]> git.karo-electronics.de Git - karo-tx-linux.git/commit
revert "mm: vmscan: do not swap anon pages just because free+file is low"
authorJohannes Weiner <hannes@cmpxchg.org>
Thu, 24 Apr 2014 22:55:26 +0000 (08:55 +1000)
committerStephen Rothwell <sfr@canb.auug.org.au>
Thu, 24 Apr 2014 22:55:26 +0000 (08:55 +1000)
commitfb6ad537d5c5f02a9c46e1ea72392b5ec919d3f4
tree75fa652781a807656a5331aa19f44eaf53295d56
parentb1d30c5a47a3922fffb535b197bfd5187dccc2c0
revert "mm: vmscan: do not swap anon pages just because free+file is low"

This reverts commit 0bf1457f0cfc ("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>
mm/vmscan.c