X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=mm%2Fvmstat.c;h=a2b6a9f96e5c6390625ec031763a74de9fa5f5b2;hb=1c0d04c9e44f4a248335c33d2be7c7f7b06ff359;hp=ad456202ff1a4ab28b0ab24e697c00d89efe5af6;hpb=f6ac2354d791195ca40822b84d73d48a4e8b7f2b;p=linux-2.6

diff --git a/mm/vmstat.c b/mm/vmstat.c
index ad456202ff..a2b6a9f96e 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -3,145 +3,380 @@
  *
  *  Manages VM statistics
  *  Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
+ *
+ *  zoned VM statistics
+ *  Copyright (C) 2006 Silicon Graphics, Inc.,
+ *		Christoph Lameter <christoph@lameter.com>
  */
 
 #include <linux/config.h>
 #include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/cpu.h>
+
+void __get_zone_counts(unsigned long *active, unsigned long *inactive,
+			unsigned long *free, struct pglist_data *pgdat)
+{
+	struct zone *zones = pgdat->node_zones;
+	int i;
+
+	*active = 0;
+	*inactive = 0;
+	*free = 0;
+	for (i = 0; i < MAX_NR_ZONES; i++) {
+		*active += zones[i].nr_active;
+		*inactive += zones[i].nr_inactive;
+		*free += zones[i].free_pages;
+	}
+}
+
+void get_zone_counts(unsigned long *active,
+		unsigned long *inactive, unsigned long *free)
+{
+	struct pglist_data *pgdat;
+
+	*active = 0;
+	*inactive = 0;
+	*free = 0;
+	for_each_online_pgdat(pgdat) {
+		unsigned long l, m, n;
+		__get_zone_counts(&l, &m, &n, pgdat);
+		*active += l;
+		*inactive += m;
+		*free += n;
+	}
+}
+
+#ifdef CONFIG_VM_EVENT_COUNTERS
+DEFINE_PER_CPU(struct vm_event_state, vm_event_states) = {{0}};
+EXPORT_PER_CPU_SYMBOL(vm_event_states);
+
+static void sum_vm_events(unsigned long *ret, cpumask_t *cpumask)
+{
+	int cpu = 0;
+	int i;
+
+	memset(ret, 0, NR_VM_EVENT_ITEMS * sizeof(unsigned long));
+
+	cpu = first_cpu(*cpumask);
+	while (cpu < NR_CPUS) {
+		struct vm_event_state *this = &per_cpu(vm_event_states, cpu);
+
+		cpu = next_cpu(cpu, *cpumask);
+
+		if (cpu < NR_CPUS)
+			prefetch(&per_cpu(vm_event_states, cpu));
+
+
+		for (i = 0; i < NR_VM_EVENT_ITEMS; i++)
+			ret[i] += this->event[i];
+	}
+}
 
 /*
- * Accumulate the page_state information across all CPUs.
+ * Accumulate the vm event counters across all CPUs.
  * The result is unavoidably approximate - it can change
  * during and after execution of this function.
+*/
+void all_vm_events(unsigned long *ret)
+{
+	sum_vm_events(ret, &cpu_online_map);
+}
+EXPORT_SYMBOL_GPL(all_vm_events);
+
+#ifdef CONFIG_HOTPLUG
+/*
+ * Fold the foreign cpu events into our own.
+ *
+ * This is adding to the events on one processor
+ * but keeps the global counts constant.
+ */
+void vm_events_fold_cpu(int cpu)
+{
+	struct vm_event_state *fold_state = &per_cpu(vm_event_states, cpu);
+	int i;
+
+	for (i = 0; i < NR_VM_EVENT_ITEMS; i++) {
+		count_vm_events(i, fold_state->event[i]);
+		fold_state->event[i] = 0;
+	}
+}
+#endif /* CONFIG_HOTPLUG */
+
+#endif /* CONFIG_VM_EVENT_COUNTERS */
+
+/*
+ * Manage combined zone based / global counters
+ *
+ * vm_stat contains the global counters
  */
-DEFINE_PER_CPU(struct page_state, page_states) = {0};
+atomic_long_t vm_stat[NR_VM_ZONE_STAT_ITEMS];
+EXPORT_SYMBOL(vm_stat);
 
-atomic_t nr_pagecache = ATOMIC_INIT(0);
-EXPORT_SYMBOL(nr_pagecache);
 #ifdef CONFIG_SMP
-DEFINE_PER_CPU(long, nr_pagecache_local) = 0;
-#endif
 
-static void __get_page_state(struct page_state *ret, int nr, cpumask_t *cpumask)
+static int calculate_threshold(struct zone *zone)
 {
-	unsigned cpu;
+	int threshold;
+	int mem;	/* memory in 128 MB units */
+
+	/*
+	 * The threshold scales with the number of processors and the amount
+	 * of memory per zone. More memory means that we can defer updates for
+	 * longer, more processors could lead to more contention.
+ 	 * fls() is used to have a cheap way of logarithmic scaling.
+	 *
+	 * Some sample thresholds:
+	 *
+	 * Threshold	Processors	(fls)	Zonesize	fls(mem+1)
+	 * ------------------------------------------------------------------
+	 * 8		1		1	0.9-1 GB	4
+	 * 16		2		2	0.9-1 GB	4
+	 * 20 		2		2	1-2 GB		5
+	 * 24		2		2	2-4 GB		6
+	 * 28		2		2	4-8 GB		7
+	 * 32		2		2	8-16 GB		8
+	 * 4		2		2	<128M		1
+	 * 30		4		3	2-4 GB		5
+	 * 48		4		3	8-16 GB		8
+	 * 32		8		4	1-2 GB		4
+	 * 32		8		4	0.9-1GB		4
+	 * 10		16		5	<128M		1
+	 * 40		16		5	900M		4
+	 * 70		64		7	2-4 GB		5
+	 * 84		64		7	4-8 GB		6
+	 * 108		512		9	4-8 GB		6
+	 * 125		1024		10	8-16 GB		8
+	 * 125		1024		10	16-32 GB	9
+	 */
+
+	mem = zone->present_pages >> (27 - PAGE_SHIFT);
+
+	threshold = 2 * fls(num_online_cpus()) * (1 + fls(mem));
+
+	/*
+	 * Maximum threshold is 125
+	 */
+	threshold = min(125, threshold);
+
+	return threshold;
+}
 
-	memset(ret, 0, nr * sizeof(unsigned long));
-	cpus_and(*cpumask, *cpumask, cpu_online_map);
+/*
+ * Refresh the thresholds for each zone.
+ */
+static void refresh_zone_stat_thresholds(void)
+{
+	struct zone *zone;
+	int cpu;
+	int threshold;
 
-	for_each_cpu_mask(cpu, *cpumask) {
-		unsigned long *in;
-		unsigned long *out;
-		unsigned off;
-		unsigned next_cpu;
+	for_each_zone(zone) {
 
-		in = (unsigned long *)&per_cpu(page_states, cpu);
+		if (!zone->present_pages)
+			continue;
 
-		next_cpu = next_cpu(cpu, *cpumask);
-		if (likely(next_cpu < NR_CPUS))
-			prefetch(&per_cpu(page_states, next_cpu));
+		threshold = calculate_threshold(zone);
 
-		out = (unsigned long *)ret;
-		for (off = 0; off < nr; off++)
-			*out++ += *in++;
+		for_each_online_cpu(cpu)
+			zone_pcp(zone, cpu)->stat_threshold = threshold;
 	}
 }
 
-void get_page_state_node(struct page_state *ret, int node)
+/*
+ * For use when we know that interrupts are disabled.
+ */
+void __mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
+				int delta)
 {
-	int nr;
-	cpumask_t mask = node_to_cpumask(node);
+	struct per_cpu_pageset *pcp = zone_pcp(zone, smp_processor_id());
+	s8 *p = pcp->vm_stat_diff + item;
+	long x;
 
-	nr = offsetof(struct page_state, GET_PAGE_STATE_LAST);
-	nr /= sizeof(unsigned long);
+	x = delta + *p;
 
-	__get_page_state(ret, nr+1, &mask);
+	if (unlikely(x > pcp->stat_threshold || x < -pcp->stat_threshold)) {
+		zone_page_state_add(x, zone, item);
+		x = 0;
+	}
+	*p = x;
 }
+EXPORT_SYMBOL(__mod_zone_page_state);
 
-void get_page_state(struct page_state *ret)
+/*
+ * For an unknown interrupt state
+ */
+void mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
+					int delta)
 {
-	int nr;
-	cpumask_t mask = CPU_MASK_ALL;
-
-	nr = offsetof(struct page_state, GET_PAGE_STATE_LAST);
-	nr /= sizeof(unsigned long);
+	unsigned long flags;
 
-	__get_page_state(ret, nr + 1, &mask);
+	local_irq_save(flags);
+	__mod_zone_page_state(zone, item, delta);
+	local_irq_restore(flags);
 }
+EXPORT_SYMBOL(mod_zone_page_state);
 
-void get_full_page_state(struct page_state *ret)
+/*
+ * Optimized increment and decrement functions.
+ *
+ * These are only for a single page and therefore can take a struct page *
+ * argument instead of struct zone *. This allows the inclusion of the code
+ * generated for page_zone(page) into the optimized functions.
+ *
+ * No overflow check is necessary and therefore the differential can be
+ * incremented or decremented in place which may allow the compilers to
+ * generate better code.
+ * The increment or decrement is known and therefore one boundary check can
+ * be omitted.
+ *
+ * NOTE: These functions are very performance sensitive. Change only
+ * with care.
+ *
+ * Some processors have inc/dec instructions that are atomic vs an interrupt.
+ * However, the code must first determine the differential location in a zone
+ * based on the processor number and then inc/dec the counter. There is no
+ * guarantee without disabling preemption that the processor will not change
+ * in between and therefore the atomicity vs. interrupt cannot be exploited
+ * in a useful way here.
+ */
+static void __inc_zone_state(struct zone *zone, enum zone_stat_item item)
 {
-	cpumask_t mask = CPU_MASK_ALL;
+	struct per_cpu_pageset *pcp = zone_pcp(zone, smp_processor_id());
+	s8 *p = pcp->vm_stat_diff + item;
+
+	(*p)++;
+
+	if (unlikely(*p > pcp->stat_threshold)) {
+		int overstep = pcp->stat_threshold / 2;
 
-	__get_page_state(ret, sizeof(*ret) / sizeof(unsigned long), &mask);
+		zone_page_state_add(*p + overstep, zone, item);
+		*p = -overstep;
+	}
 }
 
-unsigned long read_page_state_offset(unsigned long offset)
+void __inc_zone_page_state(struct page *page, enum zone_stat_item item)
 {
-	unsigned long ret = 0;
-	int cpu;
+	__inc_zone_state(page_zone(page), item);
+}
+EXPORT_SYMBOL(__inc_zone_page_state);
 
-	for_each_online_cpu(cpu) {
-		unsigned long in;
+void __dec_zone_page_state(struct page *page, enum zone_stat_item item)
+{
+	struct zone *zone = page_zone(page);
+	struct per_cpu_pageset *pcp = zone_pcp(zone, smp_processor_id());
+	s8 *p = pcp->vm_stat_diff + item;
+
+	(*p)--;
+
+	if (unlikely(*p < - pcp->stat_threshold)) {
+		int overstep = pcp->stat_threshold / 2;
 
-		in = (unsigned long)&per_cpu(page_states, cpu) + offset;
-		ret += *((unsigned long *)in);
+		zone_page_state_add(*p - overstep, zone, item);
+		*p = overstep;
 	}
-	return ret;
+}
+EXPORT_SYMBOL(__dec_zone_page_state);
+
+void inc_zone_state(struct zone *zone, enum zone_stat_item item)
+{
+	unsigned long flags;
+
+	local_irq_save(flags);
+	__inc_zone_state(zone, item);
+	local_irq_restore(flags);
 }
 
-void __mod_page_state_offset(unsigned long offset, unsigned long delta)
+void inc_zone_page_state(struct page *page, enum zone_stat_item item)
 {
-	void *ptr;
+	unsigned long flags;
+	struct zone *zone;
 
-	ptr = &__get_cpu_var(page_states);
-	*(unsigned long *)(ptr + offset) += delta;
+	zone = page_zone(page);
+	local_irq_save(flags);
+	__inc_zone_state(zone, item);
+	local_irq_restore(flags);
 }
-EXPORT_SYMBOL(__mod_page_state_offset);
+EXPORT_SYMBOL(inc_zone_page_state);
 
-void mod_page_state_offset(unsigned long offset, unsigned long delta)
+void dec_zone_page_state(struct page *page, enum zone_stat_item item)
 {
 	unsigned long flags;
-	void *ptr;
 
 	local_irq_save(flags);
-	ptr = &__get_cpu_var(page_states);
-	*(unsigned long *)(ptr + offset) += delta;
+	__dec_zone_page_state(page, item);
 	local_irq_restore(flags);
 }
-EXPORT_SYMBOL(mod_page_state_offset);
+EXPORT_SYMBOL(dec_zone_page_state);
 
-void __get_zone_counts(unsigned long *active, unsigned long *inactive,
-			unsigned long *free, struct pglist_data *pgdat)
+/*
+ * Update the zone counters for one cpu.
+ */
+void refresh_cpu_vm_stats(int cpu)
 {
-	struct zone *zones = pgdat->node_zones;
+	struct zone *zone;
 	int i;
+	unsigned long flags;
 
-	*active = 0;
-	*inactive = 0;
-	*free = 0;
-	for (i = 0; i < MAX_NR_ZONES; i++) {
-		*active += zones[i].nr_active;
-		*inactive += zones[i].nr_inactive;
-		*free += zones[i].free_pages;
+	for_each_zone(zone) {
+		struct per_cpu_pageset *pcp;
+
+		if (!populated_zone(zone))
+			continue;
+
+		pcp = zone_pcp(zone, cpu);
+
+		for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
+			if (pcp->vm_stat_diff[i]) {
+				local_irq_save(flags);
+				zone_page_state_add(pcp->vm_stat_diff[i],
+					zone, i);
+				pcp->vm_stat_diff[i] = 0;
+				local_irq_restore(flags);
+			}
 	}
 }
 
-void get_zone_counts(unsigned long *active,
-		unsigned long *inactive, unsigned long *free)
+static void __refresh_cpu_vm_stats(void *dummy)
 {
-	struct pglist_data *pgdat;
+	refresh_cpu_vm_stats(smp_processor_id());
+}
 
-	*active = 0;
-	*inactive = 0;
-	*free = 0;
-	for_each_online_pgdat(pgdat) {
-		unsigned long l, m, n;
-		__get_zone_counts(&l, &m, &n, pgdat);
-		*active += l;
-		*inactive += m;
-		*free += n;
+/*
+ * Consolidate all counters.
+ *
+ * Note that the result is less inaccurate but still inaccurate
+ * if concurrent processes are allowed to run.
+ */
+void refresh_vm_stats(void)
+{
+	on_each_cpu(__refresh_cpu_vm_stats, NULL, 0, 1);
+}
+EXPORT_SYMBOL(refresh_vm_stats);
+
+#endif
+
+#ifdef CONFIG_NUMA
+/*
+ * zonelist = the list of zones passed to the allocator
+ * z 	    = the zone from which the allocation occurred.
+ *
+ * Must be called with interrupts disabled.
+ */
+void zone_statistics(struct zonelist *zonelist, struct zone *z)
+{
+	if (z->zone_pgdat == zonelist->zones[0]->zone_pgdat) {
+		__inc_zone_state(z, NUMA_HIT);
+	} else {
+		__inc_zone_state(z, NUMA_MISS);
+		__inc_zone_state(zonelist->zones[0], NUMA_FOREIGN);
 	}
+	if (z->node == numa_node_id())
+		__inc_zone_state(z, NUMA_LOCAL);
+	else
+		__inc_zone_state(z, NUMA_OTHER);
 }
+#endif
 
 #ifdef CONFIG_PROC_FS
 
@@ -203,23 +438,51 @@ struct seq_operations fragmentation_op = {
 	.show	= frag_show,
 };
 
+#ifdef CONFIG_ZONE_DMA32
+#define TEXT_FOR_DMA32(xx) xx "_dma32",
+#else
+#define TEXT_FOR_DMA32(xx)
+#endif
+
+#ifdef CONFIG_HIGHMEM
+#define TEXT_FOR_HIGHMEM(xx) xx "_high",
+#else
+#define TEXT_FOR_HIGHMEM(xx)
+#endif
+
+#define TEXTS_FOR_ZONES(xx) xx "_dma", TEXT_FOR_DMA32(xx) xx "_normal", \
+					TEXT_FOR_HIGHMEM(xx)
+
 static char *vmstat_text[] = {
+	/* Zoned VM counters */
+	"nr_anon_pages",
+	"nr_mapped",
+	"nr_file_pages",
+	"nr_slab_reclaimable",
+	"nr_slab_unreclaimable",
+	"nr_page_table_pages",
 	"nr_dirty",
 	"nr_writeback",
 	"nr_unstable",
-	"nr_page_table_pages",
-	"nr_mapped",
-	"nr_slab",
+	"nr_bounce",
+	"nr_vmscan_write",
+
+#ifdef CONFIG_NUMA
+	"numa_hit",
+	"numa_miss",
+	"numa_foreign",
+	"numa_interleave",
+	"numa_local",
+	"numa_other",
+#endif
 
+#ifdef CONFIG_VM_EVENT_COUNTERS
 	"pgpgin",
 	"pgpgout",
 	"pswpin",
 	"pswpout",
 
-	"pgalloc_high",
-	"pgalloc_normal",
-	"pgalloc_dma32",
-	"pgalloc_dma",
+	TEXTS_FOR_ZONES("pgalloc")
 
 	"pgfree",
 	"pgactivate",
@@ -228,25 +491,10 @@ static char *vmstat_text[] = {
 	"pgfault",
 	"pgmajfault",
 
-	"pgrefill_high",
-	"pgrefill_normal",
-	"pgrefill_dma32",
-	"pgrefill_dma",
-
-	"pgsteal_high",
-	"pgsteal_normal",
-	"pgsteal_dma32",
-	"pgsteal_dma",
-
-	"pgscan_kswapd_high",
-	"pgscan_kswapd_normal",
-	"pgscan_kswapd_dma32",
-	"pgscan_kswapd_dma",
-
-	"pgscan_direct_high",
-	"pgscan_direct_normal",
-	"pgscan_direct_dma32",
-	"pgscan_direct_dma",
+	TEXTS_FOR_ZONES("pgrefill")
+	TEXTS_FOR_ZONES("pgsteal")
+	TEXTS_FOR_ZONES("pgscan_kswapd")
+	TEXTS_FOR_ZONES("pgscan_direct")
 
 	"pginodesteal",
 	"slabs_scanned",
@@ -256,7 +504,7 @@ static char *vmstat_text[] = {
 	"allocstall",
 
 	"pgrotated",
-	"nr_bounce",
+#endif
 };
 
 /*
@@ -297,6 +545,11 @@ static int zoneinfo_show(struct seq_file *m, void *arg)
 			   zone->nr_scan_active, zone->nr_scan_inactive,
 			   zone->spanned_pages,
 			   zone->present_pages);
+
+		for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
+			seq_printf(m, "\n    %-12s %lu", vmstat_text[i],
+					zone_page_state(zone, i));
+
 		seq_printf(m,
 			   "\n        protection: (%lu",
 			   zone->lowmem_reserve[0]);
@@ -327,20 +580,9 @@ static int zoneinfo_show(struct seq_file *m, void *arg)
 					   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);
+#ifdef CONFIG_SMP
+			seq_printf(m, "\n  vm stats threshold: %d",
+					pageset->stat_threshold);
 #endif
 		}
 		seq_printf(m,
@@ -368,19 +610,34 @@ struct seq_operations zoneinfo_op = {
 
 static void *vmstat_start(struct seq_file *m, loff_t *pos)
 {
-	struct page_state *ps;
+	unsigned long *v;
+#ifdef CONFIG_VM_EVENT_COUNTERS
+	unsigned long *e;
+#endif
+	int i;
 
 	if (*pos >= ARRAY_SIZE(vmstat_text))
 		return NULL;
 
-	ps = kmalloc(sizeof(*ps), GFP_KERNEL);
-	m->private = ps;
-	if (!ps)
+#ifdef CONFIG_VM_EVENT_COUNTERS
+	v = kmalloc(NR_VM_ZONE_STAT_ITEMS * sizeof(unsigned long)
+			+ sizeof(struct vm_event_state), GFP_KERNEL);
+#else
+	v = kmalloc(NR_VM_ZONE_STAT_ITEMS * sizeof(unsigned long),
+			GFP_KERNEL);
+#endif
+	m->private = v;
+	if (!v)
 		return ERR_PTR(-ENOMEM);
-	get_full_page_state(ps);
-	ps->pgpgin /= 2;		/* sectors -> kbytes */
-	ps->pgpgout /= 2;
-	return (unsigned long *)ps + *pos;
+	for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
+		v[i] = global_page_state(i);
+#ifdef CONFIG_VM_EVENT_COUNTERS
+	e = v + NR_VM_ZONE_STAT_ITEMS;
+	all_vm_events(e);
+	e[PGPGIN] /= 2;		/* sectors -> kbytes */
+	e[PGPGOUT] /= 2;
+#endif
+	return v + *pos;
 }
 
 static void *vmstat_next(struct seq_file *m, void *arg, loff_t *pos)
@@ -415,3 +672,35 @@ struct seq_operations vmstat_op = {
 
 #endif /* CONFIG_PROC_FS */
 
+#ifdef CONFIG_SMP
+/*
+ * Use the cpu notifier to insure that the thresholds are recalculated
+ * when necessary.
+ */
+static int __cpuinit vmstat_cpuup_callback(struct notifier_block *nfb,
+		unsigned long action,
+		void *hcpu)
+{
+	switch (action) {
+		case CPU_UP_PREPARE:
+		case CPU_UP_CANCELED:
+		case CPU_DEAD:
+			refresh_zone_stat_thresholds();
+			break;
+		default:
+			break;
+	}
+	return NOTIFY_OK;
+}
+
+static struct notifier_block __cpuinitdata vmstat_notifier =
+	{ &vmstat_cpuup_callback, NULL, 0 };
+
+int __init setup_vmstat(void)
+{
+	refresh_zone_stat_thresholds();
+	register_cpu_notifier(&vmstat_notifier);
+	return 0;
+}
+module_init(setup_vmstat)
+#endif