Memory controller: make charging gfp mask aware
Nick Piggin pointed out that swap cache and page cache addition routines
could be called from non GFP_KERNEL contexts. This patch makes the
charging routine aware of the gfp context. Charging might fail if the
cgroup is over it's limit, in which case a suitable error is returned.
This patch was tested on a Powerpc box. I am still looking at being able
to test the path, through which allocations happen in non GFP_KERNEL
contexts.
[kamezawa.hiroyu@jp.fujitsu.com: problem with ZONE_MOVABLE]
Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Pavel Emelianov <xemul@openvz.org>
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: 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>
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index ff7cac6..ac87744 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -261,7 +261,8 @@
* 0 if the charge was successful
* < 0 if the cgroup is over its limit
*/
-int mem_cgroup_charge(struct page *page, struct mm_struct *mm)
+int mem_cgroup_charge(struct page *page, struct mm_struct *mm,
+ gfp_t gfp_mask)
{
struct mem_cgroup *mem;
struct page_cgroup *pc, *race_pc;
@@ -293,7 +294,7 @@
unlock_page_cgroup(page);
- pc = kzalloc(sizeof(struct page_cgroup), GFP_KERNEL);
+ pc = kzalloc(sizeof(struct page_cgroup), gfp_mask);
if (pc == NULL)
goto err;
@@ -320,7 +321,14 @@
* the cgroup limit.
*/
while (res_counter_charge(&mem->res, PAGE_SIZE)) {
- if (try_to_free_mem_cgroup_pages(mem))
+ bool is_atomic = gfp_mask & GFP_ATOMIC;
+ /*
+ * We cannot reclaim under GFP_ATOMIC, fail the charge
+ */
+ if (is_atomic)
+ goto noreclaim;
+
+ if (try_to_free_mem_cgroup_pages(mem, gfp_mask))
continue;
/*
@@ -344,9 +352,10 @@
congestion_wait(WRITE, HZ/10);
continue;
}
-
+noreclaim:
css_put(&mem->css);
- mem_cgroup_out_of_memory(mem, GFP_KERNEL);
+ if (!is_atomic)
+ mem_cgroup_out_of_memory(mem, GFP_KERNEL);
goto free_pc;
}
@@ -385,7 +394,8 @@
/*
* See if the cached pages should be charged at all?
*/
-int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm)
+int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm,
+ gfp_t gfp_mask)
{
struct mem_cgroup *mem;
if (!mm)
@@ -393,7 +403,7 @@
mem = rcu_dereference(mm->mem_cgroup);
if (mem->control_type == MEM_CGROUP_TYPE_ALL)
- return mem_cgroup_charge(page, mm);
+ return mem_cgroup_charge(page, mm, gfp_mask);
else
return 0;
}