When debugging this problem I had a check for the STOP flag in my code, and it was never cleared, so just "waiting" for STOP to go away does not work. And just adding a timeout will cause the next transaction to be corrupted in the way that is descri...
@TSosi.1 I just have happened to implemented that workaround (checking and clearing STOP before starting a new transaction) and am doing a test run since yesterday (now ran for about 26 hours without causing a problem). I have a log message in my so...
I am running into exactly the same problem on a F407, every 100ms I read a I2C sensor and from time to time it does only generate a start-stop sequence. I am running Zephyr, so it seems it is not a software issue since all kinds of different software...
