Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756043Ab3HWT1p (ORCPT ); Fri, 23 Aug 2013 15:27:45 -0400 Received: from hydra.sisk.pl ([212.160.235.94]:57733 "EHLO hydra.sisk.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755049Ab3HWT1o (ORCPT ); Fri, 23 Aug 2013 15:27:44 -0400 From: "Rafael J. Wysocki" To: Neil Horman Cc: linux-kernel@vger.kernel.org, Len Brown , linux-acpi@vger.kernel.org, Bjorn Helgaas , Yinghai Lu , Linux PCI Subject: Re: [PATCH] ACPI: Fix osc flag setup ordering to allow pcie hotplug use when available Date: Fri, 23 Aug 2013 21:38:18 +0200 Message-ID: <2074111.WRY7FfU4Ni@vostro.rjw.lan> User-Agent: KMail/4.9.5 (Linux/3.11.0-rc5+; KDE/4.9.5; x86_64; ; ) In-Reply-To: <1377278379-9054-1-git-send-email-nhorman@tuxdriver.com> References: <1377278379-9054-1-git-send-email-nhorman@tuxdriver.com> MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="utf-8" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4062 Lines: 106 [CCs added] Please always send PCI-related material to linux-pci in the first place. The change that broke things for you was actually intentional: commit b8178f130e25c1bdac1c33e0996f1ff6e20ec08e Author: Bjorn Helgaas Date: Mon Apr 1 15:47:39 2013 -0600 Revert "PCI/ACPI: Request _OSC control before scanning PCI root bus" This reverts commit 8c33f51df406e1a1f7fa4e9b244845b7ebd61fa6. so I think we'll need to clean up the ASMP initialization after all. On Friday, August 23, 2013 01:19:39 PM Neil Horman wrote: > Somewhere between 3.9 and 3.10 it seems the order in which pcie and acpi probed > slots for hotplug capabilites got reversed. While this isn't a big deal, it did > uncover a bug in the ACPI bus setup path. Specifically, acpi_pci_root_add calls > pci_acpi_scan_root before setting the osc flags for the device handle. > pci_acpi_scan_root, among other things uses device_is_managed_by_native_pciehp() > to determine if a given slot has pcie hotplug capabilties, whcih checks the > devices OSC_PCI_EXPRESS_NATIVE_HP_CONTROL flag. Since that flag is not set > until after pci_acpi_scan_root_completes, the acpi code never sees that pcie > slots are hotplug capable and configures them all to use acpi instead. > > Fix is pretty simple, just defer the scan until after the osc flags have been > set on the device. Tested by myself and it seems to work well. > > Signed-off-by: Neil Horman > CC: Len Brown > CC: "Rafael J. Wysocki" > CC: linux-acpi@vger.kernel.org > --- > drivers/acpi/pci_root.c | 41 ++++++++++++++++++++--------------------- > 1 file changed, 20 insertions(+), 21 deletions(-) > > diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c > index 5917839..a2c2661 100644 > --- a/drivers/acpi/pci_root.c > +++ b/drivers/acpi/pci_root.c > @@ -437,27 +437,6 @@ static int acpi_pci_root_add(struct acpi_device *device, > flags = base_flags = OSC_PCI_SEGMENT_GROUPS_SUPPORT; > acpi_pci_osc_support(root, flags); > > - /* > - * TBD: Need PCI interface for enumeration/configuration of roots. > - */ > - > - /* > - * Scan the Root Bridge > - * -------------------- > - * Must do this prior to any attempt to bind the root device, as the > - * PCI namespace does not get created until this call is made (and > - * thus the root bridge's pci_dev does not exist). > - */ > - root->bus = pci_acpi_scan_root(root); > - if (!root->bus) { > - dev_err(&device->dev, > - "Bus %04x:%02x not present in PCI namespace\n", > - root->segment, (unsigned int)root->secondary.start); > - result = -ENODEV; > - goto end; > - } > - > - /* Indicate support for various _OSC capabilities. */ > if (pci_ext_cfg_avail()) > flags |= OSC_EXT_PCI_CONFIG_SUPPORT; > if (pcie_aspm_support_enabled()) { > @@ -520,6 +499,26 @@ static int acpi_pci_root_add(struct acpi_device *device, > "(_OSC support mask: 0x%02x)\n", flags); > } > > + /* > + * TBD: Need PCI interface for enumeration/configuration of roots. > + */ > + > + /* > + * Scan the Root Bridge > + * -------------------- > + * Must do this prior to any attempt to bind the root device, as the > + * PCI namespace does not get created until this call is made (and > + * thus the root bridge's pci_dev does not exist). > + */ > + root->bus = pci_acpi_scan_root(root); > + if (!root->bus) { > + dev_err(&device->dev, > + "Bus %04x:%02x not present in PCI namespace\n", > + root->segment, (unsigned int)root->secondary.start); > + result = -ENODEV; > + goto end; > + } > + > pci_acpi_add_bus_pm_notifier(device, root->bus); > if (device->wakeup.flags.run_wake) > device_set_run_wake(root->bus->bridge, true); > -- I speak only for myself. Rafael J. Wysocki, Intel Open Source Technology Center. -- 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/