Received-SPF: pass (google.com: domain of linux-kernel+bounces-208499-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Mon, 10 Jun 2024 17:42:29 +0200
From: Oleg Nesterov <oleg@redhat.com>
To: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Andrew Morton <akpm@linux-foundation.org>, Tejun Heo <tj@kernel.org>,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH 0/1] exit: kill signal_struct->quick_threads
Message-ID: <20240610154229.GD20640@redhat.com>
References: <20240609142342.GA11165@redhat.com>
 <87r0d5t2nt.fsf@email.froward.int.ebiederm.org>
 <20240610152902.GC20640@redhat.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240610152902.GC20640@redhat.com>
User-Agent: Mutt/1.5.24 (2015-08-30)

Forgot to mention ...

On 06/10, Oleg Nesterov wrote:
>
> I guess I was not clear. Of course, SIGNAL_GROUP_EXIT must be always
> set under ->siglock. But I think synchronize_group_exit() can just
> return if SIGNAL_GROUP_EXIT is already set? If nothing else, this is
> what do_group_exit() does.

Just in case... I guess do_group_exit() needs some cleanups too. For
example,

	if (sig->flags & SIGNAL_GROUP_EXIT)
		exit_code = sig->group_exit_code;

without siglock is racy, exit_code can be wrong (not a big problem).
LOAD(sig->flags) can make KCSAN unhappy. Etc.

But this all has nothing to do with 1/1 ?

Oleg.