00:18 <GitHub82> [smoltcp] barskern opened pull request #225: Fixes bug regarding dup_acks overflow (TcpSocket) (master...bug/fix-dup-acks-overflow) https://github.com/m-labs/smoltcp/pull/225

00:25 <GitHub159> [smoltcp] whitequark commented on issue #224: > This would trigger a fast retransmit for every segment which is a duplicate ACK when we have recived more than 3 duplicate ACKs.... https://github.com/m-labs/smoltcp/issues/224#issuecomment-394041499

00:44 <sb0> _florent_, can you look into the JESD problem?

00:44 <sb0> whitequark, not really, and fixing the compiler is more important

06:35 <_florent_> sb0: i'll at the stpl test at startup on sayma monday

09:39 <GitHub55> [smoltcp] barskern commented on issue #224: > Per my understanding, with the modified code, if e.g. first packet in a burst of 8 gets lost, the other 7 will arrive, and 4 of them will trigger retransmit of the entire burst. Is that correct?... https://github.com/m-labs/smoltcp/issues/224#issuecomment-394073919

10:02 <GitHub158> [smoltcp] barskern commented on issue #224: After some more research, I found that what should happen after a duplicate ACK is recived is that the socket should enter "Fast recovery" and every following duplicate ACK only increases the congestion window. ... https://github.com/m-labs/smoltcp/issues/224#issuecomment-394075274

10:25 <GitHub117> [smoltcp] barskern commented on issue #224: After some more research, I found that what should happen after three duplicate ACKs is recived is that the socket should enter "Fast recovery" and every following duplicate ACK only increases the congestion window. ... https://github.com/m-labs/smoltcp/issues/224#issuecomment-394075274

14:03 <sb0> hartytp, what vivado version are you using?

16:09 <GitHub-m-labs> [artiq] sbourdeauducq commented on issue #1045: Bounds checking tends to reduce programming speed. https://github.com/m-labs/artiq/issues/1045#issuecomment-394098005

16:09 <GitHub-m-labs> [artiq] sbourdeauducq commented on issue #1045: Other drivers don't have bounds checking either for this reason. https://github.com/m-labs/artiq/issues/1045#issuecomment-394098036

16:57 <GitHub139> [smoltcp] whitequark commented on issue #225: This seems OK but please add a test. You can set `local_rx_dup_acks` to 254 manually. https://github.com/m-labs/smoltcp/pull/225#issuecomment-394101502

17:00 <GitHub-m-labs> [artiq] whitequark commented on issue #1045: How about having a "debug mode" with bounds checking and "release mode" without it? Python even already has it, it's called the assert statement, and it's turned off if you pass `-O` or set `PYTHONOPTIMIZE=1`. As a bonus this will make the compiler faster (though IIRC I didn't use any particularly expensive asserts). https://github.com/m-labs/artiq/issues/1045#is