From e1f1b1572e8db87a56609fd05bef76f98f0e456a Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Fri, 20 Jul 2018 17:53:45 -0700 Subject: mm/huge_memory.c: fix data loss when splitting a file pmd __split_huge_pmd_locked() must check if the cleared huge pmd was dirty, and propagate that to PageDirty: otherwise, data may be lost when a huge tmpfs page is modified then split then reclaimed. How has this taken so long to be noticed? Because there was no problem when the huge page is written by a write system call (shmem_write_end() calls set_page_dirty()), nor when the page is allocated for a write fault (fault_dirty_shared_page() calls set_page_dirty()); but when allocated for a read fault (which MAP_POPULATE simulates), no set_page_dirty(). Link: http://lkml.kernel.org/r/alpine.LSU.2.11.1807111741430.1106@eggly.anvils Fixes: d21b9e57c74c ("thp: handle file pages in split_huge_pmd()") Signed-off-by: Hugh Dickins Reported-by: Ashwin Chaugule Reviewed-by: Yang Shi Reviewed-by: Kirill A. Shutemov Cc: "Huang, Ying" Cc: [4.8+] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/huge_memory.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 1cd7c1a57a14..25346bd99364 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -2084,6 +2084,8 @@ static void __split_huge_pmd_locked(struct vm_area_struct *vma, pmd_t *pmd, if (vma_is_dax(vma)) return; page = pmd_page(_pmd); + if (!PageDirty(page) && pmd_dirty(_pmd)) + set_page_dirty(page); if (!PageReferenced(page) && pmd_young(_pmd)) SetPageReferenced(page); page_remove_rmap(page, true); -- cgit v1.2.3-59-g8ed1b