LVS
lvs-devel
Google
 
Web LinuxVirtualServer.org

Re: [syzbot] [lvs?] possible deadlock in start_sync_thread

To: Pablo Neira Ayuso <pablo@xxxxxxxxxxxxx>, Wenjia Zhang <wenjia@xxxxxxxxxxxxx>, Jan Karcher <jaka@xxxxxxxxxxxxx>, Tony Lu <tonylu@xxxxxxxxxxxxxxxxx>, Wen Gu <guwen@xxxxxxxxxxxxxxxxx>, linux-s390@xxxxxxxxxxxxxxx
Subject: Re: [syzbot] [lvs?] possible deadlock in start_sync_thread
Cc: syzbot <syzbot+e929093395ec65f969c7@xxxxxxxxxxxxxxxxxxxxxxxxx>, coreteam@xxxxxxxxxxxxx, davem@xxxxxxxxxxxxx, edumazet@xxxxxxxxxx, horms@xxxxxxxxxxxx, ja@xxxxxx, kadlec@xxxxxxxxxxxxx, kuba@xxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, lvs-devel@xxxxxxxxxxxxxxx, netdev@xxxxxxxxxxxxxxx, netfilter-devel@xxxxxxxxxxxxxxx, pabeni@xxxxxxxxxx, syzkaller-bugs@xxxxxxxxxxxxxxxx
From: "D. Wythe" <alibuda@xxxxxxxxxxxxxxxxx>
Date: Wed, 6 Nov 2024 18:01:28 +0800


On 11/5/24 7:25 PM, Pablo Neira Ayuso wrote:
Hi,

I am Cc'ing SHARED MEMORY COMMUNICATIONS (SMC) SOCKETS maintainers.

Similar issue already reported by syzkaller here:

https://lore.kernel.org/netdev/ZyIgRmJUbnZpzXNV@calendula/T/#mf1f03a65108226102d8567c9fb6bab98c072444c

related to smc->clcsock_release_lock.

I think this is a false possible lockdep considers smc->clcsock_release_lock
is a lock of the same class sk_lock-AF_INET.

Could you please advise?

Thanks.

Hi Pablo,

Thank you for the reminder. We are currently working on it. The community has already submitted some fixes[1], and we are still reviewing them.


Best wishes,
D. Wythe


[1] https://lore.kernel.org/netdev/c2ac8e30806af319eb96f67103196b7cda22d562.1729031472.git.danielyangkang@xxxxxxxxx/





<Prev in Thread] Current Thread [Next in Thread>