Re: [PATCH v2 2/6] iommu/vt-d: Remove scalable mode context entry setup from attach_dev

[Baolu Lu]

On 12/8/23 4:56 PM, Tian, Kevin wrote:
> > From: Lu Baolu <baolu.lu@xxxxxxxxxxxxxxx>
> > Sent: Tuesday, December 5, 2023 9:22 AM
> >
> >   static int domain_context_mapping_one(struct dmar_domain *domain,
> >   				      struct intel_iommu *iommu,
> > -				      struct pasid_table *table,
> >   				      u8 bus, u8 devfn)
>
> since this is called in legacy now it's clearer to add a legacy_ prefix
> to this and other related functions.

Okay, that would make the code more readable.

> > -	/* PASID table is mandatory for a PCI device in scalable mode. */
> > -	if (sm_supported(iommu) && !dev_is_real_dma_subdevice(dev)) {
> > -		/* Setup the PASID entry for requests without PASID: */
> > -		if (hw_pass_through && domain_type_is_si(domain))
> > -			ret = intel_pasid_setup_pass_through(iommu,
> > -					dev, IOMMU_NO_PASID);
> > -		else if (domain->use_first_level)
> > -			ret = domain_setup_first_level(iommu, domain, dev,
> > -					IOMMU_NO_PASID);
> > -		else
> > -			ret = intel_pasid_setup_second_level(iommu,
> > domain,
> > -					dev, IOMMU_NO_PASID);
> > -		if (ret) {
> > -			dev_err(dev, "Setup RID2PASID failed\n");
> > -			device_block_translation(dev);
> > -			return ret;
> > -		}
> > -	}
> > +	if (dev_is_real_dma_subdevice(dev))
> > +		return 0;
>
> is it a functional change? old code doesn't early exit for such device.

This type of device is aliased to a RID on another bus. Therefore,
there's no need to set up the context and PASID table for them. I should
make the change mentioned above in a separate patch to make the
background clear.

Best regards,
baolu