From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757072AbZCWPrK (ORCPT ); Mon, 23 Mar 2009 11:47:10 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753791AbZCWPq4 (ORCPT ); Mon, 23 Mar 2009 11:46:56 -0400 Received: from g4t0014.houston.hp.com ([15.201.24.17]:8796 "EHLO g4t0014.houston.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752908AbZCWPqz (ORCPT ); Mon, 23 Mar 2009 11:46:55 -0400 From: Bjorn Helgaas To: Jesse Barnes Subject: Re: [Bug #11550] pnp: Huge number of "io resource overlap" messages Date: Mon, 23 Mar 2009 09:46:47 -0600 User-Agent: KMail/1.9.10 Cc: Frans Pop , linux-kernel@vger.kernel.org, Ingo Molnar , Thomas Gleixner , Matthew Wilcox , Benjamin Herrenschmidt , Rene Herman References: <200809091250.08620.elendil@planet.nl> <200903041453.52771.bjorn.helgaas@hp.com> <20090319190751.0ded0c4b@hobbes.lan> In-Reply-To: <20090319190751.0ded0c4b@hobbes.lan> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200903230946.48535.bjorn.helgaas@hp.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thursday 19 March 2009 08:07:51 pm Jesse Barnes wrote: > On Wed, 4 Mar 2009 14:53:51 -0700 > Bjorn Helgaas wrote: > > > On Wednesday 04 March 2009 01:17:15 pm Frans Pop wrote: > > > Original thread: > > > http://marc.info/?l=linux-kernel&m=122095745403793&w=4 > > > > Seems like we do need something, but this patch is kind of a klunky > > approach, so I'd like to come up with a better proposal. I don't > > have any better ideas yet, though. > > Patch actually seems pretty reasonable to me, though like we discussed > at kernel summit last year, there are places where a 0 resource is > assumed to mean "not assigned". And clearly we need to do something > here... Anyone else have better ideas than Bjorn's patch below? IIRC, Linus complained that it was ugly and slow to do all those config space reads, and I have to agree with him. I'd like it better if we had some sort of pci_dev "enabled" flag or if we could make it so the pci_dev resources were invalid when the device is disabled. Bjorn > > > diff --git a/drivers/pci/setup-res.c b/drivers/pci/setup-res.c > > > index 32e8d88..e63f800 100644 > > > --- a/drivers/pci/setup-res.c > > > +++ b/drivers/pci/setup-res.c > > > @@ -26,6 +26,28 @@ > > > #include "pci.h" > > > > > > > > > +int pci_resource_enabled(struct pci_dev *dev, int bar) > > > +{ > > > + u16 command = 0; > > > + u32 addr = 0; > > > + > > > + pci_read_config_word(dev, PCI_COMMAND, &command); > > > + > > > + if (pci_resource_flags(dev, bar) & IORESOURCE_IO) > > > + return command & PCI_COMMAND_IO; > > > + > > > + if (command & PCI_COMMAND_MEMORY) { > > > + if (bar == PCI_ROM_RESOURCE) { > > > + pci_read_config_dword(dev, > > > dev->rom_base_reg, &addr); > > > + return addr & PCI_ROM_ADDRESS_ENABLE; > > > + } > > > + > > > + return 1; > > > + } > > > + > > > + return 0; > > > +} > > > + > > > void pci_update_resource(struct pci_dev *dev, int resno) > > > { > > > struct pci_bus_region region; > > > diff --git a/drivers/pnp/quirks.c b/drivers/pnp/quirks.c > > > index 8473fe5..1f37988 100644 > > > --- a/drivers/pnp/quirks.c > > > +++ b/drivers/pnp/quirks.c > > > @@ -247,6 +247,9 @@ static void quirk_system_pci_resources(struct > > > pnp_dev *dev) for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) { > > > unsigned long type; > > > > > > + if (!pci_resource_enabled(pdev, i)) > > > + continue; > > > + > > > type = pci_resource_flags(pdev, i) & > > > (IORESOURCE_IO | > > > IORESOURCE_MEM); if (!type || pci_resource_len(pdev, i) == 0) > > > diff --git a/include/linux/pci.h b/include/linux/pci.h > > > index c927ae9..9848ac2 100644 > > > --- a/include/linux/pci.h > > > +++ b/include/linux/pci.h > > > @@ -870,6 +870,8 @@ static inline int pci_proc_domain(struct > > > pci_bus *bus) } > > > #endif /* CONFIG_PCI_DOMAINS */ > > > > > > +extern int pci_resource_enabled(struct pci_dev *dev, int bar); > > > + > > > #else /* CONFIG_PCI is not enabled */ > > > > > > /* > > > @@ -1050,6 +1052,9 @@ static inline struct pci_dev > > > *pci_get_bus_and_slot(unsigned int bus, unsigned int devfn) > > > { return NULL; } > > > > > > +static inline int pci_resource_enabled(struct pci_dev *dev, int > > > bar) +{ return 0; } > > > + > > > #endif /* CONFIG_PCI */ > > > > > > /* Include architecture-dependent settings and functions */ > > > > > > > > > > > > > >