mm/migrate: fixup setting UFFD_WP flag
authorAlistair Popple <alistair@popple.id.au>
Fri, 4 Sep 2020 23:35:58 +0000 (16:35 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 9 Sep 2020 17:14:33 +0000 (19:14 +0200)
commit ebdf8321eeeb623aed60f7ed16f7445363230118 upstream.

Commit f45ec5ff16a75 ("userfaultfd: wp: support swap and page migration")
introduced support for tracking the uffd wp bit during page migration.
However the non-swap PTE variant was used to set the flag for zone device
private pages which are a type of swap page.

This leads to corruption of the swap offset if the original PTE has the
uffd_wp flag set.

Fixes: f45ec5ff16a75 ("userfaultfd: wp: support swap and page migration")
Signed-off-by: Alistair Popple <alistair@popple.id.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Link: https://lkml.kernel.org/r/20200825064232.10023-1-alistair@popple.id.au
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mm/migrate.c

index 6a9ffe887e911a8db3a658c49572bcdd6a6896a5..3511f9529ea60aced2194f38123c96e259602193 100644 (file)
@@ -251,7 +251,7 @@ static bool remove_migration_pte(struct page *page, struct vm_area_struct *vma,
                                entry = make_device_private_entry(new, pte_write(pte));
                                pte = swp_entry_to_pte(entry);
                                if (pte_swp_uffd_wp(*pvmw.pte))
-                                       pte = pte_mkuffd_wp(pte);
+                                       pte = pte_swp_mkuffd_wp(pte);
                        }
                }