Fix delay on esp32h2.

[playfulFence]

closes #1509

Submission Checklist 📝

• I have updated existing examples or added new ones (if applicable).
• My changes were added to the CHANGELOG.md in the proper section.

Pull Request Details 📖

Description

Сoefficient for getting 16MHz in Delay from xtal_clock (2.5) wrong because it's related for 40MHz xtal, when esp32h2's is 32MHz

Testing

Tested delay time with DSLab LA, it's 500-501ms now.

[bjoernQ]

Thanks! Code looks good and fixes the issue

Now thinking about it I wonder why we are not using Systimer::TICKS_PER_SECOND to know the frequency?
Shouldn't we also have an issue with the 26 MHz C2?

I thought we should also have issues on S2 but apparently, we use the CPU's timer on Xtensa (i.e. no issue there) - which makes we wonder if we should use systimer everywhere it's available (i.e. not on ESP32)

[SergioGasquez]

Thanks for fixing this! Do you mind enabling the get_time test on H2 so we test it moving forward?

[SergioGasquez]

Shouldn't we also have an issue with the 26 MHz C2?

Maybe we do have them, this morning I tried locally the test suite on a 26 MHz C2 and the get_time test was failing IIRC, I didnt investigate anything as I just wanted to confirm that pins of other tests were working fine 🤔

[playfulFence]

Thanks for fixing this! Do you mind enabling the get_time test on H2 so we test it moving forward?

Done!

[jessebraham]

Now thinking about it I wonder why we are not using Systimer::TICKS_PER_SECOND to know the frequency? Shouldn't we also have an issue with the 26 MHz C2?

I thought we should also have issues on S2 but apparently, we use the CPU's timer on Xtensa (i.e. no issue there) - which makes we wonder if we should use systimer everywhere it's available (i.e. not on ESP32)

Probably worth discussing, but this PR can be merged regardless I think so let's either open an issue or make a note to discuss it in our next meeting.

[jessebraham]

Good catch, thanks for the fix!

[SergioGasquez]

LGTM! Thanks

[SergioGasquez]

Not sure what happened, but HIL on S3 failed to build the tests: https://github.com/esp-rs/esp-hal/actions/runs/8942444773/job/24565046328

Edit: CI succeeded doing the same: https://github.com/esp-rs/esp-hal/actions/runs/8942444770/job/24565051849 werid

[jessebraham]

I've seen that happen before I think, not sure what causes it. Really weird.