16:47 <abg> is there a step-by-step guide for newbies to use openocd and jlink for nrf52?

16:47 <PaulFertser> abg: hi

16:47 <abg> hello

16:47 <PaulFertser> abg: there's the official user manual. Do you have any specific questions?

16:50 <abg> so i can connect to the chip and run some commands (i think). now, i tried running this - $OPENOCD/bin/openocd -s $OPENOCD/share/openocd/scripts -f interface/jlink.cfg -c 'transport select swd' -f target/nrf52.cfg -c 'program target/thumbv7em-none-eabihf/release/rcfw.hex verify reset exit'

16:50 <abg> and the output led me to believe everything was fine, but the code was not running (no io no nothing)

16:50 <PaulFertser> abg: that looks a bit odd as it's assuming you didn't install OpenOCD the normal way but should work ok.

16:51 <PaulFertser> abg: if you're using softdevice functions you probably needed to flash the softdevice binary too? Or are you using NimBLE stack or something else like that?

16:51 <abg> i then tried the command 'nrf5 mass_erase', printed a bunch of stuff, now the last line printed is "Info : SWD DPIDR 0x2ba01477" and does nothing more :(

16:52 <abg> i installed it a bit differently yes cuz i also have some other rpi2040 fork of openocd. i am not using any softdevice functions, dunno what that means but i assume some kind of bootloader?

16:53 <abg> i only want to use the jlink for programming

16:54 <abg> i googled softdevice. no, the firmware is some rust code

16:54 <abg> based on https://github.com/nrf-rs

16:57 <abg> the code atm is very simple, just prints some stuff on a uart. recently switched devkits to the dongle pcb and want to get stuff up and running. i am pretty sure the code is fine since i can run it on another devkit (where i program stuff with some bootloader and pyocd, no jlink)

16:58 <abg> if i run openocd with the --debug flag, i see this line: Debug: 193 29 command.c:590 run_command(): Command 'init' failed with error code -4

16:58 <abg> but i cannot figure out what that means, or if related

16:59 <abg> and again, all this started happening after running nrf5 mass_erase

17:02 <PaulFertser> abg: please use some pastebin to show what happens now if you use the "program" command again.

17:05 <abg> here's two outputs, one without and one with --debug flag: https://pastebin.com/Y4uJGLcH

17:27 <PaulFertser> abg: I see. Hm, what happens if instead of program you do "adapter speed 100; init; reset halt"?

17:30 <abg> oh this is good! it goes further and i can telnet

17:30 <abg> similar behavior as before running mass_erase

17:30 <abg> what does this do?

17:32 <abg> found it in docs, and i see the value before would be 1000 from target/nrf52.cfg

17:42 <PaulFertser> abg: my guess is that 1000 is too much somehow for your configuration.

17:42 <PaulFertser> abg: so you can try to "program" again with this lower speed.

17:43 <abg> yeah programming seems to work, but still the weird behavior. i'll take a look at the binary and datasheet, at least i know that the jlink is not the problem anymore

17:43 <abg> thank you paul :)

17:43 <PaulFertser> abg: good luck! I want to also remind you that you can actually do step-by-step debugging with OpenOCD.

17:44 <PaulFertser> abg: connect gdb and debug your firmware from reset vector.

17:45 <PaulFertser> abg: probably you just have wrong memory map and your MCU has less than expected?

17:46 <PaulFertser> abg: I mean the amount of SRAM

17:47 <abg> the build system assumes an nrf52840 specifically so not sure, but I'm going for the gdb approach indeed

17:48 <PaulFertser> abg: you can try "flash info 0" and it might show your MCU model and probably SRAM size.

19:37 <PaulFertser> abg: any news?

19:46 <abg> still poking, i started from scratch on a new cargo project. i'm pretty sure the problem is with the rust side of things, in gdb i see the correct code being executed (at least same as in ida). i am trying to do sanity checks

19:50 <PaulFertser> abg: I see, hope you figure it out soon.