I implemented a in-application update similar to AN4767 with an STM32F429 CPU
My current problem is the short amount of time ( approx 200ms ) while programming the optcr register to set the BFB2 bit.
As the datasheet of STM32F429 says, a power-fail during flash programming can lead to an unpredictable behaviour.
After a bit of testing the "new" update approach i got to a situation where the SPRMOD bit was set and the device did not come up again. I'm not really sure if there is another bug in the code or if this happened while programming BFB2.
So when a power-fail occurs while setting the optcr register by issueing an OPTSTRT. What is really happening here?
As I now understand the way described in AN4767 is not power-fail-safe and therefore unusable??
What are your experiences with this feature?