STM32H563RGT6 flash HW bug even with BL=0xE5?
I've noticed some strange behavior when doing sector erase on this device.
I erase sector 0..63 on bank 2 and write my data. I can confirm that this operation is correct by reading the flash starting on 0x08080000 using stm32cubeprogrammer.
As a verification step, I read the very same address in my program and calculates a CRC, but now it seems like the reading takes place in bank 1.
Everything works fine when doing a full bank erase, but not when doing sector erase.
I am forced to use sector erase as I want to preserve the high-cycle flash which contains data that should be retained.
Bank swapping is not enabled.
Reading the ES0565 rev 5 errata, section 2.2.22, I notice that there has been errors in this field, but my device has BL = 0xE5, and should not be affected.