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, 3 May 2022 13:55:35 +0200
From:   "Jason A. Donenfeld" <Jason@zx2c4.com>
To:     Lennart Poettering <mzxreary@0pointer.de>
Cc:     Alexander Graf <graf@amazon.com>, linux-kernel@vger.kernel.org,
        linux-crypto@vger.kernel.org,
        Dominik Brodowski <linux@dominikbrodowski.net>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Theodore Ts'o <tytso@mit.edu>,
        Colm MacCarthaigh <colmmacc@amazon.com>,
        Torben Hansen <htorben@amazon.co.uk>,
        Jann Horn <jannh@google.com>
Subject: Re: [PATCH 2/2] random: add fork_event sysctl for polling VM forks
Message-ID: <YnEYN+Q3q76NVX3Q@zx2c4.com>
References: <20220502140602.130373-1-Jason@zx2c4.com>
 <20220502140602.130373-2-Jason@zx2c4.com>
 <Ym/7UlgQ5VjjC76P@gardel-login>
 <YnAC00VtU8MGb7vO@zx2c4.com>
 <YnAMBzhcJhGR5XOK@gardel-login>
 <7a1cfd1c-9f0e-f134-e544-83ee6d3cd9c9@amazon.com>
 <YnDn/d6iB0aUZkWJ@gardel-login>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <YnDn/d6iB0aUZkWJ@gardel-login>
Precedence: bulk

On Tue, May 03, 2022 at 10:29:49AM +0200, Lennart Poettering wrote:
> As mentioned earlier, I am not convinced sysctl is the right place for
> this. sysctls are understood by most people as being the place for
> tweaking kernel settings. This is not a kernel setting, but a
> notification concept, and the way Jason defined it there's nothing to
> read nor write, which strongly suggests to move it elsewhere, but not
> /proc/sys/.

I think I'm coming around to this view that having a sysctl return
-ENODATA is weird. It makes `sysctl -a` always complain to stderr, for
example, which seems bad.

> > I can see attractiveness in providing the /run/fork-id directly from the
> > kernel though, to remove the dependency on systemd for poll-less
> > notification of libraries.
> 
> I agree.

I'm still not convinced there's value in having a counter or a UUID, but
if you had to choose, would you prefer a counter or a UUID? It sounds
like the former, because you see a use for distinguishing between zero
and non-zero? Or did you finally agree with me that vmgenid isn't
granular enough for that?

Jason