From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S265006AbUELEPh (ORCPT ); Wed, 12 May 2004 00:15:37 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S264965AbUELEPh (ORCPT ); Wed, 12 May 2004 00:15:37 -0400 Received: from web14929.mail.yahoo.com ([216.136.225.94]:28854 "HELO web14929.mail.yahoo.com") by vger.kernel.org with SMTP id S265006AbUELEPg (ORCPT ); Wed, 12 May 2004 00:15:36 -0400 Message-ID: <20040512041534.55600.qmail@web14929.mail.yahoo.com> Date: Tue, 11 May 2004 21:15:34 -0700 (PDT) From: Jon Smirl Subject: first PCI probe vs hotplug app on initramfs To: lkml 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 Over on the fbdev and dri list were having a big discussion about where to do mode setting of the graphics hardware. The normal kernel boot process follow these steps... 1) kernel probes PCI hardware for built in drivers 2) driver gets probe and triggers hotplug event 3) what can happen here??? 4) hotplug event runs early user space program on initramfs The discussion is centering around what are the side effects of moving the initialization of the display monitor from step #2 to step #4. There is no consensus on what can happen in step #3. Do these early hotplug events run synchronously or asynchronously? Will all of the other drivers initialize before running the event? ===== Jon Smirl jonsmirl@yahoo.com __________________________________ Do you Yahoo!? Yahoo! Movies - Buy advance tickets for 'Shrek 2' http://movies.yahoo.com/showtimes/movie?mid=1808405861