From owner-cvs-all Thu Feb 4 15:14:17 1999 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA08999 for cvs-all-outgoing; Thu, 4 Feb 1999 15:14:17 -0800 (PST) (envelope-from owner-cvs-all@FreeBSD.ORG) Received: from critter.freebsd.dk (critter.freebsd.dk [212.242.40.131]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id PAA08933; Thu, 4 Feb 1999 15:13:43 -0800 (PST) (envelope-from phk@critter.freebsd.dk) Received: from critter.freebsd.dk (localhost [127.0.0.1]) by critter.freebsd.dk (8.9.1/8.8.5) with ESMTP id AAA01212; Fri, 5 Feb 1999 00:11:55 +0100 (CET) To: Adam David cc: nate@mt.sri.com, adam@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: Re: cvs commit: src/sys/i386/conf GENERIC In-reply-to: Your message of "Thu, 04 Feb 1999 22:49:47 GMT." <199902042249.WAA19377@veda.is> Date: Fri, 05 Feb 1999 00:11:54 +0100 Message-ID: <1210.918169914@critter.freebsd.dk> From: Poul-Henning Kamp Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk In message <199902042249.WAA19377@veda.is>, Adam David writes: >> Are you sure you don't have APM enabled in the BIOS ? > >Duh... > >Power Management = disable >PM by APM = yes > >Help says "disable,enable" and "yes,no". > >OK so disabled power management still has power management, found by trial >and error. Definitely not obvious. ... All hail Redmond! >I pulled the old comment from GENERIC. Should I pull the new comment from LINT, >leave it in, or further improve? I guess what we should really do is this: Monitor the TSC relative to the i8254 and if it strays from its "nominal" frequency (Ie, what we measured at some point during boot) relative to the i8254 counter, we drop back to the i8254 for timecounting. It's not like the TSC is designed for timekeeping in the first place, and I belive some of the Cyrix CPUs does architecturally sanctioned things which blows the use of the TSC out of the water entirely. This takes its extreeme form on multi-cpu Alphas where each CPU has its own inependent clock, generated by a SAW device which are better at measuring environmental influence than at keeping pace. I have tried to get a dialogue going with Intel for two years now about adding a counter running at a fixed frequency, readable in one atomic operation for use as timebase. It doesn't have to be in the CPU, in fact, for SMP systems it would be smarter if it were in one of the chipset functions so all CPUs can get at it. "Sorry not interested" is all that I've gotten back. At this time, the best resolution you can reliably get on the PC/AT++ architecture is still the i8254 running at: 14318318.318... Hz ------------------ = 838 nanoseconds. 12 In practice much bigger tics because accessing the i8254 is such a mess :-( Poul-Henning PS: This is one majore problem with Dave Mills new nanokernel stuff: It does UnNatural Things in the face of APM and similar crap. At least the timecounters could cope with it, if we detected the problem. -- Poul-Henning Kamp FreeBSD coreteam member phk@FreeBSD.ORG "Real hackers run -current on their laptop." FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message