From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262312AbTFOPd2 (ORCPT ); Sun, 15 Jun 2003 11:33:28 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262316AbTFOPd2 (ORCPT ); Sun, 15 Jun 2003 11:33:28 -0400 Received: from caramon.arm.linux.org.uk ([212.18.232.186]:37642 "EHLO caramon.arm.linux.org.uk") by vger.kernel.org with ESMTP id S262312AbTFOPd1 (ORCPT ); Sun, 15 Jun 2003 11:33:27 -0400 Date: Sun, 15 Jun 2003 16:47:16 +0100 From: Russell King To: linux-pcmcia@infradead.org, Linux Kernel List Subject: PCMCIA PATCHES: rsrc_mgr.c Message-ID: <20030615164716.C5417@flint.arm.linux.org.uk> Mail-Followup-To: linux-pcmcia@infradead.org, Linux Kernel List Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i X-Message-Flag: Your copy of Microsoft Outlook is vulnerable to viruses. See www.mutt.org for more details. Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I have three patches which I'm intending pushing to Linus tonight. They are: http://patches.arm.linux.org.uk/pcmcia/pcmcia-20030615-1.diff Remove the racy check_mem_resource() function. Instead, claim the region while we check it, passing a resource structure to the core validation functions. http://patches.arm.linux.org.uk/pcmcia/pcmcia-20030615-2.diff We must always allocate windows below 1MB when a socket driver indicates that it does not have "page registers". Handle this case in rsrc_mgr.c within find_mem_region rather than each use of find_mem_region(). http://patches.arm.linux.org.uk/pcmcia/pcmcia-20030615-3.diff Turn the resource management on its head. Rather than using PCMCIA's resource database as the primary object to allocate resources, use Linux's standard resource allocation instead. When we have a socket on a PCI bus, we always use the PCI resource allocation functions rather than the kernels core resource allocation, so that we can take account of any bridges. The first two should be mostly harmless. The third patch needs greater testing. Unfortunately, the PCI groundwork for this patch was recently reversed, so I just can't be bothered to separate out that change again from the other PCI changes I have. You basically need to make the function non-static, add a declaration to include/linux/pci.h and re-export it. -- Russell King (rmk@arm.linux.org.uk) The developer of ARM Linux http://www.arm.linux.org.uk/personal/aboutme.html