Stm32 cmake refactor

[safesintesi]

Fixes #12 .

FetchContent with low connection would waste time downloading the libraries.
A Flag not for FAT would force it to download freeRTOS-kernel a second time leading to a waste of time.
Since those repos are already includes through submodules, downloading is redundant.
This fix takes the configuration time down from minutes to ~0.7s.

With the occasion I refactored the Cmake file structure to fix some link dependencies and some flags:

• Added targets for drivers and core
• Fixed some libraries link
• Defined STM32F411xE as default build in Drivers/CMSIS/.../stm32f4xx.h

This now builds with cmake but needs a tweek on FAT repo.
A pull request has been already made & changes are accepted and waiting to be merged:
FreeRTOS/Lab-Project-FreeRTOS-FAT#59

Edit: Also needs Cmake version bump to 3.18

Edit:

• FAT pull request has been merged and now this repo has updated the submodule.
• We want to fix the Thread lib too since we're at it, so this won't be merged until we solve the issue.

[BIRSAx2]

As of now, your pull request to the FAT repo has not been merged. I'll wait to review it until then because, as it stands, the project doesn't build.

[BIRSAx2]

You did not include the Startup/startup_stm32f411xe.s file in the build, which specifies, among other things, the initial PC and sets the CPU in Thread mode.
As a result, the build throws the following warning: warning: cannot find entry symbol Reset_Handler; defaulting to 08000000.

[safesintesi]

Good catch, that file wasn't included even on the original cmake structure on stm32 branch. I'll fix it this afternoon.

[BIRSAx2]

I've reviewed the code, and it looks good overall.
The only issues I found are already mentioned: the inclusion of "startup_stm32f411xe.s," and the dependency on the FAT repo's pull request.
Once these are addressed, we can proceed with the merge. If there are delays with FAT repo, considering a fork might be a viable option. Thanks!

[safesintesi]

@BIRSAx2 Should be good now, can you please check?

[safesintesi]

@gekoxyz seems like this is not the only step to take to fix the Thread issue.
I suggest keeping this pull request for the original purpose (fixing the FetchContent redundancy) and opening a new one for the linux port.
To do this we should revert the last changes, before merging and closing the pull request.

[BIRSAx2]

I noticed that the file task_gianluca.h has been deleted in this pull request. However, the build doesn't compile without it.

[BIRSAx2]

On top of that, the previous two tasks Task1 and Task2 are not running when building with CMake (the led is not blinking), while they do when building with STM32CubeIDE