Matt Thomas <matt@3am-software.com> writes:

> Bucky Katz wrote:

> Again, I feel compelled to reiterate that an M:N implementation does
> not require scheduler activations.  If a pthread blocks/sleeps due a
> voluntary timeout (sleep, nanosleep, cond_timedwait, etc.) or
> waiting for a mutex or condition that can be internally to
> libpthread and may use userspace context switches to switch to next
> runnable thread (assuming it's not already running in another lwp).
> If it running in another thread and the current thread has nothing
> left to do, the cheapest thing you can do is give up the CPU and let
> the kernel scheduler do its thing.

This is why I tried to be careful and always say I wanted M:N
threading, and not scheduler activations.

I doubt very much, though, that anyone was going to volunteer to fix
M:N threading, so I asked for SA back in the uniprocessor case.

I still do not believe, having looked at the code, that it would have
been a big deal to provide, but I understood it isn't going to happen.

I have learned my lesson, and reset my expectation of the NetBSD
developer community's committment to its ueser base.

I withdraw all of my requests WRT -current, and ask that our changes
be pulled up to 4.0 as others have suggested.