[BUGFIX][PATCH 2/4] memcg: fix charge path for THP and allow earlyretirement

[KAMEZAWA Hiroyuki]

From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>

When THP is used, Hugepage size charge can happen. It's not handled
correctly in mem_cgroup_do_charge(). For example, THP can fallback
to small page allocation when HUGEPAGE allocation seems difficult
or busy, but memory cgroup doesn't understand it and continue to
try HUGEPAGE charging. And the worst thing is memory cgroup
believes 'memory reclaim succeeded' if limit - usage > PAGE_SIZE.

By this, khugepaged etc...can goes into inifinite reclaim loop
if tasks in memcg are busy.

After this patch 
 - Hugepage allocation will fail if 1st trial of page reclaim fails.

Changelog:
 - make changes small. removed renaming codes.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
---
 mm/memcontrol.c |   28 ++++++++++++++++++++++++----
 1 file changed, 24 insertions(+), 4 deletions(-)

Index: mmotm-0125/mm/memcontrol.c
===================================================================
--- mmotm-0125.orig/mm/memcontrol.c
+++ mmotm-0125/mm/memcontrol.c
@@ -1827,10 +1827,14 @@ enum {
 	CHARGE_OK,		/* success */
 	CHARGE_RETRY,		/* need to retry but retry is not bad */
 	CHARGE_NOMEM,		/* we can't do more. return -ENOMEM */
+	CHARGE_NEED_BREAK,	/* big size allocation failure */
 	CHARGE_WOULDBLOCK,	/* GFP_WAIT wasn't set and no enough res. */
 	CHARGE_OOM_DIE,		/* the current is killed because of OOM */
 };
 
+/*
+ * Now we have 3 charge size as PAGE_SIZE, HPAGE_SIZE and batched allcation.
+ */
 static int __mem_cgroup_do_charge(struct mem_cgroup *mem, gfp_t gfp_mask,
 				int csize, bool oom_check)
 {
@@ -1854,9 +1858,6 @@ static int __mem_cgroup_do_charge(struct
 	} else
 		mem_over_limit = mem_cgroup_from_res_counter(fail_res, res);
 
-	if (csize > PAGE_SIZE) /* change csize and retry */
-		return CHARGE_RETRY;
-
 	if (!(gfp_mask & __GFP_WAIT))
 		return CHARGE_WOULDBLOCK;
 
@@ -1880,6 +1881,13 @@ static int __mem_cgroup_do_charge(struct
 		return CHARGE_RETRY;
 
 	/*
+	 * if request size is larger than PAGE_SIZE, it's not OOM
+	 * and caller will do retry in smaller size.
+	 */
+	if (csize != PAGE_SIZE)
+		return CHARGE_NEED_BREAK;
+
+	/*
 	 * At task move, charge accounts can be doubly counted. So, it's
 	 * better to wait until the end of task_move if something is going on.
 	 */
@@ -1997,10 +2005,22 @@ again:
 		case CHARGE_OK:
 			break;
 		case CHARGE_RETRY: /* not in OOM situation but retry */
-			csize = page_size;
 			css_put(&mem->css);
 			mem = NULL;
 			goto again;
+		case CHARGE_NEED_BREAK: /* page_size > PAGE_SIZE */
+			css_put(&mem->css);
+			/*
+			 * We'll come here in 2 caes, batched-charge and
+			 * hugetlb alloc. batched-charge can do retry
+			 * with smaller page size. hugepage should return
+			 * NOMEM. This doesn't mean OOM.
+			 */
+			if (page_size > PAGE_SIZE)
+				goto nomem;
+			csize = page_size;
+			mem = NULL;
+			goto again;
 		case CHARGE_WOULDBLOCK: /* !__GFP_WAIT */
 			css_put(&mem->css);
 			goto nomem;

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/