There is quite a lot of inter-task communication and I find no problem with that (at the moment) on target. SW is running for days with no issues.
In this case it is “impossible” that the task releasing the mutex is different than the creator as the mutex is wrapped in a mutex object (the mutex object is just declared in the beginning of a function), so I’d rather think the mutex data which is overwritten (by me). But why would it not crash every time? Fishy.
Anyhow, to really verify if the FreeRTOS can be run as I’ve set it up in Linux with google test I problably would need to downscope the test a bit to limit the sources of possible errors.