Attendees: jtl@, kbowling@, thj@, rscheff@, tuexen@
Discussion
PR252624: tuexen@ can't reproduce it.
PR252913: tuexen@ got a tracefile and some information from Netflix.
PR238741: When using RACK in combination with CDG some TCP connections stall. tuexen@ and rrs@ will try to get a simple fix it.
PR253096: RACK does not support MD5/AO. tuexen@ will look into adding MD5 support adding to RACK. Adding AO support for the base stack and RACK is more complex. Will focus on MD5 first.
D23230: ECN++, rscheff@ will investigate the sysctl interface.
D28780: Committed to main, not stable/13 yet. Planned for releng13.0.
D28813: Fix related to PRR. Planned for releng13.0.
D28819: Another fix related to PRR. Planned for releng13.0.
D28702: Planned for main, mid March.
D28634: Fix dup acks triggering a fast retransmissions. Clearifying the SACK negotition.
D28652: Accepted.
D28686: Abandoned. Will be replaced by socket option level handling (tuexen@).
PR 253476: ipfw injecting TCP segments without timestamps even if they are required. The result was that this should be fixed in ipfw. Other implementation might also do strict checks. There is a sysctl variable to enforce less strict checks.
PR 253589: Report about a race. The race seems to be acceptable, fixing it would impact performance.
- There is a bug report on net@ about a bug TCP. rscheff@ looked into it, but has no explanation yet. Adrian might have encountered this bug, too.
Next meeting: 11 March @ 1400 UTC