From owner-freebsd-acpi@FreeBSD.ORG Tue Jul 1 05:08:13 2008 Return-Path: Delivered-To: freebsd-acpi@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 353BD1065C2C for ; Tue, 1 Jul 2008 05:08:13 +0000 (UTC) (envelope-from smithi@nimnet.asn.au) Received: from gaia.nimnet.asn.au (nimbin.lnk.telstra.net [139.130.45.143]) by mx1.freebsd.org (Postfix) with ESMTP id D94A48FC0C for ; Tue, 1 Jul 2008 05:08:10 +0000 (UTC) (envelope-from smithi@nimnet.asn.au) Received: from localhost (smithi@localhost) by gaia.nimnet.asn.au (8.8.8/8.8.8R1.5) with SMTP id PAA29082; Tue, 1 Jul 2008 15:06:51 +1000 (EST) (envelope-from smithi@nimnet.asn.au) Date: Tue, 1 Jul 2008 15:06:50 +1000 (EST) From: Ian Smith To: David Wolfskill In-Reply-To: <20080630163845.GL13924@bunrab.catwhisker.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Cc: freebsd-acpi@freebsd.org Subject: Re: How/why would dev.cpu.0.freq_levels change??!? 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: Tue, 01 Jul 2008 05:08:13 -0000 On Mon, 30 Jun 2008, David Wolfskill wrote: > On Mon, Jun 30, 2008 at 03:24:11PM +1000, Ian Smith wrote: > > ... > > > * As you can see, this can lead to the "interesting" situation that the > > > current CPU frequency is higher than the maximum "available." > > > > Perhaps just morbid curiousity, but I'm wondering which cpufreq drivers > > this machine winds up using (acpi_perf or est/p4tcc or .. ?) > > > > grep -i acpi /var/run/dmesg.boot ? > > sysctl hw.acpi ? > > I've placed copies of the dmesg.boot from each of RELENG_6, RELENG_7, > and HEAD in www.catwhisker.org:~david/public_html/FreeBSD, as well > as copies of the kernel configs (joining the ASL/DSDT stuff). I > just added output from "sysctl hw.acpi" from each, as well: > > bunrab(4.11-S)[3] ls -l laptop.i8200.* > -rw-r--r-- 1 david staff 91343 Jun 29 17:22 laptop.i8200.asl > -rw-r--r-- 1 david staff 28033 Jun 30 06:51 laptop.i8200.dmesg.boot.6 > -rw-r--r-- 1 david staff 30401 Jun 30 06:54 laptop.i8200.dmesg.boot.7 > -rw-r--r-- 1 david staff 34133 Jun 30 07:41 laptop.i8200.dmesg.boot.8 > -rw-r--r-- 1 david staff 12622 Jun 29 17:22 laptop.i8200.dsdt > -rw-r--r-- 1 david staff 975 Jun 30 08:55 laptop.i8200.hw.acpi.6 > -rw-r--r-- 1 david staff 974 Jun 30 08:57 laptop.i8200.hw.acpi.7 > -rw-r--r-- 1 david staff 976 Jun 30 09:08 laptop.i8200.hw.acpi.8 > -rw-r--r-- 1 david staff 9502 May 7 15:05 laptop.i8200.kernel.6 > -rw-r--r-- 1 david staff 9154 Jun 8 2007 laptop.i8200.kernel.7 > -rw-r--r-- 1 david staff 9399 Jan 12 17:26 laptop.i8200.kernel.8 Nothing if not thorough :) As I vaguely suspected from those freq/0 reports, the cpufreq driver attachment seems a little unusual, though I'm hoping someone who knows a lot more about this may say something. On other P4 boxes' dmesg, I often notice est+p4tcc cpufreq drivers attaching, but on yours, on 6.3-STABLE: acpi_perf0: on cpu0 acpi_throttle0: on cpu0 acpi_throttle0: P_CNT from P_BLK 0x8e0 but on 7.0-STABLE and 8-CURRENT: acpi_perf0: on cpu0 p4tcc0: on cpu0 I'm really unsure what this means, but thought it perhaps noteworthy that a) it wasn't using est? and b) I don't understand why on 6.3 it would attach acpi_throttle when p4tcc seems available - but my hours of digging through the drivers code has left me only a little the wiser, and I'm still referring to old (5.5) sources; I've just installed 7.0 on my T23 but I haven't yet setup and copied my working environment to it. I also notice that there's only a _CRT temp. value, no _PSV, and (so?) hw.acpi.thermal.tz0.passive_cooling=0 if that's relevant to this? > > . are you running the latest BIOS/ACPI upgrade available from Dell? > > "Project: DELL Mojave", > > "Date: 01/28/1998", > > "Ver: 1.00.04" > > Errr... hmmm? The machine shows "BIOS Version: A11". The above is from your .asl, dunno if it's relevant? > > . might you have any BIOS settings re performance/economy/cooling set? > > My settings under "Power Management" (in the BIOS config/setup menus): > > BATTERY AC > _______ __ > Brightness: [XXXXX ] [XXXXXXX ] > Power Management: Enabled Enabled > Display Time-Out: 4 Minutes Disabled > Disk Time-Out: 3 Minutes Disabled > Suspend Time-Out: Disabled Disabled > 22D Time-Out: Disabled Disabled > Smart CPU Mode: Enabled Enabled > Display Close: Active Active > > > Ring/Event Resume: Enabled > Alarm Resume: Enabled > Wakeup On LAN: Disabled > Intel SpeedStep(tm): Enabled > CPU on AC: Automatic > CPU on Battery: Automatic > Auto On Mode: Disabled > Auto On Time: 00:00 > > > Most of the above are defaults; I disabled Suspend & S2D on battery, > as well as told it to remain active on battery if I shut the lid. > Everything else should be a default setting. I'm really quite unsure about whether the BIOS settings do or don't affect ACPI operation under FreeBSD, or to what extent. Ie what does 'Smart CPU Mode' do here? Perhaps someone could comment on whether such BIOS settings are reflected in ACPI method invocation at all? > > . not running it in a dock are you? > > I am not running it in a docking station or port replicator. > > > From what you've described, it almost sounds like a hardware temperature > > sensor may have failed, or be reporting wrong, or something .. as this > > has only appeared recently, either something's broken, or perhaps you've > > inadvertantly changed something? You did mention having been inside .. > > did that go as far as re-pasting the CPU or other heatsinks? > > I did not see how to get the CPU heat sink off, so I didn't mess with > that. > > I did remove the keyboard, and it appears that the keyboard acts as a > secondary heat sink for the video card; I did clean both surfaces (the > chip & the underside of the keyboard) and place a thin layer of thermal > compound on the chip before re-seating the keyboard. That does not > appear to have had a noticable effect either way. > > Thanks for the help so far. Questions only help if they might lead to some answers .. not so far :) > I'm getting the distinct impression that it's likely that some of the > hardware is failing, and that I either need to have the machine repaired > by someone competent to do so (as opposed to me) or I need to consider > replacing it. (There are, after all, significant parts of the machine > that are over 5 years old. And I've been tracking various FreeBSD > branches on it just about daily as long as I've had it and it's been > working.) I saw your later message to jhb about sending it off to the shop, so hopefully something good will come of that. cheers, Ian