st link debugging issue

[manoj3]

hello team,

I am facing a debug connection issue with STM32H723VGTx using STM32CubeIDE and ST-LINK.

Issue details:

• MCU: STM32H723VGTx

• Debug tool: ST-LINK (OpenOCD via STM32CubeIDE)

• Interface: SWD

• Reset mode: Software reset

• SWD frequency: 50 kHz

Observed behavior:

• ST-LINK connects successfully.

• Cortex-M7 core is detected correctly.

• GDB server starts, but the debug session fails during memory access.

Error message:

Error: Failed to read memory at 0x5c001000 Error: auto_probe failed Error: Connect failed

Additional observations:

• Target voltage reported by ST-LINK is ~2.8 V.

• I already tried:

  • Lowering SWD frequency

  • Using software reset

  • Adding monitor gdb_memory_map disable in debug startup commands

  • Halting on reset

• Issue still persists.

Request:
Could you please confirm:

1. Whether this memory address (0x5C001000) access failure is related to power / AXI-D2 domain / debug configuration on STM32H723.

2. If there are any recommended debug settings or patches for STM32H7 devices under these conditions.

3. Whether minimum voltage requirements can affect debug access to AXI/D2 memory during GDB connection.

   if possible can we connect

Regards,
K Manoj

[Uwe Bonnes]

Look at RM0468: 0x5C001000 is DBGMCU accessed by system bus. Debugger should access as 0xE00E1000 

So the debugger probably access should access via 0xE00E1000  and not 0x5C001000/

[manoj3]

I am seeing repeated debug failures on STM32H723 where OpenOCD cannot halt the core and fails to erase flash.

Logs indicate the debugger is accessing DBGMCU at 0x5C001000, which is the system bus mapping. According to RM0468, DBGMCU is also accessible via 0xE00E1000 (core debug).

It seems the system bus is not ready when OpenOCD tries to access DBGMCU, causing the failure.

Has this issue been reported for CubeIDE 1.14.x?
Is there a way to avoid early system-bus DBGMCU access or is an IDE downgrade the correct solution?

[AScha.3]

Hi,

> SWD frequency: 50 kHz

Should work with 4 MHz , if your connections ok and not more than 20 cm about.

If not working at 4M, something is wrong.

+ > voltage reported by ST-LINK is ~2.8 V.

Thats very low...you have powered target at 3v3 ?

Can you show a pic from your setup with st-link + target ?

[Pavel A.]

Hi @manoj3 , STM32H723 is a (more or less) newer family, so be sure to update the software to handle it correctly.

Especially with OpenOCD, be sure to use the ST's branch rather than trunk version aged 7-8 years.

CubeIDE 1.14 is not too old, but better update.

Quick check: install the latest CubeProgrammer on a decent Windows 11 PC. If it can connect, the target hardware is OK and the problem is in software.

[manoj3]

hello pavel A,

as you asked photo i have sent.

your saying software problem i am using 14.1 do i need to use other version can you please suggest which one i need to use.

tell me how to over come this issue. and do i need to solder NRST button on board?

[manoj3]

[gbm]

The debug probe is definitely NOT an ST-Link. It's firmware is illegal, very likely outdated and the attempt to update it from ST software should fail because it's illegal.

Get a real ST-Link or switch to CMSIS-DAP debug interface.

[manoj3 · Accepted solution]

hello team,

I am currently facing an issue with SDMMC initialization on the STM32H723VGT6 (WeAct development board).

Setup details:

• MCU: STM32H723VGT6

• SD Card: SanDisk Ultra 256GB (SDXC, exFAT)

• Interface: SDMMC (4-bit mode)

• Connection: External SD module using jumper wires

The SD card is detected, but the system gets stuck during initialization (HAL_SD_GetCardStatus / HAL_SD_Init stage). FATFS mount and file operations are not proceeding further.

We have verified:

• Clock configuration

• SDMMC settings (1-bit and 4-bit mode)

• exFAT configuration in FatFs

• DMA settings

Even after checking all of the above, the issue is still happening. It may be related to the SDMMC connection through jumper wires or compatibility with the 256GB SDXC card.

Could you please review this setup and help suggest a solution?

If possible, we can schedule a short call to debug this together and resolve it quickly.

Please let me know your available time.

Thanks and Regards,
K Manoj 

[manoj3]

hello team,

since past from one and half month i did not get any respond from your side,

i need clarification about my doubts ,can you please slove AS soon as possible.

k Manoj