Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <A63CB783-F5AF-419A-89B9-3755E6896D41@purdue.edu>
In-Reply-To: <A63CB783-F5AF-419A-89B9-3755E6896D41@purdue.edu>
From:   Eric Dumazet <edumazet@google.com>
Date:   Thu, 15 Apr 2021 17:53:20 +0200
Message-ID: <CANn89iLdP+FFb7+TR-RHW1Kdm5XAoP1_4ypYp+Cm5wzPbx-K1w@mail.gmail.com>
Subject: Re: [PROBLEM] a data race between tcp_set_default_congestion_control()
 and tcp_set_congestion_control()
To:     "Gong, Sishuai" <sishuai@purdue.edu>
Cc:     "davem@davemloft.net" <davem@davemloft.net>,
        "yoshfuji@linux-ipv6.org" <yoshfuji@linux-ipv6.org>,
        "dsahern@kernel.org" <dsahern@kernel.org>,
        "kuba@kernel.org" <kuba@kernel.org>,
        Linux Kernel Network Developers <netdev@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Apr 15, 2021 at 5:47 PM Gong, Sishuai <sishuai@purdue.edu> wrote:
>
> Hi,
>
> We found a data race between tcp_set_default_congestion_control() and tcp=
_set_congestion_control() in linux-5.12-rc3.
> In general, when tcp_set_congestion_control() is reading ca->flags with a=
 lock grabbed, tcp_set_default_congestion_control()
> may be updating ca->flags at the same time, as shown below.
>
> When the writer and reader are running parallel, tcp_set_congestion_contr=
ol()=E2=80=99s control flow
> might be non-deterministic, either returning a -EPERM or calling tcp_rein=
it_congestion_control().
>
> We also notice in tcp_set_allowed_congestion_control(), the write to ca->=
flags is protected by tcp_cong_list_lock,
> so we want to point it out in case the data race is unexpected.
>
> Thread 1                                                        Thread 2
> //tcp_set_default_congestion_control()  //tcp_set_congestion_control()
>                                                                 // lock_s=
ock() grabbed
>                                                                 if (!((ca=
->flags & TCP_CONG_NON_RESTRICTED) || cap_net_admin))
>                                                                         e=
rr =3D -EPERM;
>                                                                 else if (=
!bpf_try_module_get(ca, ca->owner))
>                                                                         e=
rr =3D -EBUSY;
>                                                                 else
>                                                                         t=
cp_reinit_congestion_control(sk, ca);
> ca->flags |=3D TCP_CONG_NON_RESTRICTED;
>
>
>
> Thanks,
> Sishuai
>

Yes, obviously reading ca->flags while another thread might set the bit is =
racy.

This is of no consequence, if you want to silence KCSAN please a patch.