VMD device 28C0 provides native guest passthrough of the VMD endpoint
through the use of shadow registers that provide Host Physical Addresses
to correctly assign bridge windows. A quirk has been added to QEMU's
VFIO quirks to emulate the shadow registers for VMD devices which don't
support this mode natively in hardware.
The VFIO quirk assigns the VMD a subsystem vendor/device ID using the
standard QEMU vendor/device, which are typically only used for emulation
and not VFIO. There are no plans for an emulated VMD controller, but if
one is developed in the future, support for this mode can be added by
emulating the VMD VMLOCK and Shadow MEMBAR registers
Signed-off-by: Jon Derrick <jonathan.derrick@xxxxxxxxx>
---
drivers/pci/controller/vmd.c | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/drivers/pci/controller/vmd.c b/drivers/pci/controller/vmd.c
index dac91d60701d..764404b45ebb 100644
--- a/drivers/pci/controller/vmd.c
+++ b/drivers/pci/controller/vmd.c
@@ -598,6 +598,7 @@ static irqreturn_t vmd_irq(int irq, void *data)
static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)
{
struct vmd_dev *vmd;
+ unsigned long features = id->driver_data;
int i, err;
if (resource_size(&dev->resource[VMD_CFGBAR]) < (1 << 20))
@@ -648,9 +649,14 @@ static int vmd_probe(struct pci_dev *dev, const struct pci_device_id *id)
return err;
}
+ /* VFIO-quirked VMD controllers emulate the Shadow MEMBAR feature */
+ if (dev->subsystem_vendor == PCI_SUBVENDOR_ID_REDHAT_QUMRANET &&
+ dev->subsystem_device == PCI_SUBDEVICE_ID_QEMU)
+ features |= VMD_FEAT_HAS_MEMBAR_SHADOW;
+
spin_lock_init(&vmd->cfg_lock);
pci_set_drvdata(dev, vmd);
- err = vmd_enable_domain(vmd, (unsigned long) id->driver_data);
+ err = vmd_enable_domain(vmd, features);
if (err)
return err;
--
2.18.1
