Loading to two CPUs on F28379D and using the same GPIOs... CPU2 seems to be a 'master' and 'takes over' GPIOs?

Question

Mike Buba 2021-1-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/713313-loading-to-two-cpus-on-f28379d-and-using-the-same-gpios-cpu2-seems-to-be-a-master-and-takes-ov

编辑： Mike Buba 2021-1-10

Hi,

I have loaded a code to CPU2 (rectangular pulse signal with a duration

and frequency

on GPIO31 and GPIO34). On the oscilloscope, I can see the signals.

When I load a different code to CPU1 (rectangular pulse signal with a duration

(

) and frequency

(

) on GPIO31 and GPIO34 (same GPIOs!)) I can still see the signals from CPU2. Additionally, on CPU1 I am sending a rectangular pulse signal to GPIO28 and can see it on the oscilloscope, so I know the code is running, but as if CPU2 has 'taken over' shared GPIOs?!

I have tried to reset the CPUs in Code Composer Studio, but after "CPU Reset", is still runs a CPU2 code.

How to 'release' GPIOs from CPU2?
What exactly happened when I loaded the code on CPU2 and CPU1 using the same GPIOs. Why didn't the last one (CPU1) overwritten the signals sent to GPIOs? Is CPU2 master and CPU1 slave?
Is there an option to format CPUs from Simulink, i.e. delete everything and make them as from the factory? I tried compiling and loading an empty Simulink model, but it does not allow to compile an empty sheet.

As a way forward I have created a dummy code and loaded to the CPU2. This seems to do the trick and 'release' shared GPIOs. But I am not sure if this is the right solution.