[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ca59f955-dc6f-49d8-ae32-fb2d0f7e7522@gmail.com>
Date: Thu, 7 Dec 2023 10:46:35 -0800
From: Kui-Feng Lee <sinquersw@...il.com>
To: Eric Dumazet <edumazet@...gle.com>
Cc: patchwork-bot+netdevbpf@...nel.org, Kui-Feng Lee <thinker.li@...il.com>,
davem@...emloft.net, kuba@...nel.org, pabeni@...hat.com, dsahern@...nel.org,
netdev@...r.kernel.org, eric.dumazet@...il.com
Subject: Re: [PATCH net-next] ipv6: add debug checks in fib6_info_release()
On 12/7/23 10:22, Eric Dumazet wrote:
> On Thu, Dec 7, 2023 at 7:19 PM Kui-Feng Lee <sinquersw@...il.com> wrote:
>>
>>
>>
>> On 12/7/23 10:10, Eric Dumazet wrote:
>>> On Thu, Dec 7, 2023 at 7:06 PM Kui-Feng Lee <sinquersw@...il.com> wrote:
>>>
>>>> Do you happen to have a test program that can reproduce it?
>>>
>>> syzbot has a repro, let me release the bug.
>>>
>>> Of course syzbot bisection points to my last patch.
>>
>> I just looked into the code.
>> The origin issue mentioned at the thread head should be something
>> related to a GC change I made. But, the warnings you added doesn't
>> catch the the error correctly. According to your stacktrace
>>
>>
>> > ip6_route_add+0x26/0x1f0 net/ipv6/route.c:3843
>> > ipv6_route_ioctl+0x3ff/0x590 net/ipv6/route.c:4467
>> > inet6_ioctl+0x265/0x2b0 net/ipv6/af_inet6.c:575
>> > sock_do_ioctl+0x113/0x270 net/socket.c:1220
>> > sock_ioctl+0x22e/0x6b0 net/socket.c:1339
>> > vfs_ioctl fs/ioctl.c:51 [inline]
>> > __do_sys_ioctl fs/ioctl.c:871 [inline]
>> > __se_sys_ioctl fs/ioctl.c:857 [inline]
>> > __x64_sys_ioctl+0x18f/0x210 fs/ioctl.c:857
>> > do_syscall_x64 arch/x86/entry/common.c:51 [inline]
>> > do_syscall_64+0x40/0x110 arch/x86/entry/common.c:82
>> > entry_SYSCALL_64_after_hwframe+0x63/0x6b
>>
>> and warning messages you provided
>>
>> > WARNING: CPU: 0 PID: 5059 at include/net/ip6_fib.h:332
>> > fib6_info_release include/net/ip6_fib.h:332 [inline]
>> > WARNING: CPU: 0 PID: 5059 at include/net/ip6_fib.h:332
>> > ip6_route_info_create+0x1a1a/0x1f10 net/ipv6/route.c:3829
>>
>> It takes place in ip6_route_info_create() to do error handling.
>> It can be fib6_has_expires() in fib6_info_release() in this case.
>
> Feel free to amend the patch, but the issue is that we insert a fib
> gc_link to a list,
> then free the fi6 object without removing it first from the external list.
>
> I added two different warnings, and removing one or both will still
> keep the bug.
The gc_link is not inserted here actually. (see my explanation in
another message.)
According to the messages in the thread head, it is an issue of dangling
pointer, right? If I read it correctly, the original issue is gc_link
pointing to a block of memory that is already free. Am I right?
Powered by blists - more mailing lists