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
In-Reply-To: <76483220-226b-bf3c-d529-ea839d6624c6@iogearbox.net>
References: <000000000000c64842056e4b4326@google.com> <76483220-226b-bf3c-d529-ea839d6624c6@iogearbox.net>
From:   Dmitry Vyukov <dvyukov@google.com>
Date:   Mon, 11 Jun 2018 11:52:00 +0200
Message-ID: <CACT4Y+boO990R6a9kGtx-UZpZPeB2bwhgvyUjPvEFOpbzxJKrg@mail.gmail.com>
Subject: Re: KASAN: slab-out-of-bounds Read in bpf_skb_change_proto
To:     Daniel Borkmann <daniel@iogearbox.net>
Cc:     syzbot <syzbot+d2d729bdde65dee3eae6@syzkaller.appspotmail.com>,
        Alexei Starovoitov <ast@kernel.org>,
        David Miller <davem@davemloft.net>,
        LKML <linux-kernel@vger.kernel.org>,
        netdev <netdev@vger.kernel.org>,
        syzkaller-bugs <syzkaller-bugs@googlegroups.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Jun 11, 2018 at 11:42 AM, Daniel Borkmann <daniel@iogearbox.net> wrote:
> On 06/10/2018 05:27 PM, syzbot wrote:
>> Hello,
>>
>> syzbot found the following crash on:
>>
>> HEAD commit:    a16afaf7928b Merge tag 'for-v4.18' of git://git.kernel.org..
>> git tree:       upstream
>> console output: https://syzkaller.appspot.com/x/log.txt?x=1338f6bf800000
>> kernel config:  https://syzkaller.appspot.com/x/.config?x=314f2150f36c16ca
>> dashboard link: https://syzkaller.appspot.com/bug?extid=d2d729bdde65dee3eae6
>> compiler:       gcc (GCC) 8.0.1 20180413 (experimental)
>> syzkaller repro:https://syzkaller.appspot.com/x/repro.syz?x=1173381f800000
>> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=171f90cf800000
>>
>> IMPORTANT: if you fix the bug, please add the following tag to the commit:
>> Reported-by: syzbot+d2d729bdde65dee3eae6@syzkaller.appspotmail.com
>
> #syz fix: bpf: reject passing modified ctx to helper functions


On a related note, it seems that it still can unwind past
bpf_skb_change_proto. I think the "net.core.bpf_jit_kallsyms = 1"
sysctl should have been reached syzbot by the time of crash. Are you
sure that's the only thing requires? We are using frame pointer
unwinder just in case.