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>

1 files changed, 17 insertions, 7 deletions

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index ff7cac602984..ac8774426fec 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -261,7 +261,8 @@ unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan,
  * 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 @@ retry:
 
         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 @@ retry:
          * 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 @@ retry:
                         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 @@ err:
 /*
  * 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 @@ int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm)
 
         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;
 }