On Tue, Jan 18, 2022 at 10:19:21AM -0800, Peter Oskolkov wrote:

================ worker timeouts

Timeouts now are easy to explain: mutex::lock() and condvar::wait() have
timeouts, so workers waiting on these primitives naturally wait with timeouts;
if sys_umcg_wait() supports worker timeouts, this is it, all is simple; if
it does not, the userspace now has to implement the whole timeout machinery,
in order to wake these sleeping workers when their timeouts expire.

I still have absolutely no idea what you're on about here. Please reply
to the email on that subject:

  https://lkml.kernel.org/r/Ya34S2JCQg+81h4t@hirez.programming.kicks-ass.net

What would you have the timeout actually do? Talk about the state
transitions.