* [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-10 13:31 ` Vlastimil Babka
2016-06-13 17:26 ` Christoph Lameter
2016-06-09 18:04 ` [PATCH 02/27] mm, vmscan: Move lru_lock to the node Mel Gorman
` (25 subsequent siblings)
26 siblings, 2 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
References: bnc#969297 PM performance -- intel_pstate
Patch-mainline: No, expected 4.7 and queued in linux-mm
Patch-name: patches.suse/mm-vmstat-Add-infrastructure-for-per-node-vmstats.patch
VM statistic counters for reclaim decisions are zone-based. If the kernel
is to reclaim on a per-node basis then we need to track per-node statistics
but there is no infrastructure for that. The most notable change is that
the old node_page_state is renamed to sum_zone_node_page_state. The new
node_page_state takes a pglist_data and uses per-node stats but none exist
yet. There is some renaming such as vm_stat to vm_zone_stat and the addition
of vm_node_stat and the renaming of mod_state to mod_zone_state. Otherwise,
this is mostly a mechanical patch with no functional change. There is a
lot of similarity between the node and zone helpers which is unfortunate
but there was no obvious way of reusing the code and maintaining type safety.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Mel Gorman <mgorman@suse.com>
---
drivers/base/node.c | 72 +++++++------
include/linux/mmzone.h | 13 +++
include/linux/vmstat.h | 92 +++++++++++++---
mm/page_alloc.c | 10 +-
mm/vmstat.c | 282 +++++++++++++++++++++++++++++++++++++++++++++----
mm/workingset.c | 9 +-
6 files changed, 404 insertions(+), 74 deletions(-)
diff --git a/drivers/base/node.c b/drivers/base/node.c
index 560751bad294..efb81da250a8 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -74,16 +74,16 @@ static ssize_t node_read_meminfo(struct device *dev,
nid, K(i.totalram),
nid, K(i.freeram),
nid, K(i.totalram - i.freeram),
- nid, K(node_page_state(nid, NR_ACTIVE_ANON) +
- node_page_state(nid, NR_ACTIVE_FILE)),
- nid, K(node_page_state(nid, NR_INACTIVE_ANON) +
- node_page_state(nid, NR_INACTIVE_FILE)),
- nid, K(node_page_state(nid, NR_ACTIVE_ANON)),
- nid, K(node_page_state(nid, NR_INACTIVE_ANON)),
- nid, K(node_page_state(nid, NR_ACTIVE_FILE)),
- nid, K(node_page_state(nid, NR_INACTIVE_FILE)),
- nid, K(node_page_state(nid, NR_UNEVICTABLE)),
- nid, K(node_page_state(nid, NR_MLOCK)));
+ nid, K(sum_zone_node_page_state(nid, NR_ACTIVE_ANON) +
+ sum_zone_node_page_state(nid, NR_ACTIVE_FILE)),
+ nid, K(sum_zone_node_page_state(nid, NR_INACTIVE_ANON) +
+ sum_zone_node_page_state(nid, NR_INACTIVE_FILE)),
+ nid, K(sum_zone_node_page_state(nid, NR_ACTIVE_ANON)),
+ nid, K(sum_zone_node_page_state(nid, NR_INACTIVE_ANON)),
+ nid, K(sum_zone_node_page_state(nid, NR_ACTIVE_FILE)),
+ nid, K(sum_zone_node_page_state(nid, NR_INACTIVE_FILE)),
+ nid, K(sum_zone_node_page_state(nid, NR_UNEVICTABLE)),
+ nid, K(sum_zone_node_page_state(nid, NR_MLOCK)));
#ifdef CONFIG_HIGHMEM
n += sprintf(buf + n,
@@ -115,28 +115,28 @@ static ssize_t node_read_meminfo(struct device *dev,
"Node %d AnonHugePages: %8lu kB\n"
#endif
,
- nid, K(node_page_state(nid, NR_FILE_DIRTY)),
- nid, K(node_page_state(nid, NR_WRITEBACK)),
- nid, K(node_page_state(nid, NR_FILE_PAGES)),
- nid, K(node_page_state(nid, NR_FILE_MAPPED)),
- nid, K(node_page_state(nid, NR_ANON_PAGES)),
+ nid, K(sum_zone_node_page_state(nid, NR_FILE_DIRTY)),
+ nid, K(sum_zone_node_page_state(nid, NR_WRITEBACK)),
+ nid, K(sum_zone_node_page_state(nid, NR_FILE_PAGES)),
+ nid, K(sum_zone_node_page_state(nid, NR_FILE_MAPPED)),
+ nid, K(sum_zone_node_page_state(nid, NR_ANON_PAGES)),
nid, K(i.sharedram),
- nid, node_page_state(nid, NR_KERNEL_STACK) *
+ nid, sum_zone_node_page_state(nid, NR_KERNEL_STACK) *
THREAD_SIZE / 1024,
- nid, K(node_page_state(nid, NR_PAGETABLE)),
- nid, K(node_page_state(nid, NR_UNSTABLE_NFS)),
- nid, K(node_page_state(nid, NR_BOUNCE)),
- nid, K(node_page_state(nid, NR_WRITEBACK_TEMP)),
- nid, K(node_page_state(nid, NR_SLAB_RECLAIMABLE) +
- node_page_state(nid, NR_SLAB_UNRECLAIMABLE)),
- nid, K(node_page_state(nid, NR_SLAB_RECLAIMABLE)),
+ nid, K(sum_zone_node_page_state(nid, NR_PAGETABLE)),
+ nid, K(sum_zone_node_page_state(nid, NR_UNSTABLE_NFS)),
+ nid, K(sum_zone_node_page_state(nid, NR_BOUNCE)),
+ nid, K(sum_zone_node_page_state(nid, NR_WRITEBACK_TEMP)),
+ nid, K(sum_zone_node_page_state(nid, NR_SLAB_RECLAIMABLE) +
+ sum_zone_node_page_state(nid, NR_SLAB_UNRECLAIMABLE)),
+ nid, K(sum_zone_node_page_state(nid, NR_SLAB_RECLAIMABLE)),
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
- nid, K(node_page_state(nid, NR_SLAB_UNRECLAIMABLE))
+ nid, K(sum_zone_node_page_state(nid, NR_SLAB_UNRECLAIMABLE))
, nid,
- K(node_page_state(nid, NR_ANON_TRANSPARENT_HUGEPAGES) *
+ K(sum_zone_node_page_state(nid, NR_ANON_TRANSPARENT_HUGEPAGES) *
HPAGE_PMD_NR));
#else
- nid, K(node_page_state(nid, NR_SLAB_UNRECLAIMABLE)));
+ nid, K(sum_zone_node_page_state(nid, NR_SLAB_UNRECLAIMABLE)));
#endif
n += hugetlb_report_node_meminfo(nid, buf + n);
return n;
@@ -155,12 +155,12 @@ static ssize_t node_read_numastat(struct device *dev,
"interleave_hit %lu\n"
"local_node %lu\n"
"other_node %lu\n",
- node_page_state(dev->id, NUMA_HIT),
- node_page_state(dev->id, NUMA_MISS),
- node_page_state(dev->id, NUMA_FOREIGN),
- node_page_state(dev->id, NUMA_INTERLEAVE_HIT),
- node_page_state(dev->id, NUMA_LOCAL),
- node_page_state(dev->id, NUMA_OTHER));
+ sum_zone_node_page_state(dev->id, NUMA_HIT),
+ sum_zone_node_page_state(dev->id, NUMA_MISS),
+ sum_zone_node_page_state(dev->id, NUMA_FOREIGN),
+ sum_zone_node_page_state(dev->id, NUMA_INTERLEAVE_HIT),
+ sum_zone_node_page_state(dev->id, NUMA_LOCAL),
+ sum_zone_node_page_state(dev->id, NUMA_OTHER));
}
static DEVICE_ATTR(numastat, S_IRUGO, node_read_numastat, NULL);
@@ -168,12 +168,18 @@ static ssize_t node_read_vmstat(struct device *dev,
struct device_attribute *attr, char *buf)
{
int nid = dev->id;
+ struct pglist_data *pgdat = NODE_DATA(nid);
int i;
int n = 0;
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
n += sprintf(buf+n, "%s %lu\n", vmstat_text[i],
- node_page_state(nid, i));
+ sum_zone_node_page_state(nid, i));
+
+ for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
+ n += sprintf(buf+n, "%s %lu\n",
+ vmstat_text[i + NR_VM_ZONE_STAT_ITEMS],
+ node_page_state(pgdat, i));
return n;
}
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 54df45c03ba2..b299c3af798e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -158,6 +158,10 @@ enum zone_stat_item {
NR_FREE_CMA_PAGES,
NR_VM_ZONE_STAT_ITEMS };
+enum node_stat_item {
+ NR_VM_NODE_STAT_ITEMS
+};
+
/*
* We do arithmetic on the LRU lists in various places in the code,
* so it is important to keep the active lists LRU_ACTIVE higher in
@@ -265,6 +269,11 @@ struct per_cpu_pageset {
#endif
};
+struct per_cpu_nodestat {
+ s8 stat_threshold;
+ s8 vm_node_stat_diff[NR_VM_NODE_STAT_ITEMS];
+};
+
#endif /* !__GENERATING_BOUNDS.H */
enum zone_type {
@@ -693,6 +702,10 @@ typedef struct pglist_data {
struct list_head split_queue;
unsigned long split_queue_len;
#endif
+
+ /* Per-node vmstats */
+ struct per_cpu_nodestat __percpu *per_cpu_nodestats;
+ atomic_long_t vm_stat[NR_VM_NODE_STAT_ITEMS];
} pg_data_t;
#define node_present_pages(nid) (NODE_DATA(nid)->node_present_pages)
diff --git a/include/linux/vmstat.h b/include/linux/vmstat.h
index 0aa613df463e..40629576f8de 100644
--- a/include/linux/vmstat.h
+++ b/include/linux/vmstat.h
@@ -112,20 +112,38 @@ static inline void vm_events_fold_cpu(int cpu)
zone_idx(zone), delta)
/*
- * Zone based page accounting with per cpu differentials.
+ * Zone and node-based page accounting with per cpu differentials.
*/
-extern atomic_long_t vm_stat[NR_VM_ZONE_STAT_ITEMS];
+extern atomic_long_t vm_zone_stat[NR_VM_ZONE_STAT_ITEMS];
+extern atomic_long_t vm_node_stat[NR_VM_NODE_STAT_ITEMS];
static inline void zone_page_state_add(long x, struct zone *zone,
enum zone_stat_item item)
{
atomic_long_add(x, &zone->vm_stat[item]);
- atomic_long_add(x, &vm_stat[item]);
+ atomic_long_add(x, &vm_zone_stat[item]);
+}
+
+static inline void node_page_state_add(long x, struct pglist_data *pgdat,
+ enum node_stat_item item)
+{
+ atomic_long_add(x, &pgdat->vm_stat[item]);
+ atomic_long_add(x, &vm_node_stat[item]);
}
static inline unsigned long global_page_state(enum zone_stat_item item)
{
- long x = atomic_long_read(&vm_stat[item]);
+ long x = atomic_long_read(&vm_zone_stat[item]);
+#ifdef CONFIG_SMP
+ if (x < 0)
+ x = 0;
+#endif
+ return x;
+}
+
+static inline unsigned long global_node_page_state(enum node_stat_item item)
+{
+ long x = atomic_long_read(&vm_node_stat[item]);
#ifdef CONFIG_SMP
if (x < 0)
x = 0;
@@ -167,31 +185,44 @@ static inline unsigned long zone_page_state_snapshot(struct zone *zone,
}
#ifdef CONFIG_NUMA
-
-extern unsigned long node_page_state(int node, enum zone_stat_item item);
-
+extern unsigned long sum_zone_node_page_state(int node,
+ enum zone_stat_item item);
+extern unsigned long node_page_state(struct pglist_data *pgdat,
+ enum node_stat_item item);
#else
-
-#define node_page_state(node, item) global_page_state(item)
-
+#define sum_zone_node_page_state(node, item) global_node_page_state(item)
+#define node_page_state(node, item) global_node_page_state(item)
#endif /* CONFIG_NUMA */
#define add_zone_page_state(__z, __i, __d) mod_zone_page_state(__z, __i, __d)
#define sub_zone_page_state(__z, __i, __d) mod_zone_page_state(__z, __i, -(__d))
+#define add_node_page_state(__p, __i, __d) mod_node_page_state(__p, __i, __d)
+#define sub_node_page_state(__p, __i, __d) mod_node_page_state(__p, __i, -(__d))
#ifdef CONFIG_SMP
void __mod_zone_page_state(struct zone *, enum zone_stat_item item, long);
void __inc_zone_page_state(struct page *, enum zone_stat_item);
void __dec_zone_page_state(struct page *, enum zone_stat_item);
+void __mod_node_page_state(struct pglist_data *, enum node_stat_item item, long);
+void __inc_node_page_state(struct page *, enum node_stat_item);
+void __dec_node_page_state(struct page *, enum node_stat_item);
+
void mod_zone_page_state(struct zone *, enum zone_stat_item, long);
void inc_zone_page_state(struct page *, enum zone_stat_item);
void dec_zone_page_state(struct page *, enum zone_stat_item);
+void mod_node_page_state(struct pglist_data *, enum node_stat_item, long);
+void inc_node_page_state(struct page *, enum node_stat_item);
+void dec_node_page_state(struct page *, enum node_stat_item);
+
extern void inc_zone_state(struct zone *, enum zone_stat_item);
+extern void inc_node_state(struct pglist_data *, enum node_stat_item);
extern void __inc_zone_state(struct zone *, enum zone_stat_item);
+extern void __inc_node_state(struct pglist_data *, enum node_stat_item);
extern void dec_zone_state(struct zone *, enum zone_stat_item);
extern void __dec_zone_state(struct zone *, enum zone_stat_item);
+extern void __dec_node_state(struct pglist_data *, enum node_stat_item);
void quiet_vmstat(void);
void cpu_vm_stats_fold(int cpu);
@@ -219,16 +250,34 @@ static inline void __mod_zone_page_state(struct zone *zone,
zone_page_state_add(delta, zone, item);
}
+static inline void __mod_node_page_state(struct pglist_data *pgdat,
+ enum node_stat_item item, int delta)
+{
+ node_page_state_add(delta, pgdat, item);
+}
+
static inline void __inc_zone_state(struct zone *zone, enum zone_stat_item item)
{
atomic_long_inc(&zone->vm_stat[item]);
- atomic_long_inc(&vm_stat[item]);
+ atomic_long_inc(&vm_zone_stat[item]);
+}
+
+static inline void __inc_node_state(struct pglist_data *pgdat, enum node_stat_item item)
+{
+ atomic_long_inc(&pgdat->vm_stat[item]);
+ atomic_long_inc(&vm_node_stat[item]);
}
static inline void __dec_zone_state(struct zone *zone, enum zone_stat_item item)
{
atomic_long_dec(&zone->vm_stat[item]);
- atomic_long_dec(&vm_stat[item]);
+ atomic_long_dec(&vm_zone_stat[item]);
+}
+
+static inline void __dec_node_state(struct pglist_data *pgdat, enum node_stat_item item)
+{
+ atomic_long_dec(&pgdat->vm_stat[item]);
+ atomic_long_dec(&vm_node_stat[item]);
}
static inline void __inc_zone_page_state(struct page *page,
@@ -237,12 +286,26 @@ static inline void __inc_zone_page_state(struct page *page,
__inc_zone_state(page_zone(page), item);
}
+static inline void __inc_node_page_state(struct page *page,
+ enum node_stat_item item)
+{
+ __inc_node_state(page_zone(page)->zone_pgdat, item);
+}
+
+
static inline void __dec_zone_page_state(struct page *page,
enum zone_stat_item item)
{
__dec_zone_state(page_zone(page), item);
}
+static inline void __dec_node_page_state(struct page *page,
+ enum node_stat_item item)
+{
+ __dec_node_state(page_zone(page)->zone_pgdat, item);
+}
+
+
/*
* We only use atomic operations to update counters. So there is no need to
* disable interrupts.
@@ -251,7 +314,12 @@ static inline void __dec_zone_page_state(struct page *page,
#define dec_zone_page_state __dec_zone_page_state
#define mod_zone_page_state __mod_zone_page_state
+#define inc_node_page_state __inc_node_page_state
+#define dec_node_page_state __dec_node_page_state
+#define mod_node_page_state __mod_node_page_state
+
#define inc_zone_state __inc_zone_state
+#define inc_node_state __inc_node_state
#define dec_zone_state __dec_zone_state
#define set_pgdat_percpu_threshold(pgdat, callback) { }
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index a46547389e53..9d71af25acf9 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -4209,8 +4209,8 @@ void si_meminfo_node(struct sysinfo *val, int nid)
for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++)
managed_pages += pgdat->node_zones[zone_type].managed_pages;
val->totalram = managed_pages;
- val->sharedram = node_page_state(nid, NR_SHMEM);
- val->freeram = node_page_state(nid, NR_FREE_PAGES);
+ val->sharedram = sum_zone_node_page_state(nid, NR_SHMEM);
+ val->freeram = sum_zone_node_page_state(nid, NR_FREE_PAGES);
#ifdef CONFIG_HIGHMEM
for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++) {
struct zone *zone = &pgdat->node_zones[zone_type];
@@ -5316,6 +5316,11 @@ static void __meminit setup_zone_pageset(struct zone *zone)
zone->pageset = alloc_percpu(struct per_cpu_pageset);
for_each_possible_cpu(cpu)
zone_pageset_init(zone, cpu);
+
+ if (!zone->zone_pgdat->per_cpu_nodestats) {
+ zone->zone_pgdat->per_cpu_nodestats =
+ alloc_percpu(struct per_cpu_nodestat);
+ }
}
/*
@@ -6021,6 +6026,7 @@ void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
reset_deferred_meminit(pgdat);
pgdat->node_id = nid;
pgdat->node_start_pfn = node_start_pfn;
+ pgdat->per_cpu_nodestats = NULL;
#ifdef CONFIG_HAVE_MEMBLOCK_NODE_MAP
get_pfn_range_for_nid(nid, &start_pfn, &end_pfn);
pr_info("Initmem setup node %d [mem %#018Lx-%#018Lx]\n", nid,
diff --git a/mm/vmstat.c b/mm/vmstat.c
index 076c39e3ba09..e1d0deeec98e 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -98,8 +98,10 @@ void vm_events_fold_cpu(int cpu)
*
* vm_stat contains the global counters
*/
-atomic_long_t vm_stat[NR_VM_ZONE_STAT_ITEMS] __cacheline_aligned_in_smp;
-EXPORT_SYMBOL(vm_stat);
+atomic_long_t vm_zone_stat[NR_VM_ZONE_STAT_ITEMS] __cacheline_aligned_in_smp;
+atomic_long_t vm_node_stat[NR_VM_NODE_STAT_ITEMS] __cacheline_aligned_in_smp;
+EXPORT_SYMBOL(vm_zone_stat);
+EXPORT_SYMBOL(vm_node_stat);
#ifdef CONFIG_SMP
@@ -188,9 +190,13 @@ void refresh_zone_stat_thresholds(void)
threshold = calculate_normal_threshold(zone);
- for_each_online_cpu(cpu)
+ for_each_online_cpu(cpu) {
+ struct pglist_data *pgdat = zone->zone_pgdat;
per_cpu_ptr(zone->pageset, cpu)->stat_threshold
= threshold;
+ per_cpu_ptr(pgdat->per_cpu_nodestats, cpu)->stat_threshold
+ = threshold;
+ }
/*
* Only set percpu_drift_mark if there is a danger that
@@ -250,6 +256,26 @@ void __mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
}
EXPORT_SYMBOL(__mod_zone_page_state);
+void __mod_node_page_state(struct pglist_data *pgdat, enum node_stat_item item,
+ long delta)
+{
+ struct per_cpu_nodestat __percpu *pcp = pgdat->per_cpu_nodestats;
+ s8 __percpu *p = pcp->vm_node_stat_diff + item;
+ long x;
+ long t;
+
+ x = delta + __this_cpu_read(*p);
+
+ t = __this_cpu_read(pcp->stat_threshold);
+
+ if (unlikely(x > t || x < -t)) {
+ node_page_state_add(x, pgdat, item);
+ x = 0;
+ }
+ __this_cpu_write(*p, x);
+}
+EXPORT_SYMBOL(__mod_node_page_state);
+
/*
* Optimized increment and decrement functions.
*
@@ -289,12 +315,34 @@ void __inc_zone_state(struct zone *zone, enum zone_stat_item item)
}
}
+void __inc_node_state(struct pglist_data *pgdat, enum node_stat_item item)
+{
+ struct per_cpu_nodestat __percpu *pcp = pgdat->per_cpu_nodestats;
+ s8 __percpu *p = pcp->vm_node_stat_diff + item;
+ s8 v, t;
+
+ v = __this_cpu_inc_return(*p);
+ t = __this_cpu_read(pcp->stat_threshold);
+ if (unlikely(v > t)) {
+ s8 overstep = t >> 1;
+
+ node_page_state_add(v + overstep, pgdat, item);
+ __this_cpu_write(*p, -overstep);
+ }
+}
+
void __inc_zone_page_state(struct page *page, enum zone_stat_item item)
{
__inc_zone_state(page_zone(page), item);
}
EXPORT_SYMBOL(__inc_zone_page_state);
+void __inc_node_page_state(struct page *page, enum node_stat_item item)
+{
+ __inc_node_state(page_zone(page)->zone_pgdat, item);
+}
+EXPORT_SYMBOL(__inc_node_page_state);
+
void __dec_zone_state(struct zone *zone, enum zone_stat_item item)
{
struct per_cpu_pageset __percpu *pcp = zone->pageset;
@@ -311,12 +359,34 @@ void __dec_zone_state(struct zone *zone, enum zone_stat_item item)
}
}
+void __dec_node_state(struct pglist_data *pgdat, enum node_stat_item item)
+{
+ struct per_cpu_nodestat __percpu *pcp = pgdat->per_cpu_nodestats;
+ s8 __percpu *p = pcp->vm_node_stat_diff + item;
+ s8 v, t;
+
+ v = __this_cpu_dec_return(*p);
+ t = __this_cpu_read(pcp->stat_threshold);
+ if (unlikely(v < - t)) {
+ s8 overstep = t >> 1;
+
+ node_page_state_add(v - overstep, pgdat, item);
+ __this_cpu_write(*p, overstep);
+ }
+}
+
void __dec_zone_page_state(struct page *page, enum zone_stat_item item)
{
__dec_zone_state(page_zone(page), item);
}
EXPORT_SYMBOL(__dec_zone_page_state);
+void __dec_node_page_state(struct page *page, enum node_stat_item item)
+{
+ __dec_node_state(page_zone(page)->zone_pgdat, item);
+}
+EXPORT_SYMBOL(__dec_node_page_state);
+
#ifdef CONFIG_HAVE_CMPXCHG_LOCAL
/*
* If we have cmpxchg_local support then we do not need to incur the overhead
@@ -330,8 +400,8 @@ EXPORT_SYMBOL(__dec_zone_page_state);
* 1 Overstepping half of threshold
* -1 Overstepping minus half of threshold
*/
-static inline void mod_state(struct zone *zone, enum zone_stat_item item,
- long delta, int overstep_mode)
+static inline void mod_zone_state(struct zone *zone,
+ enum zone_stat_item item, long delta, int overstep_mode)
{
struct per_cpu_pageset __percpu *pcp = zone->pageset;
s8 __percpu *p = pcp->vm_stat_diff + item;
@@ -371,26 +441,88 @@ static inline void mod_state(struct zone *zone, enum zone_stat_item item,
void mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
long delta)
{
- mod_state(zone, item, delta, 0);
+ mod_zone_state(zone, item, delta, 0);
}
EXPORT_SYMBOL(mod_zone_page_state);
void inc_zone_state(struct zone *zone, enum zone_stat_item item)
{
- mod_state(zone, item, 1, 1);
+ mod_zone_state(zone, item, 1, 1);
}
void inc_zone_page_state(struct page *page, enum zone_stat_item item)
{
- mod_state(page_zone(page), item, 1, 1);
+ mod_zone_state(page_zone(page), item, 1, 1);
}
EXPORT_SYMBOL(inc_zone_page_state);
void dec_zone_page_state(struct page *page, enum zone_stat_item item)
{
- mod_state(page_zone(page), item, -1, -1);
+ mod_zone_state(page_zone(page), item, -1, -1);
}
EXPORT_SYMBOL(dec_zone_page_state);
+
+static inline void mod_node_state(struct pglist_data *pgdat,
+ enum node_stat_item item, int delta, int overstep_mode)
+{
+ struct per_cpu_nodestat __percpu *pcp = pgdat->per_cpu_nodestats;
+ s8 __percpu *p = pcp->vm_node_stat_diff + item;
+ long o, n, t, z;
+
+ do {
+ z = 0; /* overflow to zone counters */
+
+ /*
+ * The fetching of the stat_threshold is racy. We may apply
+ * a counter threshold to the wrong the cpu if we get
+ * rescheduled while executing here. However, the next
+ * counter update will apply the threshold again and
+ * therefore bring the counter under the threshold again.
+ *
+ * Most of the time the thresholds are the same anyways
+ * for all cpus in a zone.
+ */
+ t = this_cpu_read(pcp->stat_threshold);
+
+ o = this_cpu_read(*p);
+ n = delta + o;
+
+ if (n > t || n < -t) {
+ int os = overstep_mode * (t >> 1) ;
+
+ /* Overflow must be added to zone counters */
+ z = n + os;
+ n = -os;
+ }
+ } while (this_cpu_cmpxchg(*p, o, n) != o);
+
+ if (z)
+ node_page_state_add(z, pgdat, item);
+}
+
+void mod_node_page_state(struct pglist_data *pgdat, enum node_stat_item item,
+ long delta)
+{
+ mod_node_state(pgdat, item, delta, 0);
+}
+EXPORT_SYMBOL(mod_node_page_state);
+
+void inc_node_state(struct pglist_data *pgdat, enum node_stat_item item)
+{
+ mod_node_state(pgdat, item, 1, 1);
+}
+
+void inc_node_page_state(struct page *page, enum node_stat_item item)
+{
+ mod_node_state(page_zone(page)->zone_pgdat, item, 1, 1);
+}
+EXPORT_SYMBOL(inc_node_page_state);
+
+void dec_node_page_state(struct page *page, enum node_stat_item item)
+{
+ mod_node_state(page_zone(page)->zone_pgdat, item, -1, -1);
+}
+EXPORT_SYMBOL(dec_node_page_state);
#else
/*
* Use interrupt disable to serialize counter updates
@@ -436,21 +568,69 @@ void dec_zone_page_state(struct page *page, enum zone_stat_item item)
local_irq_restore(flags);
}
EXPORT_SYMBOL(dec_zone_page_state);
-#endif
+void inc_node_state(struct pglist_data *pgdat, enum node_stat_item item)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+ __inc_node_state(pgdat, item);
+ local_irq_restore(flags);
+}
+EXPORT_SYMBOL(inc_node_state);
+
+void mod_node_page_state(struct pglist_data *pgdat, enum node_stat_item item,
+ long delta)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+ __mod_node_page_state(pgdat, item, delta);
+ local_irq_restore(flags);
+}
+EXPORT_SYMBOL(mod_node_page_state);
+
+void inc_node_page_state(struct page *page, enum node_stat_item item)
+{
+ unsigned long flags;
+ struct pglist_data *pgdat;
+
+ pgdat = page_zone(page)->zone_pgdat;
+ local_irq_save(flags);
+ __inc_node_state(pgdat, item);
+ local_irq_restore(flags);
+}
+EXPORT_SYMBOL(inc_node_page_state);
+
+void dec_node_page_state(struct page *page, enum node_stat_item item)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+ __dec_node_page_state(page, item);
+ local_irq_restore(flags);
+}
+EXPORT_SYMBOL(dec_node_page_state);
+#endif
/*
* Fold a differential into the global counters.
* Returns the number of counters updated.
*/
-static int fold_diff(int *diff)
+static int fold_diff(int *zone_diff, int *node_diff)
{
int i;
int changes = 0;
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
- if (diff[i]) {
- atomic_long_add(diff[i], &vm_stat[i]);
+ if (zone_diff[i]) {
+ atomic_long_add(zone_diff[i], &vm_zone_stat[i]);
+ changes++;
+ }
+
+ for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
+ if (node_diff[i]) {
+ atomic_long_add(node_diff[i], &vm_node_stat[i]);
changes++;
}
return changes;
@@ -474,9 +654,11 @@ static int fold_diff(int *diff)
*/
static int refresh_cpu_vm_stats(bool do_pagesets)
{
+ struct pglist_data *pgdat;
struct zone *zone;
int i;
- int global_diff[NR_VM_ZONE_STAT_ITEMS] = { 0, };
+ int global_zone_diff[NR_VM_ZONE_STAT_ITEMS] = { 0, };
+ int global_node_diff[NR_VM_NODE_STAT_ITEMS] = { 0, };
int changes = 0;
for_each_populated_zone(zone) {
@@ -489,7 +671,7 @@ static int refresh_cpu_vm_stats(bool do_pagesets)
if (v) {
atomic_long_add(v, &zone->vm_stat[i]);
- global_diff[i] += v;
+ global_zone_diff[i] += v;
#ifdef CONFIG_NUMA
/* 3 seconds idle till flush */
__this_cpu_write(p->expire, 3);
@@ -528,7 +710,22 @@ static int refresh_cpu_vm_stats(bool do_pagesets)
}
#endif
}
- changes += fold_diff(global_diff);
+
+ for_each_online_pgdat(pgdat) {
+ struct per_cpu_nodestat __percpu *p = pgdat->per_cpu_nodestats;
+
+ for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++) {
+ int v;
+
+ v = this_cpu_xchg(p->vm_node_stat_diff[i], 0);
+ if (v) {
+ atomic_long_add(v, &pgdat->vm_stat[i]);
+ global_node_diff[i] += v;
+ }
+ }
+ }
+
+ changes += fold_diff(global_zone_diff, global_node_diff);
return changes;
}
@@ -539,9 +736,11 @@ static int refresh_cpu_vm_stats(bool do_pagesets)
*/
void cpu_vm_stats_fold(int cpu)
{
+ struct pglist_data *pgdat;
struct zone *zone;
int i;
- int global_diff[NR_VM_ZONE_STAT_ITEMS] = { 0, };
+ int global_zone_diff[NR_VM_ZONE_STAT_ITEMS] = { 0, };
+ int global_node_diff[NR_VM_NODE_STAT_ITEMS] = { 0, };
for_each_populated_zone(zone) {
struct per_cpu_pageset *p;
@@ -555,11 +754,27 @@ void cpu_vm_stats_fold(int cpu)
v = p->vm_stat_diff[i];
p->vm_stat_diff[i] = 0;
atomic_long_add(v, &zone->vm_stat[i]);
- global_diff[i] += v;
+ global_zone_diff[i] += v;
}
}
- fold_diff(global_diff);
+ for_each_online_pgdat(pgdat) {
+ struct per_cpu_nodestat *p;
+
+ p = per_cpu_ptr(pgdat->per_cpu_nodestats, cpu);
+
+ for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
+ if (p->vm_node_stat_diff[i]) {
+ int v;
+
+ v = p->vm_node_stat_diff[i];
+ p->vm_node_stat_diff[i] = 0;
+ atomic_long_add(v, &pgdat->vm_stat[i]);
+ global_node_diff[i] += v;
+ }
+ }
+
+ fold_diff(global_zone_diff, global_node_diff);
}
/*
@@ -575,16 +790,19 @@ void drain_zonestat(struct zone *zone, struct per_cpu_pageset *pset)
int v = pset->vm_stat_diff[i];
pset->vm_stat_diff[i] = 0;
atomic_long_add(v, &zone->vm_stat[i]);
- atomic_long_add(v, &vm_stat[i]);
+ atomic_long_add(v, &vm_zone_stat[i]);
}
}
#endif
#ifdef CONFIG_NUMA
/*
- * Determine the per node value of a stat item.
+ * Determine the per node value of a stat item. This function
+ * is called frequently in a NUMA machine, so try to be as
+ * frugal as possible.
*/
-unsigned long node_page_state(int node, enum zone_stat_item item)
+unsigned long sum_zone_node_page_state(int node,
+ enum zone_stat_item item)
{
struct zone *zones = NODE_DATA(node)->node_zones;
int i;
@@ -596,6 +814,19 @@ unsigned long node_page_state(int node, enum zone_stat_item item)
return count;
}
+/*
+ * Determine the per node value of a stat item.
+ */
+unsigned long node_page_state(struct pglist_data *pgdat,
+ enum node_stat_item item)
+{
+ long x = atomic_long_read(&pgdat->vm_stat[item]);
+#ifdef CONFIG_SMP
+ if (x < 0)
+ x = 0;
+#endif
+ return x;
+}
#endif
#ifdef CONFIG_COMPACTION
@@ -1295,6 +1526,7 @@ static void *vmstat_start(struct seq_file *m, loff_t *pos)
if (*pos >= ARRAY_SIZE(vmstat_text))
return NULL;
stat_items_size = NR_VM_ZONE_STAT_ITEMS * sizeof(unsigned long) +
+ NR_VM_NODE_STAT_ITEMS * sizeof(unsigned long) +
NR_VM_WRITEBACK_STAT_ITEMS * sizeof(unsigned long);
#ifdef CONFIG_VM_EVENT_COUNTERS
@@ -1309,6 +1541,10 @@ static void *vmstat_start(struct seq_file *m, loff_t *pos)
v[i] = global_page_state(i);
v += NR_VM_ZONE_STAT_ITEMS;
+ for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
+ v[i] = global_node_page_state(i);
+ v += NR_VM_NODE_STAT_ITEMS;
+
global_dirty_limits(v + NR_DIRTY_BG_THRESHOLD,
v + NR_DIRTY_THRESHOLD);
v += NR_VM_WRITEBACK_STAT_ITEMS;
@@ -1398,7 +1634,7 @@ int vmstat_refresh(struct ctl_table *table, int write,
if (err)
return err;
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++) {
- val = atomic_long_read(&vm_stat[i]);
+ val = atomic_long_read(&vm_zone_stat[i]);
if (val < 0) {
switch (i) {
case NR_ALLOC_BATCH:
diff --git a/mm/workingset.c b/mm/workingset.c
index 8a75f8d2916a..ac36efa8c754 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -349,12 +349,13 @@ static unsigned long count_shadow_nodes(struct shrinker *shrinker,
shadow_nodes = list_lru_shrink_count(&workingset_shadow_nodes, sc);
local_irq_enable();
- if (memcg_kmem_enabled())
+ if (memcg_kmem_enabled()) {
pages = mem_cgroup_node_nr_lru_pages(sc->memcg, sc->nid,
LRU_ALL_FILE);
- else
- pages = node_page_state(sc->nid, NR_ACTIVE_FILE) +
- node_page_state(sc->nid, NR_INACTIVE_FILE);
+ } else {
+ pages = sum_zone_node_page_state(sc->nid, NR_ACTIVE_FILE) +
+ sum_zone_node_page_state(sc->nid, NR_INACTIVE_FILE);
+ }
/*
* Active cache pages are limited to 50% of memory, and shadow
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats
2016-06-09 18:04 ` [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats Mel Gorman
@ 2016-06-10 13:31 ` Vlastimil Babka
2016-06-10 13:47 ` Mel Gorman
2016-06-13 17:26 ` Christoph Lameter
1 sibling, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-10 13:31 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> References: bnc#969297 PM performance -- intel_pstate
> Patch-mainline: No, expected 4.7 and queued in linux-mm
> Patch-name: patches.suse/mm-vmstat-Add-infrastructure-for-per-node-vmstats.patch
Remove?
> VM statistic counters for reclaim decisions are zone-based. If the kernel
> is to reclaim on a per-node basis then we need to track per-node statistics
> but there is no infrastructure for that. The most notable change is that
> the old node_page_state is renamed to sum_zone_node_page_state. The new
> node_page_state takes a pglist_data and uses per-node stats but none exist
> yet. There is some renaming such as vm_stat to vm_zone_stat and the addition
> of vm_node_stat and the renaming of mod_state to mod_zone_state. Otherwise,
> this is mostly a mechanical patch with no functional change. There is a
> lot of similarity between the node and zone helpers which is unfortunate
> but there was no obvious way of reusing the code and maintaining type safety.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> Acked-by: Johannes Weiner <hannes@cmpxchg.org>
> Signed-off-by: Mel Gorman <mgorman@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Some nitpicks below.
> @@ -237,12 +286,26 @@ static inline void __inc_zone_page_state(struct page *page,
> __inc_zone_state(page_zone(page), item);
> }
>
> +static inline void __inc_node_page_state(struct page *page,
> + enum node_stat_item item)
> +{
> + __inc_node_state(page_zone(page)->zone_pgdat, item);
This page -> node translation looks needlessly ineffective. How about
using NODE_DATA(page_to_nid(page)).
> +}
> +
> +
> static inline void __dec_zone_page_state(struct page *page,
> enum zone_stat_item item)
> {
> __dec_zone_state(page_zone(page), item);
> }
>
> +static inline void __dec_node_page_state(struct page *page,
> + enum node_stat_item item)
> +{
> + __dec_node_state(page_zone(page)->zone_pgdat, item);
> +}
Ditto.
> @@ -188,9 +190,13 @@ void refresh_zone_stat_thresholds(void)
>
> threshold = calculate_normal_threshold(zone);
>
> - for_each_online_cpu(cpu)
> + for_each_online_cpu(cpu) {
> + struct pglist_data *pgdat = zone->zone_pgdat;
Move the variable outside?
> per_cpu_ptr(zone->pageset, cpu)->stat_threshold
> = threshold;
> + per_cpu_ptr(pgdat->per_cpu_nodestats, cpu)->stat_threshold
> + = threshold;
> + }
>
> /*
> * Only set percpu_drift_mark if there is a danger that
> void __inc_zone_page_state(struct page *page, enum zone_stat_item item)
[...]
> {
> __inc_zone_state(page_zone(page), item);
> }
> EXPORT_SYMBOL(__inc_zone_page_state);
>
> +void __inc_node_page_state(struct page *page, enum node_stat_item item)
> +{
> + __inc_node_state(page_zone(page)->zone_pgdat, item);
Same page -> node thing here.
> +void __dec_node_page_state(struct page *page, enum node_stat_item item)
> +{
> + __dec_node_state(page_zone(page)->zone_pgdat, item);
And here.
>
> void dec_zone_page_state(struct page *page, enum zone_stat_item item)
> {
> - mod_state(page_zone(page), item, -1, -1);
> + mod_zone_state(page_zone(page), item, -1, -1);
> }
> EXPORT_SYMBOL(dec_zone_page_state);
> +
> +static inline void mod_node_state(struct pglist_data *pgdat,
> + enum node_stat_item item, int delta, int overstep_mode)
> +{
> + struct per_cpu_nodestat __percpu *pcp = pgdat->per_cpu_nodestats;
> + s8 __percpu *p = pcp->vm_node_stat_diff + item;
> + long o, n, t, z;
> +
> + do {
> + z = 0; /* overflow to zone counters */
s/zone/node/?
> +
> + /*
> + * The fetching of the stat_threshold is racy. We may apply
> + * a counter threshold to the wrong the cpu if we get
> + * rescheduled while executing here. However, the next
> + * counter update will apply the threshold again and
> + * therefore bring the counter under the threshold again.
> + *
> + * Most of the time the thresholds are the same anyways
> + * for all cpus in a zone.
same here.
> + */
> + t = this_cpu_read(pcp->stat_threshold);
> +
> + o = this_cpu_read(*p);
> + n = delta + o;
> +
> + if (n > t || n < -t) {
> + int os = overstep_mode * (t >> 1) ;
> +
> + /* Overflow must be added to zone counters */
and here.
> +}
> +
> +void inc_node_page_state(struct page *page, enum node_stat_item item)
> +{
> + mod_node_state(page_zone(page)->zone_pgdat, item, 1, 1);
Ditto about page -> nid.
> +}
> +EXPORT_SYMBOL(inc_node_page_state);
> +
> +void dec_node_page_state(struct page *page, enum node_stat_item item)
> +{
> + mod_node_state(page_zone(page)->zone_pgdat, item, -1, -1);
> +}
Ditto.
> +EXPORT_SYMBOL(dec_node_page_state);
> #else
> /*
> * Use interrupt disable to serialize counter updates
> @@ -436,21 +568,69 @@ void dec_zone_page_state(struct page *page, enum zone_stat_item item)
> local_irq_restore(flags);
> }
> EXPORT_SYMBOL(dec_zone_page_state);
> -#endif
>
> +void inc_node_state(struct pglist_data *pgdat, enum node_stat_item item)
> +{
> + unsigned long flags;
> +
> + local_irq_save(flags);
> + __inc_node_state(pgdat, item);
> + local_irq_restore(flags);
> +}
> +EXPORT_SYMBOL(inc_node_state);
> +
> +void mod_node_page_state(struct pglist_data *pgdat, enum node_stat_item item,
> + long delta)
> +{
> + unsigned long flags;
> +
> + local_irq_save(flags);
> + __mod_node_page_state(pgdat, item, delta);
> + local_irq_restore(flags);
> +}
> +EXPORT_SYMBOL(mod_node_page_state);
> +
> +void inc_node_page_state(struct page *page, enum node_stat_item item)
> +{
> + unsigned long flags;
> + struct pglist_data *pgdat;
> +
> + pgdat = page_zone(page)->zone_pgdat;
And here.
,9 +736,11 @@ static int refresh_cpu_vm_stats(bool do_pagesets)
> */
> void cpu_vm_stats_fold(int cpu)
> {
> + struct pglist_data *pgdat;
> struct zone *zone;
> int i;
> - int global_diff[NR_VM_ZONE_STAT_ITEMS] = { 0, };
> + int global_zone_diff[NR_VM_ZONE_STAT_ITEMS] = { 0, };
> + int global_node_diff[NR_VM_NODE_STAT_ITEMS] = { 0, };
>
> for_each_populated_zone(zone) {
> struct per_cpu_pageset *p;
> @@ -555,11 +754,27 @@ void cpu_vm_stats_fold(int cpu)
> v = p->vm_stat_diff[i];
> p->vm_stat_diff[i] = 0;
> atomic_long_add(v, &zone->vm_stat[i]);
> - global_diff[i] += v;
> + global_zone_diff[i] += v;
> }
> }
>
^ permalink raw reply [flat|nested] 70+ messages in thread* Re: [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats
2016-06-10 13:31 ` Vlastimil Babka
@ 2016-06-10 13:47 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-10 13:47 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Fri, Jun 10, 2016 at 03:31:41PM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> > References: bnc#969297 PM performance -- intel_pstate
> > Patch-mainline: No, expected 4.7 and queued in linux-mm
> > Patch-name: patches.suse/mm-vmstat-Add-infrastructure-for-per-node-vmstats.patch
>
> Remove?
>
Yes. Clearly I fat-fingers a cherry pick and used the wrong command that
added distro-specific metadata. Sorry.
> > VM statistic counters for reclaim decisions are zone-based. If the kernel
> > is to reclaim on a per-node basis then we need to track per-node statistics
> > but there is no infrastructure for that. The most notable change is that
> > the old node_page_state is renamed to sum_zone_node_page_state. The new
> > node_page_state takes a pglist_data and uses per-node stats but none exist
> > yet. There is some renaming such as vm_stat to vm_zone_stat and the addition
> > of vm_node_stat and the renaming of mod_state to mod_zone_state. Otherwise,
> > this is mostly a mechanical patch with no functional change. There is a
> > lot of similarity between the node and zone helpers which is unfortunate
> > but there was no obvious way of reusing the code and maintaining type safety.
> >
> > Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> > Acked-by: Johannes Weiner <hannes@cmpxchg.org>
> > Signed-off-by: Mel Gorman <mgorman@suse.com>
>
> Acked-by: Vlastimil Babka <vbabka@suse.cz>
>
> Some nitpicks below.
>
> > @@ -237,12 +286,26 @@ static inline void __inc_zone_page_state(struct page *page,
> > __inc_zone_state(page_zone(page), item);
> > }
> >
> > +static inline void __inc_node_page_state(struct page *page,
> > + enum node_stat_item item)
> > +{
> > + __inc_node_state(page_zone(page)->zone_pgdat, item);
>
> This page -> node translation looks needlessly ineffective. How about
> using NODE_DATA(page_to_nid(page)).
>
Yes, I will. I won't answer the individual feedbacks. They all seem
reasonable.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* Re: [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats
2016-06-09 18:04 ` [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats Mel Gorman
2016-06-10 13:31 ` Vlastimil Babka
@ 2016-06-13 17:26 ` Christoph Lameter
2016-06-14 14:25 ` Mel Gorman
1 sibling, 1 reply; 70+ messages in thread
From: Christoph Lameter @ 2016-06-13 17:26 UTC (permalink / raw)
To: Mel Gorman
Cc: Andrew Morton, Linux-MM, Rik van Riel, Vlastimil Babka,
Johannes Weiner, LKML
On Thu, 9 Jun 2016, Mel Gorman wrote:
> VM statistic counters for reclaim decisions are zone-based. If the kernel
> is to reclaim on a per-node basis then we need to track per-node statistics
> but there is no infrastructure for that. The most notable change is that
There is node_page_state() so the value of any counter per node is already
available. Note that some of the counters (NUMA_xx) for example do not
make much sense as per zone counters and are effectively used as per node
counters.
So the main effect you are looking for is to have the counters stored in
the per node structure as opposed to the per zone struct in order to
avoid the summing? Doing so duplicates a large amount of code it seems.
If you do this then also move over certain counters that have more of a
per node use from per zone to per node. Like the NUMA_xxx counters.
^ permalink raw reply [flat|nested] 70+ messages in thread
* Re: [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats
2016-06-13 17:26 ` Christoph Lameter
@ 2016-06-14 14:25 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-14 14:25 UTC (permalink / raw)
To: Christoph Lameter
Cc: Andrew Morton, Linux-MM, Rik van Riel, Vlastimil Babka,
Johannes Weiner, LKML
On Mon, Jun 13, 2016 at 12:26:13PM -0500, Christoph Lameter wrote:
> On Thu, 9 Jun 2016, Mel Gorman wrote:
>
> > VM statistic counters for reclaim decisions are zone-based. If the kernel
> > is to reclaim on a per-node basis then we need to track per-node statistics
> > but there is no infrastructure for that. The most notable change is that
>
> There is node_page_state() so the value of any counter per node is already
> available. Note that some of the counters (NUMA_xx) for example do not
> make much sense as per zone counters and are effectively used as per node
> counters.
>
> So the main effect you are looking for is to have the counters stored in
> the per node structure as opposed to the per zone struct in order to
> avoid the summing?
Yes.
> Doing so duplicates a large amount of code it seems.
>
Also yes. I considered macro magic to cover it but it turned into a
major mess. They could always be summed so it would be a minor
performance dent and a heavier cache footprint.
> If you do this then also move over certain counters that have more of a
> per node use from per zone to per node. Like the NUMA_xxx counters.
>
As the NUMA counters are consumed by userspace, I worried that it would
break some tools. If the rest of the series gets solidified then I will
do it as a single patch on top so it can be reverted if necessary
relatively easily.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 02/27] mm, vmscan: Move lru_lock to the node
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
2016-06-09 18:04 ` [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-10 16:39 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 03/27] mm, vmscan: Move LRU lists to node Mel Gorman
` (24 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Node-based reclaim requires node-based LRUs and locking. This is a
preparation patch that just moves the lru_lock to the node so later patches
are easier to review. It is a mechanical change but note this patch makes
contention worse because the LRU lock is hotter and direct reclaim and kswapd
can contend on the same lock even when reclaiming from different zones.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
---
Documentation/cgroup-v1/memcg_test.txt | 4 +--
Documentation/cgroup-v1/memory.txt | 4 +--
include/linux/mm_types.h | 2 +-
include/linux/mmzone.h | 10 +++++--
mm/compaction.c | 10 +++----
mm/filemap.c | 4 +--
mm/huge_memory.c | 4 +--
mm/memcontrol.c | 6 ++---
mm/mlock.c | 10 +++----
mm/page_alloc.c | 4 +--
mm/page_idle.c | 4 +--
mm/rmap.c | 2 +-
mm/swap.c | 30 ++++++++++-----------
mm/vmscan.c | 48 +++++++++++++++++-----------------
14 files changed, 74 insertions(+), 68 deletions(-)
diff --git a/Documentation/cgroup-v1/memcg_test.txt b/Documentation/cgroup-v1/memcg_test.txt
index 8870b0212150..78a8c2963b38 100644
--- a/Documentation/cgroup-v1/memcg_test.txt
+++ b/Documentation/cgroup-v1/memcg_test.txt
@@ -107,9 +107,9 @@ Under below explanation, we assume CONFIG_MEM_RES_CTRL_SWAP=y.
8. LRU
Each memcg has its own private LRU. Now, its handling is under global
- VM's control (means that it's handled under global zone->lru_lock).
+ VM's control (means that it's handled under global zone_lru_lock).
Almost all routines around memcg's LRU is called by global LRU's
- list management functions under zone->lru_lock().
+ list management functions under zone_lru_lock().
A special function is mem_cgroup_isolate_pages(). This scans
memcg's private LRU and call __isolate_lru_page() to extract a page
diff --git a/Documentation/cgroup-v1/memory.txt b/Documentation/cgroup-v1/memory.txt
index b14abf217239..946e69103cdd 100644
--- a/Documentation/cgroup-v1/memory.txt
+++ b/Documentation/cgroup-v1/memory.txt
@@ -267,11 +267,11 @@ When oom event notifier is registered, event will be delivered.
Other lock order is following:
PG_locked.
mm->page_table_lock
- zone->lru_lock
+ zone_lru_lock
lock_page_cgroup.
In many cases, just lock_page_cgroup() is called.
per-zone-per-cgroup LRU (cgroup's private LRU) is just guarded by
- zone->lru_lock, it has no lock of its own.
+ zone_lru_lock, it has no lock of its own.
2.7 Kernel Memory Extension (CONFIG_MEMCG_KMEM)
diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
index e093e1d3285b..ca2ed9a6c8d8 100644
--- a/include/linux/mm_types.h
+++ b/include/linux/mm_types.h
@@ -118,7 +118,7 @@ struct page {
*/
union {
struct list_head lru; /* Pageout list, eg. active_list
- * protected by zone->lru_lock !
+ * protected by zone_lru_lock !
* Can be used as a generic list
* by the page owner.
*/
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index b299c3af798e..70623b6ac833 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -93,7 +93,7 @@ struct free_area {
struct pglist_data;
/*
- * zone->lock and zone->lru_lock are two of the hottest locks in the kernel.
+ * zone->lock and the zone lru_lock are two of the hottest locks in the kernel.
* So add a wild amount of padding here to ensure that they fall into separate
* cachelines. There are very few zone structures in the machine, so space
* consumption is not a concern here.
@@ -494,7 +494,6 @@ struct zone {
/* Write-intensive fields used by page reclaim */
/* Fields commonly accessed by the page reclaim scanner */
- spinlock_t lru_lock;
struct lruvec lruvec;
/*
@@ -688,6 +687,9 @@ typedef struct pglist_data {
/* Number of pages migrated during the rate limiting time interval */
unsigned long numabalancing_migrate_nr_pages;
#endif
+ /* Write-intensive fields used from the page allocator */
+ ZONE_PADDING(_pad1_)
+ spinlock_t lru_lock;
#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
/*
@@ -719,6 +721,10 @@ typedef struct pglist_data {
#define node_start_pfn(nid) (NODE_DATA(nid)->node_start_pfn)
#define node_end_pfn(nid) pgdat_end_pfn(NODE_DATA(nid))
+static inline spinlock_t *zone_lru_lock(struct zone *zone)
+{
+ return &zone->zone_pgdat->lru_lock;
+}
static inline unsigned long pgdat_end_pfn(pg_data_t *pgdat)
{
diff --git a/mm/compaction.c b/mm/compaction.c
index fbb7b38bc0be..5aba9e6c1975 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -754,7 +754,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn,
* if contended.
*/
if (!(low_pfn % SWAP_CLUSTER_MAX)
- && compact_unlock_should_abort(&zone->lru_lock, flags,
+ && compact_unlock_should_abort(zone_lru_lock(zone), flags,
&locked, cc))
break;
@@ -816,7 +816,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn,
if (unlikely(__PageMovable(page)) &&
!PageIsolated(page)) {
if (locked) {
- spin_unlock_irqrestore(&zone->lru_lock,
+ spin_unlock_irqrestore(zone_lru_lock(zone),
flags);
locked = false;
}
@@ -839,7 +839,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn,
/* If we already hold the lock, we can skip some rechecking */
if (!locked) {
- locked = compact_trylock_irqsave(&zone->lru_lock,
+ locked = compact_trylock_irqsave(zone_lru_lock(zone),
&flags, cc);
if (!locked)
break;
@@ -902,7 +902,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn,
*/
if (nr_isolated) {
if (locked) {
- spin_unlock_irqrestore(&zone->lru_lock, flags);
+ spin_unlock_irqrestore(zone_lru_lock(zone), flags);
locked = false;
}
acct_isolated(zone, cc);
@@ -930,7 +930,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn,
low_pfn = end_pfn;
if (locked)
- spin_unlock_irqrestore(&zone->lru_lock, flags);
+ spin_unlock_irqrestore(zone_lru_lock(zone), flags);
/*
* Update the pageblock-skip information and cached scanner pfn,
diff --git a/mm/filemap.c b/mm/filemap.c
index 00ae878b2a38..bce091ed90d7 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -95,8 +95,8 @@
* ->swap_lock (try_to_unmap_one)
* ->private_lock (try_to_unmap_one)
* ->tree_lock (try_to_unmap_one)
- * ->zone.lru_lock (follow_page->mark_page_accessed)
- * ->zone.lru_lock (check_pte_range->isolate_lru_page)
+ * ->zone_lru_lock(zone) (follow_page->mark_page_accessed)
+ * ->zone_lru_lock(zone) (check_pte_range->isolate_lru_page)
* ->private_lock (page_remove_rmap->set_page_dirty)
* ->tree_lock (page_remove_rmap->set_page_dirty)
* bdi.wb->list_lock (page_remove_rmap->set_page_dirty)
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 55ad87c80e0a..8d535b778ead 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -3300,7 +3300,7 @@ static void __split_huge_page(struct page *page, struct list_head *list)
int i;
/* prevent PageLRU to go away from under us, and freeze lru stats */
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
lruvec = mem_cgroup_page_lruvec(head, zone);
/* complete memcg works before add pages to LRU */
@@ -3310,7 +3310,7 @@ static void __split_huge_page(struct page *page, struct list_head *list)
__split_huge_page_tail(head, i, lruvec, list);
ClearPageCompound(head);
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
unfreeze_page(head);
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 83a6a2b92301..605e8b5fc0db 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -2107,7 +2107,7 @@ static void lock_page_lru(struct page *page, int *isolated)
{
struct zone *zone = page_zone(page);
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
if (PageLRU(page)) {
struct lruvec *lruvec;
@@ -2131,7 +2131,7 @@ static void unlock_page_lru(struct page *page, int isolated)
SetPageLRU(page);
add_page_to_lru_list(page, lruvec, page_lru(page));
}
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
}
static void commit_charge(struct page *page, struct mem_cgroup *memcg,
@@ -2431,7 +2431,7 @@ void memcg_kmem_uncharge(struct page *page, int order)
/*
* Because tail pages are not marked as "used", set it. We're under
- * zone->lru_lock and migration entries setup in all page mappings.
+ * zone_lru_lock and migration entries setup in all page mappings.
*/
void mem_cgroup_split_huge_fixup(struct page *head)
{
diff --git a/mm/mlock.c b/mm/mlock.c
index ef8dc9f395c4..997f63082ff5 100644
--- a/mm/mlock.c
+++ b/mm/mlock.c
@@ -188,7 +188,7 @@ unsigned int munlock_vma_page(struct page *page)
* might otherwise copy PageMlocked to part of the tail pages before
* we clear it in the head page. It also stabilizes hpage_nr_pages().
*/
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
nr_pages = hpage_nr_pages(page);
if (!TestClearPageMlocked(page))
@@ -197,14 +197,14 @@ unsigned int munlock_vma_page(struct page *page)
__mod_zone_page_state(zone, NR_MLOCK, -nr_pages);
if (__munlock_isolate_lru_page(page, true)) {
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
__munlock_isolated_page(page);
goto out;
}
__munlock_isolation_failed(page);
unlock_out:
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
out:
return nr_pages - 1;
@@ -289,7 +289,7 @@ static void __munlock_pagevec(struct pagevec *pvec, struct zone *zone)
pagevec_init(&pvec_putback, 0);
/* Phase 1: page isolation */
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
for (i = 0; i < nr; i++) {
struct page *page = pvec->pages[i];
@@ -315,7 +315,7 @@ static void __munlock_pagevec(struct pagevec *pvec, struct zone *zone)
}
delta_munlocked = -nr + pagevec_count(&pvec_putback);
__mod_zone_page_state(zone, NR_MLOCK, delta_munlocked);
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
/* Now we can release pins of pages that we are not munlocking */
pagevec_release(&pvec_putback);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 9d71af25acf9..1e0ad06c33bd 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -5944,10 +5944,10 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
zone->min_slab_pages = (freesize * sysctl_min_slab_ratio) / 100;
#endif
zone->name = zone_names[j];
+ zone->zone_pgdat = pgdat;
spin_lock_init(&zone->lock);
- spin_lock_init(&zone->lru_lock);
+ spin_lock_init(zone_lru_lock(zone));
zone_seqlock_init(zone);
- zone->zone_pgdat = pgdat;
zone_pcp_init(zone);
/* For bootup, initialized properly in watermark setup */
diff --git a/mm/page_idle.c b/mm/page_idle.c
index 4ea9c4ef5146..ae11aa914e55 100644
--- a/mm/page_idle.c
+++ b/mm/page_idle.c
@@ -41,12 +41,12 @@ static struct page *page_idle_get_page(unsigned long pfn)
return NULL;
zone = page_zone(page);
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
if (unlikely(!PageLRU(page))) {
put_page(page);
page = NULL;
}
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
return page;
}
diff --git a/mm/rmap.c b/mm/rmap.c
index 0ea5d9071b32..67457f691528 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -27,7 +27,7 @@
* mapping->i_mmap_rwsem
* anon_vma->rwsem
* mm->page_table_lock or pte_lock
- * zone->lru_lock (in mark_page_accessed, isolate_lru_page)
+ * zone_lru_lock (in mark_page_accessed, isolate_lru_page)
* swap_lock (in swap_duplicate, swap_info_get)
* mmlist_lock (in mmput, drain_mmlist and others)
* mapping->private_lock (in __set_page_dirty_buffers)
diff --git a/mm/swap.c b/mm/swap.c
index 59f5fafa6e1f..1aca06977a0a 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -62,12 +62,12 @@ static void __page_cache_release(struct page *page)
struct lruvec *lruvec;
unsigned long flags;
- spin_lock_irqsave(&zone->lru_lock, flags);
+ spin_lock_irqsave(zone_lru_lock(zone), flags);
lruvec = mem_cgroup_page_lruvec(page, zone);
VM_BUG_ON_PAGE(!PageLRU(page), page);
__ClearPageLRU(page);
del_page_from_lru_list(page, lruvec, page_off_lru(page));
- spin_unlock_irqrestore(&zone->lru_lock, flags);
+ spin_unlock_irqrestore(zone_lru_lock(zone), flags);
}
mem_cgroup_uncharge(page);
}
@@ -189,16 +189,16 @@ static void pagevec_lru_move_fn(struct pagevec *pvec,
if (pagezone != zone) {
if (zone)
- spin_unlock_irqrestore(&zone->lru_lock, flags);
+ spin_unlock_irqrestore(zone_lru_lock(zone), flags);
zone = pagezone;
- spin_lock_irqsave(&zone->lru_lock, flags);
+ spin_lock_irqsave(zone_lru_lock(zone), flags);
}
lruvec = mem_cgroup_page_lruvec(page, zone);
(*move_fn)(page, lruvec, arg);
}
if (zone)
- spin_unlock_irqrestore(&zone->lru_lock, flags);
+ spin_unlock_irqrestore(zone_lru_lock(zone), flags);
release_pages(pvec->pages, pvec->nr, pvec->cold);
pagevec_reinit(pvec);
}
@@ -316,9 +316,9 @@ void activate_page(struct page *page)
{
struct zone *zone = page_zone(page);
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
__activate_page(page, mem_cgroup_page_lruvec(page, zone), NULL);
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
}
#endif
@@ -447,13 +447,13 @@ void add_page_to_unevictable_list(struct page *page)
struct zone *zone = page_zone(page);
struct lruvec *lruvec;
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
lruvec = mem_cgroup_page_lruvec(page, zone);
ClearPageActive(page);
SetPageUnevictable(page);
SetPageLRU(page);
add_page_to_lru_list(page, lruvec, LRU_UNEVICTABLE);
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
}
/**
@@ -743,7 +743,7 @@ void release_pages(struct page **pages, int nr, bool cold)
* same zone. The lock is held only if zone != NULL.
*/
if (zone && ++lock_batch == SWAP_CLUSTER_MAX) {
- spin_unlock_irqrestore(&zone->lru_lock, flags);
+ spin_unlock_irqrestore(zone_lru_lock(zone), flags);
zone = NULL;
}
@@ -758,7 +758,7 @@ void release_pages(struct page **pages, int nr, bool cold)
if (PageCompound(page)) {
if (zone) {
- spin_unlock_irqrestore(&zone->lru_lock, flags);
+ spin_unlock_irqrestore(zone_lru_lock(zone), flags);
zone = NULL;
}
__put_compound_page(page);
@@ -770,11 +770,11 @@ void release_pages(struct page **pages, int nr, bool cold)
if (pagezone != zone) {
if (zone)
- spin_unlock_irqrestore(&zone->lru_lock,
+ spin_unlock_irqrestore(zone_lru_lock(zone),
flags);
lock_batch = 0;
zone = pagezone;
- spin_lock_irqsave(&zone->lru_lock, flags);
+ spin_lock_irqsave(zone_lru_lock(zone), flags);
}
lruvec = mem_cgroup_page_lruvec(page, zone);
@@ -789,7 +789,7 @@ void release_pages(struct page **pages, int nr, bool cold)
list_add(&page->lru, &pages_to_free);
}
if (zone)
- spin_unlock_irqrestore(&zone->lru_lock, flags);
+ spin_unlock_irqrestore(zone_lru_lock(zone), flags);
mem_cgroup_uncharge_list(&pages_to_free);
free_hot_cold_page_list(&pages_to_free, cold);
@@ -825,7 +825,7 @@ void lru_add_page_tail(struct page *page, struct page *page_tail,
VM_BUG_ON_PAGE(PageCompound(page_tail), page);
VM_BUG_ON_PAGE(PageLRU(page_tail), page);
VM_BUG_ON(NR_CPUS != 1 &&
- !spin_is_locked(&lruvec_zone(lruvec)->lru_lock));
+ !spin_is_locked(zone_lru_lock(lruvec_zone(lruvec))));
if (!list)
SetPageLRU(page_tail);
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 93ba33789ac6..e5c577c40ff9 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1343,7 +1343,7 @@ int __isolate_lru_page(struct page *page, isolate_mode_t mode)
}
/*
- * zone->lru_lock is heavily contended. Some of the functions that
+ * zone_lru_lock is heavily contended. Some of the functions that
* shrink the lists perform better by taking out a batch of pages
* and working on them outside the LRU lock.
*
@@ -1438,7 +1438,7 @@ int isolate_lru_page(struct page *page)
struct zone *zone = page_zone(page);
struct lruvec *lruvec;
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
lruvec = mem_cgroup_page_lruvec(page, zone);
if (PageLRU(page)) {
int lru = page_lru(page);
@@ -1447,7 +1447,7 @@ int isolate_lru_page(struct page *page)
del_page_from_lru_list(page, lruvec, lru);
ret = 0;
}
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
}
return ret;
}
@@ -1506,9 +1506,9 @@ putback_inactive_pages(struct lruvec *lruvec, struct list_head *page_list)
VM_BUG_ON_PAGE(PageLRU(page), page);
list_del(&page->lru);
if (unlikely(!page_evictable(page))) {
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
putback_lru_page(page);
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
continue;
}
@@ -1529,10 +1529,10 @@ putback_inactive_pages(struct lruvec *lruvec, struct list_head *page_list)
del_page_from_lru_list(page, lruvec, lru);
if (unlikely(PageCompound(page))) {
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
mem_cgroup_uncharge(page);
(*get_compound_page_dtor(page))(page);
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
} else
list_add(&page->lru, &pages_to_free);
}
@@ -1594,7 +1594,7 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
if (!sc->may_writepage)
isolate_mode |= ISOLATE_CLEAN;
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &page_list,
&nr_scanned, sc, isolate_mode, lru);
@@ -1610,7 +1610,7 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
else
__count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scanned);
}
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
if (nr_taken == 0)
return 0;
@@ -1620,7 +1620,7 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
&nr_writeback, &nr_immediate,
false);
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
if (global_reclaim(sc)) {
if (current_is_kswapd())
@@ -1635,7 +1635,7 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken);
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
mem_cgroup_uncharge_list(&page_list);
free_hot_cold_page_list(&page_list, true);
@@ -1709,9 +1709,9 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
* processes, from rmap.
*
* If the pages are mostly unmapped, the processing is fast and it is
- * appropriate to hold zone->lru_lock across the whole operation. But if
+ * appropriate to hold zone_lru_lock across the whole operation. But if
* the pages are mapped, the processing is slow (page_referenced()) so we
- * should drop zone->lru_lock around each page. It's impossible to balance
+ * should drop zone_lru_lock around each page. It's impossible to balance
* this, so instead we remove the pages from the LRU while processing them.
* It is safe to rely on PG_active against the non-LRU pages in here because
* nobody will play with that bit on a non-LRU page.
@@ -1748,10 +1748,10 @@ static void move_active_pages_to_lru(struct lruvec *lruvec,
del_page_from_lru_list(page, lruvec, lru);
if (unlikely(PageCompound(page))) {
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
mem_cgroup_uncharge(page);
(*get_compound_page_dtor(page))(page);
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
} else
list_add(&page->lru, pages_to_free);
}
@@ -1786,7 +1786,7 @@ static void shrink_active_list(unsigned long nr_to_scan,
if (!sc->may_writepage)
isolate_mode |= ISOLATE_CLEAN;
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &l_hold,
&nr_scanned, sc, isolate_mode, lru);
@@ -1799,7 +1799,7 @@ static void shrink_active_list(unsigned long nr_to_scan,
__mod_zone_page_state(zone, NR_PAGES_SCANNED, nr_scanned);
__count_zone_vm_events(PGREFILL, zone, nr_scanned);
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
while (!list_empty(&l_hold)) {
cond_resched();
@@ -1844,7 +1844,7 @@ static void shrink_active_list(unsigned long nr_to_scan,
/*
* Move pages back to the lru list.
*/
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
/*
* Count referenced pages from currently used mappings as rotated,
* even though only some of them are actually re-activated. This
@@ -1856,7 +1856,7 @@ static void shrink_active_list(unsigned long nr_to_scan,
move_active_pages_to_lru(lruvec, &l_active, &l_hold, lru);
move_active_pages_to_lru(lruvec, &l_inactive, &l_hold, lru - LRU_ACTIVE);
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken);
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
mem_cgroup_uncharge_list(&l_hold);
free_hot_cold_page_list(&l_hold, true);
@@ -2071,7 +2071,7 @@ static void get_scan_count(struct lruvec *lruvec, struct mem_cgroup *memcg,
file = lruvec_lru_size(lruvec, LRU_ACTIVE_FILE) +
lruvec_lru_size(lruvec, LRU_INACTIVE_FILE);
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) {
reclaim_stat->recent_scanned[0] /= 2;
reclaim_stat->recent_rotated[0] /= 2;
@@ -2092,7 +2092,7 @@ static void get_scan_count(struct lruvec *lruvec, struct mem_cgroup *memcg,
fp = file_prio * (reclaim_stat->recent_scanned[1] + 1);
fp /= reclaim_stat->recent_rotated[1] + 1;
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
fraction[0] = ap;
fraction[1] = fp;
@@ -3785,9 +3785,9 @@ void check_move_unevictable_pages(struct page **pages, int nr_pages)
pagezone = page_zone(page);
if (pagezone != zone) {
if (zone)
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
zone = pagezone;
- spin_lock_irq(&zone->lru_lock);
+ spin_lock_irq(zone_lru_lock(zone));
}
lruvec = mem_cgroup_page_lruvec(page, zone);
@@ -3808,7 +3808,7 @@ void check_move_unevictable_pages(struct page **pages, int nr_pages)
if (zone) {
__count_vm_events(UNEVICTABLE_PGRESCUED, pgrescued);
__count_vm_events(UNEVICTABLE_PGSCANNED, pgscanned);
- spin_unlock_irq(&zone->lru_lock);
+ spin_unlock_irq(zone_lru_lock(zone));
}
}
#endif /* CONFIG_SHMEM */
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 02/27] mm, vmscan: Move lru_lock to the node
2016-06-09 18:04 ` [PATCH 02/27] mm, vmscan: Move lru_lock to the node Mel Gorman
@ 2016-06-10 16:39 ` Vlastimil Babka
2016-06-10 17:48 ` Peter Zijlstra
2016-06-14 14:41 ` Mel Gorman
0 siblings, 2 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-10 16:39 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM
Cc: Rik van Riel, Johannes Weiner, LKML, Peter Zijlstra
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> Node-based reclaim requires node-based LRUs and locking. This is a
> preparation patch that just moves the lru_lock to the node so later patches
> are easier to review. It is a mechanical change but note this patch makes
> contention worse because the LRU lock is hotter and direct reclaim and kswapd
> can contend on the same lock even when reclaiming from different zones.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
One thing...
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 9d71af25acf9..1e0ad06c33bd 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -5944,10 +5944,10 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
> zone->min_slab_pages = (freesize * sysctl_min_slab_ratio) / 100;
> #endif
> zone->name = zone_names[j];
> + zone->zone_pgdat = pgdat;
> spin_lock_init(&zone->lock);
> - spin_lock_init(&zone->lru_lock);
> + spin_lock_init(zone_lru_lock(zone));
This means the same lock will be inited MAX_NR_ZONES times. Peterz told
me it's valid but weird. Probably better to do it just once, in case
lockdep/lock debugging gains some checks for that?
> zone_seqlock_init(zone);
> - zone->zone_pgdat = pgdat;
> zone_pcp_init(zone);
>
> /* For bootup, initialized properly in watermark setup */
^ permalink raw reply [flat|nested] 70+ messages in thread
* Re: [PATCH 02/27] mm, vmscan: Move lru_lock to the node
2016-06-10 16:39 ` Vlastimil Babka
@ 2016-06-10 17:48 ` Peter Zijlstra
2016-06-14 14:41 ` Mel Gorman
1 sibling, 0 replies; 70+ messages in thread
From: Peter Zijlstra @ 2016-06-10 17:48 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Mel Gorman, Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Fri, Jun 10, 2016 at 06:39:26PM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> > Node-based reclaim requires node-based LRUs and locking. This is a
> > preparation patch that just moves the lru_lock to the node so later patches
> > are easier to review. It is a mechanical change but note this patch makes
> > contention worse because the LRU lock is hotter and direct reclaim and kswapd
> > can contend on the same lock even when reclaiming from different zones.
> >
> > Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> > Acked-by: Johannes Weiner <hannes@cmpxchg.org>
>
> Acked-by: Vlastimil Babka <vbabka@suse.cz>
>
> One thing...
>
> > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > index 9d71af25acf9..1e0ad06c33bd 100644
> > --- a/mm/page_alloc.c
> > +++ b/mm/page_alloc.c
> > @@ -5944,10 +5944,10 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
> > zone->min_slab_pages = (freesize * sysctl_min_slab_ratio) / 100;
> > #endif
> > zone->name = zone_names[j];
> > + zone->zone_pgdat = pgdat;
> > spin_lock_init(&zone->lock);
> > - spin_lock_init(&zone->lru_lock);
> > + spin_lock_init(zone_lru_lock(zone));
>
> This means the same lock will be inited MAX_NR_ZONES times. Peterz told
> me it's valid but weird. Probably better to do it just once, in case
> lockdep/lock debugging gains some checks for that?
Ah, I thought you meant using spin_lock_init() after the lock has
already been used. This is fine.
^ permalink raw reply [flat|nested] 70+ messages in thread
* Re: [PATCH 02/27] mm, vmscan: Move lru_lock to the node
2016-06-10 16:39 ` Vlastimil Babka
2016-06-10 17:48 ` Peter Zijlstra
@ 2016-06-14 14:41 ` Mel Gorman
1 sibling, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-14 14:41 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML,
Peter Zijlstra
On Fri, Jun 10, 2016 at 06:39:26PM +0200, Vlastimil Babka wrote:
> > @@ -5944,10 +5944,10 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
> > zone->min_slab_pages = (freesize * sysctl_min_slab_ratio) / 100;
> > #endif
> > zone->name = zone_names[j];
> > + zone->zone_pgdat = pgdat;
> > spin_lock_init(&zone->lock);
> > - spin_lock_init(&zone->lru_lock);
> > + spin_lock_init(zone_lru_lock(zone));
>
> This means the same lock will be inited MAX_NR_ZONES times. Peterz told
> me it's valid but weird. Probably better to do it just once, in case
> lockdep/lock debugging gains some checks for that?
>
Good point and it's an easy fix.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 03/27] mm, vmscan: Move LRU lists to node
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
2016-06-09 18:04 ` [PATCH 01/27] mm, vmstat: Add infrastructure for per-node vmstats Mel Gorman
2016-06-09 18:04 ` [PATCH 02/27] mm, vmscan: Move lru_lock to the node Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-10 18:00 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 04/27] mm, vmscan: Begin reclaiming pages on a per-node basis Mel Gorman
` (23 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
This moves the LRU lists from the zone to the node and all related data
such as counters, tracing, congestion tracking and writeback tracking.
This is mostly a mechanical patch but note that it introduces a number
of anomalies. For example, the scans are per-zone but using per-node
counters. We also mark a node as congested when a zone is congested. This
causes weird problems that are fixed later but is easier to review.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
---
arch/tile/mm/pgtable.c | 8 +-
drivers/base/node.c | 19 +--
drivers/staging/android/lowmemorykiller.c | 8 +-
include/linux/backing-dev.h | 2 +-
include/linux/memcontrol.h | 8 +-
include/linux/mm_inline.h | 2 +-
include/linux/mmzone.h | 67 ++++++-----
include/linux/swap.h | 2 +-
include/linux/vm_event_item.h | 10 +-
include/linux/vmstat.h | 17 +++
include/trace/events/vmscan.h | 12 +-
kernel/power/snapshot.c | 10 +-
mm/backing-dev.c | 15 +--
mm/compaction.c | 29 +++--
mm/huge_memory.c | 6 +-
mm/internal.h | 2 +-
mm/memcontrol.c | 18 +--
mm/memory-failure.c | 4 +-
mm/memory_hotplug.c | 2 +-
mm/mempolicy.c | 2 +-
mm/migrate.c | 21 ++--
mm/mlock.c | 2 +-
mm/page-writeback.c | 8 +-
mm/page_alloc.c | 81 ++++++++-----
mm/swap.c | 56 ++++-----
mm/vmscan.c | 194 ++++++++++++++++--------------
mm/vmstat.c | 45 +++----
mm/workingset.c | 4 +-
28 files changed, 361 insertions(+), 293 deletions(-)
diff --git a/arch/tile/mm/pgtable.c b/arch/tile/mm/pgtable.c
index c4d5bf841a7f..9e389213580d 100644
--- a/arch/tile/mm/pgtable.c
+++ b/arch/tile/mm/pgtable.c
@@ -45,10 +45,10 @@ void show_mem(unsigned int filter)
struct zone *zone;
pr_err("Active:%lu inactive:%lu dirty:%lu writeback:%lu unstable:%lu free:%lu\n slab:%lu mapped:%lu pagetables:%lu bounce:%lu pagecache:%lu swap:%lu\n",
- (global_page_state(NR_ACTIVE_ANON) +
- global_page_state(NR_ACTIVE_FILE)),
- (global_page_state(NR_INACTIVE_ANON) +
- global_page_state(NR_INACTIVE_FILE)),
+ (global_node_page_state(NR_ACTIVE_ANON) +
+ global_node_page_state(NR_ACTIVE_FILE)),
+ (global_node_page_state(NR_INACTIVE_ANON) +
+ global_node_page_state(NR_INACTIVE_FILE)),
global_page_state(NR_FILE_DIRTY),
global_page_state(NR_WRITEBACK),
global_page_state(NR_UNSTABLE_NFS),
diff --git a/drivers/base/node.c b/drivers/base/node.c
index efb81da250a8..4260c7f3ee1b 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -56,6 +56,7 @@ static ssize_t node_read_meminfo(struct device *dev,
{
int n;
int nid = dev->id;
+ struct pglist_data *pgdat = NODE_DATA(nid);
struct sysinfo i;
si_meminfo_node(&i, nid);
@@ -74,15 +75,15 @@ static ssize_t node_read_meminfo(struct device *dev,
nid, K(i.totalram),
nid, K(i.freeram),
nid, K(i.totalram - i.freeram),
- nid, K(sum_zone_node_page_state(nid, NR_ACTIVE_ANON) +
- sum_zone_node_page_state(nid, NR_ACTIVE_FILE)),
- nid, K(sum_zone_node_page_state(nid, NR_INACTIVE_ANON) +
- sum_zone_node_page_state(nid, NR_INACTIVE_FILE)),
- nid, K(sum_zone_node_page_state(nid, NR_ACTIVE_ANON)),
- nid, K(sum_zone_node_page_state(nid, NR_INACTIVE_ANON)),
- nid, K(sum_zone_node_page_state(nid, NR_ACTIVE_FILE)),
- nid, K(sum_zone_node_page_state(nid, NR_INACTIVE_FILE)),
- nid, K(sum_zone_node_page_state(nid, NR_UNEVICTABLE)),
+ nid, K(node_page_state(pgdat, NR_ACTIVE_ANON) +
+ node_page_state(pgdat, NR_ACTIVE_FILE)),
+ nid, K(node_page_state(pgdat, NR_INACTIVE_ANON) +
+ node_page_state(pgdat, NR_INACTIVE_FILE)),
+ nid, K(node_page_state(pgdat, NR_ACTIVE_ANON)),
+ nid, K(node_page_state(pgdat, NR_INACTIVE_ANON)),
+ nid, K(node_page_state(pgdat, NR_ACTIVE_FILE)),
+ nid, K(node_page_state(pgdat, NR_INACTIVE_FILE)),
+ nid, K(node_page_state(pgdat, NR_UNEVICTABLE)),
nid, K(sum_zone_node_page_state(nid, NR_MLOCK)));
#ifdef CONFIG_HIGHMEM
diff --git a/drivers/staging/android/lowmemorykiller.c b/drivers/staging/android/lowmemorykiller.c
index 24d2745e9437..93dbcc38eb0f 100644
--- a/drivers/staging/android/lowmemorykiller.c
+++ b/drivers/staging/android/lowmemorykiller.c
@@ -72,10 +72,10 @@ static unsigned long lowmem_deathpending_timeout;
static unsigned long lowmem_count(struct shrinker *s,
struct shrink_control *sc)
{
- return global_page_state(NR_ACTIVE_ANON) +
- global_page_state(NR_ACTIVE_FILE) +
- global_page_state(NR_INACTIVE_ANON) +
- global_page_state(NR_INACTIVE_FILE);
+ return global_node_page_state(NR_ACTIVE_ANON) +
+ global_node_page_state(NR_ACTIVE_FILE) +
+ global_node_page_state(NR_INACTIVE_ANON) +
+ global_node_page_state(NR_INACTIVE_FILE);
}
static unsigned long lowmem_scan(struct shrinker *s, struct shrink_control *sc)
diff --git a/include/linux/backing-dev.h b/include/linux/backing-dev.h
index c82794f20110..491a91717788 100644
--- a/include/linux/backing-dev.h
+++ b/include/linux/backing-dev.h
@@ -197,7 +197,7 @@ static inline int wb_congested(struct bdi_writeback *wb, int cong_bits)
}
long congestion_wait(int sync, long timeout);
-long wait_iff_congested(struct zone *zone, int sync, long timeout);
+long wait_iff_congested(struct pglist_data *pgdat, int sync, long timeout);
int pdflush_proc_obsolete(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp, loff_t *ppos);
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 2d03975c7dc0..9adb1bebb487 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -307,7 +307,7 @@ void mem_cgroup_uncharge_list(struct list_head *page_list);
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage);
struct lruvec *mem_cgroup_zone_lruvec(struct zone *, struct mem_cgroup *);
-struct lruvec *mem_cgroup_page_lruvec(struct page *, struct zone *);
+struct lruvec *mem_cgroup_page_lruvec(struct page *, struct pglist_data *);
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg);
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
@@ -576,13 +576,13 @@ static inline void mem_cgroup_migrate(struct page *old, struct page *new)
static inline struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone,
struct mem_cgroup *memcg)
{
- return &zone->lruvec;
+ return zone_lruvec(zone);
}
static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
- struct zone *zone)
+ struct pglist_data *pgdat)
{
- return &zone->lruvec;
+ return &pgdat->lruvec;
}
static inline bool mm_match_cgroup(struct mm_struct *mm,
diff --git a/include/linux/mm_inline.h b/include/linux/mm_inline.h
index 5bd29ba4f174..797628d69b59 100644
--- a/include/linux/mm_inline.h
+++ b/include/linux/mm_inline.h
@@ -25,7 +25,7 @@ static inline int page_is_file_cache(struct page *page)
static __always_inline void __update_lru_size(struct lruvec *lruvec,
enum lru_list lru, int nr_pages)
{
- __mod_zone_page_state(lruvec_zone(lruvec), NR_LRU_BASE + lru, nr_pages);
+ __mod_node_page_state(lruvec_pgdat(lruvec), NR_LRU_BASE + lru, nr_pages);
}
static __always_inline void update_lru_size(struct lruvec *lruvec,
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 70623b6ac833..aff621fcc6af 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -111,12 +111,6 @@ enum zone_stat_item {
/* First 128 byte cacheline (assuming 64 bit words) */
NR_FREE_PAGES,
NR_ALLOC_BATCH,
- NR_LRU_BASE,
- NR_INACTIVE_ANON = NR_LRU_BASE, /* must match order of LRU_[IN]ACTIVE */
- NR_ACTIVE_ANON, /* " " " " " */
- NR_INACTIVE_FILE, /* " " " " " */
- NR_ACTIVE_FILE, /* " " " " " */
- NR_UNEVICTABLE, /* " " " " " */
NR_MLOCK, /* mlock()ed pages found and moved off LRU */
NR_ANON_PAGES, /* Mapped anonymous pages */
NR_FILE_MAPPED, /* pagecache pages mapped into pagetables.
@@ -134,12 +128,9 @@ enum zone_stat_item {
NR_VMSCAN_WRITE,
NR_VMSCAN_IMMEDIATE, /* Prioritise for reclaim when writeback ends */
NR_WRITEBACK_TEMP, /* Writeback using temporary buffers */
- NR_ISOLATED_ANON, /* Temporary isolated pages from anon lru */
- NR_ISOLATED_FILE, /* Temporary isolated pages from file lru */
NR_SHMEM, /* shmem pages (included tmpfs/GEM pages) */
NR_DIRTIED, /* page dirtyings since bootup */
NR_WRITTEN, /* page writings since bootup */
- NR_PAGES_SCANNED, /* pages scanned since last reclaim */
#if IS_ENABLED(CONFIG_ZSMALLOC)
NR_ZSMALLOC,
#endif
@@ -159,6 +150,15 @@ enum zone_stat_item {
NR_VM_ZONE_STAT_ITEMS };
enum node_stat_item {
+ NR_LRU_BASE,
+ NR_INACTIVE_ANON = NR_LRU_BASE, /* must match order of LRU_[IN]ACTIVE */
+ NR_ACTIVE_ANON, /* " " " " " */
+ NR_INACTIVE_FILE, /* " " " " " */
+ NR_ACTIVE_FILE, /* " " " " " */
+ NR_UNEVICTABLE, /* " " " " " */
+ NR_ISOLATED_ANON, /* Temporary isolated pages from anon lru */
+ NR_ISOLATED_FILE, /* Temporary isolated pages from file lru */
+ NR_PAGES_SCANNED, /* pages scanned since last reclaim */
NR_VM_NODE_STAT_ITEMS
};
@@ -217,7 +217,7 @@ struct lruvec {
/* Evictions & activations on the inactive file list */
atomic_long_t inactive_age;
#ifdef CONFIG_MEMCG
- struct zone *zone;
+ struct pglist_data *pgdat;
#endif
};
@@ -355,13 +355,6 @@ struct zone {
#ifdef CONFIG_NUMA
int node;
#endif
-
- /*
- * The target ratio of ACTIVE_ANON to INACTIVE_ANON pages on
- * this zone's LRU. Maintained by the pageout code.
- */
- unsigned int inactive_ratio;
-
struct pglist_data *zone_pgdat;
struct per_cpu_pageset __percpu *pageset;
@@ -493,9 +486,6 @@ struct zone {
/* Write-intensive fields used by page reclaim */
- /* Fields commonly accessed by the page reclaim scanner */
- struct lruvec lruvec;
-
/*
* When free pages are below this point, additional steps are taken
* when reading the number of free pages to avoid per-cpu counter
@@ -535,17 +525,21 @@ struct zone {
enum zone_flags {
ZONE_RECLAIM_LOCKED, /* prevents concurrent reclaim */
- ZONE_CONGESTED, /* zone has many dirty pages backed by
+ ZONE_OOM_LOCKED, /* zone is in OOM killer zonelist */
+ ZONE_FAIR_DEPLETED, /* fair zone policy batch depleted */
+};
+
+enum pgdat_flags {
+ PGDAT_CONGESTED, /* zone has many dirty pages backed by
* a congested BDI
*/
- ZONE_DIRTY, /* reclaim scanning has recently found
+ PGDAT_DIRTY, /* reclaim scanning has recently found
* many dirty file pages at the tail
* of the LRU.
*/
- ZONE_WRITEBACK, /* reclaim scanning has recently found
+ PGDAT_WRITEBACK, /* reclaim scanning has recently found
* many pages under writeback
*/
- ZONE_FAIR_DEPLETED, /* fair zone policy batch depleted */
};
static inline unsigned long zone_end_pfn(const struct zone *zone)
@@ -699,12 +693,26 @@ typedef struct pglist_data {
unsigned long first_deferred_pfn;
#endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
+
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
spinlock_t split_queue_lock;
struct list_head split_queue;
unsigned long split_queue_len;
#endif
+ /* Fields commonly accessed by the page reclaim scanner */
+ struct lruvec lruvec;
+
+ /*
+ * The target ratio of ACTIVE_ANON to INACTIVE_ANON pages on
+ * this node's LRU. Maintained by the pageout code.
+ */
+ unsigned int inactive_ratio;
+
+ unsigned long flags;
+
+ ZONE_PADDING(_pad2_)
+
/* Per-node vmstats */
struct per_cpu_nodestat __percpu *per_cpu_nodestats;
atomic_long_t vm_stat[NR_VM_NODE_STAT_ITEMS];
@@ -726,6 +734,11 @@ static inline spinlock_t *zone_lru_lock(struct zone *zone)
return &zone->zone_pgdat->lru_lock;
}
+static inline struct lruvec *zone_lruvec(struct zone *zone)
+{
+ return &zone->zone_pgdat->lruvec;
+}
+
static inline unsigned long pgdat_end_pfn(pg_data_t *pgdat)
{
return pgdat->node_start_pfn + pgdat->node_spanned_pages;
@@ -777,12 +790,12 @@ extern int init_currently_empty_zone(struct zone *zone, unsigned long start_pfn,
extern void lruvec_init(struct lruvec *lruvec);
-static inline struct zone *lruvec_zone(struct lruvec *lruvec)
+static inline struct pglist_data *lruvec_pgdat(struct lruvec *lruvec)
{
#ifdef CONFIG_MEMCG
- return lruvec->zone;
+ return lruvec->pgdat;
#else
- return container_of(lruvec, struct zone, lruvec);
+ return container_of(lruvec, struct pglist_data, lruvec);
#endif
}
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 0af2bb2028fd..cc5c958c050b 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -316,7 +316,7 @@ extern void lru_cache_add_active_or_unevictable(struct page *page,
struct vm_area_struct *vma);
/* linux/mm/vmscan.c */
-extern unsigned long zone_reclaimable_pages(struct zone *zone);
+extern unsigned long pgdat_reclaimable_pages(struct pglist_data *pgdat);
extern unsigned long try_to_free_pages(struct zonelist *zonelist, int order,
gfp_t gfp_mask, nodemask_t *mask);
extern int __isolate_lru_page(struct page *page, isolate_mode_t mode);
diff --git a/include/linux/vm_event_item.h b/include/linux/vm_event_item.h
index ec084321fe09..8dcb5a813163 100644
--- a/include/linux/vm_event_item.h
+++ b/include/linux/vm_event_item.h
@@ -26,11 +26,11 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT,
PGFREE, PGACTIVATE, PGDEACTIVATE,
PGFAULT, PGMAJFAULT,
PGLAZYFREED,
- FOR_ALL_ZONES(PGREFILL),
- FOR_ALL_ZONES(PGSTEAL_KSWAPD),
- FOR_ALL_ZONES(PGSTEAL_DIRECT),
- FOR_ALL_ZONES(PGSCAN_KSWAPD),
- FOR_ALL_ZONES(PGSCAN_DIRECT),
+ PGREFILL,
+ PGSTEAL_KSWAPD,
+ PGSTEAL_DIRECT,
+ PGSCAN_KSWAPD,
+ PGSCAN_DIRECT,
PGSCAN_DIRECT_THROTTLE,
#ifdef CONFIG_NUMA
PGSCAN_ZONE_RECLAIM_FAILED,
diff --git a/include/linux/vmstat.h b/include/linux/vmstat.h
index 40629576f8de..941695e6b780 100644
--- a/include/linux/vmstat.h
+++ b/include/linux/vmstat.h
@@ -184,6 +184,23 @@ static inline unsigned long zone_page_state_snapshot(struct zone *zone,
return x;
}
+static inline unsigned long node_page_state_snapshot(pg_data_t *pgdat,
+ enum zone_stat_item item)
+{
+ long x = atomic_long_read(&pgdat->vm_stat[item]);
+
+#ifdef CONFIG_SMP
+ int cpu;
+ for_each_online_cpu(cpu)
+ x += per_cpu_ptr(pgdat->per_cpu_nodestats, cpu)->vm_node_stat_diff[item];
+
+ if (x < 0)
+ x = 0;
+#endif
+ return x;
+}
+
+
#ifdef CONFIG_NUMA
extern unsigned long sum_zone_node_page_state(int node,
enum zone_stat_item item);
diff --git a/include/trace/events/vmscan.h b/include/trace/events/vmscan.h
index 0101ef37f1ee..897f1aa1ee5f 100644
--- a/include/trace/events/vmscan.h
+++ b/include/trace/events/vmscan.h
@@ -352,15 +352,14 @@ TRACE_EVENT(mm_vmscan_writepage,
TRACE_EVENT(mm_vmscan_lru_shrink_inactive,
- TP_PROTO(struct zone *zone,
+ TP_PROTO(int nid,
unsigned long nr_scanned, unsigned long nr_reclaimed,
int priority, int file),
- TP_ARGS(zone, nr_scanned, nr_reclaimed, priority, file),
+ TP_ARGS(nid, nr_scanned, nr_reclaimed, priority, file),
TP_STRUCT__entry(
__field(int, nid)
- __field(int, zid)
__field(unsigned long, nr_scanned)
__field(unsigned long, nr_reclaimed)
__field(int, priority)
@@ -368,16 +367,15 @@ TRACE_EVENT(mm_vmscan_lru_shrink_inactive,
),
TP_fast_assign(
- __entry->nid = zone_to_nid(zone);
- __entry->zid = zone_idx(zone);
+ __entry->nid = nid;
__entry->nr_scanned = nr_scanned;
__entry->nr_reclaimed = nr_reclaimed;
__entry->priority = priority;
__entry->reclaim_flags = trace_shrink_flags(file);
),
- TP_printk("nid=%d zid=%d nr_scanned=%ld nr_reclaimed=%ld priority=%d flags=%s",
- __entry->nid, __entry->zid,
+ TP_printk("nid=%d nr_scanned=%ld nr_reclaimed=%ld priority=%d flags=%s",
+ __entry->nid,
__entry->nr_scanned, __entry->nr_reclaimed,
__entry->priority,
show_reclaim_flags(__entry->reclaim_flags))
diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c
index 3a970604308f..24a06bc23f85 100644
--- a/kernel/power/snapshot.c
+++ b/kernel/power/snapshot.c
@@ -1525,11 +1525,11 @@ static unsigned long minimum_image_size(unsigned long saveable)
unsigned long size;
size = global_page_state(NR_SLAB_RECLAIMABLE)
- + global_page_state(NR_ACTIVE_ANON)
- + global_page_state(NR_INACTIVE_ANON)
- + global_page_state(NR_ACTIVE_FILE)
- + global_page_state(NR_INACTIVE_FILE)
- - global_page_state(NR_FILE_MAPPED);
+ + global_node_page_state(NR_ACTIVE_ANON)
+ + global_node_page_state(NR_INACTIVE_ANON)
+ + global_node_page_state(NR_ACTIVE_FILE)
+ + global_node_page_state(NR_INACTIVE_FILE)
+ - global_node_page_state(NR_FILE_MAPPED);
return saveable <= size ? 0 : saveable - size;
}
diff --git a/mm/backing-dev.c b/mm/backing-dev.c
index f53b23ab7ed7..a8c3af46bd3d 100644
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -982,24 +982,24 @@ long congestion_wait(int sync, long timeout)
EXPORT_SYMBOL(congestion_wait);
/**
- * wait_iff_congested - Conditionally wait for a backing_dev to become uncongested or a zone to complete writes
- * @zone: A zone to check if it is heavily congested
+ * wait_iff_congested - Conditionally wait for a backing_dev to become uncongested or a pgdat to complete writes
+ * @pgdat: A pgdat to check if it is heavily congested
* @sync: SYNC or ASYNC IO
* @timeout: timeout in jiffies
*
* In the event of a congested backing_dev (any backing_dev) and the given
- * @zone has experienced recent congestion, this waits for up to @timeout
+ * @pgdat has experienced recent congestion, this waits for up to @timeout
* jiffies for either a BDI to exit congestion of the given @sync queue
* or a write to complete.
*
- * In the absence of zone congestion, cond_resched() is called to yield
+ * In the absence of pgdat congestion, cond_resched() is called to yield
* the processor if necessary but otherwise does not sleep.
*
* The return value is 0 if the sleep is for the full timeout. Otherwise,
* it is the number of jiffies that were still remaining when the function
* returned. return_value == timeout implies the function did not sleep.
*/
-long wait_iff_congested(struct zone *zone, int sync, long timeout)
+long wait_iff_congested(struct pglist_data *pgdat, int sync, long timeout)
{
long ret;
unsigned long start = jiffies;
@@ -1008,12 +1008,13 @@ long wait_iff_congested(struct zone *zone, int sync, long timeout)
/*
* If there is no congestion, or heavy congestion is not being
- * encountered in the current zone, yield if necessary instead
+ * encountered in the current pgdat, yield if necessary instead
* of sleeping on the congestion queue
*/
if (atomic_read(&nr_wb_congested[sync]) == 0 ||
- !test_bit(ZONE_CONGESTED, &zone->flags)) {
+ !test_bit(PGDAT_CONGESTED, &pgdat->flags)) {
cond_resched();
+
/* In case we scheduled, work out time remaining */
ret = timeout - (jiffies - start);
if (ret < 0)
diff --git a/mm/compaction.c b/mm/compaction.c
index 5aba9e6c1975..92efd98675f2 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -647,8 +647,8 @@ static void acct_isolated(struct zone *zone, struct compact_control *cc)
list_for_each_entry(page, &cc->migratepages, lru)
count[!!page_is_file_cache(page)]++;
- mod_zone_page_state(zone, NR_ISOLATED_ANON, count[0]);
- mod_zone_page_state(zone, NR_ISOLATED_FILE, count[1]);
+ mod_node_page_state(zone->zone_pgdat, NR_ISOLATED_ANON, count[0]);
+ mod_node_page_state(zone->zone_pgdat, NR_ISOLATED_FILE, count[1]);
}
/* Similar to reclaim, but different enough that they don't share logic */
@@ -656,12 +656,12 @@ static bool too_many_isolated(struct zone *zone)
{
unsigned long active, inactive, isolated;
- inactive = zone_page_state(zone, NR_INACTIVE_FILE) +
- zone_page_state(zone, NR_INACTIVE_ANON);
- active = zone_page_state(zone, NR_ACTIVE_FILE) +
- zone_page_state(zone, NR_ACTIVE_ANON);
- isolated = zone_page_state(zone, NR_ISOLATED_FILE) +
- zone_page_state(zone, NR_ISOLATED_ANON);
+ inactive = node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE) +
+ node_page_state(zone->zone_pgdat, NR_INACTIVE_ANON);
+ active = node_page_state(zone->zone_pgdat, NR_ACTIVE_FILE) +
+ node_page_state(zone->zone_pgdat, NR_ACTIVE_ANON);
+ isolated = node_page_state(zone->zone_pgdat, NR_ISOLATED_FILE) +
+ node_page_state(zone->zone_pgdat, NR_ISOLATED_ANON);
return isolated > (inactive + active) / 2;
}
@@ -859,7 +859,7 @@ isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn,
}
}
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, zone->zone_pgdat);
/* Try isolate the page */
if (__isolate_lru_page(page, isolate_mode) != 0)
@@ -1455,13 +1455,22 @@ bool compaction_zonelist_suitable(struct alloc_context *ac, int order,
enum compact_result compact_result;
/*
+ * This over-estimates the number of pages available for
+ * reclaim/compaction but walking the LRU would take too
+ * long. The consequences are that compaction may retry
+ * longer than it should for a zone-constrained allocation
+ * request.
+ */
+ available = pgdat_reclaimable_pages(zone->zone_pgdat);
+
+ /*
* Do not consider all the reclaimable memory because we do not
* want to trash just for a single high order allocation which
* is even not guaranteed to appear even if __compaction_suitable
* is happy about the watermark check.
*/
- available = zone_reclaimable_pages(zone) / order;
available += zone_page_state_snapshot(zone, NR_FREE_PAGES);
+ available = min(zone->managed_pages, available);
compact_result = __compaction_suitable(zone, order, alloc_flags,
ac_classzone_idx(ac), available);
if (compact_result != COMPACT_SKIPPED &&
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 8d535b778ead..c368db5cadca 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2035,7 +2035,7 @@ void __khugepaged_exit(struct mm_struct *mm)
static void release_pte_page(struct page *page)
{
/* 0 stands for page_is_file_cache(page) == false */
- dec_zone_page_state(page, NR_ISOLATED_ANON + 0);
+ dec_node_page_state(page, NR_ISOLATED_ANON + 0);
unlock_page(page);
putback_lru_page(page);
}
@@ -2131,7 +2131,7 @@ static int __collapse_huge_page_isolate(struct vm_area_struct *vma,
goto out;
}
/* 0 stands for page_is_file_cache(page) == false */
- inc_zone_page_state(page, NR_ISOLATED_ANON + 0);
+ inc_node_page_state(page, NR_ISOLATED_ANON + 0);
VM_BUG_ON_PAGE(!PageLocked(page), page);
VM_BUG_ON_PAGE(PageLRU(page), page);
@@ -3301,7 +3301,7 @@ static void __split_huge_page(struct page *page, struct list_head *list)
/* prevent PageLRU to go away from under us, and freeze lru stats */
spin_lock_irq(zone_lru_lock(zone));
- lruvec = mem_cgroup_page_lruvec(head, zone);
+ lruvec = mem_cgroup_page_lruvec(head, zone->zone_pgdat);
/* complete memcg works before add pages to LRU */
mem_cgroup_split_huge_fixup(head);
diff --git a/mm/internal.h b/mm/internal.h
index 420bbe300bcd..d5e4e7db141d 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -79,7 +79,7 @@ extern unsigned long highest_memmap_pfn;
*/
extern int isolate_lru_page(struct page *page);
extern void putback_lru_page(struct page *page);
-extern bool zone_reclaimable(struct zone *zone);
+extern bool pgdat_reclaimable(struct pglist_data *pgdat);
/*
* in mm/rmap.c:
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 605e8b5fc0db..12f796e52e5e 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -959,7 +959,7 @@ struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone,
struct lruvec *lruvec;
if (mem_cgroup_disabled()) {
- lruvec = &zone->lruvec;
+ lruvec = zone_lruvec(zone);
goto out;
}
@@ -971,8 +971,8 @@ struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone,
* we have to be prepared to initialize lruvec->zone here;
* and if offlined then reonlined, we need to reinitialize it.
*/
- if (unlikely(lruvec->zone != zone))
- lruvec->zone = zone;
+ if (unlikely(lruvec->pgdat != zone->zone_pgdat))
+ lruvec->pgdat = zone->zone_pgdat;
return lruvec;
}
@@ -985,14 +985,14 @@ struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone,
* and putback protocol: the LRU lock must be held, and the page must
* either be PageLRU() or the caller must have isolated/allocated it.
*/
-struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct zone *zone)
+struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct pglist_data *pgdat)
{
struct mem_cgroup_per_zone *mz;
struct mem_cgroup *memcg;
struct lruvec *lruvec;
if (mem_cgroup_disabled()) {
- lruvec = &zone->lruvec;
+ lruvec = &pgdat->lruvec;
goto out;
}
@@ -1012,8 +1012,8 @@ struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct zone *zone)
* we have to be prepared to initialize lruvec->zone here;
* and if offlined then reonlined, we need to reinitialize it.
*/
- if (unlikely(lruvec->zone != zone))
- lruvec->zone = zone;
+ if (unlikely(lruvec->pgdat != pgdat))
+ lruvec->pgdat = pgdat;
return lruvec;
}
@@ -2111,7 +2111,7 @@ static void lock_page_lru(struct page *page, int *isolated)
if (PageLRU(page)) {
struct lruvec *lruvec;
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, zone->zone_pgdat);
ClearPageLRU(page);
del_page_from_lru_list(page, lruvec, page_lru(page));
*isolated = 1;
@@ -2126,7 +2126,7 @@ static void unlock_page_lru(struct page *page, int isolated)
if (isolated) {
struct lruvec *lruvec;
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, zone->zone_pgdat);
VM_BUG_ON_PAGE(PageLRU(page), page);
SetPageLRU(page);
add_page_to_lru_list(page, lruvec, page_lru(page));
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 2fcca6b0e005..11de752ccaf5 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -1663,7 +1663,7 @@ static int __soft_offline_page(struct page *page, int flags)
put_hwpoison_page(page);
if (!ret) {
LIST_HEAD(pagelist);
- inc_zone_page_state(page, NR_ISOLATED_ANON +
+ inc_node_page_state(page, NR_ISOLATED_ANON +
page_is_file_cache(page));
list_add(&page->lru, &pagelist);
ret = migrate_pages(&pagelist, new_page, NULL, MPOL_MF_MOVE_ALL,
@@ -1671,7 +1671,7 @@ static int __soft_offline_page(struct page *page, int flags)
if (ret) {
if (!list_empty(&pagelist)) {
list_del(&page->lru);
- dec_zone_page_state(page, NR_ISOLATED_ANON +
+ dec_node_page_state(page, NR_ISOLATED_ANON +
page_is_file_cache(page));
putback_lru_page(page);
}
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index 82d0b98d27f8..c5278360ca66 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1586,7 +1586,7 @@ do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
put_page(page);
list_add_tail(&page->lru, &source);
move_pages--;
- inc_zone_page_state(page, NR_ISOLATED_ANON +
+ inc_node_page_state(page, NR_ISOLATED_ANON +
page_is_file_cache(page));
} else {
diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index fe90e5051012..e1190689634e 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -962,7 +962,7 @@ static void migrate_page_add(struct page *page, struct list_head *pagelist,
if ((flags & MPOL_MF_MOVE_ALL) || page_mapcount(page) == 1) {
if (!isolate_lru_page(page)) {
list_add_tail(&page->lru, pagelist);
- inc_zone_page_state(page, NR_ISOLATED_ANON +
+ inc_node_page_state(page, NR_ISOLATED_ANON +
page_is_file_cache(page));
}
}
diff --git a/mm/migrate.c b/mm/migrate.c
index e6daf49e224f..cff5d6aa2240 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -168,7 +168,7 @@ void putback_movable_pages(struct list_head *l)
continue;
}
list_del(&page->lru);
- dec_zone_page_state(page, NR_ISOLATED_ANON +
+ dec_node_page_state(page, NR_ISOLATED_ANON +
page_is_file_cache(page));
/*
* We isolated non-lru movable page so here we can use
@@ -1117,7 +1117,7 @@ static ICE_noinline int unmap_and_move(new_page_t get_new_page,
* restored.
*/
list_del(&page->lru);
- dec_zone_page_state(page, NR_ISOLATED_ANON +
+ dec_node_page_state(page, NR_ISOLATED_ANON +
page_is_file_cache(page));
}
@@ -1458,7 +1458,7 @@ static int do_move_page_to_node_array(struct mm_struct *mm,
err = isolate_lru_page(page);
if (!err) {
list_add_tail(&page->lru, &pagelist);
- inc_zone_page_state(page, NR_ISOLATED_ANON +
+ inc_node_page_state(page, NR_ISOLATED_ANON +
page_is_file_cache(page));
}
put_and_set:
@@ -1724,15 +1724,16 @@ static bool migrate_balanced_pgdat(struct pglist_data *pgdat,
unsigned long nr_migrate_pages)
{
int z;
+
+ if (!pgdat_reclaimable(pgdat))
+ return false;
+
for (z = pgdat->nr_zones - 1; z >= 0; z--) {
struct zone *zone = pgdat->node_zones + z;
if (!populated_zone(zone))
continue;
- if (!zone_reclaimable(zone))
- continue;
-
/* Avoid waking kswapd by allocating pages_to_migrate pages. */
if (!zone_watermark_ok(zone, 0,
high_wmark_pages(zone) +
@@ -1826,7 +1827,7 @@ static int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page)
}
page_lru = page_is_file_cache(page);
- mod_zone_page_state(page_zone(page), NR_ISOLATED_ANON + page_lru,
+ mod_node_page_state(page_zone(page)->zone_pgdat, NR_ISOLATED_ANON + page_lru,
hpage_nr_pages(page));
/*
@@ -1884,7 +1885,7 @@ int migrate_misplaced_page(struct page *page, struct vm_area_struct *vma,
if (nr_remaining) {
if (!list_empty(&migratepages)) {
list_del(&page->lru);
- dec_zone_page_state(page, NR_ISOLATED_ANON +
+ dec_node_page_state(page, NR_ISOLATED_ANON +
page_is_file_cache(page));
putback_lru_page(page);
}
@@ -1977,7 +1978,7 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm,
/* Retake the callers reference and putback on LRU */
get_page(page);
putback_lru_page(page);
- mod_zone_page_state(page_zone(page),
+ mod_node_page_state(page_zone(page)->zone_pgdat,
NR_ISOLATED_ANON + page_lru, -HPAGE_PMD_NR);
goto out_unlock;
@@ -2029,7 +2030,7 @@ int migrate_misplaced_transhuge_page(struct mm_struct *mm,
count_vm_events(PGMIGRATE_SUCCESS, HPAGE_PMD_NR);
count_vm_numa_events(NUMA_PAGE_MIGRATE, HPAGE_PMD_NR);
- mod_zone_page_state(page_zone(page),
+ mod_node_page_state(page_zone(page)->zone_pgdat,
NR_ISOLATED_ANON + page_lru,
-HPAGE_PMD_NR);
return isolated;
diff --git a/mm/mlock.c b/mm/mlock.c
index 997f63082ff5..cdd49b2e398c 100644
--- a/mm/mlock.c
+++ b/mm/mlock.c
@@ -103,7 +103,7 @@ static bool __munlock_isolate_lru_page(struct page *page, bool getpage)
if (PageLRU(page)) {
struct lruvec *lruvec;
- lruvec = mem_cgroup_page_lruvec(page, page_zone(page));
+ lruvec = mem_cgroup_page_lruvec(page, page_zone(page)->zone_pgdat);
if (getpage)
get_page(page);
ClearPageLRU(page);
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index b9956fdee8f5..3ddfa0b48436 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -285,8 +285,8 @@ static unsigned long zone_dirtyable_memory(struct zone *zone)
*/
nr_pages -= min(nr_pages, zone->totalreserve_pages);
- nr_pages += zone_page_state(zone, NR_INACTIVE_FILE);
- nr_pages += zone_page_state(zone, NR_ACTIVE_FILE);
+ nr_pages += node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE);
+ nr_pages += node_page_state(zone->zone_pgdat, NR_ACTIVE_FILE);
return nr_pages;
}
@@ -348,8 +348,8 @@ static unsigned long global_dirtyable_memory(void)
*/
x -= min(x, totalreserve_pages);
- x += global_page_state(NR_INACTIVE_FILE);
- x += global_page_state(NR_ACTIVE_FILE);
+ x += global_node_page_state(NR_INACTIVE_FILE);
+ x += global_node_page_state(NR_ACTIVE_FILE);
if (!vm_highmem_is_dirtyable)
x -= highmem_dirtyable_memory(x);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 1e0ad06c33bd..4ce578b969da 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1087,9 +1087,9 @@ static void free_pcppages_bulk(struct zone *zone, int count,
spin_lock(&zone->lock);
isolated_pageblocks = has_isolate_pageblock(zone);
- nr_scanned = zone_page_state(zone, NR_PAGES_SCANNED);
+ nr_scanned = node_page_state(zone->zone_pgdat, NR_PAGES_SCANNED);
if (nr_scanned)
- __mod_zone_page_state(zone, NR_PAGES_SCANNED, -nr_scanned);
+ __mod_node_page_state(zone->zone_pgdat, NR_PAGES_SCANNED, -nr_scanned);
while (count) {
struct page *page;
@@ -1144,9 +1144,9 @@ static void free_one_page(struct zone *zone,
{
unsigned long nr_scanned;
spin_lock(&zone->lock);
- nr_scanned = zone_page_state(zone, NR_PAGES_SCANNED);
+ nr_scanned = node_page_state(zone->zone_pgdat, NR_PAGES_SCANNED);
if (nr_scanned)
- __mod_zone_page_state(zone, NR_PAGES_SCANNED, -nr_scanned);
+ __mod_node_page_state(zone->zone_pgdat, NR_PAGES_SCANNED, -nr_scanned);
if (unlikely(has_isolate_pageblock(zone) ||
is_migrate_isolate(migratetype))) {
@@ -3486,10 +3486,19 @@ should_reclaim_retry(gfp_t gfp_mask, unsigned order,
unsigned long available;
unsigned long reclaimable;
- available = reclaimable = zone_reclaimable_pages(zone);
- available -= DIV_ROUND_UP(no_progress_loops * available,
+ /*
+ * This over-estimates the number of pages available for
+ * reclaim but walking the LRU would take too long. The
+ * consequences are that this may continue trying to
+ * reclaim for zone-constrained allocations even if those
+ * zones are already depleted.
+ */
+ reclaimable = pgdat_reclaimable_pages(zone->zone_pgdat);
+ reclaimable = min(zone->managed_pages, reclaimable);
+ available = reclaimable - DIV_ROUND_UP(no_progress_loops * reclaimable,
MAX_RECLAIM_RETRIES);
available += zone_page_state_snapshot(zone, NR_FREE_PAGES);
+ available = min(zone->managed_pages, available);
/*
* Would the allocation succeed if we reclaimed the whole
@@ -4293,6 +4302,7 @@ void show_free_areas(unsigned int filter)
unsigned long free_pcp = 0;
int cpu;
struct zone *zone;
+ pg_data_t *pgdat;
for_each_populated_zone(zone) {
if (skip_free_areas_node(filter, zone_to_nid(zone)))
@@ -4308,13 +4318,13 @@ void show_free_areas(unsigned int filter)
" slab_reclaimable:%lu slab_unreclaimable:%lu\n"
" mapped:%lu shmem:%lu pagetables:%lu bounce:%lu\n"
" free:%lu free_pcp:%lu free_cma:%lu\n",
- global_page_state(NR_ACTIVE_ANON),
- global_page_state(NR_INACTIVE_ANON),
- global_page_state(NR_ISOLATED_ANON),
- global_page_state(NR_ACTIVE_FILE),
- global_page_state(NR_INACTIVE_FILE),
- global_page_state(NR_ISOLATED_FILE),
- global_page_state(NR_UNEVICTABLE),
+ global_node_page_state(NR_ACTIVE_ANON),
+ global_node_page_state(NR_INACTIVE_ANON),
+ global_node_page_state(NR_ISOLATED_ANON),
+ global_node_page_state(NR_ACTIVE_FILE),
+ global_node_page_state(NR_INACTIVE_FILE),
+ global_node_page_state(NR_ISOLATED_FILE),
+ global_node_page_state(NR_UNEVICTABLE),
global_page_state(NR_FILE_DIRTY),
global_page_state(NR_WRITEBACK),
global_page_state(NR_UNSTABLE_NFS),
@@ -4328,6 +4338,28 @@ void show_free_areas(unsigned int filter)
free_pcp,
global_page_state(NR_FREE_CMA_PAGES));
+ for_each_online_pgdat(pgdat) {
+ printk("Node %d"
+ " active_anon:%lukB"
+ " inactive_anon:%lukB"
+ " active_file:%lukB"
+ " inactive_file:%lukB"
+ " unevictable:%lukB"
+ " isolated(anon):%lukB"
+ " isolated(file):%lukB"
+ " all_unreclaimable? %s"
+ "\n",
+ pgdat->node_id,
+ K(node_page_state(pgdat, NR_ACTIVE_ANON)),
+ K(node_page_state(pgdat, NR_INACTIVE_ANON)),
+ K(node_page_state(pgdat, NR_ACTIVE_FILE)),
+ K(node_page_state(pgdat, NR_INACTIVE_FILE)),
+ K(node_page_state(pgdat, NR_UNEVICTABLE)),
+ K(node_page_state(pgdat, NR_ISOLATED_ANON)),
+ K(node_page_state(pgdat, NR_ISOLATED_FILE)),
+ !pgdat_reclaimable(pgdat) ? "yes" : "no");
+ }
+
for_each_populated_zone(zone) {
int i;
@@ -4344,13 +4376,6 @@ void show_free_areas(unsigned int filter)
" min:%lukB"
" low:%lukB"
" high:%lukB"
- " active_anon:%lukB"
- " inactive_anon:%lukB"
- " active_file:%lukB"
- " inactive_file:%lukB"
- " unevictable:%lukB"
- " isolated(anon):%lukB"
- " isolated(file):%lukB"
" present:%lukB"
" managed:%lukB"
" mlocked:%lukB"
@@ -4368,21 +4393,13 @@ void show_free_areas(unsigned int filter)
" local_pcp:%ukB"
" free_cma:%lukB"
" writeback_tmp:%lukB"
- " pages_scanned:%lu"
- " all_unreclaimable? %s"
+ " node_pages_scanned:%lu"
"\n",
zone->name,
K(zone_page_state(zone, NR_FREE_PAGES)),
K(min_wmark_pages(zone)),
K(low_wmark_pages(zone)),
K(high_wmark_pages(zone)),
- K(zone_page_state(zone, NR_ACTIVE_ANON)),
- K(zone_page_state(zone, NR_INACTIVE_ANON)),
- K(zone_page_state(zone, NR_ACTIVE_FILE)),
- K(zone_page_state(zone, NR_INACTIVE_FILE)),
- K(zone_page_state(zone, NR_UNEVICTABLE)),
- K(zone_page_state(zone, NR_ISOLATED_ANON)),
- K(zone_page_state(zone, NR_ISOLATED_FILE)),
K(zone->present_pages),
K(zone->managed_pages),
K(zone_page_state(zone, NR_MLOCK)),
@@ -4401,9 +4418,7 @@ void show_free_areas(unsigned int filter)
K(this_cpu_read(zone->pageset->pcp.count)),
K(zone_page_state(zone, NR_FREE_CMA_PAGES)),
K(zone_page_state(zone, NR_WRITEBACK_TEMP)),
- K(zone_page_state(zone, NR_PAGES_SCANNED)),
- (!zone_reclaimable(zone) ? "yes" : "no")
- );
+ K(node_page_state(zone->zone_pgdat, NR_PAGES_SCANNED)));
printk("lowmem_reserve[]:");
for (i = 0; i < MAX_NR_ZONES; i++)
printk(" %ld", zone->lowmem_reserve[i]);
@@ -5953,7 +5968,7 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
/* For bootup, initialized properly in watermark setup */
mod_zone_page_state(zone, NR_ALLOC_BATCH, zone->managed_pages);
- lruvec_init(&zone->lruvec);
+ lruvec_init(zone_lruvec(zone));
if (!size)
continue;
diff --git a/mm/swap.c b/mm/swap.c
index 1aca06977a0a..0955396b72d0 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -63,7 +63,7 @@ static void __page_cache_release(struct page *page)
unsigned long flags;
spin_lock_irqsave(zone_lru_lock(zone), flags);
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, zone->zone_pgdat);
VM_BUG_ON_PAGE(!PageLRU(page), page);
__ClearPageLRU(page);
del_page_from_lru_list(page, lruvec, page_off_lru(page));
@@ -194,7 +194,7 @@ static void pagevec_lru_move_fn(struct pagevec *pvec,
spin_lock_irqsave(zone_lru_lock(zone), flags);
}
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, zone->zone_pgdat);
(*move_fn)(page, lruvec, arg);
}
if (zone)
@@ -314,11 +314,11 @@ static bool need_activate_page_drain(int cpu)
void activate_page(struct page *page)
{
- struct zone *zone = page_zone(page);
+ struct pglist_data *pgdat = page_zone(page)->zone_pgdat;
- spin_lock_irq(zone_lru_lock(zone));
- __activate_page(page, mem_cgroup_page_lruvec(page, zone), NULL);
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
+ __activate_page(page, mem_cgroup_page_lruvec(page, pgdat), NULL);
+ spin_unlock_irq(&pgdat->lru_lock);
}
#endif
@@ -444,16 +444,16 @@ void lru_cache_add(struct page *page)
*/
void add_page_to_unevictable_list(struct page *page)
{
- struct zone *zone = page_zone(page);
+ struct pglist_data *pgdat = page_zone(page)->zone_pgdat;
struct lruvec *lruvec;
- spin_lock_irq(zone_lru_lock(zone));
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ spin_lock_irq(&pgdat->lru_lock);
+ lruvec = mem_cgroup_page_lruvec(page, pgdat);
ClearPageActive(page);
SetPageUnevictable(page);
SetPageLRU(page);
add_page_to_lru_list(page, lruvec, LRU_UNEVICTABLE);
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_unlock_irq(&pgdat->lru_lock);
}
/**
@@ -729,7 +729,7 @@ void release_pages(struct page **pages, int nr, bool cold)
{
int i;
LIST_HEAD(pages_to_free);
- struct zone *zone = NULL;
+ struct pglist_data *pgdat = NULL;
struct lruvec *lruvec;
unsigned long uninitialized_var(flags);
unsigned int uninitialized_var(lock_batch);
@@ -740,11 +740,11 @@ void release_pages(struct page **pages, int nr, bool cold)
/*
* Make sure the IRQ-safe lock-holding time does not get
* excessive with a continuous string of pages from the
- * same zone. The lock is held only if zone != NULL.
+ * same pgdat. The lock is held only if pgdat != NULL.
*/
- if (zone && ++lock_batch == SWAP_CLUSTER_MAX) {
- spin_unlock_irqrestore(zone_lru_lock(zone), flags);
- zone = NULL;
+ if (pgdat && ++lock_batch == SWAP_CLUSTER_MAX) {
+ spin_unlock_irqrestore(&pgdat->lru_lock, flags);
+ pgdat = NULL;
}
if (is_huge_zero_page(page)) {
@@ -757,27 +757,27 @@ void release_pages(struct page **pages, int nr, bool cold)
continue;
if (PageCompound(page)) {
- if (zone) {
- spin_unlock_irqrestore(zone_lru_lock(zone), flags);
- zone = NULL;
+ if (pgdat) {
+ spin_unlock_irqrestore(&pgdat->lru_lock, flags);
+ pgdat = NULL;
}
__put_compound_page(page);
continue;
}
if (PageLRU(page)) {
- struct zone *pagezone = page_zone(page);
+ struct pglist_data *page_pgdat = page_zone(page)->zone_pgdat;
- if (pagezone != zone) {
- if (zone)
- spin_unlock_irqrestore(zone_lru_lock(zone),
+ if (page_pgdat != pgdat) {
+ if (pgdat)
+ spin_unlock_irqrestore(&pgdat->lru_lock,
flags);
lock_batch = 0;
- zone = pagezone;
- spin_lock_irqsave(zone_lru_lock(zone), flags);
+ pgdat = page_pgdat;
+ spin_lock_irqsave(&pgdat->lru_lock, flags);
}
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, pgdat);
VM_BUG_ON_PAGE(!PageLRU(page), page);
__ClearPageLRU(page);
del_page_from_lru_list(page, lruvec, page_off_lru(page));
@@ -788,8 +788,8 @@ void release_pages(struct page **pages, int nr, bool cold)
list_add(&page->lru, &pages_to_free);
}
- if (zone)
- spin_unlock_irqrestore(zone_lru_lock(zone), flags);
+ if (pgdat)
+ spin_unlock_irqrestore(&pgdat->lru_lock, flags);
mem_cgroup_uncharge_list(&pages_to_free);
free_hot_cold_page_list(&pages_to_free, cold);
@@ -825,7 +825,7 @@ void lru_add_page_tail(struct page *page, struct page *page_tail,
VM_BUG_ON_PAGE(PageCompound(page_tail), page);
VM_BUG_ON_PAGE(PageLRU(page_tail), page);
VM_BUG_ON(NR_CPUS != 1 &&
- !spin_is_locked(zone_lru_lock(lruvec_zone(lruvec))));
+ !spin_is_locked(&lruvec_pgdat(lruvec)->lru_lock));
if (!list)
SetPageLRU(page_tail);
diff --git a/mm/vmscan.c b/mm/vmscan.c
index e5c577c40ff9..f87a5a0f8793 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -191,26 +191,26 @@ static bool sane_reclaim(struct scan_control *sc)
}
#endif
-unsigned long zone_reclaimable_pages(struct zone *zone)
+unsigned long pgdat_reclaimable_pages(struct pglist_data *pgdat)
{
unsigned long nr;
- nr = zone_page_state_snapshot(zone, NR_ACTIVE_FILE) +
- zone_page_state_snapshot(zone, NR_INACTIVE_FILE) +
- zone_page_state_snapshot(zone, NR_ISOLATED_FILE);
+ nr = node_page_state_snapshot(pgdat, NR_ACTIVE_FILE) +
+ node_page_state_snapshot(pgdat, NR_INACTIVE_FILE) +
+ node_page_state_snapshot(pgdat, NR_ISOLATED_FILE);
if (get_nr_swap_pages() > 0)
- nr += zone_page_state_snapshot(zone, NR_ACTIVE_ANON) +
- zone_page_state_snapshot(zone, NR_INACTIVE_ANON) +
- zone_page_state_snapshot(zone, NR_ISOLATED_ANON);
+ nr += node_page_state_snapshot(pgdat, NR_ACTIVE_ANON) +
+ node_page_state_snapshot(pgdat, NR_INACTIVE_ANON) +
+ node_page_state_snapshot(pgdat, NR_ISOLATED_ANON);
return nr;
}
-bool zone_reclaimable(struct zone *zone)
+bool pgdat_reclaimable(struct pglist_data *pgdat)
{
- return zone_page_state_snapshot(zone, NR_PAGES_SCANNED) <
- zone_reclaimable_pages(zone) * 6;
+ return node_page_state_snapshot(pgdat, NR_PAGES_SCANNED) <
+ pgdat_reclaimable_pages(pgdat) * 6;
}
unsigned long lruvec_lru_size(struct lruvec *lruvec, enum lru_list lru)
@@ -218,7 +218,7 @@ unsigned long lruvec_lru_size(struct lruvec *lruvec, enum lru_list lru)
if (!mem_cgroup_disabled())
return mem_cgroup_get_lru_size(lruvec, lru);
- return zone_page_state(lruvec_zone(lruvec), NR_LRU_BASE + lru);
+ return node_page_state(lruvec_pgdat(lruvec), NR_LRU_BASE + lru);
}
/*
@@ -877,7 +877,7 @@ static void page_check_dirty_writeback(struct page *page,
* shrink_page_list() returns the number of reclaimed pages
*/
static unsigned long shrink_page_list(struct list_head *page_list,
- struct zone *zone,
+ struct pglist_data *pgdat,
struct scan_control *sc,
enum ttu_flags ttu_flags,
unsigned long *ret_nr_dirty,
@@ -917,7 +917,6 @@ static unsigned long shrink_page_list(struct list_head *page_list,
goto keep;
VM_BUG_ON_PAGE(PageActive(page), page);
- VM_BUG_ON_PAGE(page_zone(page) != zone, page);
sc->nr_scanned++;
@@ -996,7 +995,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
/* Case 1 above */
if (current_is_kswapd() &&
PageReclaim(page) &&
- test_bit(ZONE_WRITEBACK, &zone->flags)) {
+ test_bit(PGDAT_WRITEBACK, &pgdat->flags)) {
nr_immediate++;
goto keep_locked;
@@ -1086,7 +1085,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
*/
if (page_is_file_cache(page) &&
(!current_is_kswapd() ||
- !test_bit(ZONE_DIRTY, &zone->flags))) {
+ !test_bit(PGDAT_DIRTY, &pgdat->flags))) {
/*
* Immediately reclaim when written back.
* Similar in principal to deactivate_page()
@@ -1260,11 +1259,11 @@ unsigned long reclaim_clean_pages_from_list(struct zone *zone,
}
}
- ret = shrink_page_list(&clean_pages, zone, &sc,
+ ret = shrink_page_list(&clean_pages, zone->zone_pgdat, &sc,
TTU_UNMAP|TTU_IGNORE_ACCESS,
&dummy1, &dummy2, &dummy3, &dummy4, &dummy5, true);
list_splice(&clean_pages, page_list);
- mod_zone_page_state(zone, NR_ISOLATED_FILE, -ret);
+ mod_node_page_state(zone->zone_pgdat, NR_ISOLATED_FILE, -ret);
return ret;
}
@@ -1439,7 +1438,7 @@ int isolate_lru_page(struct page *page)
struct lruvec *lruvec;
spin_lock_irq(zone_lru_lock(zone));
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, zone->zone_pgdat);
if (PageLRU(page)) {
int lru = page_lru(page);
get_page(page);
@@ -1459,7 +1458,7 @@ int isolate_lru_page(struct page *page)
* the LRU list will go small and be scanned faster than necessary, leading to
* unnecessary swapping, thrashing and OOM.
*/
-static int too_many_isolated(struct zone *zone, int file,
+static int too_many_isolated(struct pglist_data *pgdat, int file,
struct scan_control *sc)
{
unsigned long inactive, isolated;
@@ -1471,11 +1470,11 @@ static int too_many_isolated(struct zone *zone, int file,
return 0;
if (file) {
- inactive = zone_page_state(zone, NR_INACTIVE_FILE);
- isolated = zone_page_state(zone, NR_ISOLATED_FILE);
+ inactive = node_page_state(pgdat, NR_INACTIVE_FILE);
+ isolated = node_page_state(pgdat, NR_ISOLATED_FILE);
} else {
- inactive = zone_page_state(zone, NR_INACTIVE_ANON);
- isolated = zone_page_state(zone, NR_ISOLATED_ANON);
+ inactive = node_page_state(pgdat, NR_INACTIVE_ANON);
+ isolated = node_page_state(pgdat, NR_ISOLATED_ANON);
}
/*
@@ -1493,7 +1492,7 @@ static noinline_for_stack void
putback_inactive_pages(struct lruvec *lruvec, struct list_head *page_list)
{
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat;
- struct zone *zone = lruvec_zone(lruvec);
+ struct pglist_data *pgdat = lruvec_pgdat(lruvec);
LIST_HEAD(pages_to_free);
/*
@@ -1506,13 +1505,13 @@ putback_inactive_pages(struct lruvec *lruvec, struct list_head *page_list)
VM_BUG_ON_PAGE(PageLRU(page), page);
list_del(&page->lru);
if (unlikely(!page_evictable(page))) {
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_unlock_irq(&pgdat->lru_lock);
putback_lru_page(page);
- spin_lock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
continue;
}
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, pgdat);
SetPageLRU(page);
lru = page_lru(page);
@@ -1529,10 +1528,10 @@ putback_inactive_pages(struct lruvec *lruvec, struct list_head *page_list)
del_page_from_lru_list(page, lruvec, lru);
if (unlikely(PageCompound(page))) {
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_unlock_irq(&pgdat->lru_lock);
mem_cgroup_uncharge(page);
(*get_compound_page_dtor(page))(page);
- spin_lock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
} else
list_add(&page->lru, &pages_to_free);
}
@@ -1576,10 +1575,10 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
unsigned long nr_immediate = 0;
isolate_mode_t isolate_mode = 0;
int file = is_file_lru(lru);
- struct zone *zone = lruvec_zone(lruvec);
+ struct pglist_data *pgdat = lruvec_pgdat(lruvec);
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat;
- while (unlikely(too_many_isolated(zone, file, sc))) {
+ while (unlikely(too_many_isolated(pgdat, file, sc))) {
congestion_wait(BLK_RW_ASYNC, HZ/10);
/* We are about to die and free our memory. Return now. */
@@ -1594,48 +1593,46 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
if (!sc->may_writepage)
isolate_mode |= ISOLATE_CLEAN;
- spin_lock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &page_list,
&nr_scanned, sc, isolate_mode, lru);
update_lru_size(lruvec, lru, -nr_taken);
- __mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken);
+ __mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, nr_taken);
reclaim_stat->recent_scanned[file] += nr_taken;
if (global_reclaim(sc)) {
- __mod_zone_page_state(zone, NR_PAGES_SCANNED, nr_scanned);
+ __mod_node_page_state(pgdat, NR_PAGES_SCANNED, nr_scanned);
if (current_is_kswapd())
- __count_zone_vm_events(PGSCAN_KSWAPD, zone, nr_scanned);
+ __count_vm_events(PGSCAN_KSWAPD, nr_scanned);
else
- __count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scanned);
+ __count_vm_events(PGSCAN_DIRECT, nr_scanned);
}
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_unlock_irq(&pgdat->lru_lock);
if (nr_taken == 0)
return 0;
- nr_reclaimed = shrink_page_list(&page_list, zone, sc, TTU_UNMAP,
+ nr_reclaimed = shrink_page_list(&page_list, pgdat, sc, TTU_UNMAP,
&nr_dirty, &nr_unqueued_dirty, &nr_congested,
&nr_writeback, &nr_immediate,
false);
- spin_lock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
if (global_reclaim(sc)) {
if (current_is_kswapd())
- __count_zone_vm_events(PGSTEAL_KSWAPD, zone,
- nr_reclaimed);
+ __count_vm_events(PGSTEAL_KSWAPD, nr_reclaimed);
else
- __count_zone_vm_events(PGSTEAL_DIRECT, zone,
- nr_reclaimed);
+ __count_vm_events(PGSTEAL_DIRECT, nr_reclaimed);
}
putback_inactive_pages(lruvec, &page_list);
- __mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken);
+ __mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, -nr_taken);
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_unlock_irq(&pgdat->lru_lock);
mem_cgroup_uncharge_list(&page_list);
free_hot_cold_page_list(&page_list, true);
@@ -1655,7 +1652,7 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
* are encountered in the nr_immediate check below.
*/
if (nr_writeback && nr_writeback == nr_taken)
- set_bit(ZONE_WRITEBACK, &zone->flags);
+ set_bit(PGDAT_WRITEBACK, &pgdat->flags);
/*
* Legacy memcg will stall in page writeback so avoid forcibly
@@ -1667,16 +1664,16 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
* backed by a congested BDI and wait_iff_congested will stall.
*/
if (nr_dirty && nr_dirty == nr_congested)
- set_bit(ZONE_CONGESTED, &zone->flags);
+ set_bit(PGDAT_CONGESTED, &pgdat->flags);
/*
* If dirty pages are scanned that are not queued for IO, it
* implies that flushers are not keeping up. In this case, flag
- * the zone ZONE_DIRTY and kswapd will start writing pages from
+ * the pgdat PGDAT_DIRTY and kswapd will start writing pages from
* reclaim context.
*/
if (nr_unqueued_dirty == nr_taken)
- set_bit(ZONE_DIRTY, &zone->flags);
+ set_bit(PGDAT_DIRTY, &pgdat->flags);
/*
* If kswapd scans pages marked marked for immediate
@@ -1695,9 +1692,10 @@ shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec,
*/
if (!sc->hibernation_mode && !current_is_kswapd() &&
current_may_throttle())
- wait_iff_congested(zone, BLK_RW_ASYNC, HZ/10);
+ wait_iff_congested(pgdat, BLK_RW_ASYNC, HZ/10);
- trace_mm_vmscan_lru_shrink_inactive(zone, nr_scanned, nr_reclaimed,
+ trace_mm_vmscan_lru_shrink_inactive(pgdat->node_id,
+ nr_scanned, nr_reclaimed,
sc->priority, file);
return nr_reclaimed;
}
@@ -1725,14 +1723,14 @@ static void move_active_pages_to_lru(struct lruvec *lruvec,
struct list_head *pages_to_free,
enum lru_list lru)
{
- struct zone *zone = lruvec_zone(lruvec);
+ struct pglist_data *pgdat = lruvec_pgdat(lruvec);
unsigned long pgmoved = 0;
struct page *page;
int nr_pages;
while (!list_empty(list)) {
page = lru_to_page(list);
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, pgdat);
VM_BUG_ON_PAGE(PageLRU(page), page);
SetPageLRU(page);
@@ -1748,10 +1746,10 @@ static void move_active_pages_to_lru(struct lruvec *lruvec,
del_page_from_lru_list(page, lruvec, lru);
if (unlikely(PageCompound(page))) {
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_unlock_irq(&pgdat->lru_lock);
mem_cgroup_uncharge(page);
(*get_compound_page_dtor(page))(page);
- spin_lock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
} else
list_add(&page->lru, pages_to_free);
}
@@ -1777,7 +1775,7 @@ static void shrink_active_list(unsigned long nr_to_scan,
unsigned long nr_rotated = 0;
isolate_mode_t isolate_mode = 0;
int file = is_file_lru(lru);
- struct zone *zone = lruvec_zone(lruvec);
+ struct pglist_data *pgdat = lruvec_pgdat(lruvec);
lru_add_drain();
@@ -1786,20 +1784,20 @@ static void shrink_active_list(unsigned long nr_to_scan,
if (!sc->may_writepage)
isolate_mode |= ISOLATE_CLEAN;
- spin_lock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &l_hold,
&nr_scanned, sc, isolate_mode, lru);
update_lru_size(lruvec, lru, -nr_taken);
- __mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken);
+ __mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, nr_taken);
reclaim_stat->recent_scanned[file] += nr_taken;
if (global_reclaim(sc))
- __mod_zone_page_state(zone, NR_PAGES_SCANNED, nr_scanned);
- __count_zone_vm_events(PGREFILL, zone, nr_scanned);
+ __mod_node_page_state(pgdat, NR_PAGES_SCANNED, nr_scanned);
+ __count_vm_events(PGREFILL, nr_scanned);
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_unlock_irq(&pgdat->lru_lock);
while (!list_empty(&l_hold)) {
cond_resched();
@@ -1844,7 +1842,7 @@ static void shrink_active_list(unsigned long nr_to_scan,
/*
* Move pages back to the lru list.
*/
- spin_lock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
/*
* Count referenced pages from currently used mappings as rotated,
* even though only some of them are actually re-activated. This
@@ -1855,8 +1853,8 @@ static void shrink_active_list(unsigned long nr_to_scan,
move_active_pages_to_lru(lruvec, &l_active, &l_hold, lru);
move_active_pages_to_lru(lruvec, &l_inactive, &l_hold, lru - LRU_ACTIVE);
- __mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken);
- spin_unlock_irq(zone_lru_lock(zone));
+ __mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, -nr_taken);
+ spin_unlock_irq(&pgdat->lru_lock);
mem_cgroup_uncharge_list(&l_hold);
free_hot_cold_page_list(&l_hold, true);
@@ -1950,7 +1948,7 @@ static void get_scan_count(struct lruvec *lruvec, struct mem_cgroup *memcg,
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat;
u64 fraction[2];
u64 denominator = 0; /* gcc */
- struct zone *zone = lruvec_zone(lruvec);
+ struct pglist_data *pgdat = lruvec_pgdat(lruvec);
unsigned long anon_prio, file_prio;
enum scan_balance scan_balance;
unsigned long anon, file;
@@ -1971,7 +1969,7 @@ static void get_scan_count(struct lruvec *lruvec, struct mem_cgroup *memcg,
* well.
*/
if (current_is_kswapd()) {
- if (!zone_reclaimable(zone))
+ if (!pgdat_reclaimable(pgdat))
force_scan = true;
if (!mem_cgroup_online(memcg))
force_scan = true;
@@ -2017,14 +2015,24 @@ static void get_scan_count(struct lruvec *lruvec, struct mem_cgroup *memcg,
* anon pages. Try to detect this based on file LRU size.
*/
if (global_reclaim(sc)) {
- unsigned long zonefile;
- unsigned long zonefree;
+ unsigned long pgdatfile;
+ unsigned long pgdatfree;
+ int z;
+ unsigned long total_high_wmark = 0;
+
+ pgdatfree = sum_zone_node_page_state(pgdat->node_id, NR_FREE_PAGES);
+ pgdatfile = node_page_state(pgdat, NR_ACTIVE_FILE) +
+ node_page_state(pgdat, NR_INACTIVE_FILE);
- zonefree = zone_page_state(zone, NR_FREE_PAGES);
- zonefile = zone_page_state(zone, NR_ACTIVE_FILE) +
- zone_page_state(zone, NR_INACTIVE_FILE);
+ for (z = 0; z < MAX_NR_ZONES; z++) {
+ struct zone *zone = &pgdat->node_zones[z];
+ if (!populated_zone(zone))
+ continue;
+
+ total_high_wmark += high_wmark_pages(zone);
+ }
- if (unlikely(zonefile + zonefree <= high_wmark_pages(zone))) {
+ if (unlikely(pgdatfile + pgdatfree <= total_high_wmark)) {
scan_balance = SCAN_ANON;
goto out;
}
@@ -2071,7 +2079,7 @@ static void get_scan_count(struct lruvec *lruvec, struct mem_cgroup *memcg,
file = lruvec_lru_size(lruvec, LRU_ACTIVE_FILE) +
lruvec_lru_size(lruvec, LRU_INACTIVE_FILE);
- spin_lock_irq(zone_lru_lock(zone));
+ spin_lock_irq(&pgdat->lru_lock);
if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) {
reclaim_stat->recent_scanned[0] /= 2;
reclaim_stat->recent_rotated[0] /= 2;
@@ -2092,7 +2100,7 @@ static void get_scan_count(struct lruvec *lruvec, struct mem_cgroup *memcg,
fp = file_prio * (reclaim_stat->recent_scanned[1] + 1);
fp /= reclaim_stat->recent_rotated[1] + 1;
- spin_unlock_irq(zone_lru_lock(zone));
+ spin_unlock_irq(&pgdat->lru_lock);
fraction[0] = ap;
fraction[1] = fp;
@@ -2346,9 +2354,9 @@ static inline bool should_continue_reclaim(struct zone *zone,
* inactive lists are large enough, continue reclaiming
*/
pages_for_compaction = (2UL << sc->order);
- inactive_lru_pages = zone_page_state(zone, NR_INACTIVE_FILE);
+ inactive_lru_pages = node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE);
if (get_nr_swap_pages() > 0)
- inactive_lru_pages += zone_page_state(zone, NR_INACTIVE_ANON);
+ inactive_lru_pages += node_page_state(zone->zone_pgdat, NR_INACTIVE_ANON);
if (sc->nr_reclaimed < pages_for_compaction &&
inactive_lru_pages > pages_for_compaction)
return true;
@@ -2548,7 +2556,7 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
continue;
if (sc->priority != DEF_PRIORITY &&
- !zone_reclaimable(zone))
+ !pgdat_reclaimable(zone->zone_pgdat))
continue; /* Let kswapd poll it */
/*
@@ -2686,7 +2694,7 @@ static bool pfmemalloc_watermark_ok(pg_data_t *pgdat)
for (i = 0; i <= ZONE_NORMAL; i++) {
zone = &pgdat->node_zones[i];
if (!populated_zone(zone) ||
- zone_reclaimable_pages(zone) == 0)
+ pgdat_reclaimable_pages(pgdat) == 0)
continue;
pfmemalloc_reserve += min_wmark_pages(zone);
@@ -2994,7 +3002,7 @@ static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
* DEF_PRIORITY. Effectively, it considers them balanced so
* they must be considered balanced here as well!
*/
- if (!zone_reclaimable(zone)) {
+ if (!pgdat_reclaimable(zone->zone_pgdat)) {
balanced_pages += zone->managed_pages;
continue;
}
@@ -3057,6 +3065,7 @@ static bool kswapd_shrink_zone(struct zone *zone,
{
unsigned long balance_gap;
bool lowmem_pressure;
+ struct pglist_data *pgdat = zone->zone_pgdat;
/* Reclaim above the high watermark. */
sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone));
@@ -3081,7 +3090,8 @@ static bool kswapd_shrink_zone(struct zone *zone,
shrink_zone(zone, sc, zone_idx(zone) == classzone_idx);
- clear_bit(ZONE_WRITEBACK, &zone->flags);
+ /* TODO: ANOMALY */
+ clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
/*
* If a zone reaches its high watermark, consider it to be no longer
@@ -3089,10 +3099,10 @@ static bool kswapd_shrink_zone(struct zone *zone,
* BDIs but as pressure is relieved, speculatively avoid congestion
* waits.
*/
- if (zone_reclaimable(zone) &&
+ if (pgdat_reclaimable(zone->zone_pgdat) &&
zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
- clear_bit(ZONE_CONGESTED, &zone->flags);
- clear_bit(ZONE_DIRTY, &zone->flags);
+ clear_bit(PGDAT_CONGESTED, &pgdat->flags);
+ clear_bit(PGDAT_DIRTY, &pgdat->flags);
}
return sc->nr_scanned >= sc->nr_to_reclaim;
@@ -3151,7 +3161,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
continue;
if (sc.priority != DEF_PRIORITY &&
- !zone_reclaimable(zone))
+ !pgdat_reclaimable(zone->zone_pgdat))
continue;
/*
@@ -3178,9 +3188,11 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
/*
* If balanced, clear the dirty and congested
* flags
+ *
+ * TODO: ANOMALY
*/
- clear_bit(ZONE_CONGESTED, &zone->flags);
- clear_bit(ZONE_DIRTY, &zone->flags);
+ clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
+ clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
}
}
@@ -3210,7 +3222,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
continue;
if (sc.priority != DEF_PRIORITY &&
- !zone_reclaimable(zone))
+ !pgdat_reclaimable(zone->zone_pgdat))
continue;
sc.nr_scanned = 0;
@@ -3606,8 +3618,8 @@ int sysctl_min_slab_ratio = 5;
static inline unsigned long zone_unmapped_file_pages(struct zone *zone)
{
unsigned long file_mapped = zone_page_state(zone, NR_FILE_MAPPED);
- unsigned long file_lru = zone_page_state(zone, NR_INACTIVE_FILE) +
- zone_page_state(zone, NR_ACTIVE_FILE);
+ unsigned long file_lru = node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE) +
+ node_page_state(zone->zone_pgdat, NR_ACTIVE_FILE);
/*
* It's possible for there to be more file mapped pages than
@@ -3710,7 +3722,7 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
zone_page_state(zone, NR_SLAB_RECLAIMABLE) <= zone->min_slab_pages)
return ZONE_RECLAIM_FULL;
- if (!zone_reclaimable(zone))
+ if (!pgdat_reclaimable(zone->zone_pgdat))
return ZONE_RECLAIM_FULL;
/*
@@ -3789,7 +3801,7 @@ void check_move_unevictable_pages(struct page **pages, int nr_pages)
zone = pagezone;
spin_lock_irq(zone_lru_lock(zone));
}
- lruvec = mem_cgroup_page_lruvec(page, zone);
+ lruvec = mem_cgroup_page_lruvec(page, zone->zone_pgdat);
if (!PageLRU(page) || !PageUnevictable(page))
continue;
diff --git a/mm/vmstat.c b/mm/vmstat.c
index e1d0deeec98e..1b1ea16be727 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -935,11 +935,6 @@ const char * const vmstat_text[] = {
/* enum zone_stat_item countes */
"nr_free_pages",
"nr_alloc_batch",
- "nr_inactive_anon",
- "nr_active_anon",
- "nr_inactive_file",
- "nr_active_file",
- "nr_unevictable",
"nr_mlock",
"nr_anon_pages",
"nr_mapped",
@@ -955,12 +950,9 @@ const char * const vmstat_text[] = {
"nr_vmscan_write",
"nr_vmscan_immediate_reclaim",
"nr_writeback_temp",
- "nr_isolated_anon",
- "nr_isolated_file",
"nr_shmem",
"nr_dirtied",
"nr_written",
- "nr_pages_scanned",
#if IS_ENABLED(CONFIG_ZSMALLOC)
"nr_zsmalloc",
#endif
@@ -978,6 +970,16 @@ const char * const vmstat_text[] = {
"nr_anon_transparent_hugepages",
"nr_free_cma",
+ /* Node-based counters */
+ "nr_inactive_anon",
+ "nr_active_anon",
+ "nr_inactive_file",
+ "nr_active_file",
+ "nr_unevictable",
+ "nr_isolated_anon",
+ "nr_isolated_file",
+ "nr_pages_scanned",
+
/* enum writeback_stat_item counters */
"nr_dirty_threshold",
"nr_dirty_background_threshold",
@@ -999,11 +1001,11 @@ const char * const vmstat_text[] = {
"pgmajfault",
"pglazyfreed",
- TEXTS_FOR_ZONES("pgrefill")
- TEXTS_FOR_ZONES("pgsteal_kswapd")
- TEXTS_FOR_ZONES("pgsteal_direct")
- TEXTS_FOR_ZONES("pgscan_kswapd")
- TEXTS_FOR_ZONES("pgscan_direct")
+ "pgrefill",
+ "pgsteal_kswapd",
+ "pgsteal_direct",
+ "pgscan_kswapd",
+ "pgscan_direct",
"pgscan_direct_throttle",
#ifdef CONFIG_NUMA
@@ -1429,7 +1431,7 @@ static void zoneinfo_show_print(struct seq_file *m, pg_data_t *pgdat,
"\n min %lu"
"\n low %lu"
"\n high %lu"
- "\n scanned %lu"
+ "\n node_scanned %lu"
"\n spanned %lu"
"\n present %lu"
"\n managed %lu",
@@ -1437,13 +1439,13 @@ static void zoneinfo_show_print(struct seq_file *m, pg_data_t *pgdat,
min_wmark_pages(zone),
low_wmark_pages(zone),
high_wmark_pages(zone),
- zone_page_state(zone, NR_PAGES_SCANNED),
+ node_page_state(zone->zone_pgdat, NR_PAGES_SCANNED),
zone->spanned_pages,
zone->present_pages,
zone->managed_pages);
for (i = 0; i < NR_VM_ZONE_STAT_ITEMS; i++)
- seq_printf(m, "\n %-12s %lu", vmstat_text[i],
+ seq_printf(m, "\n %-12s %lu", vmstat_text[i],
zone_page_state(zone, i));
seq_printf(m,
@@ -1473,12 +1475,12 @@ static void zoneinfo_show_print(struct seq_file *m, pg_data_t *pgdat,
#endif
}
seq_printf(m,
- "\n all_unreclaimable: %u"
- "\n start_pfn: %lu"
- "\n inactive_ratio: %u",
- !zone_reclaimable(zone),
+ "\n node_unreclaimable: %u"
+ "\n start_pfn: %lu"
+ "\n node_inactive_ratio: %u",
+ !pgdat_reclaimable(zone->zone_pgdat),
zone->zone_start_pfn,
- zone->inactive_ratio);
+ zone->zone_pgdat->inactive_ratio);
seq_putc(m, '\n');
}
@@ -1569,7 +1571,6 @@ static int vmstat_show(struct seq_file *m, void *arg)
{
unsigned long *l = arg;
unsigned long off = l - (unsigned long *)m->private;
-
seq_printf(m, "%s %lu\n", vmstat_text[off], *l);
return 0;
}
diff --git a/mm/workingset.c b/mm/workingset.c
index ac36efa8c754..c0820e06aaff 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -353,8 +353,8 @@ static unsigned long count_shadow_nodes(struct shrinker *shrinker,
pages = mem_cgroup_node_nr_lru_pages(sc->memcg, sc->nid,
LRU_ALL_FILE);
} else {
- pages = sum_zone_node_page_state(sc->nid, NR_ACTIVE_FILE) +
- sum_zone_node_page_state(sc->nid, NR_INACTIVE_FILE);
+ pages = node_page_state(NODE_DATA(sc->nid), NR_ACTIVE_FILE) +
+ node_page_state(NODE_DATA(sc->nid), NR_INACTIVE_FILE);
}
/*
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 03/27] mm, vmscan: Move LRU lists to node
2016-06-09 18:04 ` [PATCH 03/27] mm, vmscan: Move LRU lists to node Mel Gorman
@ 2016-06-10 18:00 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-10 18:00 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM
Cc: Rik van Riel, Johannes Weiner, LKML, Michal Hocko
[+CC Michal Hocko]
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> This moves the LRU lists from the zone to the node and all related data
> such as counters, tracing, congestion tracking and writeback tracking.
> This is mostly a mechanical patch but note that it introduces a number
> of anomalies. For example, the scans are per-zone but using per-node
> counters. We also mark a node as congested when a zone is congested. This
> causes weird problems that are fixed later but is easier to review.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> Acked-by: Johannes Weiner <hannes@cmpxchg.org>
> @@ -535,17 +525,21 @@ struct zone {
>
> enum zone_flags {
> ZONE_RECLAIM_LOCKED, /* prevents concurrent reclaim */
> - ZONE_CONGESTED, /* zone has many dirty pages backed by
> + ZONE_OOM_LOCKED, /* zone is in OOM killer zonelist */
This one has been zapped recently, looks like rebasing resurrected it.
> @@ -1455,13 +1455,22 @@ bool compaction_zonelist_suitable(struct alloc_context *ac, int order,
> enum compact_result compact_result;
>
> /*
> + * This over-estimates the number of pages available for
> + * reclaim/compaction but walking the LRU would take too
> + * long. The consequences are that compaction may retry
> + * longer than it should for a zone-constrained allocation
> + * request.
> + */
> + available = pgdat_reclaimable_pages(zone->zone_pgdat);
I'm worried if "longer than it should" means "potentially forever", as
the limit on retries in should_compact_retry() doesn't apply when this
function returns true. Unless some later patches change that.
I'm starting to wonder if it's a good idea to give up per-zone LRU
accounting, because we still have per-zone watermarks that we are trying
to satisfy. How will we even recognize situation where a small zone is
so depleted of LRU pages that it can't even reach its watermarks,
causing a massive whole-node reclaim? Couldn't we have a combination of
per-node lru with per-zone accounting?
> +
> + /*
> * Do not consider all the reclaimable memory because we do not
> * want to trash just for a single high order allocation which
> * is even not guaranteed to appear even if __compaction_suitable
> * is happy about the watermark check.
> */
> - available = zone_reclaimable_pages(zone) / order;
This removed the scaling by order. Accidentally I guess, as the comment
is still there.
> available += zone_page_state_snapshot(zone, NR_FREE_PAGES);
> + available = min(zone->managed_pages, available);
> compact_result = __compaction_suitable(zone, order, alloc_flags,
> ac_classzone_idx(ac), available);
> if (compact_result != COMPACT_SKIPPED &&
[...]
> @@ -1826,7 +1827,7 @@ static int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page)
> }
>
> page_lru = page_is_file_cache(page);
> - mod_zone_page_state(page_zone(page), NR_ISOLATED_ANON + page_lru,
> + mod_node_page_state(page_zone(page)->zone_pgdat, NR_ISOLATED_ANON + page_lru,
This again, I won't point out further. But I think a page_node() (or
page_pgdat()?) function is called for?
> @@ -3486,10 +3486,19 @@ should_reclaim_retry(gfp_t gfp_mask, unsigned order,
> unsigned long available;
> unsigned long reclaimable;
>
> - available = reclaimable = zone_reclaimable_pages(zone);
> - available -= DIV_ROUND_UP(no_progress_loops * available,
> + /*
> + * This over-estimates the number of pages available for
> + * reclaim but walking the LRU would take too long. The
> + * consequences are that this may continue trying to
> + * reclaim for zone-constrained allocations even if those
> + * zones are already depleted.
> + */
> + reclaimable = pgdat_reclaimable_pages(zone->zone_pgdat);
> + reclaimable = min(zone->managed_pages, reclaimable);
> + available = reclaimable - DIV_ROUND_UP(no_progress_loops * reclaimable,
> MAX_RECLAIM_RETRIES);
> available += zone_page_state_snapshot(zone, NR_FREE_PAGES);
> + available = min(zone->managed_pages, available);
>
> /*
> * Would the allocation succeed if we reclaimed the whole
This adds to my worries about per-node LRU accounting :/
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 04/27] mm, vmscan: Begin reclaiming pages on a per-node basis
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (2 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 03/27] mm, vmscan: Move LRU lists to node Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-15 12:52 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 05/27] mm, vmscan: Have kswapd only scan based on the highest requested zone Mel Gorman
` (22 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
This patch makes reclaim decisions on a per-node basis. A reclaimer knows
what zone is required by the allocation request and skips pages from
higher zones. In many cases this will be ok because it's a GFP_HIGHMEM
request of some description. On 64-bit, ZONE_DMA32 requests will cause
some problems but 32-bit devices on 64-bit platforms are increasingly
rare. Historically it would have been a major problem on 32-bit with big
Highmem:Lowmem ratios but such configurations are also now rare and even
where they exist, they are not encouraged. If it really becomes a problem,
it'll manifest as very low reclaim efficiencies.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/vmscan.c | 72 ++++++++++++++++++++++++++++++++++++++++---------------------
1 file changed, 47 insertions(+), 25 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index f87a5a0f8793..ab1b28e7e20a 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -84,6 +84,9 @@ struct scan_control {
/* Scan (total_size >> priority) pages at once */
int priority;
+ /* The highest zone to isolate pages for reclaim from */
+ enum zone_type reclaim_idx;
+
unsigned int may_writepage:1;
/* Can mapped pages be reclaimed? */
@@ -1369,6 +1372,7 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
struct list_head *src = &lruvec->lists[lru];
unsigned long nr_taken = 0;
unsigned long scan;
+ LIST_HEAD(pages_skipped);
for (scan = 0; scan < nr_to_scan && nr_taken < nr_to_scan &&
!list_empty(src); scan++) {
@@ -1379,6 +1383,11 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
VM_BUG_ON_PAGE(!PageLRU(page), page);
+ if (page_zonenum(page) > sc->reclaim_idx) {
+ list_move(&page->lru, &pages_skipped);
+ continue;
+ }
+
switch (__isolate_lru_page(page, mode)) {
case 0:
nr_taken += hpage_nr_pages(page);
@@ -1395,6 +1404,15 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
}
}
+ /*
+ * Splice any skipped pages to the start of the LRU list. Note that
+ * this disrupts the LRU order when reclaiming for lower zones but
+ * we cannot splice to the tail. If we did then the SWAP_CLUSTER_MAX
+ * scanning would soon rescan the same pages to skip and put the
+ * system at risk of premature OOM.
+ */
+ if (!list_empty(&pages_skipped))
+ list_splice(&pages_skipped, src);
*nr_scanned = scan;
trace_mm_vmscan_lru_isolate(sc->order, nr_to_scan, scan,
nr_taken, mode, is_file_lru(lru));
@@ -1557,7 +1575,7 @@ static int current_may_throttle(void)
}
/*
- * shrink_inactive_list() is a helper for shrink_zone(). It returns the number
+ * shrink_inactive_list() is a helper for shrink_node(). It returns the number
* of reclaimed pages
*/
static noinline_for_stack unsigned long
@@ -2371,12 +2389,13 @@ static inline bool should_continue_reclaim(struct zone *zone,
}
}
-static bool shrink_zone(struct zone *zone, struct scan_control *sc,
- bool is_classzone)
+static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
+ enum zone_type classzone_idx)
{
struct reclaim_state *reclaim_state = current->reclaim_state;
unsigned long nr_reclaimed, nr_scanned;
bool reclaimable = false;
+ struct zone *zone = &pgdat->node_zones[classzone_idx];
do {
struct mem_cgroup *root = sc->target_mem_cgroup;
@@ -2408,7 +2427,7 @@ static bool shrink_zone(struct zone *zone, struct scan_control *sc,
shrink_zone_memcg(zone, memcg, sc, &lru_pages);
zone_lru_pages += lru_pages;
- if (memcg && is_classzone)
+ if (!global_reclaim(sc) && sc->reclaim_idx == classzone_idx)
shrink_slab(sc->gfp_mask, zone_to_nid(zone),
memcg, sc->nr_scanned - scanned,
lru_pages);
@@ -2439,7 +2458,7 @@ static bool shrink_zone(struct zone *zone, struct scan_control *sc,
* Shrink the slab caches in the same proportion that
* the eligible LRU pages were scanned.
*/
- if (global_reclaim(sc) && is_classzone)
+ if (global_reclaim(sc) && sc->reclaim_idx == classzone_idx)
shrink_slab(sc->gfp_mask, zone_to_nid(zone), NULL,
sc->nr_scanned - nr_scanned,
zone_lru_pages);
@@ -2516,14 +2535,14 @@ static inline bool compaction_ready(struct zone *zone, int order, int classzone_
* If a zone is deemed to be full of pinned pages then just give it a light
* scan then give up on it.
*/
-static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
+static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc,
+ enum zone_type classzone_idx)
{
struct zoneref *z;
struct zone *zone;
unsigned long nr_soft_reclaimed;
unsigned long nr_soft_scanned;
gfp_t orig_mask;
- enum zone_type requested_highidx = gfp_zone(sc->gfp_mask);
/*
* If the number of buffer_heads in the machine exceeds the maximum
@@ -2536,15 +2555,15 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
for_each_zone_zonelist_nodemask(zone, z, zonelist,
gfp_zone(sc->gfp_mask), sc->nodemask) {
- enum zone_type classzone_idx;
-
if (!populated_zone(zone))
continue;
- classzone_idx = requested_highidx;
while (!populated_zone(zone->zone_pgdat->node_zones +
- classzone_idx))
+ classzone_idx)) {
+ sc->reclaim_idx--;
classzone_idx--;
+ continue;
+ }
/*
* Take care memory controller reclaiming has small influence
@@ -2570,8 +2589,8 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
*/
if (IS_ENABLED(CONFIG_COMPACTION) &&
sc->order > PAGE_ALLOC_COSTLY_ORDER &&
- zonelist_zone_idx(z) <= requested_highidx &&
- compaction_ready(zone, sc->order, requested_highidx)) {
+ zonelist_zone_idx(z) <= classzone_idx &&
+ compaction_ready(zone, sc->order, classzone_idx)) {
sc->compaction_ready = true;
continue;
}
@@ -2591,7 +2610,7 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
/* need some check for avoid more shrink_zone() */
}
- shrink_zone(zone, sc, zone_idx(zone) == classzone_idx);
+ shrink_node(zone->zone_pgdat, sc, classzone_idx);
}
/*
@@ -2623,6 +2642,7 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist,
int initial_priority = sc->priority;
unsigned long total_scanned = 0;
unsigned long writeback_threshold;
+ enum zone_type classzone_idx = gfp_zone(sc->gfp_mask);
retry:
delayacct_freepages_start();
@@ -2633,7 +2653,7 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist,
vmpressure_prio(sc->gfp_mask, sc->target_mem_cgroup,
sc->priority);
sc->nr_scanned = 0;
- shrink_zones(zonelist, sc);
+ shrink_zones(zonelist, sc, classzone_idx);
total_scanned += sc->nr_scanned;
if (sc->nr_reclaimed >= sc->nr_to_reclaim)
@@ -3088,7 +3108,7 @@ static bool kswapd_shrink_zone(struct zone *zone,
balance_gap, classzone_idx))
return true;
- shrink_zone(zone, sc, zone_idx(zone) == classzone_idx);
+ shrink_node(zone->zone_pgdat, sc, classzone_idx);
/* TODO: ANOMALY */
clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
@@ -3137,6 +3157,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
unsigned long nr_soft_scanned;
struct scan_control sc = {
.gfp_mask = GFP_KERNEL,
+ .reclaim_idx = MAX_NR_ZONES - 1,
.order = order,
.priority = DEF_PRIORITY,
.may_writepage = !laptop_mode,
@@ -3207,15 +3228,14 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
sc.may_writepage = 1;
/*
- * Now scan the zone in the dma->highmem direction, stopping
- * at the last zone which needs scanning.
- *
- * We do this because the page allocator works in the opposite
- * direction. This prevents the page allocator from allocating
- * pages behind kswapd's direction of progress, which would
- * cause too much scanning of the lower zones.
+ * Continue scanning in the highmem->dma direction stopping at
+ * the last zone which needs scanning. This may reclaim lowmem
+ * pages that are not necessary for zone balancing but it
+ * preserves LRU ordering. It is assumed that the bulk of
+ * allocation requests can use arbitrary zones with the
+ * possible exception of big highmem:lowmem configurations.
*/
- for (i = 0; i <= end_zone; i++) {
+ for (i = end_zone; i >= end_zone; i--) {
struct zone *zone = pgdat->node_zones + i;
if (!populated_zone(zone))
@@ -3226,6 +3246,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
continue;
sc.nr_scanned = 0;
+ sc.reclaim_idx = i;
nr_soft_scanned = 0;
/*
@@ -3674,6 +3695,7 @@ static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
.may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE),
.may_unmap = !!(zone_reclaim_mode & RECLAIM_UNMAP),
.may_swap = 1,
+ .reclaim_idx = zone_idx(zone),
};
cond_resched();
@@ -3693,7 +3715,7 @@ static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
* priorities until we have enough memory freed.
*/
do {
- shrink_zone(zone, &sc, true);
+ shrink_node(zone->zone_pgdat, &sc, zone_idx(zone));
} while (sc.nr_reclaimed < nr_pages && --sc.priority >= 0);
}
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 04/27] mm, vmscan: Begin reclaiming pages on a per-node basis
2016-06-09 18:04 ` [PATCH 04/27] mm, vmscan: Begin reclaiming pages on a per-node basis Mel Gorman
@ 2016-06-15 12:52 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-15 12:52 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> This patch makes reclaim decisions on a per-node basis. A reclaimer knows
> what zone is required by the allocation request and skips pages from
> higher zones. In many cases this will be ok because it's a GFP_HIGHMEM
> request of some description. On 64-bit, ZONE_DMA32 requests will cause
> some problems but 32-bit devices on 64-bit platforms are increasingly
> rare. Historically it would have been a major problem on 32-bit with big
> Highmem:Lowmem ratios but such configurations are also now rare and even
> where they exist, they are not encouraged. If it really becomes a problem,
> it'll manifest as very low reclaim efficiencies.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> ---
> mm/vmscan.c | 72 ++++++++++++++++++++++++++++++++++++++++---------------------
> 1 file changed, 47 insertions(+), 25 deletions(-)
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index f87a5a0f8793..ab1b28e7e20a 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -84,6 +84,9 @@ struct scan_control {
> /* Scan (total_size >> priority) pages at once */
> int priority;
>
> + /* The highest zone to isolate pages for reclaim from */
> + enum zone_type reclaim_idx;
> +
> unsigned int may_writepage:1;
>
> /* Can mapped pages be reclaimed? */
> @@ -1369,6 +1372,7 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
> struct list_head *src = &lruvec->lists[lru];
> unsigned long nr_taken = 0;
> unsigned long scan;
> + LIST_HEAD(pages_skipped);
>
> for (scan = 0; scan < nr_to_scan && nr_taken < nr_to_scan &&
> !list_empty(src); scan++) {
> @@ -1379,6 +1383,11 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
>
> VM_BUG_ON_PAGE(!PageLRU(page), page);
>
> + if (page_zonenum(page) > sc->reclaim_idx) {
> + list_move(&page->lru, &pages_skipped);
> + continue;
> + }
> +
> switch (__isolate_lru_page(page, mode)) {
> case 0:
> nr_taken += hpage_nr_pages(page);
> @@ -1395,6 +1404,15 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
> }
> }
>
> + /*
> + * Splice any skipped pages to the start of the LRU list. Note that
> + * this disrupts the LRU order when reclaiming for lower zones but
> + * we cannot splice to the tail. If we did then the SWAP_CLUSTER_MAX
> + * scanning would soon rescan the same pages to skip and put the
> + * system at risk of premature OOM.
> + */
> + if (!list_empty(&pages_skipped))
> + list_splice(&pages_skipped, src);
Hmm, that's unfortunate. But probably better than reclaiming the pages
in the name of LRU order, even though it wouldn't help the allocation at
hand.
[...]
> @@ -2516,14 +2535,14 @@ static inline bool compaction_ready(struct zone *zone, int order, int classzone_
> * If a zone is deemed to be full of pinned pages then just give it a light
> * scan then give up on it.
> */
> -static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
> +static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc,
> + enum zone_type classzone_idx)
> {
> struct zoneref *z;
> struct zone *zone;
> unsigned long nr_soft_reclaimed;
> unsigned long nr_soft_scanned;
> gfp_t orig_mask;
> - enum zone_type requested_highidx = gfp_zone(sc->gfp_mask);
>
> /*
> * If the number of buffer_heads in the machine exceeds the maximum
> @@ -2536,15 +2555,15 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
>
> for_each_zone_zonelist_nodemask(zone, z, zonelist,
> gfp_zone(sc->gfp_mask), sc->nodemask) {
> - enum zone_type classzone_idx;
> -
> if (!populated_zone(zone))
> continue;
>
> - classzone_idx = requested_highidx;
> while (!populated_zone(zone->zone_pgdat->node_zones +
> - classzone_idx))
> + classzone_idx)) {
> + sc->reclaim_idx--;
> classzone_idx--;
> + continue;
Isn't this wrong to do this across whole zonelist which will contain
multiple nodes? Example: a small node 0 without Normal zone will get us
sc->reclaim_idx == classzone_idx == dma32. Node 1 won't have dma/dma32
zones so we won't see classzone_idx populated, and the while loop will
lead to underflow?
And sc->reclaim_idx seems to be unitialized when called via
try_to_free_pages() -> do_try_to_free_pages() -> shrink_zones() ?
Which means it's zero and we underflow immediately?
> @@ -3207,15 +3228,14 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
> sc.may_writepage = 1;
>
> /*
> - * Now scan the zone in the dma->highmem direction, stopping
> - * at the last zone which needs scanning.
> - *
> - * We do this because the page allocator works in the opposite
> - * direction. This prevents the page allocator from allocating
> - * pages behind kswapd's direction of progress, which would
> - * cause too much scanning of the lower zones.
> + * Continue scanning in the highmem->dma direction stopping at
> + * the last zone which needs scanning. This may reclaim lowmem
> + * pages that are not necessary for zone balancing but it
> + * preserves LRU ordering. It is assumed that the bulk of
> + * allocation requests can use arbitrary zones with the
> + * possible exception of big highmem:lowmem configurations.
> */
> - for (i = 0; i <= end_zone; i++) {
> + for (i = end_zone; i >= end_zone; i--) {
i >= 0 ?
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 05/27] mm, vmscan: Have kswapd only scan based on the highest requested zone
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (3 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 04/27] mm, vmscan: Begin reclaiming pages on a per-node basis Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-15 13:13 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Mel Gorman
` (21 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
kswapd checks all eligible zones to see if they need balancing even if it was
woken for a lower zone. This made sense when we reclaimed on a per-zone basis
because we wanted to shrink zones fairly so avoid age-inversion problems.
Ideally this is completely unnecessary when reclaiming on a per-node basis.
In theory, there may still be anomalies when all requests are for lower
zones and very old pages are preserved in higher zones but this should be
the exceptional case.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/vmscan.c | 7 ++-----
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index ab1b28e7e20a..0a619241c576 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3171,11 +3171,8 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
sc.nr_reclaimed = 0;
- /*
- * Scan in the highmem->dma direction for the highest
- * zone which needs scanning
- */
- for (i = pgdat->nr_zones - 1; i >= 0; i--) {
+ /* Scan from the highest requested zone to dma */
+ for (i = classzone_idx; i >= 0; i--) {
struct zone *zone = pgdat->node_zones + i;
if (!populated_zone(zone))
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 05/27] mm, vmscan: Have kswapd only scan based on the highest requested zone
2016-06-09 18:04 ` [PATCH 05/27] mm, vmscan: Have kswapd only scan based on the highest requested zone Mel Gorman
@ 2016-06-15 13:13 ` Vlastimil Babka
2016-06-16 7:44 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-15 13:13 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> kswapd checks all eligible zones to see if they need balancing even if it was
> woken for a lower zone. This made sense when we reclaimed on a per-zone basis
> because we wanted to shrink zones fairly so avoid age-inversion problems.
Now we reclaim a single lru, but still will skip over pages from the
higher zones than reclaim_idx, so this is not much different from
per-zone basis wrt age-inversion?
> Ideally this is completely unnecessary when reclaiming on a per-node basis.
> In theory, there may still be anomalies when all requests are for lower
> zones and very old pages are preserved in higher zones but this should be
> the exceptional case.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
I don't see the argument, but agree it should be exceptional in any
case, and if there's such a case, it's better to focus on pages from the
zone(s) where a pending (potentially atomic) allocation is restricted
to. Or rather, this is the only way we can focus reclaim on such pages
now that there's a single lru list.
Acked-by: Vlastimil Babka <vbabka@suse.cz>
> ---
> mm/vmscan.c | 7 ++-----
> 1 file changed, 2 insertions(+), 5 deletions(-)
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index ab1b28e7e20a..0a619241c576 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -3171,11 +3171,8 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
>
> sc.nr_reclaimed = 0;
>
> - /*
> - * Scan in the highmem->dma direction for the highest
> - * zone which needs scanning
> - */
> - for (i = pgdat->nr_zones - 1; i >= 0; i--) {
> + /* Scan from the highest requested zone to dma */
> + for (i = classzone_idx; i >= 0; i--) {
> struct zone *zone = pgdat->node_zones + i;
>
> if (!populated_zone(zone))
>
^ permalink raw reply [flat|nested] 70+ messages in thread* Re: [PATCH 05/27] mm, vmscan: Have kswapd only scan based on the highest requested zone
2016-06-15 13:13 ` Vlastimil Babka
@ 2016-06-16 7:44 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-16 7:44 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Wed, Jun 15, 2016 at 03:13:13PM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> >kswapd checks all eligible zones to see if they need balancing even if it was
> >woken for a lower zone. This made sense when we reclaimed on a per-zone basis
> >because we wanted to shrink zones fairly so avoid age-inversion problems.
>
> Now we reclaim a single lru, but still will skip over pages from the higher
> zones than reclaim_idx, so this is not much different from per-zone basis
> wrt age-inversion?
>
Yes, but it only applies in the case where the allocation request is zone
restricted. Previously, even with fair zone allocation policy, we had
problems with a high zone with recently allocated pages being reclaimed
simply because the low watermark was reached. Think of bugs in the past
where the normal zone was a small percentage of memory.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (4 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 05/27] mm, vmscan: Have kswapd only scan based on the highest requested zone Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-15 14:23 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 07/27] mm, vmscan: Remove balance gap Mel Gorman
` (20 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Patch "mm: vmscan: Begin reclaiming pages on a per-node basis" started
thinking of reclaim in terms of nodes but kswapd is still zone-centric. This
patch gets rid of many of the node-based versus zone-based decisions.
o A node is considered balanced when any eligible lower zone is balanced.
This eliminates one class of age-inversion problem because we avoid
reclaiming a newer page just because it's in the wrong zone
o pgdat_balanced disappears because we now only care about one zone being
balanced.
o Some anomalies related to writeback and congestion tracking being based on
zones disappear.
o kswapd no longer has to take care to reclaim zones in the reverse order
that the page allocator uses.
o Most importantly of all, reclaim from node 0 with multiple zones will
have similar aging and reclaiming characteristics as every
other node.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
---
mm/vmscan.c | 292 +++++++++++++++++++++---------------------------------------
1 file changed, 101 insertions(+), 191 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 0a619241c576..9368af4cfb06 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2942,7 +2942,8 @@ unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
}
#endif
-static void age_active_anon(struct zone *zone, struct scan_control *sc)
+static void age_active_anon(struct pglist_data *pgdat,
+ struct zone *zone, struct scan_control *sc)
{
struct mem_cgroup *memcg;
@@ -2961,85 +2962,15 @@ static void age_active_anon(struct zone *zone, struct scan_control *sc)
} while (memcg);
}
-static bool zone_balanced(struct zone *zone, int order, bool highorder,
+static bool zone_balanced(struct zone *zone, int order,
unsigned long balance_gap, int classzone_idx)
{
unsigned long mark = high_wmark_pages(zone) + balance_gap;
- /*
- * When checking from pgdat_balanced(), kswapd should stop and sleep
- * when it reaches the high order-0 watermark and let kcompactd take
- * over. Other callers such as wakeup_kswapd() want to determine the
- * true high-order watermark.
- */
- if (IS_ENABLED(CONFIG_COMPACTION) && !highorder) {
- mark += (1UL << order);
- order = 0;
- }
-
return zone_watermark_ok_safe(zone, order, mark, classzone_idx);
}
/*
- * pgdat_balanced() is used when checking if a node is balanced.
- *
- * For order-0, all zones must be balanced!
- *
- * For high-order allocations only zones that meet watermarks and are in a
- * zone allowed by the callers classzone_idx are added to balanced_pages. The
- * total of balanced pages must be at least 25% of the zones allowed by
- * classzone_idx for the node to be considered balanced. Forcing all zones to
- * be balanced for high orders can cause excessive reclaim when there are
- * imbalanced zones.
- * The choice of 25% is due to
- * o a 16M DMA zone that is balanced will not balance a zone on any
- * reasonable sized machine
- * o On all other machines, the top zone must be at least a reasonable
- * percentage of the middle zones. For example, on 32-bit x86, highmem
- * would need to be at least 256M for it to be balance a whole node.
- * Similarly, on x86-64 the Normal zone would need to be at least 1G
- * to balance a node on its own. These seemed like reasonable ratios.
- */
-static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
-{
- unsigned long managed_pages = 0;
- unsigned long balanced_pages = 0;
- int i;
-
- /* Check the watermark levels */
- for (i = 0; i <= classzone_idx; i++) {
- struct zone *zone = pgdat->node_zones + i;
-
- if (!populated_zone(zone))
- continue;
-
- managed_pages += zone->managed_pages;
-
- /*
- * A special case here:
- *
- * balance_pgdat() skips over all_unreclaimable after
- * DEF_PRIORITY. Effectively, it considers them balanced so
- * they must be considered balanced here as well!
- */
- if (!pgdat_reclaimable(zone->zone_pgdat)) {
- balanced_pages += zone->managed_pages;
- continue;
- }
-
- if (zone_balanced(zone, order, false, 0, i))
- balanced_pages += zone->managed_pages;
- else if (!order)
- return false;
- }
-
- if (order)
- return balanced_pages >= (managed_pages >> 2);
- else
- return true;
-}
-
-/*
* Prepare kswapd for sleeping. This verifies that there are no processes
* waiting in throttle_direct_reclaim() and that watermarks have been met.
*
@@ -3048,6 +2979,8 @@ static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx)
static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
int classzone_idx)
{
+ int i;
+
/* If a direct reclaimer woke kswapd within HZ/10, it's premature */
if (remaining)
return false;
@@ -3068,101 +3001,90 @@ static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
if (waitqueue_active(&pgdat->pfmemalloc_wait))
wake_up_all(&pgdat->pfmemalloc_wait);
- return pgdat_balanced(pgdat, order, classzone_idx);
+ for (i = 0; i <= classzone_idx; i++) {
+ struct zone *zone = pgdat->node_zones + i;
+
+ if (!populated_zone(zone))
+ continue;
+
+ if (zone_balanced(zone, order, 0, classzone_idx))
+ return true;
+ }
+
+ return false;
}
/*
- * kswapd shrinks the zone by the number of pages required to reach
- * the high watermark.
+ * kswapd shrinks a node of pages that are at or below the highest usable
+ * zone that is currently unbalanced.
*
* Returns true if kswapd scanned at least the requested number of pages to
* reclaim or if the lack of progress was due to pages under writeback.
* This is used to determine if the scanning priority needs to be raised.
*/
-static bool kswapd_shrink_zone(struct zone *zone,
+static bool kswapd_shrink_node(pg_data_t *pgdat,
int classzone_idx,
struct scan_control *sc)
{
- unsigned long balance_gap;
- bool lowmem_pressure;
- struct pglist_data *pgdat = zone->zone_pgdat;
+ struct zone *zone;
+ unsigned long nr_to_reclaim = 0;
+ int z;
- /* Reclaim above the high watermark. */
- sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone));
+ /* Reclaim a number of pages proportional to the number of zones */
+ for (z = 0; z <= classzone_idx; z++) {
+ zone = pgdat->node_zones + z;
+ if (!populated_zone(zone))
+ continue;
- /*
- * We put equal pressure on every zone, unless one zone has way too
- * many pages free already. The "too many pages" is defined as the
- * high wmark plus a "gap" where the gap is either the low
- * watermark or 1% of the zone, whichever is smaller.
- */
- balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP(
- zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO));
+ nr_to_reclaim += max(high_wmark_pages(zone), SWAP_CLUSTER_MAX);
+ }
/*
- * If there is no low memory pressure or the zone is balanced then no
- * reclaim is necessary
+ * Historically care was taken to put equal pressure on all zones but
+ * now pressure is applied based on node LRU order.
*/
- lowmem_pressure = (buffer_heads_over_limit && is_highmem(zone));
- if (!lowmem_pressure && zone_balanced(zone, sc->order, false,
- balance_gap, classzone_idx))
- return true;
-
- shrink_node(zone->zone_pgdat, sc, classzone_idx);
-
- /* TODO: ANOMALY */
- clear_bit(PGDAT_WRITEBACK, &pgdat->flags);
+ shrink_node(pgdat, sc, classzone_idx);
/*
- * If a zone reaches its high watermark, consider it to be no longer
- * congested. It's possible there are dirty pages backed by congested
- * BDIs but as pressure is relieved, speculatively avoid congestion
- * waits.
+ * Fragmentation may mean that the system cannot be rebalanced for
+ * high-order allocations. If twice the allocation size has been
+ * reclaimed then recheck watermarks only at order-0 to prevent
+ * excessive reclaim. Assume that a process requested a high-order
+ * can direct reclaim/compact.
*/
- if (pgdat_reclaimable(zone->zone_pgdat) &&
- zone_balanced(zone, sc->order, false, 0, classzone_idx)) {
- clear_bit(PGDAT_CONGESTED, &pgdat->flags);
- clear_bit(PGDAT_DIRTY, &pgdat->flags);
- }
+ if (sc->order && sc->nr_reclaimed >= 2UL << sc->order)
+ sc->order = 0;
return sc->nr_scanned >= sc->nr_to_reclaim;
}
/*
- * For kswapd, balance_pgdat() will work across all this node's zones until
- * they are all at high_wmark_pages(zone).
- *
- * Returns the highest zone idx kswapd was reclaiming at
+ * For kswapd, balance_pgdat() will reclaim pages across a node from zones
+ * that are eligible for use by the caller until at least one zone is
+ * balanced.
*
- * There is special handling here for zones which are full of pinned pages.
- * This can happen if the pages are all mlocked, or if they are all used by
- * device drivers (say, ZONE_DMA). Or if they are all in use by hugetlb.
- * What we do is to detect the case where all pages in the zone have been
- * scanned twice and there has been zero successful reclaim. Mark the zone as
- * dead and from now on, only perform a short scan. Basically we're polling
- * the zone for when the problem goes away.
+ * Returns the order kswapd finished reclaiming at.
*
* kswapd scans the zones in the highmem->normal->dma direction. It skips
* zones which have free_pages > high_wmark_pages(zone), but once a zone is
- * found to have free_pages <= high_wmark_pages(zone), we scan that zone and the
- * lower zones regardless of the number of free pages in the lower zones. This
- * interoperates with the page allocator fallback scheme to ensure that aging
- * of pages is balanced across the zones.
+ * found to have free_pages <= high_wmark_pages(zone), any page is that zone
+ * or lower is eligible for reclaim until at least one usable zone is
+ * balanced.
*/
static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
{
int i;
- int end_zone = 0; /* Inclusive. 0 = ZONE_DMA */
unsigned long nr_soft_reclaimed;
unsigned long nr_soft_scanned;
+ struct zone *zone;
struct scan_control sc = {
.gfp_mask = GFP_KERNEL,
- .reclaim_idx = MAX_NR_ZONES - 1,
.order = order,
.priority = DEF_PRIORITY,
.may_writepage = !laptop_mode,
.may_unmap = 1,
.may_swap = 1,
+ .reclaim_idx = classzone_idx,
};
count_vm_event(PAGEOUTRUN);
@@ -3173,21 +3095,10 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
/* Scan from the highest requested zone to dma */
for (i = classzone_idx; i >= 0; i--) {
- struct zone *zone = pgdat->node_zones + i;
-
+ zone = pgdat->node_zones + i;
if (!populated_zone(zone))
continue;
- if (sc.priority != DEF_PRIORITY &&
- !pgdat_reclaimable(zone->zone_pgdat))
- continue;
-
- /*
- * Do some background aging of the anon list, to give
- * pages a chance to be referenced before reclaiming.
- */
- age_active_anon(zone, &sc);
-
/*
* If the number of buffer_heads in the machine
* exceeds the maximum allowed level and this node
@@ -3195,19 +3106,17 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
* it to relieve lowmem pressure.
*/
if (buffer_heads_over_limit && is_highmem_idx(i)) {
- end_zone = i;
+ classzone_idx = i;
break;
}
- if (!zone_balanced(zone, order, false, 0, 0)) {
- end_zone = i;
+ if (!zone_balanced(zone, order, 0, 0)) {
+ classzone_idx = i;
break;
} else {
/*
- * If balanced, clear the dirty and congested
- * flags
- *
- * TODO: ANOMALY
+ * If any eligible zone is balanced then the
+ * node is not considered congested or dirty.
*/
clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
@@ -3218,51 +3127,34 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
goto out;
/*
+ * Do some background aging of the anon list, to give
+ * pages a chance to be referenced before reclaiming. All
+ * pages are rotated regardless of classzone as this is
+ * about consistent aging.
+ */
+ age_active_anon(pgdat, &pgdat->node_zones[MAX_NR_ZONES - 1], &sc);
+
+ /*
* If we're getting trouble reclaiming, start doing writepage
* even in laptop mode.
*/
- if (sc.priority < DEF_PRIORITY - 2)
+ if (sc.priority < DEF_PRIORITY - 2 || !pgdat_reclaimable(pgdat))
sc.may_writepage = 1;
+ /* Call soft limit reclaim before calling shrink_node. */
+ sc.nr_scanned = 0;
+ nr_soft_scanned = 0;
+ nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, sc.order,
+ sc.gfp_mask, &nr_soft_scanned);
+ sc.nr_reclaimed += nr_soft_reclaimed;
+
/*
- * Continue scanning in the highmem->dma direction stopping at
- * the last zone which needs scanning. This may reclaim lowmem
- * pages that are not necessary for zone balancing but it
- * preserves LRU ordering. It is assumed that the bulk of
- * allocation requests can use arbitrary zones with the
- * possible exception of big highmem:lowmem configurations.
+ * There should be no need to raise the scanning priority if
+ * enough pages are already being scanned that that high
+ * watermark would be met at 100% efficiency.
*/
- for (i = end_zone; i >= end_zone; i--) {
- struct zone *zone = pgdat->node_zones + i;
-
- if (!populated_zone(zone))
- continue;
-
- if (sc.priority != DEF_PRIORITY &&
- !pgdat_reclaimable(zone->zone_pgdat))
- continue;
-
- sc.nr_scanned = 0;
- sc.reclaim_idx = i;
-
- nr_soft_scanned = 0;
- /*
- * Call soft limit reclaim before calling shrink_zone.
- */
- nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
- order, sc.gfp_mask,
- &nr_soft_scanned);
- sc.nr_reclaimed += nr_soft_reclaimed;
-
- /*
- * There should be no need to raise the scanning
- * priority if enough pages are already being scanned
- * that that high watermark would be met at 100%
- * efficiency.
- */
- if (kswapd_shrink_zone(zone, end_zone, &sc))
- raise_priority = false;
- }
+ if (kswapd_shrink_node(pgdat, classzone_idx, &sc))
+ raise_priority = false;
/*
* If the low watermark is met there is no need for processes
@@ -3278,20 +3170,37 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
break;
/*
+ * Stop reclaiming if any eligible zone is balanced and clear
+ * node writeback or congested.
+ */
+ for (i = 0; i <= classzone_idx; i++) {
+ zone = pgdat->node_zones + i;
+ if (!populated_zone(zone))
+ continue;
+
+ if (zone_balanced(zone, sc.order, 0, classzone_idx)) {
+ clear_bit(PGDAT_CONGESTED, &pgdat->flags);
+ clear_bit(PGDAT_DIRTY, &pgdat->flags);
+ goto out;
+ }
+ }
+
+ /*
* Raise priority if scanning rate is too low or there was no
* progress in reclaiming pages
*/
if (raise_priority || !sc.nr_reclaimed)
sc.priority--;
- } while (sc.priority >= 1 &&
- !pgdat_balanced(pgdat, order, classzone_idx));
+ } while (sc.priority >= 1);
out:
/*
- * Return the highest zone idx we were reclaiming at so
- * prepare_kswapd_sleep() makes the same decisions as here.
+ * Return the order kswapd stopped reclaiming at as
+ * prepare_kswapd_sleep() takes it into account. If another caller
+ * entered the allocator slow path while kswapd was awake, order will
+ * remain at the higher level.
*/
- return end_zone;
+ return sc.order;
}
static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
@@ -3448,8 +3357,9 @@ static int kswapd(void *p)
*/
if (!ret) {
trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
- balanced_classzone_idx = balance_pgdat(pgdat, order,
- classzone_idx);
+
+ /* return value ignored until next patch */
+ balance_pgdat(pgdat, order, classzone_idx);
}
}
@@ -3479,7 +3389,7 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
}
if (!waitqueue_active(&pgdat->kswapd_wait))
return;
- if (zone_balanced(zone, order, true, 0, 0))
+ if (zone_balanced(zone, order, 0, 0))
return;
trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes
2016-06-09 18:04 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Mel Gorman
@ 2016-06-15 14:23 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-15 14:23 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> Patch "mm: vmscan: Begin reclaiming pages on a per-node basis" started
> thinking of reclaim in terms of nodes but kswapd is still zone-centric. This
> patch gets rid of many of the node-based versus zone-based decisions.
>
> o A node is considered balanced when any eligible lower zone is balanced.
> This eliminates one class of age-inversion problem because we avoid
> reclaiming a newer page just because it's in the wrong zone
> o pgdat_balanced disappears because we now only care about one zone being
> balanced.
> o Some anomalies related to writeback and congestion tracking being based on
> zones disappear.
> o kswapd no longer has to take care to reclaim zones in the reverse order
> that the page allocator uses.
> o Most importantly of all, reclaim from node 0 with multiple zones will
> have similar aging and reclaiming characteristics as every
> other node.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 07/27] mm, vmscan: Remove balance gap
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (5 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 06/27] mm, vmscan: Make kswapd reclaim in terms of nodes Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-15 14:28 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 08/27] mm, vmscan: Simplify the logic deciding whether kswapd sleeps Mel Gorman
` (19 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
The balance gap was introduced to apply equal pressure to all zones when
reclaiming for a higher zone. With node-based LRU, the need for the balance
gap is removed and the code is dead so remove it.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/vmscan.c | 19 ++++++++-----------
1 file changed, 8 insertions(+), 11 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 9368af4cfb06..96bf841f9352 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2488,7 +2488,7 @@ static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
*/
static inline bool compaction_ready(struct zone *zone, int order, int classzone_idx)
{
- unsigned long balance_gap, watermark;
+ unsigned long watermark;
bool watermark_ok;
/*
@@ -2497,9 +2497,7 @@ static inline bool compaction_ready(struct zone *zone, int order, int classzone_
* there is a buffer of free pages available to give compaction
* a reasonable chance of completing and allocating the page
*/
- balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP(
- zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO));
- watermark = high_wmark_pages(zone) + balance_gap + (2UL << order);
+ watermark = high_wmark_pages(zone) + (2UL << order);
watermark_ok = zone_watermark_ok_safe(zone, 0, watermark, classzone_idx);
/*
@@ -2962,10 +2960,9 @@ static void age_active_anon(struct pglist_data *pgdat,
} while (memcg);
}
-static bool zone_balanced(struct zone *zone, int order,
- unsigned long balance_gap, int classzone_idx)
+static bool zone_balanced(struct zone *zone, int order, int classzone_idx)
{
- unsigned long mark = high_wmark_pages(zone) + balance_gap;
+ unsigned long mark = high_wmark_pages(zone);
return zone_watermark_ok_safe(zone, order, mark, classzone_idx);
}
@@ -3007,7 +3004,7 @@ static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining,
if (!populated_zone(zone))
continue;
- if (zone_balanced(zone, order, 0, classzone_idx))
+ if (zone_balanced(zone, order, classzone_idx))
return true;
}
@@ -3110,7 +3107,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
break;
}
- if (!zone_balanced(zone, order, 0, 0)) {
+ if (!zone_balanced(zone, order, 0)) {
classzone_idx = i;
break;
} else {
@@ -3178,7 +3175,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
if (!populated_zone(zone))
continue;
- if (zone_balanced(zone, sc.order, 0, classzone_idx)) {
+ if (zone_balanced(zone, sc.order, classzone_idx)) {
clear_bit(PGDAT_CONGESTED, &pgdat->flags);
clear_bit(PGDAT_DIRTY, &pgdat->flags);
goto out;
@@ -3389,7 +3386,7 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
}
if (!waitqueue_active(&pgdat->kswapd_wait))
return;
- if (zone_balanced(zone, order, 0, 0))
+ if (zone_balanced(zone, order, 0))
return;
trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 07/27] mm, vmscan: Remove balance gap
2016-06-09 18:04 ` [PATCH 07/27] mm, vmscan: Remove balance gap Mel Gorman
@ 2016-06-15 14:28 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-15 14:28 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> The balance gap was introduced to apply equal pressure to all zones when
> reclaiming for a higher zone. With node-based LRU, the need for the balance
> gap is removed and the code is dead so remove it.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
> ---
> mm/vmscan.c | 19 ++++++++-----------
> 1 file changed, 8 insertions(+), 11 deletions(-)
Also this:
include/linux/swap.h:#define KSWAPD_ZONE_BALANCE_GAP_RATIO 100
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 08/27] mm, vmscan: Simplify the logic deciding whether kswapd sleeps
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (6 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 07/27] mm, vmscan: Remove balance gap Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-15 15:18 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 09/27] mm, vmscan: By default have direct reclaim only shrink once per node Mel Gorman
` (18 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
kswapd goes through some complex steps trying to figure out if it
should stay awake based on the classzone_idx and the requested order.
It is unnecessarily complex and passes in an invalid classzone_idx to
balance_pgdat(). What matters most of all is whether a larger order has
been requsted and whether kswapd successfully reclaimed at the previous
order. This patch irons out the logic to check just that and the end result
is less headache inducing.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
---
include/linux/mmzone.h | 5 ++--
mm/memory_hotplug.c | 5 ++--
mm/page_alloc.c | 2 +-
mm/vmscan.c | 79 +++++++++++++++++++++++++-------------------------
4 files changed, 46 insertions(+), 45 deletions(-)
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index aff621fcc6af..15ce454d0d59 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -663,8 +663,9 @@ typedef struct pglist_data {
wait_queue_head_t pfmemalloc_wait;
struct task_struct *kswapd; /* Protected by
mem_hotplug_begin/end() */
- int kswapd_max_order;
- enum zone_type classzone_idx;
+ int kswapd_order;
+ enum zone_type kswapd_classzone_idx;
+
#ifdef CONFIG_COMPACTION
int kcompactd_max_order;
enum zone_type kcompactd_classzone_idx;
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index c5278360ca66..42d758bffd0a 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1209,9 +1209,10 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
arch_refresh_nodedata(nid, pgdat);
} else {
- /* Reset the nr_zones and classzone_idx to 0 before reuse */
+ /* Reset the nr_zones, order and classzone_idx before reuse */
pgdat->nr_zones = 0;
- pgdat->classzone_idx = 0;
+ pgdat->kswapd_order = 0;
+ pgdat->kswapd_classzone_idx = -1;
}
/* we can use NODE_DATA(nid) from here */
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 4ce578b969da..d8cb483d5cad 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6036,7 +6036,7 @@ void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
unsigned long end_pfn = 0;
/* pg_data_t should be reset to zero when it's allocated */
- WARN_ON(pgdat->nr_zones || pgdat->classzone_idx);
+ WARN_ON(pgdat->nr_zones || pgdat->kswapd_classzone_idx);
reset_deferred_meminit(pgdat);
pgdat->node_id = nid;
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 96bf841f9352..14b34eebedff 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2727,7 +2727,7 @@ static bool pfmemalloc_watermark_ok(pg_data_t *pgdat)
/* kswapd must be awake if processes are being throttled */
if (!wmark_ok && waitqueue_active(&pgdat->kswapd_wait)) {
- pgdat->classzone_idx = min(pgdat->classzone_idx,
+ pgdat->kswapd_classzone_idx = min(pgdat->kswapd_classzone_idx,
(enum zone_type)ZONE_NORMAL);
wake_up_interruptible(&pgdat->kswapd_wait);
}
@@ -3211,6 +3211,12 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
+ /* If kswapd has not been woken recently, then full sleep */
+ if (classzone_idx == -1) {
+ classzone_idx = balanced_classzone_idx = MAX_NR_ZONES - 1;
+ goto full_sleep;
+ }
+
/* Try to sleep for a short interval */
if (prepare_kswapd_sleep(pgdat, order, remaining,
balanced_classzone_idx)) {
@@ -3233,6 +3239,7 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
}
+full_sleep:
/*
* After a short sleep, check if it was a premature sleep. If not, then
* go fully to sleep until explicitly woken up.
@@ -3279,9 +3286,7 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
*/
static int kswapd(void *p)
{
- unsigned long order, new_order;
- int classzone_idx, new_classzone_idx;
- int balanced_classzone_idx;
+ unsigned int order, classzone_idx;
pg_data_t *pgdat = (pg_data_t*)p;
struct task_struct *tsk = current;
@@ -3311,38 +3316,25 @@ static int kswapd(void *p)
tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD;
set_freezable();
- order = new_order = 0;
- classzone_idx = new_classzone_idx = pgdat->nr_zones - 1;
- balanced_classzone_idx = classzone_idx;
+ pgdat->kswapd_order = order = 0;
+ pgdat->kswapd_classzone_idx = classzone_idx = -1;
for ( ; ; ) {
bool ret;
+kswapd_try_sleep:
+ kswapd_try_to_sleep(pgdat, order, classzone_idx, classzone_idx);
+
/*
- * While we were reclaiming, there might have been another
- * wakeup, so check the values.
+ * Read the new order and classzone_idx which may be -1 if
+ * kswapd_try_to_sleep() woke up after a short timeout instead
+ * of being woken by the page allocator.
*/
- new_order = pgdat->kswapd_max_order;
- new_classzone_idx = pgdat->classzone_idx;
- pgdat->kswapd_max_order = 0;
- pgdat->classzone_idx = pgdat->nr_zones - 1;
-
- if (order < new_order || classzone_idx > new_classzone_idx) {
- /*
- * Don't sleep if someone wants a larger 'order'
- * allocation or has tigher zone constraints
- */
- order = new_order;
- classzone_idx = new_classzone_idx;
- } else {
- kswapd_try_to_sleep(pgdat, order, classzone_idx,
- balanced_classzone_idx);
- order = pgdat->kswapd_max_order;
- classzone_idx = pgdat->classzone_idx;
- new_order = order;
- new_classzone_idx = classzone_idx;
- pgdat->kswapd_max_order = 0;
- pgdat->classzone_idx = pgdat->nr_zones - 1;
- }
+ order = pgdat->kswapd_order;
+ classzone_idx = pgdat->kswapd_classzone_idx;
+ if (classzone_idx == -1)
+ classzone_idx = MAX_NR_ZONES - 1;
+ pgdat->kswapd_order = 0;
+ pgdat->kswapd_classzone_idx = -1;
ret = try_to_freeze();
if (kthread_should_stop())
@@ -3352,12 +3344,19 @@ static int kswapd(void *p)
* We can speed up thawing tasks if we don't call balance_pgdat
* after returning from the refrigerator
*/
- if (!ret) {
- trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
+ if (ret)
+ continue;
- /* return value ignored until next patch */
- balance_pgdat(pgdat, order, classzone_idx);
- }
+ /*
+ * Try reclaim the requested order but if that fails
+ * then try sleeping on the basis of the order reclaimed.
+ */
+ trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
+ if (balance_pgdat(pgdat, order, classzone_idx) < order)
+ goto kswapd_try_sleep;
+
+ order = pgdat->kswapd_order;
+ classzone_idx = pgdat->kswapd_classzone_idx;
}
tsk->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD);
@@ -3380,10 +3379,10 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
if (!cpuset_zone_allowed(zone, GFP_KERNEL | __GFP_HARDWALL))
return;
pgdat = zone->zone_pgdat;
- if (pgdat->kswapd_max_order < order) {
- pgdat->kswapd_max_order = order;
- pgdat->classzone_idx = min(pgdat->classzone_idx, classzone_idx);
- }
+ if (pgdat->kswapd_classzone_idx == -1)
+ pgdat->kswapd_classzone_idx = classzone_idx;
+ pgdat->kswapd_classzone_idx = max(pgdat->kswapd_classzone_idx, classzone_idx);
+ pgdat->kswapd_order = max(pgdat->kswapd_order, order);
if (!waitqueue_active(&pgdat->kswapd_wait))
return;
if (zone_balanced(zone, order, 0))
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 08/27] mm, vmscan: Simplify the logic deciding whether kswapd sleeps
2016-06-09 18:04 ` [PATCH 08/27] mm, vmscan: Simplify the logic deciding whether kswapd sleeps Mel Gorman
@ 2016-06-15 15:18 ` Vlastimil Babka
2016-06-16 8:30 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-15 15:18 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> kswapd goes through some complex steps trying to figure out if it
> should stay awake based on the classzone_idx and the requested order.
> It is unnecessarily complex and passes in an invalid classzone_idx to
> balance_pgdat(). What matters most of all is whether a larger order has
> been requsted and whether kswapd successfully reclaimed at the previous
> order. This patch irons out the logic to check just that and the end result
> is less headache inducing.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> Acked-by: Johannes Weiner <hannes@cmpxchg.org>
> ---
> include/linux/mmzone.h | 5 ++--
> mm/memory_hotplug.c | 5 ++--
> mm/page_alloc.c | 2 +-
> mm/vmscan.c | 79 +++++++++++++++++++++++++-------------------------
> 4 files changed, 46 insertions(+), 45 deletions(-)
>
> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
> index aff621fcc6af..15ce454d0d59 100644
> --- a/include/linux/mmzone.h
> +++ b/include/linux/mmzone.h
> @@ -663,8 +663,9 @@ typedef struct pglist_data {
> wait_queue_head_t pfmemalloc_wait;
> struct task_struct *kswapd; /* Protected by
> mem_hotplug_begin/end() */
> - int kswapd_max_order;
> - enum zone_type classzone_idx;
> + int kswapd_order;
> + enum zone_type kswapd_classzone_idx;
> +
> #ifdef CONFIG_COMPACTION
> int kcompactd_max_order;
> enum zone_type kcompactd_classzone_idx;
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index c5278360ca66..42d758bffd0a 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -1209,9 +1209,10 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
>
> arch_refresh_nodedata(nid, pgdat);
> } else {
> - /* Reset the nr_zones and classzone_idx to 0 before reuse */
> + /* Reset the nr_zones, order and classzone_idx before reuse */
> pgdat->nr_zones = 0;
> - pgdat->classzone_idx = 0;
> + pgdat->kswapd_order = 0;
> + pgdat->kswapd_classzone_idx = -1;
> }
>
> /* we can use NODE_DATA(nid) from here */
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 4ce578b969da..d8cb483d5cad 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -6036,7 +6036,7 @@ void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
> unsigned long end_pfn = 0;
>
> /* pg_data_t should be reset to zero when it's allocated */
> - WARN_ON(pgdat->nr_zones || pgdat->classzone_idx);
> + WARN_ON(pgdat->nr_zones || pgdat->kswapd_classzone_idx);
Above you changed the reset value of kswapd_classzone_idx from 0 to -1,
so won't this trigger? Also should we check kswapd_order that's newly
reset too?
>
> reset_deferred_meminit(pgdat);
> pgdat->node_id = nid;
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 96bf841f9352..14b34eebedff 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -2727,7 +2727,7 @@ static bool pfmemalloc_watermark_ok(pg_data_t *pgdat)
>
> /* kswapd must be awake if processes are being throttled */
> if (!wmark_ok && waitqueue_active(&pgdat->kswapd_wait)) {
> - pgdat->classzone_idx = min(pgdat->classzone_idx,
> + pgdat->kswapd_classzone_idx = min(pgdat->kswapd_classzone_idx,
> (enum zone_type)ZONE_NORMAL);
> wake_up_interruptible(&pgdat->kswapd_wait);
> }
> @@ -3211,6 +3211,12 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
>
> prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
>
> + /* If kswapd has not been woken recently, then full sleep */
> + if (classzone_idx == -1) {
> + classzone_idx = balanced_classzone_idx = MAX_NR_ZONES - 1;
> + goto full_sleep;
This will skip the wakeup_kcompactd() part.
> + }
> +
> /* Try to sleep for a short interval */
> if (prepare_kswapd_sleep(pgdat, order, remaining,
> balanced_classzone_idx)) {
> @@ -3233,6 +3239,7 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
> prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
> }
>
> +full_sleep:
> /*
> * After a short sleep, check if it was a premature sleep. If not, then
> * go fully to sleep until explicitly woken up.
> @@ -3279,9 +3286,7 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
> */
> static int kswapd(void *p)
> {
> - unsigned long order, new_order;
> - int classzone_idx, new_classzone_idx;
> - int balanced_classzone_idx;
> + unsigned int order, classzone_idx;
> pg_data_t *pgdat = (pg_data_t*)p;
> struct task_struct *tsk = current;
>
> @@ -3311,38 +3316,25 @@ static int kswapd(void *p)
> tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD;
> set_freezable();
>
> - order = new_order = 0;
> - classzone_idx = new_classzone_idx = pgdat->nr_zones - 1;
> - balanced_classzone_idx = classzone_idx;
> + pgdat->kswapd_order = order = 0;
> + pgdat->kswapd_classzone_idx = classzone_idx = -1;
> for ( ; ; ) {
> bool ret;
>
> +kswapd_try_sleep:
> + kswapd_try_to_sleep(pgdat, order, classzone_idx, classzone_idx);
The last two parameters are now the same, remove one?
> +
> /*
> - * While we were reclaiming, there might have been another
> - * wakeup, so check the values.
> + * Read the new order and classzone_idx which may be -1 if
> + * kswapd_try_to_sleep() woke up after a short timeout instead
> + * of being woken by the page allocator.
> */
> - new_order = pgdat->kswapd_max_order;
> - new_classzone_idx = pgdat->classzone_idx;
> - pgdat->kswapd_max_order = 0;
> - pgdat->classzone_idx = pgdat->nr_zones - 1;
> -
> - if (order < new_order || classzone_idx > new_classzone_idx) {
> - /*
> - * Don't sleep if someone wants a larger 'order'
> - * allocation or has tigher zone constraints
> - */
> - order = new_order;
> - classzone_idx = new_classzone_idx;
> - } else {
> - kswapd_try_to_sleep(pgdat, order, classzone_idx,
> - balanced_classzone_idx);
> - order = pgdat->kswapd_max_order;
> - classzone_idx = pgdat->classzone_idx;
> - new_order = order;
> - new_classzone_idx = classzone_idx;
> - pgdat->kswapd_max_order = 0;
> - pgdat->classzone_idx = pgdat->nr_zones - 1;
> - }
> + order = pgdat->kswapd_order;
> + classzone_idx = pgdat->kswapd_classzone_idx;
> + if (classzone_idx == -1)
> + classzone_idx = MAX_NR_ZONES - 1;
> + pgdat->kswapd_order = 0;
> + pgdat->kswapd_classzone_idx = -1;
>
> ret = try_to_freeze();
> if (kthread_should_stop())
> @@ -3352,12 +3344,19 @@ static int kswapd(void *p)
> * We can speed up thawing tasks if we don't call balance_pgdat
> * after returning from the refrigerator
> */
> - if (!ret) {
> - trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
> + if (ret)
> + continue;
>
> - /* return value ignored until next patch */
> - balance_pgdat(pgdat, order, classzone_idx);
> - }
> + /*
> + * Try reclaim the requested order but if that fails
> + * then try sleeping on the basis of the order reclaimed.
Is the last word really meant to be "reclaimed", or "requested"?
> + */
> + trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
> + if (balance_pgdat(pgdat, order, classzone_idx) < order)
> + goto kswapd_try_sleep;
AFAICS now kswapd_try_to_sleep() will use the "requested" order. That's
needed for proper wakeup_kcompactd(), but won't it prevent kswapd from
actually going to sleep, because zone_balanced() in prepare-sleep will
be false? So I think you need to give it both orders to do the right thing?
> +
> + order = pgdat->kswapd_order;
> + classzone_idx = pgdat->kswapd_classzone_idx;
> }
>
> tsk->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD);
> @@ -3380,10 +3379,10 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
> if (!cpuset_zone_allowed(zone, GFP_KERNEL | __GFP_HARDWALL))
> return;
> pgdat = zone->zone_pgdat;
> - if (pgdat->kswapd_max_order < order) {
> - pgdat->kswapd_max_order = order;
> - pgdat->classzone_idx = min(pgdat->classzone_idx, classzone_idx);
> - }
> + if (pgdat->kswapd_classzone_idx == -1)
> + pgdat->kswapd_classzone_idx = classzone_idx;
> + pgdat->kswapd_classzone_idx = max(pgdat->kswapd_classzone_idx, classzone_idx);
> + pgdat->kswapd_order = max(pgdat->kswapd_order, order);
> if (!waitqueue_active(&pgdat->kswapd_wait))
> return;
> if (zone_balanced(zone, order, 0))
>
^ permalink raw reply [flat|nested] 70+ messages in thread* Re: [PATCH 08/27] mm, vmscan: Simplify the logic deciding whether kswapd sleeps
2016-06-15 15:18 ` Vlastimil Babka
@ 2016-06-16 8:30 ` Mel Gorman
2016-06-16 9:16 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-16 8:30 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Wed, Jun 15, 2016 at 05:18:00PM +0200, Vlastimil Babka wrote:
> >@@ -1209,9 +1209,10 @@ static pg_data_t __ref *hotadd_new_pgdat(int nid, u64 start)
> >
> > arch_refresh_nodedata(nid, pgdat);
> > } else {
> >- /* Reset the nr_zones and classzone_idx to 0 before reuse */
> >+ /* Reset the nr_zones, order and classzone_idx before reuse */
> > pgdat->nr_zones = 0;
> >- pgdat->classzone_idx = 0;
> >+ pgdat->kswapd_order = 0;
> >+ pgdat->kswapd_classzone_idx = -1;
> > }
> >
> > /* we can use NODE_DATA(nid) from here */
> >diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> >index 4ce578b969da..d8cb483d5cad 100644
> >--- a/mm/page_alloc.c
> >+++ b/mm/page_alloc.c
> >@@ -6036,7 +6036,7 @@ void __paginginit free_area_init_node(int nid, unsigned long *zones_size,
> > unsigned long end_pfn = 0;
> >
> > /* pg_data_t should be reset to zero when it's allocated */
> >- WARN_ON(pgdat->nr_zones || pgdat->classzone_idx);
> >+ WARN_ON(pgdat->nr_zones || pgdat->kswapd_classzone_idx);
>
> Above you changed the reset value of kswapd_classzone_idx from 0 to -1, so
> won't this trigger? Also should we check kswapd_order that's newly reset
> too?
>
Good spot. The memory initialisation paths are ok but node memory hotplug
was broken.
> >
> > reset_deferred_meminit(pgdat);
> > pgdat->node_id = nid;
> >diff --git a/mm/vmscan.c b/mm/vmscan.c
> >index 96bf841f9352..14b34eebedff 100644
> >--- a/mm/vmscan.c
> >+++ b/mm/vmscan.c
> >@@ -2727,7 +2727,7 @@ static bool pfmemalloc_watermark_ok(pg_data_t *pgdat)
> >
> > /* kswapd must be awake if processes are being throttled */
> > if (!wmark_ok && waitqueue_active(&pgdat->kswapd_wait)) {
> >- pgdat->classzone_idx = min(pgdat->classzone_idx,
> >+ pgdat->kswapd_classzone_idx = min(pgdat->kswapd_classzone_idx,
> > (enum zone_type)ZONE_NORMAL);
> > wake_up_interruptible(&pgdat->kswapd_wait);
> > }
> >@@ -3211,6 +3211,12 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
> >
> > prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
> >
> >+ /* If kswapd has not been woken recently, then full sleep */
> >+ if (classzone_idx == -1) {
> >+ classzone_idx = balanced_classzone_idx = MAX_NR_ZONES - 1;
> >+ goto full_sleep;
>
> This will skip the wakeup_kcompactd() part.
>
I wrestled with this one. I decided to leave it alone on the grounds
that if kswapd has not been woken recently then compaction efforts also
have not failed and kcompactd is not required.
> >@@ -3311,38 +3316,25 @@ static int kswapd(void *p)
> > tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD;
> > set_freezable();
> >
> >- order = new_order = 0;
> >- classzone_idx = new_classzone_idx = pgdat->nr_zones - 1;
> >- balanced_classzone_idx = classzone_idx;
> >+ pgdat->kswapd_order = order = 0;
> >+ pgdat->kswapd_classzone_idx = classzone_idx = -1;
> > for ( ; ; ) {
> > bool ret;
> >
> >+kswapd_try_sleep:
> >+ kswapd_try_to_sleep(pgdat, order, classzone_idx, classzone_idx);
>
> The last two parameters are now the same, remove one?
>
Yes. A few more basic simplifications are then possible.
> >@@ -3352,12 +3344,19 @@ static int kswapd(void *p)
> > * We can speed up thawing tasks if we don't call balance_pgdat
> > * after returning from the refrigerator
> > */
> >- if (!ret) {
> >- trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
> >+ if (ret)
> >+ continue;
> >
> >- /* return value ignored until next patch */
> >- balance_pgdat(pgdat, order, classzone_idx);
> >- }
> >+ /*
> >+ * Try reclaim the requested order but if that fails
> >+ * then try sleeping on the basis of the order reclaimed.
>
> Is the last word really meant to be "reclaimed", or "requested"?
>
No, I really meant reclaimed.
> >+ */
> >+ trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
> >+ if (balance_pgdat(pgdat, order, classzone_idx) < order)
> >+ goto kswapd_try_sleep;
>
> AFAICS now kswapd_try_to_sleep() will use the "requested" order. That's
> needed for proper wakeup_kcompactd(), but won't it prevent kswapd from
> actually going to sleep, because zone_balanced() in prepare-sleep will be
> false? So I think you need to give it both orders to do the right thing?
>
You're right. There is a risk that kswapd stays awake longer in high
fragmentation scenarios.
Should be fixed now by passing in both orders.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread* Re: [PATCH 08/27] mm, vmscan: Simplify the logic deciding whether kswapd sleeps
2016-06-16 8:30 ` Mel Gorman
@ 2016-06-16 9:16 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-16 9:16 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Thu, Jun 16, 2016 at 09:30:33AM +0100, Mel Gorman wrote:
> > >@@ -2727,7 +2727,7 @@ static bool pfmemalloc_watermark_ok(pg_data_t *pgdat)
> > >
> > > /* kswapd must be awake if processes are being throttled */
> > > if (!wmark_ok && waitqueue_active(&pgdat->kswapd_wait)) {
> > >- pgdat->classzone_idx = min(pgdat->classzone_idx,
> > >+ pgdat->kswapd_classzone_idx = min(pgdat->kswapd_classzone_idx,
> > > (enum zone_type)ZONE_NORMAL);
> > > wake_up_interruptible(&pgdat->kswapd_wait);
> > > }
> > >@@ -3211,6 +3211,12 @@ static void kswapd_try_to_sleep(pg_data_t *pgdat, int order,
> > >
> > > prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE);
> > >
> > >+ /* If kswapd has not been woken recently, then full sleep */
> > >+ if (classzone_idx == -1) {
> > >+ classzone_idx = balanced_classzone_idx = MAX_NR_ZONES - 1;
> > >+ goto full_sleep;
> >
> > This will skip the wakeup_kcompactd() part.
> >
>
> I wrestled with this one. I decided to leave it alone on the grounds
> that if kswapd has not been woken recently then compaction efforts also
> have not failed and kcompactd is not required.
>
And I was wrong. There needs to be a call to wakeup_kcompactd there to
cover the case where there was a single high-order allocation request
that failed and woke kswapd.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 09/27] mm, vmscan: By default have direct reclaim only shrink once per node
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (7 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 08/27] mm, vmscan: Simplify the logic deciding whether kswapd sleeps Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-16 8:59 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 10/27] mm, vmscan: Clear congestion, dirty and need for compaction on a per-node basis Mel Gorman
` (17 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Direct reclaim iterates over all zones in the zonelist and shrinking them
but this is in conflict with node-based reclaim. In the default case,
only shrink once per node.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
---
mm/vmscan.c | 20 +++++++++++---------
1 file changed, 11 insertions(+), 9 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 14b34eebedff..dd68e3154732 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2522,14 +2522,6 @@ static inline bool compaction_ready(struct zone *zone, int order, int classzone_
* try to reclaim pages from zones which will satisfy the caller's allocation
* request.
*
- * We reclaim from a zone even if that zone is over high_wmark_pages(zone).
- * Because:
- * a) The caller may be trying to free *extra* pages to satisfy a higher-order
- * allocation or
- * b) The target zone may be at high_wmark_pages(zone) but the lower zones
- * must go *over* high_wmark_pages(zone) to satisfy the `incremental min'
- * zone defense algorithm.
- *
* If a zone is deemed to be full of pinned pages then just give it a light
* scan then give up on it.
*/
@@ -2541,6 +2533,7 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc,
unsigned long nr_soft_reclaimed;
unsigned long nr_soft_scanned;
gfp_t orig_mask;
+ pg_data_t *last_pgdat = NULL;
/*
* If the number of buffer_heads in the machine exceeds the maximum
@@ -2560,10 +2553,19 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc,
classzone_idx)) {
sc->reclaim_idx--;
classzone_idx--;
- continue;
}
/*
+ * Shrink each node in the zonelist once. If the zonelist is
+ * ordered by zone (not the default) then a node may be
+ * shrunk multiple times but in that case the user prefers
+ * lower zones being preserved
+ */
+ if (zone->zone_pgdat == last_pgdat)
+ continue;
+ last_pgdat = zone->zone_pgdat;
+
+ /*
* Take care memory controller reclaiming has small influence
* to global LRU.
*/
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* [PATCH 10/27] mm, vmscan: Clear congestion, dirty and need for compaction on a per-node basis
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (8 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 09/27] mm, vmscan: By default have direct reclaim only shrink once per node Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-16 9:29 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 11/27] mm: vmscan: Do not reclaim from kswapd if there is any eligible zone Mel Gorman
` (16 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Congested and dirty tracking of a node and whether reclaim should stall
is still based on zone activity. This patch considers whether the kernel
should stall based on node-based reclaim activity.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/vmscan.c | 24 ++++++++++++------------
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index dd68e3154732..e4f3e068b7a0 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2966,7 +2966,17 @@ static bool zone_balanced(struct zone *zone, int order, int classzone_idx)
{
unsigned long mark = high_wmark_pages(zone);
- return zone_watermark_ok_safe(zone, order, mark, classzone_idx);
+ if (!zone_watermark_ok_safe(zone, order, mark, classzone_idx))
+ return false;
+
+ /*
+ * If any eligible zone is balanced then the node is not considered
+ * to be congested or dirty
+ */
+ clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
+ clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
+
+ return true;
}
/*
@@ -3112,13 +3122,6 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
if (!zone_balanced(zone, order, 0)) {
classzone_idx = i;
break;
- } else {
- /*
- * If any eligible zone is balanced then the
- * node is not considered congested or dirty.
- */
- clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
- clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
}
}
@@ -3177,11 +3180,8 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
if (!populated_zone(zone))
continue;
- if (zone_balanced(zone, sc.order, classzone_idx)) {
- clear_bit(PGDAT_CONGESTED, &pgdat->flags);
- clear_bit(PGDAT_DIRTY, &pgdat->flags);
+ if (zone_balanced(zone, sc.order, classzone_idx))
goto out;
- }
}
/*
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 10/27] mm, vmscan: Clear congestion, dirty and need for compaction on a per-node basis
2016-06-09 18:04 ` [PATCH 10/27] mm, vmscan: Clear congestion, dirty and need for compaction on a per-node basis Mel Gorman
@ 2016-06-16 9:29 ` Vlastimil Babka
2016-06-16 10:29 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-16 9:29 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> Congested and dirty tracking of a node and whether reclaim should stall
> is still based on zone activity. This patch considers whether the kernel
> should stall based on node-based reclaim activity.
I'm a bit confused about the description vs actual code.
It appears to move some duplicated code to a related function, which is
fine. The rest of callsites that didn't perform the clearing before
(prepare_kswapd_sleep() and wakeup_kswapd()) might be a bit overkill,
but won't hurt. But I don't see the part "considers whether the kernel
should stall based on node-based reclaim activity". Is something missing?
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> ---
> mm/vmscan.c | 24 ++++++++++++------------
> 1 file changed, 12 insertions(+), 12 deletions(-)
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index dd68e3154732..e4f3e068b7a0 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -2966,7 +2966,17 @@ static bool zone_balanced(struct zone *zone, int order, int classzone_idx)
> {
> unsigned long mark = high_wmark_pages(zone);
>
> - return zone_watermark_ok_safe(zone, order, mark, classzone_idx);
> + if (!zone_watermark_ok_safe(zone, order, mark, classzone_idx))
> + return false;
> +
> + /*
> + * If any eligible zone is balanced then the node is not considered
> + * to be congested or dirty
> + */
> + clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
> + clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
> +
> + return true;
> }
>
> /*
> @@ -3112,13 +3122,6 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
> if (!zone_balanced(zone, order, 0)) {
> classzone_idx = i;
> break;
> - } else {
> - /*
> - * If any eligible zone is balanced then the
> - * node is not considered congested or dirty.
> - */
> - clear_bit(PGDAT_CONGESTED, &zone->zone_pgdat->flags);
> - clear_bit(PGDAT_DIRTY, &zone->zone_pgdat->flags);
> }
> }
>
> @@ -3177,11 +3180,8 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
> if (!populated_zone(zone))
> continue;
>
> - if (zone_balanced(zone, sc.order, classzone_idx)) {
> - clear_bit(PGDAT_CONGESTED, &pgdat->flags);
> - clear_bit(PGDAT_DIRTY, &pgdat->flags);
> + if (zone_balanced(zone, sc.order, classzone_idx))
> goto out;
> - }
> }
>
> /*
>
^ permalink raw reply [flat|nested] 70+ messages in thread* Re: [PATCH 10/27] mm, vmscan: Clear congestion, dirty and need for compaction on a per-node basis
2016-06-16 9:29 ` Vlastimil Babka
@ 2016-06-16 10:29 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-16 10:29 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Thu, Jun 16, 2016 at 11:29:00AM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> >Congested and dirty tracking of a node and whether reclaim should stall
> >is still based on zone activity. This patch considers whether the kernel
> >should stall based on node-based reclaim activity.
>
> I'm a bit confused about the description vs actual code.
> It appears to move some duplicated code to a related function, which is
> fine. The rest of callsites that didn't perform the clearing before
> (prepare_kswapd_sleep() and wakeup_kswapd()) might be a bit overkill, but
> won't hurt. But I don't see the part "considers whether the kernel
> should stall based on node-based reclaim activity". Is something missing?
>
Tired when writing the changelog. Does this make more sense?
mm, vmscan: Remove duplicate logic clearing node congestion and dirty state
Reclaim may stall if there is too much dirty or congested data on a node.
This was previously based on zone flags and the logic for clearing the
flags is in two places. As congestion/dirty tracking is now tracked on
a per-node basis, we can remove some duplicate logic.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 11/27] mm: vmscan: Do not reclaim from kswapd if there is any eligible zone
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (9 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 10/27] mm, vmscan: Clear congestion, dirty and need for compaction on a per-node basis Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-16 10:08 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 12/27] mm, vmscan: Make shrink_node decisions more node-centric Mel Gorman
` (15 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
kswapd scans from highest to lowest for a zone that requires balancing.
This was necessary when reclaim was per-zone to fairly age pages on
lower zones. Now that we are reclaiming on a per-node basis, any eligible
zone can be used and pages will still be aged fairly. This patch avoids
reclaiming excessively unless buffer_heads are over the limit and it's
necessary to reclaim from a higher zone than requested by the waker of
kswapd to relieve low memory pressure.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/vmscan.c | 32 +++++++++++++++++++-------------
1 file changed, 19 insertions(+), 13 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index e4f3e068b7a0..6663fc75c3bc 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3102,24 +3102,30 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
sc.nr_reclaimed = 0;
- /* Scan from the highest requested zone to dma */
+ /*
+ * If the number of buffer_heads in the machine exceeds the
+ * maximum allowed level and this node has a highmem zone,
+ * force kswapd to reclaim from it to relieve lowmem pressure.
+ */
+ if (buffer_heads_over_limit) {
+ for (i = MAX_NR_ZONES - 1; i >= 0; i++) {
+ zone = pgdat->node_zones + i;
+ if (!populated_zone(zone))
+ continue;
+
+ if (is_highmem_idx(i))
+ classzone_idx = i;
+ break;
+ }
+ }
+
+ /* Only reclaim if there are no eligible zones */
for (i = classzone_idx; i >= 0; i--) {
zone = pgdat->node_zones + i;
if (!populated_zone(zone))
continue;
- /*
- * If the number of buffer_heads in the machine
- * exceeds the maximum allowed level and this node
- * has a highmem zone, force kswapd to reclaim from
- * it to relieve lowmem pressure.
- */
- if (buffer_heads_over_limit && is_highmem_idx(i)) {
- classzone_idx = i;
- break;
- }
-
- if (!zone_balanced(zone, order, 0)) {
+ if (!zone_balanced(zone, sc.order, classzone_idx)) {
classzone_idx = i;
break;
}
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 11/27] mm: vmscan: Do not reclaim from kswapd if there is any eligible zone
2016-06-09 18:04 ` [PATCH 11/27] mm: vmscan: Do not reclaim from kswapd if there is any eligible zone Mel Gorman
@ 2016-06-16 10:08 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-16 10:08 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> kswapd scans from highest to lowest for a zone that requires balancing.
> This was necessary when reclaim was per-zone to fairly age pages on
> lower zones. Now that we are reclaiming on a per-node basis, any eligible
> zone can be used and pages will still be aged fairly. This patch avoids
> reclaiming excessively unless buffer_heads are over the limit and it's
> necessary to reclaim from a higher zone than requested by the waker of
> kswapd to relieve low memory pressure.
Looks like the code was even wrong before... if classzone_idx wasn't
already set to the highmem zone in the first place, it wouldn't look at it.
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
After fixing the bug below,
Acked-by: Vlastimil Babka <vbabka@suse.cz>
> ---
> mm/vmscan.c | 32 +++++++++++++++++++-------------
> 1 file changed, 19 insertions(+), 13 deletions(-)
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index e4f3e068b7a0..6663fc75c3bc 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -3102,24 +3102,30 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
>
> sc.nr_reclaimed = 0;
>
> - /* Scan from the highest requested zone to dma */
> + /*
> + * If the number of buffer_heads in the machine exceeds the
> + * maximum allowed level and this node has a highmem zone,
> + * force kswapd to reclaim from it to relieve lowmem pressure.
> + */
> + if (buffer_heads_over_limit) {
> + for (i = MAX_NR_ZONES - 1; i >= 0; i++) {
i--
> + zone = pgdat->node_zones + i;
> + if (!populated_zone(zone))
> + continue;
> +
> + if (is_highmem_idx(i))
> + classzone_idx = i;
> + break;
> + }
> + }
> +
> + /* Only reclaim if there are no eligible zones */
> for (i = classzone_idx; i >= 0; i--) {
> zone = pgdat->node_zones + i;
> if (!populated_zone(zone))
> continue;
>
> - /*
> - * If the number of buffer_heads in the machine
> - * exceeds the maximum allowed level and this node
> - * has a highmem zone, force kswapd to reclaim from
> - * it to relieve lowmem pressure.
> - */
> - if (buffer_heads_over_limit && is_highmem_idx(i)) {
> - classzone_idx = i;
> - break;
> - }
> -
> - if (!zone_balanced(zone, order, 0)) {
> + if (!zone_balanced(zone, sc.order, classzone_idx)) {
> classzone_idx = i;
> break;
> }
>
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 12/27] mm, vmscan: Make shrink_node decisions more node-centric
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (10 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 11/27] mm: vmscan: Do not reclaim from kswapd if there is any eligible zone Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-16 13:35 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 13/27] mm, memcg: Move memcg limit enforcement from zones to nodes Mel Gorman
` (14 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Earlier patches focused on having direct reclaim and kswapd use data that
is node-centric for reclaiming but shrink_node() itself still uses too much
zone information. This patch removes unnecessary zone-based information
with the most important decision being whether to continue reclaim or
not. Some memcg APIs are adjusted as a result even though memcg itself
still uses some zone information.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
include/linux/memcontrol.h | 9 +++----
include/linux/mmzone.h | 4 ++--
include/linux/swap.h | 2 +-
mm/memcontrol.c | 17 +++++++-------
mm/page_alloc.c | 2 +-
mm/vmscan.c | 58 ++++++++++++++++++++++++++--------------------
mm/workingset.c | 6 ++---
7 files changed, 54 insertions(+), 44 deletions(-)
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index 9adb1bebb487..beee5be5d5d5 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -306,7 +306,8 @@ void mem_cgroup_uncharge_list(struct list_head *page_list);
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage);
-struct lruvec *mem_cgroup_zone_lruvec(struct zone *, struct mem_cgroup *);
+struct lruvec *mem_cgroup_lruvec(struct pglist_data *, struct zone *zone,
+ struct mem_cgroup *);
struct lruvec *mem_cgroup_page_lruvec(struct page *, struct pglist_data *);
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg);
@@ -573,10 +574,10 @@ static inline void mem_cgroup_migrate(struct page *old, struct page *new)
{
}
-static inline struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone,
- struct mem_cgroup *memcg)
+static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
+ struct zone *zone, struct mem_cgroup *memcg)
{
- return zone_lruvec(zone);
+ return node_lruvec(pgdat);
}
static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 15ce454d0d59..16da23c095a0 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -735,9 +735,9 @@ static inline spinlock_t *zone_lru_lock(struct zone *zone)
return &zone->zone_pgdat->lru_lock;
}
-static inline struct lruvec *zone_lruvec(struct zone *zone)
+static inline struct lruvec *node_lruvec(struct pglist_data *pgdat)
{
- return &zone->zone_pgdat->lruvec;
+ return &pgdat->lruvec;
}
static inline unsigned long pgdat_end_pfn(pg_data_t *pgdat)
diff --git a/include/linux/swap.h b/include/linux/swap.h
index cc5c958c050b..575efae79a66 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -324,7 +324,7 @@ extern unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
unsigned long nr_pages,
gfp_t gfp_mask,
bool may_swap);
-extern unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *mem,
+extern unsigned long mem_cgroup_shrink_node(struct mem_cgroup *mem,
gfp_t gfp_mask, bool noswap,
struct zone *zone,
unsigned long *nr_scanned);
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 12f796e52e5e..cda2342af10e 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -944,22 +944,23 @@ static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg)
iter = mem_cgroup_iter(NULL, iter, NULL))
/**
- * mem_cgroup_zone_lruvec - get the lru list vector for a zone and memcg
+ * mem_cgroup_lruvec - get the lru list vector for a node or a memcg zone
+ * @node: node of the wanted lruvec
* @zone: zone of the wanted lruvec
* @memcg: memcg of the wanted lruvec
*
- * Returns the lru list vector holding pages for the given @zone and
- * @mem. This can be the global zone lruvec, if the memory controller
+ * Returns the lru list vector holding pages for a given @node or a given
+ * @memcg and @zone. This can be the node lruvec, if the memory controller
* is disabled.
*/
-struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone,
- struct mem_cgroup *memcg)
+struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
+ struct zone *zone, struct mem_cgroup *memcg)
{
struct mem_cgroup_per_zone *mz;
struct lruvec *lruvec;
if (mem_cgroup_disabled()) {
- lruvec = zone_lruvec(zone);
+ lruvec = node_lruvec(pgdat);
goto out;
}
@@ -1473,8 +1474,8 @@ static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg,
}
continue;
}
- total += mem_cgroup_shrink_node_zone(victim, gfp_mask, false,
- zone, &nr_scanned);
+ total += mem_cgroup_shrink_node(victim, gfp_mask, false,
+ zone, &nr_scanned);
*total_scanned += nr_scanned;
if (!soft_limit_excess(root_memcg))
break;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index d8cb483d5cad..fc5bad895732 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -5905,6 +5905,7 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
init_waitqueue_head(&pgdat->kcompactd_wait);
#endif
pgdat_page_ext_init(pgdat);
+ lruvec_init(node_lruvec(pgdat));
for (j = 0; j < MAX_NR_ZONES; j++) {
struct zone *zone = pgdat->node_zones + j;
@@ -5968,7 +5969,6 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
/* For bootup, initialized properly in watermark setup */
mod_zone_page_state(zone, NR_ALLOC_BATCH, zone->managed_pages);
- lruvec_init(zone_lruvec(zone));
if (!size)
continue;
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 6663fc75c3bc..93523944abb0 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2196,10 +2196,11 @@ static inline void init_tlb_ubc(void)
/*
* This is a basic per-zone page freer. Used by both kswapd and direct reclaim.
*/
-static void shrink_zone_memcg(struct zone *zone, struct mem_cgroup *memcg,
+static void shrink_node_memcg(struct pglist_data *pgdat, struct mem_cgroup *memcg,
struct scan_control *sc, unsigned long *lru_pages)
{
- struct lruvec *lruvec = mem_cgroup_zone_lruvec(zone, memcg);
+ struct zone *zone = &pgdat->node_zones[sc->reclaim_idx];
+ struct lruvec *lruvec = mem_cgroup_lruvec(pgdat, zone, memcg);
unsigned long nr[NR_LRU_LISTS];
unsigned long targets[NR_LRU_LISTS];
unsigned long nr_to_scan;
@@ -2332,13 +2333,14 @@ static bool in_reclaim_compaction(struct scan_control *sc)
* calls try_to_compact_zone() that it will have enough free pages to succeed.
* It will give up earlier than that if there is difficulty reclaiming pages.
*/
-static inline bool should_continue_reclaim(struct zone *zone,
+static inline bool should_continue_reclaim(struct pglist_data *pgdat,
unsigned long nr_reclaimed,
unsigned long nr_scanned,
struct scan_control *sc)
{
unsigned long pages_for_compaction;
unsigned long inactive_lru_pages;
+ int z;
/* If not in reclaim/compaction mode, stop */
if (!in_reclaim_compaction(sc))
@@ -2372,21 +2374,27 @@ static inline bool should_continue_reclaim(struct zone *zone,
* inactive lists are large enough, continue reclaiming
*/
pages_for_compaction = (2UL << sc->order);
- inactive_lru_pages = node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE);
+ inactive_lru_pages = node_page_state(pgdat, NR_INACTIVE_FILE);
if (get_nr_swap_pages() > 0)
- inactive_lru_pages += node_page_state(zone->zone_pgdat, NR_INACTIVE_ANON);
+ inactive_lru_pages += node_page_state(pgdat, NR_INACTIVE_ANON);
if (sc->nr_reclaimed < pages_for_compaction &&
inactive_lru_pages > pages_for_compaction)
return true;
/* If compaction would go ahead or the allocation would succeed, stop */
- switch (compaction_suitable(zone, sc->order, 0, 0)) {
- case COMPACT_PARTIAL:
- case COMPACT_CONTINUE:
- return false;
- default:
- return true;
+ for (z = 0; z <= sc->reclaim_idx; z++) {
+ struct zone *zone = &pgdat->node_zones[z];
+
+ switch (compaction_suitable(zone, sc->order, 0, 0)) {
+ case COMPACT_PARTIAL:
+ case COMPACT_CONTINUE:
+ return false;
+ default:
+ /* check next zone */
+ ;
+ }
}
+ return true;
}
static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
@@ -2395,15 +2403,14 @@ static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
struct reclaim_state *reclaim_state = current->reclaim_state;
unsigned long nr_reclaimed, nr_scanned;
bool reclaimable = false;
- struct zone *zone = &pgdat->node_zones[classzone_idx];
do {
struct mem_cgroup *root = sc->target_mem_cgroup;
struct mem_cgroup_reclaim_cookie reclaim = {
- .zone = zone,
+ .zone = &pgdat->node_zones[classzone_idx],
.priority = sc->priority,
};
- unsigned long zone_lru_pages = 0;
+ unsigned long node_lru_pages = 0;
struct mem_cgroup *memcg;
nr_reclaimed = sc->nr_reclaimed;
@@ -2424,11 +2431,11 @@ static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
reclaimed = sc->nr_reclaimed;
scanned = sc->nr_scanned;
- shrink_zone_memcg(zone, memcg, sc, &lru_pages);
- zone_lru_pages += lru_pages;
+ shrink_node_memcg(pgdat, memcg, sc, &lru_pages);
+ node_lru_pages += lru_pages;
if (!global_reclaim(sc) && sc->reclaim_idx == classzone_idx)
- shrink_slab(sc->gfp_mask, zone_to_nid(zone),
+ shrink_slab(sc->gfp_mask, pgdat->node_id,
memcg, sc->nr_scanned - scanned,
lru_pages);
@@ -2440,7 +2447,7 @@ static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
/*
* Direct reclaim and kswapd have to scan all memory
* cgroups to fulfill the overall scan target for the
- * zone.
+ * node.
*
* Limit reclaim, on the other hand, only cares about
* nr_to_reclaim pages to be reclaimed and it will
@@ -2459,9 +2466,9 @@ static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
* the eligible LRU pages were scanned.
*/
if (global_reclaim(sc) && sc->reclaim_idx == classzone_idx)
- shrink_slab(sc->gfp_mask, zone_to_nid(zone), NULL,
+ shrink_slab(sc->gfp_mask, pgdat->node_id, NULL,
sc->nr_scanned - nr_scanned,
- zone_lru_pages);
+ node_lru_pages);
if (reclaim_state) {
sc->nr_reclaimed += reclaim_state->reclaimed_slab;
@@ -2476,7 +2483,7 @@ static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
if (sc->nr_reclaimed - nr_reclaimed)
reclaimable = true;
- } while (should_continue_reclaim(zone, sc->nr_reclaimed - nr_reclaimed,
+ } while (should_continue_reclaim(pgdat, sc->nr_reclaimed - nr_reclaimed,
sc->nr_scanned - nr_scanned, sc));
return reclaimable;
@@ -2866,7 +2873,7 @@ unsigned long try_to_free_pages(struct zonelist *zonelist, int order,
#ifdef CONFIG_MEMCG
-unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *memcg,
+unsigned long mem_cgroup_shrink_node(struct mem_cgroup *memcg,
gfp_t gfp_mask, bool noswap,
struct zone *zone,
unsigned long *nr_scanned)
@@ -2876,6 +2883,7 @@ unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *memcg,
.target_mem_cgroup = memcg,
.may_writepage = !laptop_mode,
.may_unmap = 1,
+ .reclaim_idx = zone_idx(zone),
.may_swap = !noswap,
};
unsigned long lru_pages;
@@ -2890,11 +2898,11 @@ unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *memcg,
/*
* NOTE: Although we can get the priority field, using it
* here is not a good idea, since it limits the pages we can scan.
- * if we don't reclaim here, the shrink_zone from balance_pgdat
+ * if we don't reclaim here, the shrink_node from balance_pgdat
* will pick up pages from other mem cgroup's as well. We hack
* the priority and make it zero.
*/
- shrink_zone_memcg(zone, memcg, &sc, &lru_pages);
+ shrink_node_memcg(zone->zone_pgdat, memcg, &sc, &lru_pages);
trace_mm_vmscan_memcg_softlimit_reclaim_end(sc.nr_reclaimed);
@@ -2952,7 +2960,7 @@ static void age_active_anon(struct pglist_data *pgdat,
memcg = mem_cgroup_iter(NULL, NULL, NULL);
do {
- struct lruvec *lruvec = mem_cgroup_zone_lruvec(zone, memcg);
+ struct lruvec *lruvec = mem_cgroup_lruvec(pgdat, zone, memcg);
if (inactive_list_is_low(lruvec, false))
shrink_active_list(SWAP_CLUSTER_MAX, lruvec,
diff --git a/mm/workingset.c b/mm/workingset.c
index c0820e06aaff..10ddf707782a 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -218,7 +218,7 @@ void *workingset_eviction(struct address_space *mapping, struct page *page)
VM_BUG_ON_PAGE(page_count(page), page);
VM_BUG_ON_PAGE(!PageLocked(page), page);
- lruvec = mem_cgroup_zone_lruvec(zone, memcg);
+ lruvec = mem_cgroup_lruvec(zone->zone_pgdat, zone, memcg);
eviction = atomic_long_inc_return(&lruvec->inactive_age);
return pack_shadow(memcgid, zone, eviction);
}
@@ -267,7 +267,7 @@ bool workingset_refault(void *shadow)
rcu_read_unlock();
return false;
}
- lruvec = mem_cgroup_zone_lruvec(zone, memcg);
+ lruvec = mem_cgroup_lruvec(zone->zone_pgdat, zone, memcg);
refault = atomic_long_read(&lruvec->inactive_age);
active_file = lruvec_lru_size(lruvec, LRU_ACTIVE_FILE);
rcu_read_unlock();
@@ -317,7 +317,7 @@ void workingset_activation(struct page *page)
*/
if (!mem_cgroup_disabled() && !page_memcg(page))
goto out;
- lruvec = mem_cgroup_zone_lruvec(page_zone(page), page_memcg(page));
+ lruvec = mem_cgroup_lruvec(page_zone(page)->zone_pgdat, page_zone(page), page_memcg(page));
atomic_long_inc(&lruvec->inactive_age);
out:
unlock_page_memcg(page);
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 12/27] mm, vmscan: Make shrink_node decisions more node-centric
2016-06-09 18:04 ` [PATCH 12/27] mm, vmscan: Make shrink_node decisions more node-centric Mel Gorman
@ 2016-06-16 13:35 ` Vlastimil Babka
2016-06-16 14:47 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-16 13:35 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> Earlier patches focused on having direct reclaim and kswapd use data that
> is node-centric for reclaiming but shrink_node() itself still uses too much
> zone information. This patch removes unnecessary zone-based information
> with the most important decision being whether to continue reclaim or
> not. Some memcg APIs are adjusted as a result even though memcg itself
> still uses some zone information.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
[...]
> @@ -2372,21 +2374,27 @@ static inline bool should_continue_reclaim(struct zone *zone,
> * inactive lists are large enough, continue reclaiming
> */
> pages_for_compaction = (2UL << sc->order);
> - inactive_lru_pages = node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE);
> + inactive_lru_pages = node_page_state(pgdat, NR_INACTIVE_FILE);
> if (get_nr_swap_pages() > 0)
> - inactive_lru_pages += node_page_state(zone->zone_pgdat, NR_INACTIVE_ANON);
> + inactive_lru_pages += node_page_state(pgdat, NR_INACTIVE_ANON);
> if (sc->nr_reclaimed < pages_for_compaction &&
> inactive_lru_pages > pages_for_compaction)
> return true;
>
> /* If compaction would go ahead or the allocation would succeed, stop */
> - switch (compaction_suitable(zone, sc->order, 0, 0)) {
> - case COMPACT_PARTIAL:
> - case COMPACT_CONTINUE:
> - return false;
> - default:
> - return true;
> + for (z = 0; z <= sc->reclaim_idx; z++) {
> + struct zone *zone = &pgdat->node_zones[z];
> +
> + switch (compaction_suitable(zone, sc->order, 0, 0)) {
Using 0 for classzone_idx here was sort of OK when each zone was
reclaimed separately, as a Normal allocation not passing appropriate
classzone_idx (and thus subtracting lowmem reserve from free pages)
means that a false COMPACT_PARTIAL (or COMPACT_CONTINUE) could be
returned for e.g. DMA zone. It means a premature end of reclaim for this
single zone, which is relatively small anyway, so no big deal (and we
might avoid useless over-reclaim, when even reclaiming everything
wouldn't get us above the lowmem_reserve).
But in node-centric reclaim, such premature "return false" from a DMA
zone stops reclaiming the whole node. So I think we should involve the
real classzone_idx here.
> + case COMPACT_PARTIAL:
> + case COMPACT_CONTINUE:
> + return false;
> + default:
> + /* check next zone */
> + ;
> + }
> }
> + return true;
> }
>
^ permalink raw reply [flat|nested] 70+ messages in thread* Re: [PATCH 12/27] mm, vmscan: Make shrink_node decisions more node-centric
2016-06-16 13:35 ` Vlastimil Babka
@ 2016-06-16 14:47 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-16 14:47 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Thu, Jun 16, 2016 at 03:35:15PM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> >Earlier patches focused on having direct reclaim and kswapd use data that
> >is node-centric for reclaiming but shrink_node() itself still uses too much
> >zone information. This patch removes unnecessary zone-based information
> >with the most important decision being whether to continue reclaim or
> >not. Some memcg APIs are adjusted as a result even though memcg itself
> >still uses some zone information.
> >
> >Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
>
> [...]
>
> >@@ -2372,21 +2374,27 @@ static inline bool should_continue_reclaim(struct zone *zone,
> > * inactive lists are large enough, continue reclaiming
> > */
> > pages_for_compaction = (2UL << sc->order);
> >- inactive_lru_pages = node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE);
> >+ inactive_lru_pages = node_page_state(pgdat, NR_INACTIVE_FILE);
> > if (get_nr_swap_pages() > 0)
> >- inactive_lru_pages += node_page_state(zone->zone_pgdat, NR_INACTIVE_ANON);
> >+ inactive_lru_pages += node_page_state(pgdat, NR_INACTIVE_ANON);
> > if (sc->nr_reclaimed < pages_for_compaction &&
> > inactive_lru_pages > pages_for_compaction)
> > return true;
> >
> > /* If compaction would go ahead or the allocation would succeed, stop */
> >- switch (compaction_suitable(zone, sc->order, 0, 0)) {
> >- case COMPACT_PARTIAL:
> >- case COMPACT_CONTINUE:
> >- return false;
> >- default:
> >- return true;
> >+ for (z = 0; z <= sc->reclaim_idx; z++) {
> >+ struct zone *zone = &pgdat->node_zones[z];
> >+
> >+ switch (compaction_suitable(zone, sc->order, 0, 0)) {
>
> Using 0 for classzone_idx here was sort of OK when each zone was reclaimed
> separately, as a Normal allocation not passing appropriate classzone_idx
> (and thus subtracting lowmem reserve from free pages) means that a false
> COMPACT_PARTIAL (or COMPACT_CONTINUE) could be returned for e.g. DMA zone.
> It means a premature end of reclaim for this single zone, which is
> relatively small anyway, so no big deal (and we might avoid useless
> over-reclaim, when even reclaiming everything wouldn't get us above the
> lowmem_reserve).
>
> But in node-centric reclaim, such premature "return false" from a DMA zone
> stops reclaiming the whole node. So I think we should involve the real
> classzone_idx here.
>
Fair point although for compaction, it'll occur for a marginal corner
case. Premature allowed compaction for ZONE_DMA is unfortunate but
bizarre to think there would be a high-order allocation restricted to
just that zone too.
I'll pass in sc->reclaim_idx as it represents the allocating order. That
highlights that direct reclaim was not setting reclaim_idx but that's
been corrected.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 13/27] mm, memcg: Move memcg limit enforcement from zones to nodes
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (11 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 12/27] mm, vmscan: Make shrink_node decisions more node-centric Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-16 15:06 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 14/27] mm, workingset: Make working set detection node-aware Mel Gorman
` (13 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Memcg was broken by the move of all LRUs to nodes because it is tracking
limits on a per-zone basis while receiving reclaim requests on a per-node
basis. This patch moves limit enforcement to the nodes. Technically, all
the variable names should also change but people are already familiar by
the meaning of "mz" even if "mn" would be a more appropriate name now.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
include/linux/memcontrol.h | 21 ++---
include/linux/swap.h | 2 +-
mm/memcontrol.c | 204 +++++++++++++++++++--------------------------
mm/vmscan.c | 21 +++--
mm/workingset.c | 6 +-
5 files changed, 110 insertions(+), 144 deletions(-)
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h
index beee5be5d5d5..1b47fdbc0e6d 100644
--- a/include/linux/memcontrol.h
+++ b/include/linux/memcontrol.h
@@ -60,7 +60,7 @@ enum mem_cgroup_stat_index {
};
struct mem_cgroup_reclaim_cookie {
- struct zone *zone;
+ pg_data_t *pgdat;
int priority;
unsigned int generation;
};
@@ -113,7 +113,7 @@ struct mem_cgroup_reclaim_iter {
/*
* per-zone information in memory controller.
*/
-struct mem_cgroup_per_zone {
+struct mem_cgroup_per_node {
struct lruvec lruvec;
unsigned long lru_size[NR_LRU_LISTS];
@@ -127,10 +127,6 @@ struct mem_cgroup_per_zone {
/* use container_of */
};
-struct mem_cgroup_per_node {
- struct mem_cgroup_per_zone zoneinfo[MAX_NR_ZONES];
-};
-
struct mem_cgroup_threshold {
struct eventfd_ctx *eventfd;
unsigned long threshold;
@@ -306,8 +302,7 @@ void mem_cgroup_uncharge_list(struct list_head *page_list);
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage);
-struct lruvec *mem_cgroup_lruvec(struct pglist_data *, struct zone *zone,
- struct mem_cgroup *);
+struct lruvec *mem_cgroup_lruvec(struct pglist_data *, struct mem_cgroup *);
struct lruvec *mem_cgroup_page_lruvec(struct page *, struct pglist_data *);
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg);
@@ -410,9 +405,9 @@ unsigned long mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
static inline
unsigned long mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
{
- struct mem_cgroup_per_zone *mz;
+ struct mem_cgroup_per_node *mz;
- mz = container_of(lruvec, struct mem_cgroup_per_zone, lruvec);
+ mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
return mz->lru_size[lru];
}
@@ -483,7 +478,7 @@ static inline void mem_cgroup_dec_page_stat(struct page *page,
mem_cgroup_update_page_stat(page, idx, -1);
}
-unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
+unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
gfp_t gfp_mask,
unsigned long *total_scanned);
@@ -575,7 +570,7 @@ static inline void mem_cgroup_migrate(struct page *old, struct page *new)
}
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
- struct zone *zone, struct mem_cgroup *memcg)
+ struct mem_cgroup *memcg)
{
return node_lruvec(pgdat);
}
@@ -687,7 +682,7 @@ static inline void mem_cgroup_dec_page_stat(struct page *page,
}
static inline
-unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
+unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
gfp_t gfp_mask,
unsigned long *total_scanned)
{
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 575efae79a66..5d1eac6259d2 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -326,7 +326,7 @@ extern unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
bool may_swap);
extern unsigned long mem_cgroup_shrink_node(struct mem_cgroup *mem,
gfp_t gfp_mask, bool noswap,
- struct zone *zone,
+ pg_data_t *pgdat,
unsigned long *nr_scanned);
extern unsigned long shrink_all_memory(unsigned long nr_pages);
extern int vm_swappiness;
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index cda2342af10e..79d0bd27fdb8 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -132,15 +132,11 @@ static const char * const mem_cgroup_lru_names[] = {
* their hierarchy representation
*/
-struct mem_cgroup_tree_per_zone {
+struct mem_cgroup_tree_per_node {
struct rb_root rb_root;
spinlock_t lock;
};
-struct mem_cgroup_tree_per_node {
- struct mem_cgroup_tree_per_zone rb_tree_per_zone[MAX_NR_ZONES];
-};
-
struct mem_cgroup_tree {
struct mem_cgroup_tree_per_node *rb_tree_per_node[MAX_NUMNODES];
};
@@ -323,13 +319,10 @@ EXPORT_SYMBOL(memcg_kmem_enabled_key);
#endif /* !CONFIG_SLOB */
-static struct mem_cgroup_per_zone *
-mem_cgroup_zone_zoneinfo(struct mem_cgroup *memcg, struct zone *zone)
+static struct mem_cgroup_per_node *
+mem_cgroup_nodeinfo(struct mem_cgroup *memcg, pg_data_t *pgdat)
{
- int nid = zone_to_nid(zone);
- int zid = zone_idx(zone);
-
- return &memcg->nodeinfo[nid]->zoneinfo[zid];
+ return memcg->nodeinfo[pgdat->node_id];
}
/**
@@ -383,37 +376,35 @@ ino_t page_cgroup_ino(struct page *page)
return ino;
}
-static struct mem_cgroup_per_zone *
+static struct mem_cgroup_per_node *
mem_cgroup_page_zoneinfo(struct mem_cgroup *memcg, struct page *page)
{
int nid = page_to_nid(page);
- int zid = page_zonenum(page);
- return &memcg->nodeinfo[nid]->zoneinfo[zid];
+ return memcg->nodeinfo[nid];
}
-static struct mem_cgroup_tree_per_zone *
-soft_limit_tree_node_zone(int nid, int zid)
+static struct mem_cgroup_tree_per_node *
+soft_limit_tree_node(int nid)
{
- return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid];
+ return soft_limit_tree.rb_tree_per_node[nid];
}
-static struct mem_cgroup_tree_per_zone *
+static struct mem_cgroup_tree_per_node *
soft_limit_tree_from_page(struct page *page)
{
int nid = page_to_nid(page);
- int zid = page_zonenum(page);
- return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid];
+ return soft_limit_tree.rb_tree_per_node[nid];
}
-static void __mem_cgroup_insert_exceeded(struct mem_cgroup_per_zone *mz,
- struct mem_cgroup_tree_per_zone *mctz,
+static void __mem_cgroup_insert_exceeded(struct mem_cgroup_per_node *mz,
+ struct mem_cgroup_tree_per_node *mctz,
unsigned long new_usage_in_excess)
{
struct rb_node **p = &mctz->rb_root.rb_node;
struct rb_node *parent = NULL;
- struct mem_cgroup_per_zone *mz_node;
+ struct mem_cgroup_per_node *mz_node;
if (mz->on_tree)
return;
@@ -423,7 +414,7 @@ static void __mem_cgroup_insert_exceeded(struct mem_cgroup_per_zone *mz,
return;
while (*p) {
parent = *p;
- mz_node = rb_entry(parent, struct mem_cgroup_per_zone,
+ mz_node = rb_entry(parent, struct mem_cgroup_per_node,
tree_node);
if (mz->usage_in_excess < mz_node->usage_in_excess)
p = &(*p)->rb_left;
@@ -439,8 +430,8 @@ static void __mem_cgroup_insert_exceeded(struct mem_cgroup_per_zone *mz,
mz->on_tree = true;
}
-static void __mem_cgroup_remove_exceeded(struct mem_cgroup_per_zone *mz,
- struct mem_cgroup_tree_per_zone *mctz)
+static void __mem_cgroup_remove_exceeded(struct mem_cgroup_per_node *mz,
+ struct mem_cgroup_tree_per_node *mctz)
{
if (!mz->on_tree)
return;
@@ -448,8 +439,8 @@ static void __mem_cgroup_remove_exceeded(struct mem_cgroup_per_zone *mz,
mz->on_tree = false;
}
-static void mem_cgroup_remove_exceeded(struct mem_cgroup_per_zone *mz,
- struct mem_cgroup_tree_per_zone *mctz)
+static void mem_cgroup_remove_exceeded(struct mem_cgroup_per_node *mz,
+ struct mem_cgroup_tree_per_node *mctz)
{
unsigned long flags;
@@ -473,8 +464,8 @@ static unsigned long soft_limit_excess(struct mem_cgroup *memcg)
static void mem_cgroup_update_tree(struct mem_cgroup *memcg, struct page *page)
{
unsigned long excess;
- struct mem_cgroup_per_zone *mz;
- struct mem_cgroup_tree_per_zone *mctz;
+ struct mem_cgroup_per_node *mz;
+ struct mem_cgroup_tree_per_node *mctz;
mctz = soft_limit_tree_from_page(page);
/*
@@ -507,24 +498,22 @@ static void mem_cgroup_update_tree(struct mem_cgroup *memcg, struct page *page)
static void mem_cgroup_remove_from_trees(struct mem_cgroup *memcg)
{
- struct mem_cgroup_tree_per_zone *mctz;
- struct mem_cgroup_per_zone *mz;
- int nid, zid;
+ struct mem_cgroup_tree_per_node *mctz;
+ struct mem_cgroup_per_node *mz;
+ int nid;
for_each_node(nid) {
- for (zid = 0; zid < MAX_NR_ZONES; zid++) {
- mz = &memcg->nodeinfo[nid]->zoneinfo[zid];
- mctz = soft_limit_tree_node_zone(nid, zid);
- mem_cgroup_remove_exceeded(mz, mctz);
- }
+ mz = mem_cgroup_nodeinfo(memcg, NODE_DATA(nid));
+ mctz = soft_limit_tree_node(nid);
+ mem_cgroup_remove_exceeded(mz, mctz);
}
}
-static struct mem_cgroup_per_zone *
-__mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz)
+static struct mem_cgroup_per_node *
+__mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_node *mctz)
{
struct rb_node *rightmost = NULL;
- struct mem_cgroup_per_zone *mz;
+ struct mem_cgroup_per_node *mz;
retry:
mz = NULL;
@@ -532,7 +521,7 @@ __mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz)
if (!rightmost)
goto done; /* Nothing to reclaim from */
- mz = rb_entry(rightmost, struct mem_cgroup_per_zone, tree_node);
+ mz = rb_entry(rightmost, struct mem_cgroup_per_node, tree_node);
/*
* Remove the node now but someone else can add it back,
* we will to add it back at the end of reclaim to its correct
@@ -546,10 +535,10 @@ __mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz)
return mz;
}
-static struct mem_cgroup_per_zone *
-mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz)
+static struct mem_cgroup_per_node *
+mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_node *mctz)
{
- struct mem_cgroup_per_zone *mz;
+ struct mem_cgroup_per_node *mz;
spin_lock_irq(&mctz->lock);
mz = __mem_cgroup_largest_soft_limit_node(mctz);
@@ -643,20 +632,16 @@ unsigned long mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
int nid, unsigned int lru_mask)
{
unsigned long nr = 0;
- int zid;
+ struct mem_cgroup_per_node *mz;
+ enum lru_list lru;
VM_BUG_ON((unsigned)nid >= nr_node_ids);
- for (zid = 0; zid < MAX_NR_ZONES; zid++) {
- struct mem_cgroup_per_zone *mz;
- enum lru_list lru;
-
- for_each_lru(lru) {
- if (!(BIT(lru) & lru_mask))
- continue;
- mz = &memcg->nodeinfo[nid]->zoneinfo[zid];
- nr += mz->lru_size[lru];
- }
+ for_each_lru(lru) {
+ if (!(BIT(lru) & lru_mask))
+ continue;
+ mz = mem_cgroup_nodeinfo(memcg, NODE_DATA(nid));
+ nr += mz->lru_size[lru];
}
return nr;
}
@@ -809,9 +794,9 @@ struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
rcu_read_lock();
if (reclaim) {
- struct mem_cgroup_per_zone *mz;
+ struct mem_cgroup_per_node *mz;
- mz = mem_cgroup_zone_zoneinfo(root, reclaim->zone);
+ mz = mem_cgroup_nodeinfo(root, reclaim->pgdat);
iter = &mz->iter[reclaim->priority];
if (prev && reclaim->generation != iter->generation)
@@ -910,19 +895,17 @@ static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg)
{
struct mem_cgroup *memcg = dead_memcg;
struct mem_cgroup_reclaim_iter *iter;
- struct mem_cgroup_per_zone *mz;
- int nid, zid;
+ struct mem_cgroup_per_node *mz;
+ int nid;
int i;
while ((memcg = parent_mem_cgroup(memcg))) {
for_each_node(nid) {
- for (zid = 0; zid < MAX_NR_ZONES; zid++) {
- mz = &memcg->nodeinfo[nid]->zoneinfo[zid];
- for (i = 0; i <= DEF_PRIORITY; i++) {
- iter = &mz->iter[i];
- cmpxchg(&iter->position,
- dead_memcg, NULL);
- }
+ mz = mem_cgroup_nodeinfo(memcg, NODE_DATA(nid));
+ for (i = 0; i <= DEF_PRIORITY; i++) {
+ iter = &mz->iter[i];
+ cmpxchg(&iter->position,
+ dead_memcg, NULL);
}
}
}
@@ -946,7 +929,6 @@ static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg)
/**
* mem_cgroup_lruvec - get the lru list vector for a node or a memcg zone
* @node: node of the wanted lruvec
- * @zone: zone of the wanted lruvec
* @memcg: memcg of the wanted lruvec
*
* Returns the lru list vector holding pages for a given @node or a given
@@ -954,9 +936,9 @@ static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg)
* is disabled.
*/
struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
- struct zone *zone, struct mem_cgroup *memcg)
+ struct mem_cgroup *memcg)
{
- struct mem_cgroup_per_zone *mz;
+ struct mem_cgroup_per_node *mz;
struct lruvec *lruvec;
if (mem_cgroup_disabled()) {
@@ -964,7 +946,7 @@ struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
goto out;
}
- mz = mem_cgroup_zone_zoneinfo(memcg, zone);
+ mz = mem_cgroup_nodeinfo(memcg, pgdat);
lruvec = &mz->lruvec;
out:
/*
@@ -972,8 +954,8 @@ struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
* we have to be prepared to initialize lruvec->zone here;
* and if offlined then reonlined, we need to reinitialize it.
*/
- if (unlikely(lruvec->pgdat != zone->zone_pgdat))
- lruvec->pgdat = zone->zone_pgdat;
+ if (unlikely(lruvec->pgdat != pgdat))
+ lruvec->pgdat = pgdat;
return lruvec;
}
@@ -988,7 +970,7 @@ struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
*/
struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct pglist_data *pgdat)
{
- struct mem_cgroup_per_zone *mz;
+ struct mem_cgroup_per_node *mz;
struct mem_cgroup *memcg;
struct lruvec *lruvec;
@@ -1031,7 +1013,7 @@ struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct pglist_data *pgd
void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
int nr_pages)
{
- struct mem_cgroup_per_zone *mz;
+ struct mem_cgroup_per_node *mz;
unsigned long *lru_size;
long size;
bool empty;
@@ -1041,7 +1023,7 @@ void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
if (mem_cgroup_disabled())
return;
- mz = container_of(lruvec, struct mem_cgroup_per_zone, lruvec);
+ mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
lru_size = mz->lru_size + lru;
empty = list_empty(lruvec->lists + lru);
@@ -1434,7 +1416,7 @@ int mem_cgroup_select_victim_node(struct mem_cgroup *memcg)
#endif
static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg,
- struct zone *zone,
+ pg_data_t *pgdat,
gfp_t gfp_mask,
unsigned long *total_scanned)
{
@@ -1444,7 +1426,7 @@ static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg,
unsigned long excess;
unsigned long nr_scanned;
struct mem_cgroup_reclaim_cookie reclaim = {
- .zone = zone,
+ .pgdat = pgdat,
.priority = 0,
};
@@ -1475,7 +1457,7 @@ static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg,
continue;
}
total += mem_cgroup_shrink_node(victim, gfp_mask, false,
- zone, &nr_scanned);
+ pgdat, &nr_scanned);
*total_scanned += nr_scanned;
if (!soft_limit_excess(root_memcg))
break;
@@ -2602,22 +2584,22 @@ static int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg,
return ret;
}
-unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
+unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
gfp_t gfp_mask,
unsigned long *total_scanned)
{
unsigned long nr_reclaimed = 0;
- struct mem_cgroup_per_zone *mz, *next_mz = NULL;
+ struct mem_cgroup_per_node *mz, *next_mz = NULL;
unsigned long reclaimed;
int loop = 0;
- struct mem_cgroup_tree_per_zone *mctz;
+ struct mem_cgroup_tree_per_node *mctz;
unsigned long excess;
unsigned long nr_scanned;
if (order > 0)
return 0;
- mctz = soft_limit_tree_node_zone(zone_to_nid(zone), zone_idx(zone));
+ mctz = soft_limit_tree_node(pgdat->node_id);
/*
* This loop can run a while, specially if mem_cgroup's continuously
* keep exceeding their soft limit and putting the system under
@@ -2632,7 +2614,7 @@ unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
break;
nr_scanned = 0;
- reclaimed = mem_cgroup_soft_reclaim(mz->memcg, zone,
+ reclaimed = mem_cgroup_soft_reclaim(mz->memcg, pgdat,
gfp_mask, &nr_scanned);
nr_reclaimed += reclaimed;
*total_scanned += nr_scanned;
@@ -3253,22 +3235,21 @@ static int memcg_stat_show(struct seq_file *m, void *v)
#ifdef CONFIG_DEBUG_VM
{
- int nid, zid;
- struct mem_cgroup_per_zone *mz;
+ pg_data_t *pgdat;
+ struct mem_cgroup_per_node *mz;
struct zone_reclaim_stat *rstat;
unsigned long recent_rotated[2] = {0, 0};
unsigned long recent_scanned[2] = {0, 0};
- for_each_online_node(nid)
- for (zid = 0; zid < MAX_NR_ZONES; zid++) {
- mz = &memcg->nodeinfo[nid]->zoneinfo[zid];
- rstat = &mz->lruvec.reclaim_stat;
+ for_each_online_pgdat(pgdat) {
+ mz = mem_cgroup_nodeinfo(memcg, pgdat);
+ rstat = &mz->lruvec.reclaim_stat;
- recent_rotated[0] += rstat->recent_rotated[0];
- recent_rotated[1] += rstat->recent_rotated[1];
- recent_scanned[0] += rstat->recent_scanned[0];
- recent_scanned[1] += rstat->recent_scanned[1];
- }
+ recent_rotated[0] += rstat->recent_rotated[0];
+ recent_rotated[1] += rstat->recent_rotated[1];
+ recent_scanned[0] += rstat->recent_scanned[0];
+ recent_scanned[1] += rstat->recent_scanned[1];
+ }
seq_printf(m, "recent_rotated_anon %lu\n", recent_rotated[0]);
seq_printf(m, "recent_rotated_file %lu\n", recent_rotated[1]);
seq_printf(m, "recent_scanned_anon %lu\n", recent_scanned[0]);
@@ -4094,11 +4075,10 @@ static struct cftype mem_cgroup_legacy_files[] = {
{ }, /* terminate */
};
-static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node)
+static int alloc_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node)
{
struct mem_cgroup_per_node *pn;
- struct mem_cgroup_per_zone *mz;
- int zone, tmp = node;
+ int tmp = node;
/*
* This routine is called against possible nodes.
* But it's BUG to call kmalloc() against offline node.
@@ -4113,18 +4093,16 @@ static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node)
if (!pn)
return 1;
- for (zone = 0; zone < MAX_NR_ZONES; zone++) {
- mz = &pn->zoneinfo[zone];
- lruvec_init(&mz->lruvec);
- mz->usage_in_excess = 0;
- mz->on_tree = false;
- mz->memcg = memcg;
- }
+ lruvec_init(&pn->lruvec);
+ pn->usage_in_excess = 0;
+ pn->on_tree = false;
+ pn->memcg = memcg;
+
memcg->nodeinfo[node] = pn;
return 0;
}
-static void free_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node)
+static void free_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node)
{
kfree(memcg->nodeinfo[node]);
}
@@ -4135,7 +4113,7 @@ static void mem_cgroup_free(struct mem_cgroup *memcg)
memcg_wb_domain_exit(memcg);
for_each_node(node)
- free_mem_cgroup_per_zone_info(memcg, node);
+ free_mem_cgroup_per_node_info(memcg, node);
free_percpu(memcg->stat);
kfree(memcg);
}
@@ -4158,7 +4136,7 @@ static struct mem_cgroup *mem_cgroup_alloc(void)
goto fail;
for_each_node(node)
- if (alloc_mem_cgroup_per_zone_info(memcg, node))
+ if (alloc_mem_cgroup_per_node_info(memcg, node))
goto fail;
if (memcg_wb_domain_init(memcg, GFP_KERNEL))
@@ -5754,18 +5732,12 @@ static int __init mem_cgroup_init(void)
for_each_node(node) {
struct mem_cgroup_tree_per_node *rtpn;
- int zone;
rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL,
node_online(node) ? node : NUMA_NO_NODE);
- for (zone = 0; zone < MAX_NR_ZONES; zone++) {
- struct mem_cgroup_tree_per_zone *rtpz;
-
- rtpz = &rtpn->rb_tree_per_zone[zone];
- rtpz->rb_root = RB_ROOT;
- spin_lock_init(&rtpz->lock);
- }
+ rtpn->rb_root = RB_ROOT;
+ spin_lock_init(&rtpn->lock);
soft_limit_tree.rb_tree_per_node[node] = rtpn;
}
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 93523944abb0..cad6409e798e 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2199,8 +2199,7 @@ static inline void init_tlb_ubc(void)
static void shrink_node_memcg(struct pglist_data *pgdat, struct mem_cgroup *memcg,
struct scan_control *sc, unsigned long *lru_pages)
{
- struct zone *zone = &pgdat->node_zones[sc->reclaim_idx];
- struct lruvec *lruvec = mem_cgroup_lruvec(pgdat, zone, memcg);
+ struct lruvec *lruvec = mem_cgroup_lruvec(pgdat, memcg);
unsigned long nr[NR_LRU_LISTS];
unsigned long targets[NR_LRU_LISTS];
unsigned long nr_to_scan;
@@ -2407,7 +2406,7 @@ static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc,
do {
struct mem_cgroup *root = sc->target_mem_cgroup;
struct mem_cgroup_reclaim_cookie reclaim = {
- .zone = &pgdat->node_zones[classzone_idx],
+ .pgdat = pgdat,
.priority = sc->priority,
};
unsigned long node_lru_pages = 0;
@@ -2609,7 +2608,7 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc,
* and balancing, not for a memcg's limit.
*/
nr_soft_scanned = 0;
- nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
+ nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone->zone_pgdat,
sc->order, sc->gfp_mask,
&nr_soft_scanned);
sc->nr_reclaimed += nr_soft_reclaimed;
@@ -2875,7 +2874,7 @@ unsigned long try_to_free_pages(struct zonelist *zonelist, int order,
unsigned long mem_cgroup_shrink_node(struct mem_cgroup *memcg,
gfp_t gfp_mask, bool noswap,
- struct zone *zone,
+ pg_data_t *pgdat,
unsigned long *nr_scanned)
{
struct scan_control sc = {
@@ -2883,7 +2882,7 @@ unsigned long mem_cgroup_shrink_node(struct mem_cgroup *memcg,
.target_mem_cgroup = memcg,
.may_writepage = !laptop_mode,
.may_unmap = 1,
- .reclaim_idx = zone_idx(zone),
+ .reclaim_idx = MAX_NR_ZONES - 1,
.may_swap = !noswap,
};
unsigned long lru_pages;
@@ -2902,7 +2901,7 @@ unsigned long mem_cgroup_shrink_node(struct mem_cgroup *memcg,
* will pick up pages from other mem cgroup's as well. We hack
* the priority and make it zero.
*/
- shrink_node_memcg(zone->zone_pgdat, memcg, &sc, &lru_pages);
+ shrink_node_memcg(pgdat, memcg, &sc, &lru_pages);
trace_mm_vmscan_memcg_softlimit_reclaim_end(sc.nr_reclaimed);
@@ -2951,7 +2950,7 @@ unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg,
#endif
static void age_active_anon(struct pglist_data *pgdat,
- struct zone *zone, struct scan_control *sc)
+ struct scan_control *sc)
{
struct mem_cgroup *memcg;
@@ -2960,7 +2959,7 @@ static void age_active_anon(struct pglist_data *pgdat,
memcg = mem_cgroup_iter(NULL, NULL, NULL);
do {
- struct lruvec *lruvec = mem_cgroup_lruvec(pgdat, zone, memcg);
+ struct lruvec *lruvec = mem_cgroup_lruvec(pgdat, memcg);
if (inactive_list_is_low(lruvec, false))
shrink_active_list(SWAP_CLUSTER_MAX, lruvec,
@@ -3148,7 +3147,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
* pages are rotated regardless of classzone as this is
* about consistent aging.
*/
- age_active_anon(pgdat, &pgdat->node_zones[MAX_NR_ZONES - 1], &sc);
+ age_active_anon(pgdat, &sc);
/*
* If we're getting trouble reclaiming, start doing writepage
@@ -3160,7 +3159,7 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx)
/* Call soft limit reclaim before calling shrink_node. */
sc.nr_scanned = 0;
nr_soft_scanned = 0;
- nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, sc.order,
+ nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(pgdat, sc.order,
sc.gfp_mask, &nr_soft_scanned);
sc.nr_reclaimed += nr_soft_reclaimed;
diff --git a/mm/workingset.c b/mm/workingset.c
index 10ddf707782a..4e18ffd26967 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -218,7 +218,7 @@ void *workingset_eviction(struct address_space *mapping, struct page *page)
VM_BUG_ON_PAGE(page_count(page), page);
VM_BUG_ON_PAGE(!PageLocked(page), page);
- lruvec = mem_cgroup_lruvec(zone->zone_pgdat, zone, memcg);
+ lruvec = mem_cgroup_lruvec(zone->zone_pgdat, memcg);
eviction = atomic_long_inc_return(&lruvec->inactive_age);
return pack_shadow(memcgid, zone, eviction);
}
@@ -267,7 +267,7 @@ bool workingset_refault(void *shadow)
rcu_read_unlock();
return false;
}
- lruvec = mem_cgroup_lruvec(zone->zone_pgdat, zone, memcg);
+ lruvec = mem_cgroup_lruvec(zone->zone_pgdat, memcg);
refault = atomic_long_read(&lruvec->inactive_age);
active_file = lruvec_lru_size(lruvec, LRU_ACTIVE_FILE);
rcu_read_unlock();
@@ -317,7 +317,7 @@ void workingset_activation(struct page *page)
*/
if (!mem_cgroup_disabled() && !page_memcg(page))
goto out;
- lruvec = mem_cgroup_lruvec(page_zone(page)->zone_pgdat, page_zone(page), page_memcg(page));
+ lruvec = mem_cgroup_lruvec(page_zone(page)->zone_pgdat, page_memcg(page));
atomic_long_inc(&lruvec->inactive_age);
out:
unlock_page_memcg(page);
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 13/27] mm, memcg: Move memcg limit enforcement from zones to nodes
2016-06-09 18:04 ` [PATCH 13/27] mm, memcg: Move memcg limit enforcement from zones to nodes Mel Gorman
@ 2016-06-16 15:06 ` Vlastimil Babka
2016-06-16 15:53 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-16 15:06 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> Memcg was broken by the move of all LRUs to nodes because it is tracking
> limits on a per-zone basis while receiving reclaim requests on a per-node
> basis. This patch moves limit enforcement to the nodes. Technically, all
> the variable names should also change but people are already familiar by
> the meaning of "mz" even if "mn" would be a more appropriate name now.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Didn't spot bugs, but I'm not that familiar with memcg. Noticed some
things below to further optimize/cleanup.
[...]
> @@ -323,13 +319,10 @@ EXPORT_SYMBOL(memcg_kmem_enabled_key);
>
> #endif /* !CONFIG_SLOB */
>
> -static struct mem_cgroup_per_zone *
> -mem_cgroup_zone_zoneinfo(struct mem_cgroup *memcg, struct zone *zone)
> +static struct mem_cgroup_per_node *
> +mem_cgroup_nodeinfo(struct mem_cgroup *memcg, pg_data_t *pgdat)
> {
> - int nid = zone_to_nid(zone);
> - int zid = zone_idx(zone);
> -
> - return &memcg->nodeinfo[nid]->zoneinfo[zid];
> + return memcg->nodeinfo[pgdat->node_id];
I've noticed most callers pass NODE_DATA(nid) as second parameter, which
is quite wasteful to just obtain back the node_id (I doubt the compiler
can know that they will be the same?). So it would be more efficient to
use nid instead of pg_data_t pointer in the signature.
> }
>
> /**
> @@ -383,37 +376,35 @@ ino_t page_cgroup_ino(struct page *page)
> return ino;
> }
>
> -static struct mem_cgroup_per_zone *
> +static struct mem_cgroup_per_node *
> mem_cgroup_page_zoneinfo(struct mem_cgroup *memcg, struct page *page)
This could be renamed to _nodeinfo()?
> {
> int nid = page_to_nid(page);
> - int zid = page_zonenum(page);
>
> - return &memcg->nodeinfo[nid]->zoneinfo[zid];
> + return memcg->nodeinfo[nid];
> }
>
^ permalink raw reply [flat|nested] 70+ messages in thread* Re: [PATCH 13/27] mm, memcg: Move memcg limit enforcement from zones to nodes
2016-06-16 15:06 ` Vlastimil Babka
@ 2016-06-16 15:53 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-16 15:53 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Thu, Jun 16, 2016 at 05:06:46PM +0200, Vlastimil Babka wrote:
> >@@ -323,13 +319,10 @@ EXPORT_SYMBOL(memcg_kmem_enabled_key);
> >
> > #endif /* !CONFIG_SLOB */
> >
> >-static struct mem_cgroup_per_zone *
> >-mem_cgroup_zone_zoneinfo(struct mem_cgroup *memcg, struct zone *zone)
> >+static struct mem_cgroup_per_node *
> >+mem_cgroup_nodeinfo(struct mem_cgroup *memcg, pg_data_t *pgdat)
> > {
> >- int nid = zone_to_nid(zone);
> >- int zid = zone_idx(zone);
> >-
> >- return &memcg->nodeinfo[nid]->zoneinfo[zid];
> >+ return memcg->nodeinfo[pgdat->node_id];
>
> I've noticed most callers pass NODE_DATA(nid) as second parameter, which is
> quite wasteful to just obtain back the node_id (I doubt the compiler can
> know that they will be the same?). So it would be more efficient to use nid
> instead of pg_data_t pointer in the signature.
>
No harm in making the conversion, done now.
> > }
> >
> > /**
> >@@ -383,37 +376,35 @@ ino_t page_cgroup_ino(struct page *page)
> > return ino;
> > }
> >
> >-static struct mem_cgroup_per_zone *
> >+static struct mem_cgroup_per_node *
> > mem_cgroup_page_zoneinfo(struct mem_cgroup *memcg, struct page *page)
>
> This could be renamed to _nodeinfo()?
>
Renamed.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 14/27] mm, workingset: Make working set detection node-aware
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (12 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 13/27] mm, memcg: Move memcg limit enforcement from zones to nodes Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-16 15:13 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 15/27] mm, page_alloc: Consider dirtyable memory in terms of nodes Mel Gorman
` (12 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Working set and refault detection is still zone-based, fix it.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
---
include/linux/mmzone.h | 6 +++---
include/linux/vmstat.h | 1 -
mm/vmstat.c | 20 +++-----------------
mm/workingset.c | 39 ++++++++++++++++++---------------------
4 files changed, 24 insertions(+), 42 deletions(-)
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 16da23c095a0..19a7adf0a734 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -142,9 +142,6 @@ enum zone_stat_item {
NUMA_LOCAL, /* allocation from local node */
NUMA_OTHER, /* allocation from other node */
#endif
- WORKINGSET_REFAULT,
- WORKINGSET_ACTIVATE,
- WORKINGSET_NODERECLAIM,
NR_ANON_TRANSPARENT_HUGEPAGES,
NR_FREE_CMA_PAGES,
NR_VM_ZONE_STAT_ITEMS };
@@ -159,6 +156,9 @@ enum node_stat_item {
NR_ISOLATED_ANON, /* Temporary isolated pages from anon lru */
NR_ISOLATED_FILE, /* Temporary isolated pages from file lru */
NR_PAGES_SCANNED, /* pages scanned since last reclaim */
+ WORKINGSET_REFAULT,
+ WORKINGSET_ACTIVATE,
+ WORKINGSET_NODERECLAIM,
NR_VM_NODE_STAT_ITEMS
};
diff --git a/include/linux/vmstat.h b/include/linux/vmstat.h
index 941695e6b780..533948c93550 100644
--- a/include/linux/vmstat.h
+++ b/include/linux/vmstat.h
@@ -233,7 +233,6 @@ void mod_node_page_state(struct pglist_data *, enum node_stat_item, long);
void inc_node_page_state(struct page *, enum node_stat_item);
void dec_node_page_state(struct page *, enum node_stat_item);
-extern void inc_zone_state(struct zone *, enum zone_stat_item);
extern void inc_node_state(struct pglist_data *, enum node_stat_item);
extern void __inc_zone_state(struct zone *, enum zone_stat_item);
extern void __inc_node_state(struct pglist_data *, enum node_stat_item);
diff --git a/mm/vmstat.c b/mm/vmstat.c
index 1b1ea16be727..fff9614044c7 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -445,11 +445,6 @@ void mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
}
EXPORT_SYMBOL(mod_zone_page_state);
-void inc_zone_state(struct zone *zone, enum zone_stat_item item)
-{
- mod_zone_state(zone, item, 1, 1);
-}
-
void inc_zone_page_state(struct page *page, enum zone_stat_item item)
{
mod_zone_state(page_zone(page), item, 1, 1);
@@ -538,15 +533,6 @@ void mod_zone_page_state(struct zone *zone, enum zone_stat_item item,
}
EXPORT_SYMBOL(mod_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 inc_zone_page_state(struct page *page, enum zone_stat_item item)
{
unsigned long flags;
@@ -964,9 +950,6 @@ const char * const vmstat_text[] = {
"numa_local",
"numa_other",
#endif
- "workingset_refault",
- "workingset_activate",
- "workingset_nodereclaim",
"nr_anon_transparent_hugepages",
"nr_free_cma",
@@ -979,6 +962,9 @@ const char * const vmstat_text[] = {
"nr_isolated_anon",
"nr_isolated_file",
"nr_pages_scanned",
+ "workingset_refault",
+ "workingset_activate",
+ "workingset_nodereclaim",
/* enum writeback_stat_item counters */
"nr_dirty_threshold",
diff --git a/mm/workingset.c b/mm/workingset.c
index 4e18ffd26967..368d8234e3e3 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -16,7 +16,7 @@
/*
* Double CLOCK lists
*
- * Per zone, two clock lists are maintained for file pages: the
+ * Per node, two clock lists are maintained for file pages: the
* inactive and the active list. Freshly faulted pages start out at
* the head of the inactive list and page reclaim scans pages from the
* tail. Pages that are accessed multiple times on the inactive list
@@ -141,11 +141,11 @@
*
* Implementation
*
- * For each zone's file LRU lists, a counter for inactive evictions
- * and activations is maintained (zone->inactive_age).
+ * For each node's file LRU lists, a counter for inactive evictions
+ * and activations is maintained (node->inactive_age).
*
* On eviction, a snapshot of this counter (along with some bits to
- * identify the zone) is stored in the now empty page cache radix tree
+ * identify the node) is stored in the now empty page cache radix tree
* slot of the evicted page. This is called a shadow entry.
*
* On cache misses for which there are shadow entries, an eligible
@@ -153,7 +153,7 @@
*/
#define EVICTION_SHIFT (RADIX_TREE_EXCEPTIONAL_ENTRY + \
- ZONES_SHIFT + NODES_SHIFT + \
+ NODES_SHIFT + \
MEM_CGROUP_ID_SHIFT)
#define EVICTION_MASK (~0UL >> EVICTION_SHIFT)
@@ -167,33 +167,30 @@
*/
static unsigned int bucket_order __read_mostly;
-static void *pack_shadow(int memcgid, struct zone *zone, unsigned long eviction)
+static void *pack_shadow(int memcgid, pg_data_t *pgdat, unsigned long eviction)
{
eviction >>= bucket_order;
eviction = (eviction << MEM_CGROUP_ID_SHIFT) | memcgid;
- eviction = (eviction << NODES_SHIFT) | zone_to_nid(zone);
- eviction = (eviction << ZONES_SHIFT) | zone_idx(zone);
+ eviction = (eviction << NODES_SHIFT) | pgdat->node_id;
eviction = (eviction << RADIX_TREE_EXCEPTIONAL_SHIFT);
return (void *)(eviction | RADIX_TREE_EXCEPTIONAL_ENTRY);
}
-static void unpack_shadow(void *shadow, int *memcgidp, struct zone **zonep,
+static void unpack_shadow(void *shadow, int *memcgidp, pg_data_t **pgdat,
unsigned long *evictionp)
{
unsigned long entry = (unsigned long)shadow;
- int memcgid, nid, zid;
+ int memcgid, nid;
entry >>= RADIX_TREE_EXCEPTIONAL_SHIFT;
- zid = entry & ((1UL << ZONES_SHIFT) - 1);
- entry >>= ZONES_SHIFT;
nid = entry & ((1UL << NODES_SHIFT) - 1);
entry >>= NODES_SHIFT;
memcgid = entry & ((1UL << MEM_CGROUP_ID_SHIFT) - 1);
entry >>= MEM_CGROUP_ID_SHIFT;
*memcgidp = memcgid;
- *zonep = NODE_DATA(nid)->node_zones + zid;
+ *pgdat = NODE_DATA(nid);
*evictionp = entry << bucket_order;
}
@@ -220,7 +217,7 @@ void *workingset_eviction(struct address_space *mapping, struct page *page)
lruvec = mem_cgroup_lruvec(zone->zone_pgdat, memcg);
eviction = atomic_long_inc_return(&lruvec->inactive_age);
- return pack_shadow(memcgid, zone, eviction);
+ return pack_shadow(memcgid, zone->zone_pgdat, eviction);
}
/**
@@ -228,7 +225,7 @@ void *workingset_eviction(struct address_space *mapping, struct page *page)
* @shadow: shadow entry of the evicted page
*
* Calculates and evaluates the refault distance of the previously
- * evicted page in the context of the zone it was allocated in.
+ * evicted page in the context of the node it was allocated in.
*
* Returns %true if the page should be activated, %false otherwise.
*/
@@ -240,10 +237,10 @@ bool workingset_refault(void *shadow)
unsigned long eviction;
struct lruvec *lruvec;
unsigned long refault;
- struct zone *zone;
+ struct pglist_data *pgdat;
int memcgid;
- unpack_shadow(shadow, &memcgid, &zone, &eviction);
+ unpack_shadow(shadow, &memcgid, &pgdat, &eviction);
rcu_read_lock();
/*
@@ -267,7 +264,7 @@ bool workingset_refault(void *shadow)
rcu_read_unlock();
return false;
}
- lruvec = mem_cgroup_lruvec(zone->zone_pgdat, memcg);
+ lruvec = mem_cgroup_lruvec(pgdat, memcg);
refault = atomic_long_read(&lruvec->inactive_age);
active_file = lruvec_lru_size(lruvec, LRU_ACTIVE_FILE);
rcu_read_unlock();
@@ -290,10 +287,10 @@ bool workingset_refault(void *shadow)
*/
refault_distance = (refault - eviction) & EVICTION_MASK;
- inc_zone_state(zone, WORKINGSET_REFAULT);
+ inc_node_state(pgdat, WORKINGSET_REFAULT);
if (refault_distance <= active_file) {
- inc_zone_state(zone, WORKINGSET_ACTIVATE);
+ inc_node_state(pgdat, WORKINGSET_ACTIVATE);
return true;
}
return false;
@@ -434,7 +431,7 @@ static enum lru_status shadow_lru_isolate(struct list_head *item,
}
}
BUG_ON(node->count);
- inc_zone_state(page_zone(virt_to_page(node)), WORKINGSET_NODERECLAIM);
+ inc_node_state(page_zone(virt_to_page(node))->zone_pgdat, WORKINGSET_NODERECLAIM);
if (!__radix_tree_delete_node(&mapping->page_tree, node))
BUG();
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 14/27] mm, workingset: Make working set detection node-aware
2016-06-09 18:04 ` [PATCH 14/27] mm, workingset: Make working set detection node-aware Mel Gorman
@ 2016-06-16 15:13 ` Vlastimil Babka
2016-06-16 15:56 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-16 15:13 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> Working set and refault detection is still zone-based, fix it.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
If you wanted, workingset_eviction() could obtain pgdat without going
through zone.
^ permalink raw reply [flat|nested] 70+ messages in thread
* Re: [PATCH 14/27] mm, workingset: Make working set detection node-aware
2016-06-16 15:13 ` Vlastimil Babka
@ 2016-06-16 15:56 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-16 15:56 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Thu, Jun 16, 2016 at 05:13:51PM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> >Working set and refault detection is still zone-based, fix it.
> >
> >Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> >Acked-by: Johannes Weiner <hannes@cmpxchg.org>
>
> Acked-by: Vlastimil Babka <vbabka@suse.cz>
>
> If you wanted, workingset_eviction() could obtain pgdat without going
> through zone.
Yeah, saves a few lookups. Thanks.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 15/27] mm, page_alloc: Consider dirtyable memory in terms of nodes
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (13 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 14/27] mm, workingset: Make working set detection node-aware Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-16 15:45 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 16/27] mm: Move page mapped accounting to the node Mel Gorman
` (11 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Historically dirty pages were spread among zones but now that LRUs are
per-node it is more appropriate to consider dirty pages in a node.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
---
include/linux/mmzone.h | 12 +++---
include/linux/writeback.h | 2 +-
mm/page-writeback.c | 94 ++++++++++++++++++++++++-----------------------
mm/page_alloc.c | 26 ++++++-------
4 files changed, 67 insertions(+), 67 deletions(-)
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 19a7adf0a734..5de2fc2dd0e6 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -358,12 +358,6 @@ struct zone {
struct pglist_data *zone_pgdat;
struct per_cpu_pageset __percpu *pageset;
- /*
- * This is a per-zone reserve of pages that are not available
- * to userspace allocations.
- */
- unsigned long totalreserve_pages;
-
#ifndef CONFIG_SPARSEMEM
/*
* Flags for a pageblock_nr_pages block. See pageblock-flags.h.
@@ -682,6 +676,12 @@ typedef struct pglist_data {
/* Number of pages migrated during the rate limiting time interval */
unsigned long numabalancing_migrate_nr_pages;
#endif
+ /*
+ * This is a per-node reserve of pages that are not available
+ * to userspace allocations.
+ */
+ unsigned long totalreserve_pages;
+
/* Write-intensive fields used from the page allocator */
ZONE_PADDING(_pad1_)
spinlock_t lru_lock;
diff --git a/include/linux/writeback.h b/include/linux/writeback.h
index d0b5ca5d4e08..44b4422ae57f 100644
--- a/include/linux/writeback.h
+++ b/include/linux/writeback.h
@@ -320,7 +320,7 @@ void laptop_mode_timer_fn(unsigned long data);
static inline void laptop_sync_completion(void) { }
#endif
void throttle_vm_writeout(gfp_t gfp_mask);
-bool zone_dirty_ok(struct zone *zone);
+bool node_dirty_ok(struct pglist_data *pgdat);
int wb_domain_init(struct wb_domain *dom, gfp_t gfp);
#ifdef CONFIG_CGROUP_WRITEBACK
void wb_domain_exit(struct wb_domain *dom);
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index 3ddfa0b48436..4db3be44bb1d 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -267,26 +267,35 @@ static void wb_min_max_ratio(struct bdi_writeback *wb,
*/
/**
- * zone_dirtyable_memory - number of dirtyable pages in a zone
- * @zone: the zone
+ * node_dirtyable_memory - number of dirtyable pages in a node
+ * @pgdat: the node
*
- * Returns the zone's number of pages potentially available for dirty
- * page cache. This is the base value for the per-zone dirty limits.
+ * Returns the node's number of pages potentially available for dirty
+ * page cache. This is the base value for the per-node dirty limits.
*/
-static unsigned long zone_dirtyable_memory(struct zone *zone)
+static unsigned long node_dirtyable_memory(struct pglist_data *pgdat)
{
- unsigned long nr_pages;
+ unsigned long nr_pages = 0;
+ int z;
+
+ for (z = 0; z < MAX_NR_ZONES; z++) {
+ struct zone *zone = pgdat->node_zones + z;
+
+ if (!populated_zone(zone))
+ continue;
+
+ nr_pages += zone_page_state(zone, NR_FREE_PAGES);
+ }
- nr_pages = zone_page_state(zone, NR_FREE_PAGES);
/*
* Pages reserved for the kernel should not be considered
* dirtyable, to prevent a situation where reclaim has to
* clean pages in order to balance the zones.
*/
- nr_pages -= min(nr_pages, zone->totalreserve_pages);
+ nr_pages -= min(nr_pages, pgdat->totalreserve_pages);
- nr_pages += node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE);
- nr_pages += node_page_state(zone->zone_pgdat, NR_ACTIVE_FILE);
+ nr_pages += node_page_state(pgdat, NR_INACTIVE_FILE);
+ nr_pages += node_page_state(pgdat, NR_ACTIVE_FILE);
return nr_pages;
}
@@ -294,29 +303,13 @@ static unsigned long zone_dirtyable_memory(struct zone *zone)
static unsigned long highmem_dirtyable_memory(unsigned long total)
{
#ifdef CONFIG_HIGHMEM
- int node;
unsigned long x = 0;
- int i;
-
- for_each_node_state(node, N_HIGH_MEMORY) {
- for (i = 0; i < MAX_NR_ZONES; i++) {
- struct zone *z = &NODE_DATA(node)->node_zones[i];
- if (is_highmem(z))
- x += zone_dirtyable_memory(z);
- }
- }
/*
- * Unreclaimable memory (kernel memory or anonymous memory
- * without swap) can bring down the dirtyable pages below
- * the zone's dirty balance reserve and the above calculation
- * will underflow. However we still want to add in nodes
- * which are below threshold (negative values) to get a more
- * accurate calculation but make sure that the total never
- * underflows.
+ * LRU lists are per-node so there is no fast and accurate means of
+ * calculating dirtyable memory in the highmem zone.
*/
- if ((long)x < 0)
- x = 0;
+ x = totalhigh_pages;
/*
* Make sure that the number of highmem pages is never larger
@@ -442,23 +435,23 @@ void global_dirty_limits(unsigned long *pbackground, unsigned long *pdirty)
}
/**
- * zone_dirty_limit - maximum number of dirty pages allowed in a zone
- * @zone: the zone
+ * node_dirty_limit - maximum number of dirty pages allowed in a node
+ * @pgdat: the node
*
- * Returns the maximum number of dirty pages allowed in a zone, based
- * on the zone's dirtyable memory.
+ * Returns the maximum number of dirty pages allowed in a node, based
+ * on the node's dirtyable memory.
*/
-static unsigned long zone_dirty_limit(struct zone *zone)
+static unsigned long node_dirty_limit(struct pglist_data *pgdat)
{
- unsigned long zone_memory = zone_dirtyable_memory(zone);
+ unsigned long node_memory = node_dirtyable_memory(pgdat);
struct task_struct *tsk = current;
unsigned long dirty;
if (vm_dirty_bytes)
dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE) *
- zone_memory / global_dirtyable_memory();
+ node_memory / global_dirtyable_memory();
else
- dirty = vm_dirty_ratio * zone_memory / 100;
+ dirty = vm_dirty_ratio * node_memory / 100;
if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk))
dirty += dirty / 4;
@@ -467,19 +460,30 @@ static unsigned long zone_dirty_limit(struct zone *zone)
}
/**
- * zone_dirty_ok - tells whether a zone is within its dirty limits
- * @zone: the zone to check
+ * node_dirty_ok - tells whether a node is within its dirty limits
+ * @pgdat: the node to check
*
- * Returns %true when the dirty pages in @zone are within the zone's
+ * Returns %true when the dirty pages in @pgdat are within the node's
* dirty limit, %false if the limit is exceeded.
*/
-bool zone_dirty_ok(struct zone *zone)
+bool node_dirty_ok(struct pglist_data *pgdat)
{
- unsigned long limit = zone_dirty_limit(zone);
+ int z;
+ unsigned long limit = node_dirty_limit(pgdat);
+ unsigned long nr_pages = 0;
+
+ for (z = 0; z < MAX_NR_ZONES; z++) {
+ struct zone *zone = pgdat->node_zones + z;
+
+ if (!populated_zone(zone))
+ continue;
+
+ nr_pages += zone_page_state(zone, NR_FILE_DIRTY);
+ nr_pages += zone_page_state(zone, NR_UNSTABLE_NFS);
+ nr_pages += zone_page_state(zone, NR_WRITEBACK);
+ }
- return zone_page_state(zone, NR_FILE_DIRTY) +
- zone_page_state(zone, NR_UNSTABLE_NFS) +
- zone_page_state(zone, NR_WRITEBACK) <= limit;
+ return nr_pages <= limit;
}
int dirty_background_ratio_handler(struct ctl_table *table, int write,
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index fc5bad895732..37e7b909ba48 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2917,31 +2917,24 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
}
/*
* When allocating a page cache page for writing, we
- * want to get it from a zone that is within its dirty
- * limit, such that no single zone holds more than its
+ * want to get it from a node that is within its dirty
+ * limit, such that no single node holds more than its
* proportional share of globally allowed dirty pages.
- * The dirty limits take into account the zone's
+ * The dirty limits take into account the node's
* lowmem reserves and high watermark so that kswapd
* should be able to balance it without having to
* write pages from its LRU list.
*
- * This may look like it could increase pressure on
- * lower zones by failing allocations in higher zones
- * before they are full. But the pages that do spill
- * over are limited as the lower zones are protected
- * by this very same mechanism. It should not become
- * a practical burden to them.
- *
* XXX: For now, allow allocations to potentially
- * exceed the per-zone dirty limit in the slowpath
+ * exceed the per-node dirty limit in the slowpath
* (spread_dirty_pages unset) before going into reclaim,
* which is important when on a NUMA setup the allowed
- * zones are together not big enough to reach the
+ * nodes are together not big enough to reach the
* global limit. The proper fix for these situations
- * will require awareness of zones in the
+ * will require awareness of nodes in the
* dirty-throttling and the flusher threads.
*/
- if (ac->spread_dirty_pages && !zone_dirty_ok(zone))
+ if (ac->spread_dirty_pages && !node_dirty_ok(zone->zone_pgdat))
continue;
mark = zone->watermark[alloc_flags & ALLOC_WMARK_MASK];
@@ -6701,6 +6694,9 @@ static void calculate_totalreserve_pages(void)
enum zone_type i, j;
for_each_online_pgdat(pgdat) {
+
+ pgdat->totalreserve_pages = 0;
+
for (i = 0; i < MAX_NR_ZONES; i++) {
struct zone *zone = pgdat->node_zones + i;
long max = 0;
@@ -6717,7 +6713,7 @@ static void calculate_totalreserve_pages(void)
if (max > zone->managed_pages)
max = zone->managed_pages;
- zone->totalreserve_pages = max;
+ pgdat->totalreserve_pages += max;
reserve_pages += max;
}
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* [PATCH 16/27] mm: Move page mapped accounting to the node
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (14 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 15/27] mm, page_alloc: Consider dirtyable memory in terms of nodes Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-16 15:52 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 17/27] mm: Rename NR_ANON_PAGES to NR_ANON_MAPPED Mel Gorman
` (10 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
Reclaim makes decisions based on the number of file pages that are mapped but
it's mixing node and zone information. Account NR_FILE_MAPPED pages on the node.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
arch/tile/mm/pgtable.c | 2 +-
drivers/base/node.c | 4 ++--
fs/proc/meminfo.c | 4 ++--
include/linux/mmzone.h | 6 +++---
mm/page_alloc.c | 6 +++---
mm/rmap.c | 13 +++++++------
mm/vmscan.c | 2 +-
mm/vmstat.c | 4 ++--
8 files changed, 21 insertions(+), 20 deletions(-)
diff --git a/arch/tile/mm/pgtable.c b/arch/tile/mm/pgtable.c
index 9e389213580d..c606b0ef2f7e 100644
--- a/arch/tile/mm/pgtable.c
+++ b/arch/tile/mm/pgtable.c
@@ -55,7 +55,7 @@ void show_mem(unsigned int filter)
global_page_state(NR_FREE_PAGES),
(global_page_state(NR_SLAB_RECLAIMABLE) +
global_page_state(NR_SLAB_UNRECLAIMABLE)),
- global_page_state(NR_FILE_MAPPED),
+ global_node_page_state(NR_FILE_MAPPED),
global_page_state(NR_PAGETABLE),
global_page_state(NR_BOUNCE),
global_page_state(NR_FILE_PAGES),
diff --git a/drivers/base/node.c b/drivers/base/node.c
index 4260c7f3ee1b..66aed68a0fdc 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -119,8 +119,8 @@ static ssize_t node_read_meminfo(struct device *dev,
nid, K(sum_zone_node_page_state(nid, NR_FILE_DIRTY)),
nid, K(sum_zone_node_page_state(nid, NR_WRITEBACK)),
nid, K(sum_zone_node_page_state(nid, NR_FILE_PAGES)),
- nid, K(sum_zone_node_page_state(nid, NR_FILE_MAPPED)),
- nid, K(sum_zone_node_page_state(nid, NR_ANON_PAGES)),
+ nid, K(node_page_state(pgdat, NR_FILE_MAPPED)),
+ nid, K(node_page_state(pgdat, NR_ANON_PAGES)),
nid, K(i.sharedram),
nid, sum_zone_node_page_state(nid, NR_KERNEL_STACK) *
THREAD_SIZE / 1024,
diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index 83720460c5bc..54e039682ec9 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -138,8 +138,8 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
K(i.freeswap),
K(global_page_state(NR_FILE_DIRTY)),
K(global_page_state(NR_WRITEBACK)),
- K(global_page_state(NR_ANON_PAGES)),
- K(global_page_state(NR_FILE_MAPPED)),
+ K(global_node_page_state(NR_ANON_PAGES)),
+ K(global_node_page_state(NR_FILE_MAPPED)),
K(i.sharedram),
K(global_page_state(NR_SLAB_RECLAIMABLE) +
global_page_state(NR_SLAB_UNRECLAIMABLE)),
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 5de2fc2dd0e6..eac00a55e098 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -112,9 +112,6 @@ enum zone_stat_item {
NR_FREE_PAGES,
NR_ALLOC_BATCH,
NR_MLOCK, /* mlock()ed pages found and moved off LRU */
- NR_ANON_PAGES, /* Mapped anonymous pages */
- NR_FILE_MAPPED, /* pagecache pages mapped into pagetables.
- only modified from process context */
NR_FILE_PAGES,
NR_FILE_DIRTY,
NR_WRITEBACK,
@@ -159,6 +156,9 @@ enum node_stat_item {
WORKINGSET_REFAULT,
WORKINGSET_ACTIVATE,
WORKINGSET_NODERECLAIM,
+ NR_ANON_PAGES, /* Mapped anonymous pages */
+ NR_FILE_MAPPED, /* pagecache pages mapped into pagetables.
+ only modified from process context */
NR_VM_NODE_STAT_ITEMS
};
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 37e7b909ba48..1ba0c98d9461 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -4323,7 +4323,7 @@ void show_free_areas(unsigned int filter)
global_page_state(NR_UNSTABLE_NFS),
global_page_state(NR_SLAB_RECLAIMABLE),
global_page_state(NR_SLAB_UNRECLAIMABLE),
- global_page_state(NR_FILE_MAPPED),
+ global_node_page_state(NR_FILE_MAPPED),
global_page_state(NR_SHMEM),
global_page_state(NR_PAGETABLE),
global_page_state(NR_BOUNCE),
@@ -4340,6 +4340,7 @@ void show_free_areas(unsigned int filter)
" unevictable:%lukB"
" isolated(anon):%lukB"
" isolated(file):%lukB"
+ " mapped:%lukB"
" all_unreclaimable? %s"
"\n",
pgdat->node_id,
@@ -4350,6 +4351,7 @@ void show_free_areas(unsigned int filter)
K(node_page_state(pgdat, NR_UNEVICTABLE)),
K(node_page_state(pgdat, NR_ISOLATED_ANON)),
K(node_page_state(pgdat, NR_ISOLATED_FILE)),
+ K(node_page_state(pgdat, NR_FILE_MAPPED)),
!pgdat_reclaimable(pgdat) ? "yes" : "no");
}
@@ -4374,7 +4376,6 @@ void show_free_areas(unsigned int filter)
" mlocked:%lukB"
" dirty:%lukB"
" writeback:%lukB"
- " mapped:%lukB"
" shmem:%lukB"
" slab_reclaimable:%lukB"
" slab_unreclaimable:%lukB"
@@ -4398,7 +4399,6 @@ void show_free_areas(unsigned int filter)
K(zone_page_state(zone, NR_MLOCK)),
K(zone_page_state(zone, NR_FILE_DIRTY)),
K(zone_page_state(zone, NR_WRITEBACK)),
- K(zone_page_state(zone, NR_FILE_MAPPED)),
K(zone_page_state(zone, NR_SHMEM)),
K(zone_page_state(zone, NR_SLAB_RECLAIMABLE)),
K(zone_page_state(zone, NR_SLAB_UNRECLAIMABLE)),
diff --git a/mm/rmap.c b/mm/rmap.c
index 67457f691528..ab3d7a3151c3 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -1219,7 +1219,8 @@ void do_page_add_anon_rmap(struct page *page,
__inc_zone_page_state(page,
NR_ANON_TRANSPARENT_HUGEPAGES);
}
- __mod_zone_page_state(page_zone(page), NR_ANON_PAGES, nr);
+ __mod_node_page_state(page_zone(page)->zone_pgdat,
+ NR_ANON_PAGES, nr);
}
if (unlikely(PageKsm(page)))
return;
@@ -1263,7 +1264,7 @@ void page_add_new_anon_rmap(struct page *page,
/* increment count (starts at -1) */
atomic_set(&page->_mapcount, 0);
}
- __mod_zone_page_state(page_zone(page), NR_ANON_PAGES, nr);
+ __mod_node_page_state(page_zone(page)->zone_pgdat, NR_ANON_PAGES, nr);
__page_set_anon_rmap(page, vma, address, 1);
}
@@ -1277,7 +1278,7 @@ void page_add_file_rmap(struct page *page)
{
lock_page_memcg(page);
if (atomic_inc_and_test(&page->_mapcount)) {
- __inc_zone_page_state(page, NR_FILE_MAPPED);
+ __inc_node_page_state(page, NR_FILE_MAPPED);
mem_cgroup_inc_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED);
}
unlock_page_memcg(page);
@@ -1303,7 +1304,7 @@ static void page_remove_file_rmap(struct page *page)
* these counters are not modified in interrupt context, and
* pte lock(a spinlock) is held, which implies preemption disabled.
*/
- __dec_zone_page_state(page, NR_FILE_MAPPED);
+ __dec_node_page_state(page, NR_FILE_MAPPED);
mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_FILE_MAPPED);
if (unlikely(PageMlocked(page)))
@@ -1345,7 +1346,7 @@ static void page_remove_anon_compound_rmap(struct page *page)
clear_page_mlock(page);
if (nr) {
- __mod_zone_page_state(page_zone(page), NR_ANON_PAGES, -nr);
+ __mod_node_page_state(page_zone(page)->zone_pgdat, NR_ANON_PAGES, -nr);
deferred_split_huge_page(page);
}
}
@@ -1377,7 +1378,7 @@ void page_remove_rmap(struct page *page, bool compound)
* these counters are not modified in interrupt context, and
* pte lock(a spinlock) is held, which implies preemption disabled.
*/
- __dec_zone_page_state(page, NR_ANON_PAGES);
+ __dec_node_page_state(page, NR_ANON_PAGES);
if (unlikely(PageMlocked(page)))
clear_page_mlock(page);
diff --git a/mm/vmscan.c b/mm/vmscan.c
index cad6409e798e..56482c3fbc76 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3556,7 +3556,7 @@ int sysctl_min_slab_ratio = 5;
static inline unsigned long zone_unmapped_file_pages(struct zone *zone)
{
- unsigned long file_mapped = zone_page_state(zone, NR_FILE_MAPPED);
+ unsigned long file_mapped = node_page_state(zone->zone_pgdat, NR_FILE_MAPPED);
unsigned long file_lru = node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE) +
node_page_state(zone->zone_pgdat, NR_ACTIVE_FILE);
diff --git a/mm/vmstat.c b/mm/vmstat.c
index fff9614044c7..960f549a7b1a 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -922,8 +922,6 @@ const char * const vmstat_text[] = {
"nr_free_pages",
"nr_alloc_batch",
"nr_mlock",
- "nr_anon_pages",
- "nr_mapped",
"nr_file_pages",
"nr_dirty",
"nr_writeback",
@@ -965,6 +963,8 @@ const char * const vmstat_text[] = {
"workingset_refault",
"workingset_activate",
"workingset_nodereclaim",
+ "nr_anon_pages",
+ "nr_mapped",
/* enum writeback_stat_item counters */
"nr_dirty_threshold",
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 16/27] mm: Move page mapped accounting to the node
2016-06-09 18:04 ` [PATCH 16/27] mm: Move page mapped accounting to the node Mel Gorman
@ 2016-06-16 15:52 ` Vlastimil Babka
2016-06-16 16:04 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-16 15:52 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> Reclaim makes decisions based on the number of file pages that are mapped but
> it's mixing node and zone information. Account NR_FILE_MAPPED pages on the node.
And NR_ANON_PAGES.
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
I've noticed some new "page_zone(page)->zone_pgdat" instances here.
^ permalink raw reply [flat|nested] 70+ messages in thread
* Re: [PATCH 16/27] mm: Move page mapped accounting to the node
2016-06-16 15:52 ` Vlastimil Babka
@ 2016-06-16 16:04 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-16 16:04 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Thu, Jun 16, 2016 at 05:52:59PM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> >Reclaim makes decisions based on the number of file pages that are mapped but
> >it's mixing node and zone information. Account NR_FILE_MAPPED pages on the node.
>
> And NR_ANON_PAGES.
>
Yes.
> >Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
>
> Acked-by: Vlastimil Babka <vbabka@suse.cz>
>
> I've noticed some new "page_zone(page)->zone_pgdat" instances here.
>
FWIW, I created a page_pgdat helper and converted all patterns like this
introduced by the series to the helper. There are other potential users
of the helper in the tree that are not touched by the patches but I left
them to preserve history.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 17/27] mm: Rename NR_ANON_PAGES to NR_ANON_MAPPED
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (15 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 16/27] mm: Move page mapped accounting to the node Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 8:28 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 18/27] mm: Move most file-based accounting to the node Mel Gorman
` (9 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
NR_FILE_PAGES is the number of file pages.
NR_FILE_MAPPED is the number of mapped file pages.
NR_ANON_PAGES is the number of mapped anon pages.
This is unhelpful naming as it's easy to confuse NR_FILE_MAPPED and NR_ANON_PAGES for
mapped pages. This patch renames NR_ANON_PAGES so we have
NR_FILE_PAGES is the number of file pages.
NR_FILE_MAPPED is the number of mapped file pages.
NR_ANON_MAPPED is the number of mapped anon pages.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
drivers/base/node.c | 2 +-
fs/proc/meminfo.c | 2 +-
include/linux/mmzone.h | 2 +-
mm/migrate.c | 2 +-
mm/rmap.c | 8 ++++----
5 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/drivers/base/node.c b/drivers/base/node.c
index 66aed68a0fdc..897b6bcb36be 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -120,7 +120,7 @@ static ssize_t node_read_meminfo(struct device *dev,
nid, K(sum_zone_node_page_state(nid, NR_WRITEBACK)),
nid, K(sum_zone_node_page_state(nid, NR_FILE_PAGES)),
nid, K(node_page_state(pgdat, NR_FILE_MAPPED)),
- nid, K(node_page_state(pgdat, NR_ANON_PAGES)),
+ nid, K(node_page_state(pgdat, NR_ANON_MAPPED)),
nid, K(i.sharedram),
nid, sum_zone_node_page_state(nid, NR_KERNEL_STACK) *
THREAD_SIZE / 1024,
diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index 54e039682ec9..076afb43fc56 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -138,7 +138,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
K(i.freeswap),
K(global_page_state(NR_FILE_DIRTY)),
K(global_page_state(NR_WRITEBACK)),
- K(global_node_page_state(NR_ANON_PAGES)),
+ K(global_node_page_state(NR_ANON_MAPPED)),
K(global_node_page_state(NR_FILE_MAPPED)),
K(i.sharedram),
K(global_page_state(NR_SLAB_RECLAIMABLE) +
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index eac00a55e098..8fbb2a1f7a0f 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -156,7 +156,7 @@ enum node_stat_item {
WORKINGSET_REFAULT,
WORKINGSET_ACTIVATE,
WORKINGSET_NODERECLAIM,
- NR_ANON_PAGES, /* Mapped anonymous pages */
+ NR_ANON_MAPPED, /* Mapped anonymous pages */
NR_FILE_MAPPED, /* pagecache pages mapped into pagetables.
only modified from process context */
NR_VM_NODE_STAT_ITEMS
diff --git a/mm/migrate.c b/mm/migrate.c
index cff5d6aa2240..133772d8f08f 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -501,7 +501,7 @@ int migrate_page_move_mapping(struct address_space *mapping,
* new page and drop references to the old page.
*
* Note that anonymous pages are accounted for
- * via NR_FILE_PAGES and NR_ANON_PAGES if they
+ * via NR_FILE_PAGES and NR_ANON_MAPPED if they
* are mapped to swap space.
*/
if (newzone != oldzone) {
diff --git a/mm/rmap.c b/mm/rmap.c
index ab3d7a3151c3..45443d488e92 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -1220,7 +1220,7 @@ void do_page_add_anon_rmap(struct page *page,
NR_ANON_TRANSPARENT_HUGEPAGES);
}
__mod_node_page_state(page_zone(page)->zone_pgdat,
- NR_ANON_PAGES, nr);
+ NR_ANON_MAPPED, nr);
}
if (unlikely(PageKsm(page)))
return;
@@ -1264,7 +1264,7 @@ void page_add_new_anon_rmap(struct page *page,
/* increment count (starts at -1) */
atomic_set(&page->_mapcount, 0);
}
- __mod_node_page_state(page_zone(page)->zone_pgdat, NR_ANON_PAGES, nr);
+ __mod_node_page_state(page_zone(page)->zone_pgdat, NR_ANON_MAPPED, nr);
__page_set_anon_rmap(page, vma, address, 1);
}
@@ -1346,7 +1346,7 @@ static void page_remove_anon_compound_rmap(struct page *page)
clear_page_mlock(page);
if (nr) {
- __mod_node_page_state(page_zone(page)->zone_pgdat, NR_ANON_PAGES, -nr);
+ __mod_node_page_state(page_zone(page)->zone_pgdat, NR_ANON_MAPPED, -nr);
deferred_split_huge_page(page);
}
}
@@ -1378,7 +1378,7 @@ void page_remove_rmap(struct page *page, bool compound)
* these counters are not modified in interrupt context, and
* pte lock(a spinlock) is held, which implies preemption disabled.
*/
- __dec_node_page_state(page, NR_ANON_PAGES);
+ __dec_node_page_state(page, NR_ANON_MAPPED);
if (unlikely(PageMlocked(page)))
clear_page_mlock(page);
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 17/27] mm: Rename NR_ANON_PAGES to NR_ANON_MAPPED
2016-06-09 18:04 ` [PATCH 17/27] mm: Rename NR_ANON_PAGES to NR_ANON_MAPPED Mel Gorman
@ 2016-06-17 8:28 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 8:28 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> NR_FILE_PAGES is the number of file pages.
> NR_FILE_MAPPED is the number of mapped file pages.
> NR_ANON_PAGES is the number of mapped anon pages.
>
> This is unhelpful naming as it's easy to confuse NR_FILE_MAPPED and NR_ANON_PAGES for
> mapped pages. This patch renames NR_ANON_PAGES so we have
>
> NR_FILE_PAGES is the number of file pages.
> NR_FILE_MAPPED is the number of mapped file pages.
> NR_ANON_MAPPED is the number of mapped anon pages.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 18/27] mm: Move most file-based accounting to the node
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (16 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 17/27] mm: Rename NR_ANON_PAGES to NR_ANON_MAPPED Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 8:35 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 19/27] mm: Move vmscan writes and file write " Mel Gorman
` (8 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
There are now a number of accounting oddities such as mapped file pages
being accounted for on the node while the total number of file pages are
accounted on the zone. This can be coped with to some extent but it's
confusing so this patch moves the relevant file-based accounted.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
arch/s390/appldata/appldata_mem.c | 2 +-
arch/tile/mm/pgtable.c | 8 +++---
drivers/base/node.c | 10 ++++----
drivers/staging/android/lowmemorykiller.c | 4 +--
fs/fs-writeback.c | 4 +--
fs/fuse/file.c | 8 +++---
fs/nfs/internal.h | 2 +-
fs/nfs/write.c | 2 +-
fs/proc/meminfo.c | 10 ++++----
include/linux/mmzone.h | 12 ++++-----
include/trace/events/writeback.h | 6 ++---
mm/filemap.c | 10 ++++----
mm/migrate.c | 12 ++++-----
mm/page-writeback.c | 42 +++++++++++++------------------
mm/page_alloc.c | 34 ++++++++++++-------------
mm/shmem.c | 12 ++++-----
mm/swap_state.c | 4 +--
mm/util.c | 4 +--
mm/vmscan.c | 16 ++++++------
mm/vmstat.c | 12 ++++-----
20 files changed, 103 insertions(+), 111 deletions(-)
diff --git a/arch/s390/appldata/appldata_mem.c b/arch/s390/appldata/appldata_mem.c
index edcf2a706942..598df5708501 100644
--- a/arch/s390/appldata/appldata_mem.c
+++ b/arch/s390/appldata/appldata_mem.c
@@ -102,7 +102,7 @@ static void appldata_get_mem_data(void *data)
mem_data->totalhigh = P2K(val.totalhigh);
mem_data->freehigh = P2K(val.freehigh);
mem_data->bufferram = P2K(val.bufferram);
- mem_data->cached = P2K(global_page_state(NR_FILE_PAGES)
+ mem_data->cached = P2K(global_node_page_state(NR_FILE_PAGES)
- val.bufferram);
si_swapinfo(&val);
diff --git a/arch/tile/mm/pgtable.c b/arch/tile/mm/pgtable.c
index c606b0ef2f7e..7cc6ee7f1a58 100644
--- a/arch/tile/mm/pgtable.c
+++ b/arch/tile/mm/pgtable.c
@@ -49,16 +49,16 @@ void show_mem(unsigned int filter)
global_node_page_state(NR_ACTIVE_FILE)),
(global_node_page_state(NR_INACTIVE_ANON) +
global_node_page_state(NR_INACTIVE_FILE)),
- global_page_state(NR_FILE_DIRTY),
- global_page_state(NR_WRITEBACK),
- global_page_state(NR_UNSTABLE_NFS),
+ global_node_page_state(NR_FILE_DIRTY),
+ global_node_page_state(NR_WRITEBACK),
+ global_node_page_state(NR_UNSTABLE_NFS),
global_page_state(NR_FREE_PAGES),
(global_page_state(NR_SLAB_RECLAIMABLE) +
global_page_state(NR_SLAB_UNRECLAIMABLE)),
global_node_page_state(NR_FILE_MAPPED),
global_page_state(NR_PAGETABLE),
global_page_state(NR_BOUNCE),
- global_page_state(NR_FILE_PAGES),
+ global_node_page_state(NR_FILE_PAGES),
get_nr_swap_pages());
for_each_zone(zone) {
diff --git a/drivers/base/node.c b/drivers/base/node.c
index 897b6bcb36be..ec733919bc6b 100644
--- a/drivers/base/node.c
+++ b/drivers/base/node.c
@@ -116,18 +116,18 @@ static ssize_t node_read_meminfo(struct device *dev,
"Node %d AnonHugePages: %8lu kB\n"
#endif
,
- nid, K(sum_zone_node_page_state(nid, NR_FILE_DIRTY)),
- nid, K(sum_zone_node_page_state(nid, NR_WRITEBACK)),
- nid, K(sum_zone_node_page_state(nid, NR_FILE_PAGES)),
+ nid, K(node_page_state(pgdat, NR_FILE_DIRTY)),
+ nid, K(node_page_state(pgdat, NR_WRITEBACK)),
+ nid, K(node_page_state(pgdat, NR_FILE_PAGES)),
nid, K(node_page_state(pgdat, NR_FILE_MAPPED)),
nid, K(node_page_state(pgdat, NR_ANON_MAPPED)),
nid, K(i.sharedram),
nid, sum_zone_node_page_state(nid, NR_KERNEL_STACK) *
THREAD_SIZE / 1024,
nid, K(sum_zone_node_page_state(nid, NR_PAGETABLE)),
- nid, K(sum_zone_node_page_state(nid, NR_UNSTABLE_NFS)),
+ nid, K(node_page_state(pgdat, NR_UNSTABLE_NFS)),
nid, K(sum_zone_node_page_state(nid, NR_BOUNCE)),
- nid, K(sum_zone_node_page_state(nid, NR_WRITEBACK_TEMP)),
+ nid, K(node_page_state(pgdat, NR_WRITEBACK_TEMP)),
nid, K(sum_zone_node_page_state(nid, NR_SLAB_RECLAIMABLE) +
sum_zone_node_page_state(nid, NR_SLAB_UNRECLAIMABLE)),
nid, K(sum_zone_node_page_state(nid, NR_SLAB_RECLAIMABLE)),
diff --git a/drivers/staging/android/lowmemorykiller.c b/drivers/staging/android/lowmemorykiller.c
index 93dbcc38eb0f..45a1b4ec4ca3 100644
--- a/drivers/staging/android/lowmemorykiller.c
+++ b/drivers/staging/android/lowmemorykiller.c
@@ -91,8 +91,8 @@ static unsigned long lowmem_scan(struct shrinker *s, struct shrink_control *sc)
short selected_oom_score_adj;
int array_size = ARRAY_SIZE(lowmem_adj);
int other_free = global_page_state(NR_FREE_PAGES) - totalreserve_pages;
- int other_file = global_page_state(NR_FILE_PAGES) -
- global_page_state(NR_SHMEM) -
+ int other_file = global_node_page_state(NR_FILE_PAGES) -
+ global_node_page_state(NR_SHMEM) -
total_swapcache_pages();
if (lowmem_adj_size < array_size)
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 989a2cef6b76..fd68f8efb440 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -1771,8 +1771,8 @@ static struct wb_writeback_work *get_next_work_item(struct bdi_writeback *wb)
*/
static unsigned long get_nr_dirty_pages(void)
{
- return global_page_state(NR_FILE_DIRTY) +
- global_page_state(NR_UNSTABLE_NFS) +
+ return global_node_page_state(NR_FILE_DIRTY) +
+ global_node_page_state(NR_UNSTABLE_NFS) +
get_nr_dirty_inodes();
}
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index 9154f8679024..2382f22a2a8b 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -1452,7 +1452,7 @@ static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req)
list_del(&req->writepages_entry);
for (i = 0; i < req->num_pages; i++) {
dec_wb_stat(&bdi->wb, WB_WRITEBACK);
- dec_zone_page_state(req->pages[i], NR_WRITEBACK_TEMP);
+ dec_node_page_state(req->pages[i], NR_WRITEBACK_TEMP);
wb_writeout_inc(&bdi->wb);
}
wake_up(&fi->page_waitq);
@@ -1642,7 +1642,7 @@ static int fuse_writepage_locked(struct page *page)
req->inode = inode;
inc_wb_stat(&inode_to_bdi(inode)->wb, WB_WRITEBACK);
- inc_zone_page_state(tmp_page, NR_WRITEBACK_TEMP);
+ inc_node_page_state(tmp_page, NR_WRITEBACK_TEMP);
spin_lock(&fc->lock);
list_add(&req->writepages_entry, &fi->writepages);
@@ -1756,7 +1756,7 @@ static bool fuse_writepage_in_flight(struct fuse_req *new_req,
spin_unlock(&fc->lock);
dec_wb_stat(&bdi->wb, WB_WRITEBACK);
- dec_zone_page_state(page, NR_WRITEBACK_TEMP);
+ dec_node_page_state(page, NR_WRITEBACK_TEMP);
wb_writeout_inc(&bdi->wb);
fuse_writepage_free(fc, new_req);
fuse_request_free(new_req);
@@ -1855,7 +1855,7 @@ static int fuse_writepages_fill(struct page *page,
req->page_descs[req->num_pages].length = PAGE_SIZE;
inc_wb_stat(&inode_to_bdi(inode)->wb, WB_WRITEBACK);
- inc_zone_page_state(tmp_page, NR_WRITEBACK_TEMP);
+ inc_node_page_state(tmp_page, NR_WRITEBACK_TEMP);
err = 0;
if (is_writeback && fuse_writepage_in_flight(req, page)) {
diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index 5154fa65a2f2..5ea04d87fc65 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -623,7 +623,7 @@ void nfs_mark_page_unstable(struct page *page, struct nfs_commit_info *cinfo)
if (!cinfo->dreq) {
struct inode *inode = page_file_mapping(page)->host;
- inc_zone_page_state(page, NR_UNSTABLE_NFS);
+ inc_node_page_state(page, NR_UNSTABLE_NFS);
inc_wb_stat(&inode_to_bdi(inode)->wb, WB_RECLAIMABLE);
__mark_inode_dirty(inode, I_DIRTY_DATASYNC);
}
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index e1c74d3db64d..593fa21a02c0 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -898,7 +898,7 @@ nfs_mark_request_commit(struct nfs_page *req, struct pnfs_layout_segment *lseg,
static void
nfs_clear_page_commit(struct page *page)
{
- dec_zone_page_state(page, NR_UNSTABLE_NFS);
+ dec_node_page_state(page, NR_UNSTABLE_NFS);
dec_wb_stat(&inode_to_bdi(page_file_mapping(page)->host)->wb,
WB_RECLAIMABLE);
}
diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
index 076afb43fc56..6cb9ea36d0fc 100644
--- a/fs/proc/meminfo.c
+++ b/fs/proc/meminfo.c
@@ -40,7 +40,7 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
si_swapinfo(&i);
committed = percpu_counter_read_positive(&vm_committed_as);
- cached = global_page_state(NR_FILE_PAGES) -
+ cached = global_node_page_state(NR_FILE_PAGES) -
total_swapcache_pages() - i.bufferram;
if (cached < 0)
cached = 0;
@@ -136,8 +136,8 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
#endif
K(i.totalswap),
K(i.freeswap),
- K(global_page_state(NR_FILE_DIRTY)),
- K(global_page_state(NR_WRITEBACK)),
+ K(global_node_page_state(NR_FILE_DIRTY)),
+ K(global_node_page_state(NR_WRITEBACK)),
K(global_node_page_state(NR_ANON_MAPPED)),
K(global_node_page_state(NR_FILE_MAPPED)),
K(i.sharedram),
@@ -150,9 +150,9 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
#ifdef CONFIG_QUICKLIST
K(quicklist_total_size()),
#endif
- K(global_page_state(NR_UNSTABLE_NFS)),
+ K(global_node_page_state(NR_UNSTABLE_NFS)),
K(global_page_state(NR_BOUNCE)),
- K(global_page_state(NR_WRITEBACK_TEMP)),
+ K(global_node_page_state(NR_WRITEBACK_TEMP)),
K(vm_commit_limit()),
K(committed),
(unsigned long)VMALLOC_TOTAL >> 10,
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 8fbb2a1f7a0f..f36af7e9f536 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -112,20 +112,14 @@ enum zone_stat_item {
NR_FREE_PAGES,
NR_ALLOC_BATCH,
NR_MLOCK, /* mlock()ed pages found and moved off LRU */
- NR_FILE_PAGES,
- NR_FILE_DIRTY,
- NR_WRITEBACK,
NR_SLAB_RECLAIMABLE,
NR_SLAB_UNRECLAIMABLE,
NR_PAGETABLE, /* used for pagetables */
NR_KERNEL_STACK,
/* Second 128 byte cacheline */
- NR_UNSTABLE_NFS, /* NFS unstable pages */
NR_BOUNCE,
NR_VMSCAN_WRITE,
NR_VMSCAN_IMMEDIATE, /* Prioritise for reclaim when writeback ends */
- NR_WRITEBACK_TEMP, /* Writeback using temporary buffers */
- NR_SHMEM, /* shmem pages (included tmpfs/GEM pages) */
NR_DIRTIED, /* page dirtyings since bootup */
NR_WRITTEN, /* page writings since bootup */
#if IS_ENABLED(CONFIG_ZSMALLOC)
@@ -159,6 +153,12 @@ enum node_stat_item {
NR_ANON_MAPPED, /* Mapped anonymous pages */
NR_FILE_MAPPED, /* pagecache pages mapped into pagetables.
only modified from process context */
+ NR_FILE_PAGES,
+ NR_FILE_DIRTY,
+ NR_WRITEBACK,
+ NR_WRITEBACK_TEMP, /* Writeback using temporary buffers */
+ NR_SHMEM, /* shmem pages (included tmpfs/GEM pages) */
+ NR_UNSTABLE_NFS, /* NFS unstable pages */
NR_VM_NODE_STAT_ITEMS
};
diff --git a/include/trace/events/writeback.h b/include/trace/events/writeback.h
index 73614ce1d204..c581d9c04ca5 100644
--- a/include/trace/events/writeback.h
+++ b/include/trace/events/writeback.h
@@ -412,9 +412,9 @@ TRACE_EVENT(global_dirty_state,
),
TP_fast_assign(
- __entry->nr_dirty = global_page_state(NR_FILE_DIRTY);
- __entry->nr_writeback = global_page_state(NR_WRITEBACK);
- __entry->nr_unstable = global_page_state(NR_UNSTABLE_NFS);
+ __entry->nr_dirty = global_node_page_state(NR_FILE_DIRTY);
+ __entry->nr_writeback = global_node_page_state(NR_WRITEBACK);
+ __entry->nr_unstable = global_node_page_state(NR_UNSTABLE_NFS);
__entry->nr_dirtied = global_page_state(NR_DIRTIED);
__entry->nr_written = global_page_state(NR_WRITTEN);
__entry->background_thresh = background_thresh;
diff --git a/mm/filemap.c b/mm/filemap.c
index bce091ed90d7..479d52fbc422 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -209,9 +209,9 @@ void __delete_from_page_cache(struct page *page, void *shadow)
/* hugetlb pages do not participate in page cache accounting. */
if (!PageHuge(page))
- __dec_zone_page_state(page, NR_FILE_PAGES);
+ __dec_node_page_state(page, NR_FILE_PAGES);
if (PageSwapBacked(page))
- __dec_zone_page_state(page, NR_SHMEM);
+ __dec_node_page_state(page, NR_SHMEM);
/*
* At this point page must be either written or cleaned by truncate.
@@ -549,9 +549,9 @@ int replace_page_cache_page(struct page *old, struct page *new, gfp_t gfp_mask)
* hugetlb pages do not participate in page cache accounting.
*/
if (!PageHuge(new))
- __inc_zone_page_state(new, NR_FILE_PAGES);
+ __inc_node_page_state(new, NR_FILE_PAGES);
if (PageSwapBacked(new))
- __inc_zone_page_state(new, NR_SHMEM);
+ __inc_node_page_state(new, NR_SHMEM);
spin_unlock_irqrestore(&mapping->tree_lock, flags);
mem_cgroup_migrate(old, new);
radix_tree_preload_end();
@@ -658,7 +658,7 @@ static int __add_to_page_cache_locked(struct page *page,
/* hugetlb pages do not participate in page cache accounting. */
if (!huge)
- __inc_zone_page_state(page, NR_FILE_PAGES);
+ __inc_node_page_state(page, NR_FILE_PAGES);
spin_unlock_irq(&mapping->tree_lock);
if (!huge)
mem_cgroup_commit_charge(page, memcg, false, false);
diff --git a/mm/migrate.c b/mm/migrate.c
index 133772d8f08f..16add0d43994 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -505,15 +505,15 @@ int migrate_page_move_mapping(struct address_space *mapping,
* are mapped to swap space.
*/
if (newzone != oldzone) {
- __dec_zone_state(oldzone, NR_FILE_PAGES);
- __inc_zone_state(newzone, NR_FILE_PAGES);
+ __dec_node_state(oldzone->zone_pgdat, NR_FILE_PAGES);
+ __inc_node_state(newzone->zone_pgdat, NR_FILE_PAGES);
if (PageSwapBacked(page) && !PageSwapCache(page)) {
- __dec_zone_state(oldzone, NR_SHMEM);
- __inc_zone_state(newzone, NR_SHMEM);
+ __dec_node_state(oldzone->zone_pgdat, NR_SHMEM);
+ __inc_node_state(newzone->zone_pgdat, NR_SHMEM);
}
if (dirty && mapping_cap_account_dirty(mapping)) {
- __dec_zone_state(oldzone, NR_FILE_DIRTY);
- __inc_zone_state(newzone, NR_FILE_DIRTY);
+ __dec_node_state(oldzone->zone_pgdat, NR_FILE_DIRTY);
+ __inc_node_state(newzone->zone_pgdat, NR_FILE_DIRTY);
}
}
local_irq_enable();
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index 4db3be44bb1d..af8e620ead97 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -468,20 +468,12 @@ static unsigned long node_dirty_limit(struct pglist_data *pgdat)
*/
bool node_dirty_ok(struct pglist_data *pgdat)
{
- int z;
unsigned long limit = node_dirty_limit(pgdat);
unsigned long nr_pages = 0;
- for (z = 0; z < MAX_NR_ZONES; z++) {
- struct zone *zone = pgdat->node_zones + z;
-
- if (!populated_zone(zone))
- continue;
-
- nr_pages += zone_page_state(zone, NR_FILE_DIRTY);
- nr_pages += zone_page_state(zone, NR_UNSTABLE_NFS);
- nr_pages += zone_page_state(zone, NR_WRITEBACK);
- }
+ nr_pages += node_page_state(pgdat, NR_FILE_DIRTY);
+ nr_pages += node_page_state(pgdat, NR_UNSTABLE_NFS);
+ nr_pages += node_page_state(pgdat, NR_WRITEBACK);
return nr_pages <= limit;
}
@@ -1571,10 +1563,10 @@ static void balance_dirty_pages(struct address_space *mapping,
* written to the server's write cache, but has not yet
* been flushed to permanent storage.
*/
- nr_reclaimable = global_page_state(NR_FILE_DIRTY) +
- global_page_state(NR_UNSTABLE_NFS);
+ nr_reclaimable = global_node_page_state(NR_FILE_DIRTY) +
+ global_node_page_state(NR_UNSTABLE_NFS);
gdtc->avail = global_dirtyable_memory();
- gdtc->dirty = nr_reclaimable + global_page_state(NR_WRITEBACK);
+ gdtc->dirty = nr_reclaimable + global_node_page_state(NR_WRITEBACK);
domain_dirty_limits(gdtc);
@@ -1911,8 +1903,8 @@ bool wb_over_bg_thresh(struct bdi_writeback *wb)
* as we're trying to decide whether to put more under writeback.
*/
gdtc->avail = global_dirtyable_memory();
- gdtc->dirty = global_page_state(NR_FILE_DIRTY) +
- global_page_state(NR_UNSTABLE_NFS);
+ gdtc->dirty = global_node_page_state(NR_FILE_DIRTY) +
+ global_node_page_state(NR_UNSTABLE_NFS);
domain_dirty_limits(gdtc);
if (gdtc->dirty > gdtc->bg_thresh)
@@ -1956,8 +1948,8 @@ void throttle_vm_writeout(gfp_t gfp_mask)
*/
dirty_thresh += dirty_thresh / 10; /* wheeee... */
- if (global_page_state(NR_UNSTABLE_NFS) +
- global_page_state(NR_WRITEBACK) <= dirty_thresh)
+ if (global_node_page_state(NR_UNSTABLE_NFS) +
+ global_node_page_state(NR_WRITEBACK) <= dirty_thresh)
break;
congestion_wait(BLK_RW_ASYNC, HZ/10);
@@ -1985,8 +1977,8 @@ int dirty_writeback_centisecs_handler(struct ctl_table *table, int write,
void laptop_mode_timer_fn(unsigned long data)
{
struct request_queue *q = (struct request_queue *)data;
- int nr_pages = global_page_state(NR_FILE_DIRTY) +
- global_page_state(NR_UNSTABLE_NFS);
+ int nr_pages = global_node_page_state(NR_FILE_DIRTY) +
+ global_node_page_state(NR_UNSTABLE_NFS);
struct bdi_writeback *wb;
/*
@@ -2437,7 +2429,7 @@ void account_page_dirtied(struct page *page, struct address_space *mapping)
wb = inode_to_wb(inode);
mem_cgroup_inc_page_stat(page, MEM_CGROUP_STAT_DIRTY);
- __inc_zone_page_state(page, NR_FILE_DIRTY);
+ __inc_node_page_state(page, NR_FILE_DIRTY);
__inc_zone_page_state(page, NR_DIRTIED);
__inc_wb_stat(wb, WB_RECLAIMABLE);
__inc_wb_stat(wb, WB_DIRTIED);
@@ -2458,7 +2450,7 @@ void account_page_cleaned(struct page *page, struct address_space *mapping,
{
if (mapping_cap_account_dirty(mapping)) {
mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_DIRTY);
- dec_zone_page_state(page, NR_FILE_DIRTY);
+ dec_node_page_state(page, NR_FILE_DIRTY);
dec_wb_stat(wb, WB_RECLAIMABLE);
task_io_account_cancelled_write(PAGE_SIZE);
}
@@ -2713,7 +2705,7 @@ int clear_page_dirty_for_io(struct page *page)
wb = unlocked_inode_to_wb_begin(inode, &locked);
if (TestClearPageDirty(page)) {
mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_DIRTY);
- dec_zone_page_state(page, NR_FILE_DIRTY);
+ dec_node_page_state(page, NR_FILE_DIRTY);
dec_wb_stat(wb, WB_RECLAIMABLE);
ret = 1;
}
@@ -2754,7 +2746,7 @@ int test_clear_page_writeback(struct page *page)
}
if (ret) {
mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_WRITEBACK);
- dec_zone_page_state(page, NR_WRITEBACK);
+ dec_node_page_state(page, NR_WRITEBACK);
inc_zone_page_state(page, NR_WRITTEN);
}
unlock_page_memcg(page);
@@ -2795,7 +2787,7 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
}
if (!ret) {
mem_cgroup_inc_page_stat(page, MEM_CGROUP_STAT_WRITEBACK);
- inc_zone_page_state(page, NR_WRITEBACK);
+ inc_node_page_state(page, NR_WRITEBACK);
}
unlock_page_memcg(page);
return ret;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 1ba0c98d9461..8faeb0cb21f0 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -4189,7 +4189,7 @@ EXPORT_SYMBOL_GPL(si_mem_available);
void si_meminfo(struct sysinfo *val)
{
val->totalram = totalram_pages;
- val->sharedram = global_page_state(NR_SHMEM);
+ val->sharedram = global_node_page_state(NR_SHMEM);
val->freeram = global_page_state(NR_FREE_PAGES);
val->bufferram = nr_blockdev_pages();
val->totalhigh = totalhigh_pages;
@@ -4211,7 +4211,7 @@ void si_meminfo_node(struct sysinfo *val, int nid)
for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++)
managed_pages += pgdat->node_zones[zone_type].managed_pages;
val->totalram = managed_pages;
- val->sharedram = sum_zone_node_page_state(nid, NR_SHMEM);
+ val->sharedram = node_page_state(pgdat, NR_SHMEM);
val->freeram = sum_zone_node_page_state(nid, NR_FREE_PAGES);
#ifdef CONFIG_HIGHMEM
for (zone_type = 0; zone_type < MAX_NR_ZONES; zone_type++) {
@@ -4318,13 +4318,13 @@ void show_free_areas(unsigned int filter)
global_node_page_state(NR_INACTIVE_FILE),
global_node_page_state(NR_ISOLATED_FILE),
global_node_page_state(NR_UNEVICTABLE),
- global_page_state(NR_FILE_DIRTY),
- global_page_state(NR_WRITEBACK),
- global_page_state(NR_UNSTABLE_NFS),
+ global_node_page_state(NR_FILE_DIRTY),
+ global_node_page_state(NR_WRITEBACK),
+ global_node_page_state(NR_UNSTABLE_NFS),
global_page_state(NR_SLAB_RECLAIMABLE),
global_page_state(NR_SLAB_UNRECLAIMABLE),
global_node_page_state(NR_FILE_MAPPED),
- global_page_state(NR_SHMEM),
+ global_node_page_state(NR_SHMEM),
global_page_state(NR_PAGETABLE),
global_page_state(NR_BOUNCE),
global_page_state(NR_FREE_PAGES),
@@ -4341,6 +4341,11 @@ void show_free_areas(unsigned int filter)
" isolated(anon):%lukB"
" isolated(file):%lukB"
" mapped:%lukB"
+ " dirty:%lukB"
+ " writeback:%lukB"
+ " shmem:%lukB"
+ " writeback_tmp:%lukB"
+ " unstable:%lukB"
" all_unreclaimable? %s"
"\n",
pgdat->node_id,
@@ -4352,6 +4357,11 @@ void show_free_areas(unsigned int filter)
K(node_page_state(pgdat, NR_ISOLATED_ANON)),
K(node_page_state(pgdat, NR_ISOLATED_FILE)),
K(node_page_state(pgdat, NR_FILE_MAPPED)),
+ K(node_page_state(pgdat, NR_FILE_DIRTY)),
+ K(node_page_state(pgdat, NR_WRITEBACK)),
+ K(node_page_state(pgdat, NR_SHMEM)),
+ K(node_page_state(pgdat, NR_WRITEBACK_TEMP)),
+ K(node_page_state(pgdat, NR_UNSTABLE_NFS)),
!pgdat_reclaimable(pgdat) ? "yes" : "no");
}
@@ -4374,19 +4384,14 @@ void show_free_areas(unsigned int filter)
" present:%lukB"
" managed:%lukB"
" mlocked:%lukB"
- " dirty:%lukB"
- " writeback:%lukB"
- " shmem:%lukB"
" slab_reclaimable:%lukB"
" slab_unreclaimable:%lukB"
" kernel_stack:%lukB"
" pagetables:%lukB"
- " unstable:%lukB"
" bounce:%lukB"
" free_pcp:%lukB"
" local_pcp:%ukB"
" free_cma:%lukB"
- " writeback_tmp:%lukB"
" node_pages_scanned:%lu"
"\n",
zone->name,
@@ -4397,20 +4402,15 @@ void show_free_areas(unsigned int filter)
K(zone->present_pages),
K(zone->managed_pages),
K(zone_page_state(zone, NR_MLOCK)),
- K(zone_page_state(zone, NR_FILE_DIRTY)),
- K(zone_page_state(zone, NR_WRITEBACK)),
- K(zone_page_state(zone, NR_SHMEM)),
K(zone_page_state(zone, NR_SLAB_RECLAIMABLE)),
K(zone_page_state(zone, NR_SLAB_UNRECLAIMABLE)),
zone_page_state(zone, NR_KERNEL_STACK) *
THREAD_SIZE / 1024,
K(zone_page_state(zone, NR_PAGETABLE)),
- K(zone_page_state(zone, NR_UNSTABLE_NFS)),
K(zone_page_state(zone, NR_BOUNCE)),
K(free_pcp),
K(this_cpu_read(zone->pageset->pcp.count)),
K(zone_page_state(zone, NR_FREE_CMA_PAGES)),
- K(zone_page_state(zone, NR_WRITEBACK_TEMP)),
K(node_page_state(zone->zone_pgdat, NR_PAGES_SCANNED)));
printk("lowmem_reserve[]:");
for (i = 0; i < MAX_NR_ZONES; i++)
@@ -4453,7 +4453,7 @@ void show_free_areas(unsigned int filter)
hugetlb_show_meminfo();
- printk("%ld total pagecache pages\n", global_page_state(NR_FILE_PAGES));
+ printk("%ld total pagecache pages\n", global_node_page_state(NR_FILE_PAGES));
show_swap_cache_info();
}
diff --git a/mm/shmem.c b/mm/shmem.c
index a36144909b28..745d6d62ea1f 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -312,8 +312,8 @@ static int shmem_add_to_page_cache(struct page *page,
page);
if (!error) {
mapping->nrpages++;
- __inc_zone_page_state(page, NR_FILE_PAGES);
- __inc_zone_page_state(page, NR_SHMEM);
+ __inc_node_page_state(page, NR_FILE_PAGES);
+ __inc_node_page_state(page, NR_SHMEM);
spin_unlock_irq(&mapping->tree_lock);
} else {
page->mapping = NULL;
@@ -335,8 +335,8 @@ static void shmem_delete_from_page_cache(struct page *page, void *radswap)
error = shmem_radix_tree_replace(mapping, page->index, page, radswap);
page->mapping = NULL;
mapping->nrpages--;
- __dec_zone_page_state(page, NR_FILE_PAGES);
- __dec_zone_page_state(page, NR_SHMEM);
+ __dec_node_page_state(page, NR_FILE_PAGES);
+ __dec_node_page_state(page, NR_SHMEM);
spin_unlock_irq(&mapping->tree_lock);
put_page(page);
BUG_ON(error);
@@ -1084,8 +1084,8 @@ static int shmem_replace_page(struct page **pagep, gfp_t gfp,
error = shmem_radix_tree_replace(swap_mapping, swap_index, oldpage,
newpage);
if (!error) {
- __inc_zone_page_state(newpage, NR_FILE_PAGES);
- __dec_zone_page_state(oldpage, NR_FILE_PAGES);
+ __inc_node_page_state(newpage, NR_FILE_PAGES);
+ __dec_node_page_state(oldpage, NR_FILE_PAGES);
}
spin_unlock_irq(&swap_mapping->tree_lock);
diff --git a/mm/swap_state.c b/mm/swap_state.c
index c99463ac02fb..c8310a37be3a 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -95,7 +95,7 @@ int __add_to_swap_cache(struct page *page, swp_entry_t entry)
entry.val, page);
if (likely(!error)) {
address_space->nrpages++;
- __inc_zone_page_state(page, NR_FILE_PAGES);
+ __inc_node_page_state(page, NR_FILE_PAGES);
INC_CACHE_INFO(add_total);
}
spin_unlock_irq(&address_space->tree_lock);
@@ -147,7 +147,7 @@ void __delete_from_swap_cache(struct page *page)
set_page_private(page, 0);
ClearPageSwapCache(page);
address_space->nrpages--;
- __dec_zone_page_state(page, NR_FILE_PAGES);
+ __dec_node_page_state(page, NR_FILE_PAGES);
INC_CACHE_INFO(del_total);
}
diff --git a/mm/util.c b/mm/util.c
index b756ee36f7f0..21cbc0fc2fd4 100644
--- a/mm/util.c
+++ b/mm/util.c
@@ -522,7 +522,7 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin)
if (sysctl_overcommit_memory == OVERCOMMIT_GUESS) {
free = global_page_state(NR_FREE_PAGES);
- free += global_page_state(NR_FILE_PAGES);
+ free += global_node_page_state(NR_FILE_PAGES);
/*
* shmem pages shouldn't be counted as free in this
@@ -530,7 +530,7 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin)
* that won't affect the overall amount of available
* memory in the system.
*/
- free -= global_page_state(NR_SHMEM);
+ free -= global_node_page_state(NR_SHMEM);
free += get_nr_swap_pages();
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 56482c3fbc76..71d3330a3d33 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3554,11 +3554,11 @@ int sysctl_min_unmapped_ratio = 1;
*/
int sysctl_min_slab_ratio = 5;
-static inline unsigned long zone_unmapped_file_pages(struct zone *zone)
+static inline unsigned long node_unmapped_file_pages(struct pglist_data *pgdat)
{
- unsigned long file_mapped = node_page_state(zone->zone_pgdat, NR_FILE_MAPPED);
- unsigned long file_lru = node_page_state(zone->zone_pgdat, NR_INACTIVE_FILE) +
- node_page_state(zone->zone_pgdat, NR_ACTIVE_FILE);
+ unsigned long file_mapped = node_page_state(pgdat, NR_FILE_MAPPED);
+ unsigned long file_lru = node_page_state(pgdat, NR_INACTIVE_FILE) +
+ node_page_state(pgdat, NR_ACTIVE_FILE);
/*
* It's possible for there to be more file mapped pages than
@@ -3577,17 +3577,17 @@ static unsigned long zone_pagecache_reclaimable(struct zone *zone)
/*
* If RECLAIM_UNMAP is set, then all file pages are considered
* potentially reclaimable. Otherwise, we have to worry about
- * pages like swapcache and zone_unmapped_file_pages() provides
+ * pages like swapcache and node_unmapped_file_pages() provides
* a better estimate
*/
if (zone_reclaim_mode & RECLAIM_UNMAP)
- nr_pagecache_reclaimable = zone_page_state(zone, NR_FILE_PAGES);
+ nr_pagecache_reclaimable = node_page_state(zone->zone_pgdat, NR_FILE_PAGES);
else
- nr_pagecache_reclaimable = zone_unmapped_file_pages(zone);
+ nr_pagecache_reclaimable = node_unmapped_file_pages(zone->zone_pgdat);
/* If we can't clean pages, remove dirty pages from consideration */
if (!(zone_reclaim_mode & RECLAIM_WRITE))
- delta += zone_page_state(zone, NR_FILE_DIRTY);
+ delta += node_page_state(zone->zone_pgdat, NR_FILE_DIRTY);
/* Watch for any possible underflows due to delta */
if (unlikely(delta > nr_pagecache_reclaimable))
diff --git a/mm/vmstat.c b/mm/vmstat.c
index 960f549a7b1a..a4473c69342b 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -922,19 +922,13 @@ const char * const vmstat_text[] = {
"nr_free_pages",
"nr_alloc_batch",
"nr_mlock",
- "nr_file_pages",
- "nr_dirty",
- "nr_writeback",
"nr_slab_reclaimable",
"nr_slab_unreclaimable",
"nr_page_table_pages",
"nr_kernel_stack",
- "nr_unstable",
"nr_bounce",
"nr_vmscan_write",
"nr_vmscan_immediate_reclaim",
- "nr_writeback_temp",
- "nr_shmem",
"nr_dirtied",
"nr_written",
#if IS_ENABLED(CONFIG_ZSMALLOC)
@@ -965,6 +959,12 @@ const char * const vmstat_text[] = {
"workingset_nodereclaim",
"nr_anon_pages",
"nr_mapped",
+ "nr_file_pages",
+ "nr_dirty",
+ "nr_writeback",
+ "nr_writeback_temp",
+ "nr_shmem",
+ "nr_unstable",
/* enum writeback_stat_item counters */
"nr_dirty_threshold",
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 18/27] mm: Move most file-based accounting to the node
2016-06-09 18:04 ` [PATCH 18/27] mm: Move most file-based accounting to the node Mel Gorman
@ 2016-06-17 8:35 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 8:35 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> There are now a number of accounting oddities such as mapped file pages
> being accounted for on the node while the total number of file pages are
> accounted on the zone. This can be coped with to some extent but it's
> confusing so this patch moves the relevant file-based accounted.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 19/27] mm: Move vmscan writes and file write accounting to the node
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (17 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 18/27] mm: Move most file-based accounting to the node Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 8:51 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 20/27] mm, vmscan: Update classzone_idx if buffer_heads_over_limit Mel Gorman
` (7 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
As reclaim is now node-based, it follows that page write activity
due to page reclaim should also be accounted for on the node. For
consistency, also account page writes and page dirtying on a per-node
basis.
After this patch, there are a few remaining zone counters that may
appear strange but are fine. NUMA stats are still per-zone as this is a
user-space interface that tools consume. NR_MLOCK, NR_SLAB_*, NR_PAGETABLE,
NR_KERNEL_STACK and NR_BOUNCE are all allocations that potentially pin
low memory and cannot trivially be reclaimed on demand. This information
is still useful for debugging a page allocation failure warning.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
include/linux/mmzone.h | 8 ++++----
include/trace/events/writeback.h | 4 ++--
mm/page-writeback.c | 6 +++---
mm/vmscan.c | 4 ++--
mm/vmstat.c | 8 ++++----
5 files changed, 15 insertions(+), 15 deletions(-)
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index f36af7e9f536..9cc75f88bb8e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -118,10 +118,6 @@ enum zone_stat_item {
NR_KERNEL_STACK,
/* Second 128 byte cacheline */
NR_BOUNCE,
- NR_VMSCAN_WRITE,
- NR_VMSCAN_IMMEDIATE, /* Prioritise for reclaim when writeback ends */
- NR_DIRTIED, /* page dirtyings since bootup */
- NR_WRITTEN, /* page writings since bootup */
#if IS_ENABLED(CONFIG_ZSMALLOC)
NR_ZSMALLOC,
#endif
@@ -159,6 +155,10 @@ enum node_stat_item {
NR_WRITEBACK_TEMP, /* Writeback using temporary buffers */
NR_SHMEM, /* shmem pages (included tmpfs/GEM pages) */
NR_UNSTABLE_NFS, /* NFS unstable pages */
+ NR_VMSCAN_WRITE,
+ NR_VMSCAN_IMMEDIATE, /* Prioritise for reclaim when writeback ends */
+ NR_DIRTIED, /* page dirtyings since bootup */
+ NR_WRITTEN, /* page writings since bootup */
NR_VM_NODE_STAT_ITEMS
};
diff --git a/include/trace/events/writeback.h b/include/trace/events/writeback.h
index c581d9c04ca5..a1002cfede1d 100644
--- a/include/trace/events/writeback.h
+++ b/include/trace/events/writeback.h
@@ -415,8 +415,8 @@ TRACE_EVENT(global_dirty_state,
__entry->nr_dirty = global_node_page_state(NR_FILE_DIRTY);
__entry->nr_writeback = global_node_page_state(NR_WRITEBACK);
__entry->nr_unstable = global_node_page_state(NR_UNSTABLE_NFS);
- __entry->nr_dirtied = global_page_state(NR_DIRTIED);
- __entry->nr_written = global_page_state(NR_WRITTEN);
+ __entry->nr_dirtied = global_node_page_state(NR_DIRTIED);
+ __entry->nr_written = global_node_page_state(NR_WRITTEN);
__entry->background_thresh = background_thresh;
__entry->dirty_thresh = dirty_thresh;
__entry->dirty_limit = global_wb_domain.dirty_limit;
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index af8e620ead97..67b17aafc63b 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -2430,7 +2430,7 @@ void account_page_dirtied(struct page *page, struct address_space *mapping)
mem_cgroup_inc_page_stat(page, MEM_CGROUP_STAT_DIRTY);
__inc_node_page_state(page, NR_FILE_DIRTY);
- __inc_zone_page_state(page, NR_DIRTIED);
+ __inc_node_page_state(page, NR_DIRTIED);
__inc_wb_stat(wb, WB_RECLAIMABLE);
__inc_wb_stat(wb, WB_DIRTIED);
task_io_account_write(PAGE_SIZE);
@@ -2518,7 +2518,7 @@ void account_page_redirty(struct page *page)
wb = unlocked_inode_to_wb_begin(inode, &locked);
current->nr_dirtied--;
- dec_zone_page_state(page, NR_DIRTIED);
+ dec_node_page_state(page, NR_DIRTIED);
dec_wb_stat(wb, WB_DIRTIED);
unlocked_inode_to_wb_end(inode, locked);
}
@@ -2747,7 +2747,7 @@ int test_clear_page_writeback(struct page *page)
if (ret) {
mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_WRITEBACK);
dec_node_page_state(page, NR_WRITEBACK);
- inc_zone_page_state(page, NR_WRITTEN);
+ inc_node_page_state(page, NR_WRITTEN);
}
unlock_page_memcg(page);
return ret;
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 71d3330a3d33..3edf941e9965 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -596,7 +596,7 @@ static pageout_t pageout(struct page *page, struct address_space *mapping,
ClearPageReclaim(page);
}
trace_mm_vmscan_writepage(page);
- inc_zone_page_state(page, NR_VMSCAN_WRITE);
+ inc_node_page_state(page, NR_VMSCAN_WRITE);
return PAGE_SUCCESS;
}
@@ -1095,7 +1095,7 @@ static unsigned long shrink_page_list(struct list_head *page_list,
* except we already have the page isolated
* and know it's dirty
*/
- inc_zone_page_state(page, NR_VMSCAN_IMMEDIATE);
+ inc_node_page_state(page, NR_VMSCAN_IMMEDIATE);
SetPageReclaim(page);
goto keep_locked;
diff --git a/mm/vmstat.c b/mm/vmstat.c
index a4473c69342b..26e9873be7a2 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -927,10 +927,6 @@ const char * const vmstat_text[] = {
"nr_page_table_pages",
"nr_kernel_stack",
"nr_bounce",
- "nr_vmscan_write",
- "nr_vmscan_immediate_reclaim",
- "nr_dirtied",
- "nr_written",
#if IS_ENABLED(CONFIG_ZSMALLOC)
"nr_zsmalloc",
#endif
@@ -965,6 +961,10 @@ const char * const vmstat_text[] = {
"nr_writeback_temp",
"nr_shmem",
"nr_unstable",
+ "nr_vmscan_write",
+ "nr_vmscan_immediate_reclaim",
+ "nr_dirtied",
+ "nr_written",
/* enum writeback_stat_item counters */
"nr_dirty_threshold",
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 19/27] mm: Move vmscan writes and file write accounting to the node
2016-06-09 18:04 ` [PATCH 19/27] mm: Move vmscan writes and file write " Mel Gorman
@ 2016-06-17 8:51 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 8:51 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> As reclaim is now node-based, it follows that page write activity
> due to page reclaim should also be accounted for on the node. For
> consistency, also account page writes and page dirtying on a per-node
> basis.
>
> After this patch, there are a few remaining zone counters that may
> appear strange but are fine. NUMA stats are still per-zone as this is a
> user-space interface that tools consume. NR_MLOCK, NR_SLAB_*, NR_PAGETABLE,
> NR_KERNEL_STACK and NR_BOUNCE are all allocations that potentially pin
> low memory and cannot trivially be reclaimed on demand. This information
> is still useful for debugging a page allocation failure warning.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 20/27] mm, vmscan: Update classzone_idx if buffer_heads_over_limit
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (18 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 19/27] mm: Move vmscan writes and file write " Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 8:53 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 21/27] mm, vmscan: Only wakeup kswapd once per node for the requested classzone Mel Gorman
` (6 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
If buffer heads are over the limit then the direct reclaim gfp_mask
is promoted to __GFP_HIGHMEM so that lowmem is indirectly freed. With
node-based reclaim, it is also required that the classzone_idx be updated
or the pages will be skipped.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/vmscan.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 3edf941e9965..7a2d69612231 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -2547,8 +2547,10 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc,
* highmem pages could be pinning lowmem pages storing buffer_heads
*/
orig_mask = sc->gfp_mask;
- if (buffer_heads_over_limit)
+ if (buffer_heads_over_limit) {
sc->gfp_mask |= __GFP_HIGHMEM;
+ classzone_idx = gfp_zone(sc->gfp_mask);
+ }
for_each_zone_zonelist_nodemask(zone, z, zonelist,
gfp_zone(sc->gfp_mask), sc->nodemask) {
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 20/27] mm, vmscan: Update classzone_idx if buffer_heads_over_limit
2016-06-09 18:04 ` [PATCH 20/27] mm, vmscan: Update classzone_idx if buffer_heads_over_limit Mel Gorman
@ 2016-06-17 8:53 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 8:53 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> If buffer heads are over the limit then the direct reclaim gfp_mask
> is promoted to __GFP_HIGHMEM so that lowmem is indirectly freed. With
> node-based reclaim, it is also required that the classzone_idx be updated
> or the pages will be skipped.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
> ---
> mm/vmscan.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 3edf941e9965..7a2d69612231 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -2547,8 +2547,10 @@ static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc,
> * highmem pages could be pinning lowmem pages storing buffer_heads
> */
> orig_mask = sc->gfp_mask;
> - if (buffer_heads_over_limit)
> + if (buffer_heads_over_limit) {
> sc->gfp_mask |= __GFP_HIGHMEM;
> + classzone_idx = gfp_zone(sc->gfp_mask);
> + }
>
> for_each_zone_zonelist_nodemask(zone, z, zonelist,
> gfp_zone(sc->gfp_mask), sc->nodemask) {
>
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 21/27] mm, vmscan: Only wakeup kswapd once per node for the requested classzone
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (19 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 20/27] mm, vmscan: Update classzone_idx if buffer_heads_over_limit Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 10:46 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 22/27] mm: Convert zone_reclaim to node_reclaim Mel Gorman
` (5 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
kswapd is woken when zones are below the low watermark but the wakeup
decision is not taking the classzone into account. Now that reclaim is
node-based, it is only required to wake kswapd once per node and only if
all zones are unbalanced for the requested classzone.
Note that one node might be checked multiple times but there is no cheap
way of tracking what nodes have already been visited for zoneslists that
be ordered by either zone or node.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/vmscan.c | 13 +++++++++++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 7a2d69612231..b9cff9047ac0 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3389,6 +3389,7 @@ static int kswapd(void *p)
void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
{
pg_data_t *pgdat;
+ int z;
if (!populated_zone(zone))
return;
@@ -3402,8 +3403,16 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
pgdat->kswapd_order = max(pgdat->kswapd_order, order);
if (!waitqueue_active(&pgdat->kswapd_wait))
return;
- if (zone_balanced(zone, order, 0))
- return;
+
+ /* Only wake kswapd if all zones are unbalanced */
+ for (z = 0; z <= classzone_idx; z++) {
+ zone = pgdat->node_zones + z;
+ if (!populated_zone(zone))
+ continue;
+
+ if (zone_balanced(zone, order, classzone_idx))
+ return;
+ }
trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
wake_up_interruptible(&pgdat->kswapd_wait);
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 21/27] mm, vmscan: Only wakeup kswapd once per node for the requested classzone
2016-06-09 18:04 ` [PATCH 21/27] mm, vmscan: Only wakeup kswapd once per node for the requested classzone Mel Gorman
@ 2016-06-17 10:46 ` Vlastimil Babka
2016-06-17 12:03 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 10:46 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> kswapd is woken when zones are below the low watermark but the wakeup
> decision is not taking the classzone into account. Now that reclaim is
> node-based, it is only required to wake kswapd once per node and only if
> all zones are unbalanced for the requested classzone.
>
> Note that one node might be checked multiple times but there is no cheap
> way of tracking what nodes have already been visited for zoneslists that
> be ordered by either zone or node.
Wouldn't it be possible to optimize for node order as you did in direct reclaim?
Do the zone_balanced checks when going through zonelist, and once node changes
in iteration, wake up if no eligible zones visited so far were balanced.
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> ---
> mm/vmscan.c | 13 +++++++++++--
> 1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 7a2d69612231..b9cff9047ac0 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -3389,6 +3389,7 @@ static int kswapd(void *p)
> void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
> {
> pg_data_t *pgdat;
> + int z;
>
> if (!populated_zone(zone))
> return;
> @@ -3402,8 +3403,16 @@ void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx)
> pgdat->kswapd_order = max(pgdat->kswapd_order, order);
> if (!waitqueue_active(&pgdat->kswapd_wait))
> return;
> - if (zone_balanced(zone, order, 0))
> - return;
> +
> + /* Only wake kswapd if all zones are unbalanced */
> + for (z = 0; z <= classzone_idx; z++) {
> + zone = pgdat->node_zones + z;
> + if (!populated_zone(zone))
> + continue;
> +
> + if (zone_balanced(zone, order, classzone_idx))
> + return;
> + }
>
> trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order);
> wake_up_interruptible(&pgdat->kswapd_wait);
>
^ permalink raw reply [flat|nested] 70+ messages in thread* Re: [PATCH 21/27] mm, vmscan: Only wakeup kswapd once per node for the requested classzone
2016-06-17 10:46 ` Vlastimil Babka
@ 2016-06-17 12:03 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-17 12:03 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML
On Fri, Jun 17, 2016 at 12:46:05PM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> >kswapd is woken when zones are below the low watermark but the wakeup
> >decision is not taking the classzone into account. Now that reclaim is
> >node-based, it is only required to wake kswapd once per node and only if
> >all zones are unbalanced for the requested classzone.
> >
> >Note that one node might be checked multiple times but there is no cheap
> >way of tracking what nodes have already been visited for zoneslists that
> >be ordered by either zone or node.
>
> Wouldn't it be possible to optimize for node order as you did in direct
> reclaim? Do the zone_balanced checks when going through zonelist, and once
> node changes in iteration, wake up if no eligible zones visited so far were
> balanced.
>
Yeah, it is. I'll chuck it in.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 22/27] mm: Convert zone_reclaim to node_reclaim
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (20 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 21/27] mm, vmscan: Only wakeup kswapd once per node for the requested classzone Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 10:55 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 23/27] mm, vmscan: Add classzone information to tracepoints Mel Gorman
` (4 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
As reclaim is now per-node based, convert zone_reclaim to be node_reclaim.
It is possible that a node will be reclaimed multiple times if it has
multiple zones but this is unavoidable without caching all nodes traversed
so far. The documentation and interface to userspace is the same from
a configuration perspective and will will be similar in behaviour unless
the node-local allocation requests were also limited to lower zones.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
include/linux/mmzone.h | 18 +++++------
include/linux/swap.h | 9 +++---
include/linux/topology.h | 2 +-
kernel/sysctl.c | 4 +--
mm/huge_memory.c | 4 +--
mm/internal.h | 8 ++---
mm/page_alloc.c | 24 ++++++++++-----
mm/vmscan.c | 77 ++++++++++++++++++++++++------------------------
8 files changed, 77 insertions(+), 69 deletions(-)
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 9cc75f88bb8e..d826d203185e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -366,14 +366,6 @@ struct zone {
unsigned long *pageblock_flags;
#endif /* CONFIG_SPARSEMEM */
-#ifdef CONFIG_NUMA
- /*
- * zone reclaim becomes active if more unmapped pages exist.
- */
- unsigned long min_unmapped_pages;
- unsigned long min_slab_pages;
-#endif /* CONFIG_NUMA */
-
/* zone_start_pfn == zone_start_paddr >> PAGE_SHIFT */
unsigned long zone_start_pfn;
@@ -518,7 +510,6 @@ struct zone {
} ____cacheline_internodealigned_in_smp;
enum zone_flags {
- ZONE_RECLAIM_LOCKED, /* prevents concurrent reclaim */
ZONE_OOM_LOCKED, /* zone is in OOM killer zonelist */
ZONE_FAIR_DEPLETED, /* fair zone policy batch depleted */
};
@@ -534,6 +525,7 @@ enum pgdat_flags {
PGDAT_WRITEBACK, /* reclaim scanning has recently found
* many pages under writeback
*/
+ PGDAT_RECLAIM_LOCKED, /* prevents concurrent reclaim */
};
static inline unsigned long zone_end_pfn(const struct zone *zone)
@@ -682,6 +674,14 @@ typedef struct pglist_data {
*/
unsigned long totalreserve_pages;
+#ifdef CONFIG_NUMA
+ /*
+ * zone reclaim becomes active if more unmapped pages exist.
+ */
+ unsigned long min_unmapped_pages;
+ unsigned long min_slab_pages;
+#endif /* CONFIG_NUMA */
+
/* Write-intensive fields used from the page allocator */
ZONE_PADDING(_pad1_)
spinlock_t lru_lock;
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 5d1eac6259d2..bae1615e698d 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -334,13 +334,14 @@ extern int remove_mapping(struct address_space *mapping, struct page *page);
extern unsigned long vm_total_pages;
#ifdef CONFIG_NUMA
-extern int zone_reclaim_mode;
+extern int node_reclaim_mode;
extern int sysctl_min_unmapped_ratio;
extern int sysctl_min_slab_ratio;
-extern int zone_reclaim(struct zone *, gfp_t, unsigned int);
+extern int node_reclaim(struct pglist_data *, gfp_t, unsigned int);
#else
-#define zone_reclaim_mode 0
-static inline int zone_reclaim(struct zone *z, gfp_t mask, unsigned int order)
+#define node_reclaim_mode 0
+static inline int node_reclaim(struct pglist_data *pgdat, gfp_t mask,
+ unsigned int order)
{
return 0;
}
diff --git a/include/linux/topology.h b/include/linux/topology.h
index afce69296ac0..cb0775e1ee4b 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -54,7 +54,7 @@ int arch_update_cpu_topology(void);
/*
* If the distance between nodes in a system is larger than RECLAIM_DISTANCE
* (in whatever arch specific measurement units returned by node_distance())
- * and zone_reclaim_mode is enabled then the VM will only call zone_reclaim()
+ * and node_reclaim_mode is enabled then the VM will only call node_reclaim()
* on nodes within this distance.
*/
#define RECLAIM_DISTANCE 30
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 35f0dcb1cb4f..53954631a4e1 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -1508,8 +1508,8 @@ static struct ctl_table vm_table[] = {
#ifdef CONFIG_NUMA
{
.procname = "zone_reclaim_mode",
- .data = &zone_reclaim_mode,
- .maxlen = sizeof(zone_reclaim_mode),
+ .data = &node_reclaim_mode,
+ .maxlen = sizeof(node_reclaim_mode),
.mode = 0644,
.proc_handler = proc_dointvec,
.extra1 = &zero,
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index c368db5cadca..0512b863a441 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2227,10 +2227,10 @@ static bool khugepaged_scan_abort(int nid)
int i;
/*
- * If zone_reclaim_mode is disabled, then no extra effort is made to
+ * If node_reclaim_mode is disabled, then no extra effort is made to
* allocate memory locally.
*/
- if (!zone_reclaim_mode)
+ if (!node_reclaim_mode)
return false;
/* If there is a count for this node already, it must be acceptable */
diff --git a/mm/internal.h b/mm/internal.h
index d5e4e7db141d..5231344a9e52 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -434,10 +434,10 @@ static inline void mminit_validate_memmodel_limits(unsigned long *start_pfn,
}
#endif /* CONFIG_SPARSEMEM */
-#define ZONE_RECLAIM_NOSCAN -2
-#define ZONE_RECLAIM_FULL -1
-#define ZONE_RECLAIM_SOME 0
-#define ZONE_RECLAIM_SUCCESS 1
+#define NODE_RECLAIM_NOSCAN -2
+#define NODE_RECLAIM_FULL -1
+#define NODE_RECLAIM_SOME 0
+#define NODE_RECLAIM_SUCCESS 1
extern int hwpoison_filter(struct page *p);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 8faeb0cb21f0..a249128999a8 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2947,16 +2947,16 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
if (alloc_flags & ALLOC_NO_WATERMARKS)
goto try_this_zone;
- if (zone_reclaim_mode == 0 ||
+ if (node_reclaim_mode == 0 ||
!zone_allows_reclaim(ac->preferred_zoneref->zone, zone))
continue;
- ret = zone_reclaim(zone, gfp_mask, order);
+ ret = node_reclaim(zone->zone_pgdat, gfp_mask, order);
switch (ret) {
- case ZONE_RECLAIM_NOSCAN:
+ case NODE_RECLAIM_NOSCAN:
/* did not scan */
continue;
- case ZONE_RECLAIM_FULL:
+ case NODE_RECLAIM_FULL:
/* scanned but unreclaimable */
continue;
default:
@@ -5948,9 +5948,9 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
zone->managed_pages = is_highmem_idx(j) ? realsize : freesize;
#ifdef CONFIG_NUMA
zone->node = nid;
- zone->min_unmapped_pages = (freesize*sysctl_min_unmapped_ratio)
+ pgdat->min_unmapped_pages += (freesize*sysctl_min_unmapped_ratio)
/ 100;
- zone->min_slab_pages = (freesize * sysctl_min_slab_ratio) / 100;
+ pgdat->min_slab_pages += (freesize * sysctl_min_slab_ratio) / 100;
#endif
zone->name = zone_names[j];
zone->zone_pgdat = pgdat;
@@ -6928,6 +6928,7 @@ int watermark_scale_factor_sysctl_handler(struct ctl_table *table, int write,
int sysctl_min_unmapped_ratio_sysctl_handler(struct ctl_table *table, int write,
void __user *buffer, size_t *length, loff_t *ppos)
{
+ struct pglist_data *pgdat;
struct zone *zone;
int rc;
@@ -6935,8 +6936,11 @@ int sysctl_min_unmapped_ratio_sysctl_handler(struct ctl_table *table, int write,
if (rc)
return rc;
+ for_each_online_pgdat(pgdat)
+ pgdat->min_slab_pages = 0;
+
for_each_zone(zone)
- zone->min_unmapped_pages = (zone->managed_pages *
+ zone->zone_pgdat->min_unmapped_pages += (zone->managed_pages *
sysctl_min_unmapped_ratio) / 100;
return 0;
}
@@ -6944,6 +6948,7 @@ int sysctl_min_unmapped_ratio_sysctl_handler(struct ctl_table *table, int write,
int sysctl_min_slab_ratio_sysctl_handler(struct ctl_table *table, int write,
void __user *buffer, size_t *length, loff_t *ppos)
{
+ struct pglist_data *pgdat;
struct zone *zone;
int rc;
@@ -6951,8 +6956,11 @@ int sysctl_min_slab_ratio_sysctl_handler(struct ctl_table *table, int write,
if (rc)
return rc;
+ for_each_online_pgdat(pgdat)
+ pgdat->min_slab_pages = 0;
+
for_each_zone(zone)
- zone->min_slab_pages = (zone->managed_pages *
+ zone->zone_pgdat->min_slab_pages += (zone->managed_pages *
sysctl_min_slab_ratio) / 100;
return 0;
}
diff --git a/mm/vmscan.c b/mm/vmscan.c
index b9cff9047ac0..4e6bb656afb3 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -3534,12 +3534,12 @@ module_init(kswapd_init)
#ifdef CONFIG_NUMA
/*
- * Zone reclaim mode
+ * Node reclaim mode
*
- * If non-zero call zone_reclaim when the number of free pages falls below
+ * If non-zero call node_reclaim when the number of free pages falls below
* the watermarks.
*/
-int zone_reclaim_mode __read_mostly;
+int node_reclaim_mode __read_mostly;
#define RECLAIM_OFF 0
#define RECLAIM_ZONE (1<<0) /* Run shrink_inactive_list on the zone */
@@ -3547,14 +3547,14 @@ int zone_reclaim_mode __read_mostly;
#define RECLAIM_UNMAP (1<<2) /* Unmap pages during reclaim */
/*
- * Priority for ZONE_RECLAIM. This determines the fraction of pages
+ * Priority for NODE_RECLAIM. This determines the fraction of pages
* of a node considered for each zone_reclaim. 4 scans 1/16th of
* a zone.
*/
-#define ZONE_RECLAIM_PRIORITY 4
+#define NODE_RECLAIM_PRIORITY 4
/*
- * Percentage of pages in a zone that must be unmapped for zone_reclaim to
+ * Percentage of pages in a zone that must be unmapped for node_reclaim to
* occur.
*/
int sysctl_min_unmapped_ratio = 1;
@@ -3580,7 +3580,7 @@ static inline unsigned long node_unmapped_file_pages(struct pglist_data *pgdat)
}
/* Work out how many page cache pages we can reclaim in this reclaim_mode */
-static unsigned long zone_pagecache_reclaimable(struct zone *zone)
+static unsigned long zone_pagecache_reclaimable(struct pglist_data *pgdat)
{
unsigned long nr_pagecache_reclaimable;
unsigned long delta = 0;
@@ -3591,14 +3591,14 @@ static unsigned long zone_pagecache_reclaimable(struct zone *zone)
* pages like swapcache and node_unmapped_file_pages() provides
* a better estimate
*/
- if (zone_reclaim_mode & RECLAIM_UNMAP)
- nr_pagecache_reclaimable = node_page_state(zone->zone_pgdat, NR_FILE_PAGES);
+ if (node_reclaim_mode & RECLAIM_UNMAP)
+ nr_pagecache_reclaimable = node_page_state(pgdat, NR_FILE_PAGES);
else
- nr_pagecache_reclaimable = node_unmapped_file_pages(zone->zone_pgdat);
+ nr_pagecache_reclaimable = node_unmapped_file_pages(pgdat);
/* If we can't clean pages, remove dirty pages from consideration */
- if (!(zone_reclaim_mode & RECLAIM_WRITE))
- delta += node_page_state(zone->zone_pgdat, NR_FILE_DIRTY);
+ if (!(node_reclaim_mode & RECLAIM_WRITE))
+ delta += node_page_state(pgdat, NR_FILE_DIRTY);
/* Watch for any possible underflows due to delta */
if (unlikely(delta > nr_pagecache_reclaimable))
@@ -3608,23 +3608,24 @@ static unsigned long zone_pagecache_reclaimable(struct zone *zone)
}
/*
- * Try to free up some pages from this zone through reclaim.
+ * Try to free up some pages from this node through reclaim.
*/
-static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
+static int __node_reclaim(struct pglist_data *pgdat, gfp_t gfp_mask, unsigned int order)
{
/* Minimum pages needed in order to stay on node */
const unsigned long nr_pages = 1 << order;
struct task_struct *p = current;
struct reclaim_state reclaim_state;
+ int classzone_idx = gfp_zone(gfp_mask);
struct scan_control sc = {
.nr_to_reclaim = max(nr_pages, SWAP_CLUSTER_MAX),
.gfp_mask = (gfp_mask = memalloc_noio_flags(gfp_mask)),
.order = order,
- .priority = ZONE_RECLAIM_PRIORITY,
- .may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE),
- .may_unmap = !!(zone_reclaim_mode & RECLAIM_UNMAP),
+ .priority = NODE_RECLAIM_PRIORITY,
+ .may_writepage = !!(node_reclaim_mode & RECLAIM_WRITE),
+ .may_unmap = !!(node_reclaim_mode & RECLAIM_UNMAP),
.may_swap = 1,
- .reclaim_idx = zone_idx(zone),
+ .reclaim_idx = classzone_idx,
};
cond_resched();
@@ -3638,13 +3639,13 @@ static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
reclaim_state.reclaimed_slab = 0;
p->reclaim_state = &reclaim_state;
- if (zone_pagecache_reclaimable(zone) > zone->min_unmapped_pages) {
+ if (zone_pagecache_reclaimable(pgdat) > pgdat->min_unmapped_pages) {
/*
* Free memory by calling shrink zone with increasing
* priorities until we have enough memory freed.
*/
do {
- shrink_node(zone->zone_pgdat, &sc, zone_idx(zone));
+ shrink_node(pgdat, &sc, classzone_idx);
} while (sc.nr_reclaimed < nr_pages && --sc.priority >= 0);
}
@@ -3654,49 +3655,47 @@ static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
return sc.nr_reclaimed >= nr_pages;
}
-int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
+int node_reclaim(struct pglist_data *pgdat, gfp_t gfp_mask, unsigned int order)
{
- int node_id;
int ret;
/*
- * Zone reclaim reclaims unmapped file backed pages and
+ * Node reclaim reclaims unmapped file backed pages and
* slab pages if we are over the defined limits.
*
* A small portion of unmapped file backed pages is needed for
* file I/O otherwise pages read by file I/O will be immediately
- * thrown out if the zone is overallocated. So we do not reclaim
- * if less than a specified percentage of the zone is used by
+ * thrown out if the node is overallocated. So we do not reclaim
+ * if less than a specified percentage of the node is used by
* unmapped file backed pages.
*/
- if (zone_pagecache_reclaimable(zone) <= zone->min_unmapped_pages &&
- zone_page_state(zone, NR_SLAB_RECLAIMABLE) <= zone->min_slab_pages)
- return ZONE_RECLAIM_FULL;
+ if (zone_pagecache_reclaimable(pgdat) <= pgdat->min_unmapped_pages &&
+ sum_zone_node_page_state(pgdat->node_id, NR_SLAB_RECLAIMABLE) <= pgdat->min_slab_pages)
+ return NODE_RECLAIM_FULL;
- if (!pgdat_reclaimable(zone->zone_pgdat))
- return ZONE_RECLAIM_FULL;
+ if (!pgdat_reclaimable(pgdat))
+ return NODE_RECLAIM_FULL;
/*
* Do not scan if the allocation should not be delayed.
*/
if (!gfpflags_allow_blocking(gfp_mask) || (current->flags & PF_MEMALLOC))
- return ZONE_RECLAIM_NOSCAN;
+ return NODE_RECLAIM_NOSCAN;
/*
- * Only run zone reclaim on the local zone or on zones that do not
+ * Only run node reclaim on the local node or on nodes that do not
* have associated processors. This will favor the local processor
* over remote processors and spread off node memory allocations
* as wide as possible.
*/
- node_id = zone_to_nid(zone);
- if (node_state(node_id, N_CPU) && node_id != numa_node_id())
- return ZONE_RECLAIM_NOSCAN;
+ if (node_state(pgdat->node_id, N_CPU) && pgdat->node_id != numa_node_id())
+ return NODE_RECLAIM_NOSCAN;
- if (test_and_set_bit(ZONE_RECLAIM_LOCKED, &zone->flags))
- return ZONE_RECLAIM_NOSCAN;
+ if (test_and_set_bit(PGDAT_RECLAIM_LOCKED, &pgdat->flags))
+ return NODE_RECLAIM_NOSCAN;
- ret = __zone_reclaim(zone, gfp_mask, order);
- clear_bit(ZONE_RECLAIM_LOCKED, &zone->flags);
+ ret = __node_reclaim(pgdat, gfp_mask, order);
+ clear_bit(PGDAT_RECLAIM_LOCKED, &pgdat->flags);
if (!ret)
count_vm_event(PGSCAN_ZONE_RECLAIM_FAILED);
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 22/27] mm: Convert zone_reclaim to node_reclaim
2016-06-09 18:04 ` [PATCH 22/27] mm: Convert zone_reclaim to node_reclaim Mel Gorman
@ 2016-06-17 10:55 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 10:55 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> As reclaim is now per-node based, convert zone_reclaim to be node_reclaim.
> It is possible that a node will be reclaimed multiple times if it has
> multiple zones but this is unavoidable without caching all nodes traversed
> so far. The documentation and interface to userspace is the same from
> a configuration perspective and will will be similar in behaviour unless
> the node-local allocation requests were also limited to lower zones.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
[...]
> @@ -682,6 +674,14 @@ typedef struct pglist_data {
> */
> unsigned long totalreserve_pages;
>
> +#ifdef CONFIG_NUMA
> + /*
> + * zone reclaim becomes active if more unmapped pages exist.
node reclaim
> + */
> + unsigned long min_unmapped_pages;
> + unsigned long min_slab_pages;
> +#endif /* CONFIG_NUMA */
> +
> /* Write-intensive fields used from the page allocator */
> ZONE_PADDING(_pad1_)
> spinlock_t lru_lock;
[...]
> @@ -3580,7 +3580,7 @@ static inline unsigned long node_unmapped_file_pages(struct pglist_data *pgdat)
> }
>
> /* Work out how many page cache pages we can reclaim in this reclaim_mode */
> -static unsigned long zone_pagecache_reclaimable(struct zone *zone)
> +static unsigned long zone_pagecache_reclaimable(struct pglist_data *pgdat)
Rename to node_pagecache_reclaimable?
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 23/27] mm, vmscan: Add classzone information to tracepoints
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (21 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 22/27] mm: Convert zone_reclaim to node_reclaim Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 10:57 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 24/27] mm, page_alloc: Remove fair zone allocation policy Mel Gorman
` (3 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
This is convenient when tracking down why the skip count is high because it'll
show what classzone kswapd woke up at and what zones are being isolated.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
include/trace/events/vmscan.h | 28 ++++++++++++++++++----------
mm/vmscan.c | 4 ++--
2 files changed, 20 insertions(+), 12 deletions(-)
diff --git a/include/trace/events/vmscan.h b/include/trace/events/vmscan.h
index 897f1aa1ee5f..3d242fb8910a 100644
--- a/include/trace/events/vmscan.h
+++ b/include/trace/events/vmscan.h
@@ -55,21 +55,23 @@ TRACE_EVENT(mm_vmscan_kswapd_sleep,
TRACE_EVENT(mm_vmscan_kswapd_wake,
- TP_PROTO(int nid, int order),
+ TP_PROTO(int nid, int zid, int order),
- TP_ARGS(nid, order),
+ TP_ARGS(nid, zid, order),
TP_STRUCT__entry(
__field( int, nid )
+ __field( int, zid )
__field( int, order )
),
TP_fast_assign(
__entry->nid = nid;
+ __entry->zid = zid;
__entry->order = order;
),
- TP_printk("nid=%d order=%d", __entry->nid, __entry->order)
+ TP_printk("nid=%d zid=%d order=%d", __entry->nid, __entry->zid, __entry->order)
);
TRACE_EVENT(mm_vmscan_wakeup_kswapd,
@@ -266,16 +268,18 @@ TRACE_EVENT(mm_shrink_slab_end,
DECLARE_EVENT_CLASS(mm_vmscan_lru_isolate_template,
- TP_PROTO(int order,
+ TP_PROTO(int classzone_idx,
+ int order,
unsigned long nr_requested,
unsigned long nr_scanned,
unsigned long nr_taken,
isolate_mode_t isolate_mode,
int file),
- TP_ARGS(order, nr_requested, nr_scanned, nr_taken, isolate_mode, file),
+ TP_ARGS(classzone_idx, order, nr_requested, nr_scanned, nr_taken, isolate_mode, file),
TP_STRUCT__entry(
+ __field(int, classzone_idx)
__field(int, order)
__field(unsigned long, nr_requested)
__field(unsigned long, nr_scanned)
@@ -285,6 +289,7 @@ DECLARE_EVENT_CLASS(mm_vmscan_lru_isolate_template,
),
TP_fast_assign(
+ __entry->classzone_idx = classzone_idx;
__entry->order = order;
__entry->nr_requested = nr_requested;
__entry->nr_scanned = nr_scanned;
@@ -293,8 +298,9 @@ DECLARE_EVENT_CLASS(mm_vmscan_lru_isolate_template,
__entry->file = file;
),
- TP_printk("isolate_mode=%d order=%d nr_requested=%lu nr_scanned=%lu nr_taken=%lu file=%d",
+ TP_printk("isolate_mode=%d classzone=%d order=%d nr_requested=%lu nr_scanned=%lu nr_taken=%lu file=%d",
__entry->isolate_mode,
+ __entry->classzone_idx,
__entry->order,
__entry->nr_requested,
__entry->nr_scanned,
@@ -304,27 +310,29 @@ DECLARE_EVENT_CLASS(mm_vmscan_lru_isolate_template,
DEFINE_EVENT(mm_vmscan_lru_isolate_template, mm_vmscan_lru_isolate,
- TP_PROTO(int order,
+ TP_PROTO(int classzone_idx,
+ int order,
unsigned long nr_requested,
unsigned long nr_scanned,
unsigned long nr_taken,
isolate_mode_t isolate_mode,
int file),
- TP_ARGS(order, nr_requested, nr_scanned, nr_taken, isolate_mode, file)
+ TP_ARGS(classzone_idx, order, nr_requested, nr_scanned, nr_taken, isolate_mode, file)
);
DEFINE_EVENT(mm_vmscan_lru_isolate_template, mm_vmscan_memcg_isolate,
- TP_PROTO(int order,
+ TP_PROTO(int classzone_idx,
+ int order,
unsigned long nr_requested,
unsigned long nr_scanned,
unsigned long nr_taken,
isolate_mode_t isolate_mode,
int file),
- TP_ARGS(order, nr_requested, nr_scanned, nr_taken, isolate_mode, file)
+ TP_ARGS(classzone_idx, order, nr_requested, nr_scanned, nr_taken, isolate_mode, file)
);
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 4e6bb656afb3..9e12f5d75c06 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1414,7 +1414,7 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
if (!list_empty(&pages_skipped))
list_splice(&pages_skipped, src);
*nr_scanned = scan;
- trace_mm_vmscan_lru_isolate(sc->order, nr_to_scan, scan,
+ trace_mm_vmscan_lru_isolate(sc->reclaim_idx, sc->order, nr_to_scan, scan,
nr_taken, mode, is_file_lru(lru));
return nr_taken;
}
@@ -3368,7 +3368,7 @@ static int kswapd(void *p)
* Try reclaim the requested order but if that fails
* then try sleeping on the basis of the order reclaimed.
*/
- trace_mm_vmscan_kswapd_wake(pgdat->node_id, order);
+ trace_mm_vmscan_kswapd_wake(pgdat->node_id, classzone_idx, order);
if (balance_pgdat(pgdat, order, classzone_idx) < order)
goto kswapd_try_sleep;
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* [PATCH 24/27] mm, page_alloc: Remove fair zone allocation policy
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (22 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 23/27] mm, vmscan: Add classzone information to tracepoints Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 11:27 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 25/27] mm: page_alloc: Cache the last node whose dirty limit is reached Mel Gorman
` (2 subsequent siblings)
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
The fair zone allocation policy interleaves allocation requests between
zones to avoid an age inversion problem whereby new pages are reclaimed
to balance a zone. Reclaim is now node-based so this should no longer be
an issue and the fair zone allocation policy is not free. This patch
removes it.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
include/linux/mmzone.h | 2 --
mm/internal.h | 1 -
mm/page_alloc.c | 75 +-------------------------------------------------
mm/vmstat.c | 4 +--
4 files changed, 2 insertions(+), 80 deletions(-)
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index d826d203185e..347a13a62d55 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -110,7 +110,6 @@ struct zone_padding {
enum zone_stat_item {
/* First 128 byte cacheline (assuming 64 bit words) */
NR_FREE_PAGES,
- NR_ALLOC_BATCH,
NR_MLOCK, /* mlock()ed pages found and moved off LRU */
NR_SLAB_RECLAIMABLE,
NR_SLAB_UNRECLAIMABLE,
@@ -511,7 +510,6 @@ struct zone {
enum zone_flags {
ZONE_OOM_LOCKED, /* zone is in OOM killer zonelist */
- ZONE_FAIR_DEPLETED, /* fair zone policy batch depleted */
};
enum pgdat_flags {
diff --git a/mm/internal.h b/mm/internal.h
index 5231344a9e52..51505914d7c9 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -468,7 +468,6 @@ unsigned long reclaim_clean_pages_from_list(struct zone *zone,
#define ALLOC_HIGH 0x20 /* __GFP_HIGH set */
#define ALLOC_CPUSET 0x40 /* check for correct cpuset */
#define ALLOC_CMA 0x80 /* allow allocations from CMA areas */
-#define ALLOC_FAIR 0x100 /* fair zone allocation */
enum ttu_flags;
struct tlbflush_unmap_batch;
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index a249128999a8..17fa3f361f01 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2592,7 +2592,6 @@ struct page *buffered_rmqueue(struct zone *preferred_zone,
else
page = list_first_entry(list, struct page, lru);
- __dec_zone_state(zone, NR_ALLOC_BATCH);
list_del(&page->lru);
pcp->count--;
@@ -2618,15 +2617,10 @@ struct page *buffered_rmqueue(struct zone *preferred_zone,
spin_unlock(&zone->lock);
if (!page)
goto failed;
- __mod_zone_page_state(zone, NR_ALLOC_BATCH, -(1 << order));
__mod_zone_freepage_state(zone, -(1 << order),
get_pcppage_migratetype(page));
}
- if (atomic_long_read(&zone->vm_stat[NR_ALLOC_BATCH]) <= 0 &&
- !test_bit(ZONE_FAIR_DEPLETED, &zone->flags))
- set_bit(ZONE_FAIR_DEPLETED, &zone->flags);
-
__count_zone_vm_events(PGALLOC, zone, 1 << order);
zone_statistics(preferred_zone, zone, gfp_flags);
local_irq_restore(flags);
@@ -2837,40 +2831,18 @@ bool zone_watermark_ok_safe(struct zone *z, unsigned int order,
}
#ifdef CONFIG_NUMA
-static bool zone_local(struct zone *local_zone, struct zone *zone)
-{
- return local_zone->node == zone->node;
-}
-
static bool zone_allows_reclaim(struct zone *local_zone, struct zone *zone)
{
return node_distance(zone_to_nid(local_zone), zone_to_nid(zone)) <
RECLAIM_DISTANCE;
}
#else /* CONFIG_NUMA */
-static bool zone_local(struct zone *local_zone, struct zone *zone)
-{
- return true;
-}
-
static bool zone_allows_reclaim(struct zone *local_zone, struct zone *zone)
{
return true;
}
#endif /* CONFIG_NUMA */
-static void reset_alloc_batches(struct zone *preferred_zone)
-{
- struct zone *zone = preferred_zone->zone_pgdat->node_zones;
-
- do {
- mod_zone_page_state(zone, NR_ALLOC_BATCH,
- high_wmark_pages(zone) - low_wmark_pages(zone) -
- atomic_long_read(&zone->vm_stat[NR_ALLOC_BATCH]));
- clear_bit(ZONE_FAIR_DEPLETED, &zone->flags);
- } while (zone++ != preferred_zone);
-}
-
/*
* get_page_from_freelist goes through the zonelist trying to allocate
* a page.
@@ -2881,10 +2853,6 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
{
struct zoneref *z = ac->preferred_zoneref;
struct zone *zone;
- bool fair_skipped = false;
- bool apply_fair = (alloc_flags & ALLOC_FAIR);
-
-zonelist_scan:
/*
* Scan zonelist, looking for a zone with enough free.
* See also __cpuset_node_allowed() comment in kernel/cpuset.c.
@@ -2899,23 +2867,6 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
!__cpuset_zone_allowed(zone, gfp_mask))
continue;
/*
- * Distribute pages in proportion to the individual
- * zone size to ensure fair page aging. The zone a
- * page was allocated in should have no effect on the
- * time the page has in memory before being reclaimed.
- */
- if (apply_fair) {
- if (test_bit(ZONE_FAIR_DEPLETED, &zone->flags)) {
- fair_skipped = true;
- continue;
- }
- if (!zone_local(ac->preferred_zoneref->zone, zone)) {
- if (fair_skipped)
- goto reset_fair;
- apply_fair = false;
- }
- }
- /*
* When allocating a page cache page for writing, we
* want to get it from a node that is within its dirty
* limit, such that no single node holds more than its
@@ -2986,23 +2937,6 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
}
}
- /*
- * The first pass makes sure allocations are spread fairly within the
- * local node. However, the local node might have free pages left
- * after the fairness batches are exhausted, and remote zones haven't
- * even been considered yet. Try once more without fairness, and
- * include remote zones now, before entering the slowpath and waking
- * kswapd: prefer spilling to a remote zone over swapping locally.
- */
- if (fair_skipped) {
-reset_fair:
- apply_fair = false;
- fair_skipped = false;
- reset_alloc_batches(ac->preferred_zoneref->zone);
- z = ac->preferred_zoneref;
- goto zonelist_scan;
- }
-
return NULL;
}
@@ -3758,7 +3692,7 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order,
{
struct page *page;
unsigned int cpuset_mems_cookie;
- unsigned int alloc_flags = ALLOC_WMARK_LOW|ALLOC_FAIR;
+ unsigned int alloc_flags = ALLOC_WMARK_LOW;
gfp_t alloc_mask = gfp_mask; /* The gfp_t that was actually used for allocation */
struct alloc_context ac = {
.high_zoneidx = gfp_zone(gfp_mask),
@@ -5959,9 +5893,6 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat)
zone_seqlock_init(zone);
zone_pcp_init(zone);
- /* For bootup, initialized properly in watermark setup */
- mod_zone_page_state(zone, NR_ALLOC_BATCH, zone->managed_pages);
-
if (!size)
continue;
@@ -6814,10 +6745,6 @@ static void __setup_per_zone_wmarks(void)
zone->watermark[WMARK_LOW] = min_wmark_pages(zone) + tmp;
zone->watermark[WMARK_HIGH] = min_wmark_pages(zone) + tmp * 2;
- __mod_zone_page_state(zone, NR_ALLOC_BATCH,
- high_wmark_pages(zone) - low_wmark_pages(zone) -
- atomic_long_read(&zone->vm_stat[NR_ALLOC_BATCH]));
-
spin_unlock_irqrestore(&zone->lock, flags);
}
diff --git a/mm/vmstat.c b/mm/vmstat.c
index 26e9873be7a2..dd60fa3ca66b 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -920,7 +920,6 @@ int fragmentation_index(struct zone *zone, unsigned int order)
const char * const vmstat_text[] = {
/* enum zone_stat_item countes */
"nr_free_pages",
- "nr_alloc_batch",
"nr_mlock",
"nr_slab_reclaimable",
"nr_slab_unreclaimable",
@@ -1624,10 +1623,9 @@ int vmstat_refresh(struct ctl_table *table, int write,
val = atomic_long_read(&vm_zone_stat[i]);
if (val < 0) {
switch (i) {
- case NR_ALLOC_BATCH:
case NR_PAGES_SCANNED:
/*
- * These are often seen to go negative in
+ * This is often seen to go negative in
* recent kernels, but not to go permanently
* negative. Whilst it would be nicer not to
* have exceptions, rooting them out would be
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 24/27] mm, page_alloc: Remove fair zone allocation policy
2016-06-09 18:04 ` [PATCH 24/27] mm, page_alloc: Remove fair zone allocation policy Mel Gorman
@ 2016-06-17 11:27 ` Vlastimil Babka
2016-06-17 12:07 ` Mel Gorman
0 siblings, 1 reply; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 11:27 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM
Cc: Rik van Riel, Johannes Weiner, LKML, Michal Hocko
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> The fair zone allocation policy interleaves allocation requests between
> zones to avoid an age inversion problem whereby new pages are reclaimed
> to balance a zone. Reclaim is now node-based so this should no longer be
> an issue and the fair zone allocation policy is not free. This patch
> removes it.
I wonder if fair zone allocation had the side effect of preventing e.g. a small
Normal zone to be almost fully occupied by long-lived unreclaimable allocations
early in the kernel lifetime. So that might be one thing to watch out for. But
otherwise I would agree it should be no longer needed with node-based reclaim.
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
^ permalink raw reply [flat|nested] 70+ messages in thread
* Re: [PATCH 24/27] mm, page_alloc: Remove fair zone allocation policy
2016-06-17 11:27 ` Vlastimil Babka
@ 2016-06-17 12:07 ` Mel Gorman
0 siblings, 0 replies; 70+ messages in thread
From: Mel Gorman @ 2016-06-17 12:07 UTC (permalink / raw)
To: Vlastimil Babka
Cc: Andrew Morton, Linux-MM, Rik van Riel, Johannes Weiner, LKML,
Michal Hocko
On Fri, Jun 17, 2016 at 01:27:09PM +0200, Vlastimil Babka wrote:
> On 06/09/2016 08:04 PM, Mel Gorman wrote:
> >The fair zone allocation policy interleaves allocation requests between
> >zones to avoid an age inversion problem whereby new pages are reclaimed
> >to balance a zone. Reclaim is now node-based so this should no longer be
> >an issue and the fair zone allocation policy is not free. This patch
> >removes it.
>
> I wonder if fair zone allocation had the side effect of preventing e.g. a
> small Normal zone to be almost fully occupied by long-lived unreclaimable
> allocations early in the kernel lifetime. So that might be one thing to
> watch out for.
It's a marginal corner case and the zonelist scan is slightly inefficient
as the first zone is always skipped but the impact is light.
> But otherwise I would agree it should be no longer needed
> with node-based reclaim.
>
> >Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
>
> Acked-by: Vlastimil Babka <vbabka@suse.cz>
>
Thanks.
--
Mel Gorman
SUSE Labs
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 25/27] mm: page_alloc: Cache the last node whose dirty limit is reached
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (23 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 24/27] mm, page_alloc: Remove fair zone allocation policy Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 11:31 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 26/27] mm: vmstat: Replace __count_zone_vm_events with a zone id equivalent Mel Gorman
2016-06-09 18:04 ` [PATCH 27/27] mm: vmstat: Account per-zone stalls and pages skipped during reclaim Mel Gorman
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
If a page is about to be dirtied then the page allocator attempts to limit
the total number of dirty pages that exists in any given zone. The call
to node_dirty_ok is expensive so this patch records if the last pgdat
examined hit the dirty limits. In some cases, this reduces the number
of calls to node_dirty_ok().
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/page_alloc.c | 13 +++++++++++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 17fa3f361f01..7c6c18a314a1 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2853,6 +2853,8 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
{
struct zoneref *z = ac->preferred_zoneref;
struct zone *zone;
+ struct pglist_data *last_pgdat_dirty_limit = NULL;
+
/*
* Scan zonelist, looking for a zone with enough free.
* See also __cpuset_node_allowed() comment in kernel/cpuset.c.
@@ -2885,8 +2887,15 @@ get_page_from_freelist(gfp_t gfp_mask, unsigned int order, int alloc_flags,
* will require awareness of nodes in the
* dirty-throttling and the flusher threads.
*/
- if (ac->spread_dirty_pages && !node_dirty_ok(zone->zone_pgdat))
- continue;
+ if (ac->spread_dirty_pages) {
+ if (last_pgdat_dirty_limit == zone->zone_pgdat)
+ continue;
+
+ if (!node_dirty_ok(zone->zone_pgdat)) {
+ last_pgdat_dirty_limit = zone->zone_pgdat;
+ continue;
+ }
+ }
mark = zone->watermark[alloc_flags & ALLOC_WMARK_MASK];
if (!zone_watermark_fast(zone, order, mark,
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 25/27] mm: page_alloc: Cache the last node whose dirty limit is reached
2016-06-09 18:04 ` [PATCH 25/27] mm: page_alloc: Cache the last node whose dirty limit is reached Mel Gorman
@ 2016-06-17 11:31 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 11:31 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> If a page is about to be dirtied then the page allocator attempts to limit
> the total number of dirty pages that exists in any given zone. The call
> to node_dirty_ok is expensive so this patch records if the last pgdat
> examined hit the dirty limits. In some cases, this reduces the number
> of calls to node_dirty_ok().
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 26/27] mm: vmstat: Replace __count_zone_vm_events with a zone id equivalent
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (24 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 25/27] mm: page_alloc: Cache the last node whose dirty limit is reached Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 11:36 ` Vlastimil Babka
2016-06-09 18:04 ` [PATCH 27/27] mm: vmstat: Account per-zone stalls and pages skipped during reclaim Mel Gorman
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
This is partially a preparation patch for more vmstat work but it also
has the slight advantage that __count_zid_vm_events is cheaper to
calculate than __count_zone_vm_events().
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
include/linux/vmstat.h | 5 ++---
mm/page_alloc.c | 2 +-
2 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/include/linux/vmstat.h b/include/linux/vmstat.h
index 533948c93550..2feab717704d 100644
--- a/include/linux/vmstat.h
+++ b/include/linux/vmstat.h
@@ -107,9 +107,8 @@ static inline void vm_events_fold_cpu(int cpu)
#define count_vm_vmacache_event(x) do {} while (0)
#endif
-#define __count_zone_vm_events(item, zone, delta) \
- __count_vm_events(item##_NORMAL - ZONE_NORMAL + \
- zone_idx(zone), delta)
+#define __count_zid_vm_events(item, zid, delta) \
+ __count_vm_events(item##_NORMAL - ZONE_NORMAL + zid, delta)
/*
* Zone and node-based page accounting with per cpu differentials.
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 7c6c18a314a1..028d088633c4 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2621,7 +2621,7 @@ struct page *buffered_rmqueue(struct zone *preferred_zone,
get_pcppage_migratetype(page));
}
- __count_zone_vm_events(PGALLOC, zone, 1 << order);
+ __count_zid_vm_events(PGALLOC, page_zonenum(page), 1 << order);
zone_statistics(preferred_zone, zone, gfp_flags);
local_irq_restore(flags);
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 26/27] mm: vmstat: Replace __count_zone_vm_events with a zone id equivalent
2016-06-09 18:04 ` [PATCH 26/27] mm: vmstat: Replace __count_zone_vm_events with a zone id equivalent Mel Gorman
@ 2016-06-17 11:36 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 11:36 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> This is partially a preparation patch for more vmstat work but it also
> has the slight advantage that __count_zid_vm_events is cheaper to
> calculate than __count_zone_vm_events().
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
> ---
> include/linux/vmstat.h | 5 ++---
> mm/page_alloc.c | 2 +-
> 2 files changed, 3 insertions(+), 4 deletions(-)
>
> diff --git a/include/linux/vmstat.h b/include/linux/vmstat.h
> index 533948c93550..2feab717704d 100644
> --- a/include/linux/vmstat.h
> +++ b/include/linux/vmstat.h
> @@ -107,9 +107,8 @@ static inline void vm_events_fold_cpu(int cpu)
> #define count_vm_vmacache_event(x) do {} while (0)
> #endif
>
> -#define __count_zone_vm_events(item, zone, delta) \
> - __count_vm_events(item##_NORMAL - ZONE_NORMAL + \
> - zone_idx(zone), delta)
> +#define __count_zid_vm_events(item, zid, delta) \
> + __count_vm_events(item##_NORMAL - ZONE_NORMAL + zid, delta)
>
> /*
> * Zone and node-based page accounting with per cpu differentials.
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 7c6c18a314a1..028d088633c4 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -2621,7 +2621,7 @@ struct page *buffered_rmqueue(struct zone *preferred_zone,
> get_pcppage_migratetype(page));
> }
>
> - __count_zone_vm_events(PGALLOC, zone, 1 << order);
> + __count_zid_vm_events(PGALLOC, page_zonenum(page), 1 << order);
> zone_statistics(preferred_zone, zone, gfp_flags);
> local_irq_restore(flags);
>
>
^ permalink raw reply [flat|nested] 70+ messages in thread
* [PATCH 27/27] mm: vmstat: Account per-zone stalls and pages skipped during reclaim
2016-06-09 18:04 [PATCH 00/27] Move LRU page reclaim from zones to nodes v6 Mel Gorman
` (25 preceding siblings ...)
2016-06-09 18:04 ` [PATCH 26/27] mm: vmstat: Replace __count_zone_vm_events with a zone id equivalent Mel Gorman
@ 2016-06-09 18:04 ` Mel Gorman
2016-06-17 12:39 ` Vlastimil Babka
26 siblings, 1 reply; 70+ messages in thread
From: Mel Gorman @ 2016-06-09 18:04 UTC (permalink / raw)
To: Andrew Morton, Linux-MM
Cc: Rik van Riel, Vlastimil Babka, Johannes Weiner, LKML, Mel Gorman
The vmstat allocstall was fairly useful in the general sense but
node-based LRUs change that. It's important to know if a stall was for an
address-limited allocation request as this will require skipping pages from
other zones. This patch adds pgstall_* counters to replace allocstall. The
sum of the counters will equal the old allocstall so it can be trivially
recalculated. A high number of address-limited allocation requests may
result in a lot of useless LRU scanning for suitable pages.
As address-limited allocations require pages to be skipped, it's important
to know how much useless LRU scanning took place so this patch adds
pgskip* counters. This yields the following model
1. The number of address-space limited stalls can be accounted for (pgstall)
2. The amount of useless work required to reclaim the data is accounted (pgskip)
3. The total number of scans is available from pgscan_kswapd and pgscan_direct
so from that the ratio of useful to useless scans can be calculated.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
include/linux/vm_event_item.h | 4 +++-
mm/huge_memory.c | 19 +++++++++++++++----
mm/vmscan.c | 15 +++++++++++++--
mm/vmstat.c | 3 ++-
4 files changed, 33 insertions(+), 8 deletions(-)
diff --git a/include/linux/vm_event_item.h b/include/linux/vm_event_item.h
index 8dcb5a813163..0a0503da8c3b 100644
--- a/include/linux/vm_event_item.h
+++ b/include/linux/vm_event_item.h
@@ -23,6 +23,8 @@
enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT,
FOR_ALL_ZONES(PGALLOC),
+ FOR_ALL_ZONES(PGSTALL),
+ FOR_ALL_ZONES(PGSCAN_SKIP),
PGFREE, PGACTIVATE, PGDEACTIVATE,
PGFAULT, PGMAJFAULT,
PGLAZYFREED,
@@ -37,7 +39,7 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT,
#endif
PGINODESTEAL, SLABS_SCANNED, KSWAPD_INODESTEAL,
KSWAPD_LOW_WMARK_HIT_QUICKLY, KSWAPD_HIGH_WMARK_HIT_QUICKLY,
- PAGEOUTRUN, ALLOCSTALL, PGROTATED,
+ PAGEOUTRUN, PGROTATED,
DROP_PAGECACHE, DROP_SLAB,
#ifdef CONFIG_NUMA_BALANCING
NUMA_PTE_UPDATES,
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 0512b863a441..e554ca7d095e 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -2459,6 +2459,17 @@ static bool __collapse_huge_page_swapin(struct mm_struct *mm,
return true;
}
+static unsigned long sum_alloc_stalls(void)
+{
+ int zid;
+ unsigned long allocstall = 0;
+
+ for (zid = 0; zid < MAX_NR_ZONES - 1; zid++)
+ allocstall += sum_vm_event(PGSTALL_NORMAL - ZONE_NORMAL + zid);
+
+ return allocstall;
+}
+
static void collapse_huge_page(struct mm_struct *mm,
unsigned long address,
struct page **hpage,
@@ -2495,7 +2506,7 @@ static void collapse_huge_page(struct mm_struct *mm,
}
swap = get_mm_counter(mm, MM_SWAPENTS);
- curr_allocstall = sum_vm_event(ALLOCSTALL);
+ curr_allocstall = sum_alloc_stalls();;
down_read(&mm->mmap_sem);
result = hugepage_vma_revalidate(mm, address);
if (result) {
@@ -2929,7 +2940,7 @@ static void khugepaged_wait_work(void)
if (!scan_sleep_jiffies)
return;
- allocstall = sum_vm_event(ALLOCSTALL);
+ allocstall = sum_alloc_stalls();
khugepaged_sleep_expire = jiffies + scan_sleep_jiffies;
wait_event_freezable_timeout(khugepaged_wait,
khugepaged_should_wakeup(),
@@ -2938,7 +2949,7 @@ static void khugepaged_wait_work(void)
}
if (khugepaged_enabled()) {
- allocstall = sum_vm_event(ALLOCSTALL);
+ allocstall = sum_alloc_stalls();
wait_event_freezable(khugepaged_wait, khugepaged_wait_event());
}
}
@@ -2949,7 +2960,7 @@ static int khugepaged(void *none)
set_freezable();
set_user_nice(current, MAX_NICE);
- allocstall = sum_vm_event(ALLOCSTALL);
+ allocstall = sum_alloc_stalls();
while (!kthread_should_stop()) {
khugepaged_do_scan();
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 9e12f5d75c06..ce7f4e54ae26 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1372,6 +1372,7 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
struct list_head *src = &lruvec->lists[lru];
unsigned long nr_taken = 0;
unsigned long scan;
+ unsigned long nr_skipped[MAX_NR_ZONES] = { 0, };
LIST_HEAD(pages_skipped);
for (scan = 0; scan < nr_to_scan && nr_taken < nr_to_scan &&
@@ -1385,6 +1386,7 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
if (page_zonenum(page) > sc->reclaim_idx) {
list_move(&page->lru, &pages_skipped);
+ nr_skipped[page_zonenum(page)]++;
continue;
}
@@ -1411,8 +1413,17 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
* scanning would soon rescan the same pages to skip and put the
* system at risk of premature OOM.
*/
- if (!list_empty(&pages_skipped))
+ if (!list_empty(&pages_skipped)) {
+ int zid;
+
list_splice(&pages_skipped, src);
+ for (zid = 0; zid < MAX_NR_ZONES; zid++) {
+ if (!nr_skipped[zid])
+ continue;
+
+ __count_zid_vm_events(PGSCAN_SKIP, zid, nr_skipped[zid]);
+ }
+ }
*nr_scanned = scan;
trace_mm_vmscan_lru_isolate(sc->reclaim_idx, sc->order, nr_to_scan, scan,
nr_taken, mode, is_file_lru(lru));
@@ -2655,7 +2666,7 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist,
delayacct_freepages_start();
if (global_reclaim(sc))
- count_vm_event(ALLOCSTALL);
+ __count_zid_vm_events(PGSTALL, classzone_idx, 1);
do {
vmpressure_prio(sc->gfp_mask, sc->target_mem_cgroup,
diff --git a/mm/vmstat.c b/mm/vmstat.c
index dd60fa3ca66b..2de51e9d7e73 100644
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -977,6 +977,8 @@ const char * const vmstat_text[] = {
"pswpout",
TEXTS_FOR_ZONES("pgalloc")
+ TEXTS_FOR_ZONES("pgstall")
+ TEXTS_FOR_ZONES("pgskip")
"pgfree",
"pgactivate",
@@ -1002,7 +1004,6 @@ const char * const vmstat_text[] = {
"kswapd_low_wmark_hit_quickly",
"kswapd_high_wmark_hit_quickly",
"pageoutrun",
- "allocstall",
"pgrotated",
--
2.6.4
^ permalink raw reply related [flat|nested] 70+ messages in thread* Re: [PATCH 27/27] mm: vmstat: Account per-zone stalls and pages skipped during reclaim
2016-06-09 18:04 ` [PATCH 27/27] mm: vmstat: Account per-zone stalls and pages skipped during reclaim Mel Gorman
@ 2016-06-17 12:39 ` Vlastimil Babka
0 siblings, 0 replies; 70+ messages in thread
From: Vlastimil Babka @ 2016-06-17 12:39 UTC (permalink / raw)
To: Mel Gorman, Andrew Morton, Linux-MM; +Cc: Rik van Riel, Johannes Weiner, LKML
On 06/09/2016 08:04 PM, Mel Gorman wrote:
> The vmstat allocstall was fairly useful in the general sense but
> node-based LRUs change that. It's important to know if a stall was for an
> address-limited allocation request as this will require skipping pages from
> other zones. This patch adds pgstall_* counters to replace allocstall. The
> sum of the counters will equal the old allocstall so it can be trivially
> recalculated. A high number of address-limited allocation requests may
> result in a lot of useless LRU scanning for suitable pages.
>
> As address-limited allocations require pages to be skipped, it's important
> to know how much useless LRU scanning took place so this patch adds
> pgskip* counters. This yields the following model
>
> 1. The number of address-space limited stalls can be accounted for (pgstall)
> 2. The amount of useless work required to reclaim the data is accounted (pgskip)
> 3. The total number of scans is available from pgscan_kswapd and pgscan_direct
> so from that the ratio of useful to useless scans can be calculated.
>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Nice work!
^ permalink raw reply [flat|nested] 70+ messages in thread