This is a note to let you know that I've just added the patch titled
mm: exclude memoryless nodes from zone_reclaim
to the 3.14-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
mm-exclude-memoryless-nodes-from-zone_reclaim.patch
and it can be found in the queue-3.14 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
>From 70ef57e6c22c3323dce179b7d0d433c479266612 Mon Sep 17 00:00:00 2001
From: Michal Hocko <mhocko@xxxxxxx>
Date: Mon, 7 Apr 2014 15:37:01 -0700
Subject: mm: exclude memoryless nodes from zone_reclaim
From: Michal Hocko <mhocko@xxxxxxx>
commit 70ef57e6c22c3323dce179b7d0d433c479266612 upstream.
We had a report about strange OOM killer strikes on a PPC machine
although there was a lot of swap free and a tons of anonymous memory
which could be swapped out. In the end it turned out that the OOM was a
side effect of zone reclaim which wasn't unmapping and swapping out and
so the system was pushed to the OOM. Although this sounds like a bug
somewhere in the kswapd vs. zone reclaim vs. direct reclaim
interaction numactl on the said hardware suggests that the zone reclaim
should not have been set in the first place:
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
node 0 size: 0 MB
node 0 free: 0 MB
node 2 cpus:
node 2 size: 7168 MB
node 2 free: 6019 MB
node distances:
node 0 2
0: 10 40
2: 40 10
So all the CPUs are associated with Node0 which doesn't have any memory
while Node2 contains all the available memory. Node distances cause an
automatic zone_reclaim_mode enabling.
Zone reclaim is intended to keep the allocations local but this doesn't
make any sense on the memoryless nodes. So let's exclude such nodes for
init_zone_allows_reclaim which evaluates zone reclaim behavior and
suitable reclaim_nodes.
Signed-off-by: Michal Hocko <mhocko@xxxxxxx>
Acked-by: David Rientjes <rientjes@xxxxxxxxxx>
Acked-by: Nishanth Aravamudan <nacc@xxxxxxxxxxxxxxxxxx>
Tested-by: Nishanth Aravamudan <nacc@xxxxxxxxxxxxxxxxxx>
Acked-by: Mel Gorman <mgorman@xxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
mm/page_alloc.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1869,7 +1869,7 @@ static void __paginginit init_zone_allow
{
int i;
- for_each_online_node(i)
+ for_each_node_state(i, N_MEMORY)
if (node_distance(nid, i) <= RECLAIM_DISTANCE)
node_set(i, NODE_DATA(nid)->reclaim_nodes);
else
@@ -4933,7 +4933,8 @@ void __paginginit free_area_init_node(in
pgdat->node_id = nid;
pgdat->node_start_pfn = node_start_pfn;
- init_zone_allows_reclaim(nid);
+ if (node_state(nid, N_MEMORY))
+ init_zone_allows_reclaim(nid);
#ifdef CONFIG_HAVE_MEMBLOCK_NODE_MAP
get_pfn_range_for_nid(nid, &start_pfn, &end_pfn);
#endif
Patches currently in stable-queue which might be from mhocko@xxxxxxx are
queue-3.14/swap-change-swap_info-singly-linked-list-to-list_head.patch
queue-3.14/swap-change-swap_list_head-to-plist-add-swap_avail_head.patch
queue-3.14/mm-exclude-memoryless-nodes-from-zone_reclaim.patch
queue-3.14/lib-plist-add-helper-functions.patch
queue-3.14/lib-plist-add-plist_requeue.patch
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html