From owner-freebsd-acpi@FreeBSD.ORG Mon Feb 14 16:41:48 2005 Return-Path: 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 10E4116A4CE; Mon, 14 Feb 2005 16:41:48 +0000 (GMT) Received: from postal3.es.net (postal3.es.net [198.128.3.207]) by mx1.FreeBSD.org (Postfix) with ESMTP id A02C243D49; Mon, 14 Feb 2005 16:41:47 +0000 (GMT) (envelope-from oberman@es.net) Received: from ptavv.es.net ([198.128.4.29]) by postal3.es.net (Postal Node 3) with ESMTP (SSL) id IBA74465; Mon, 14 Feb 2005 08:41:47 -0800 Received: from ptavv (localhost [127.0.0.1]) by ptavv.es.net (Tachyon Server) with ESMTP id 7BB8B5D07; Mon, 14 Feb 2005 08:41:46 -0800 (PST) To: Nate Lawson In-reply-to: Your message of "Sun, 13 Feb 2005 15:33:27 PST." <420FE3C7.6020003@root.org> Date: Mon, 14 Feb 2005 08:41:46 -0800 From: "Kevin Oberman" Message-Id: <20050214164146.7BB8B5D07@ptavv.es.net> cc: acpi@freebsd.org cc: current@freebsd.org Subject: Re: HEADSUP: cpufreq import complete, acpi_throttling changed X-BeenThere: freebsd-acpi@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: ACPI and power management development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 14 Feb 2005 16:41:48 -0000 > Date: Sun, 13 Feb 2005 15:33:27 -0800 > From: Nate Lawson > > Kevin Oberman wrote: > >>Date: Sun, 06 Feb 2005 13:21:32 -0800 > >>From: Nate Lawson > >>Sender: owner-freebsd-acpi@freebsd.org > >> > >> > >>If you have throttling, please test the new configuration to be sure it > >>still works as before. Final upcoming work will be manpage support and > >>bugfixing as necessary. > > > > > > On my T30, throttling has simply vanished. Kernel sources as of this > > afternoon at about 11:00 PST. > > > > sysctl hw.acpi does not list any throttling entries at all. > > It shouldn't, they were merged into the sysctl dev.cpu output as you > mention below. > > > It does list an amazing number of frequency settings, but only 1800 and > > 1200 seem to actually work. Perhaps the others are derived by mixing the > > two capabilities? On the earlier versions of cpufreq I was getting only > > the two frequencies listed along with the 8 throttling states. > > > > Did I miss a message on this? > > Try cvsupping to now. I did some commits about an hour or two ago that > should address throttling not attaching. > > > I am especially concerned because my CPU is now running VERY hot when > > busy. It never used to exceed about 180F and now it quickly jumps to > > 190+ when the system is working (such as a buildkernel). Since it was > > previously running without throttling, I don't understand why things are > > suddenly worse. > > > > Any idea on what is happening? I don't want to fry my T30. > > One person reported Cx states being broken by the cpufreq import. > (Well, actually he got a C3 state that he didn't have before but it > didn't work.) Try setting hw.acpi.cpu.cx_lowest to C1 or something > you're sure works. > > Send me the output of sysctl dev.cpu, dmesg, and devinfo -rv. Things are better now, and it was not really an ACPI issue. For about the millionth time I remind myself: Only change one thing at a time! At the same time that I started running cpufreq and acpi_perf, I also switched from 4BSD to ULE. This is why the system started running so much hotter! Throttling is now working correctly and I can keep my CPU at about 175(F) degrees or a kernel build by lowering the "frequency" from 1800 to 1350. My dmesg shows ACPI throttling setting up fine: cpu0: on acpi0 acpi_perf0: on cpu0 acpi_throttle0: on cpu0 I guess all I have really done is demonstrate that ULE is much more efficient than 4BSD. Thanks for the quick response and sorry for the false alarm. -- R. Kevin Oberman, Network Engineer Energy Sciences Network (ESnet) Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab) E-mail: oberman@es.net Phone: +1 510 486-8634