Re: [RFC PATCH v6 08/11] iommufd: Add vIOMMU provider support

[Aneesh Kumar K.V <aneesh.kumar@kernel.org>]

Jason Gunthorpe <jgg@nvidia.com> writes:

> On Fri, Sep 25, 2026 at 11:38:15AM +0530, Aneesh Kumar K.V wrote:
>> Jason Gunthorpe <jgg@nvidia.com> writes:
>> 
>> >> vIOMMU creation currently dispatches directly through the physical IOMMU
>> >> driver. Some vIOMMU types need an implementation supplied by another
>> >> subsystem, together with private data whose lifetime extends across all
>> >> vIOMMUs using that implementation.
>> >> 
>> >> Introduce a vIOMMU provider that associates implementation operations,
>> >> their module owner and private data. Allow one external provider to
>> >> register for each non-default vIOMMU type.
>> >
>> > This seems much more extensive than I imagined when I proposed this.
>> >
>> > I imagined you'd just have iommufd see there is a TSM attached the pci
>> > device with a function call, and check if that TSM handled the
>> > requested iommu ID.
>> >
>> > No need for a registration system because we already know the only
>> > possible TSM driver that can handle it,
>> >
>> 
>> Agreed, there will be only one vIOMMU provider. But, without a
>> registration mechanism, what would the module dependency look like?
>> 
>> I currently have the TSM driver depending on IOMMUFD. Therefore, the TSM
>> driver needs to tell iommufd which vIOMMU type it provides and register
>> the corresponding provider. IOMMUFD does not call into TSM directly.
>
> Just put the two new ops in:
>
> struct pci_tsm_ops {
> 	size_t (*viommu_get_size)(struct device *dev, enum iommu_viommu_type type);
> 	int (*viommu_init)(struct iommufd_viommu *viommu, struct device *dev,
> 		    struct iommu_domain *parent,
> 		    const struct iommu_user_data *user_data);
> }
>
> And bounce them with a static inline
>
> static inline
> ssize_t pci_tsm_viommu_get_size(struct device *dev, enum iommu_viommu_type type)
> {
> 	const struct pci_tsm_ops *ops;
> 	struct pci_dev *pdev;
>
> 	if (!dev_is_pci(dev))
> 		return -EINVAL;
>
> 	pdev = to_pci_dev(dev);
> 	if (!pdev->tsm)
> 		return -ENODEV;
>
> 	ops = to_pci_tsm_ops(pdev->tsm);
> 	if (!ops->viommu_get_size)
> 		return 0;
> 	return ops->viommu_get_size(dev, type);
> }
>
> No module dependency.
>
> The iommufd side is simple, it just calls this before calling the
> iommu driver. First one to return a size wins and creates the viommu.
>
> We'd want to have a reasonable lifetime model where the the pdev->tsm
> is guarenteed stable while a driver is bound.
>

And the init side will be as below.

bool use_tsm = false;

viommu_size = tsm_viommu_get_size(dev, type);
if (viommu_size) {
	use_tsm = true;
} else {
	viommu_size = iommu_ops->get_viommu_size(dev, type);
	if (!viommu_size)
		return -EOPNOTSUPP;
}


Initialization becomes:


if (use_tsm)
	ret = tsm_viommu_init(viommu, dev, parent, user_data);
else
	ret = iommu_ops->viommu_init(viommu, parent, user_data);


-aneesh