[PATCH v7 13/16] iommufd/viommu: Select vIOMMU operations before allocation

[Aneesh Kumar K.V]

Jason Gunthorpe <jgg at nvidia.com> writes:

>> [ ... 62 lines skipped ... ]
>> --- a/drivers/iommu/amd/iommufd.c
>> +++ b/drivers/iommu/amd/iommufd.c
>> @@ -32,13 +32,21 @@ void *amd_iommufd_hw_info(struct device *dev, u32 *length, enum iommu_hw_info_ty
>>  	return hwinfo;
>>  }
>>  
>> -size_t amd_iommufd_get_viommu_size(struct device *dev, enum iommu_viommu_type viommu_type)
>> +static size_t amd_iommufd_get_viommu_size(struct device *dev,
>> +		enum iommu_viommu_type viommu_type)
>>  {
>>  	return VIOMMU_STRUCT_SIZE(struct amd_iommu_viommu, core);
>>  }
>>  
>> -int amd_iommufd_viommu_init(struct iommufd_viommu *viommu, struct iommu_domain *parent,
>> -			    const struct iommu_user_data *user_data)
>> +const struct iommufd_viommu_ops *
>> +amd_iommufd_get_viommu_ops(struct device *dev, enum iommu_viommu_type viommu_type)
>> +{
>> +	return &amd_viommu_ops;
>> +}
>
> This needs to fail not retain the broken behavior if this is getting
> touched..
>

Do you mean return NULL until we add an AMD-specific type equivalent to
IOMMU_VIOMMU_TYPE_ARM_SMMUV3?

modified    drivers/iommu/amd/iommufd.c
@@ -41,7 +41,7 @@ static size_t amd_iommufd_get_viommu_size(struct device *dev,
 const struct iommufd_viommu_ops *
 amd_iommufd_get_viommu_ops(struct device *dev, enum iommu_viommu_type viommu_type)
 {
-	return &amd_viommu_ops;
+	return NULL:
 }
 
 int amd_iommufd_viommu_init(struct iommufd_viommu *viommu,


Alexey,

Could you look at this and suggest the AMD equivalent of
arm_smmu_get_viommu_ops()? We could add it as part of this patch.


-aneesh