Overflow in settime posix function

[AndreyDodonov-EH]

Describe the bug
Overflow in settime posix function

This is a follow-up of #41111
and #41814

In this PR gettime function was changed: #41814
Settime uses same conversion and has same effect.
So API should be symmetrical.

[AndreyDodonov-EH]

Obviously it will reduce precision, but it was already done anyway by changing gettime function.

[cfriedt]

Thanks for finding this. We should probably have a test that checks for rollovers.

[nordic-krch]

Obviously it will reduce precision, but it was already done anyway by changing gettime function.

It does not reduce the precision. It's just that calculation is performed on seconds and only remainder is in nanoseconds. Contrary to the approach where all is calculated in nanoseconds and that can result in overflow.

[AndreyDodonov-EH]

Obviously it will reduce precision, but it was already done anyway by changing gettime function.

It does not reduce the precision. It's just that calculation is performed on seconds and only remainder is in nanoseconds. Contrary to the approach where all is calculated in nanoseconds and that can result in overflow.

But doesn't in this case remainder have milliseconds precision? (even though saved in ns member of the struct)
I'm talking about using ms in settime.

[github-actions]

This issue has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this issue will automatically be closed in 14 days. Note, that you can always re-open a closed issue at any time.

[cfriedt]

Still trying to sort the last few clock-related items out with nrf & nxp in #53605 .

Whatever that fix ends up being should also be backported to v2.7

[cfriedt]

@AndreyDodonov-EH - can you verify that this is also fixed by #52977