ext4: prevent right-shifting extents beyond EXT_MAX_BLOCKS
authorEric Biggers <ebiggers@google.com>
Thu, 12 Apr 2018 15:48:09 +0000 (11:48 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 1 May 2018 19:58:06 +0000 (12:58 -0700)
commit1a538cb0879d8646437836d540b6e952d9103f63
tree9d34846d2ad7daa408d3488b0d3b31c45fb1578f
parenta87463f7420cd33c8cb3ece420468023c16f27b4
ext4: prevent right-shifting extents beyond EXT_MAX_BLOCKS

commit 349fa7d6e1935f49bf4161c4900711b2989180a9 upstream.

During the "insert range" fallocate operation, extents starting at the
range offset are shifted "right" (to a higher file offset) by the range
length.  But, as shown by syzbot, it's not validated that this doesn't
cause extents to be shifted beyond EXT_MAX_BLOCKS.  In that case
->ee_block can wrap around, corrupting the extent tree.

Fix it by returning an error if the space between the end of the last
extent and EXT4_MAX_BLOCKS is smaller than the range being inserted.

This bug can be reproduced by running the following commands when the
current directory is on an ext4 filesystem with a 4k block size:

        fallocate -l 8192 file
        fallocate --keep-size -o 0xfffffffe000 -l 4096 -n file
        fallocate --insert-range -l 8192 file

Then after unmounting the filesystem, e2fsck reports corruption.

Reported-by: syzbot+06c885be0edcdaeab40c@syzkaller.appspotmail.com
Fixes: 331573febb6a ("ext4: Add support FALLOC_FL_INSERT_RANGE for fallocate")
Cc: stable@vger.kernel.org # v4.2+
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/ext4/extents.c