
commite5d46141d1d18c66c7e7a343f485cc73651cde4f
tree94d42e42a3b1eb693ba387b6b297351ef6b03ed2
parente4fe4aa1df4a2a641be948a13e1f02b42bdc8271
revert "mm: vmscan: do not swap anon pages just because free+file is low"

This reverts commit 4b77e252a7ea ("mm: vmscan: do not swap anon pages
just because free+file is low") because it introduced a regression in
mostly-anonymous workloads, where reclaim would become ineffective and
trap every allocating task in direct reclaim.

The problem is that there is a runaway feedback loop in the scan balance
between file and anon, where the balance tips heavily towards a tiny
thrashing file LRU and anonymous pages are no longer being looked at.
The commit in question removed the safe guard that would detect such
situations and respond with forced anonymous reclaim.

This commit was part of a series to fix premature swapping in loads with
relatively little cache, and while it made a small difference, the cure
is obviously worse than the disease.  Revert it.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Rafael Aquini <aquini@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: <stable@kernel.org> [3.12+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/vmscan.c