[PATCH] zone_reclaim: minor fixes
authorChristoph Lameter <clameter@engr.sgi.com>
Wed, 1 Feb 2006 11:05:29 +0000 (03:05 -0800)
committerLinus Torvalds <torvalds@g5.osdl.org>
Wed, 1 Feb 2006 16:53:15 +0000 (08:53 -0800)
- If we only reclaim nr_pages then its okay to stay on node.
  Switch from > to >= for the comparison.

- vm_table[] entry for zone_reclaim_mode is a bit screwed up.

- Add empty lines around shrink_zone to show that this is the
  central function to be called.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
kernel/sysctl.c
mm/vmscan.c

index cb99a42..c74f03b 100644 (file)
@@ -878,7 +878,8 @@ static ctl_table vm_table[] = {
                .maxlen         = sizeof(zone_reclaim_mode),
                .mode           = 0644,
                .proc_handler   = &proc_dointvec,
-               .strategy       = &zero,
+               .strategy       = &sysctl_intvec,
+               .extra1         = &zero,
        },
 #endif
        { .ctl_name = 0 }
index a29efb2..61ca009 100644 (file)
@@ -1636,14 +1636,16 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
        p->flags |= PF_MEMALLOC;
        reclaim_state.reclaimed_slab = 0;
        p->reclaim_state = &reclaim_state;
+
        shrink_zone(zone, &sc);
+
        p->reclaim_state = NULL;
        current->flags &= ~PF_MEMALLOC;
 
        if (sc.nr_reclaimed == 0)
                zone->last_unsuccessful_zone_reclaim = jiffies;
 
-       return sc.nr_reclaimed > nr_pages;
+       return sc.nr_reclaimed >= nr_pages;
 }
 #endif