From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S966882AbWKYRiO (ORCPT ); Sat, 25 Nov 2006 12:38:14 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S966883AbWKYRiO (ORCPT ); Sat, 25 Nov 2006 12:38:14 -0500 Received: from [212.33.161.223] ([212.33.161.223]:61312 "EHLO localhost.localdomain") by vger.kernel.org with ESMTP id S966882AbWKYRiN (ORCPT ); Sat, 25 Nov 2006 12:38:13 -0500 From: Al Boldi To: Pavel Machek Subject: Re: [patch] PM: suspend/resume debugging should depend on Date: Sat, 25 Nov 2006 20:40:45 +0300 User-Agent: KMail/1.5 Cc: linux-kernel@vger.kernel.org References: <200611192130.11542.a1426z@gawab.com> <20061124224130.GA4782@ucw.cz> In-Reply-To: <20061124224130.GA4782@ucw.cz> MIME-Version: 1.0 Content-Type: text/plain; charset="windows-1256" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200611252040.45048.a1426z@gawab.com> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Pavel Machek wrote: > On Sun 19-11-06 21:30:11, Al Boldi wrote: > > Mike Galbraith wrote: > > > On Sun, 2006-11-19 at 09:33 -0800, Linus Torvalds wrote: > > > > On Sun, 19 Nov 2006, Chuck Ebbert wrote: > > > > > When doing 'make oldconfig' we should ask about suspend/resume > > > > > debug features when SOFTWARE_SUSPEND is not enabled. > > > > > > > > That's wrong. > > > > > > > > I never use SOFTWARE_SUSPEND, and I think the whole concept is > > > > totally broken. > > > > > > > > Sane people use suspend-to-ram, and that's when you need the suspend > > > > and resume debugging. > > > > > > Here I am wishing I had the _opportunity_ to be sane. With my ATI > > > X850 AGP card, I have no choices except swsusp or reboot. > > > > You make it sound like suspend-to-ram is dependent on a correct display > > adapter to work; it doesn't even work on a machine without a display > > adapter. > > > > see http://bugzilla.kernel.org/show_bug.cgi?id=7037 > > Well, on 'normal' machines, video card is the most common problem. Ok, but it's normally well understood that when debugging, you should try to reduce the number of elements to a bare minimum for the problem to surface. Are you saying that the ACPI STR spec requires a display-card to be attached? If not, then all efforts should go into fixing STR with a bare minimum .config without even a display adapter. Thanks! -- Al