Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <0000000000005e47a2057d0edc49@google.com> <20181216190412.GE4170@linux.ibm.com>
 <CACT4Y+b+mSs2FV3BY+MsHDTPPgvPaCdMmJONd9c4zEU67CdmOA@mail.gmail.com>
 <20181217112916.GG4170@linux.ibm.com> <1583d5fc-34bf-3a81-363d-01a1085a7363@linux.intel.com>
 <CACT4Y+aKn4XsGVgt3ELGnnJcg6H-m+dSdV018dLvJhrpshycWA@mail.gmail.com>
 <20641819-e4fb-f3bd-34c8-c68106cccd0e@gmail.com> <20181217162421.6d636ee5@redhat.com>
 <CACT4Y+YXFrnQh9h1KQf8=NcHugmxTj+d7AQdGCvRX_xsHwr1qA@mail.gmail.com>
 <20181218001828.49cea463@redhat.com> <CACT4Y+ZFpO26Tgqpi+Jies9XOiiCgAdp2DV0z_54hCNTqkJsiw@mail.gmail.com>
 <20181218134024.45d2d5e3@redhat.com>
In-Reply-To: <20181218134024.45d2d5e3@redhat.com>
From:   Cong Wang <xiyou.wangcong@gmail.com>
Date:   Tue, 18 Dec 2018 20:12:04 -0800
Message-ID: <CAM_iQpUYeF1vLRX=TryfH6ruPi2qL-j8Y2shbs7s9cKiy4xvmA@mail.gmail.com>
Subject: Re: WARNING in __rcu_read_unlock
To:     Stefano Brivio <sbrivio@redhat.com>
Cc:     Dmitry Vyukov <dvyukov@google.com>,
        Eric Dumazet <eric.dumazet@gmail.com>,
        Arjan van de Ven <arjan@linux.intel.com>,
        "Paul E. McKenney" <paulmck@linux.ibm.com>,
        syzbot <syzbot+43f6755d1c2e62743468@syzkaller.appspotmail.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Josh Triplett <josh@joshtriplett.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Ingo Molnar <mingo@kernel.org>,
        syzkaller-bugs <syzkaller-bugs@googlegroups.com>,
        netdev <netdev@vger.kernel.org>, Xin Long <lucien.xin@gmail.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Dec 18, 2018 at 4:40 AM Stefano Brivio <sbrivio@redhat.com> wrote:
> Not as far as I know. The selftests checking this path, by design, only
> use supported configurations, they don't forge packets.
>
> Maybe it would be nice to have a semi-automated way to isolate and
> describe/name specific conditions found by syzbot via fuzzing and turn
> those into tests that are then repeated periodically. I'm not sure how
> that would look like, but I think it's still more maintainable than a
> pile of C reproducers with forged packets in selftests/net.
>
> Eric, Cong, Xin, as you also recently fixed a nice deal of similar cases
> reported by syzbot, what do you think? Did you ever feel the need to
> turn a syzbot reproducer into a regression test case?

I think it is a very good idea to archive these C reproducers in
tools/testing/selftests/. After this is done, kbuild bot could do these
regression tests, no longer need to bother syzbot to run them again. :)

Thanks.