The part that confused me is the name and implementation of ktime_divns(): I wonder if ktime_divns() should "convert" its ktime_t argument to nanoseconds first with ktime_to_ns(). It would be a no-op but just to clarify that it works with and returns nanoseconds, even though the argument is a ktime_t.

This revision is now accepted and ready to land.May 8 2025, 9:42 AM

In D50005#1146250, @dumbbell wrote:

This looks good to me.

The part that confused me is the name and implementation of ktime_divns(): I wonder if ktime_divns() should "convert" its ktime_t argument to nanoseconds first with ktime_to_ns(). It would be a no-op but just to clarify that it works with and returns nanoseconds, even though the argument is a ktime_t.

Yes, ktime_t == ns seems to be implied. There's a few type mis-matches in the file I think but that's unrelated to this change. I wonder if we should do a pass and fix some of them but I'd rather let other things settle first a bit.

Closed by commit rG93b82146c43b: LinuxKPI: add ktime_get_boottime_seconds() (authored by bz). · Explain WhyMay 9 2025, 7:46 PM

This revision was automatically updated to reflect the committed changes.

bz added a commit: rG93b82146c43b: LinuxKPI: add ktime_get_boottime_seconds().

bz added a commit: rG1d70da6aa294: LinuxKPI: add ktime_get_boottime_seconds().May 15 2025, 9:53 AM