change the locking order for namespace_sem
authorAl Viro <viro@zeniv.linux.org.uk>
Fri, 18 Mar 2011 12:55:38 +0000 (08:55 -0400)
committerAl Viro <viro@zeniv.linux.org.uk>
Fri, 18 Mar 2011 12:55:38 +0000 (08:55 -0400)
commitb12cea9198fa99ffd3de1776c323bc7464d26b44
tree3ae5818b2690e45c8a71432ed681751683091287
parent27cb1572e3e6bb1f8cf6bb3d74c914a87b131792
change the locking order for namespace_sem

Have it nested inside ->i_mutex.  Instead of using follow_down()
under namespace_sem, followed by grabbing i_mutex and checking that
mountpoint to be is not dead, do the following:
grab i_mutex
check that it's not dead
grab namespace_sem
see if anything is mounted there
if not, we've won
otherwise
drop locks
put_path on what we had
replace with what's mounted
retry everything with new mountpoint to be

New helper (lock_mount()) does that.  do_add_mount(), do_move_mount(),
do_loopback() and pivot_root() switched to it; in case of the last
two that eliminates a race we used to have - original code didn't
do follow_down().

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
fs/namespace.c