Re: [PATCH] mm: compaction: Abort compaction if too many pages are isolated and caller is asynchronous

[Andrea Arcangeli <aarcange@redhat.com>]

On Tue, May 31, 2011 at 11:38:30PM +0900, Minchan Kim wrote:
> > Yes. You find a new BUG.
> > It seems to be related to this problem but it should be solved although
> 
>  typo : It doesn't seem to be.

This should fix it, but I doubt it matters for this problem.

===
Subject: mm: no page_count without a page pin

From: Andrea Arcangeli <aarcange@redhat.com>

It's unsafe to run page_count during the physical pfn scan.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
---

diff --git a/mm/vmscan.c b/mm/vmscan.c
index faa0a08..e41e78a 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1124,8 +1124,18 @@ static unsigned long isolate_lru_pages(unsigned long nr_to_scan,
 					nr_lumpy_dirty++;
 				scan++;
 			} else {
-				/* the page is freed already. */
-				if (!page_count(cursor_page))
+				/*
+				 * We can't use page_count() as that
+				 * requires compound_head and we don't
+				 * have a pin on the page here. If a
+				 * page is tail, we may or may not
+				 * have isolated the head, so assume
+				 * it's not free, it'd be tricky to
+				 * track the head status without a
+				 * page pin.
+				 */
+				if (!PageTail(cursor_page) &&
+				    !atomic_read(&cursor_page->_count))
 					continue;
 				break;
 			}