Received-SPF: pass (google.com: domain of linux-kernel+bounces-208825-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Sender: Tejun Heo <htejun@gmail.com>
Date: Mon, 10 Jun 2024 10:02:42 -1000
From: Tejun Heo <tj@kernel.org>
To: Oleg Nesterov <oleg@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	"Eric W. Biederman" <ebiederm@xmission.com>,
	linux-kernel@vger.kernel.org, Zefan Li <lizefan.x@bytedance.com>,
	Johannes Weiner <hannes@cmpxchg.org>
Subject: Re: Q: css_task_iter_advance() && dying_tasks
Message-ID: <Zmdb4moXhvkU8P0R@slm.duckdns.org>
References: <20240609142342.GA11165@redhat.com>
 <20240610105028.GA21586@redhat.com>
 <20240610110852.GA21828@redhat.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240610110852.GA21828@redhat.com>

On Mon, Jun 10, 2024 at 01:08:52PM +0200, Oleg Nesterov wrote:
> Sorry for the spam, forgot to mention. Either way the usage of group_dead
> or atomic_read(signal->live) in these paths doesn't look "perfect", but
> this is another thing. The pseudo code below tries to mimic the current
> logic but again, I'm afraid I misread this code completely.

The usage of signal->live there is something I added without much thinking.
I just needed something which goes off after all sub-threads are gone.
Getting @group_dead from do_exit() sounds perfectly fine to me.

Thanks.

-- 
tejun