Re: [PATCH v4 04/12] iommu/vt-d: Add helper to setup pasid nested translation

[Baolu Lu]

On 2023/8/2 15:10, Tian, Kevin wrote:
> > From: Liu, Yi L<yi.l.liu@xxxxxxxxx>
> > Sent: Monday, July 24, 2023 7:13 PM
> >   }
> > +
> > +/**
> > + * intel_pasid_setup_nested() - Set up PASID entry for nested translation.
> > + * This could be used for nested translation based vIOMMU. e.g. guest IOVA
> s/could be/is/

Ack.

> > + * and guest shared virtual address. In this case, the first level page
> > + * tables are used for GVA/GIOVA-GPA translation in the guest, second level
> > + * page tables are used for GPA-HPA translation.
> let's be consistent on using stage-1/stage-2
>
> btw the convention is to have 1-line summary, then the list of
> parameters followed by detail explanation of the function.

This patch just follows the existing code style in this file. Need a
separated patch to cleanup this.

> > + *
> > + * @iommu:      IOMMU which the device belong to
> > + * @dev:        Device to be set up for translation
> > + * @pasid:      PASID to be programmed in the device PASID table
> > + * @domain:     User stage-1 domain nested on a s2 domain
> > + */
> > +int intel_pasid_setup_nested(struct intel_iommu *iommu, struct device
> > *dev,
> > +			     u32 pasid, struct dmar_domain *domain)
> > +{
> > +	struct iommu_hwpt_vtd_s1 *s1_cfg = &domain->s1_cfg;
> > +	pgd_t *s1_gpgd = (pgd_t *)(uintptr_t)domain->s1_pgtbl;
> > +	struct dmar_domain *s2_domain = domain->s2_domain;
> > +	u16 did = domain_id_iommu(domain, iommu);
> > +	struct dma_pte *pgd = s2_domain->pgd;
> > +	struct pasid_entry *pte;
> > +
> > +	if (!ecap_nest(iommu->ecap)) {
> > +		pr_err_ratelimited("%s: No nested translation support\n",
> > +				   iommu->name);
> > +		return -ENODEV;
> > +	}
> -EINVAL

This is in the attach domain path. -EINVAL has the special meaning of
"this domain is not compatible with iommu for the device".

So here, I still think we should return -ENODEV and the caller doesn't
need to retry anymore.

> > +
> > +	if (s2_domain->agaw > iommu->agaw) {
> > +		pr_err_ratelimited("Incompatible agaw %s\n", iommu-
> > > name);
> > +		return -EINVAL;
> > +	}
> there is a duplicated check in intel_nested_attach_dev().

Yeah, should be removed.

Best regards,
baolu