stm32 qspi in memory-mapped mode with dual quad-NOR on stm32H7 disco kits

[FRASTM]

Taking the #72339 to run the samples/application_development/code_relocation_nocopy/ on stm32h750b_dk and stm32h747i_disco_m7

Both targets have two qspi banks connected to 2 identical quad-spi NOR flash memories (mt25ql512ab)

[FRASTM]

Running the sample on the stm32h750b_dk.
The Dual-flash mode must be configured and the west flash must download the piece of function_in_ext_flash directly into the quad-NOR memory.

The STM32CubeProgrammer shows that

and memory-mapped examination

0x90000000 <function_in_ext_flash>:	0x49074a06	0xb5084807	0xf812f000	0x48074906
0x90000010 <function_in_ext_flash+16>:	0xe8bd680a	0xf0004008	0xbf00b80b	0x90000001
0x90000020 <function_in_ext_flash+32>:	0x08009dde	0x08009da6	0x240000e0	0x08009db8

are similar and samples execution is correct :

*** Booting Zephyr OS build v3.6.0-3069-gb4c50fa37515 ***                                                                                              
Address of main function 0x800066d                                                                                                                     
Address of function_in_ext_flash 0x90000001                                                                                                            
Address of var_ext_sram_data 0x240000e0 (10)                                                                                                           
Address of function_in_sram 0x24000001                                                                                                                 
Address of var_sram_data 0x240000e4 (10)                                                                                                               
Hello World! stm32h750b_dk

[FRASTM]

However this is still not enough to make the stm32h747_disco_m7 work correctly.
Even if the STM32CubeProgrammer shows code in the 0x90000000 area, the execution fails on wrong instruction.
--> bytes are not similar

[arbrauns]

The bytes are similar, are you sure the QSPI clock isn't too fast? I saw similar things in my testing when the flash was being overclocked.

[FRASTM]

The bytes are similar, are you sure the QSPI clock isn't too fast? I saw similar things in my testing when the flash was being overclocked.

Reducing the qspi-max-frequency = <8000000>; does not change the problem