[PATCH] mm/hwpoison: fix wrong num_poisoned_pages account

[Minchan Kim]

Currently, migration code increases num_poisoned_pages on failed
migration page as well as successfully migrated one at the trial
of memory-failure. It will make the stat wrong.

As well, it marks page as PG_HWPoison even if the migration trial
failed. It would make we cannot recover the corrupted page using
memory-failure facility.

This patches fixes it.

Cc: stable@xxxxxxxxxxxxxxx
Reported-by: Vlastimil Babka <vbabka@xxxxxxx>
Acked-by: Naoya Horiguchi <n-horiguchi@xxxxxxxxxxxxx>
Signed-off-by: Minchan Kim <minchan@xxxxxxxxxx>
---
 mm/migrate.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/mm/migrate.c b/mm/migrate.c
index 6c822a7b27e0..f9dfb18a4eba 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -975,7 +975,13 @@ static ICE_noinline int unmap_and_move(new_page_t get_new_page,
 		dec_zone_page_state(page, NR_ISOLATED_ANON +
 				page_is_file_cache(page));
 		/* Soft-offlined page shouldn't go through lru cache list */
-		if (reason == MR_MEMORY_FAILURE) {
+		if (reason == MR_MEMORY_FAILURE && rc == MIGRATEPAGE_SUCCESS) {
+			/*
+			 * With this release, we free successfully migrated
+			 * page and set PG_HWPoison on just freed page
+			 * intentionally. Although it's rather weird, it's how
+			 * HWPoison flag works at the moment.
+			 */
 			put_page(page);
 			if (!test_set_page_hwpoison(page))
 				num_poisoned_pages_inc();
-- 
1.9.1