Jason Thorpe <thorpej@nas.nasa.gov> writes:

> What we should do here, instead, is provide those primitives in our
> libc, and then make the pthreads package build on top of those!

I think that will be fairly difficult, since locks don't exist in a
pure vacuum.  They are assiciated with a scheduler, have a list of
waiting threads, etc.

I have a feeling by the time all the bits are split out it would have
been simplier to just include pthreads directly into libc.

--Michael