Top and limit stack addresses

[kamtom480]

This PR changes the way to get stack addresses. Now instead of using only values from the linker file, you can define the top and limit address of the stack in the code.

Two new functions have been added: port_stack_get_limit and port_stack_get_top.

For atmel-samd, nrf and stm32f4 stack addresses are returned using values from the linker file. For cxd56 the stack addresses have been taken from the task where CircuitPython is running.

[dhalbert]

@kamtom480 I'm not sure why you resubmitted the PR multiple times, but don't worry about have a "clean" set of commits. It's fine with us to push new commits that fix build problems, typos, etc.

[kamtom480]

@dhalbert I just wanted to rerun tests. I didn't want to add new commits. There is a problem with these tests because sometimes they do not work properly. Maybe this is caused by the -j2 parameter when building?

[dhalbert]

Do you have this button on, say https://github.com/adafruit/circuitpython/pull/2230/checks. I'm not sure if you have to have a certain privilege to have it.

[dhalbert]

Your changes should not be affecting the micropython tests at all, I think, because your code is not used when building unix micropython.

We have seen these kinds of thread test failures occasionally, but not for a long time, so I'm not sure why they have reappeared. I'm not sure -j2 is involved because the test runner is a Python script, and I'm not sure anything from the makefiles are running in parallel at that point. But we'd have to take a closer look.

[tannewt]

Looks good to me! Thank you for the refactor.

[kamtom480]

@dhalbert I don't have this button.

[dhalbert]

@dhalbert I don't have this button.

OK, if it happens again, just ping one of us in the PR and we will re-run the build.