Comment 18 for bug 719446

Ok, so this commit is related to this bug (after bisection):

58e15f5029c75b45adbcb25cd76a36d26a6c4297 is the first bad commit
commit 58e15f5029c75b45adbcb25cd76a36d26a6c4297
Author: Christoph Lameter <email address hidden>
Date: Thu Sep 9 16:38:17 2010 -0700

    mm: page allocator: calculate a better estimate of NR_FREE_PAGES when memory is low and kswapd is awake

    commit aa45484031ddee09b06350ab8528bfe5b2c76d1c upstream.

    Ordinarily watermark checks are based on the vmstat NR_FREE_PAGES as it is
    cheaper than scanning a number of lists. To avoid synchronization
    overhead, counter deltas are maintained on a per-cpu basis and drained
    both periodically and when the delta is above a threshold. On large CPU
    systems, the difference between the estimated and real value of
    NR_FREE_PAGES can be very high. If NR_FREE_PAGES is much higher than
    number of real free page in buddy, the VM can allocate pages below min
    watermark, at worst reducing the real number of pages to zero. Even if
    the OOM killer kills some victim for freeing memory, it may not free
    memory if the exit path requires a new page resulting in livelock.

    This patch introduces a zone_page_state_snapshot() function (courtesy of
    Christoph) that takes a slightly more accurate view of an arbitrary vmstat
    counter. It is used to read NR_FREE_PAGES while kswapd is awake to avoid
    the watermark being accidentally broken. The estimate is not perfect and
    may result in cache line bounces but is expected to be lighter than the
    IPI calls necessary to continually drain the per-cpu counters while kswapd
    is awake.

    Signed-off-by: Christoph Lameter <email address hidden>
    Signed-off-by: Mel Gorman <email address hidden>
    Signed-off-by: Andrew Morton <email address hidden>
    Signed-off-by: Linus Torvalds <email address hidden>
    Signed-off-by: Greg Kroah-Hartman <email address hidden>

    (cherry picked from commit 87f1cbdee91c60af6dd255226e792a6410d77fbb 2.6.35.6)

    Signed-off-by: Leann Ogasawara <email address hidden>

:040000 040000 56b962edbe8815533041d28a1b4c744bf810c302 b93faeb21bff790a5e90f1370509a4903bb4fd9a M include
:040000 040000 a3d6ce0e31b4552a22ef291e0a66441b78744ac2 4664c82b2f1aa2452751144455f9f133cbf71270 M mm

It *MIGHT* be also connected to the previous commit (7dd373d47c6e0dc124924348265be61990ba0fb6) as I observed some short (0.5 second) freezes while testing it, but no errors.

I'll see if I can revert it in HEAD and whether it helps - will post soon.