Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <1f91fe12-f9ff-06c8-4a5b-52dc21e6df05@linux.dev>
Date:   Wed, 26 Jul 2023 16:33:40 -0700
MIME-Version: 1.0
Subject: Re: [syzbot] [bpf?] WARNING: ODEBUG bug in tcx_uninstall
Content-Language: en-US
From:   Martin KaFai Lau <martin.lau@linux.dev>
To:     Leon Romanovsky <leon@kernel.org>
Cc:     Jakub Kicinski <kuba@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        syzbot <syzbot+14736e249bce46091c18@syzkaller.appspotmail.com>,
        andrii@kernel.org, ast@kernel.org, bpf@vger.kernel.org,
        haoluo@google.com, john.fastabend@gmail.com, jolsa@kernel.org,
        kpsingh@kernel.org, linux-kernel@vger.kernel.org,
        netdev@vger.kernel.org, sdf@google.com, song@kernel.org,
        syzkaller-bugs@googlegroups.com, yhs@fb.com,
        Gal Pressman <gal@nvidia.com>
References: <000000000000ee69e80600ec7cc7@google.com>
 <91396dc0-23e4-6c81-f8d8-f6427eaa52b0@iogearbox.net>
 <20230726071254.GA1380402@unreal> <20230726082312.1600053e@kernel.org>
 <20230726170133.GX11388@unreal>
 <896cbaf8-c23d-e51a-6f5e-1e6d0383aed0@linux.dev>
In-Reply-To: <896cbaf8-c23d-e51a-6f5e-1e6d0383aed0@linux.dev>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 7/26/23 11:16 AM, Martin KaFai Lau wrote:
> On 7/26/23 10:01 AM, Leon Romanovsky wrote:
>> On Wed, Jul 26, 2023 at 08:23:12AM -0700, Jakub Kicinski wrote:
>>> On Wed, 26 Jul 2023 10:12:54 +0300 Leon Romanovsky wrote:
>>>>> Thanks, I'll take a look this evening.
>>>>
>>>> Did anybody post a fix for that?
>>>>
>>>> We are experiencing the following kernel panic in netdev commit
>>>> b57e0d48b300 (net-next/main) Merge branch '100GbE' of 
>>>> git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue
>>>
>>> Not that I know, looks like this is with Daniel's previous fix already
>>> present, and syzbot is hitting it, too :(
>>
>> My naive workaround which restored our regression runs is:
>>
>> diff --git a/kernel/bpf/tcx.c b/kernel/bpf/tcx.c
>> index 69a272712b29..10c9ab830702 100644
>> --- a/kernel/bpf/tcx.c
>> +++ b/kernel/bpf/tcx.c
>> @@ -111,6 +111,7 @@ void tcx_uninstall(struct net_device *dev, bool ingress)
>>                          bpf_prog_put(tuple.prog);
>>                  tcx_skeys_dec(ingress);
>>          }
>> -       WARN_ON_ONCE(tcx_entry(entry)->miniq_active);
>> +       tcx_miniq_set_active(entry, false);
> 
> Thanks for the report. I will look into it.

I don't see how that may be triggered for now after Daniel's recent fix in 
commit dc644b540a2d ("tcx: Fix splat in ingress_destroy upon tcx_entry_free"). 
Do you have a small reproducible case? Thanks.

> 
>>          tcx_entry_free(entry);
>>   }
>>
> 
>