kernel/git/stable/linux.git/drivers/pci, branch linux-3.9.y

PCI: Fix refcount issue in pci_create_root_bus() error recovery path

2013-07-21T00:16:08+00:00

commit 343df771e671d821478dd3ef525a0610b808dbf8 upstream. After calling device_register(&bridge->dev), the bridge is reference- counted, and it is illegal to call kfree() on it except in the release function. [bhelgaas: changelog, use put_device() after device_register() failure] Signed-off-by: Jiang Liu Signed-off-by: Bjorn Helgaas Signed-off-by: Greg Kroah-Hartman

PCI: Finish SR-IOV VF setup before adding the device

2013-07-21T00:16:08+00:00

commit fbf33f516bdbcc2ab1ba1e54dfb720b0cfaa6874 upstream. Commit 4f535093cf "PCI: Put pci_dev in device tree as early as possible" moves device registering from pci_bus_add_devices() to pci_device_add(). That causes problems for virtual functions because device_add(&virtfn->dev) is called before setting the virtfn->is_virtfn flag, which then causes Xen to report PCI virtual functions as PCI physical functions. Fix it by setting virtfn->is_virtfn before calling pci_device_add(). [Jiang Liu]: Move the setting of virtfn->is_virtfn ahead further for better readability and modify changelog. Signed-off-by: Xudong Hao Signed-off-by: Jiang Liu Signed-off-by: Bjorn Helgaas Signed-off-by: Greg Kroah-Hartman

xen/pcifront: Deal with toolstack missing 'XenbusStateClosing' state.

2013-07-21T00:16:07+00:00

commit 098b1aeaf4d6149953b8f1f8d55c21d85536fbff upstream. There are two tool-stack that can instruct the Xen PCI frontend and backend to change states: 'xm' (Python code with a daemon), and 'xl' (C library - does not keep state changes). With the 'xm', the path to disconnect a single PCI device (xm pci-detach ) is: 4(Connected)->7(Reconfiguring*)-> 8(Reconfigured)-> 4(Connected)->5(Closing*). The * is for states that the tool-stack sets. For 'xl', it is similar: 4(Connected)->7(Reconfiguring*)-> 8(Reconfigured)-> 4(Connected) Both of them also tear down the XenBus structure, so the backend state ends up going in the 3(Initialised) and calls pcifront_xenbus_remove. When a PCI device is plugged back in (xm pci-attach ) both of them follow the same pattern: 2(InitWait*), 3(Initialized*), 4(Connected*)->4(Connected). [xen-pcifront ignores the 2,3 state changes and only acts when 4 (Connected) has been reached] Note that this is for a _single_ PCI device. If there were two PCI devices and only one was disconnected 'xm' would show the same state changes. The problem is that git commit 3d925320e9e2de162bd138bf97816bda8c3f71be ("xen/pcifront: Use Xen-SWIOTLB when initting if required") introduced a mechanism to initialize the SWIOTLB when the Xen PCI front moves to Connected state. It also had some aggressive seatbelt code check that would warn the user if one tried to change to Connected state without hitting first the Closing state: pcifront pci-0: PCI frontend already installed! However, that code can be relaxed and we can continue on working even if the frontend is instructed to be the 'Connected' state with no devices and then gets tickled to be in 'Connected' state again. In other words, this 4(Connected)->5(Closing)->4(Connected) state was expected, while 4(Connected)->.... anything but 5(Closing)->4(Connected) was not. This patch removes that aggressive check and allows Xen pcifront to work with the 'xl' toolstack (for one or more PCI devices) and with 'xm' toolstack (for more than two PCI devices). Acked-by: Bjorn Helgaas Cc: linux-pci@vger.kernel.org [v2: Added in the description about two PCI devices] Signed-off-by: Konrad Rzeszutek Wilk Signed-off-by: Greg Kroah-Hartman

pci: Set dev->dev.type in alloc_pci_dev

2013-06-13T17:49:41+00:00

commit 88e7b167a079f090405ab4390b629b5effdab41a upstream. Set dev->dev.type in alloc_pci_dev so that archs that have their own versions of pci_setup_device get this set properly in order to ensure things like the boot_vga sysfs parameter get created as expected. Signed-off-by: Brian King Acked-by: Bjorn Helgaas Signed-off-by: Michael Ellerman Signed-off-by: Kleber Sacilotto de Souza Signed-off-by: Greg Kroah-Hartman

PCI: Delay final fixups until resources are assigned

2013-05-11T14:18:38+00:00

commit e253aaf0af51c1e4dc7dd3b26ea8e666bf9a2d8d upstream. Commit 4f535093cf "PCI: Put pci_dev in device tree as early as possible" moved final fixups from pci_bus_add_device() to pci_device_add(). But pci_device_add() happens before resource assignment, so BARs may not be valid yet. Typical flow for hot-add: pciehp_configure_device pci_scan_slot pci_scan_single_device pci_device_add pci_fixup_device(pci_fixup_final, dev) # previous location # resource assignment happens here pci_bus_add_devices pci_bus_add_device pci_fixup_device(pci_fixup_final, dev) # new location [bhelgaas: changelog, move fixups to pci_bus_add_device()] Reference: https://lkml.kernel.org/r/20130415182614.GB9224@xanatos Reported-by: David Bulkow Tested-by: David Bulkow Signed-off-by: Yinghai Lu Signed-off-by: Bjorn Helgaas Signed-off-by: Greg Kroah-Hartman

PCI/PM: Fix fallback to PCI_D0 in pci_platform_power_transition()

2013-05-08T03:33:08+00:00

commit 769ba7212f2059ca9fe0c73371e3d415c8c1c529 upstream. Commit b51306c (PCI: Set device power state to PCI_D0 for device without native PM support) modified pci_platform_power_transition() by adding code causing dev->current_state for devices that don't support native PCI PM but are power-manageable by the platform to be changed to PCI_D0 regardless of the value returned by the preceding platform_pci_set_power_state(). In particular, that also is done if the platform_pci_set_power_state() has been successful, which causes the correct power state of the device set by pci_update_current_state() in that case to be overwritten by PCI_D0. Fix that mistake by making the fallback to PCI_D0 only happen if the platform_pci_set_power_state() has returned an error. [bhelgaas: folded in Yinghai's simplification, added URL & stable info] Reference: http://lkml.kernel.org/r/27806FC4E5928A408B78E88BBC67A2306F466BBA@ORSMSX101.amr.corp.intel.com Reported-by: Chris J. Benenati Signed-off-by: Rafael J. Wysocki Signed-off-by: Bjorn Helgaas Acked-by: Yinghai Lu Signed-off-by: Greg Kroah-Hartman

PCI/PM: Disable runtime PM of PCIe ports

2013-04-03T21:54:59+00:00

The runtime PM of PCIe ports turns out to be quite fragile, as in some cases things work while in some other cases they don't and we don't seem to have a good way to determine whether or not they are going to work in advance. For this reason, avoid enabling runtime PM for PCIe ports by keeping their runtime PM reference counters always above 0 for the time being. When a PCIe port is suspended, it can no longer report events like hotplug, so hotplug below the port may not work, as in the bug report below. [bhelgaas: changelog, stable] Reference: https://bugzilla.kernel.org/show_bug.cgi?id=53811 Signed-off-by: Rafael J. Wysocki Signed-off-by: Bjorn Helgaas CC: stable@vger.kernel.org # v3.6+

PCI/ACPI: Always resume devices on ACPI wakeup notifications

2013-04-03T21:49:52+00:00

It turns out that the _Lxx control methods provided by some BIOSes clear the PME Status bit of PCI devices they handle, which means that pci_acpi_wake_dev() cannot really use that bit to check whether or not the device has signalled wakeup. One symptom of the problem is, for example, that when an affected PCI USB controller is runtime-suspended, then plugging in a new USB device into one of the controller's ports will not wake up the controller, which should happen. For this reason, make pci_acpi_wake_dev() always attempt to resume the device it is called for regardless of the device's PME Status bit value (that bit still has to be cleared if set at this point, though). Reported-by: Sarah Sharp Signed-off-by: Rafael J. Wysocki Signed-off-by: Bjorn Helgaas Acked-by: Matthew Garrett CC: stable@vger.kernel.org # v3.7+

PCI: Don't try to disable Bus Master on disconnected PCI devices

2013-04-03T00:03:09+00:00

This is a fix for commit 7897e60227 ("PCI: Disable Bus Master unconditionally in pci_device_shutdown()"). Vivek reported that with this commit, kexec failed because none of his SATA disks came up. A ->shutdown() callback might put the device in D3cold, which means config space is no longer available. [bhelgaas: changelog] Link: https://lkml.org/lkml/2013/3/12/529 Reported-and-Tested-by: Vivek Goyal Signed-off-by: Konstantin Khlebnikov Signed-off-by: Bjorn Helgaas Cc: Rafael J. Wysocki

PCI: Add PCI ROM helper for platform-provided ROM images

2013-03-26T23:19:41+00:00

It turns out that some UEFI systems provide apparently an apparently valid PCI ROM BAR that turns out to contain garbage, so the attempt in 547b52463 to prefer the ROM from the BAR actually breaks a different set of machines. As Linus pointed out, the graphics drivers are probably in the best position to make this judgement, so this basically reverts 547b52463 and f9a37be0f and adds a new helper function. Followup patches will add support to nouveau and radeon for probing this ROM source if they can't find a ROM from some other source. [bhelgaas: added reporter and bugzilla pointers, s/f4eb5ff05/547b52463] Reference: https://bugzilla.redhat.com/show_bug.cgi?id=927451 Reference: http://lkml.kernel.org/r/kg69ef$vdb$1@ger.gmane.org Reported-by: Mantas Mikulėnas Reported-by: Chris Murphy Signed-off-by: Matthew Garrett Signed-off-by: Bjorn Helgaas