]> git.proxmox.com Git - mirror_ubuntu-eoan-kernel.git/commit
vmscan: clear ZONE_CONGESTED for zone with good watermark
authorShaohua Li <shaohua.li@intel.com>
Thu, 25 Aug 2011 22:59:12 +0000 (15:59 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 25 Aug 2011 23:25:34 +0000 (16:25 -0700)
commit439423f6894aa0dec22187526827456f5004baed
tree1f22238a24b83e0beb1467a1f73b0db952c69894
parent4c30c6f566c0989ddaee3407da44751e340a63ed
vmscan: clear ZONE_CONGESTED for zone with good watermark

ZONE_CONGESTED is only cleared in kswapd, but pages can be freed in any
task.  It's possible ZONE_CONGESTED isn't cleared in some cases:

 1. the zone is already balanced just entering balance_pgdat() for
    order-0 because concurrent tasks free memory.  In this case, later
    check will skip the zone as it's balanced so the flag isn't cleared.

 2. high order balance fallbacks to order-0.  quote from Mel: At the
    end of balance_pgdat(), kswapd uses the following logic;

If reclaiming at high order {
for each zone {
if all_unreclaimable
skip
if watermark is not met
order = 0
loop again

/* watermark is met */
clear congested
}
}

    i.e. it clears ZONE_CONGESTED if it the zone is balanced.  if not,
    it restarts balancing at order-0.  However, if the higher zones are
    balanced for order-0, kswapd will miss clearing ZONE_CONGESTED as
    that only happens after a zone is shrunk.  This can mean that
    wait_iff_congested() stalls unnecessarily.

This patch makes kswapd clear ZONE_CONGESTED during its initial
highmem->dma scan for zones that are already balanced.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/vmscan.c