On Mon, Jun 15, 2015 at 05:06:39PM +0000, jakeo@xxxxxxxxxxxxx wrote: > From: Jake Oshins <jakeo@xxxxxxxxxxxxx> > > This patch supplies a new driver, hv_pcifront, which exposes new root PCI > buses. When a Hyper-V VM running Linux is offered a paravirtual PCI bus > in order to expose a device passed through on that bus, this drver > registers a new bus with the PCI driver. This new bus is bus "zero" > within a new PCI domain. This is done so that there's no overlap between > the emulated, or fully virtualized, PCI that may be in the VM and a real > PCIe device that will be exposed. > > This PCI front-end only supports PCIe devices which do not use I/O BARs and > which do not need a legacy INTx# interrupt (relying on MSI or MSI-X). > Multifunction devices are supported, but devices with Type 1 or Type 2 > config headers are not supported. (These are bridges to other PCI buses.) > > The content of this patch differs from the one sent previously in that it > incorporates feedback related to unused code (which has been removed) and > GPL license (which has been updated.) It also removes Kconfig lines that > had been copied from the Xen PCI front-end entry, but which shouldn't be > used here. This paragraph goes below the --- line, we don't need it in the changelog :( _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
