From owner-freebsd-stable Sat Aug 18 22:55:53 2001 Delivered-To: freebsd-stable@freebsd.org Received: from peter3.wemm.org (c1315225-a.plstn1.sfba.home.com [24.14.150.180]) by hub.freebsd.org (Postfix) with ESMTP id 5C0AE37B403 for ; Sat, 18 Aug 2001 22:55:45 -0700 (PDT) (envelope-from peter@wemm.org) Received: from overcee.netplex.com.au (overcee.wemm.org [10.0.0.3]) by peter3.wemm.org (8.11.0/8.11.0) with ESMTP id f7J5tjM62450 for ; Sat, 18 Aug 2001 22:55:45 -0700 (PDT) (envelope-from peter@wemm.org) Received: from wemm.org (localhost [127.0.0.1]) by overcee.netplex.com.au (Postfix) with ESMTP id 1306B38FD; Sat, 18 Aug 2001 22:55:45 -0700 (PDT) (envelope-from peter@wemm.org) X-Mailer: exmh version 2.3.1 01/18/2001 with nmh-1.0.4 To: freebsd-stable@freebsd.org Cc: Kris Kennaway Subject: Re: New kernel option CPU_ENABLE_SSE Date: Sat, 18 Aug 2001 22:55:44 -0700 From: Peter Wemm Message-Id: <20010819055545.1306B38FD@overcee.netplex.com.au> Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Thu Aug 16, 2001 3:03 pm, Kris Kennaway wrote: > On Thu, Aug 16, 2001 at 06:00:40PM -0400, Kenneth W Cochran wrote: > > Ok, so how can I tell if a given CPU supports that feature? > > > I see a "feature list" in the kernel startup, and in the > > case of a Pentium-III, SSE is at the "end" of that list; is > > that the Definitive Indicator? If so, what's to keep the > > kernel from detecting and enabling it (per config-option, > > for example)? (kernel complexity, for example...) > > Yeah, it could probably be done that way. It *is* done that way. CPU_ENABLE_SSE only compiles in the support code. If you compile it in, it looks for the 'SSE' flag in the cpu_features bits and adapts at runtime to the presense of the FXSR and SSE bits. You cannot use SSE without FXSR. It was done as an option for now for conservative reasons. After 4.4, I think the support code should be made a negative option (ie: XX_NO_SSE to compile *out* the support code), and a tunable to disable support for it at boot time as well. However, the issue of this affecting FreeBSD as a vmware "guest" OS has been raised, even without CPU_ENABLE_SSE active. This is not supposed to happen, the only practical change that the code does without the option active is increase the size of the PCB. I am wondering if that is causing stack overflows under vmware now. Can anybody else confirm or deny whether 4.4-RC with (and without) CPU_ENABLE_SSE works under vmware? Cheers, -Peter -- Peter Wemm - peter@FreeBSD.org; peter@yahoo-inc.com; peter@netplex.com.au "All of this is for nothing if we don't go to the stars" - JMS/B5 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message