From: Michal Hocko mhocko@suse.com Subject: Revert "mm/gup: check page posion status for coredump."
While reviewing http://lkml.kernel.org/r/20210429122519.15183-4-david@redhat.com I have crossed d3378e86d182 ("mm/gup: check page posion status for coredump.") and noticed that this patch is broken in two ways. First it doesn't really prevent hwpoison pages from being dumped because hwpoison pages can be marked asynchornously at any time after the check. Secondly, and more importantly, the patch introduces a ref count leak because get_dump_page takes a reference on the page which is not released.
It also seems that the patch was merged incorrectly because there were follow up changes not included as well as discussions on how to address the underlying problem http://lkml.kernel.org/r/57ac524c-b49a-99ec-c1e4-ef5027bfb61b@redhat.com
Therefore revert the original patch.
Link: https://lkml.kernel.org/r/20210505135407.31590-1-mhocko@kernel.org Fixes: d3378e86d182 ("mm/gup: check page posion status for coredump.") Signed-off-by: Michal Hocko mhocko@suse.com Reviewed-by: David Hildenbrand david@redhat.com Cc: Aili Yao yaoaili@kingsoft.com Cc: stable@vger.kernel.org Signed-off-by: Andrew Morton akpm@linux-foundation.org ---
mm/gup.c | 4 ---- mm/internal.h | 20 -------------------- 2 files changed, 24 deletions(-)
--- a/mm/gup.c~revert-mm-gup-check-page-posion-status-for-coredump +++ a/mm/gup.c @@ -1593,10 +1593,6 @@ struct page *get_dump_page(unsigned long FOLL_FORCE | FOLL_DUMP | FOLL_GET); if (locked) mmap_read_unlock(mm); - - if (ret == 1 && is_page_poisoned(page)) - return NULL; - return (ret == 1) ? page : NULL; } #endif /* CONFIG_ELF_CORE */ --- a/mm/internal.h~revert-mm-gup-check-page-posion-status-for-coredump +++ a/mm/internal.h @@ -96,26 +96,6 @@ static inline void set_page_refcounted(s set_page_count(page, 1); }
-/* - * When kernel touch the user page, the user page may be have been marked - * poison but still mapped in user space, if without this page, the kernel - * can guarantee the data integrity and operation success, the kernel is - * better to check the posion status and avoid touching it, be good not to - * panic, coredump for process fatal signal is a sample case matching this - * scenario. Or if kernel can't guarantee the data integrity, it's better - * not to call this function, let kernel touch the poison page and get to - * panic. - */ -static inline bool is_page_poisoned(struct page *page) -{ - if (PageHWPoison(page)) - return true; - else if (PageHuge(page) && PageHWPoison(compound_head(page))) - return true; - - return false; -} - extern unsigned long highest_memmap_pfn;
/* _