From owner-freebsd-acpi@FreeBSD.ORG Wed Feb 20 17:22:13 2008 Return-Path: Delivered-To: acpi@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id D0D5716A409 for ; Wed, 20 Feb 2008 17:22:13 +0000 (UTC) (envelope-from alex.kovalenko@verizon.net) Received: from vms048pub.verizon.net (vms048pub.verizon.net [206.46.252.48]) by mx1.freebsd.org (Postfix) with ESMTP id AC00213C448 for ; Wed, 20 Feb 2008 17:22:13 +0000 (UTC) (envelope-from alex.kovalenko@verizon.net) Received: from [10.0.3.231] ([70.111.176.151]) by vms048.mailsrvcs.net (Sun Java System Messaging Server 6.2-6.01 (built Apr 3 2006)) with ESMTPA id <0JWJ001HXS8PZ890@vms048.mailsrvcs.net> for acpi@freebsd.org; Wed, 20 Feb 2008 11:22:03 -0600 (CST) Date: Wed, 20 Feb 2008 12:22:01 -0500 From: "Alexandre \"Sunny\" Kovalenko" To: acpi@freebsd.org Message-id: <1203528121.1019.25.camel@RabbitsDen> MIME-version: 1.0 X-Mailer: Evolution 2.12.3 FreeBSD GNOME Team Port Content-type: text/plain; charset=UTF-8 Content-transfer-encoding: 8BIT Cc: Subject: Is halting of the CPU0 on SMP supposed to slow down time? X-BeenThere: freebsd-acpi@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: ACPI and power management development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 20 Feb 2008 17:22:13 -0000 While trying to understand why setting hw.acpi.cpu.cx_lowest=C3 makes my laptop (ThinkPad X60 with Intel T2400 @ 1.8GHz) literally crawl, I have run across the following observation: if dev.cpu.0.cx_lowest is set to C3 and machdep.cpu_idle_hlt is set to 1 time slows down at about two orders of magnitude. The same result could be achieved by setting machdep.hlt_cpus to 1. Setting dev.cpu.1.cx_lowest to C3 or machdep.hlt_cpus to 2 does not cause any negative side effects. Setting machdep.cpu_idle_hlt to 0 and hw.acpi.cpu.cx_lowest to C3 gets machine warming all the way up to _PSV rather rapidly. This was observed with the HPET and ACPI-fast timecounters without discernible difference. The system is RELENG_7 as of Feb 15, cpufreq is loaded and powerd is running. I would like to have C3 working as it gives me much cooler machine in the idle state, but not at the cost of losing time accounting. Any suggestions will be greatly appreciated. -- Alexandre "Sunny" Kovalenko (Олександр Коваленко)