ext4: handle EOF correctly in ext4_bio_write_page()
authorYongqiang Yang <xiaoqiangnk@gmail.com>
Wed, 14 Dec 2011 03:29:12 +0000 (22:29 -0500)
committerTheodore Ts'o <tytso@mit.edu>
Wed, 14 Dec 2011 03:29:12 +0000 (22:29 -0500)
We need to zero out part of a page which beyond EOF before setting uptodate,
otherwise, mapread or write will see non-zero data beyond EOF.

Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
fs/ext4/page-io.c

index 7ce1d0b19c94576892d0d80f016f7db44110ccb7..7e106c810c62bb18435d9a0fcb6d5066a3813fad 100644 (file)
@@ -385,6 +385,18 @@ int ext4_bio_write_page(struct ext4_io_submit *io,
 
                block_end = block_start + blocksize;
                if (block_start >= len) {
+                       /*
+                        * Comments copied from block_write_full_page_endio:
+                        *
+                        * The page straddles i_size.  It must be zeroed out on
+                        * each and every writepage invocation because it may
+                        * be mmapped.  "A file is mapped in multiples of the
+                        * page size.  For a file that is not a multiple of
+                        * the  page size, the remaining memory is zeroed when
+                        * mapped, and writes to that region are not written
+                        * out to the file."
+                        */
+                       zero_user_segment(page, block_start, block_end);
                        clear_buffer_dirty(bh);
                        set_buffer_uptodate(bh);
                        continue;