aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorChristoph Lameter <clameter@sgi.com>2007-05-09 02:35:14 -0700
committerLinus Torvalds <torvalds@woody.linux-foundation.org>2007-05-09 12:30:56 -0700
commit4037d452202e34214e8a939fa5621b2b3bbb45b7 (patch)
tree31b59c0ca94fba4d53b6738b0bad3d1e9fde3063
parent77461ab33229d48614402decfb1b2eaa6d446861 (diff)
Move remote node draining out of slab allocators
Currently the slab allocators contain callbacks into the page allocator to perform the draining of pagesets on remote nodes. This requires SLUB to have a whole subsystem in order to be compatible with SLAB. Moving node draining out of the slab allocators avoids a section of code in SLUB. Move the node draining so that is is done when the vm statistics are updated. At that point we are already touching all the cachelines with the pagesets of a processor. Add a expire counter there. If we have to update per zone or global vm statistics then assume that the pageset will require subsequent draining. The expire counter will be decremented on each vm stats update pass until it reaches zero. Then we will drain one batch from the pageset. The draining will cause vm counter updates which will then cause another expiration until the pcp is empty. So we will drain a batch every 3 seconds. Note that remote node draining is a somewhat esoteric feature that is required on large NUMA systems because otherwise significant portions of system memory can become trapped in pcp queues. The number of pcp is determined by the number of processors and nodes in a system. A system with 4 processors and 2 nodes has 8 pcps which is okay. But a system with 1024 processors and 512 nodes has 512k pcps with a high potential for large amount of memory being caught in them. Signed-off-by: Christoph Lameter <clameter@sgi.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
-rw-r--r--include/linux/gfp.h6
-rw-r--r--include/linux/mmzone.h3
-rw-r--r--mm/page_alloc.c45
-rw-r--r--mm/slab.c6
-rw-r--r--mm/slub.c84
-rw-r--r--mm/vmstat.c54
6 files changed, 67 insertions, 131 deletions
diff --git a/include/linux/gfp.h b/include/linux/gfp.h
index 97a36c3d96e..0d2ef0b082a 100644
--- a/include/linux/gfp.h
+++ b/include/linux/gfp.h
@@ -176,10 +176,6 @@ extern void FASTCALL(free_cold_page(struct page *page));
#define free_page(addr) free_pages((addr),0)
void page_alloc_init(void);
-#ifdef CONFIG_NUMA
-void drain_node_pages(int node);
-#else
-static inline void drain_node_pages(int node) { };
-#endif
+void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp);
#endif /* __LINUX_GFP_H */
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 2f1544e8304..d09b1345a3a 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -83,6 +83,9 @@ struct per_cpu_pages {
struct per_cpu_pageset {
struct per_cpu_pages pcp[2]; /* 0: hot. 1: cold */
+#ifdef CONFIG_NUMA
+ s8 expire;
+#endif
#ifdef CONFIG_SMP
s8 stat_threshold;
s8 vm_stat_diff[NR_VM_ZONE_STAT_ITEMS];
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index d53cbf8acb8..f9b5d6d5f4d 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -691,43 +691,26 @@ static void __init setup_nr_node_ids(void) {}
#ifdef CONFIG_NUMA
/*
- * Called from the slab reaper to drain pagesets on a particular node that
- * belongs to the currently executing processor.
+ * Called from the vmstat counter updater to drain pagesets of this
+ * currently executing processor on remote nodes after they have
+ * expired.
+ *
* Note that this function must be called with the thread pinned to
* a single processor.
*/
-void drain_node_pages(int nodeid)
+void drain_zone_pages(struct zone *zone, struct per_cpu_pages *pcp)
{
- int i;
- enum zone_type z;
unsigned long flags;
+ int to_drain;
- for (z = 0; z < MAX_NR_ZONES; z++) {
- struct zone *zone = NODE_DATA(nodeid)->node_zones + z;
- struct per_cpu_pageset *pset;
-
- if (!populated_zone(zone))
- continue;
-
- pset = zone_pcp(zone, smp_processor_id());
- for (i = 0; i < ARRAY_SIZE(pset->pcp); i++) {
- struct per_cpu_pages *pcp;
-
- pcp = &pset->pcp[i];
- if (pcp->count) {
- int to_drain;
-
- local_irq_save(flags);
- if (pcp->count >= pcp->batch)
- to_drain = pcp->batch;
- else
- to_drain = pcp->count;
- free_pages_bulk(zone, to_drain, &pcp->list, 0);
- pcp->count -= to_drain;
- local_irq_restore(flags);
- }
- }
- }
+ local_irq_save(flags);
+ if (pcp->count >= pcp->batch)
+ to_drain = pcp->batch;
+ else
+ to_drain = pcp->count;
+ free_pages_bulk(zone, to_drain, &pcp->list, 0);
+ pcp->count -= to_drain;
+ local_irq_restore(flags);
}
#endif
diff --git a/mm/slab.c b/mm/slab.c
index e50908b2bfa..944b20581f8 100644
--- a/mm/slab.c
+++ b/mm/slab.c
@@ -928,12 +928,6 @@ static void next_reap_node(void)
{
int node = __get_cpu_var(reap_node);
- /*
- * Also drain per cpu pages on remote zones
- */
- if (node != numa_node_id())
- drain_node_pages(node);
-
node = next_node(node, node_online_map);
if (unlikely(node >= MAX_NUMNODES))
node = first_node(node_online_map);
diff --git a/mm/slub.c b/mm/slub.c
index dbb206503a8..bd2efae02bc 100644
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -2530,90 +2530,6 @@ static struct notifier_block __cpuinitdata slab_notifier =
#endif
-#ifdef CONFIG_NUMA
-
-/*****************************************************************
- * Generic reaper used to support the page allocator
- * (the cpu slabs are reaped by a per slab workqueue).
- *
- * Maybe move this to the page allocator?
- ****************************************************************/
-
-static DEFINE_PER_CPU(unsigned long, reap_node);
-
-static void init_reap_node(int cpu)
-{
- int node;
-
- node = next_node(cpu_to_node(cpu), node_online_map);
- if (node == MAX_NUMNODES)
- node = first_node(node_online_map);
-
- __get_cpu_var(reap_node) = node;
-}
-
-static void next_reap_node(void)
-{
- int node = __get_cpu_var(reap_node);
-
- /*
- * Also drain per cpu pages on remote zones
- */
- if (node != numa_node_id())
- drain_node_pages(node);
-
- node = next_node(node, node_online_map);
- if (unlikely(node >= MAX_NUMNODES))
- node = first_node(node_online_map);
- __get_cpu_var(reap_node) = node;
-}
-#else
-#define init_reap_node(cpu) do { } while (0)
-#define next_reap_node(void) do { } while (0)
-#endif
-
-#define REAPTIMEOUT_CPUC (2*HZ)
-
-#ifdef CONFIG_SMP
-static DEFINE_PER_CPU(struct delayed_work, reap_work);
-
-static void cache_reap(struct work_struct *unused)
-{
- next_reap_node();
- schedule_delayed_work(&__get_cpu_var(reap_work),
- REAPTIMEOUT_CPUC);
-}
-
-static void __devinit start_cpu_timer(int cpu)
-{
- struct delayed_work *reap_work = &per_cpu(reap_work, cpu);
-
- /*
- * When this gets called from do_initcalls via cpucache_init(),
- * init_workqueues() has already run, so keventd will be setup
- * at that time.
- */
- if (keventd_up() && reap_work->work.func == NULL) {
- init_reap_node(cpu);
- INIT_DELAYED_WORK(reap_work, cache_reap);
- schedule_delayed_work_on(cpu, reap_work, HZ + 3 * cpu);
- }
-}
-
-static int __init cpucache_init(void)
-{
- int cpu;
-
- /*
- * Register the timers that drain pcp pages and update vm statistics
- */
- for_each_online_cpu(cpu)
- start_cpu_timer(cpu);
- return 0;
-}
-__initcall(cpucache_init);
-#endif
-
void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, void *caller)
{
struct kmem_cache *s = get_slab(size, gfpflags);
diff --git a/mm/vmstat.c b/mm/vmstat.c
index 006eb762186..9832d9a41d8 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -281,6 +281,17 @@ EXPORT_SYMBOL(dec_zone_page_state);
/*
* Update the zone counters for one cpu.
+ *
+ * Note that refresh_cpu_vm_stats strives to only access
+ * node local memory. The per cpu pagesets on remote zones are placed
+ * in the memory local to the processor using that pageset. So the
+ * loop over all zones will access a series of cachelines local to
+ * the processor.
+ *
+ * The call to zone_page_state_add updates the cachelines with the
+ * statistics in the remote zone struct as well as the global cachelines
+ * with the global counters. These could cause remote node cache line
+ * bouncing and will have to be only done when necessary.
*/
void refresh_cpu_vm_stats(int cpu)
{
@@ -289,21 +300,54 @@ void refresh_cpu_vm_stats(int cpu)
unsigned long flags;
for_each_zone(zone) {
- struct per_cpu_pageset *pcp;
+ struct per_cpu_pageset *p;
if (!populated_zone(zone))
continue;
- pcp = zone_pcp(zone, cpu);
+ p = zone_pcp(zone, cpu);
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
- if (pcp->vm_stat_diff[i]) {
+ if (p->vm_stat_diff[i]) {
local_irq_save(flags);
- zone_page_state_add(pcp->vm_stat_diff[i],
+ zone_page_state_add(p->vm_stat_diff[i],
zone, i);
- pcp->vm_stat_diff[i] = 0;
+ p->vm_stat_diff[i] = 0;
+#ifdef CONFIG_NUMA
+ /* 3 seconds idle till flush */
+ p->expire = 3;
+#endif
local_irq_restore(flags);
}
+#ifdef CONFIG_NUMA
+ /*
+ * Deal with draining the remote pageset of this
+ * processor
+ *
+ * Check if there are pages remaining in this pageset
+ * if not then there is nothing to expire.
+ */
+ if (!p->expire || (!p->pcp[0].count && !p->pcp[1].count))
+ continue;
+
+ /*
+ * We never drain zones local to this processor.
+ */
+ if (zone_to_nid(zone) == numa_node_id()) {
+ p->expire = 0;
+ continue;
+ }
+
+ p->expire--;
+ if (p->expire)
+ continue;
+
+ if (p->pcp[0].count)
+ drain_zone_pages(zone, p->pcp + 0);
+
+ if (p->pcp[1].count)
+ drain_zone_pages(zone, p->pcp + 1);
+#endif
}
}