The patch titled
zone reclaim with slab: avoid unecessary off node allocations
has been added to the -mm tree. Its filename is
zone-reclaim-with-slab-avoid-unecessary-off-node-allocations.patch
See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this
------------------------------------------------------
Subject: zone reclaim with slab: avoid unecessary off node allocations
From: Christoph Lameter <clameter@xxxxxxx>
Minor performance fix.
If we reclaimed enough slab pages from a zone then we can avoid going off
node with the current allocation. Take care of updating nr_reclaimed when
reclaiming from the slab.
Signed-off-by: Christoph Lameter <clameter@xxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxx>
---
mm/vmscan.c | 17 ++++++++++++-----
1 files changed, 12 insertions(+), 5 deletions(-)
diff -puN mm/vmscan.c~zone-reclaim-with-slab-avoid-unecessary-off-node-allocations mm/vmscan.c
--- a/mm/vmscan.c~zone-reclaim-with-slab-avoid-unecessary-off-node-allocations
+++ a/mm/vmscan.c
@@ -1566,6 +1566,7 @@ static int __zone_reclaim(struct zone *z
.gfp_mask = gfp_mask,
.swappiness = vm_swappiness,
};
+ unsigned long slab_reclaimable;
disable_swap_token();
cond_resched();
@@ -1592,7 +1593,8 @@ static int __zone_reclaim(struct zone *z
} while (priority >= 0 && nr_reclaimed < nr_pages);
}
- if (zone_page_state(zone, NR_SLAB_RECLAIMABLE) > zone->min_slab_pages) {
+ slab_reclaimable = zone_page_state(zone, NR_SLAB_RECLAIMABLE);
+ if (slab_reclaimable > zone->min_slab_pages) {
/*
* shrink_slab() does not currently allow us to determine how
* many pages were freed in this zone. So we take the current
@@ -1603,12 +1605,17 @@ static int __zone_reclaim(struct zone *z
* Note that shrink_slab will free memory on all zones and may
* take a long time.
*/
- unsigned long limit = zone_page_state(zone,
- NR_SLAB_RECLAIMABLE) - nr_pages;
-
while (shrink_slab(sc.nr_scanned, gfp_mask, order) &&
- zone_page_state(zone, NR_SLAB_RECLAIMABLE) > limit)
+ zone_page_state(zone, NR_SLAB_RECLAIMABLE) >
+ slab_reclaimable - nr_pages)
;
+
+ /*
+ * Update nr_reclaimed by the number of slab pages we
+ * reclaimed from this zone.
+ */
+ nr_reclaimed += slab_reclaimable -
+ zone_page_state(zone, NR_SLAB_RECLAIMABLE);
}
p->reclaim_state = NULL;
_
Patches currently in -mm which might be from clameter@xxxxxxx are
zvc-overstep-counters.patch
zvc-scale-thresholds-depending-on-the-size-of-the-system.patch
page-migration-replace-radix_tree_lookup_slot-with-radix_tree_lockup.patch
reduce-max_nr_zones-remove-two-strange-uses-of-max_nr_zones.patch
reduce-max_nr_zones-fix-max_nr_zones-array-initializations.patch
reduce-max_nr_zones-make-display-of-highmem-counters-conditional-on-config_highmem.patch
reduce-max_nr_zones-make-display-of-highmem-counters-conditional-on-config_highmem-tidy.patch
reduce-max_nr_zones-move-highmem-counters-into-highmemc-h.patch
reduce-max_nr_zones-move-highmem-counters-into-highmemc-h-fix.patch
reduce-max_nr_zones-page-allocator-zone_highmem-cleanup.patch
reduce-max_nr_zones-use-enum-to-define-zones-reformat-and-comment.patch
reduce-max_nr_zones-use-enum-to-define-zones-reformat-and-comment-cleanup.patch
reduce-max_nr_zones-make-zone_dma32-optional.patch
reduce-max_nr_zones-make-zone_highmem-optional.patch
reduce-max_nr_zones-make-zone_highmem-optional-fix.patch
reduce-max_nr_zones-make-zone_highmem-optional-fix-fix.patch
reduce-max_nr_zones-remove-display-of-counters-for-unconfigured-zones.patch
reduce-max_nr_zones-fix-i386-srat-check-for-max_nr_zones.patch
mempolicies-fix-policy_zone-check.patch
apply-type-enum-zone_type.patch
apply-type-enum-zone_type-fix.patch
linearly-index-zone-node_zonelists.patch
slab-respect-architecture-and-caller-mandated-alignment.patch
slab-optimize-kmalloc_node-the-same-way-as-kmalloc.patch
slab-optimize-kmalloc_node-the-same-way-as-kmalloc-fix.patch
slab-extract-__kmem_cache_destroy-from-kmem_cache_destroy.patch
slab-do-not-panic-when-alloc_kmemlist-fails-and-slab-is-up.patch
add-__gfp_thisnode-to-avoid-fallback-to-other-nodes-and-ignore.patch
add-__gfp_thisnode-to-avoid-fallback-to-other-nodes-and-ignore-fix.patch
sys_move_pages-do-not-fall-back-to-other-nodes.patch
guarantee-that-the-uncached-allocator-gets-pages-on-the-correct.patch
cleanup-add-zone-pointer-to-get_page_from_freelist.patch
profiling-require-buffer-allocation-on-the-correct-node.patch
define-easier-to-handle-gfp_thisnode.patch
optimize-free_one_page.patch
do-not-check-unpopulated-zones-for-draining-and-counter.patch
extract-the-allocpercpu-functions-from-the-slab-allocator.patch
replace-min_unmapped_ratio-by-min_unmapped_pages-in-struct-zone.patch
zvc-support-nr_slab_reclaimable--nr_slab_unreclaimable.patch
zone_reclaim-dynamic-slab-reclaim.patch
zone_reclaim-dynamic-slab-reclaim-tidy.patch
zone-reclaim-with-slab-avoid-unecessary-off-node-allocations.patch
x86-implement-always-locked-bit-ops-for-memory-shared-with-an-smp-hypervisor.patch
zvc-support-nr_slab_reclaimable--nr_slab_unreclaimable-swap_prefetch.patch
reduce-max_nr_zones-swap_prefetch-remove-incorrect-use-of-zone_highmem.patch
readahead-state-based-method-aging-accounting-apply-type-enum-zone_type-readahead.patch
-
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
