From mboxrd@z Thu Jan 1 00:00:00 1970 From: Zhang Rui Subject: Re: [PATCH 6/8] [-mm] ACPI: add ACPI Processor_throttling sysfs interface Date: Thu, 22 Mar 2007 14:12:44 +0800 Message-ID: <1174543964.8833.159.camel@localhost.localdomain> References: <1174382493.8833.85.camel@localhost.localdomain> <200703220055.38215.lenb@kernel.org> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from mga01.intel.com ([192.55.52.88]:37314 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751275AbXCVGNV (ORCPT ); Thu, 22 Mar 2007 02:13:21 -0400 In-Reply-To: <200703220055.38215.lenb@kernel.org> Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: Len Brown Cc: "linux-acpi@vger" On Thu, 2007-03-22 at 00:55 -0400, Len Brown wrote: > On Tuesday 20 March 2007 05:21, Zhang Rui wrote: > > From: Zhang Rui > > > > Add ACPI Processor throttling control sysfs interface. > > > > Attribute Mode Description > > state_count RO Maximum throttling state supported by this > > processor. > > active_state RW the current throttling state. > > I like this better than the procfs interface. > In particular, this is simple levels and doesn't get tangled up in percentages. > > However, this gives the user the impression that they can actually > change these values and they will stick -- which may be a lie. > > Firmware sometimes messes with throttling behind our back. > Thermal management can change throttling levels on us > and it isn't clear what happens to the user-request in that context. > Finally, something like the p4-clockmod driver can take over throttling > at the request of the cpufreq sub-system. > > Plus, we have the _TCP enhancements coming from Luming. > Yep. I've discussed with Luming and we only need to add another attribute like "platform_limit". It exports a value "n" which means that only Tn~Tmax throttling states are available currently. > I think we need to sort through these issues before we > can possibly propose the appropriate API for exporting > processor throttling control to user-space. > Agree. Maybe we need a processor throttling state driver just like we do for processor power/performance state control(cpui_dle and cpu_freq). Thanks, Rui