tests/periph_timer_periodic: spice up test

[maribu]

Contribution description

The test should detect some bugs regarding incorrect behavior regarding timer_start() not resuming periodic timers as expected.

Testing procedure

The test should still pass (hopefully) for all non-AVR boards, but fail on AVR due to the bug fixed in #17387

Issues/PRs references

Used to test #17387

[fjmolinas]

I ran it on arduino-uno arduino-zero atmega256rfr2-xpro b-l072z-lrwan1 frdm-k64f frdm-kw41z hifive1b i-nucleo-lrwan1 nrf52dk nucleo-f030r8 nucleo-f091rc nucleo-f103rb nucleo-f207zg nucleo-f303k8 nucleo-f303re nucleo-f334r8 nucleo-f767zi nucleo-g071rb nucleo-g474re nucleo-l073rz nucleo-l152re nucleo-l433rc nucleo-l452re nucleo-l496zg nucleo-l4r5zi nucleo-l552ze-q nucleo-wl55jc openmote-b p-nucleo-wb55 samr21-xpro slstk3402a z1, all good but the AVR ones see here

[kaspar030]

BTW, this is the first build that was limited by can_fast_ci_run.py. Worked fine.
Here's the output (cut from https://ci.riot-os.org/RIOT-OS/RIOT/17388/d84e6515fb189e9c9af61880037f47e40d4e8ac0/output.html):

--- static tests: passed
-- running on worker mobi6.inet.haw-hamburg.de-21 thread 1, build number 103295.
--- can_fast_ci_run:
{
  "apps": [
    "tests/periph_timer_periodic"
  ],
  "boards": []
}
BOARDS_CHANGED=""
APPS_CHANGED="tests/periph_timer_periodic"

[maribu]

I changed the wording a bit: s/repetition/iteration/g since the first run is not a repetition.

[maribu]

This is btw. cppcheck approved (I locally still run it in my editor - with appropriate include paths). The chan > 0 looked suspiciously like a bound check to cppcheck and cppecheck pointed out that chan < MAX_CHANNELS would also be needed for full bounds checking.

[fjmolinas]

ACK

[maribu]

Thanks for the review!