From owner-freebsd-hackers Fri Apr 3 08:09:09 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id IAA21277 for freebsd-hackers-outgoing; Fri, 3 Apr 1998 08:09:09 -0800 (PST) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from penrose.isocor.ie (penrose.isocor.ie [194.106.155.117]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA21269 for ; Fri, 3 Apr 1998 08:09:03 -0800 (PST) (envelope-from peter.edwards@isocor.ie) Received: from isocor.ie (194.106.155.26) by penrose.isocor.ie; 3 Apr 1998 17:08:41 +0100 Message-ID: <35250964.F89A1631@isocor.ie> Date: Fri, 03 Apr 1998 17:08:04 +0100 From: Peter Edwards Organization: ISOCOR X-Mailer: Mozilla 4.04 [en] (WinNT; I) MIME-Version: 1.0 To: Nate Williams CC: freebsd-hackers@FreeBSD.ORG Subject: Re: bcopy implementation References: <857m57c8tr.fsf@localhost.zilker.net> <199804031420.QAA05837@ocean.campus.luth.se> <199804031547.IAA22090@mt.sri.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > (The kernel doesn't save the contents of the FP registers, but since all > kernel calls are guaranteed to 'complete' before getting swapped, then > this isn't a problem in kernel-land.) > > Nate Huh? Surely the FP registers are saved? To get anything like decent speed math, I assume user programs can use floating point instructions directly. A quick look at /usr/src/sys/i386/i386/swtch.s and the save87 struct in /usr/include/sys/pcb.h seem to bear this out. Cheers, Peter. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message