Eric,
> From: Auger Eric <eric.auger@xxxxxxxxxx>
> Sent: Monday, March 30, 2020 5:36 PM
> To: Liu, Yi L <yi.l.liu@xxxxxxxxx>; qemu-devel@xxxxxxxxxx;
> Subject: Re: [PATCH v2 03/22] vfio: check VFIO_TYPE1_NESTING_IOMMU support
>
> Yi,
>
> On 3/30/20 6:24 AM, Liu Yi L wrote:
> > VFIO needs to check VFIO_TYPE1_NESTING_IOMMU support with Kernel before
> > further using it. e.g. requires to check IOMMU UAPI version.
> >
> > Cc: Kevin Tian <kevin.tian@xxxxxxxxx>
> > Cc: Jacob Pan <jacob.jun.pan@xxxxxxxxxxxxxxx>
> > Cc: Peter Xu <peterx@xxxxxxxxxx>
> > Cc: Eric Auger <eric.auger@xxxxxxxxxx>
> > Cc: Yi Sun <yi.y.sun@xxxxxxxxxxxxxxx>
> > Cc: David Gibson <david@xxxxxxxxxxxxxxxxxxxxx>
> > Cc: Alex Williamson <alex.williamson@xxxxxxxxxx>
> > Signed-off-by: Liu Yi L <yi.l.liu@xxxxxxxxx>
> > Signed-off-by: Yi Sun <yi.y.sun@xxxxxxxxxxxxxxx>
> > ---
> > hw/vfio/common.c | 14 ++++++++++++--
> > 1 file changed, 12 insertions(+), 2 deletions(-)
> >
> > diff --git a/hw/vfio/common.c b/hw/vfio/common.c
> > index 0b3593b..c276732 100644
> > --- a/hw/vfio/common.c
> > +++ b/hw/vfio/common.c
> > @@ -1157,12 +1157,21 @@ static void
> vfio_put_address_space(VFIOAddressSpace *space)
> > static int vfio_get_iommu_type(VFIOContainer *container,
> > Error **errp)
> > {
> > - int iommu_types[] = { VFIO_TYPE1v2_IOMMU, VFIO_TYPE1_IOMMU,
> > + int iommu_types[] = { VFIO_TYPE1_NESTING_IOMMU,
> > + VFIO_TYPE1v2_IOMMU, VFIO_TYPE1_IOMMU,
> > VFIO_SPAPR_TCE_v2_IOMMU, VFIO_SPAPR_TCE_IOMMU };
> > - int i;
> > + int i, version;
> >
> > for (i = 0; i < ARRAY_SIZE(iommu_types); i++) {
> > if (ioctl(container->fd, VFIO_CHECK_EXTENSION, iommu_types[i])) {
> > + if (iommu_types[i] == VFIO_TYPE1_NESTING_IOMMU) {
> > + version = ioctl(container->fd, VFIO_CHECK_EXTENSION,
> > + VFIO_NESTING_IOMMU_UAPI);
> > + if (version < IOMMU_UAPI_VERSION) {
> > + info_report("IOMMU UAPI incompatible for nesting");
> > + continue;
> > + }
> > + }
> This means that by default VFIO_TYPE1_NESTING_IOMMU wwould be chosen. I
> don't think this what we want. On ARM this would mean that for a
> standard VFIO assignment without vIOMMU, SL will be used instead of FL.
> This may not be harmless.
>
> For instance, in "[RFC v6 09/24] vfio: Force nested if iommu requires
> it", I use nested only if I detect we have a vSMMU. Otherwise I keep the
> legacy VFIO_TYPE1v2_IOMMU.
>
Good point. I also replied in your patch.
Regards,
Yi Liu
