From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751508AbaEQWTG (ORCPT ); Sat, 17 May 2014 18:19:06 -0400 Received: from mail.active-venture.com ([67.228.131.205]:54156 "EHLO mail.active-venture.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750964AbaEQWTF (ORCPT ); Sat, 17 May 2014 18:19:05 -0400 X-Originating-IP: 108.223.40.66 Message-ID: <5377E04E.2070505@roeck-us.net> Date: Sat, 17 May 2014 15:18:54 -0700 From: Guenter Roeck User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.4.0 MIME-Version: 1.0 To: Jean Delvare CC: =?ISO-8859-1?Q?Pali_Roh=E1r?= , linux-kernel@vger.kernel.org Subject: Re: Dell Latitude E6440 & i8k References: <201405162037.41782@pali> <20140516211117.0c766287@endymion.delvare> <53777F72.3000907@roeck-us.net> <20140517230940.25c36f07@endymion.delvare> In-Reply-To: <20140517230940.25c36f07@endymion.delvare> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 05/17/2014 02:09 PM, Jean Delvare wrote: > On Sat, 17 May 2014 08:25:38 -0700, Guenter Roeck wrote: >> On 05/16/2014 12:11 PM, Jean Delvare wrote: >>> Load the i8k driver with fan_mult=1. >> >> Would it make sense to change the default multiplier to 1 ? >> Lots of people have problems with it, and trying to figure out >> affected machines one by one would be an all but impossible task. > > That would cause a regression on many (presumably older) machines. I > doubt this is acceptable. One option would be to use the ACPI year to > change the default, if indeed all new machines need fan_mult=1. I don't > know if this is the case. > > One this I had in mind was auto-detecting the scaling factor. AFAIK > only 30 and 1 are possible values, so any value above ~300 would imply > scaling factor of 1 (30 would make it > 9000 RPM which is unrealistic.) > But I don't know if we can actually do that, as such a simple heuristic > could easily fail is the fan is stopped (30 * 0 == 1 * 0) or if the > returned raw speed is temporarily unreliable for whatever reason. > Sounds like an idea. We could make the cutoff higher, such as 500 or even 1000. I am not much concerned about 0 rpm - the code could simply check the returned rpm and adjust the scaling factor to 1 if the reading is too high. Since 30*0 is still 0, there is no problem if the fan is stopped. > I have to admit that working on a reverse engineered driver for > hardware I don't even have isn't quite at the top of my to-do list. > ;-). We have several of those systems, so there is at least some interest on my side. If I have time, I'll play around with it. The driver does need major cleanup, though, so that may take a while. Guenter