From owner-freebsd-hardware@FreeBSD.ORG Mon Dec 29 14:15:18 2003 Return-Path: Delivered-To: freebsd-hardware@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 1211416A4D0; Mon, 29 Dec 2003 14:15:18 -0800 (PST) Received: from vp4.netgate.net (vp4.netgate.net [205.214.170.248]) by mx1.FreeBSD.org (Postfix) with ESMTP id 67A8043D2D; Mon, 29 Dec 2003 14:15:15 -0800 (PST) (envelope-from admin@forkthepenguin.com) Received: from localhost (ibrew@localhost) by vp4.netgate.net (8.11.6/8.11.7) with ESMTP id hBTMFF101620; Mon, 29 Dec 2003 14:15:15 -0800 (PST) Date: Mon, 29 Dec 2003 14:15:15 -0800 (PST) From: admin@forkthepenguin.com X-X-Sender: ibrew@vp4.netgate.net To: John Baldwin In-Reply-To: Message-ID: References: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: Damian Gerow cc: freebsd-hardware@FreeBSD.org Subject: Re: Kernel Optimizations for Processors X-BeenThere: freebsd-hardware@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: General discussion of FreeBSD hardware List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 29 Dec 2003 22:15:18 -0000 On Mon, 29 Dec 2003, John Baldwin wrote: > > On 29-Dec-2003 admin@forkthepenguin.com wrote: > >> > # NO_F00F_HACK disables the hack that prevents Pentiums (and ONLY > >> > # Pentiums) from locking up when a LOCK CMPXCHG8B instruction is > >> > # executed. This option is only needed if I586_CPU is also defined, > >> > # and should be included for any non-Pentium CPU that defines it. > >> > >> 'and ONLY Pentiums' > > > > It also says "should be included for any non-Pentium CPU that defines it" > > but how do you know if the C# defines this? Is this another item that > > would show up in the "Features" list? I looked at a PI-133 I have and > > don't see it : > > It means that if you have 'cpu I586_CPU' in your kernel config and > you aren't running a Pentium, then you can define this option. However, > it won't really buy you much since we only install the F00F hack once > during bootup and all adding this option will do is just not check for > it at bootup. Cool, thanks for the clarification. I do not have I586_CPU defined in my kernel conf (only I686_CPU) and had no problems building the kernel. Under Linux you have to define -march=i586 to build the kernel properly for the C3. I bring this up because you can specify CPUTYPE in the /etc/defaults/make.conf. According to the notes in that file, the following cpus can be defined : # Intel x86 architecture: # (AMD CPUs) k7 k6-2 k6 k5 # (Intel CPUs) p4 p3 p2 i686 i586/mmx i586 i486 i386 # Alpha/AXP architecture: ev6 pca56 ev56 ev5 ev45 ev4 Does CPUTYPE affect the -march flag? I ask because gnu apparently supports the VIA C3 using -march=c3. As a test I set CPUTYPE=c3 and had no problems building the kernel nor do I see any evidence that CPUTYPE was pulled from make.conf during the kernel build. Is CPUTYPE only used for application optimizations? I suspect that adding CPUTYPE=c3 was a misguided assumption based on the gnu docs (see below). http://gcc.gnu.org/onlinedocs/gcc-3.3.1/gcc/i386-and-x86-64-Options.html#i386%20and%20x86-64%20Options Chris