Revision 6b7339f4c31ad69c8e9c0b2859276e22cf72176d authored by Kirill A. Shutemov on 06 July 2015, 20:18:37 UTC, committed by Linus Torvalds on 09 July 2015, 18:12:48 UTC
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>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent 883a2df
History
File Mode Size
Kconfig -rw-r--r-- 3.1 KB
Makefile -rw-r--r-- 416 bytes
acl.c -rw-r--r-- 8.8 KB
acl.h -rw-r--r-- 1.2 KB
checkpoint.c -rw-r--r-- 27.3 KB
crypto.c -rw-r--r-- 12.7 KB
crypto_fname.c -rw-r--r-- 11.0 KB
crypto_key.c -rw-r--r-- 6.7 KB
crypto_policy.c -rw-r--r-- 5.9 KB
data.c -rw-r--r-- 50.0 KB
debug.c -rw-r--r-- 13.1 KB
dir.c -rw-r--r-- 20.8 KB
f2fs.h -rw-r--r-- 63.0 KB
f2fs_crypto.h -rw-r--r-- 3.8 KB
file.c -rw-r--r-- 37.3 KB
gc.c -rw-r--r-- 20.2 KB
gc.h -rw-r--r-- 3.0 KB
hash.c -rw-r--r-- 2.0 KB
inline.c -rw-r--r-- 12.7 KB
inode.c -rw-r--r-- 10.4 KB
namei.c -rw-r--r-- 24.0 KB
node.c -rw-r--r-- 50.1 KB
node.h -rw-r--r-- 10.5 KB
recovery.c -rw-r--r-- 13.7 KB
segment.c -rw-r--r-- 61.2 KB
segment.h -rw-r--r-- 22.3 KB
super.c -rw-r--r-- 35.3 KB
trace.c -rw-r--r-- 3.4 KB
trace.h -rw-r--r-- 1013 bytes
xattr.c -rw-r--r-- 14.8 KB
xattr.h -rw-r--r-- 4.6 KB

back to top