linux-kernel - Re: [PATCH net] vsock: Ignore signal/timeout on connect() if already established

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <69a3d223-dd95-43df-af3a-522968d6b850@rbox.co>
Date: Sun, 23 Nov 2025 22:46:22 +0100
From: Michal Luczaj <mhal@...x.co>
To: Stefano Garzarella <sgarzare@...hat.com>
Cc: "David S. Miller" <davem@...emloft.net>,
 Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>,
 Paolo Abeni <pabeni@...hat.com>, Simon Horman <horms@...nel.org>,
 virtualization@...ts.linux.dev, netdev@...r.kernel.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH net] vsock: Ignore signal/timeout on connect() if already
 established

On 11/21/25 10:21, Stefano Garzarella wrote:
> On Thu, Nov 20, 2025 at 10:12:20PM +0100, Michal Luczaj wrote:
>> On 11/19/25 20:52, Michal Luczaj wrote:
>>> ...
>>> To follow up, should I add a version of syzkaller's lockdep warning repro
>>> to vsock test suite? In theory it could test this fix here as well, but in
>>> practice the race window is small and hitting it (the brute way) takes
>>> prohibitively long.
>>
>> Replying to self to add more data.
>>
>> After reverting
>>
>> f7c877e75352 ("vsock: fix lock inversion in vsock_assign_transport()")
>> 002541ef650b ("vsock: Ignore signal/timeout on connect() if already
>> established")
>>
>> adding
>>
>> --- a/tools/testing/vsock/vsock_test.c
>> +++ b/tools/testing/vsock/vsock_test.c
>> @@ -2014,6 +2014,7 @@ static void test_stream_transport_change_client(const
>> struct test_opts *opts)
>>                        perror("socket");
>>                        exit(EXIT_FAILURE);
>>                }
>> +               enable_so_linger(s, 1);
>>
>>                ret = connect(s, (struct sockaddr *)&sa, sizeof(sa));
>>                /* The connect can fail due to signals coming from the
>>
>> is enough for vsock_test to trigger the lockdep warning syzkaller found.
>>
> 
> cool, so if it's only that, maybe is worth adding.

Ok, there it is:
https://lore.kernel.org/netdev/20251123-vsock_test-linger-lockdep-warn-v1-1-4b1edf9d8cdc@rbox.co/

And circling back to [1], let me know if you think it's worth adding to the
suit. I guess it would test the case #2 from [2], but it'd take another 2s
and would require both h2g and non-h2g transports enabled.

[1]:
https://lore.kernel.org/netdev/fjy4jaww6xualdudevfuyoavnrbu45cg4d7erv4rttde363xfc@nahglijbl2eg/
[2]:
https://lore.kernel.org/netdev/20251119-vsock-interrupted-connect-v2-1-70734cf1233f@rbox.co/