Re: [PATCH v2 10/12] iommu/amd: Add support for nested domain allocation

[Suthikulpanit, Suravee]

On 10/6/2025 9:49 AM, Jason Gunthorpe wrote:
> On Wed, Oct 01, 2025 at 06:09:52AM +0000, Suravee Suthikulpanit wrote:
> ....
> > +
> > +	ret = iommu_copy_struct_from_user(&gdte, user_data,
> > +					  IOMMU_HWPT_DATA_AMD_GUEST,
> > +					  dte);
> > +	if (ret)
> > +		return ERR_PTR(ret);
> > +
> > +	ndom = kzalloc(sizeof(*ndom), GFP_KERNEL);
> > +	if (IS_ERR(ndom))
> > +		return ERR_PTR(-ENOMEM);
> > +
> > +	ndom->domain.ops = &nested_domain_ops;
> > +	ndom->domain.type = IOMMU_DOMAIN_NESTED;
> > +	memcpy(&ndom->guest_dte, &gdte, sizeof(struct dev_table_entry));
> > +
> > +	/* Due to possible aliasing issue use per-device nested domain ID */
> > +	ndom->id = amd_iommu_pdom_id_alloc();
>
> I've forgotten the details, but doesn't the gdet have the virtual
> domain ID in side it? Shouldn't this be going to the viommu struct and
> mapping virtual domain IDs to physical ones so they can be shared if
> the guest says it is safe to share them? I guess that is an
> optimization, but it should have a note here explaining it.

The gDTE[DomainID] field contains guest Domain ID (gDomID). The host 
IOMMU driver uses the gDomId and guest ID (gid) to index the Domain ID 
mapping table, and store the host Domain ID (hDomID) in the table entry. 
This data structure is required by hw to translation gDomID->hDomID to 
virtualize guest invalidation command. This will be part of the upcoming 
series to enable hw-vIOMMU.

This ndom->id is the hDomID, which is currently allocated per-device to 
avoid TLB aliasing i.e. A guest w/ multiple pass-through devices w/ the 
same hDomID (same stage 2 table) and different stage-1 tables with same 
PASID. IOMMU would use the same TLB tag, which results in TLB aliasing 
issue. Therefore, we workaround the issue by allocating per-device 
hDomID for nested domain.

Thanks,
Suravee



Thanks,
Suravee