ext4: Reorder fs/Makefile so that ext2 root fs's are mounted using ext2
authorTheodore Ts'o <tytso@mit.edu>
Sat, 28 Feb 2009 14:50:01 +0000 (09:50 -0500)
committerTheodore Ts'o <tytso@mit.edu>
Sat, 28 Feb 2009 14:50:01 +0000 (09:50 -0500)
commitd8ae4601a4b7ea1fa17fa395c3468c0e144d1275
treece4a6593aaa883377fdd2754655ff5815d363111
parent8b1a8ff8b321a9384304aeea4dbdb9747daf7ee8
ext4: Reorder fs/Makefile so that ext2 root fs's are mounted using ext2

In fs/Makefile, ext3 was placed before ext2 so that a root filesystem
that possessed a journal, it would be mounted as ext3 instead of ext2.
This was necessary because a cleanly unmounted ext3 filesystem was
fully backwards compatible with ext2, and could be mounted by ext2 ---
but it was desirable that it be mounted with ext3 so that the
journaling would be enabled.

The ext4 filesystem supports new incompatible features, so there is no
danger of an ext4 filesystem being mistaken for an ext2 filesystem.
At that point, the relative ordering of ext4 with respect to ext2
didn't matter until ext4 gained the ability to mount filesystems
without a journal starting in 2.6.29-rc1.  Now that this is the case,
given that ext4 is before ext2, it means that root filesystems that
were using the plain-jane ext2 format are getting mounted using the
ext4 filesystem driver, which is a change in behavior which could be
surprising to users.

It's doubtful that there are that many ext2-only root filesystem users
that would also have ext4 compiled into the kernel, but to adhere to
the principle of least surprise, the correct ordering in fs/Makefile
is ext3, followed by ext2, and finally ext4.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
fs/Makefile