Date: Mon, 25 Apr 2016 15:42:49 -0400
From: Tejun Heo <tj@kernel.org>
To: Michal Hocko <mhocko@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>, Petr Mladek <pmladek@suse.com>,
        cgroups@vger.kernel.org, Cyril Hrubis <chrubis@suse.cz>,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/2] memcg: relocate charge moving from ->attach to
 ->post_attach
Message-ID: <20160425194249.GG7822@mtj.duckdns.org>
References: <20160413094216.GC5774@pathway.suse.cz>
 <20160413183309.GG3676@htj.duckdns.org>
 <20160413192313.GA30260@dhcp22.suse.cz>
 <20160414175055.GA6794@cmpxchg.org>
 <20160415191719.GK12583@htj.duckdns.org>
 <20160417120747.GC21757@dhcp22.suse.cz>
 <20160421230648.GI4775@htj.duckdns.org>
 <20160421230902.GJ4775@htj.duckdns.org>
 <20160425082542.GB23933@dhcp22.suse.cz>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20160425082542.GB23933@dhcp22.suse.cz>
User-Agent: Mutt/1.5.24 (2015-08-30)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1049
Lines: 27

Hello, Michal.

On Mon, Apr 25, 2016 at 10:25:42AM +0200, Michal Hocko wrote:
> > * move_charge_struct->mm is added and ->can_attach is now responsible
> >   for pinning and recording the target mm.  mem_cgroup_clear_mc() is
> >   updated accordingly.  This also simplifies mem_cgroup_move_task().
> > 
> > * mem_cgroup_move_task() is now called from ->post_attach instead of
> >   ->attach.
> 
> This looks much better than the previous quick&dirty workaround. Thanks
> for taking an extra step!
> 
> Sorry for being so pushy but shouldn't we at least document those
> callbacks which depend on cgroup_threadgroup_rwsem to never ever block
> on WQ directly or indirectly. Maybe even enforce they have to be
> non-blocking. This would be out of scope of this particular patch of
> course but it would fit nicely into the series.

Yeah, I'll look into whether I can annotate this dependency with
lockdep.  If that isn't doable, I'll document the requirements.
cgroup methods have changed a lot and need to be redocumented anyway.

Thanks.

-- 
tejun