From mboxrd@z Thu Jan 1 00:00:00 1970 From: Corentin Chary Subject: Re: eeepc-laptop: bugreport Date: Mon, 7 Dec 2009 22:23:39 +0100 Message-ID: <71cd59b00912071323p58626576y3a9538b375f95861@mail.gmail.com> References: <71cd59b00911150139i5d7377cawda04a17e5e50d05e@mail.gmail.com> <71cd59b00911260703r284b5bfdn3144de4db74b7c3a@mail.gmail.com> <20091126164856.GA19208@srcf.ucam.org> <4B0FB7B9.9040902@tuffmail.co.uk> <20091127172522.GA5422@srcf.ucam.org> <20091127215815.GA9173@srcf.ucam.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Return-path: Received: from mail-vw0-f197.google.com ([209.85.212.197]:51458 "EHLO mail-vw0-f197.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S935596AbZLGVXh (ORCPT ); Mon, 7 Dec 2009 16:23:37 -0500 Received: by vws35 with SMTP id 35so2192440vws.4 for ; Mon, 07 Dec 2009 13:23:39 -0800 (PST) In-Reply-To: <20091127215815.GA9173@srcf.ucam.org> Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: Matthew Garrett Cc: andrej.gelenberg@udo.edu, Alan Jenkins , linux acpi , acpi4asus-user@lists.sourceforge.net On Fri, Nov 27, 2009 at 10:58 PM, Matthew Garrett wrote: > On Fri, Nov 27, 2009 at 09:26:23PM +0100, andrej.gelenberg@udo.edu wrote: > >> The eepc-latop-rfkill should be hard block. If the LED is not on, >> my wlan-card won't transmit. The eeepc-laptop is only driver >> which use hotplug-subsystem to hide the hardware. > > No. The *hardware* disables the PCI device, at which point all reads > return errors and drivers fall over. The eeepc-laptop driver then hides > the device because it's no longer there. > > Now, it's entirely possible that this behaviour is no longer present on > the 1005h. That's fine, and it necessitates changing the behaviour of > the driver. But it's not a reason for removing the functionality > entirely, because the 700s, 900s and earlier 1000 series *do* require > that PCI hotplugging be peformed. > > (Well. There's a separate situation where the PCI runtime power > management code is going to interfere with the way eeepc-laptop does > things, and this functionality is going to need to be added to the PCI > core and removed from eeepc-laptop) Matthew, do you think that we could skip the hotplug code when all rfkill notifier (_SB.PCI0.P0P[5,6,7]) fail to register ? This should work at least for 1005ha. -- Corentin Chary http://xf.iksaif.net