Rtc tests refactor

[jaenrig-ifx]

By creating this pull request you agree to the terms in CONTRIBUTING.md.
https://github.com/Infineon/.github/blob/master/CONTRIBUTING.md
--- DO NOT DELETE ANYTHING ABOVE THIS LINE ---

CONTRIBUTING.md also tells you what to expect in the PR process.

@NikhitaR-IFX here are the changes.

• Moved the rtc_init.c to soft_reset: in main.c (The only real necessary change).
• Unified alarm setup, which is now called within alarm() and if repeat in the internal irq.
• irq() is only setting the callback
• Replaced sleep() in test for condition check with timeout in tests (this would be less sensitive to time fluctuation and minimize that waiting time).
• Added __ del __() (** this might be removed again).

Things that are still missing:

1. consider cyhal_init()/free() to be delegated to the main(), as the rtc() is required by the time module. Which means, this RTC API does not provide any chance of calling those functions. It will be always inited in the main() and freed when we do a soft reset.
2. check arguments in irq() -> wake is unused, and we might inform about it? trigger can only be RTC.ALARM0 (even if there is no other).
3. Remove commented initialization of rtc_obj in constructor? Let´s imagine we want to call a second time RTC() constructor. What should be the behavior? If it is a singleton, it should complain... Otherwise it should return the same object in the same state? That would not make much sense to me. Shall it return again the RTC() but fresh and clean to reconfigure? We have the deinit() cancel()... So I would say they should use that...
   I would say: second call to RTC() should return error. What do you think? Anything I am missing?
4. Maybe refactor the modtime function in a way that they reuse functions of rtc. Instead of directly calling cyhal_rtc functions.

[NikhitaR-IFX]

1. init() and free() are handled in main right now also no or? Only a check is made in respective modules if RTC is already inited through rtc_enable() to ensure there is no access to RTC without initializing. Did I get your point right?
2. Agreed.
3. So I tried calling init twice from mtb and it does it fine without any complaints. But ideally if there is only 1RTC, I would expect second init() to fail. The datasheet does not mention of more RTC's available neither it says there is only 1. But with diagram in "Block and Functionality" section in Page 6, I would believe there is only 1. So on MPY side, the constructor should be ideally called only once and we can keep a track of that with a flag, return error on second call. BTW, I did not find "commented initialization of rtc_obj in constructor".
4. Yes we can. Probably we might need to add/extend some helper functions but all doable.

[NikhitaR-IFX]

I am happy with the rest of the changes. Just some improvements/questions we can discuss over.

[NikhitaR-IFX]

But where is this initialized as NULL? Should be done in cases where there is no alarm or there is alarm but no callback associated.

[NikhitaR-IFX]

Ideally callback can be NULL or a pointer to handler function. For ex: A NULL is valid, when wake-up condition is set. You only want to wake up the system from deep-sleep/hibernate modes and set the wakeup source as RTC ALARM and then you do not need user-level cback. The only difference is if handler is NULL, we don't need to call register function. Rest remains the same.