OpenOCD - Open On-Chip Debugger / Tickets / #398 Status of STM fork and adding support for new STM32 silicon

Tommy Murphy - 2023-05-31

What is the status of that fork?

You probably need to ask STM that question?

I am struggling with creating a working target config file for the STM32WBA chip.

What's wrong with the target script that's in the mainline OpenOCD?

https://github.com/openocd-org/openocd/blob/master/tcl/target/stm32wbx.cfg

Last edit: Tommy Murphy 2023-05-31

alternate
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Jan Henke - 2023-05-31
  
  The current script does not work, for starters the STM32WBA has different JTAG/SWD IDs. But also the chip architecture is different. STM32WBx is a dual core chip with one M4 and one M+, the new STM32WBA is a single core M33 chip. So there is definitely some difference that needs to be accounted for.
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
  - Tommy Murphy - 2023-05-31
    
    The current script does not work, for starters the STM32WBA has different JTAG/SWD IDs.
    
    Have you tried adding the relevant STM32WBA ID(s) to the stm32wbx.cfg script?
    
    https://github.com/openocd-org/openocd/blob/78688fea984b69d1986b6b730ff2935668cc9208/tcl/target/stm32wbx.cfg#L27
    
    (I presume that the 'x' in the name is a placeholder so that script should - eventually, if not right now - also support STM32WBA?).
    
    STM32WBx is a dual core chip with one M4 and one M+, the new STM32WBA is a single core M33 chip
    
    The stm32wbx.cfg script only defines a single Cortex-M target.
    
    https://github.com/openocd-org/openocd/blob/78688fea984b69d1986b6b730ff2935668cc9208/tcl/target/stm32wbx.cfg#L47
    
    Last edit: Tommy Murphy 2023-05-31
    
    If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Tarek BOCHKATI - 2023-05-31

I will push a gerrit change in a few minutes with the needed cfg file

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

please test using this patch
https://review.openocd.org/c/openocd/+/7694

Thank you for your work, unfortunately it runs into an error:

/usr/bin/openocd -c "tcl_port disabled" -c "gdb_port 3333" -c "telnet_port 4444" -s /usr/share/openocd/scripts -f /home/jan/Projekte/Experimente/Elektronik/stm32wba52cg-test/openocd.cfg -c "program /home/jan/Projekte/Experimente/Elektronik/stm32wba52cg-test/cmake-build-debug/stm32wba52cg-test.elf" -c "init;reset init;" -c "echo (((READY)))"
Open On-Chip Debugger 0.12.0
Licensed under GNU GPL v2
For bug reports, read
    http://openocd.org/doc/doxygen/bugs.html
Info : The selected transport took over low-level target control. The results might differ compared to plain JTAG/SWD
Error: hla_target: invalid parameter -ap-num (> 0)
/usr/share/openocd/scripts/target/stm32wbax.cfg:45: Error: 
in procedure 'script' 
at file "embedded:startup.tcl", line 28
at file "/home/jan/Projekte/Experimente/Elektronik/stm32wba52cg-test/openocd.cfg", line 11
at file "/usr/share/openocd/scripts/target/stm32wbax.cfg", line 45
GDB Server stopped, exit code 1

Using this config file:

#
# Configuration for STM32WBA52 Nucleo board (STM32WBA52CG)
#

source [find interface/stlink.cfg]

transport select hla_swd

source [find target/stm32wbax.cfg]

reset_config srst_only

Changing the ap-num in the cfg file manually to 0 yields another error:

/usr/bin/openocd -c "tcl_port disabled" -c "gdb_port 3333" -c "telnet_port 4444" -s /usr/share/openocd/scripts -f /home/jan/Projekte/Experimente/Elektronik/stm32wba52cg-test/openocd.cfg -c "init;reset init;" -c "echo (((READY)))"
Open On-Chip Debugger 0.12.0
Licensed under GNU GPL v2
For bug reports, read
    http://openocd.org/doc/doxygen/bugs.html
Info : The selected transport took over low-level target control. The results might differ compared to plain JTAG/SWD
srst_only separate srst_nogate srst_open_drain connect_deassert_srst

Info : clock speed 500 kHz
Info : STLINK V3J11M3B5S1 (API v3) VID:PID 0483:374F
Info : Target voltage: 3.312922
Warn : UNEXPECTED idcode: 0x0be12477
Error: expected 1 of 1: 0x6ba02477


GDB Server stopped, exit code 1

Process finished with exit code 0

I see that you are using "interface/stlink.cfg"
we recommend using "interface/stlink-dap.cfg" to work with cortex-M33 on Access Port != 0

I did not know that, switching to dap interface got more output, but still an error:

/usr/bin/openocd -c "tcl_port disabled" -c "gdb_port 3333" -c "telnet_port 4444" -s /usr/share/openocd/scripts -f /home/jan/Projekte/Experimente/Elektronik/stm32wba52cg-test/openocd.cfg -c "init;reset init;" -c "echo (((READY)))"
Open On-Chip Debugger 0.12.0
Licensed under GNU GPL v2
For bug reports, read
    http://openocd.org/doc/doxygen/bugs.html
srst_only separate srst_nogate srst_open_drain connect_deassert_srst

Info : STLINK V3J11M3B5S1 (API v3) VID:PID 0483:374F
Info : Target voltage: 3.273161
Info : Unable to match requested speed 500 kHz, using 200 kHz
Info : Unable to match requested speed 500 kHz, using 200 kHz
Info : clock speed 200 kHz
Info : stlink_dap_op_connect(connect)
Info : SWD DPIDR 0x0be12477
Error: [stm32wbax.cpu] Cortex-M PARTNO 0x0 is unrecognized
Warn : target stm32wbax.cpu examination failed
Info : starting gdb server for stm32wbax.cpu on 3333
Info : Listening on port 3333 for gdb connections
Error: [stm32wbax.cpu] Cortex-M PARTNO 0x0 is unrecognized
Error: [stm32wbax.cpu] Cortex-M PARTNO 0x0 is unrecognized
Error: [stm32wbax.cpu] Cortex-M PARTNO 0x0 is unrecognized
TARGET: stm32wbax.cpu - Not examined
GDB Server stopped, exit code 1

Sorry my bad, I still had the cfg file edited, with ap_num 1 we get further, still not entirely happy:

/usr/bin/openocd -c "tcl_port disabled" -c "gdb_port 3333" -c "telnet_port 4444" -s /usr/share/openocd/scripts -f /home/jan/Projekte/Experimente/Elektronik/stm32wba52cg-test/openocd.cfg -c "init;reset init;" -c "echo (((READY)))"
Open On-Chip Debugger 0.12.0
Licensed under GNU GPL v2
For bug reports, read
    http://openocd.org/doc/doxygen/bugs.html
srst_only separate srst_nogate srst_open_drain connect_deassert_srst

Info : STLINK V3J11M3B5S1 (API v3) VID:PID 0483:374F
Info : Target voltage: 3.244533
Info : Unable to match requested speed 500 kHz, using 200 kHz
Info : Unable to match requested speed 500 kHz, using 200 kHz
Info : clock speed 200 kHz
Info : stlink_dap_op_connect(connect)
Info : SWD DPIDR 0x0be12477
Info : [stm32wbax.cpu] Cortex-M33 r0p4 processor detected
Info : [stm32wbax.cpu] target has 8 breakpoints, 4 watchpoints
Info : starting gdb server for stm32wbax.cpu on 3333
Info : Listening on port 3333 for gdb connections
[stm32wbax.cpu] halted due to debug-request, current mode: Thread 
xPSR: 0xf9000000 pc: 0x08029ff8 msp: 0x20001000
(((READY)))
Info : tcl server disabled
Info : Listening on port 4444 for telnet connections
Info : accepting 'gdb' connection on tcp/3333
Error: Failed to read memory at 0x40015800
Info : device idcode = 0x10006492 (STM32WBA5x - Rev A : 0x1000)
Info : TZEN = 0 : TrustZone disabled by option bytes
Info : RDP level 0 (0xAA)
Info : flash size = 1024 KiB
Info : flash mode : single-bank
Error: Failed to read memory at 0x40015800
Info : device idcode = 0x10006492 (STM32WBA5x - Rev A : 0x1000)
Info : TZEN = 0 : TrustZone disabled by option bytes
Info : RDP level 0 (0xAA)
Error: invalid bank base address
Error: auto_probe failed
Error: Connect failed. Consider setting up a gdb-attach event for the target to prepare target for GDB connect, or use 'gdb_memory_map disable'.
Error: attempted 'gdb' connection rejected
com.jetbrains.cidr.execution.debugger.backend.gdb.GDBDriver$GDBCommandException: Remote connection closed
Debugger disconnected

Process finished with exit code 0
GDB Server stopped, exit code 137

Jan Henke - 2023-06-04

To verbalize the last post: With the change I was able to flash my MCU, but I cannot attach a debugger. Can you look into why the debugging does not work?

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Tommy Murphy - 2023-06-04
  
  You should post (attach) a verbose OpenOCD log for the failed attempt to debug to assist with investigation. Use openocd -d3 ... to get the verbose log.
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Jan Henke - 2023-06-04

Attached is the debug log. Thank you for looking into it.

debug.log

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Tommy Murphy - 2023-06-08
  
  Thanks - these seem to be the key errors in the verbose log file:
  
  Line 357: Error: 308 224554 arm_adi_v5.c:570 mem_ap_read(): Failed to read memory at 0x40015800 Line 370: Error: 321 224559 arm_adi_v5.c:570 mem_ap_read(): Failed to read memory at 0x40015800 Line 377: Error: 328 224560 stm32l4x.c:1851 stm32l4_probe(): invalid bank base address Line 378: Error: 329 224560 core.c:291 get_flash_bank_by_num(): auto_probe failed Line 379: Error: 330 224560 gdb_server.c:1061 gdb_new_connection(): Connect failed. Consider setting up a gdb-attach event for the target to prepare target for GDB connect, or use 'gdb_memory_map disable'. Line 380: Error: 331 224560 server.c:94 add_connection(): attempted 'gdb' connection rejected
  
  So it would make sense to drill into these in more detail first.
  
  Last edit: Tommy Murphy 2023-06-08
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
  - Jan Henke - 2023-06-23
    
    Hi,
    
    were you able to make some progress on this issue?
    
    Best regards,
    Jan Henke
    
    If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Jan Henke - 2023-07-23

Would it be possible to get the changes merged? Being able to upload is better then nothing, even if debugging is still broken.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Tommy Murphy - 2023-07-23
  
  OpenOCD is open-source so anybody is free to submit changes for consideration and inclusion. Including any changes in the STM fork - or any other fork. Of course, it would be nice if STM would do this themselves (and maybe they do from time to time?) but it's open to anybody else, in particular stakeholders in the STM mods, to do this unilaterally. Why not contact STM ask them if/when they might upstream the changes in their fork?
  
  Edit: I just noticed this:
  
  https://github.com/STMicroelectronics/OpenOCD
  
  STMicroelectronics is actively working at merging these modifications in the official OpenOCD. In the mean time, these modifications are available in this repository.
  
  And I see that you did contact them here but got no response so far:
  
  https://github.com/STMicroelectronics/OpenOCD/issues/20
  
  Last edit: Tommy Murphy 2023-07-23
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
  - Jan Henke - 2023-07-23
    
    I already did contact them (https://github.com/STMicroelectronics/OpenOCD/issues/20), but there is zero activity from STM to this issue.
    I am also referring to an open change set here: https://review.openocd.org/c/openocd/+/7694
    I tested it and it allows flashing the device. Attaching a debugger unfortunately does not work (see debug logs above). Therefore I was hoping the patch's original author could look into why debugging does not work. But alas since there is no activity at all, I would prefer to add at least the flashing ability.
    
    If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Tommy Murphy - 2023-07-23

I am also referring to an open change set here: https://review.openocd.org/c/openocd/+/7694

Maybe if you post review feedback on that change set it might help?
As far as I can see nobody has reviewed it so that may be one reason why it's lying dormant.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Jan Henke - 2023-07-23
  
  I am simply not familiar enough with the codebase to give a meaningful review. So far I have been a user only.
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Tomas Vanek - 2023-07-23

Tarek wrote 2 moths ago:

please test using this patch
https://review.openocd.org/c/openocd/+/7694

But I see in all logs:
Open On-Chip Debugger 0.12.0

instead of
Open On-Chip Debugger 0.12.0+dev-00xxx-gd3949b4af339

Jan,
testing a patched config file with an unpatched release OpenOCD binary has no value for us
and the reported error likely originates here.

Please recompile OpenOCD from source
git fetch https://review.openocd.org/openocd refs/changes/94/7694/2 && git checkout FETCH_HEAD
and retry.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Jan Henke - 2023-07-24
  
  Hi,
  I am using Gentoo, so everything is compiled from source. I have added the patch from https://review.openocd.org/c/openocd/+/7694 to my user patches and reemerged (recompiled) the package. So the changes are included, even if the version does not change it.
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Jan Henke - 2023-07-24
  
  For the sake of the argument I recompiled from Git master + the patch from https://review.openocd.org/c/openocd/+/7694.
  
  Attached you find the debug log with the same problem
  
  debug.log
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Tomas Vanek - 2023-07-24

There seems to be a mismatch in OTP bank address.
Could you try comment out OTP definition in tcl/target/stm32wbax.cfg
# flash bank $_CHIPNAME.otp stm32l4x 0x1fff7000 0 0 0 $_TARGETNAME

It should allow gdb attach.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Jan Henke - 2023-07-24
  
  Thanks, that worked indeed. Is that line necessary for anything?
  
  It would be great to get this patch into the next release, as it seems to be working with that line commented out.
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Status of STM fork and adding support for new STM32 silicon

The Open On-Chip Debugger

Milestone

Searches

Help

#398 Status of STM fork and adding support for new STM32 silicon

Discussion