Remove direct use of us and lp ticker from tests

[c1728p9]

Remove the direct use of the microsecond and low power ticker from the tests. This enforces that sleep mode is properly locked when using timers. Furthermore, this prepares the codebase for new ticker changes which allow differing clock frequencies and timer bit widths.

[c1728p9]

/morph test

[mbed-bot]

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph test

Output

mbed Build Number: 1304

All builds and test passed!

[0xc0170]

@pan- @bulislaw

[pan-]

In the lp_ticker test, it is a bit strange to use a combination of the C and C++ API. I'd suggest to explain when one has to be used over, that would help future maintainers.

[0xc0170]

In the lp_ticker test, it is a bit strange to use a combination of the C and C++ API. I'd suggest to explain when one has to be used over, that would help future maintainers.

+1. how shall this be suggested? In the test, explaine why Timer is used instead of directly tickers?

[c1728p9]

In general (not including the code under test) we need to encourage tests to be written using the C++ API rather than the HAL API. This is because the C++ API's provide thread and sleep safety and mbed-os is committed to keeping these APIs backwards compatible. With the HAL APIs there is more flexibility to change over time.

In this PR the HAL API is used only in TESTS/mbed_hal/lp_ticker/main.cpp. This is because it is the code being tested. The C++ API is also used to aid in the testing, but it itself is not the focus of the testing. All other files in this PR should be using the C++ APIs only.

[c1728p9]

/morph test-nightly

[theotherjimmy]

@c1728p9 Will this be needed for 5.6.1?

[c1728p9]

nope

[mbed-bot]

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph test-nightly

Output

mbed Build Number: 1349

All builds and test passed!