On Sat, 2016-12-03 at 05:13 +0000, Al Viro wrote: > FWIW, I've folded that pile into vfs.git#work.autofs. > > Problems: > * (fixed) __path_is_mountpoint() should _not_ treat NULL from > __lookup_mnt() as "nothing's mounted there" until it has checked > that mount_lock hadn't been touched - mount --move on something unrelated > can race with lockless hash lookup and lead to false negatives. Right, looking at what you've done there the mistake is so obvious now! Thanks for helping with it. > * linux/mount.h might be the wrong place for path_is_mountpoint(). > Or it shouldn't be inlined. I don't like the includes you've added there. > * path_has_submounts() is broken. At the very least, it's > AB-BA between mount_lock and rename_lock. I would suggest trying to > put read_seqlock_excl(&mount_lock) around the call of d_walk() in there, > and using __lookup_mnt() in the callback (without retries on the mount_lock, > of course - read_seqlock_excl done on the outside is enough). I'm not sure > if it won't cause trouble with contention, though; that needs testing. As > it is, that function is broken in #work.autofs, same as it is in -mm and > -next. Umm ... that's a much more obvious dumb mistake and what you've done there didn't occur to me even after you spelled it out, I'll take some time to digest it. And thanks for that one too. Ian -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
