[PATCH] Zone reclaim: resurrect may_swap

Zone reclaim has a huge impact on NUMA performance (f.e.  our maximum
throughput with XFS is raised from 4GB to 6GB/sec / page cache contamination
of numa nodes destroys locality if one just does a large copy operation which
results in performance dropping for good until reboot).

This patch:

Resurrect may_swap in struct scan_control

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

mm/vmscan.c

@@ -71,6 +71,9 @@ struct scan_control {
 
 	int may_writepage;
 
+	/* Can pages be swapped as part of reclaim? */
+	int may_swap;
+
 	/* This context's SWAP_CLUSTER_MAX. If freeing memory for
 	 * suspend, we effectively ignore SWAP_CLUSTER_MAX.
 	 * In this context, it doesn't matter that we scan the
@@ -458,6 +461,8 @@ static int shrink_list(struct list_head *page_list, struct scan_control *sc)
 		 * Try to allocate it some swap space here.
 		 */
 		if (PageAnon(page) && !PageSwapCache(page)) {
+			if (!sc->may_swap)
+				goto keep_locked;
 			if (!add_to_swap(page, GFP_ATOMIC))
 				goto activate_locked;
 		}
@@ -1166,6 +1171,7 @@ int try_to_free_pages(struct zone **zones, gfp_t gfp_mask)
 
 	sc.gfp_mask = gfp_mask;
 	sc.may_writepage = 0;
+	sc.may_swap = 1;
 
 	inc_page_state(allocstall);
 
@@ -1268,6 +1274,7 @@ static int balance_pgdat(pg_data_t *pgdat, int nr_pages, int order)
 	total_reclaimed = 0;
 	sc.gfp_mask = GFP_KERNEL;
 	sc.may_writepage = 0;
+	sc.may_swap = 1;
 	sc.nr_mapped = read_page_state(nr_mapped);
 
 	inc_page_state(pageoutrun);