"Nathan J. Williams" wrote:
> 
> I've been working on some kernel infrastructure (scheduler
> activations) for a particular kind of kernel-assisted userlevel thread
> system. That may be what people were talking about. It's still very
> much a work in progress, but I hope to have much more to report by
> early February.

Sounds interesting... Just out of curiosity, whats the status of
this?  Would your approach allow other "kernel assisted userland
threads" to run while one thread of the process is in a
(blocking) syscall?

It would be great to have kernel-level-type threads available to
applications.  It seems like something that is increasing in importance in
todays applications.  I've got some software (designed originaly for
linux) that (stupidly) makes the assumption that pthreads are
kernel-level, and so just deadlocks under NetBSD.  I was planning on
going through it and adding "I/O Jacket" routines to make all syscalls 
non-blocking, but that will probably be a pain.

Does anyone know if any of the other *BSDs make kernel level threads
available to applications?  I'm aware of the threading model of Linux and
Solaris (and Windoze of course), but I don't know much about any other
OS...

Thanks,
	Rick