00:10 <GitHub4> [sinara] marmeladapk pushed 1 new commit to master: https://github.com/m-labs/sinara/commit/36d7a16af4999949cb9ca741c35a12f87cf44c90

00:10 <GitHub4> sinara/master 36d7a16 marmeladapk: Kasli: Changed 3V3 rail to 3,2 A maximum....

07:51 <sb0> hm, the media converter won't talk to the ethernet switch

07:51 <sb0> the rj45 link led stays off

09:25 <GitHub162> [ionpak] sbourdeauducq pushed 2 new commits to master: https://github.com/m-labs/ionpak/compare/916e9407803b...34ae0901aee5

09:25 <GitHub162> ionpak/master 30746d0 Sebastien Bourdeauducq: update smoltcp

09:25 <GitHub162> ionpak/master 34ae090 Sebastien Bourdeauducq: README: add note about firewall

10:07 <sb0> okay, the new smoltcp works a lot better on the mcu :)

10:44 <sb0> rjo, should I replace the openocd in /usr/local/bin with the m-labs fork that has KU support?

11:11 <sb0> but nothing works anyway, because of the jtag bug when the rtm is not present

11:11 <sb0> sigh

11:12 <sb0> and OF COURSE I still don't have RTMs

11:13 <sb0> and power cycling doesn't seem to help

11:14 <sb0> ffs

11:21 <sb0> managed to load another board

11:22 <sb0> that media converter doesn't work at all, no link on rj45, no link on sfp

11:24 <sb0> aah

11:24 <sb0> it wants 1G

11:24 <sb0> and the router is 100M

11:25 <sb0> okay now only the SFP end is broken

15:49 <GitHub26> [artiq] sbourdeauducq commented on issue #828: @whitequark Is this the proper solution:... https://github.com/m-labs/artiq/issues/828#issuecomment-328285422

16:54 <rjo> sb0: don't replace it yet.

16:55 <rjo> sb0: also we don't have openocd packages yet.

16:55 <rjo> sb0: what bug when the RTM is not there? with some openocd TCL magic, one can probably detect RTM presence and adjust the taps and DR_LENGTH setting.

16:56 <rjo> _florent_: could you give us a quick update on jesd204b?

20:19 <whitequark> sb0: a lot better?

20:19 <whitequark> but not perfect?

20:20 <whitequark> sb0: ohhh I just realized something

20:20 <whitequark> the reason I sometimes have trouble reproducing cjbe's results is that our router is indeed 100M, and an ACK roundtrip is slower than it would be on 1G

20:20 <whitequark> well anyway I can reproduce it on the simulator on the host, so that's fine

20:24 <GitHub15> [artiq] whitequark commented on issue #828: @sbourdeauducq This is the proper solution. Likely what happened is LLVM decided to not inline or not unroll something, and an expression that was previously constant-folded is not, anymore. This should be backported to ARTIQ-2. https://github.com/m-labs/artiq/issues/828#issuecomment-328301425