On Fri, 28 Jan 2011 09:02:13 +0100
Johannes Weiner [off-list ref] wrote:

On Fri, Jan 28, 2011 at 12:27:29PM +0900, KAMEZAWA Hiroyuki wrote:

quoted

Thanks to Johanns and Daisuke for suggestion.
=
Hugepage allocation shouldn't trigger oom.
Allocation failure is not fatal.

Orignal-patch-by: Johannes Weiner [off-list ref]
Signed-off-by: KAMEZAWA Hiroyuki <redacted>
---
 mm/memcontrol.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

Index: mmotm-0125/mm/memcontrol.c
===================================================================

--- mmotm-0125.orig/mm/memcontrol.c
+++ mmotm-0125/mm/memcontrol.c

@@ -2369,11 +2369,14 @@ static int mem_cgroup_charge_common(stru
 	struct page_cgroup *pc;
 	int ret;
 	int page_size = PAGE_SIZE;
+	bool oom;
 
 	if (PageTransHuge(page)) {
 		page_size <<= compound_order(page);
 		VM_BUG_ON(!PageTransHuge(page));
-	}
+		oom = false;
+	} else
+		oom = true;

That needs a comment.  You can take the one from my patch if you like.

How about this ?
==
Hugepage allocation shouldn't trigger oom.
Allocation failure is not fatal.

Changelog:
 - added comments.

Orignal-patch-by: Johannes Weiner [off-list ref]
Signed-off-by: KAMEZAWA Hiroyuki <redacted>
---
 mm/memcontrol.c |   11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

Index: mmotm-0125/mm/memcontrol.c
===================================================================

--- mmotm-0125.orig/mm/memcontrol.c
+++ mmotm-0125/mm/memcontrol.c

@@ -2366,11 +2366,18 @@ static int mem_cgroup_charge_common(stru
 	struct page_cgroup *pc;
 	int ret;
 	int page_size = PAGE_SIZE;
+	bool oom;
 
 	if (PageTransHuge(page)) {
 		page_size <<= compound_order(page);
 		VM_BUG_ON(!PageTransHuge(page));
-	}
+		/*
+		 * Hugepage allocation will retry in small pages even if
+		 * this allocation fails.
+		 */
+		oom = false;
+	} else
+		oom = true;
 
 	pc = lookup_page_cgroup(page);
 	/* can happen at boot */

@@ -2378,7 +2385,7 @@ static int mem_cgroup_charge_common(stru
 		return 0;
 	prefetchw(pc);
 
-	ret = __mem_cgroup_try_charge(mm, gfp_mask, &mem, true, page_size);
+	ret = __mem_cgroup_try_charge(mm, gfp_mask, &mem, oom, page_size);
 	if (ret || !mem)
 		return ret;
 

--

To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>