Revert "Merge pull request #8272 from NXPmicro/Ensure_RTC_OSC_Start"

[cmonr]

Description

This reverts commit 3d859ca, reversing changes made to 03f4623.

One set of CI issues that have been blocking PRs this week have revolved around lp_ticker activity with the K64F, K22F, and K82F devices in our Test CI.

We don't yet know how to exactly put the devices in a bad state, but we did transplant a device that was in a bad state into a local debugging env and found that the lp_ticker was in a bad state. Since the particular timer used in the Kinetis devices is only reset on POR and currently our CI does not have the ability to power cycle devices, once the devices were put into a bad state, they would fail lp_ticker tests until the devices were power cycled. Special thanks to @deepikabhavnani for looking at the transplanted device.

Part of what makes reproduction of the issue difficult is that our CI has a load-balancing feature where the Test CI can run tests on different devices of the same board. This allows us to 1) run tests in parallel, and 2) swap out boards if one is disabled, but the downside is that from a device's point of view, the tests aren't run in sequence; the order that tests are run are undeterministic. @studavekar is looking into disabling this in the short term to see if this helps aleviate issues.

Pull request type

[x] Fix
[ ] Refactor
[ ] Target update
[ ] Functionality change
[ ] Docs update
[ ] Test update
[ ] Breaking change

[cmonr]

/morph build

[mbed-ci]

Build : SUCCESS

Build number : 3694
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/8826/

Triggering tests

/morph test
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 3293
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/8826/

[mbed-ci]

Test : FAILURE

Build number : 3466
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/8826/3466

[deepikabhavnani]

K82F IAR failure does not look related to lp ticker issue.. please confirm

[cmonr]

@deepikabhavnani That looks correct, although it's odd that new issues would crop up like this...

@ARMmbed/mbed-os-maintainers fyi.

Restarting.
/morph build

[deepikabhavnani]

:139::FAIL: Values Not Within Delta 2000 Expected 19989 Was 213735.

@cmonr - looking more in detail issue didn;t just pop up, it is the same issue which original PR was trying to resolve. Inconsistent delay first time after boot up (hard boot), will not be seen after soft reset.

Reopened #5348.

Please note low power timer/ticker test on Kinetics device will have high failure rate on hard boot in CI.

[cmonr]

#8826 (comment)

Ooph. Not good news at all. Thanks for the clarification.

@ARMmbed/mbed-os-hal @ARMmbed/team-nxp Fyi.

[mbed-ci]

Build : SUCCESS

Build number : 3695
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/8826/

Triggering tests

/morph test
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 3294
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/8826/

[mbed-ci]

Test : SUCCESS

Build number : 3467
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/8826/3467

[0xc0170]

By reverting this we still need to fix the issue with tickers + powering up . Thanks for reopening that issue. I'll talk to @fkjagodzinski

[ithinuel]

@deepikabhavnani

[...] capacitor configuration is skipped and default RTC configuration is not fetched.

And rtc_basic_config is actually never initialized so the behaviour of that is unpredictable.
That might be the cause of intermittent failures we have seen.

The code in https://community.nxp.com/docs/DOC-335568?q=32 looks good but that is not exactly what was done in this PR.

@maclobdell @ARMmbed/team-nxp