I observed that hugetlb does not support/expect write-faults in shared mappings that would have to map the R/O-mapped page writable -- and I found one case where we could currently get such faults and would erroneously map an anon page into a shared mapping, by triggering clear_refs to clear soft-dirty tracking at the wrong point in time on a process. I propose to backport the fix to stable trees, as the fix for write-notify should be straight forward. Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> Cc: Mike Kravetz <mike.kravetz@xxxxxxxxxx> Cc: Muchun Song <songmuchun@xxxxxxxxxxxxx> Cc: Peter Xu <peterx@xxxxxxxxxx> Cc: Peter Feiner <pfeiner@xxxxxxxxxx> Cc: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx> David Hildenbrand (2): mm/hugetlb: fix hugetlb not supporting write-notify mm/hugetlb: support write-faults in shared mappings mm/hugetlb.c | 21 ++++++++++++++------- mm/mmap.c | 7 +++++++ 2 files changed, 21 insertions(+), 7 deletions(-) -- 2.35.3
