Re: Make condition variables take absolute wakeup time

[Mindaugas Rasiukevicius <rmind%NetBSD.org@localhost>]

Andrew Doran <ad%netbsd.org@localhost> wrote:
> For a sleep ~497 days into the future at 100Hz, yes. Perhaps uint64_t would
> be better.

Just a detail, but is not that ~497 days for 1000Hz, and ~4971 days for 100Hz?

yamt%mwd.biglobe.ne.jp@localhost (YAMAMOTO Takashi) wrote:
> is it necessary for eg. cv_time_sec to return errors?
> isn't it enough to simply return the value of hardclock_ticks and
> let cv_timedwait handle it?

Separate thing, but while changing this.. Could hardclock_ticks be made
unsigned? I would say uint32_t. Perhaps same for other *hz counters.

-- 
Best regards,
Mindaugas
www.NetBSD.org