11:28 <kpedersen> zapb_: Many thanks. I will see what it uses (there isn't much documentation). Otherwise, I am assuming the company is just recommending keeping an ancient install of Windows around? Shame really.

12:54 <PaulFertser> kpedersen: have you checked if the adapter is not cmsis-dap/daplink compatible? Probably ask the company? Either way, you can get a reasonable JTAG adapter for real cheap from China these days.

13:26 <zapb_> Yes, I think it's not worth the effort to implement support for this adapter and IMO it is also questionable if we should accept a driver for such an old (and niche?) adapter

15:04 <clever> PaulFertser: ive been messing with the debug on the pi5 lately, and doing `halt ; reg ; continue` over telnet to openocd causes the target to crash, i assume that isnt supposd to happen? any idea why it might?

15:16 <PaulFertser> clever: just a crash, nothing on serial?

15:16 <PaulFertser> clever: what if you do not do "reg" there?

15:16 <clever> PaulFertser: serial wasnt hooked up, but there was an SPI based display connected, and dmesg was barfing a lot of errors, and ssh didnt respond

15:17 <PaulFertser> clever: so not just a total crash. And I assume you can't reproduce?

15:17 <clever> i havent tried reproducing yet

15:17 <clever> let me try now

15:17 <clever> did a halt, then continue, and now ssh is dead again

15:17 <PaulFertser> Probably it stopped in some really unusual state which was not handled by aarch64 code properly.

15:18 <clever> User : 19562 74978677 armv8.c:1316 armv8_aarch64_state(): bcm2712.cpu0 halted in AArch64 state due to debug-request, current mode: EL2H

15:18 <clever> cpsr: 0x604000c9 pc: 0xffffd0000cbf86bc

15:18 <clever> User : 19563 74978677 armv8.c:1348 armv8_arch_state(): MMU: enabled, D-Cache: enabled, I-Cache: enabled

15:18 <clever> User : 19564 74980896 command.c:601 command_run_line():

15:19 <clever> https://gist.github.com/cleverca22/5ad40b80fd4f6393b1e12f6889871253

15:19 <clever> it looks like openocd still decided to read every register, even without the reg command

15:20 <clever> no complaints on the spi display this time, but i understand that requires it to be fairly functional

15:28 <clever> PaulFertser: any other ideas?

15:56 <PaulFertser> clever: I'm afraid not, no ideas here.

22:34 <kpedersen> PaulFertser: zapb_: I have fired the company an email (its still alive). I will update you if they respond with anything useful. Its an old board (though we have 50+ of them laying around) and fundamentally it is annoying if they use some vendor specific way of uploading the hex code.

22:41 <PaulFertser> kpedersen: where's that Mikroe EasyARM schematics?

22:43 <PaulFertser> kpedersen: what vid/pid do you get when you attach that board's integrated debugger to USB?

22:45 <kpedersen> I have some ratty paper docs in the box. Online, is this useful:

22:45 <kpedersen> https://www.mouser.com/ds/2/272/easyarmv6_manual_v100-34167.pdf

22:45 <kpedersen> (I have a bunch of EasyPIC boards too if that is interesting).

22:46 <PaulFertser> kpedersen: this v6 shows regular ft232rl which is supported by OpenOCD.

22:46 <kpedersen> I don't get anything particulaly useful when I plug it in. Plugging it in to FreeBSD I get a generix ugen which doesn't attach to anything.

22:47 <kpedersen> Oh right. It doesn't attach as a uart / /dev/cuaU0 or anything like that. Perhaps a different version?

22:48 <PaulFertser> kpedersen: which vid pid though? And what do you see on the board itself visually observing the ICs there, what's connected to USB?

22:48 <kpedersen> And if that is just a serial interface, surely it has a specific protocol to send the hex code? (apologies a bit out of my depth here)

22:49 <kpedersen> PaulFertiser: vid / pid: Is that detected via udevadm? I will check that.

22:50 <PaulFertser> kpedersen: you see it in kernel messages

22:50 <PaulFertser> udevadm is Linux-specific and not needed here

22:51 <PaulFertser> kpedersen: do you have those small boards that they call mikroBoard? What does it have and how do you plan to develop for a dev board if you do not have schematics?

22:55 <kpedersen> Annoyingly I didn't bring a box home with me. I will update you on the vid / pid on Monday.

22:56 <PaulFertser> kpedersen: probably you just talk to target's UART directly and use its bootloader to flash with e.g. lpc21isp or lpc2k_pgm tools?

22:56 <kpedersen> Really I wanted to check if I could uplaod hex code *before* I put in time to research the hardware in any kind of detail. I have loads of sample code on a CD-ROM.

22:57 <PaulFertser> Well what's the target MCU for starters?

22:58 <kpedersen> OK. That is really useful. I was unaware of those tools.

22:59 <PaulFertser> Like stm32flash for ST parts.

22:59 <kpedersen> The MCU that comes attached to these boards is an LPC2148

23:00 <PaulFertser> And on the boards themselves?

23:01 <kpedersen> I was hoping there was an equivalent to avrdude but actually had no idea where to start. I assumed OpenOCD was an approach to this.

23:03 <kpedersen> Admittedly I am not sure. It has a bunch of "stuff" attached. Including the on chip flash programmer (ARM prog). If you are local around South England, I am happy to give you one (or a few ;)

23:04 <kpedersen> The crusty compiler itself is GCC based. Looks fairly easy to interact with the pins (and the provided LCD screen, etc).

23:05 <PaulFertser> avrdude is only for flashing. OpenOCD is for talking to the debug module inside the MCU or CPU.

23:05 <kpedersen> Once I was sure I could upload hex files easily I was going to tackle building a newer cross compiler.

23:05 <PaulFertser> Alas, not local at all.

23:06 <kpedersen> Am I correct in thinking that OpenOCD can also upload hex code?

23:06 <PaulFertser> Yes, but it does so by manipulating target registers directly and (optionally) running additional helpers on it.

23:07 <kpedersen> Ah I see (I was pointed towards this project by one of my colleagues as a starting point).

23:07 <PaulFertser> While with avrdude/stm32flash/lpc21isp you'd be talking to the ROM bootloader (embedded in the target by the vendor).

23:07 <PaulFertser> Of course having actual debug facilities is really nice and saves plenty of time compared to the old school recompile/reflash/test loop.

23:11 <kpedersen> Agreed. My first priority is getting code on there but it would be very slick if I could hook up GDB, etc.

23:14 <kpedersen> Thankyou for your advice so far. Very helpful. I will explore those projects for flashing it and will provide updates (if you are interested in this relatively old board).