l4-hurd
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Pthread assumption + starting real port


From: Farid Hajji
Subject: Pthread assumption + starting real port
Date: Mon, 13 Nov 2000 03:01:28 +0100

Hi,

would it be safe to assume the presence of a pthread library in each
target environment for the new Hurd implementation?

I'm seriously considering starting to rewrite auth (and probably later
proc) mostly from scratch, defining a preliminary libmom API while
proceeding. As I outlined before, I'll try to completely avoid glibc
at this stage and will restrict myself to POSIX C Library calls, iff they
are needed at all. My first shot at libmom would be
libmom-{guest-os-with-posix-c-library} running on top of FreeBSD.
The second step would be to port this part of libmom to libmom-l4,
then run auth and proc on top of L4, linking against the reentrant
FreeBSD C-Library bundled with OSKit.

[If this doesn't make sense, please let me know before I start hacking
in the wrong direction].

The main problem I'm facing right now (apart from general lack of time)
is how to do threading in a most portable way. Sticking with cthreads is
silly, since it's completely mach-related. IMO, we should really move to
pthread-API from the very beginning, leaving the implementation of a
pthread library running on top of L4 to people with enough expertise.
This way, development can start right now (at least for libmom-{guestos}),
since most modern guest-os provide a pthread API nowadays.

I have no idea how much time this preliminary port will require nor if
it will succeed at all (I can currently only dedicate an extremely slim
portion of brain-time to it). But we really need to start at least
_something_ to get going. I'll keep you updated and I'll certainly have
lots of questions (especially about Hurd internals). If you're willing
to help in this precise area, please coordinate with me.

Thanks,

-Farid.

-- 
Farid Hajji -- Unix Systems and Network Admin | Phone: +49-2131-67-555
Broicherdorfstr. 83, D-41564 Kaarst, Germany  | address@hidden
- - - - - - - - - - - - - - - - - - - - - - - + - - - - - - - - - - - -
Murphy's Law fails only when you try to demonstrate it, and thus succeeds.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]