Commit c772be93 authored by KOSAKI Motohiro's avatar KOSAKI Motohiro Committed by Linus Torvalds

memcg: fix calculation of active_ratio

Currently, inactive_ratio of memcg is calculated at setting limit.
because page_alloc.c does so and current implementation is straightforward
porting.

However, memcg introduced hierarchy feature recently.  In hierarchy
restriction, memory limit is not only decided memory.limit_in_bytes of
current cgroup, but also parent limit and sibling memory usage.

Then, The optimal inactive_ratio is changed frequently.  So, everytime
calculation is better.
Tested-by: default avatarKAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: default avatarKAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: default avatarKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent a7885eb8
...@@ -97,8 +97,7 @@ extern void mem_cgroup_note_reclaim_priority(struct mem_cgroup *mem, ...@@ -97,8 +97,7 @@ extern void mem_cgroup_note_reclaim_priority(struct mem_cgroup *mem,
int priority); int priority);
extern void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, extern void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem,
int priority); int priority);
int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg, int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg);
struct zone *zone);
unsigned long mem_cgroup_zone_nr_pages(struct mem_cgroup *memcg, unsigned long mem_cgroup_zone_nr_pages(struct mem_cgroup *memcg,
struct zone *zone, struct zone *zone,
enum lru_list lru); enum lru_list lru);
...@@ -252,7 +251,7 @@ static inline bool mem_cgroup_oom_called(struct task_struct *task) ...@@ -252,7 +251,7 @@ static inline bool mem_cgroup_oom_called(struct task_struct *task)
} }
static inline int static inline int
mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg, struct zone *zone) mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg)
{ {
return 1; return 1;
} }
......
...@@ -166,9 +166,6 @@ struct mem_cgroup { ...@@ -166,9 +166,6 @@ struct mem_cgroup {
unsigned int swappiness; unsigned int swappiness;
unsigned int inactive_ratio;
/* /*
* statistics. This must be placed at the end of memcg. * statistics. This must be placed at the end of memcg.
*/ */
...@@ -432,15 +429,43 @@ void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, int priority) ...@@ -432,15 +429,43 @@ void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, int priority)
spin_unlock(&mem->reclaim_param_lock); spin_unlock(&mem->reclaim_param_lock);
} }
int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg, struct zone *zone) static int calc_inactive_ratio(struct mem_cgroup *memcg, unsigned long *present_pages)
{ {
unsigned long active; unsigned long active;
unsigned long inactive; unsigned long inactive;
unsigned long gb;
unsigned long inactive_ratio;
inactive = mem_cgroup_get_all_zonestat(memcg, LRU_INACTIVE_ANON); inactive = mem_cgroup_get_all_zonestat(memcg, LRU_INACTIVE_ANON);
active = mem_cgroup_get_all_zonestat(memcg, LRU_ACTIVE_ANON); active = mem_cgroup_get_all_zonestat(memcg, LRU_ACTIVE_ANON);
if (inactive * memcg->inactive_ratio < active) gb = (inactive + active) >> (30 - PAGE_SHIFT);
if (gb)
inactive_ratio = int_sqrt(10 * gb);
else
inactive_ratio = 1;
if (present_pages) {
present_pages[0] = inactive;
present_pages[1] = active;
}
return inactive_ratio;
}
int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg)
{
unsigned long active;
unsigned long inactive;
unsigned long present_pages[2];
unsigned long inactive_ratio;
inactive_ratio = calc_inactive_ratio(memcg, present_pages);
inactive = present_pages[0];
active = present_pages[1];
if (inactive * inactive_ratio < active)
return 1; return 1;
return 0; return 0;
...@@ -1432,29 +1457,6 @@ int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask) ...@@ -1432,29 +1457,6 @@ int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask)
return 0; return 0;
} }
/*
* The inactive anon list should be small enough that the VM never has to
* do too much work, but large enough that each inactive page has a chance
* to be referenced again before it is swapped out.
*
* this calculation is straightforward porting from
* page_alloc.c::setup_per_zone_inactive_ratio().
* it describe more detail.
*/
static void mem_cgroup_set_inactive_ratio(struct mem_cgroup *memcg)
{
unsigned int gb, ratio;
gb = res_counter_read_u64(&memcg->res, RES_LIMIT) >> 30;
if (gb)
ratio = int_sqrt(10 * gb);
else
ratio = 1;
memcg->inactive_ratio = ratio;
}
static DEFINE_MUTEX(set_limit_mutex); static DEFINE_MUTEX(set_limit_mutex);
static int mem_cgroup_resize_limit(struct mem_cgroup *memcg, static int mem_cgroup_resize_limit(struct mem_cgroup *memcg,
...@@ -1496,9 +1498,6 @@ static int mem_cgroup_resize_limit(struct mem_cgroup *memcg, ...@@ -1496,9 +1498,6 @@ static int mem_cgroup_resize_limit(struct mem_cgroup *memcg,
if (!progress) retry_count--; if (!progress) retry_count--;
} }
if (!ret)
mem_cgroup_set_inactive_ratio(memcg);
return ret; return ret;
} }
...@@ -1858,7 +1857,7 @@ static int mem_control_stat_show(struct cgroup *cont, struct cftype *cft, ...@@ -1858,7 +1857,7 @@ static int mem_control_stat_show(struct cgroup *cont, struct cftype *cft,
} }
#ifdef CONFIG_DEBUG_VM #ifdef CONFIG_DEBUG_VM
cb->fill(cb, "inactive_ratio", mem_cont->inactive_ratio); cb->fill(cb, "inactive_ratio", calc_inactive_ratio(mem_cont, NULL));
{ {
int nid, zid; int nid, zid;
...@@ -2150,7 +2149,6 @@ mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont) ...@@ -2150,7 +2149,6 @@ mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont)
res_counter_init(&mem->res, NULL); res_counter_init(&mem->res, NULL);
res_counter_init(&mem->memsw, NULL); res_counter_init(&mem->memsw, NULL);
} }
mem_cgroup_set_inactive_ratio(mem);
mem->last_scanned_child = NULL; mem->last_scanned_child = NULL;
spin_lock_init(&mem->reclaim_param_lock); spin_lock_init(&mem->reclaim_param_lock);
......
...@@ -1340,7 +1340,7 @@ static int inactive_anon_is_low(struct zone *zone, struct scan_control *sc) ...@@ -1340,7 +1340,7 @@ static int inactive_anon_is_low(struct zone *zone, struct scan_control *sc)
if (scanning_global_lru(sc)) if (scanning_global_lru(sc))
low = inactive_anon_is_low_global(zone); low = inactive_anon_is_low_global(zone);
else else
low = mem_cgroup_inactive_anon_is_low(sc->mem_cgroup, zone); low = mem_cgroup_inactive_anon_is_low(sc->mem_cgroup);
return low; return low;
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment