Discussion
- Any update on blackhole detection?
- No update on implementation within our tree.
- Gorry may be interested in improving the RFCs on blackhole detection.
- Update on the interesting problem with packet lookup returning the wrong TCB which tuexen@ reported at the last call. This is a regression due to epoch.
D18954, D18982, and D19118 are now committed.
One patch got missed during refactoring: after idle, ssthresh is the max of former ssthresh or 3/4 of the former cwnd. This is in D22438. rrs@/jtl@ recommend a sysctl and/or socket option so people can test with this and then decide whether to use it.
Richard has Accurate ECN support refactored into several reviews: D22426, D22428, D22429, D22436. rrs@/jtl@ request that Accurate ECN bits move to t_flags2 instead of using spare bits in t_flags. (It will conflict with other patches in progress.)
Next Meeting
Next meeting is 5 Dec 2019 @ 1400 UTC.