Re: [PATCH 2/7] iommu/vt-d: Add blocking domain support

[Baolu Lu <baolu.lu@linux.intel.com>]

On 2022/11/4 10:11, Tian, Kevin wrote:
>> From: Lu Baolu <baolu.lu@linux.intel.com>
>> Sent: Thursday, November 3, 2022 1:53 PM
>>
>> +/*
>> + * Clear the page table pointer in context or pasid table entries so that
>> + * all DMA requests without PASID from the device are blocked. If the page
>> + * table has been set, clean up the data structures.
>> + */
>> +static void device_block_translation(struct device *dev)
> 
> given this helper will be used both by existing paths and the new block
> domain, it makes more sense to have it work with existing paths first
> i.e. merging with patch3 and then add block domain after.

Yes. Sounds good.

> 
>> +{
>> +	struct device_domain_info *info = dev_iommu_priv_get(dev);
>> +	struct intel_iommu *iommu = info->iommu;
>> +	unsigned long flags;
>> +
>> +	if (!dev_is_real_dma_subdevice(dev)) {
>> +		if (sm_supported(iommu))
>> +			intel_pasid_tear_down_entry(iommu, dev,
>> +						    PASID_RID2PASID, false);
> 
> Since commit 4140d77a aliases devices don't share pasid table,
> which implies that pci_for_each_dma_alias() is required as did
> in domain_context_clear().

The PCI alias devices have already been covered by the iommu group
concept in the iommu core. On the contrary, I've been thinking about
retiring pci_for_each_dma_alias() in domain_context_clear().



> 
>> +		else
>> +			domain_context_clear(info);
>> +	}
>> +
>> +	if (!info->domain)
>> +		return;
>> +
>> +	spin_lock_irqsave(&info->domain->lock, flags);
>> +	list_del(&info->link);
>> +	spin_unlock_irqrestore(&info->domain->lock, flags);
>> +
>> +	domain_detach_iommu(info->domain, iommu);
>> +	info->domain = NULL;
>> +}
>> +

Best regards,
baolu