From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750956AbWKBUzI (ORCPT ); Thu, 2 Nov 2006 15:55:08 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750957AbWKBUzI (ORCPT ); Thu, 2 Nov 2006 15:55:08 -0500 Received: from smtp.osdl.org ([65.172.181.4]:59817 "EHLO smtp.osdl.org") by vger.kernel.org with ESMTP id S1750956AbWKBUzG (ORCPT ); Thu, 2 Nov 2006 15:55:06 -0500 Date: Thu, 2 Nov 2006 12:54:02 -0800 (PST) From: Linus Torvalds To: "Rafael J. Wysocki" cc: Adrian Bunk , Andrew Morton , LKML , Laurent Riffard , Rajesh Shah , toralf.foerster@gmx.de, Jeff Garzik , Pavel Machek , Greg KH , Auke Kok , Jesse Brandeburg Subject: Re: 2.6.19-rc4: known unfixed regressions In-Reply-To: <200611022102.02302.rjw@sisk.pl> Message-ID: References: <20061031195654.GV27968@stusta.de> <200611022102.02302.rjw@sisk.pl> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 2 Nov 2006, Rafael J. Wysocki wrote: > Can we please add the following two to the list of known regressions: > > http://bugzilla.kernel.org/show_bug.cgi?id=7082 Ok, I think I'll just revert it. Decoding the PCI IO range is fine - even if a driver has detached, the kernel knows where the PCI devices are, and won't re-use the range. So while the patch that triggers the problem seems valid in itself, it's probably not worth the pain to apply it at this point. So I think I'll revert it - the rationale for the patch was fairly weak. Greg, or would you prefer to do the honors? > http://bugzilla.kernel.org/show_bug.cgi?id=7207 This one is apparently purely an e1000 driver problem. Can't help you with that one. Although I find it suspicious that the "e1000_resume()" path doesn't seem to be calling "e1000_power_up_phy()" before e1000_up(). Linus