-
Al Viro authored
* if you do dput() of root dentry, do *not* follow that with iput() of root inode. * while we are at it, don't do that dput() at all - you are leaving the pointer in ->s_root and your ->kill_sb() will be very unhappy with that. It will do proper dput(), though, so the easiest way is to leave that to it entirely. * freeing ->s_fs_info is also best left to ->kill_sb() (which will do it anyway), especially since we leave the pointer in place. * that xchg() in ->kill_sb() is not a bug per se, but it's a plain and simple masturbation with fewer excuses than Onan had... Signed-off-by:
Al Viro <viro@zeniv.linux.org.uk>
5b5f9560