Attendees: rrs@, rscheff@, tuexen@, jtl@
Discussion
PR 252913: tuexen@ got a tracefile and some information from Netflix; impacts 12, waiting for 13 to be released.
D23230: ECN++, rscheff@ will investigate the sysctl interface. Still in progress.
D28702: Planned for main, mid March. Still in progress, waiting for 13 release. rrs@ will review.
Lost retransmission detection: D28931 is an idea for doing SACK loss recovery in the base stack. We will plan to discuss the I-D next meeeting.
PR 254333: Fixed by clean up of host cache code.
D29515: Proposal to switch hash function used for host cache. rscheff@ will update to use a random salt and allow an administrator to re-salt with a sysctl
Support for TCP over UDP is up for review and is planned to be committed to main after 13.0 is released. tuexen@ will do this.
- RACK will be updated after TCP over UDP support has been committed. rrs@ will do this.
D29411: PRR for ECN and congestion window reduction with NewReno. No plans to MFC.
D29576: glebius@ proposes changes to SYN cache to avoid serializing on acquiring a write lock for the listen socket.
- We had a discussion on the use of sockaddr_in on IPv6 sockets and mapped IPv4 addresses. We agreed to look at the code and try to standardize on a set of behavior for dealing with IPv4 addresses on IPv6 sockets which are not marked as IPv6-only. We will discuss this further in two weeks.
254725: LRO panic. Peer ACKs up to last data byte. Peer SACKs FIN. snd_una is correct, but send buffer is empty. Not able to reproduce artificially. rscheff@ will ask reporter to try to reproduce.
- tuexen@ reports on issue reported by NFS developer. We send a challenge ACK. Remote side sends a RST. We don't process the RST. The developer will re-run the test with debug messages enabled.
Next meeting: 22 April @ 1400 UTC