Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Tue, 15 Feb 2022 11:25:18 +0100
From:   Michal =?iso-8859-1?Q?Koutn=FD?= <mkoutny@suse.com>
To:     "Eric W. Biederman" <ebiederm@xmission.com>
Cc:     Solar Designer <solar@openwall.com>, linux-kernel@vger.kernel.org,
        Alexey Gladkov <legion@kernel.org>,
        Kees Cook <keescook@chromium.org>,
        Shuah Khan <shuah@kernel.org>,
        Christian Brauner <brauner@kernel.org>,
        Ran Xiaokai <ran.xiaokai@zte.com.cn>, stable@vger.kernel.org
Subject: Re: [PATCH 3/8] ucounts: Fix and simplify RLIMIT_NPROC handling
 during setuid()+execve
Message-ID: <20220215102518.GE21589@blackbody.suse.cz>
References: <87o83e2mbu.fsf@email.froward.int.ebiederm.org>
 <20220211021324.4116773-3-ebiederm@xmission.com>
 <20220212231701.GA29483@openwall.com>
 <87ee45wkjq.fsf@email.froward.int.ebiederm.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <87ee45wkjq.fsf@email.froward.int.ebiederm.org>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: bulk

On Mon, Feb 14, 2022 at 09:10:49AM -0600, "Eric W. Biederman" <ebiederm@xmission.com> wrote:
> I really like how cleanly this patch seems to be.  Unfortunately it is
> wrong.

It seems [1] so:

setuid()		// RLIMIT_NPROC is fine at this moment
...		fork()
		...
...		fork()
execve()		// eh, oh

This "punishes" the exec'ing task although the cause is elsewhere.

Michal

[1] The decoupled setuid()+execve() check can be interpretted both ways.
I understood historically the excess at the setuid() moment is relevant.