[PATCH] Zone reclaim: Allow modification of zone reclaim behavior

In some situations one may want zone_reclaim to behave differently.  For
example a process writing large amounts of memory will spew unto other nodes
to cache the writes if many pages in a zone become dirty.  This may impact the
performance of processes running on other nodes.

Allowing writes during reclaim puts a stop to that behavior and throttles the
process by restricting the pages to the local zone.

Similarly one may want to contain processes to local memory by enabling
regular swap behavior during zone_reclaim.  Off node memory allocation can
then be controlled through memory policies and cpusets.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/Documentation/sysctl/vm.txt b/Documentation/sysctl/vm.txt
index 44518c0239490134bd230cd5ecdb6e1d8e1cf566..4bca2a3d9174677b3275d62c93d01b8d99217b1a 100644 (file)
--- a/Documentation/sysctl/vm.txt
+++ b/Documentation/sysctl/vm.txt
@@ -127,17 +127,39 @@ the high water marks for each per cpu page list.
 
 zone_reclaim_mode:
 
-This is set during bootup to 1 if it is determined that pages from
-remote zones will cause a significant performance reduction. The
+Zone_reclaim_mode allows to set more or less agressive approaches to
+reclaim memory when a zone runs out of memory. If it is set to zero then no
+zone reclaim occurs. Allocations will be satisfied from other zones / nodes
+in the system.
+
+This is value ORed together of
+
+1      = Zone reclaim on
+2      = Zone reclaim writes dirty pages out
+4      = Zone reclaim swaps pages
+
+zone_reclaim_mode is set during bootup to 1 if it is determined that pages
+from remote zones will cause a measurable performance reduction. The
 page allocator will then reclaim easily reusable pages (those page
-cache pages that are currently not used) before going off node.
+cache pages that are currently not used) before allocating off node pages.
+
+It may be beneficial to switch off zone reclaim if the system is
+used for a file server and all of memory should be used for caching files
+from disk. In that case the caching effect is more important than
+data locality.
+
+Allowing zone reclaim to write out pages stops processes that are
+writing large amounts of data from dirtying pages on other nodes. Zone
+reclaim will write out dirty pages if a zone fills up and so effectively
+throttle the process. This may decrease the performance of a single process
+since it cannot use all of system memory to buffer the outgoing writes
+anymore but it preserve the memory on other nodes so that the performance
+of other processes running on other nodes will not be affected.
 
-The user can override this setting. It may be beneficial to switch
-off zone reclaim if the system is used for a file server and all
-of memory should be used for caching files from disk.
+Allowing regular swap effectively restricts allocations to the local
+node unless explicitly overridden by memory policies or cpuset
+configurations.
 
-It may be beneficial to switch this on if one wants to do zone
-reclaim regardless of the numa distances in the system.
 ================================================================
 
 zone_reclaim_interval:

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 8760a4abfa1f8b7d11fd04e119563c409f49a07a..9e2ef3624d77a0dbe411faa2f59b3f449661c101 100644 (file)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1592,6 +1592,11 @@ module_init(kswapd_init)
  */
 int zone_reclaim_mode __read_mostly;
 
+#define RECLAIM_OFF 0
+#define RECLAIM_ZONE (1<<0)    /* Run shrink_cache on the zone */
+#define RECLAIM_WRITE (1<<1)   /* Writeout pages during reclaim */
+#define RECLAIM_SWAP (1<<2)    /* Swap pages out during reclaim */
+
 /*
  * Mininum time between zone reclaim scans
  */
@@ -1630,8 +1635,8 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
        if (!cpus_empty(mask) && node_id != numa_node_id())
                return 0;
 
-       sc.may_writepage = 0;
-       sc.may_swap = 0;
+       sc.may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE);
+       sc.may_swap = !!(zone_reclaim_mode & RECLAIM_SWAP);
        sc.nr_scanned = 0;
        sc.nr_reclaimed = 0;
        sc.priority = ZONE_RECLAIM_PRIORITY + 1;