01:39 <mithro> ysionneau: ping?

02:00 <mithro> actually I think I need rjo ?

02:04 <mithro> sb0: Hey - you were talking about migen programmers and openocd. Is openocd an actually actively developed jtag programmer? IE Should we be making sure the Opsis is well supported using it?

02:11 <sb0> mithro, Robert added FPGA support to openocd, including flash writing via proxy bitstream. I'd say yes

02:11 <mithro> sb0: It looks like openocd already has support for Altera USB Blasters? We have firmware for our FX2 which pretends to be one of them for urjtag

02:12 <sb0> looks like it... let us know how it goes

02:18 <mithro> sb0: I'll give it a go in an hour or two probably

02:41 <sb0> rjo, https://www.apexanalog.com/apex-products/pa98/ "Very High Slew Rate 1000V/µs, High Voltage 450V Power Amplifier"

02:52 <sb0> https://www.apexanalog.com/apex-products/pa107/ may be even better

04:02 <mithro> sb0: what timezone is rjo in?

04:35 <mithro> openocd desperately needs a better website

08:18 <cr1901_modern> openocd's meritocracy mentality concerns me a bit (though sb0 says we could fork if it became a problem)

14:38 <cr1901_modern> sb0: https://github.com/m-labs/migen/blob/new/migen/genlib/record.py#L147 Is there a particular reason you use reduce(or_, etc) instead of using optree to or the slaves together?

16:10 <whitequark> sb0: what did you recommend pa98 to rjo for?

16:10 <whitequark> well. i was looking for a power amplifier for a quadrupole ion guide

16:12 <whitequark> haha wow pa107 is 413@1

16:12 <whitequark> but the gain-bandwidth product is sooooo tasty

16:12 <whitequark> and power bandwidth of 2MHz, too

19:03 <mithro> rjo: ping?

20:23 <rjo> mithro: pong

20:24 <rjo> sb0: pa98 and relatives have less than a 20th of the slew rat I get from my crt drivers. and less than a 7th the slew rate of a good opamp.

20:25 <rjo> sb0: they are good for dust traps and piezo drivers.

20:26 <rjo> but they have a beryllium substrate and thus are a bit of a pain to dispose of. and their failure mode is a short circuit...

20:27 <rjo> mithro: yes. i wrote strace_tailor.sh ysionneau figured out that a fiew files need relocating still.

20:28 <mithro> rjo: Did he document that anywhere?

20:31 <rjo> the code is pretty explicit.

20:32 <rjo> https://github.com/m-labs/artiq/blob/master/.travis/get-xilinx.sh

20:48 <rjo> mithro: is there something missing?

20:49 <mithro> rjo: well, I can't use your packages because I don't know the passphrase :) - I would also like to make sure I can generate my own

20:50 <rjo> strace_tailor.sh has documentation how to use it in the file. does that not work for you?

20:51 <rjo> i can imagine xilinx being anal enough to "dislike" distributing unencrypted licenses and software...

20:55 <mithro> rjo: yes, they are

21:18 <mithro> rjo / ysionneau: I'm just rewriting the strace_tailor.sh script in Python to fix up some stuff it misses

21:18 <mithro> bblr

21:34 <rjo> mithro: what does it miss?

22:57 <felix_> i'm currently writing lm32 support for radare2; thought that might be interesting for you

22:59 <felix_> the lm32 isa is definitely one of the better ones

23:04 <mithro> felix_: what is radare2?

23:04 <mithro> felix_: seems to be some sort of text editor?

23:06 <felix_> its a tool like ida pro

23:06 <felix_> quite powerful disassembler

23:09 <mithro> ahh

23:12 <felix_> i'll also write a code analyzer plugin for it, but i'm quite happy that at least the disassembler plugin sort of works (still with at least a few bugs)

23:12 <rjo> felix_: can i haz radare-or1k?

23:13 <felix_> rjo: you could maybe hire me as a consultant...

23:14 <rjo> interesting!

23:15 <felix_> i'm writing the support for lm32, cause i need that for some reverse-engineering project; so that was a spare-time project for me

23:15 <rjo> ouh. there are closed software things for lm32?

23:23 <felix_> no, ida pro currently doesn't have support for lm32

23:23 <felix_> if it had, i'd have used that

23:23 <felix_> ah, you mean designs

23:23 <felix_> yes, lm32 is used on some asics

23:30 <mithro> rjo: I ended up with the following -> https://github.com/timvideos/HDMI2USB-misoc-firmware/blob/travis-xilinx/.travis/package-xilinx.sh and https://github.com/timvideos/HDMI2USB-misoc-firmware/blob/travis-xilinx/.travis/package-xilinx-filter-strace.py

23:32 <felix_> e.g. amd uses a lm32 for power, clock, and thermal management; there was a talk on that subject at the 30c3 ;)

23:34 <mithro> rjo: still trying to figure out why it likes the xc6slx45 but not the xc6slx45t

23:35 <rjo> mithro: what did the original script miss? it has worked fine for us on several S3, several S6 and zynq/kintex.

23:37 <mithro> rjo: I'm unsure exactly now - it complained a lot about missing files when doing the cp and I got a couple "file not found" errors on the travis output

23:39 <mithro> rjo: Some of it was also the sed command not dealing with double slashes ending up in the path names - so trying to copy the file twice

23:39 <rjo> did you run strace_tailor.sh for that very target chip?

23:40 <rjo> copying the same file twice is very harmless.

23:40 <mithro> rjo: I'll be interested to compare the tar generated with the python script verse the strace_tailor.sh script

23:42 <mithro> rjo: I'm more confident now that things are working as expected now there are no complaints from the cp

23:43 <mithro> rjo: I'm also interested to see which commands are accessing which files - which the Python script will be easy to extend to do