From owner-freebsd-bugs@FreeBSD.ORG Tue Mar 12 22:03:25 2013 Return-Path: Delivered-To: freebsd-bugs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by hub.freebsd.org (Postfix) with ESMTP id 4D8FA21E; Tue, 12 Mar 2013 22:03:25 +0000 (UTC) (envelope-from kob6558@gmail.com) Received: from mail-ob0-x232.google.com (mail-ob0-x232.google.com [IPv6:2607:f8b0:4003:c01::232]) by mx1.freebsd.org (Postfix) with ESMTP id 00F70AC4; Tue, 12 Mar 2013 22:03:24 +0000 (UTC) Received: by mail-ob0-f178.google.com with SMTP id wd20so378754obb.9 for ; Tue, 12 Mar 2013 15:03:24 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:x-received:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type; bh=8taRkigNtAMlx7OlAmuodJNwYtOirB5jyIsR40elJT8=; b=vkAqZ390mjUF1FLjcNmgUUQIzU9qxRnI0Ii/PIZQHz+GKpyo3ijhykGMVzPYMyrpKj XNQL8JUhxQlJSc2zwPtPw1r7MEJgZXAsDCgrA75r5shgaGcV3+EyRa4HuRIXNJC8MRs0 1g4NsEaZFCOkVtM9YIYl/JyN/2rt6jjwgFClBUgFpgS56nldqSeY+lk7q5DTL2B+CYGT /Yzxojv/OU1OJdq302sj63KbV5f/KaKdAYog56WYlzgWzN2Y6q4jtV7ARR6bGsPF4C2f KK7Evcx61ERwPyNkepZgBPTQZwGH1Mmcoy/Qh1zmlz6iDNVLyWYsO29KT13zy23Z4VGH oQ8w== MIME-Version: 1.0 X-Received: by 10.60.22.69 with SMTP id b5mr14161672oef.38.1363125804558; Tue, 12 Mar 2013 15:03:24 -0700 (PDT) Sender: kob6558@gmail.com Received: by 10.76.11.165 with HTTP; Tue, 12 Mar 2013 15:03:24 -0700 (PDT) In-Reply-To: <201303122116.r2CLGLOJ082410@freefall.freebsd.org> References: <201303122116.r2CLGLOJ082410@freefall.freebsd.org> Date: Tue, 12 Mar 2013 15:03:24 -0700 X-Google-Sender-Auth: PtOXDiKU0YQScAs1cQOsvU5aVwg Message-ID: Subject: Re: kern/165381: [cpufreq] powerd(8) eats CPUs for breakfast From: Kevin Oberman To: crees@freebsd.org Content-Type: text/plain; charset=UTF-8 X-Content-Filtered-By: Mailman/MimeDel 2.1.14 Cc: freebsd-bugs@freebsd.org, freebsd-acpi@freebsd.org X-BeenThere: freebsd-bugs@freebsd.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: Bug reports List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 12 Mar 2013 22:03:25 -0000 On Tue, Mar 12, 2013 at 2:16 PM, wrote: > Synopsis: [cpufreq] powerd(8) eats CPUs for breakfast > > Responsible-Changed-From-To: freebsd-bugs->freebsd-acpi > Responsible-Changed-By: crees > Responsible-Changed-When: Tue Mar 12 21:16:07 UTC 2013 > Responsible-Changed-Why: > This appears to be a rather serious issue-- anyone know about powerd on > amd? > > http://www.freebsd.org/cgi/query-pr.cgi?pr=165381 > _______________________________________________ > > I suspect that you are getting bitten by issues of conflict between C-states and cpufreq. I suggest that you disable TCC (which is intended for thermal control, not power management) and see if that helps. TCC/throttling is designed to slow down the CPU dramatically when the CPU temperature approaches unsafe levels and normally is relevant onl,y when the CPU is busy. Throttling, either by TCC, also called P4TCC even though its only association with the Pentium4 is that it was introduced on those chips, or the older manual throttling can reduce the CPU to using only 1 in 8 clock cycles if the CPU is idle. This should be fine, but when combined with SpeedStep or the AMD equivalent (PowerNOW!) and C-states, it can cause the system to deadlock. It essentially goes to sleep and never wakes up again, the effective equivalent of halting. Since C-states are the biggest win for power saving, it is especially important that these be used. To disable TCC and throttling, add the following two lines to your /boot/loader.conf file: hint.p4tcc.0.disabled=1 hint.acpu_throttle.0.disabled=1 You should also confirm that the hw.acpi.cx_lowest sysctl is set to "C8". Since both SpeedStep/PowerNOW! and C-states do reduce both power consumption and idle temperature quite effectively, they should be used. -- R. Kevin Oberman, Network Engineer E-mail: rkoberman@gmail.com