Messages in this thread | | | From | "Khimenko Victor" <> | Date | Sun, 18 Jul 1999 14:03:21 +0400 (MSD) | Subject | Re: kernel thread support - LWP's |
| |
In <m3g12mfm7t.fsf@fred.muc.de> Andi Kleen (ak@muc.de) wrote: > drepper@cygnus.com (Ulrich Drepper) writes:
>> Tim Hockin <thockin@isunix.it.ilstu.edu> writes: >> >> > I like it too. I'm doing it right now. Only problem is notifying the >> > parent that it had a child without knowing it! (Next on Jerry Springer..) >> >> While you are at it please also take the appended patch into accound. >> It was written by Richard Henderson after my continuous whining about >> missing kernel features. It is also needed to correctly start >> threads. What it does is introducing a feature available in all other >> kernels, namely, creating a thread which is not starting to run >> immediately.
> Why can't you do that by clone()ing to a trampoline that sends SIGSTOP > to itself ?
The whole idea is to avoid context switches after process creation (it's not even related to LWP's). I can not understood how clone()ing to a trampoline will help to do this.
- 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/
| |