20:43 <orzel> Hello. I have a circuit with a HC32F030 (a chinese ARM kinda like STM32f030, i guess). The circuit has a five-hole debug port, and i could identify them as +vcc/gnd/swdio/swdclk and "resetB", according to the chinese datasheet. My only 'debugging' card is a ft232h breakout board. I tried connecting that and using opencd, but i keep getting this https://dpaste.org/vyyr no matter how i connect things. Any idea what i'm missing ? I'm

20:43 <orzel> afraid something needs to be done with reset that the scripts i'm using don't do.

20:44 <Krazubu> hi orzel I'm in the same boat, trying to communicate with a cortex M33

20:45 <PaulFertser> orzel: hi

20:45 <Krazubu> once in a while I get the correct addr

20:45 <orzel> Krazubu: you're luckier than me so far :)

20:45 <PaulFertser> orzel: how exactly are you connecting ft232h breakout for SWD?

20:46 <PaulFertser> orzel: also please show -d3 log

20:47 <orzel> vcc/gnd : obvious, swdclk to d0, swdio to d2, a resistor between d1 and d2 (i tried 220 and 470 ohms). I tried not using the reset, or connecting to d4 or d5

20:47 <orzel> PaulFertser: because ft232h-module-swd.cfg mentions RESET on those pins.

20:47 <Krazubu> orzel that took me 3 days to get something though :) and it started to appear yesterday… sometimes, and it's gone since then

20:48 <orzel> PaulFertser: i tried inverting d1/d2 for swdio and the resistor

20:48 <PaulFertser> orzel: TDI should be connected via a resistor to SWDIO, TDO should be connected to SWDIO directly.

20:50 <PaulFertser> orzel: then if it gives the same result try starting OpenOCD while connecting resetB to GND manually.

20:50 <orzel> ok. tdi is d2, do that's the the second option.

20:51 <orzel> The mcu is powered by the ftdi breakout. I 'assume' it's ok, and enough.

20:52 <PaulFertser> orzel: if it's a bare MCU, probably. You can check the Vcc with voltmeter to see if it's not sunk.

20:54 <orzel> It's not bare, there's a board, with buttons, screen and all. The mcu is soldered, smd, can't 'easily' take it out.

20:55 <PaulFertser> orzel: I mean powering probably is ok, not sure how much ft232h can provide, it's about 50 mA probably.

20:55 <orzel> i tried while the board was powered. Slighlty dangerous, but everything survived, and i didn't work any better.

20:55 <orzel> "it didn't"

20:56 <PaulFertser> orzel: it's not dangerous if it's powered by 3.3 V supply.

20:56 <orzel> with the ftdi only, vcc is still stable at 3.3v

20:56 <PaulFertser> orzel: so retry with resetB pulled down. And show -d3 please.

20:56 <orzel> yes, both are 3.3V, but you never now if there's a slight difference.

20:57 <PaulFertser> btw, since 1003bc7be7c471d28794280be69632c5f074789f on a real ftdi breakout board you can just short TDO and TDI together without resistor if you're only using it in SWD mode, see 1003bc7be7c471d28794280be69632c5f074789f

20:59 <orzel> i had already tried just connecting them together

21:01 <orzel> PaulFertser: i connected resetB to to gnd : https://dpaste.org/KSGt

21:02 <PaulFertser> orzel: is it possible SWSCK and SWDIO are mixed up, have you tried tracing the signals to the chip?

21:03 <orzel> https://github.com/openocd-org/openocd/commit/1003bc7be7c471d28794280be69632c5f074789f (i have it i guess, i'm using 0.11)

21:03 <PaulFertser> orzel: yep

21:03 <orzel> PaulFertser: i identified the port using a continuity tester between those holes (not even pins) and the mcu

21:06 <PaulFertser> orzel: do you have a logic analyser? Sigrok has support for SWD.

21:06 <orzel> I just tried inverting them, doesn't help

21:07 <orzel> PaulFertser: i do, and I know how to use sigrok

21:07 <orzel> I didn't expect to go down to this :)

21:08 <PaulFertser> orzel: I can't spot anything wrong in what you're doing so far, the target seems to be not replying.

21:08 <PaulFertser> orzel: on your breakout board where do you have Vccio connected to?

21:09 <PaulFertser> In other words, does it get its 3.3 V?

21:09 <PaulFertser> orzel: btw, original stm32 will always reply while its reset pin is pulled low.

21:11 <PaulFertser> orzel: do you by any chance have some other cortex-m board so that you could check if the adapter itself is all right?

21:17 <orzel> PaulFertser: the breakout is connected by usb, and i've verified 3.3v on vcc

21:17 <orzel> not sure about cortex-m, will check. possible.

21:18 <orzel> i have those cheap blue pill clones

21:19 <orzel> pa13=swdio / pa14=swclk, it's the same as my chinese chip

21:19 <orzel> gonna try this

21:24 <PaulFertser> orzel: checking on a blue pill would be an excellent test

21:52 <orzel> PaulFertser: I still need to connect the resistor, but i got something like this : https://dpaste.org/5dxk

21:52 <orzel> Far from perfect, but at least it recognizes the mcu

21:52 <orzel> I didn't change the jumpers, i dont need to, right ?

22:01 <Krazubu> is my clock supposed to look like this while openocd is trying to communicate ? https://ibb.co/bsC3ppn

22:01 <Krazubu> using raspberry pi 3 with cortex m33 target

22:02 <PaulFertser> orzel: is that with a bluepill?

22:02 <orzel> PaulFertser: yes

22:02 <orzel> (a clone)

22:03 <PaulFertser> orzel: the connection seems unstable.

22:03 <PaulFertser> Krazubu: looks like a clock...

22:04 <Krazubu> why is it irregular ?

22:05 <PaulFertser> Krazubu: OpenOCD has delays here and there, and it's all bitbanging anyway.

22:07 <Krazubu> ok, I'm trying to find why I can't communicate with my MCU

22:08 <Krazubu> I managed to get the IDR once in a while when launching many times

22:08 <Krazubu> but it won't give it anymore now

22:08 <orzel> PaulFertser: i've noticed. I have short cables, but through a breadboard. I tried different usb ports.

22:10 <PaulFertser> orzel: I'd try lowering adapter speed too.

22:10 <PaulFertser> orzel: and probably your clock should go via a resistor to suppress ringing.

22:10 <PaulFertser> orzel: sometimes it also helps to connect more ground wires between adapter and the target.

22:26 <orzel> 1kHz seems rather slow already, not ?

22:28 <orzel> PaulFertser: how can i slow the adapter ? Through openocd options ? Modifying the script ?

22:33 <orzel> ah, ouch, it's 1MHz :)

22:33 <orzel> Indeed, probably too fast.

22:33 <orzel> What a weird default :)

22:37 <karlp> if you have reasonable wiring, 1MHz is a _slow_ defualt :)

22:38 <orzel> er.. ok

22:39 <orzel> found for speed, i need "adapter speed ..."

22:39 <orzel> let's move command line to openocd.cfg file

22:45 <orzel> Ok, at 100kHz, it's better, but i keep getting some "Error: SWD Read data parity mismatch"

22:45 <orzel> at 200kHz it's ok (i think). Go figure..

22:46 <orzel> PaulFertser: give me that (still with blupill) https://dpaste.org/Xb7f

22:46 <orzel> Is the setup now 'validated' ? :-)

22:51 <orzel> (i also added a resistor on clock)

23:03 <orzel> PaulFertser: I can connect to another board with a similar mcu.

23:08 <orzel> But still nothing on the hc32f030, reset open or pulled to ground, whatever the speed

23:08 <orzel> reset is high (3.3v when kept open)