<wkoszek>
Hi folks. What's the best cable for OpenOCD? I'm looking at the repo for the most frequently changed files, and it's stlink and jlink.
<antto>
"cable" ?
<antto>
do you mean debugger device?
<wkoszek>
Yes.
<wkoszek>
JTAG dongle
<antto>
that probably depends on the targets you wanna deal with and the $$$$
<antto>
i deal with cortex-M mostly, kamami zl33prog is nice (needs a saner connector/adapter but i'm fine with that)
<wkoszek>
My friends made a custom board and I need 2x10 PIN connector with a ribbon cable and OpenOCD support - one CPU is Layerscape ARM and another one is 4-core RISC-V
<zear>
wkoszek, I think we have the same friends :)
<zear>
it definitely works with SAMA5D27 one board and I don't see why it wouldn't with the other ones
<wkoszek>
I need something eBayable :)
<antto>
tme never has it in stock, i bought it directly from kamami.pl
<zear>
I think the question is - what bandwidth do you need? The majority of cheaper probes are limited by USB 2.0's bandwidth, which is good enough for stepping through gdb
<wkoszek>
I think speed doesn't matter that much. I just need to flash this board.
<zear>
then anything will work, but keep in mind that, unless your flash maps directly into addressable memory in rw mode, you'll need some sort of flashing script/driver
<zear>
the proprietary jtag probes usually come with dedicated flashing firmwares for various boards
<zear>
where, under the hood, they will upload a small program into RAM, which in turn talks to the flash chip using QSPI/whatever
<PaulFertser>
You can often use U-Boot for that.
<zear>
I think wkoszek needs to flash U-Boot firmware :)
<PaulFertser>
I mean upload U-Boot to RAM and let it flash itself.
<zear>
the guys at work have a fixation on those Lauterbach probes, because they paid so much for them, they have to use them
<zear>
but I prefer to stick to free software
<wkoszek>
We have who is who of Polish embedded systems engineering here
<wkoszek>
I remember you from 15yrs ago when I visited Semihalf
<wkoszek>
I have 2 boards, and I need to get some servicing on them, and I'm in the US and I don't want to spend much $ on this stuff, as I'll likely debug them sporadically once I get Ubuntu on both of them
<zear>
in general openocd is the way to go
<wkoszek>
Yeah
<zear>
just keep in mind openocd uses those jtag probes as "dumb" JTAG/SWD pass-throughs
<zear>
the expensive proprietary probes usually have an FPGA w/ a lot of RAM inside, so they talk to the board directly
<zear>
and spit out results over USB/ethernet back to the debugger
<zear>
openocd doesn't work like that
<zear>
with openocd the entire processing is done on your computer's side and needs to fit in the probe's bandwidth
<zear>
so if you want to perform real time profiling, etc., openocd will not be a good choice
bvernoux has joined #openocd
<zear>
<wkoszek> I remember you from 15yrs ago when I visited Semihalf
<zear>
I did work for semihalf, but not 15 years ago
<zear>
so perhaps you're misremembering me :)
<wkoszek>
Maybe I'm overestimating the time. But I think we briefly met during my rare visits there. And then I remember you from FreeBSD patches.
<zear>
that latter part checks out :)
<zear>
(I wish that I eventually get back to doing that)
<wkoszek>
I think adding FreeBSD support to Conclusive boards could make sense.
<PaulFertser>
I have to note J-Link was nice enough and provided OpenOCD with some protocol documentation and their pricing seems to be reasonably fair so for business I would probably suggest buying devices from them rather than shady clones.
<PaulFertser>
Olimex OCD-H is cheaper than legit J-Link and I'd say it should be about the same from OpenOCD performance point of view.
<PaulFertser>
wkoszek: there's also J-Link EDU for for non-commercial projects
wmat has quit [Server closed connection]
wmat has joined #openocd
bvernoux has quit [Read error: Connection reset by peer]