In vfs_busy(), lockmgr() cannot legitimately sleep, because code checked

MNTK_UNMOUNT before, and mnt_mtx is used as interlock. vfs_busy() always
tries to obtain a shared lock on mnt_lock, the other user is unmount who
tries to drain it, setting MNTK_UNMOUNT before.

Approved by:	re (kensmith)

sys/kern/vfs_subr.c

@@ -361,7 +361,7 @@ vfs_busy(struct mount *mp, int flags, struct mtx *interlkp,
 	}
 	if (interlkp)
 		mtx_unlock(interlkp);
-	lkflags = LK_SHARED | LK_INTERLOCK;
+	lkflags = LK_SHARED | LK_INTERLOCK | LK_NOWAIT;
 	if (lockmgr(&mp->mnt_lock, lkflags, MNT_MTX(mp), td))
 		panic("vfs_busy: unexpected lock failure");
 	return (0);