Thanks for the extra details. I tried to reproduce it locally today, and I couldn't. While checking the code and your logs, I suspect this is IPv6-specific, and my testing env is IPv4-only. I'll come back when I manage to reproduce it. My current suspicion is:
Thanks for the extra details. I tried to reproduce it locally today, and I couldn't. While checking the code and your logs, I suspect this is IPv6-specific, and my testing env is IPv4-only. I'll come back when I manage to reproduce it. My current suspicion is:
commit e198bdd7ad2834c 77567d59158e883 63d2f0e05d pktoptions. /bugs.launchpad .net/bugs/ 2015399 3fc290073d5603a 755fd12eed upstream.
Author: Kuniyuki Iwashima <email address hidden>
Date: Thu Feb 9 16:22:01 2023 -0800
dccp/tcp: Avoid negative sk_forward_alloc by ipv6_pinfo.
BugLink: https:/
commit ca43ccf41224b02
This would confirm the above IPv6 theory.
PS: The warning is generated by WARN_ON( sk->sk_ forward_ alloc) in sk_stream_ kill_queues.