mm: avoid setting up anonymous pages into file mapping
authorKirill A. Shutemov <kirill.shutemov@linux.intel.com>
Mon, 6 Jul 2015 20:18:37 +0000 (23:18 +0300)
committerJiri Slaby <jslaby@suse.cz>
Wed, 19 Aug 2015 06:36:27 +0000 (08:36 +0200)
commitbf653833fc72668cadcb1f54310f3d0bd6b36393
treefabe83c3a78daf38ee430aa3b721b7530c67039c
parentc707649e400a18e0b2fa96e533270673cfa4687f
mm: avoid setting up anonymous pages into file mapping

commit 6b7339f4c31ad69c8e9c0b2859276e22cf72176d upstream.

Reading page fault handler code I've noticed that under right
circumstances kernel would map anonymous pages into file mappings: if
the VMA doesn't have vm_ops->fault() and the VMA wasn't fully populated
on ->mmap(), kernel would handle page fault to not populated pte with
do_anonymous_page().

Let's change page fault handler to use do_anonymous_page() only on
anonymous VMA (->vm_ops == NULL) and make sure that the VMA is not
shared.

For file mappings without vm_ops->fault() or shred VMA without vm_ops,
page fault on pte_none() entry would lead to SIGBUS.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
mm/memory.c