From owner-freebsd-acpi@FreeBSD.ORG Thu Oct 12 04:46:04 2006 Return-Path: X-Original-To: freebsd-acpi@freebsd.org Delivered-To: freebsd-acpi@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id F145F16A407 for ; Thu, 12 Oct 2006 04:46:04 +0000 (UTC) (envelope-from nate@root.org) Received: from root.org (root.org [67.118.192.226]) by mx1.FreeBSD.org (Postfix) with ESMTP id D24C043D79 for ; Thu, 12 Oct 2006 04:46:02 +0000 (GMT) (envelope-from nate@root.org) Received: (qmail 47561 invoked from network); 12 Oct 2006 04:46:10 -0000 Received: from ppp-71-139-39-204.dsl.snfc21.pacbell.net (HELO ?10.0.5.50?) (nate-mail@71.139.39.204) by root.org with ESMTPA; 12 Oct 2006 04:46:10 -0000 Message-ID: <452DC86F.7060102@root.org> Date: Wed, 11 Oct 2006 21:45:35 -0700 From: Nate Lawson User-Agent: Thunderbird 1.5.0.7 (Windows/20060909) MIME-Version: 1.0 To: Jacques Garrigue References: <20061010.165404.74754637.garrigue@math.nagoya-u.ac.jp> <20061010.173707.08072970.garrigue@math.nagoya-u.ac.jp> <452BD9D2.5030609@root.org> <20061012.120703.59465247.garrigue@math.nagoya-u.ac.jp> In-Reply-To: <20061012.120703.59465247.garrigue@math.nagoya-u.ac.jp> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-acpi@freebsd.org Subject: Re: Sony Vaio VGN-TX92S 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: Thu, 12 Oct 2006 04:46:05 -0000 Jacques Garrigue wrote: > From: Nate Lawson > >>>>> Also, see the acpi man page for cpu idling, namely >>>>> setting hw.acpi.cpu.cx_lowest=C3. >>>> I tried this, with only C2, as C3 does not seem to be available for >>>> this cpu currently. (The Intel docs talk of a "very deep sleep" mode, >>>> so this may be lacking in the FreeBSD support.) >>> I was wrong. When I tried first the supported modes were only C1 and >>> C2, but using sysctl again later I found that C3 and even C4 were >>> available. I wonder why it changes while running. I tried with C4, but >>> I get no significant change in temperature. >> It changes based on AC line status. See the acpi man page or >> /etc/defaults/rc.conf ("cx" lines) to see. > > Thanks. With this I can keep it low. > But this does not change the original strange behaviour: if I boot > with AC on, I only get C1 and C2 listed, but if I boot on battery I > get C1 to C4. The supported modes do not seem to change after boot. The BIOS decides that. We should get a notify that Cx states changed and re-evaluate the settings. That support is not done yet so you're stuck with whatever you get on boot. >>> Since the main problem is hang-up from overheat, is there any progress >>> in having powerd slow the CPU when the temperature is too high? >> Passive cooling is already supported in 6-stable and 7-current. > > OK, I've set _PSV to 60.0C, which should have protected me against > overheat. I verified that the CPU slows when over 60.0C. > But just as I was writing this mail (and extracting the Linux kernel > sources, don't ask me why...) I experienced again a strange unexpected > reboot. Just before reboot, the sysctl node dev.cpu.0 had disappeared! > The temperature was over 60C, but not that high. Sounds like a bug or hardware problem. > By the way, the acpi_thermal man page seems to be wrong about setting > temperatures through sysctl: it says that you should use an integer > temperature in celsius, but if I do that I get "Invalid argument". > Actually I had to set in 1/10 of Kelvin (ie 3332), after reading > kernel sources. I think you need a "C" to indicate celsius. Example: ="60C" The sysctl user program translates for Celsius, Fahrenheit ("120F"), and Kelvin (no letter). -- Nate