netdev - Re: mptcp splat

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: <8410a6f61e7a778117819ebeda667687353ffb21.camel@redhat.com>
Date: Wed, 27 Mar 2024 17:56:04 +0100
From: Paolo Abeni <pabeni@...hat.com>
To: Alexei Starovoitov <alexei.starovoitov@...il.com>, Network Development
 <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>, MPTCP Upstream
 <mptcp@...ts.linux.dev>, Matthieu Baerts <matttbe@...nel.org>, Mat
 Martineau <martineau@...nel.org>, Jakub Kicinski <kuba@...nel.org>, Martin
 KaFai Lau <martin.lau@...nel.org>
Subject: Re: mptcp splat

On Wed, 2024-03-27 at 09:43 -0700, Alexei Starovoitov wrote:
> I ffwded bpf tree with the recent net fixes and caught this:
> 
> [   48.386337] WARNING: CPU: 32 PID: 3276 at net/mptcp/subflow.c:1430
> subflow_data_ready+0x147/0x1c0
> [   48.392012] Modules linked in: dummy bpf_testmod(O) [last unloaded:
> bpf_test_no_cfi(O)]
> [   48.396609] CPU: 32 PID: 3276 Comm: test_progs Tainted: G
> O       6.8.0-12873-g2c43c33bfd23 #1014
> #[   48.467143] Call Trace:
> [   48.469094]  <TASK>
> [   48.472159]  ? __warn+0x80/0x180
> [   48.475019]  ? subflow_data_ready+0x147/0x1c0
> [   48.478068]  ? report_bug+0x189/0x1c0
> [   48.480725]  ? handle_bug+0x36/0x70
> [   48.483061]  ? exc_invalid_op+0x13/0x60
> [   48.485809]  ? asm_exc_invalid_op+0x16/0x20
> [   48.488754]  ? subflow_data_ready+0x147/0x1c0
> [   48.492159]  mptcp_set_rcvlowat+0x79/0x1d0
> [   48.495026]  sk_setsockopt+0x6c0/0x1540
> 
> It doesn't reproduce all the time though.
> Some race?
> Known issue?

It was not known to me. Looks like something related to not so recent
changes (rcvlowat support).

Definitely looks lie a race.

If you could share more info about the running context and/or a full
decoded splat it could help, thanks!

Paolo