07:01 <keshav>  Hi karlp, I have multiple STM32 Microcontrollers individually connected in a linear daisy chain. TDI --> TDI1, TDO1 --> TDI2, TDO2--> TDO.

07:01 <keshav> I would like to mass flash those microcontrollers in the chain with one Hex file. It would always be STM32 in a chain.

07:05 <keshav> If that works, I  will make a PCB which would connect to those multiple Microcontrollers with one connector for the adapter and flash multiple Microcontrollers at a time which would make my work a whole lot easier

07:12 <Hawk777> keshav: Does it make sense to construct a config file that uses a TCL loop to define a series of chips, and then defines a TCL proc that contains a loop to flash code to all of them? You could then use a TCL variable to define how many chips there are, which could be set externally by outside tooling. I’m not sure whether you could do that on the command line, but even if not, you could have your external tooling write a tiny conf

07:12 <Hawk777> ile that just sets the variable and then includes the main file, and then uses the RPC API to call the function to flash the targets (or even just include a call to that function using -c on the command line).

07:19 <keshav> Unfortunately I have been working on OpenOCD only for the past week and I am unsure on how the program works. I have been testing with command line codes with existing Config files and I am able to flash single Microcontrollers through those config files. But regrettably I am starting to learn TCL and how it works.

07:20 <keshav> Is there an example cfg file in OpenOCD that could help me understand the loop that you mean?

07:31 <Hawk777> Hm, I’m not sure if there are any examples of loops in the shipped config files.

07:34 <Hawk777> But TCL is a regular programming language, so there are definitely plenty of options available. There’s some documentation at <http://jim.tcl.tk> that might be useful.

07:40 <keshav> Thank you.

12:27 <keshav> Hi.

12:27 <keshav> Update on the JTAG Olimex STM32 chain flashing.

12:27 <keshav> I connected two STM32 CPUs in a daisy chain as mentioned yesterday with TDI --> TDI1, TDO1 --> TDI2, TDO2--> TDO.

12:27 <keshav> The command code used is

12:27 <keshav> "C:\Program Files\GNU ARM Eclipse\OpenOCD\xpack-openocd-0.11.0-3\bin\openocd.exe"  -f interface\ftdi\olimex-arm-usb-ocd-h.cfg -f target\stm32f3x.cfg -c init -c targets -c "halt" -c "flash  erase_sector 0 0 127" -c "reset halt" -c "flash write_image C:/Users/Buero-1/Desktop/init.hex" -c "verify_image C:/Users/Buero-1/Desktop/init.hex" -c "reset

12:27 <keshav> halt" -c shutdown

12:27 <keshav> There were no error messages. Two Taps were found. But the last Tap on the chain was flashed. The first one on the chain was not flashed. There was no change in Tap 1. Where could I be doing this wrong? Is there something I am missing in the command code?

12:28 <keshav> xPack OpenOCD x86_64 Open On-Chip Debugger 0.11.0+dev (2021-12-07-17:33)

12:28 <keshav> Licensed under GNU GPL v2

12:28 <keshav> For bug reports, read

12:28 <keshav>         http://openocd.org/doc/doxygen/bugs.html

12:28 <keshav> Info : auto-selecting first available session transport "jtag". To override use 'transport select <transport>'.

12:28 <keshav> Info : DEPRECATED target event trace-config; use TPIU events {pre,post}-{enable,disable}

12:28 <keshav> Info : clock speed 1000 kHz

12:28 <keshav> Info : JTAG tap: stm32f3x.cpu tap/device found: 0x4ba00477 (mfg: 0x23b (ARM Ltd), part: 0xba00, ver: 0x4)

12:28 <keshav> Info : JTAG tap: stm32f3x.bs tap/device found: 0x06432041 (mfg: 0x020 (STMicroelectronics), part: 0x6432, ver: 0x0)

12:28 <keshav> Info : JTAG tap: auto0.tap tap/device found: 0x4ba00477 (mfg: 0x23b (ARM Ltd), part: 0xba00, ver: 0x4)

12:28 <keshav> Info : JTAG tap: auto1.tap tap/device found: 0x06432041 (mfg: 0x020 (STMicroelectronics), part: 0x6432, ver: 0x0)

12:28 <keshav> Warn : AUTO auto0.tap - use "jtag newtap auto0 tap -irlen 4 -expected-id 0x4ba00477"

12:28 <keshav> Warn : AUTO auto1.tap - use "jtag newtap auto1 tap -irlen 5 -expected-id 0x06432041"

12:28 <keshav> Info : stm32f3x.cpu: Cortex-M4 r0p1 processor detected

12:28 <keshav> Info : stm32f3x.cpu: target has 6 breakpoints, 4 watchpoints

12:28 <keshav> Info : stm32f3x.cpu: external reset detected

12:28 <keshav> Info : starting gdb server for stm32f3x.cpu on 3333

12:28 <keshav> Info : Listening on port 3333 for gdb connections

12:28 <keshav> Info : JTAG tap: auto1.tap tap/device found: 0x06432041 (mfg: 0x020 (STMicroelectronics), part: 0x6432, ver: 0x0)

12:28 <keshav> target halted due to debug-request, current mode: Thread

12:28 <keshav> xPSR: 0x01000000 pc: 0x080001d4 msp: 0x20008000

12:28 <keshav> shutdown command invoked

12:28 <keshav> This was the result of the execution

12:28 <PaulFertser> keshav: use some pastebin, do not paste to channel!

12:41 <PaulFertser> keshav: yes, by default the last target is selected so it gets flashed.

12:42 <PaulFertser> keshav: you can select another target with "targets" command.

12:44 <keshav> Hi. Sorry about that. Here is the result in Pastebin

12:44 <keshav> https://pastebin.com/SQ5GjXpz

12:45 <PaulFertser> keshav: so it looks like it does what you're asking it for

12:47 <keshav> When I run the targets -c command, it shows only one target and that Target 0 is selected. That should be the last tap in my chain right?

12:47 <keshav> TargetName         Type       Endian TapName            State

12:47 <keshav> --  ------------------ ---------- ------ ------------------ ------------

12:47 <keshav>  0* stm32f3x.cpu       cortex_m   little stm32f3x.cpu       halted

12:52 <keshav> Is it showing just one target because both the taps are the same CPUs?

12:56 <PaulFertser> keshav: btw, is it olimex on windows 11 or did you install GNU/Linux?

12:56 <PaulFertser> Hm, one target, I see you sourced the target config just once, why?

12:57 <keshav> It is Olimex on Windows 10. I havent installed Linux yet and I havent tested on Windows 11 so far. I wanted to make this work on Windows 10 so that when this is successful, I will try it on Windows 11.

12:57 <PaulFertser> Makes sense

12:57 <PaulFertser> keshav: you have one target because you sourced it only once.

12:59 <keshav> I am a bit blank on sourcing it twice. I tried it today and the program finds the second one as a command of the first target config.

12:59 <keshav> cortex_m reset_config sysresetreq

12:59 <keshav> invalid command name "targetstm32f3x.cfg"

12:59 <PaulFertser> keshav: show the invocation command line

13:00 <keshav> "C:\Program Files\GNU ARM Eclipse\OpenOCD\xpack-openocd-0.11.0-3\bin\openocd.exe"  -f interface\ftdi\olimex-arm-usb-ocd-h.cfg -f target\stm32f3x.cfg -c target\stm32f3x.cfg -c init -c targets -c "halt" -c "flash  erase_sector 0 0 127" -c "reset halt" -c "flash write_image C:/Users/Buero-1/Desktop/init.hex" -c "verify_image

13:00 <keshav> C:/Users/Buero-1/Desktop/init.hex" -c "reset halt" -c shutdown

13:01 <PaulFertser> keshav: why do you have -c before target there?

13:02 <keshav> Should it be -f?

13:02 <PaulFertser> keshav: of course. And before that you probably want -c "set CHIPNAME stm32f3x-2"

13:04 <keshav> Is it possible to show me in the invocation command line?

13:05 <PaulFertser> keshav: please rephrase

13:07 <keshav> Could you edit my invocation command line? I tried with -f, but it throws an error  Error: Command/target: stm32f3x.cpu Exists

13:17 <PaulFertser> keshav: yes, it now says you need to add that -c command before -f so that it would be using different CHIPNAME

13:23 <keshav> PaulFertser: You mean like this?

13:23 <keshav> "C:\Program Files\GNU ARM Eclipse\OpenOCD\xpack-openocd-0.11.0-3\bin\openocd.exe"  -f interface\ftdi\olimex-arm-usb-ocd-h.cfg -f target\stm32f3x.cfg -f "set CHIPNAME stm32f3x-2" -c target\stm32f3x.cfg -c init -c targets -c "halt" -c "flash  erase_sector 0 0 127" -c "reset halt" -c "flash write_image C:/Users/Buero-1/Desktop/init.hex" -c

13:23 <keshav> "verify_image C:/Users/Buero-1/Desktop/init.hex" -c "reset halt" -c shutdown

13:24 <PaulFertser> keshav: not -f before set, -c

13:24 <PaulFertser> keshav: other than that, yes, like that.

13:32 <keshav> It throws the error

13:32 <keshav> Info : DEPRECATED target event trace-config; use TPIU events {pre,post}-{enable,disable}

13:32 <keshav> stm32f3x-2

13:32 <keshav> invalid command name "targetstm32f3x.cfg"

13:32 <keshav> I think on the second line it prints the name that we provided. But after that, it is unable to find the target cfg as a config data.

13:32 <keshav> I ran the invocation command line,

13:32 <keshav> "C:\Program Files\GNU ARM Eclipse\OpenOCD\xpack-openocd-0.11.0-3\bin\openocd.exe"  -f interface\ftdi\olimex-arm-usb-ocd-h.cfg -f target\stm32f3x.cfg -c "set CHIPNAME stm32f3x-2" -c target\stm32f3x.cfg -c init -c targets -c "halt" -c "flash  erase_sector 0 0 127" -c "reset halt" -c "flash write_image

13:32 <keshav> Z:/Elektronik/GSV13/Fertigung_GSV-13iu/Init/GSV13init_Ver1_6.hex" -c "verify_image Z:/Elektronik/GSV13/Fertigung_GSV-13iu/Init/GSV13init_Ver1_6.hex" -c "reset halt" -c shutdown

13:36 <PaulFertser> keshav: -c is for Tcl commands. -f is for files.

13:37 <PaulFertser> keshav: first you had -c for the file. Then you changed it to -f so it was correct but then we figured another command was to be added before that, and you added it with -f (which was wrong). Now you changed it to -c (correct) but also for whatever reason changed correct -f for the file to incorrect -c.

13:41 <keshav> I understand now. I was a bit stupid and understood that the -c corresponded to the next data in the command.

13:41 <keshav> I have altered it again. Thank you.

13:41 <keshav> "C:\Program Files\GNU ARM Eclipse\OpenOCD\xpack-openocd-0.11.0-3\bin\openocd.exe" -f interface\ftdi\olimex-arm-usb-ocd-h.cfg -f target\stm32f3x.cfg -c "set CHIPNAME stm32f3x-2" -f target\stm32f3x.cfg -c init -c targets -c "halt" -c "flash  erase_sector 0 0 127" -c "reset halt" -c "flash write_image C:/Users/Buero-1/Desktop/init.hex" -c

13:41 <keshav> "verify_image C:/Users/Buero-1/Desktop/init.hex" -c "reset halt" -c shutdown

13:41 <keshav> Somehow I got a bit flustered and understood it wrongly. Thank you. Now I will test if it works

13:42 <PaulFertser> keshav: that still writes only to one part, you'll need to also add commands to select another target and to flash it too after the first flash verify_image retuns.

13:51 <keshav> It works with "target CHIPNAME" right?

13:51 <keshav> To select the other target

13:51 <PaulFertser> keshav: I think the command is "targets stm32f3x-2" or "targets stm32f3x".

13:51 <PaulFertser> I know, sounds confusing

13:51 <PaulFertser> keshav: you can experiment interactively with telnet.

13:57 <keshav> Thank you so much. I will test it out and update you.

14:02 <karlp> it' also wonderful that the CHIPNAME is just a convention, entirely implemented as copypasta inside _most_ of the target tcl file too :)

14:03 <bencoh> :)

14:03 <bencoh> well, this convention is mentionned in the documentation ... I guess it helps :)

14:04 <bencoh> -n

14:04 <karlp> where is it in the docs?

14:04 <karlp> it's not in tap declaration

14:05 <bencoh> "variable naming convention"

14:06 <karlp> oh right, that's totally discoverable...

14:07 <karlp> the section "Communication Between Config files" which is really "example for multiple target support".... is super easy to find

14:09 <bencoh> I kinda feel like one needs to real the whole documentation to get a good grasp of what it can or cant do :)

14:10 <bencoh> (apart from trial&error)

14:10 <bencoh> to read*

14:10 <karlp> yeah, but it's splattered all over the place, even after having readit a few times I could never find it,

14:11 <bencoh> I usually search in the (full) pdf, I think it helps

14:17 <PaulFertser> When you search in the Info manual you get all the hits too

15:19 <karlp> I find info cumbersome, so I'm normally using the html. and the single page one loads badly, so I get no useful search :|

15:21 <PaulFertser> karlp: do you use GNOME or KDE probably? Their GUI Help apps should be showing Info pages too.

16:33 <karlp> gnome, but I've never had any luck with the built in help app honestly.

16:34 <karlp> it's either installed separately, or only handles the packaged versions, or whatever

16:34 <karlp> I mean, I can install to prefix=$HOME/.local and the cli does the right thing for unning, but whether man/info do the right thing? I've no idea.

16:35 <karlp> and the installed man/info pages I hve don't have any version info so I can't tell either

16:35 <PaulFertser> OpenOCD info has version on the first page iirc.

22:44 <karlp> not here, nor at work...

22:44 <karlp> I did look...

22:46 <PaulFertser> This User's Guide documents release 0.10.0+dev, dated 9 December 2020,

22:46 <PaulFertser> of the Open On-Chip Debugger (OpenOCD).

22:53 <karlp> hrm, I suspect info on my machine is showing me a man page.

22:53 <karlp> the _man_ page doesn't have a verison at least...

22:53 <karlp> in fact, it's dated 2009

22:54 <PaulFertser> man page didn't receive any love, that's true.

22:54 <karlp> well, "info openocd" is showing me the man page of my local custom build at least.

22:55 <karlp> I guess configure && make && make install didn't find the right things to build info?

22:56 <PaulFertser> it ends up in doc/openocd.info* and then installed in /usr/local/share/info normally

22:56 <karlp> yeah, I guess I don't have makeinfo or something, but that's not there.

22:56 <karlp> and "info openocd" is apparently falling back to showing me a man page

22:58 <PaulFertser> Lack of makeinfo would explain it.