From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1762213AbZCaPX0 (ORCPT ); Tue, 31 Mar 2009 11:23:26 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1758854AbZCaPXJ (ORCPT ); Tue, 31 Mar 2009 11:23:09 -0400 Received: from g4t0014.houston.hp.com ([15.201.24.17]:14817 "EHLO g4t0014.houston.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758736AbZCaPXH (ORCPT ); Tue, 31 Mar 2009 11:23:07 -0400 Subject: [PATCH 00/15] PCI Hotplug: cpqphp cleanups To: jbarnes@virtuousgeek.org From: Alex Chiang Cc: linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org Date: Tue, 31 Mar 2009 09:23:05 -0600 Message-ID: <20090331150158.26622.45423.stgit@bob.kio> User-Agent: StGit/0.14.3.347.g594a MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Jesse, This is my first pass at cleaning up cpqphp. Why? I dunno, I found some old hardware and thought it might be fun to clean up a crufty old driver. This series is mostly whitespace and style fixes. A few functions have been reworked so they aren't indented quite so much. I expect that the first few patches will be very hard to read because there's so much code movement, but it does help with readability towards the end of the series. At the end of the series, we start to see some more interesting changes. For one, we no longer call pcibios_get_irq_routing_table nearly as often. The other is removing the pci_find_slot() interface. I'm still in the process of untangling exactly what the lifetime rules are for a pci_dev in cpqphp, but this is a good step forward, and also finally gets rid of the last in-tree caller of pci_find_slot(), which would be nice for 2.6.30. Also, this means that cpqphp no longer depends on PCI_LEGACY. The current functional state of cpqphp is broken. We are able to load the module and successfully power off cards. We have some issues when we try to power them back on though, because commit b81d988c "broke" cpqphp and we get BAR collisions when we try to hot-add. However, I use "broke" very loosely, as I agree with Bjorn's commit, and cpqphp is doing horrible things like essentially duplicating a lot of the PCI core, incorrectly, I might add. Once I rip out some more of the PCI code in cpqphp, I expect hot-add to work again. I plan on untangling that in the next phase of cleanups. I've successfully tested my changes on my DL580g2. No functional regressions (compared to behavior since b81d988c). Anyhow, please consider applying this series for the current merge window. I'd like to be able to turn off PCI_LEGACY. :) Thanks! /ac --- Alex Chiang (15): PCI: remove deprecated pci_find_slot() interface PCI Hotplug: cpqphp: don't use pci_find_slot() PCI Hotplug: cpqphp: constify slot_name() PCI Hotplug: cpqphp: eliminate dead code - PCI_ScanBusNonBridge PCI Hotplug: cpqphp: clean up accesses to pcibios_get_irq_routing_table() PCI Hotplug: cpqphp: refactor cpqhp_save_config PCI Hotplug: cpqphp: style cleanups PCI Hotplug: cpqphp: refactor cpqphp_save_slot_config PCI Hotplug: cpqphp: clean up cpqphp_ctrl.c PCI Hotplug: cpqphp: refactor cpqhp_probe PCI Hotplug: cpqphp: eliminate stray braces PCI Hotplug: cpqphp: remove useless prototypes in cpqphp_core.c PCI Hotplug: cpqphp: obey 80 column convention in cpqphp.h PCI Hotplug: cpqphp: fix comment style PCI Hotplug: cpqphp: stray whitespace cleanups drivers/pci/hotplug/Kconfig | 2 drivers/pci/hotplug/cpqphp.h | 167 +++-- drivers/pci/hotplug/cpqphp_core.c | 1101 +++++++++++++++++------------------- drivers/pci/hotplug/cpqphp_ctrl.c | 371 ++++++------ drivers/pci/hotplug/cpqphp_nvram.c | 97 ++- drivers/pci/hotplug/cpqphp_pci.c | 599 ++++++++++---------- drivers/pci/search.c | 30 - include/linux/pci.h | 8 8 files changed, 1162 insertions(+), 1213 deletions(-)