summaryrefslogtreecommitdiff
path: root/mm/memcontrol.c
diff options
context:
space:
mode:
authorBalbir Singh <balbir@linux.vnet.ibm.com>2008-02-07 00:13:56 -0800
committerLinus Torvalds <torvalds@woody.linux-foundation.org>2008-02-07 08:42:18 -0800
commit66e1707bc34609f626e2e7b4fe7e454c9748bad5 (patch)
treed850a729887485874c976ba64eb85e3406e488a1 /mm/memcontrol.c
parent67e465a77ba658635309ee00b367bec6555ea544 (diff)
Memory controller: add per cgroup LRU and reclaim
Add the page_cgroup to the per cgroup LRU. The reclaim algorithm has been modified to make the isolate_lru_pages() as a pluggable component. The scan_control data structure now accepts the cgroup on behalf of which reclaims are carried out. try_to_free_pages() has been extended to become cgroup aware. [akpm@linux-foundation.org: fix warning] [Lee.Schermerhorn@hp.com: initialize all scan_control's isolate_pages member] [bunk@kernel.org: make do_try_to_free_pages() static] [hugh@veritas.com: memcgroup: fix try_to_free order] [kamezawa.hiroyu@jp.fujitsu.com: this unlock_page_cgroup() is unnecessary] Signed-off-by: Pavel Emelianov <xemul@openvz.org> Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Paul Menage <menage@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Kirill Korotaev <dev@sw.ru> Cc: Herbert Poetzl <herbert@13thfloor.at> Cc: David Rientjes <rientjes@google.com> Cc: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'mm/memcontrol.c')
-rw-r--r--mm/memcontrol.c148
1 files changed, 143 insertions, 5 deletions
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index b25df2a9d024..9e9ff914c0f1 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -22,10 +22,15 @@
#include <linux/cgroup.h>
#include <linux/mm.h>
#include <linux/page-flags.h>
+#include <linux/backing-dev.h>
#include <linux/bit_spinlock.h>
#include <linux/rcupdate.h>
+#include <linux/swap.h>
+#include <linux/spinlock.h>
+#include <linux/fs.h>
struct cgroup_subsys mem_cgroup_subsys;
+static const int MEM_CGROUP_RECLAIM_RETRIES = 5;
/*
* The memory controller data structure. The memory controller controls both
@@ -51,6 +56,10 @@ struct mem_cgroup {
*/
struct list_head active_list;
struct list_head inactive_list;
+ /*
+ * spin_lock to protect the per cgroup LRU
+ */
+ spinlock_t lru_lock;
};
/*
@@ -141,6 +150,94 @@ void __always_inline unlock_page_cgroup(struct page *page)
bit_spin_unlock(PAGE_CGROUP_LOCK_BIT, &page->page_cgroup);
}
+void __mem_cgroup_move_lists(struct page_cgroup *pc, bool active)
+{
+ if (active)
+ list_move(&pc->lru, &pc->mem_cgroup->active_list);
+ else
+ list_move(&pc->lru, &pc->mem_cgroup->inactive_list);
+}
+
+/*
+ * This routine assumes that the appropriate zone's lru lock is already held
+ */
+void mem_cgroup_move_lists(struct page_cgroup *pc, bool active)
+{
+ struct mem_cgroup *mem;
+ if (!pc)
+ return;
+
+ mem = pc->mem_cgroup;
+
+ spin_lock(&mem->lru_lock);
+ __mem_cgroup_move_lists(pc, active);
+ spin_unlock(&mem->lru_lock);
+}
+
+unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan,
+ struct list_head *dst,
+ unsigned long *scanned, int order,
+ int mode, struct zone *z,
+ struct mem_cgroup *mem_cont,
+ int active)
+{
+ unsigned long nr_taken = 0;
+ struct page *page;
+ unsigned long scan;
+ LIST_HEAD(pc_list);
+ struct list_head *src;
+ struct page_cgroup *pc;
+
+ if (active)
+ src = &mem_cont->active_list;
+ else
+ src = &mem_cont->inactive_list;
+
+ spin_lock(&mem_cont->lru_lock);
+ for (scan = 0; scan < nr_to_scan && !list_empty(src); scan++) {
+ pc = list_entry(src->prev, struct page_cgroup, lru);
+ page = pc->page;
+ VM_BUG_ON(!pc);
+
+ if (PageActive(page) && !active) {
+ __mem_cgroup_move_lists(pc, true);
+ scan--;
+ continue;
+ }
+ if (!PageActive(page) && active) {
+ __mem_cgroup_move_lists(pc, false);
+ scan--;
+ continue;
+ }
+
+ /*
+ * Reclaim, per zone
+ * TODO: make the active/inactive lists per zone
+ */
+ if (page_zone(page) != z)
+ continue;
+
+ /*
+ * Check if the meta page went away from under us
+ */
+ if (!list_empty(&pc->lru))
+ list_move(&pc->lru, &pc_list);
+ else
+ continue;
+
+ if (__isolate_lru_page(page, mode) == 0) {
+ list_move(&page->lru, dst);
+ nr_taken++;
+ }
+ }
+
+ list_splice(&pc_list, src);
+ spin_unlock(&mem_cont->lru_lock);
+
+ *scanned = scan;
+ return nr_taken;
+}
+
/*
* Charge the memory controller for page usage.
* Return
@@ -151,6 +248,8 @@ int mem_cgroup_charge(struct page *page, struct mm_struct *mm)
{
struct mem_cgroup *mem;
struct page_cgroup *pc, *race_pc;
+ unsigned long flags;
+ unsigned long nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
/*
* Should page_cgroup's go to their own slab?
@@ -159,14 +258,20 @@ int mem_cgroup_charge(struct page *page, struct mm_struct *mm)
* to see if the cgroup page already has a page_cgroup associated
* with it
*/
+retry:
lock_page_cgroup(page);
pc = page_get_page_cgroup(page);
/*
* The page_cgroup exists and the page has already been accounted
*/
if (pc) {
- atomic_inc(&pc->ref_cnt);
- goto done;
+ if (unlikely(!atomic_inc_not_zero(&pc->ref_cnt))) {
+ /* this page is under being uncharged ? */
+ unlock_page_cgroup(page);
+ cpu_relax();
+ goto retry;
+ } else
+ goto done;
}
unlock_page_cgroup(page);
@@ -197,7 +302,32 @@ int mem_cgroup_charge(struct page *page, struct mm_struct *mm)
* If we created the page_cgroup, we should free it on exceeding
* the cgroup limit.
*/
- if (res_counter_charge(&mem->res, 1)) {
+ while (res_counter_charge(&mem->res, 1)) {
+ if (try_to_free_mem_cgroup_pages(mem))
+ continue;
+
+ /*
+ * try_to_free_mem_cgroup_pages() might not give us a full
+ * picture of reclaim. Some pages are reclaimed and might be
+ * moved to swap cache or just unmapped from the cgroup.
+ * Check the limit again to see if the reclaim reduced the
+ * current usage of the cgroup before giving up
+ */
+ if (res_counter_check_under_limit(&mem->res))
+ continue;
+ /*
+ * Since we control both RSS and cache, we end up with a
+ * very interesting scenario where we end up reclaiming
+ * memory (essentially RSS), since the memory is pushed
+ * to swap cache, we eventually end up adding those
+ * pages back to our list. Hence we give ourselves a
+ * few chances before we fail
+ */
+ else if (nr_retries--) {
+ congestion_wait(WRITE, HZ/10);
+ continue;
+ }
+
css_put(&mem->css);
goto free_pc;
}
@@ -221,14 +351,16 @@ int mem_cgroup_charge(struct page *page, struct mm_struct *mm)
pc->page = page;
page_assign_page_cgroup(page, pc);
+ spin_lock_irqsave(&mem->lru_lock, flags);
+ list_add(&pc->lru, &mem->active_list);
+ spin_unlock_irqrestore(&mem->lru_lock, flags);
+
done:
unlock_page_cgroup(page);
return 0;
free_pc:
kfree(pc);
- return -ENOMEM;
err:
- unlock_page_cgroup(page);
return -ENOMEM;
}
@@ -240,6 +372,7 @@ void mem_cgroup_uncharge(struct page_cgroup *pc)
{
struct mem_cgroup *mem;
struct page *page;
+ unsigned long flags;
if (!pc)
return;
@@ -252,6 +385,10 @@ void mem_cgroup_uncharge(struct page_cgroup *pc)
page_assign_page_cgroup(page, NULL);
unlock_page_cgroup(page);
res_counter_uncharge(&mem->res, 1);
+
+ spin_lock_irqsave(&mem->lru_lock, flags);
+ list_del_init(&pc->lru);
+ spin_unlock_irqrestore(&mem->lru_lock, flags);
kfree(pc);
}
}
@@ -310,6 +447,7 @@ mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont)
res_counter_init(&mem->res);
INIT_LIST_HEAD(&mem->active_list);
INIT_LIST_HEAD(&mem->inactive_list);
+ spin_lock_init(&mem->lru_lock);
return &mem->css;
}