Proper testing shows yet another problem in balloon migration: it works only once for each page. balloon_page_movable() check page flags and page_count. In __unmap_and_move page is locked, reference counter is elevated, so balloon_page_movable() _always_ fails here. As result in __unmap_and_move() migration goes to the normal migration path.
Balloon ->migratepage() is so special, it returns MIGRATEPAGE_BALLOON_SUCCESS instead of MIGRATEPAGE_SUCCESS. After that in move_to_new_page() successfully migrated page got NULL into its mapping pointer and loses connectivity with balloon and ability for further migration. It's safe to use __is_movable_balloon_page here: page is isolated and pinned. Signed-off-by: Konstantin Khlebnikov <k.khlebni...@samsung.com> Cc: stable <sta...@vger.kernel.org> # v3.8 --- mm/migrate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/migrate.c b/mm/migrate.c index f78ec9b..161d044 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -873,7 +873,7 @@ static int __unmap_and_move(struct page *page, struct page *newpage, } } - if (unlikely(balloon_page_movable(page))) { + if (unlikely(__is_movable_balloon_page(page))) { /* * A ballooned page does not need any special attention from * physical to virtual reverse mapping procedures. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/