From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750726AbWFKSHr (ORCPT ); Sun, 11 Jun 2006 14:07:47 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750732AbWFKSHr (ORCPT ); Sun, 11 Jun 2006 14:07:47 -0400 Received: from anchor-post-35.mail.demon.net ([194.217.242.85]:57093 "EHLO anchor-post-35.mail.demon.net") by vger.kernel.org with ESMTP id S1750726AbWFKSHq (ORCPT ); Sun, 11 Jun 2006 14:07:46 -0400 Message-ID: <448C5BF0.7070601@superbug.demon.co.uk> Date: Sun, 11 Jun 2006 19:07:44 +0100 From: James Courtier-Dutton User-Agent: Thunderbird 1.5.0.4 (X11/20060609) MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: Video drivers and System Management mode. X-Enigmail-Version: 0.94.0.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi, I know we all laugh about the windows blue screen of death, but to be fair, when Linux oops, it is not even able to display anything on the screen, unless in dump terminal mode. I.e. Not X or some other GUI. Are there any plans to implement a sort of interactive system management mode, that would pop up a window when Linux oops. Something like the program called SoftICE for windows would be a nice addition to Linux, and help with kernel development. For those who don't know what SoftICE does, when windows wishes to display a blue screen of death (BSOD), SoftICE pops up a window showing a disassembly of the point where the crash happened, and allows users to type commands on the keyboard that will display extra information. SoftICE also has a hot key, so the user can cause it to pop up at any time they wish. During the pop-up, the entire OS is halted, and the only commands possible are within the pop-up window. Another command exits the pop-up and returns control to the OS. I imagine that this feature would be tightly linked with the work currently being done to unite all the different video drivers. Unfortunately, I think work could take a long time. For example, on my nvidia based card, even the kernel frame buffer drivers either fail to display anything, or fail to scroll up the screen when a new printk happens. I am not attending the KS this year, but I do hope that all the interested parties can use a BOF session to come to some agreement on the way forward, I would then be able to make useful contributions to at least get my system working with whatever new model is decided on. James