From owner-freebsd-hackers Tue Nov 21 10:17:43 1995 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id KAA11111 for hackers-outgoing; Tue, 21 Nov 1995 10:17:43 -0800 Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id KAA11086 for ; Tue, 21 Nov 1995 10:17:27 -0800 Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id LAA04316; Tue, 21 Nov 1995 11:12:40 -0700 From: Terry Lambert Message-Id: <199511211812.LAA04316@phaeton.artisoft.com> Subject: Re: int type in jmpbuf To: cimaxp1!jb@werple.net.au (John Birrell) Date: Tue, 21 Nov 1995 11:12:39 -0700 (MST) Cc: terry@lambert.org, hackers@FreeBSD.org, jb@cimlogic.com.au In-Reply-To: <199511210106.MAA29490@werple.net.au> from "John Birrell" at Nov 21, 95 12:10:30 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 853 Sender: owner-hackers@FreeBSD.org Precedence: bulk [ ... preeemptive thread scheduling using signals ... ] > > If you need it, you should add a system call (like SunOS has) to save > > and restore register state, including the FP registers. > > We're doing as the MIT pthreads code does (__asm__ fsave/frstor) to save and > restore the FP state on entering the scheduler from the real signal handler. > So far we don't see a problem on the 486 we are using. It's the coprocessor > problem that Bruce was pointing out that we can't cope with. We're treating > that as a limitation and pressing on... As long as you are hacking, putting the state save/restore for the signal handler below a system call interface would save your portability to non x86 platforms. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.