Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751356AbaJAGtG (ORCPT ); Wed, 1 Oct 2014 02:49:06 -0400 Received: from mx1.redhat.com ([209.132.183.28]:33917 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750997AbaJAGtC (ORCPT ); Wed, 1 Oct 2014 02:49:02 -0400 Date: Wed, 1 Oct 2014 09:52:27 +0300 From: "Michael S. Tsirkin" To: Marcel Apfelbaum Cc: linux-pci@vger.kernel.org, bhelgaas@google.com, linux-kernel@vger.kernel.org, marcel@redhat.com, alex.williamson@redhat.com Subject: Re: [PATCH] PCI: add kernel parameter to override devid<->driver mapping. Message-ID: <20141001065227.GB31859@redhat.com> References: <1412095115-28982-1-git-send-email-marcel.a@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1412095115-28982-1-git-send-email-marcel.a@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Sep 30, 2014 at 07:38:35PM +0300, Marcel Apfelbaum wrote: > Scanning a lot of devices during boot requires a lot of time. > On other scenarios there is a need to bind a driver to a specific slot. > > Binding devices to pci-stub driver does not work, > as it will not differentiate between devices of the > same type. Using some start scripts is error prone. > > The solution leverages driver_override functionality introduced by > > commit: 782a985d7af26db39e86070d28f987cad21313c0 > Author: Alex Williamson > Date: Tue May 20 08:53:21 2014 -0600 > > PCI: Introduce new device binding path using pci_dev.driver_override > > In order to bind PCI slot 0001:02:03.4 to pci-stub use: > driver[0001:02:03.4]=pci-stub > > Signed-off-by: Marcel Apfelbaum > --- > Notes: > - For the moment only 32 slots can be assigned specific driver. > - I have further ideas on top of this patch based on your reviews. > I thought of: > - Use wildcards to specify entire buses/devices, something like: > driver[0001:02:*.*]=pci-stub > - Use comma to separate several devices: > driver[0001:02:03.4,0001:02:04.0,...]=pci-stub > - Make domain optional: > driver[00:03.0]=pci-stub > > Comments will be appreciated, > Thanks, > Marcel > > Documentation/kernel-parameters.txt | 4 +++ > drivers/pci/bus.c | 1 + > drivers/pci/pci.c | 61 +++++++++++++++++++++++++++++++++++++ > drivers/pci/pci.h | 1 + > 4 files changed, 67 insertions(+) > > diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt > index 5ae8608..c1cbb4c 100644 > --- a/Documentation/kernel-parameters.txt > +++ b/Documentation/kernel-parameters.txt > @@ -2631,6 +2631,10 @@ bytes respectively. Such letter suffixes can also be entirely omitted. > pcie_scan_all Scan all possible PCIe devices. Otherwise we > only look for one device below a PCIe downstream > port. > + driver Provide an override to the devid<->driver mapping > + for a specific slot. > + Bind PCI slot 0001:02:03.4 to pci-stub by: > + driver[0001:02:03.4]=pci-stub > > pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power > Management. > diff --git a/drivers/pci/bus.c b/drivers/pci/bus.c > index 73aef51..7a7ed85 100644 > --- a/drivers/pci/bus.c > +++ b/drivers/pci/bus.c > @@ -245,6 +245,7 @@ void pci_bus_add_device(struct pci_dev *dev) > * are not assigned yet for some devices. > */ > pci_fixup_device(pci_fixup_final, dev); > + pci_device_setup_driver_override(dev); > pci_create_sysfs_dev_files(dev); > pci_proc_attach_device(dev); > > diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c > index b7678be..0a3e07c 100644 > --- a/drivers/pci/pci.c > +++ b/drivers/pci/pci.c > @@ -4426,6 +4426,65 @@ void __weak pci_fixup_cardbus(struct pci_bus *bus) > } > EXPORT_SYMBOL(pci_fixup_cardbus); > > +#define DRIVER_OVERRIDE_NAME_LENGTH 64 > +#define DRIVER_OVERRIDE_MAP_SIZE 32 > + > +struct driver_override_entry { > + unsigned int domain; > + unsigned char bus; > + unsigned int devfn; > + char driver_name[DRIVER_OVERRIDE_NAME_LENGTH]; > +}; > + > +static struct driver_override_entry __initdata > + driver_override_map[DRIVER_OVERRIDE_MAP_SIZE]; > +static int driver_override_num_entries __initdata; > + > +void pci_device_setup_driver_override(struct pci_dev *dev) > +{ > + int i; > + > + for (i = 0; i < driver_override_num_entries; i++) { > + if (pci_domain_nr(dev->bus) != driver_override_map[i].domain || > + dev->bus->number != driver_override_map[i].bus || > + dev->devfn != driver_override_map[i].devfn) > + continue; > + > + dev->driver_override = kstrndup(driver_override_map[i].driver_name, > + DRIVER_OVERRIDE_NAME_LENGTH, > + GFP_KERNEL); You allocate this entry but never free it? Shouldn't allocation failure be handled in some way? > + break; > + } > +} > + > +static void __init pci_parse_driver_override(char *str) > +{ > + unsigned int domain, bus, dev, fn; > + char driver_name[DRIVER_OVERRIDE_NAME_LENGTH]; > + int ret, next_entry; > + char format[] = "[%4x:%2x:%2x.%2x]=%%%ds"; > + > + sprintf(format + 18, "%%%ds", DRIVER_OVERRIDE_NAME_LENGTH); > + ret = sscanf(str, format, &domain, &bus, &dev, &fn, driver_name); > + > + if (ret != 5) { > + pr_err("PCI: Invalid command line: driver%s\n", str); > + return; > + } > + > + if (driver_override_num_entries == DRIVER_OVERRIDE_MAP_SIZE) { > + pr_err("PCI: Driver map overflow - ignoring driver%s\n", str); > + return; > + } > + > + next_entry = driver_override_num_entries++; > + driver_override_map[next_entry].domain = domain; > + driver_override_map[next_entry].bus = bus && 0xff; > + driver_override_map[next_entry].devfn = ((dev & 0x1f) << 3) | (fn & 0x7); > + memcpy(driver_override_map[next_entry].driver_name, > + driver_name, sizeof(driver_name)); > +} > + > static int __init pci_setup(char *str) > { > while (str) { > @@ -4468,6 +4527,8 @@ static int __init pci_setup(char *str) > pcie_bus_config = PCIE_BUS_PEER2PEER; > } else if (!strncmp(str, "pcie_scan_all", 13)) { > pci_add_flags(PCI_SCAN_ALL_PCIE_DEVS); > + } else if (!strncmp(str, "driver", 6)) { > + pci_parse_driver_override(str + 6); > } else { > printk(KERN_ERR "PCI: Unknown option `%s'\n", > str); > diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h > index 0601890..023d78b 100644 > --- a/drivers/pci/pci.h > +++ b/drivers/pci/pci.h > @@ -197,6 +197,7 @@ enum pci_bar_type { > bool pci_bus_read_dev_vendor_id(struct pci_bus *bus, int devfn, u32 *pl, > int crs_timeout); > int pci_setup_device(struct pci_dev *dev); > +void pci_device_setup_driver_override(struct pci_dev *dev); > int __pci_read_base(struct pci_dev *dev, enum pci_bar_type type, > struct resource *res, unsigned int reg); > int pci_resource_bar(struct pci_dev *dev, int resno, enum pci_bar_type *type); > -- > 1.8.3.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/