[v2] intel_iommu: Enable stage-1 translation for passthrough device

[PATCH v2 04/19] vfio/iommufd: Force creating nested parent domain

Posted by Zhenzhong Duan 4 months, 4 weeks ago

Call pci_device_get_viommu_cap() to get if vIOMMU supports VIOMMU_CAP_STAGE1,
if yes, create nested parent domain which could be reused by vIOMMU to create
nested domain.

Suggested-by: Nicolin Chen <nicolinc@nvidia.com>
Suggested-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
---
 hw/vfio/iommufd.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/hw/vfio/iommufd.c b/hw/vfio/iommufd.c
index d3efef71af..83a632bdee 100644
--- a/hw/vfio/iommufd.c
+++ b/hw/vfio/iommufd.c
@@ -20,6 +20,7 @@
 #include "trace.h"
 #include "qapi/error.h"
 #include "system/iommufd.h"
+#include "hw/iommu.h"
 #include "hw/qdev-core.h"
 #include "hw/vfio/vfio-cpr.h"
 #include "system/reset.h"
@@ -352,6 +353,19 @@ static bool iommufd_cdev_autodomains_get(VFIODevice *vbasedev,
         flags = IOMMU_HWPT_ALLOC_DIRTY_TRACKING;
     }
 
+    /*
+     * If vIOMMU supports stage-1 translation, force to create nested parent
+     * domain which could be reused by vIOMMU to create nested domain.
+     */
+    if (vbasedev->type == VFIO_DEVICE_TYPE_PCI) {
+        VFIOPCIDevice *vdev = container_of(vbasedev, VFIOPCIDevice, vbasedev);
+
+        hw_caps = pci_device_get_viommu_cap(&vdev->pdev);
+        if (hw_caps & VIOMMU_CAP_STAGE1) {
+            flags |= IOMMU_HWPT_ALLOC_NEST_PARENT;
+        }
+    }
+
     if (!iommufd_backend_alloc_hwpt(iommufd, vbasedev->devid,
                                     container->ioas_id, flags,
                                     IOMMU_HWPT_DATA_NONE, 0, NULL,
-- 
2.34.1

Re: [PATCH v2 04/19] vfio/iommufd: Force creating nested parent domain

Posted by Eric Auger 4 months, 4 weeks ago


On 6/20/25 9:17 AM, Zhenzhong Duan wrote:
> Call pci_device_get_viommu_cap() to get if vIOMMU supports VIOMMU_CAP_STAGE1,
> if yes, create nested parent domain which could be reused by vIOMMU to create
could? Is there a choice?
> nested domain.

I would add in the commit msg that it is safe because

hw_caps & VIOMMU_CAP_STAGE1 cannot be set yet because s->flts is forbidden until last patch of the series.

Eric

>
> Suggested-by: Nicolin Chen <nicolinc@nvidia.com>
> Suggested-by: Yi Liu <yi.l.liu@intel.com>
> Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
> ---
>  hw/vfio/iommufd.c | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
>
> diff --git a/hw/vfio/iommufd.c b/hw/vfio/iommufd.c
> index d3efef71af..83a632bdee 100644
> --- a/hw/vfio/iommufd.c
> +++ b/hw/vfio/iommufd.c
> @@ -20,6 +20,7 @@
>  #include "trace.h"
>  #include "qapi/error.h"
>  #include "system/iommufd.h"
> +#include "hw/iommu.h"
>  #include "hw/qdev-core.h"
>  #include "hw/vfio/vfio-cpr.h"
>  #include "system/reset.h"
> @@ -352,6 +353,19 @@ static bool iommufd_cdev_autodomains_get(VFIODevice *vbasedev,
>          flags = IOMMU_HWPT_ALLOC_DIRTY_TRACKING;
>      }
>  
> +    /*
> +     * If vIOMMU supports stage-1 translation, force to create nested parent
> +     * domain which could be reused by vIOMMU to create nested domain.
> +     */
> +    if (vbasedev->type == VFIO_DEVICE_TYPE_PCI) {
> +        VFIOPCIDevice *vdev = container_of(vbasedev, VFIOPCIDevice, vbasedev);
> +
> +        hw_caps = pci_device_get_viommu_cap(&vdev->pdev);
> +        if (hw_caps & VIOMMU_CAP_STAGE1) {
> +            flags |= IOMMU_HWPT_ALLOC_NEST_PARENT;
> +        }
> +    }
> +
>      if (!iommufd_backend_alloc_hwpt(iommufd, vbasedev->devid,
>                                      container->ioas_id, flags,
>                                      IOMMU_HWPT_DATA_NONE, 0, NULL,

RE: [PATCH v2 04/19] vfio/iommufd: Force creating nested parent domain

Posted by Duan, Zhenzhong 4 months, 3 weeks ago

Hi Eric,

>-----Original Message-----
>From: Eric Auger <eric.auger@redhat.com>
>Subject: Re: [PATCH v2 04/19] vfio/iommufd: Force creating nested parent
>domain
>
>
>
>On 6/20/25 9:17 AM, Zhenzhong Duan wrote:
>> Call pci_device_get_viommu_cap() to get if vIOMMU supports
>VIOMMU_CAP_STAGE1,
>> if yes, create nested parent domain which could be reused by vIOMMU to
>create
>could? Is there a choice?

I think no other choice.
If vIOMMU returns VIOMMU_CAP_STAGE1, both smmuv3 and vtd want nested
translation.

>> nested domain.
>
>I would add in the commit msg that it is safe because
>
>hw_caps & VIOMMU_CAP_STAGE1 cannot be set yet because s->flts is forbidden
>until last patch of the series.

Will do.

Thanks
Zhenzhong