From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753493AbZI3KA5 (ORCPT ); Wed, 30 Sep 2009 06:00:57 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752153AbZI3KA4 (ORCPT ); Wed, 30 Sep 2009 06:00:56 -0400 Received: from 8bytes.org ([88.198.83.132]:52316 "EHLO 8bytes.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751987AbZI3KA4 (ORCPT ); Wed, 30 Sep 2009 06:00:56 -0400 Date: Wed, 30 Sep 2009 12:00:59 +0200 From: Joerg Roedel To: Mark Langsdorf Cc: linux-kernel@vger.kernel.org, bhavna.sarathy@amd.com, davej@redhat.com Subject: Re: [PATCH] Option to synchronize P-states for AMD family 0xf Message-ID: <20090930100059.GG29634@8bytes.org> References: <200909291623.56130.mark.langsdorf@amd.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200909291623.56130.mark.langsdorf@amd.com> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Sep 29, 2009 at 04:23:55PM -0500, Mark Langsdorf wrote: > AMD Family 0xf processors (Athlon 64s and Opterons) do not > have P-State or C-State invariant TSCs. Instead, the TSC > increments at the current processor frequency, which makes > it unstable and unreliable if cpufreq is enabled. > > This patch introduces a new command parameter for the > PowerNow! driver called "tscsync". When tscsync is > set to 1, the driver forces all processors in the system > to change frequency as the same time to the same frequency. > This unified change means that all the TSCs in the system > increment at the same rate. Cool feature. Does Linux consider the TSC as stable on these machines even after a longer uptime (say after a week)? Joerg