From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261678AbVFKK1Y (ORCPT ); Sat, 11 Jun 2005 06:27:24 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261680AbVFKK1X (ORCPT ); Sat, 11 Jun 2005 06:27:23 -0400 Received: from jurassic.park.msu.ru ([195.208.223.243]:31636 "EHLO jurassic.park.msu.ru") by vger.kernel.org with ESMTP id S261677AbVFKK1O (ORCPT ); Sat, 11 Jun 2005 06:27:14 -0400 Date: Sat, 11 Jun 2005 14:26:58 +0400 From: Ivan Kokshaysky To: linux@horizon.com Cc: linux-kernel@vger.kernel.org, linux-pci@atrey.karlin.mff.cuni.cz, torvalds@osdl.org Subject: Re: PROBLEM: Devices behind PCI Express-to-PCI bridge not mapped Message-ID: <20050611142658.A4952@jurassic.park.msu.ru> References: <20050611053331.9203.qmail@science.horizon.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20050611053331.9203.qmail@science.horizon.com>; from linux@horizon.com on Sat, Jun 11, 2005 at 05:33:31AM -0000 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sat, Jun 11, 2005 at 05:33:31AM -0000, linux@horizon.com wrote: > Actually, that's not possible, because no difference is permitted. Not possible on only on _real_ PCI bus. Example: virtually every laptop has mobile bridge and ISA/LPC/whatever bridge on the same bus segment. Both are subtractive decode devices - impossible situation from classic PCI point of view. But it's just happens because it's _not_ PCI but some proprietary bus, even if it looks like PCI from software point of view. And sort of priority decode is certainly takes place there. Ivan.