Re: pthreads

[i_khavki]

Quoting Jeroen Dekkers <jeroen@dekkers.cx>:

> Is there anyone working on pthreads? If not I want to try to finish it.
> (I have holidays this week. :) I already got the code from the
> sourceforgetit CVS.

All I can say is that I'm not working on it at the moment (or anything
Hurd related for that matter). Several people have expressed interest
to me privately, but not much has come out of it.

> I got a couple of questions:
> What's the status of the code?
> Is there anything I need to know other than what's in the TODO, NOTES
> and
> SIGNAL files?

Those files are notes that I took to be able to have a clear picture
of glibc <-> Hurd interaction in my head, I also tried to be complete
so that other would be able to benefit from it. However the glibc source
tree is the defining reference for these things. There's also a crucial
(at least it was to me) piece of documentation missing, a "walk through"
for the initialization (the stuff that happens before main()) for
glibc dynamic and static executables. Without it I was at a loss
as to how to write the initializatoin code for pthreads itself.

> I also want to know to which standards it needs to conform. I've got the
> SUSv2 and austin draft7 here, are there any others?
> Also which extensions of austin needs to be implemented?
> Anything other then the X/OPEN extensions?
> And what's the meaning of pthread_cond_{imply,unimply}?

Those are the same standards that I was using. Originally I was intdending
to implement everything except the real-time extensions.

pthread_cond_{imply,unimply} was introduced by Mark Kettenis on top of the
standard, they allow one condition to activate (imply) many other ones.
There was some ambiguity as to how they should be implemented, that is why
I put it off so I could ask Mark later.

The first step is to read the SUSv2 standard (it's not very long). Then
pick something that's not implemented yet, and do it. If confused, check
the glibc docs, glibc code, or ask.

Good luck.

Igor