From Konrad Wilk on xen-devel:
This patch set contains the groundwork and the driver itself for Xen Paravirtualized (PV) domains to use PCI pass-through devices. These patches utilize the Xen-SWIOTLB library (http://lkml.org/lkml/2010/7/27/246) and I don't expect them to get in the 2.6.36 merge window. Some of these groundwork patches had been posted in the past, and gotten some review (and some Acks). Since then work has been done to expand them, fix bugs, squish them, etc, so that they are feature complete. I was not sure if I should carry the Ack/Reviewed-by flags, so I've just reset them to Cc. Also some of the authors of these patches have moved on, and their e-mails are bouncing. Any thoughts of what to do about that? I've squished some of my patches in their work as I want to have the least amount of patches, and I am de facto the maintainer at this point. First of Xen PCI frontend driver can be used by PV guests on hardware that with or without hardware IOMMU. Without an hardware IOMMU you have a potential security hole wherein a guest domain can use the hardware to map pages outside its memory range and slurp pages up. As such, this is more restricted to a Privileged PV domain, aka - device driver domain (similar to Qubes but a poor-man mechanism [1]). I am not sure how many folks are actually interested in the internal workings of the Xen PCI frontend, so I've skipped on that and concentrated on the other patches (X86 related) which would be more interesting to folks I think. The first four expose functionality for module drivers to be able to enumerate and also have the appropriate flag set on PCI devices (_PAGE_IOMAP): [PATCH 04/20] x86/io_apic: add get_nr_irqs_gsi() [PATCH 12/20] x86/PCI: Clean up pci_cache_line_size [PATCH 13/20] x86/PCI: make sure _PAGE_IOMAP it set on pci mappings [PATCH 14/20] x86/PCI: Export pci_walk_bus function. The MSI/MSI-X ones are a mechanism for the MSI/MSI-X framework 'arch_*' to call into a function pointer structure and call from there the appropriate sub-platform MSI implementation. On bare-metal this is still the old code, except that first part of the name has changed from 'arch' to 'native': [PATCH 15/20] x86: Copy-n-paste arch_teardown_msi_irqs from msi.c to io_apic.c. [PATCH 16/20] x86: Introduce x86_msi_ops And lastly, the Xen PCI stub driver. I've put it in the same location where other sub-platform PCI drivers are (say, OLPC). It hooks up to the PCI legacy IRQ setup ('pcibios_enable_irq'), and MSI/MSI-X allocation/de-allocation (via the x86_msi_ops). [PATCH 17/20] xen/x86/PCI: Add support for the Xen PCI subsystem The Xen PCI driver by itself is responsible for hooking up to the PCI configuration read/write methods via the 'pci_scan_bus_parented' call. In essence all pci_conf_read/write would be tunneled via pcifront_bus_[read|write] methods: [PATCH 20/20] xen-pcifront: Xen PCI frontend driver. The rest of the patches are related to the Xen framework and tweaking it to properly work. Those patches have/are in Jeremy's tree too since 2.6.31, but it made more sense to have them logically separated in this patch-set since they are utilized by this driver. And here is the full list of patches and the diffstat: Alex Nixon (3): xen: Don't disable the I/O space x86/PCI: Clean up pci_cache_line_size xen/x86/PCI: Add support for the Xen PCI subsystem Gerd Hoffmann (2): xen: implement pirq type event channels xen: set pirq name to something useful. Jeremy Fitzhardinge (5): xen: define BIOVEC_PHYS_MERGEABLE() x86/io_apic: add get_nr_irqs_gsi() xen: identity map gsi->irqs xen: statically initialize cpu_evtchn_mask_p x86/PCI: make sure _PAGE_IOMAP it set on pci mappings Konrad Rzeszutek Wilk (5): xen: dynamically allocate irq & event structures xen: Find an unbound irq number in reverse order (high to low). xen: Provide a variant of xen_poll_irq with timeout. x86/PCI: Export pci_walk_bus function. x86: Copy-n-paste arch_teardown_msi_irqs from msi.c to io_apic.c. Noboru Iwamatsu (1): xenbus: prevent warnings on unhandled enumeration values Ryan Wilson (1): xen-pcifront: Xen PCI frontend driver. Stefano Stabellini (1): x86: Introduce x86_msi_ops Weidong Han (1): xen: fix shared irq device passthrough Yosuke Iwamatsu (1): xenbus: Xen paravirtualised PCI hotplug support. P.S. [1]. In case you prefer to look at git tree: git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git devel/xen-pcifront-0.5 To get it working, you would need to checkout first: git://git.kernel.org/pub/scm/linux/kernel/git/konrad/swiotlb-2.6.git devel/linux-next and then merge the above mentioned branch in it. [2]: http://qubes-os.org/ which utilizes hardware IOMMU to run separate domains wherein each has specific access to hardware.