On Wed 31-10-12 09:41:23, Tejun Heo wrote:

Hey, Michal.

On Wed, Oct 31, 2012 at 03:37:51PM +0100, Michal Hocko wrote:

quoted

quoted

+	for_each_subsys(cgrp->root, ss)
+		if (ss->pre_destroy)
+			WARN_ON_ONCE(ss->pre_destroy(cgrp));

Hmm, I am not sure I like this WARN_ON_ONCE. First it can happen for
more than one controller and second we can just clear CGRP_WAIT_ON_RMDIR
and return with EBUSY. The only possible failure at the moment is when a
new task or a child group appear.
I know it is not a big deal because it will disappear later in the
series but it would be more readable IMO.

The WARN_ON_ONCE() is just moved from the original
cgroup_call_pre_destroy().  We can add an error out there but that
makes future changes difficult.  It's a chicken and egg problem.  You
gotta break the loop somewhere.

I do not think this is that hard. You can simply change the return path
on pre_destroy error by BUG_ON in "cgroup: deactivate CSS's and mark
cgroup dead before invoking ->pre_destroy()". There is no chicke&egg
problem here because once the group is marked dead and css frozen then
the existing callbacks cannot possibly fail.
Or am I missing your point?
-- 
Michal Hocko
SUSE Labs