01:59 <GyrosGeier> hi

02:00 <GyrosGeier> can I send an SVF file over a network socket?

02:00 <GyrosGeier> basically, I have a gitlab pipeline that generates an SVF for my FPGA

02:00 <GyrosGeier> now I'd like to add a pipeline stage where the FPGA is programmed

02:01 <GyrosGeier> I can define a "service" in gitlab, i.e. a demand-started docker container that has a specific hardware config (allowing programmer access) and exposed ports

02:03 <GyrosGeier> my two blockers right now are that the server does not accept commands (immediately closes the connection after accepting), and that sending "svf test.svf" accesses the local file system

06:28 <PaulFertser> GyrosGeier: hi. Interesting task.

06:28 <PaulFertser> GyrosGeier: server closing connection should be investigated separately. You should check -d3 log for starters.

06:29 <PaulFertser> GyrosGeier: I think you can use Tcl commands to save SVF file locally.

14:54 <hanetzer> heyo. doing a bit of research. I currently own a flyswatter2 (ftdi based jtag/serial) and a knock off jlink (forget exact model). Was wondering if anyone had experience programming avr chips with these and openocd (specfically the atmega32u4 family)

14:58 <antto> i think there was some support for flashing atmegas, no idea at what state, but i wonder why you wouldn't use avrdude

15:01 <PaulFertser> hanetzer: yes, I used JTAG to flash atmega128 once. But why not avrdude and ISP (SPI) interface?

15:04 <hanetzer> PaulFertser: because I already have the other equipment :P

15:04 <hanetzer> I do not have any of the atmega ecosystem hardware

15:05 <PaulFertser> hanetzer: this flyswatter2 can work with avrdude , probably with some minimal tweaking to enable the output buffer.

15:08 <PaulFertser> hanetzer: yes, I see how it can be done without any code changes, just appropriate avrdude config.

15:09 <PaulFertser> hanetzer: found for you http://savannah.nongnu.org/bugs/?47043

15:11 <hanetzer> PaulFertser: <3

15:17 <hanetzer> now I just need to see what pins map to what

15:26 <PaulFertser> hanetzer: man avrdude tells

15:26 <hanetzer> ah

15:26 <PaulFertser> hanetzer: but really you should just open ARM jtag pinout

15:27 <PaulFertser> hanetzer: clock is clock, tdo/tdi is miso/mosi, tms is /cs

15:27 <hanetzer> oh, that makes it easier, then. I was confused because pins don't number from 0. What is buff, then?

15:27 <PaulFertser> hanetzer: pin controlling output buffer power

15:29 <hanetzer> honestly don't know what to make of that, I'm still a relative amature at this stuff

15:39 <PaulFertser> hanetzer: you do not need to, as I provided you to a config someone else has already prepared and tested.

15:40 <hanetzer> well I need to know what to hook up physically

15:43 <PaulFertser> hanetzer: told you, it's 4 wires.

15:43 <PaulFertser> hanetzer: TMS to /RESET, TCK to SCK, TDI to MOSI, TDO to MISO.

15:43 <PaulFertser> And GND

15:49 <hanetzer> huh. aight.

16:01 <hanetzer> well in that case, I just need to run those out to pin headers or the tag connect header :)

16:18 <hanetzer> I assume avrdude, using these isp lines, can burn bootloaders to 'empty' chips?

16:27 <urja> yup

16:29 <hanetzer> excellent. many thanks. tbh I'd rather use some kinetis chips but you can't fucking find them anywhere right now :/

17:21 <PaulFertser> hanetzer: how about gd32 or other stm32 clones? :)

17:30 <hanetzer> PaulFertser: heh, yeah there's a lot of mcu options.