From owner-freebsd-mobile@FreeBSD.ORG Tue Aug 26 11:01:40 2008 Return-Path: Delivered-To: freebsd-mobile@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 5DF961065671 for ; Tue, 26 Aug 2008 11:01:40 +0000 (UTC) (envelope-from smithi@nimnet.asn.au) Received: from sola.nimnet.asn.au (paqi.nimnet.asn.au [220.233.188.227]) by mx1.freebsd.org (Postfix) with ESMTP id 6CDD38FC15 for ; Tue, 26 Aug 2008 11:01:39 +0000 (UTC) (envelope-from smithi@nimnet.asn.au) Received: from localhost (localhost [127.0.0.1]) by sola.nimnet.asn.au (8.14.2/8.14.2) with ESMTP id m7QB1bbs045870; Tue, 26 Aug 2008 21:01:37 +1000 (EST) (envelope-from smithi@nimnet.asn.au) Date: Tue, 26 Aug 2008 21:01:36 +1000 (EST) From: Ian Smith To: freebsd_user@guice.ath.cx In-Reply-To: <20080826084027.GA8703@WORKSTATION.guice.ath.cx> Message-ID: <20080826203021.W14827@sola.nimnet.asn.au> References: <489E9531.2090200@guice.ath.cx> <20080825025833.GB3301@WORKSTATION.guice.ath.cx> <20080826002657.B14827@sola.nimnet.asn.au> <20080826084027.GA8703@WORKSTATION.guice.ath.cx> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Cc: freebsd-mobile@freebsd.org Subject: Re: TECRA A9-S9017 -- Idles too hot -- Hardware Support X-BeenThere: freebsd-mobile@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Mobile computing with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 26 Aug 2008 11:01:40 -0000 On Tue, 26 Aug 2008, freebsd_user@guice.ath.cx wrote: > For those just starting to follow this thread, you can somewhat start at the > begining here: In-Reply-To: <489E9531.2090200@guice.ath.cx> > - > - > Here's more data to append onto my last message; In-Reply-To: > <20080826002657.B14827@sola.nimnet.asn.au> -- in response to your: > > > However we need some empirical data about what it's doing. Showing > > your /var/run/dmesg.boot and 'sysctl hw.acpi' output would be a good start. > > > > With respects to 'powered' we ran in to a speed bump or two (2). > > IAN: > > (if it's running) then run 'powerd -v' which runs in foreground and says > > exactly what it's doing re shifting CPU frequency under various loads. > > > > freebsd_user: > Tue Aug 26 03:30:40 EDT 2008 --> /etc/rc.d/powerd stop > powerd not running? Right, it wasn't running because cpufreq isn't loaded, as below. Check 'kldstat -v | grep cpufreq' > Tue Aug 26 03:30:40 EDT 2008 --> > Tue Aug 26 03:30:40 EDT 2008 --> /etc/rc.d/powerd -v > /etc/rc.d/powerd: unknown directive '-v'. > Usage: /etc/rc.d/powerd > [fast|force|one](start|stop|restart|rcvar|status|poll) No, I'd have said; not /etc/rc.d/powerd, but just 'powerd -v' % which powerd /usr/sbin/powerd The idea is to start powerd manually, with -v switch, in foreground in its own window or vty, as shown in powerd(4) > Tue Aug 26 03:30:40 EDT 2008 --> /etc/rc.d/powerd start -v > Starting powerd. > powerd: lookup freq: No such file or directory Yes; it won't even startup without finding the cpu freq sysctls. I'm surprised that powerd failing to start hasn't been logged in /var/log/messages every time you've tried .. ? > Tue Aug 26 03:30:40 EDT 2008 --> ps auxww | grep -i powerd > root 9190 0.0 0.0 372 208 p3 R+ 3:32AM 0:00.00 grep -i > powerd > Tue Aug 26 03:30:40 EDT 2008 --> ps auxww | grep -i powerd > Tue Aug 26 03:30:40 EDT 2008 --> /etc/rc.d/powerd start -v > Starting powerd. > powerd: lookup freq: No such file or directory > Tue Aug 26 03:30:40 EDT 2008 --> ps auxww | grep -i powerd > Tue Aug 26 03:30:40 EDT 2008 --> /etc/rc.d/powerd start -v > Starting powerd. > powerd: lookup freq: No such file or directory > > Now I'm curious about the contents of /etc/rc.d/powerd ... Curious is good :) > powerd_poststop() > { > sysctl dev.cpu.0.freq=`sysctl -n dev.cpu.0.freq_levels | > sed -e 's:/.*::'` > /dev/null > } That'll be to reset dev.cpu.0.freq to its highest speed after quitting. > which prompts me to look at the following 'sysctl' ... > > Tue Aug 26 03:30:40 EDT 2008 --> sysctl dev.cpu > dev.cpu.0.%desc: ACPI CPU > dev.cpu.0.%driver: cpu > dev.cpu.0.%location: handle=\_PR_.CPU0 > dev.cpu.0.%pnpinfo: _HID=none _UID=0 > dev.cpu.0.%parent: acpi0 > dev.cpu.0.cx_supported: C1/157 > dev.cpu.0.cx_lowest: C1 > dev.cpu.0.cx_usage: 100.00% Yep, and all the freq stuff is missing. Refer previous message. > Tue Aug 26 03:30:40 EDT 2008 --> sysctl -a |grep -i freq > kern.acct_chkfreq: 15 > debug.cpufreq.lowest: 0 > debug.cpufreq.verbose: 0 > machdep.tsc_freq: 2194521505 > machdep.i8254_freq: 1193182 > machdep.acpi_timer_freq: 3579545 > Tue Aug 26 03:30:40 EDT 2008 --> > > Unless I've missed or mistyped something, the file /etc/rc.d/powerd is > trying to set a variable (dev.cpu.0.freq=) using the value(s) of "sysctl dev.cpu.0.freq_levels". Right. See /usr/src/usr.sbin/powerd/powerd.c to see what/how it does. > Once again unless I've missed or mistyped something, and please correct me if I'm wrong,"sysctl > dev.cpu.0.freq_levels" doesn't seem to exist within the machine. Yes, that's the (possibly but not necessarily whole) problem. > If UPGRADING from 6.3-p3 to 7.X will save us all some time with the issues stated in this > thread, then so be it. I don't mind trouble-shooting or customizing issues such > as this, but it may be a bit much given my mobile nature. I can't tell if upgrading would help. Try the GENERIC kernel first - or at least one having cpufreq, but perhaps you left out something else you didn't know was necessary? Unsure whether you can just kldload cpufreq in /boot/loader.conf with your current kernel, but it's an easy test. > Time permiting I'll get to your next suggetion shown; just below this > line: > > > It's also useful to watch the temperature(s) directly over the time, see ug > > acpi_thermal(4) and try logging those sysctls periodically in a script. All good for debugging, but with cpufreq, maybe it will just go? :) [..] cheers, Ian