struct zone *zone_table[1 << (ZONES_SHIFT + NODES_SHIFT)];
EXPORT_SYMBOL(zone_table);
+#ifdef CONFIG_NUMA
+static struct per_cpu_pageset
+ pageset_table[MAX_NR_ZONES*MAX_NUMNODES*NR_CPUS] __initdata;
+#endif
+
static char *zone_names[MAX_NR_ZONES] = { "DMA", "Normal", "HighMem" };
int min_free_kbytes = 1024;
for_each_zone(zone) {
struct per_cpu_pageset *pset;
- pset = &zone->pageset[cpu];
+ pset = zone_pcp(zone, cpu);
for (i = 0; i < ARRAY_SIZE(pset->pcp); i++) {
struct per_cpu_pages *pcp;
local_irq_save(flags);
cpu = smp_processor_id();
- p = &z->pageset[cpu];
+ p = zone_pcp(z,cpu);
if (pg == orig) {
- z->pageset[cpu].numa_hit++;
+ p->numa_hit++;
} else {
p->numa_miss++;
- zonelist->zones[0]->pageset[cpu].numa_foreign++;
+ zone_pcp(zonelist->zones[0], cpu)->numa_foreign++;
}
if (pg == NODE_DATA(numa_node_id()))
p->local_node++;
if (PageAnon(page))
page->mapping = NULL;
free_pages_check(__FUNCTION__, page);
- pcp = &zone->pageset[get_cpu()].pcp[cold];
+ pcp = &zone_pcp(zone, get_cpu())->pcp[cold];
local_irq_save(flags);
if (pcp->count >= pcp->high)
pcp->count -= free_pages_bulk(zone, pcp->batch, &pcp->list, 0);
if (order == 0) {
struct per_cpu_pages *pcp;
- pcp = &zone->pageset[get_cpu()].pcp[cold];
+ pcp = &zone_pcp(zone, get_cpu())->pcp[cold];
local_irq_save(flags);
if (pcp->count <= pcp->low)
pcp->count += rmqueue_bulk(zone, 0,
return 1;
}
+static inline int
+should_reclaim_zone(struct zone *z, unsigned int gfp_mask)
+{
+ if (!z->reclaim_pages)
+ return 0;
+ if (gfp_mask & __GFP_NORECLAIM)
+ return 0;
+ return 1;
+}
+
/*
* This is the 'heart' of the zoned buddy allocator.
*/
classzone_idx = zone_idx(zones[0]);
- restart:
+restart:
/* Go through the zonelist once, looking for a zone with enough free */
for (i = 0; (z = zones[i]) != NULL; i++) {
-
- if (!zone_watermark_ok(z, order, z->pages_low,
- classzone_idx, 0, 0))
- continue;
+ int do_reclaim = should_reclaim_zone(z, gfp_mask);
if (!cpuset_zone_allowed(z))
continue;
+ /*
+ * If the zone is to attempt early page reclaim then this loop
+ * will try to reclaim pages and check the watermark a second
+ * time before giving up and falling back to the next zone.
+ */
+zone_reclaim_retry:
+ if (!zone_watermark_ok(z, order, z->pages_low,
+ classzone_idx, 0, 0)) {
+ if (!do_reclaim)
+ continue;
+ else {
+ zone_reclaim(z, gfp_mask, order);
+ /* Only try reclaim once */
+ do_reclaim = 0;
+ goto zone_reclaim_retry;
+ }
+ }
+
page = buffered_rmqueue(z, order, gfp_mask);
if (page)
goto got_pg;
reclaim_state.reclaimed_slab = 0;
p->reclaim_state = &reclaim_state;
- did_some_progress = try_to_free_pages(zones, gfp_mask, order);
+ did_some_progress = try_to_free_pages(zones, gfp_mask);
p->reclaim_state = NULL;
p->flags &= ~PF_MEMALLOC;
if (!cpu_possible(cpu))
continue;
- pageset = zone->pageset + cpu;
+ pageset = zone_pcp(zone, cpu);
for (temperature = 0; temperature < 2; temperature++)
printk("cpu %d %s: low %d, high %d, batch %d\n",
memmap_init_zone((size), (nid), (zone), (start_pfn))
#endif
+static int __devinit zone_batchsize(struct zone *zone)
+{
+ int batch;
+
+ /*
+ * The per-cpu-pages pools are set to around 1000th of the
+ * size of the zone. But no more than 1/4 of a meg - there's
+ * no point in going beyond the size of L2 cache.
+ *
+ * OK, so we don't know how big the cache is. So guess.
+ */
+ batch = zone->present_pages / 1024;
+ if (batch * PAGE_SIZE > 256 * 1024)
+ batch = (256 * 1024) / PAGE_SIZE;
+ batch /= 4; /* We effectively *= 4 below */
+ if (batch < 1)
+ batch = 1;
+
+ /*
+ * Clamp the batch to a 2^n - 1 value. Having a power
+ * of 2 value was found to be more likely to have
+ * suboptimal cache aliasing properties in some cases.
+ *
+ * For example if 2 tasks are alternately allocating
+ * batches of pages, one task can end up with a lot
+ * of pages of one half of the possible page colors
+ * and the other with pages of the other colors.
+ */
+ batch = (1 << fls(batch + batch/2)) - 1;
+ return batch;
+}
+
+#ifdef CONFIG_NUMA
+/*
+ * Dynamicaly allocate memory for the
+ * per cpu pageset array in struct zone.
+ */
+static int __devinit process_zones(int cpu)
+{
+ struct zone *zone, *dzone;
+ int i;
+
+ for_each_zone(zone) {
+ struct per_cpu_pageset *npageset = NULL;
+
+ npageset = kmalloc_node(sizeof(struct per_cpu_pageset),
+ GFP_KERNEL, cpu_to_node(cpu));
+ if (!npageset) {
+ zone->pageset[cpu] = NULL;
+ goto bad;
+ }
+
+ if (zone->pageset[cpu]) {
+ memcpy(npageset, zone->pageset[cpu],
+ sizeof(struct per_cpu_pageset));
+
+ /* Relocate lists */
+ for (i = 0; i < 2; i++) {
+ INIT_LIST_HEAD(&npageset->pcp[i].list);
+ list_splice(&zone->pageset[cpu]->pcp[i].list,
+ &npageset->pcp[i].list);
+ }
+ } else {
+ struct per_cpu_pages *pcp;
+ unsigned long batch;
+
+ batch = zone_batchsize(zone);
+
+ pcp = &npageset->pcp[0]; /* hot */
+ pcp->count = 0;
+ pcp->low = 2 * batch;
+ pcp->high = 6 * batch;
+ pcp->batch = 1 * batch;
+ INIT_LIST_HEAD(&pcp->list);
+
+ pcp = &npageset->pcp[1]; /* cold*/
+ pcp->count = 0;
+ pcp->low = 0;
+ pcp->high = 2 * batch;
+ pcp->batch = 1 * batch;
+ INIT_LIST_HEAD(&pcp->list);
+ }
+ zone->pageset[cpu] = npageset;
+ }
+
+ return 0;
+bad:
+ for_each_zone(dzone) {
+ if (dzone == zone)
+ break;
+ kfree(dzone->pageset[cpu]);
+ dzone->pageset[cpu] = NULL;
+ }
+ return -ENOMEM;
+}
+
+static inline void free_zone_pagesets(int cpu)
+{
+#ifdef CONFIG_NUMA
+ struct zone *zone;
+
+ for_each_zone(zone) {
+ struct per_cpu_pageset *pset = zone_pcp(zone, cpu);
+
+ zone_pcp(zone, cpu) = NULL;
+ kfree(pset);
+ }
+#endif
+}
+
+static int __devinit pageset_cpuup_callback(struct notifier_block *nfb,
+ unsigned long action,
+ void *hcpu)
+{
+ int cpu = (long)hcpu;
+ int ret = NOTIFY_OK;
+
+ switch (action) {
+ case CPU_UP_PREPARE:
+ if (process_zones(cpu))
+ ret = NOTIFY_BAD;
+ break;
+#ifdef CONFIG_HOTPLUG_CPU
+ case CPU_DEAD:
+ free_zone_pagesets(cpu);
+ break;
+#endif
+ default:
+ break;
+ }
+ return ret;
+}
+
+static struct notifier_block pageset_notifier =
+ { &pageset_cpuup_callback, NULL, 0 };
+
+void __init setup_per_cpu_pageset()
+{
+ int err;
+
+ /* Initialize per_cpu_pageset for cpu 0.
+ * A cpuup callback will do this for every cpu
+ * as it comes online
+ */
+ err = process_zones(smp_processor_id());
+ BUG_ON(err);
+ register_cpu_notifier(&pageset_notifier);
+}
+
+#endif
+
/*
* Set up the zone data structures:
* - mark all pages reserved
zone->temp_priority = zone->prev_priority = DEF_PRIORITY;
- /*
- * The per-cpu-pages pools are set to around 1000th of the
- * size of the zone. But no more than 1/4 of a meg - there's
- * no point in going beyond the size of L2 cache.
- *
- * OK, so we don't know how big the cache is. So guess.
- */
- batch = zone->present_pages / 1024;
- if (batch * PAGE_SIZE > 256 * 1024)
- batch = (256 * 1024) / PAGE_SIZE;
- batch /= 4; /* We effectively *= 4 below */
- if (batch < 1)
- batch = 1;
-
- /*
- * Clamp the batch to a 2^n - 1 value. Having a power
- * of 2 value was found to be more likely to have
- * suboptimal cache aliasing properties in some cases.
- *
- * For example if 2 tasks are alternately allocating
- * batches of pages, one task can end up with a lot
- * of pages of one half of the possible page colors
- * and the other with pages of the other colors.
- */
- batch = (1 << fls(batch + batch/2)) - 1;
+ batch = zone_batchsize(zone);
for (cpu = 0; cpu < NR_CPUS; cpu++) {
struct per_cpu_pages *pcp;
+#ifdef CONFIG_NUMA
+ struct per_cpu_pageset *pgset;
+ pgset = &pageset_table[nid*MAX_NR_ZONES*NR_CPUS +
+ (j * NR_CPUS) + cpu];
+
+ zone->pageset[cpu] = pgset;
+#else
+ struct per_cpu_pageset *pgset = zone_pcp(zone, cpu);
+#endif
- pcp = &zone->pageset[cpu].pcp[0]; /* hot */
+ pcp = &pgset->pcp[0]; /* hot */
pcp->count = 0;
pcp->low = 2 * batch;
pcp->high = 6 * batch;
pcp->batch = 1 * batch;
INIT_LIST_HEAD(&pcp->list);
- pcp = &zone->pageset[cpu].pcp[1]; /* cold */
+ pcp = &pgset->pcp[1]; /* cold */
pcp->count = 0;
pcp->low = 0;
pcp->high = 2 * batch;
zone->nr_scan_inactive = 0;
zone->nr_active = 0;
zone->nr_inactive = 0;
+ atomic_set(&zone->reclaim_in_progress, -1);
if (!size)
continue;
.show = frag_show,
};
+/*
+ * Output information about zones in @pgdat.
+ */
+static int zoneinfo_show(struct seq_file *m, void *arg)
+{
+ pg_data_t *pgdat = arg;
+ struct zone *zone;
+ struct zone *node_zones = pgdat->node_zones;
+ unsigned long flags;
+
+ for (zone = node_zones; zone - node_zones < MAX_NR_ZONES; zone++) {
+ int i;
+
+ if (!zone->present_pages)
+ continue;
+
+ spin_lock_irqsave(&zone->lock, flags);
+ seq_printf(m, "Node %d, zone %8s", pgdat->node_id, zone->name);
+ seq_printf(m,
+ "\n pages free %lu"
+ "\n min %lu"
+ "\n low %lu"
+ "\n high %lu"
+ "\n active %lu"
+ "\n inactive %lu"
+ "\n scanned %lu (a: %lu i: %lu)"
+ "\n spanned %lu"
+ "\n present %lu",
+ zone->free_pages,
+ zone->pages_min,
+ zone->pages_low,
+ zone->pages_high,
+ zone->nr_active,
+ zone->nr_inactive,
+ zone->pages_scanned,
+ zone->nr_scan_active, zone->nr_scan_inactive,
+ zone->spanned_pages,
+ zone->present_pages);
+ seq_printf(m,
+ "\n protection: (%lu",
+ zone->lowmem_reserve[0]);
+ for (i = 1; i < ARRAY_SIZE(zone->lowmem_reserve); i++)
+ seq_printf(m, ", %lu", zone->lowmem_reserve[i]);
+ seq_printf(m,
+ ")"
+ "\n pagesets");
+ for (i = 0; i < ARRAY_SIZE(zone->pageset); i++) {
+ struct per_cpu_pageset *pageset;
+ int j;
+
+ pageset = zone_pcp(zone, i);
+ for (j = 0; j < ARRAY_SIZE(pageset->pcp); j++) {
+ if (pageset->pcp[j].count)
+ break;
+ }
+ if (j == ARRAY_SIZE(pageset->pcp))
+ continue;
+ for (j = 0; j < ARRAY_SIZE(pageset->pcp); j++) {
+ seq_printf(m,
+ "\n cpu: %i pcp: %i"
+ "\n count: %i"
+ "\n low: %i"
+ "\n high: %i"
+ "\n batch: %i",
+ i, j,
+ pageset->pcp[j].count,
+ pageset->pcp[j].low,
+ pageset->pcp[j].high,
+ pageset->pcp[j].batch);
+ }
+#ifdef CONFIG_NUMA
+ seq_printf(m,
+ "\n numa_hit: %lu"
+ "\n numa_miss: %lu"
+ "\n numa_foreign: %lu"
+ "\n interleave_hit: %lu"
+ "\n local_node: %lu"
+ "\n other_node: %lu",
+ pageset->numa_hit,
+ pageset->numa_miss,
+ pageset->numa_foreign,
+ pageset->interleave_hit,
+ pageset->local_node,
+ pageset->other_node);
+#endif
+ }
+ seq_printf(m,
+ "\n all_unreclaimable: %u"
+ "\n prev_priority: %i"
+ "\n temp_priority: %i"
+ "\n start_pfn: %lu",
+ zone->all_unreclaimable,
+ zone->prev_priority,
+ zone->temp_priority,
+ zone->zone_start_pfn);
+ spin_unlock_irqrestore(&zone->lock, flags);
+ seq_putc(m, '\n');
+ }
+ return 0;
+}
+
+struct seq_operations zoneinfo_op = {
+ .start = frag_start, /* iterate over all zones. The same as in
+ * fragmentation. */
+ .next = frag_next,
+ .stop = frag_stop,
+ .show = zoneinfo_show,
+};
+
static char *vmstat_text[] = {
"nr_dirty",
"nr_writeback",
min_pages = 128;
zone->pages_min = min_pages;
} else {
- /* if it's a lowmem zone, reserve a number of pages
+ /* if it's a lowmem zone, reserve a number of pages
* proportionate to the zone's size.
*/
- zone->pages_min = (pages_min * zone->present_pages) /
+ zone->pages_min = (pages_min * zone->present_pages) /
lowmem_pages;
}