lkml.org 
[lkml]   [1999]   [Jul]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: Patch: CLONE_PPID (was kernel thread support - LWP's)
From
Date
Tim Hockin <thockin@isunix.it.ilstu.edu> writes:
> I'm not trying to scatter efforts. My stuff just got coded yesterday :) I
> agree completely with what you say about 1 patch. As for Pthreads, Ulrich
> - you have some say in the glibc maintaining, correct?

Is anyone working on the LinuxThreads component?

Last year sometime I talked about my CLONE_SIBLING idea to make thread
creation cheaper. I haven't seen the CLONE_PPID code yet, but the
intention sounds similar. I did an initial implementation of
CLONE_SIBLING, which worked well enough that I could do the necessary
work on LinuxThreads. Back around the start of the year I got a proof
of concept version working, based on glibc 2.0 LinuxThreads, that runs
LinuxThreads programs happily, but with much faster pthread_create. I
have gradually tidied that version up, and I was planning to put it
out for testing/feedback after fixing a couple of things:

- Porting my changes to glibc 2.1 LinuxThreads, to make a glibc
submission feasible.

- The modified LinuxThreads ought to work just as well as the existing
LinuxThreads when running on existing kernels (my code already follows
a design that makes this easy, it just needs finishing off).

- Revisit my kernel changes (I suspect subtle security problems).

I expected to get the code out for testing later this summer, but I've
been pre-empted. So now I'm wondering how I can best contribute to
this flurry of interest in the area. It sounds like Tim's CLONE_PPID
work will match up closely to what I'm doing to LinuxThreads, so I
won't need to work on the kernel side any more. I'm not sure yet if
and how it relates to Chris Smith's work. What I'll try to do is
finish off the LinuxThreads stuff I'm in the middle of (I'll get it
done next weekend probably). People working on the kernel side might
find that useful, and can give me feedback, and then I'll see what the
glibc folks think of it.


David Wragg

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:53    [W:0.052 / U:0.104 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site