On Mon, Mar 22, 2021 at 10:51 AM Peter Xu <peterx@xxxxxxxxxx> wrote:
>
> When uffd-minor enabled, we need to put the page cache before handling the
> userfault in hugetlb_no_page(), otherwise the page refcount got leaked.
>
> This can be reproduced by running userfaultfd selftest with hugetlb_shared
> mode, then cat /proc/meminfo.
>
> Cc: Axel Rasmussen <axelrasmussen@xxxxxxxxxx>
> Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx>
> Cc: Mike Kravetz <mike.kravetz@xxxxxxxxxx>
> Cc: Mike Rapoport <rppt@xxxxxxxxxxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Fixes: f2bf15fb0969 ("userfaultfd: add minor fault registration mode")
> Signed-off-by: Peter Xu <peterx@xxxxxxxxxx>
> ---
> mm/hugetlb.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> index 408dbc08298a..56b78a206913 100644
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -4449,6 +4449,7 @@ static vm_fault_t hugetlb_no_page(struct mm_struct *mm,
> /* Check for page in userfault range. */
> if (userfaultfd_minor(vma)) {
> unlock_page(page);
> + put_page(page);
> ret = hugetlb_handle_userfault(vma, mapping, idx,
> flags, haddr,
> VM_UFFD_MINOR);
> --
> 2.26.2
>
Thanks for the fix, Peter! I applied the patch and verified it does
indeed fix the issue.
Reviewed-by: Axel Rasmussen <axelrasmussen@xxxxxxxxxx>
