From: Andrea Arcangeli Date: Fri, 28 Sep 2012 12:35:31 +0000 (+0200) Subject: thp: avoid VM_BUG_ON page_count(page) false positives in __collapse_huge_page_copy X-Git-Tag: MMI-PSA29.97-13-9~15941 X-Git-Url: https://git.stricted.de/?a=commitdiff_plain;h=99a1300e1d84709f419182bb5189760e78234882;p=GitHub%2FMotorolaMobilityLLC%2Fkernel-slsi.git thp: avoid VM_BUG_ON page_count(page) false positives in __collapse_huge_page_copy Speculative cache pagecache lookups can elevate the refcount from under us, so avoid the false positive. If the refcount is < 2 we'll be notified by a VM_BUG_ON in put_page_testzero as there are two put_page(src_page) in a row before returning from this function. Signed-off-by: Andrea Arcangeli Reviewed-by: Rik van Riel Reviewed-by: Johannes Weiner Cc: Hugh Dickins Cc: Mel Gorman Cc: Petr Holasek Cc: Andrew Morton Signed-off-by: Linus Torvalds --- diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 57c4b9309015..141dbb695097 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -1811,7 +1811,6 @@ static void __collapse_huge_page_copy(pte_t *pte, struct page *page, src_page = pte_page(pteval); copy_user_highpage(page, src_page, address, vma); VM_BUG_ON(page_mapcount(src_page) != 1); - VM_BUG_ON(page_count(src_page) != 2); release_pte_page(src_page); /* * ptl mostly unnecessary, but preempt has to