SWO-related differences between STM32H723ZG and STM32H743ZI

[PTiha]

Hello,

I made a minimal project with the NUCLEO-H723ZG board, in which I also used SWO printf.

The project is all about generating and writing out a random number every 500 ms in the main cycle.

Debugging in STM32CubeIDE with ST-LINK/V3 and J-Link Edu also provides completely invaluable data.

Although I only use ITM channel 0, DWT-Comparator data is displayed in LOG ... ???

The MCU clock is 400MHz and the SWO divider has also been carefully set.

I did the same project on a NUCLEO-H743ZI2 board (the newer ST-LINK/V3 release) and it works perfectly with ST-LINK/V3 and J-Link EDU as well.

I extended both projects with SEGGER RTT printf and tried to debug with Ozone:Only RTT is visible with NUCLEO-H723ZG, SWO also works with NUCLEO-H743ZI2.

At this point, I rule out an error in the STM32CubeIDE or ST-LINK/V3, the problem must be somewhere in the difference between the two MCUs.

Software versions used: J-Link V6.98a, Ozone V3.22b

I attach the two .elf files, both loaded into RAM.

Thank you very much in advance!

[PTiha]

Anyone have any idea about the problem?

The debug infrastructure block diagram shows that TSG is missing from STM32H723ZG.

There are other differences too. Can these differences cause that problem?

[Tesla DeLorean · Accepted solution]

Cross linking related posts/topics https://community.st.com/s/question/0D53W00000bebkzSAA/nucleoh723zg-itm-printf-swv-problems

The H743 was originally conceived as a dual core device, there is a lot of clutter and complexity in there related to that. The H723 is likely a cleaner single core implementation. Expect it to have subtly different registers and plumbing.

There isn't a SWTF also

Expect different DBGMCU registers, addresses and units. Would definitely suggest unpacking the ROM Tables and seeing exactly what units/options where implemented.