netdev - Re: [PATCH net-next v3 2/2] tcp: Free TCP-AO/TCP-MD5 info/keys without RCU

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAGrbwDTT-T=v672DR4wJU0qw_yO2QCMQ4OyuLjw+6Y=zSu5xfw@mail.gmail.com>
Date: Thu, 4 Sep 2025 00:17:34 +0100
From: Dmitry Safonov <dima@...sta.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: Dmitry Safonov via B4 Relay <devnull+dima.arista.com@...nel.org>, Eric Dumazet <edumazet@...gle.com>, 
	Neal Cardwell <ncardwell@...gle.com>, Kuniyuki Iwashima <kuniyu@...gle.com>, 
	"David S. Miller" <davem@...emloft.net>, David Ahern <dsahern@...nel.org>, 
	Paolo Abeni <pabeni@...hat.com>, Simon Horman <horms@...nel.org>, Bob Gilligan <gilligan@...sta.com>, 
	Salam Noureddine <noureddine@...sta.com>, Dmitry Safonov <0x7f454c46@...il.com>, netdev@...r.kernel.org, 
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next v3 2/2] tcp: Free TCP-AO/TCP-MD5 info/keys
 without RCU

On Wed, Sep 3, 2025 at 11:23 PM Jakub Kicinski <kuba@...nel.org> wrote:
> On Wed, 3 Sep 2025 18:41:39 +0100 Dmitry Safonov wrote:
> > > On Sat, 30 Aug 2025 05:31:47 +0100 Dmitry Safonov via B4 Relay wrote:
> > > > Now that the destruction of info/keys is delayed until the socket
> > > > destructor, it's safe to use kfree() without an RCU callback.
> > > > As either socket was yet in TCP_CLOSE state or the socket refcounter is
> > > > zero and no one can discover it anymore, it's safe to release memory
> > > > straight away.
> > > > Similar thing was possible for twsk already.
> > >
> > > After this patch the rcu members of struct tcp_ao* seem to no longer
> > > be used?
> >
> > Right. I'll remove tcp_ao_info::rcu in v4.
> > For tcp_ao_key it's needed for the regular key rotation, as well as
> > for tcp_md5sig_key.
>
> Hm, maybe I missed something. I did a test allmodconfig build yesterday
> and while the md5sig_key rcu was still needed, removing the ao_key
> didn't cause issues. But it was just a quick test I didn't even config
> kconfig is sane.

Hmm, probably CONFIG_TCP_AO was off?
tcp_ao_delete_key() does call_rcu(&key->rcu, tcp_ao_key_free_rcu).

Looking at the code now, I guess what I could have done even more is
migrating tcp_sock::ao_info (and tcp_timewait_sock::ao_info) from
rcu_*() helpers to acquire/release ones. Somewhat feeling uneasy about
going that far just yet. Should I do it with another cleanup on the
top, what do you think?

Thanks,
           Dmitry