From mboxrd@z Thu Jan 1 00:00:00 1970 From: Niel Lambrechts Subject: Re: 2.6.30-rc1 (latest git): thinkpad-acpi: cannot control brightness with hotkeys (BISECTED) Date: Wed, 15 Apr 2009 22:57:12 +0200 Message-ID: <49E64A28.8070402@gmail.com> References: <49E10684.5090101@gmail.com> <20090411150555.8b13ed18.akpm@linux-foundation.org> <20090412161448.GC26502@khazad-dum.debian.net> <49E234ED.5010105@gmail.com> <1239585695.5564.5.camel@localhost.localdomain> <49E3575A.60204@gmail.com> <20090413203450.GC16152@khazad-dum.debian.net> <49E5234E.5070303@gmail.com> <20090415164939.GA27807@srcf.ucam.org> <49E63BCB.9090801@gmail.com> <20090415200125.GA31483@srcf.ucam.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Return-path: Received: from mail-fx0-f158.google.com ([209.85.220.158]:38622 "EHLO mail-fx0-f158.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753248AbZDOU5f (ORCPT ); Wed, 15 Apr 2009 16:57:35 -0400 In-Reply-To: <20090415200125.GA31483@srcf.ucam.org> Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: Matthew Garrett Cc: Henrique de Moraes Holschuh , yakui_zhao , Andrew Morton , "linux.kernel" , "linux-acpi@vger.kernel.org" On 04/15/2009 10:01 PM, Matthew Garrett wrote: > On Wed, Apr 15, 2009 at 09:55:55PM +0200, Niel Lambrechts wrote: > > >> The behaviour seems slightly different from before your 74a365b3f354 >> commit in that /sys/class/backlight remains empty until i915 gets loaded >> for the first time, the contents then persist even if i915 is unloaded. >> Perhaps the directory was populated by a different kernel module, since >> when I did the bisect yesterday the directory would be there on boot >> before I started X or loaded i915. >> > > In the past the directory would be created when the acpi video driver > was loaded, but probably wouldn't work. Doing it that way is a violation > of the opregion spec and causes some machines to break in a horrible > manner, so I'd prefer not to go back to that version... > Cool, I suspected it was intended behaviour... :) So is there any chance of getting this fix merged any time soon (at least before 2.6.30 is out)? > If you use modesetting then you shouldn't need to launch X to get > working brightness control. The problem with the old-style world is that > you won't get interrupts until X comes up. I'd love to use KMS, but I have no idea how to troubleshoot the freeze I'm getting after the main gdm login screen. Or is there still a future i915 milestone(s) to be reached before I can realistically do i915.modeset=1, given the large amount of activity around this in recent times? Thanks for the help! Niel