From owner-freebsd-realtime Wed Mar 11 08:52:53 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA07402 for freebsd-realtime-outgoing; Wed, 11 Mar 1998 08:52:53 -0800 (PST) (envelope-from owner-freebsd-realtime@FreeBSD.ORG) Received: from hda.hda.com (hda-bicnet.bicnet.net [208.220.66.37]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA07394 for ; Wed, 11 Mar 1998 08:52:50 -0800 (PST) (envelope-from dufault@hda.hda.com) Received: (from dufault@localhost) by hda.hda.com (8.8.5/8.8.5) id LAA25788; Wed, 11 Mar 1998 11:52:25 -0500 (EST) From: Peter Dufault Message-Id: <199803111652.LAA25788@hda.hda.com> Subject: Re: threads questions In-Reply-To: <199803032151.QAA05739@immrc.eng.mcmaster.ca> from Richard Teltz at "Mar 3, 98 04:51:04 pm" To: rich@immrc1.Eng.McMaster.CA (Richard Teltz) Date: Wed, 11 Mar 1998 11:52:25 -0500 (EST) Cc: realtime@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL25 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-realtime@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org There is no rt-FreeBSD. I've been putting the infrastructure for P1003.1B into -current, and that along with the scheduler pieces should be there this weekend. Once that is there my understanding is that the threads package will notice the support from the header files and you can specify the alternative schedulers. I haven't looked at this, though, so you'll have to research how it works for yourself. You still have the many issues of the non-reentrancy of the kernel, etc, that make FreeBSD inappropriate for most serious uses. If I had the time I'd follow this route: 1. First set it up so that you could have the scheduler policy apply only within a process group, making the system a lot better for developing and simulating real time processes, though not delivering them. For example, you could then use the Ptolemy system's real time scheduling to drive simulations without having to be root and taking over the system. 2. Develop an approach of adding new pieces that were reentrant and targeted at real time, and a way of detecting when regular interrupts, system code, etc was activated and use make that information available to the process as a "real time fault". 3. Work toward an MP system where you'd have a dedicated realtime system. Peter -- Peter Dufault (dufault@hda.com) Realtime development, Machine control, HD Associates, Inc. Safety critical systems, Agency approval To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-realtime" in the body of the message