From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754697Ab1GBNE1 (ORCPT ); Sat, 2 Jul 2011 09:04:27 -0400 Received: from e4.ny.us.ibm.com ([32.97.182.144]:34918 "EHLO e4.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754370Ab1GBNEZ (ORCPT ); Sat, 2 Jul 2011 09:04:25 -0400 Date: Sat, 2 Jul 2011 06:04:17 -0700 From: Ram Pai To: Ben Hutchings Cc: jbarnes@virtuousgeek.org, torvalds@linux-foundation.org, linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org, yinghai@kernel.org, socketcan@hartkopp.net, bhelgaas@google.com, linux@dominikbrodowski.net Subject: Re: [PATCH 0/5 v2] PCI: fix cardbus and sriov regressions Message-ID: <20110702130417.GA4016@ram-laptop> Reply-To: Ram Pai References: <1309477662-18680-1-git-send-email-linuxram@us.ibm.com> <1309561669.2693.20.camel@bwh-desktop> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1309561669.2693.20.camel@bwh-desktop> User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, Jul 02, 2011 at 12:07:48AM +0100, Ben Hutchings wrote: > On Thu, 2011-06-30 at 16:47 -0700, Ram Pai wrote: > > The following patch-set fixes regressions caused by: > > > > the commit "PCI: update bridge resources to get more big ranges when allocating space (again)" > > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=da7822e5ad71ec9b745b412639f1e5e0ba795a20 > > > > patch 1/5: fix calculation of additional resource size for hotplug bridges > > patch 2/5: ability to resize assigned pci-resource > > patch 3/5: make SRIOV BARs resources optional > > patch 4/5: make cardbus bridge resources optional > > patch 5/5: code and terminology cleanup > > > > The regression was caused on some platforms with limited i/o and memory > > resources, the optional resources were allocated ahead of required resources, > > thus starving the latter. The patchset ensures that all the required resources > > are satisfied before any optional resources are satisfied. > > This certainly fixes the problem I originally reported: all the basic > BARs for devices with SR-IOV enabled are mapped. > > When testing this, I noticed that a BAR which we fail to allocate space > for may be left configured with an address range that overlaps that of > other BARs. However, that does not appear to be a regression. overlaps the other SRIOV BARs of the same device? Can you send me the dmesg output? RP