#168 Address Problems on STM32H742xG/743xG devices

[Jonathan Krämer]

There is currently a problem with the address ranges on the STM32H742xG/743xG devices. These have not the full 2MB flash storage capacity defined in the dev_table.c line 121, but only 1MB, which would be no problem if the address range would be continously. The flash storage on these devices however is seperated into 4 sectors of 128kB on each bank, starting once at 0x08000000 (the normal start of the flash addresses for all STM32H7 devices, and the other 512kB are located at the second bank at 0x08100000.

Due to the fact that the addresses of the given .hex file in the extended linear address record lines are only evaluated once at the start a jump from bank 1 to bank 2 is impossible as I understand it.

As Workaround I defined the first flash bank to have 1MB storage and fill up unused space with 0s to get continously data writing over all addresses from 0x08000000 to 0x08100000 and then the other data in the hex file. 2 Problems arising from this workaround: The created hex file is larger then needed and the flashing process takes up more time; more problematic is, that addresses are written to, that are non existant and the -v option to verify the written content can not be used with this solution.
Can there be another solution or fix in the code of this package?

[Tormod Volden]

Please don't post images of text, just copy and paste the text.

Can't you just flash the part at 0x08000000 with one stm32flash invocation and the part at 0x08100000 with another?