search for: assigned_device

The OpRegion shouldn''t be mapped 1:1 because the address in the host can''t be used in the guest directly. This patch traps read and write access to the opregion of the Intel GPU config space (offset 0xfc). To work correctly this patch needs a change in hvmloader. HVMloader will allocate 2 pages for the OpRegion and write this address on the config space of the Intel GPU. Qemu

...en-traditional/hw/pass-through.c. and found pt_iomul_init is called if CONFIG_STUBDOM not defined. so should CONFIG_STUBDOM be defined ? Below is part of pass-through.c. #ifndef CONFIG_STUBDOM #define PCI_IOMUL_DEV_PATH "/dev/xen/pci_iomul" static void pt_iomul_init(struct pt_dev *assigned_device, uint8_t r_bus, uint8_t r_dev, uint8_t r_func) { int fd = PCI_IOMUL_INVALID_FD; struct pci_iomul_setup setup = { .segment = 0, .bus = r_bus, .dev = r_dev, .func = r_func, }; fd = open(PCI_IOMUL_DEV_PATH, O_RDWR); if ( fd...

...e_device = PCI_SLOT(ptdev->dev.devfn); + e_intx = pci_intx(ptdev); if (ptdev->msi_trans_en == 0 && ptdev->machine_irq) { @@ -4131,7 +4130,6 @@ static struct pt_dev * register_real_dev if (rc < 0 && machine_irq != 0) { e_device = PCI_SLOT(assigned_device->dev.devfn); - /* fix virtual interrupt pin to INTA# */ e_intx = pci_intx(assigned_device); rc = xc_domain_bind_pt_pci_irq(xc_handle, domid, machine_irq, 0, _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lis...

Hi List, many People seem to be interested in the Graphic-Card Passthrough Feature (for more or less obvious reasons). Official Support is still under development, and i hope not to interfere with it in any (bad) way ... But i remember my own painfull and timeconsuming research when i wanted this feature to work, so i thought perhaps this spares some time for the unpatient users like me :)