ext4: fix data corruption for mmap writes
authorJan Kara <jack@suse.cz>
Fri, 26 May 2017 21:45:45 +0000 (17:45 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 12 Oct 2017 09:51:26 +0000 (11:51 +0200)
commit2d605d9188d64e7d2bd96b4ec5bf91434024a25a
tree4177f7a4f531c22a18fc736226e763d7215da847
parent27db1f020373a0681d483cc2d304e018cbd15723
ext4: fix data corruption for mmap writes

commit a056bdaae7a181f7dcc876cfab2f94538e508709 upstream.

mpage_submit_page() can race with another process growing i_size and
writing data via mmap to the written-back page. As mpage_submit_page()
samples i_size too early, it may happen that ext4_bio_write_page()
zeroes out too large tail of the page and thus corrupts user data.

Fix the problem by sampling i_size only after the page has been
write-protected in page tables by clear_page_dirty_for_io() call.

Reported-by: Michael Zimmer <michael@swarm64.com>
Fixes: cb20d5188366f04d96d2e07b1240cc92170ade40
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/ext4/inode.c