kernel.git/drivers/pci/setup-res.c, branch linux-2.6.31.y

PCI SR-IOV: correct broken resource alignment calculations

2009-08-30T15:37:25Z

An SR-IOV capable device includes an SR-IOV PCIe capability which describes the Virtual Function (VF) BAR requirements. A typical SR-IOV device can support multiple VFs whose BARs must be in a contiguous region, effectively an array of VF BARs. The BAR reports the size requirement for a single VF. We calculate the full range needed by simply multiplying the VF BAR size with the number of possible VFs and create a resource spanning the full range. This all seems sane enough except it artificially inflates the alignment requirement for the VF BAR. The VF BAR need only be aligned to the size of a single BAR not the contiguous range of VF BARs. This can cause us to fail to allocate resources for the BAR despite the fact that we actually have enough space. This patch adds a thin PCI specific layer over the generic resource_alignment() function which is aware of the special nature of VF BARs and does sorting and allocation based on the smaller alignment requirement. I recognize that while resource_alignment is generic, it's basically a PCI helper. An alternative to this patch is to add PCI VF BAR specific information to struct resource. I opted for the extra layer rather than adding such PCI specific information to struct resource. This does have the slight downside that we don't cache the BAR size and re-read for each alignment query (happens a small handful of times during boot for each VF BAR). Signed-off-by: Chris Wright Cc: Ivan Kokshaysky Cc: Linus Torvalds Cc: Matthew Wilcox Cc: Yu Zhao Cc: stable@kernel.org Signed-off-by: Jesse Barnes

Make pci_claim_resource() use request_resource() rather than insert_resource()

2009-08-02T21:10:18Z

This function has traditionally used "insert_resource()", because before commit cebd78a8c5 ("Fix pci_claim_resource") it used to just insert the resource into whatever root resource tree that was indicated by "pcibios_select_root()". So there Matthew fixed it to actually look up the proper parent resource, which means that now it's actively wrong to then traverse the resource tree any more: we already know exactly where the new resource should go. And when we then did commit a76117dfd6 ("x86: Use pci_claim_resource"), which changed the x86 PCI code from the open-coded pr = pci_find_parent_resource(dev, r); if (!pr || request_resource(pr, r) < 0) { to using if (pci_claim_resource(dev, idx) < 0) { that "insert_resource()" now suddenly became a problem, and causes a regression covered by http://bugzilla.kernel.org/show_bug.cgi?id=13891 which this fixes. Reported-and-tested-by: Rafael J. Wysocki Cc: Matthew Wilcox Cc: Andrew Patterson Cc: Linux PCI Signed-off-by: Linus Torvalds

Merge branch 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6

2009-06-22T18:59:51Z

* 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6: (74 commits) PCI: make msi_free_irqs() to use msix_mask_irq() instead of open coded write PCI: Fix the NIU MSI-X problem in a better way PCI ASPM: remove get_root_port_link PCI ASPM: cleanup pcie_aspm_sanity_check PCI ASPM: remove has_switch field PCI ASPM: cleanup calc_Lx_latency PCI ASPM: cleanup pcie_aspm_get_cap_device PCI ASPM: cleanup clkpm checks PCI ASPM: cleanup __pcie_aspm_check_state_one PCI ASPM: cleanup initialization PCI ASPM: cleanup change input argument of aspm functions PCI ASPM: cleanup misc in struct pcie_link_state PCI ASPM: cleanup clkpm state in struct pcie_link_state PCI ASPM: cleanup latency field in struct pcie_link_state PCI ASPM: cleanup aspm state field in struct pcie_link_state PCI ASPM: fix typo in struct pcie_link_state PCI: drivers/pci/slot.c should depend on CONFIG_SYSFS PCI: remove redundant __msi_set_enable() PCI PM: consistently use type bool for wake enable variable x86/ACPI: Correct maximum allowed _CRS returned resources and warn if exceeded ...

Fix pci_claim_resource

2009-06-17T21:04:42Z

Instead of starting from the iomem or ioport roots, start from the parent bus' resources. This fixes a bug where child resources would appear above their parents resources if they had the same size. Signed-off-by: Matthew Wilcox Tested-by: Andrew Patterson Signed-off-by: Linus Torvalds

PCI: improve resource allocation under transparent bridges

2009-06-11T19:04:07Z

We could run out of space under under 4g, but devices under transparent bridges can use 64bit resources, so keep trying on the parent bus until we hit a non-transparent bridge. Impact: better support for assigning unassigned resources Reviewed-by: Ivan Kokshaysky Signed-off-by: Yinghai Lu Signed-off-by: Jesse Barnes

PCI: allow assignment of memory resources with a specified alignment

2009-03-20T17:48:15Z

This patch allows memory resources to be assigned with a specified alignment at boot-time or run-time. The patch is useful when we use PCI pass-through, because page-aligned memory resources are required to securely share PCI resources with guest drivers. If you want to assign the resource at boot time, please set "pci=resource_alignment=" boot parameter. This is format of "pci=resource_alignment=" boot parameter: [@][:]:.[; ...] Specifies alignment and device to reassign aligned memory resources. If is not specified, PAGE_SIZE is used as alignment. PCI-PCI bridge can be specified, if resource windows need to be expanded. This is example: pci=resource_alignment=20@07:00.0;18@0f:00.0;00:1d.7 If you want to assign the resource at run-time, please set "/sys/bus/pci/resource_alignment" file, and hot-remove the device and hot-add the device. For this purpose, fakephp or PCI hotplug interfaces can be used. The format of "/sys/bus/pci/resource_alignment" file is the same with boot parameter. You can use "," instead of ";". For example: # cd /sys/bus/pci # echo -n 20@12:00.0 > resource_alignment # echo 1 > devices/0000:12:00.0/remove # echo 1 > rescan Reviewed-by: Alex Chiang Reviewed-by: Yu Zhao Signed-off-by: Yuji Shimada Signed-off-by: Jesse Barnes

PCI: add a new function to map BAR offsets

2009-01-07T19:13:04Z

Add a function to map a given resource number to a corresponding register so drivers can get the offset and type of device specific BARs. Signed-off-by: Yu Zhao Signed-off-by: Jesse Barnes

PCI: remove unnecessary arg of pci_update_resource()

2009-01-07T19:13:00Z

This cleanup removes unnecessary argument 'struct resource *res' in pci_update_resource(), so it takes same arguments as other companion functions (pci_assign_resource(), etc.). Signed-off-by: Yu Zhao Signed-off-by: Jesse Barnes

PCI: Don't carp about BAR allocation failures in quiet boot

2009-01-07T19:12:42Z

These are easy to trigger (more or less harmlessly) with multiple video cards, since the ROM BAR will typically not be given any space by the BIOS bridge setup. No reason to punish quiet boot for this. Signed-off-by: Adam Jackson Signed-off-by: Jesse Barnes

PCI: use resource_size() everywhere.

2008-10-20T17:54:28Z

This is a cleanup that replaces the resource calculation formula with resource_size(). Signed-off-by: Yu Zhao Signed-off-by: Jesse Barnes