Re: [PATCH v3 01/12] iommu/vt-d: Add cache tag assignment interface

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

On 4/23/24 5:01 PM, Tian, Kevin wrote:
>> From: Lu Baolu <baolu.lu@linux.intel.com>
>> Sent: Tuesday, April 16, 2024 4:07 PM
>>
>> @@ -1757,6 +1759,9 @@ int domain_attach_iommu(struct dmar_domain
>> *domain, struct intel_iommu *iommu)
>>   	unsigned long ndomains;
>>   	int num, ret = -ENOSPC;
>>
>> +	if (domain->domain.type == IOMMU_DOMAIN_SVA)
>> +		return 0;
>> +
>>   	info = kzalloc(sizeof(*info), GFP_KERNEL);
>>   	if (!info)
>>   		return -ENOMEM;
>> @@ -1804,6 +1809,9 @@ void domain_detach_iommu(struct dmar_domain
>> *domain, struct intel_iommu *iommu)
>>   {
>>   	struct iommu_domain_info *info;
>>
>> +	if (domain->domain.type == IOMMU_DOMAIN_SVA)
>> +		return;
>> +
>>   	spin_lock(&iommu->lock);
>>   	info = xa_load(&domain->iommu_array, iommu->seq_id);
>>   	if (--info->refcnt == 0) {
> 
> above two are not called for SVA. Why do they start checking
> SVA type now?

domain_detach_iommu() is called in remove_dev_pasid path which covers
SVA domain as well.

I am moving the check to the domain_attach/detach_iommu() helpers so
that the call sites could be generic. I will later move the domain ID
management to the cache tag code, this kind of check will be dropped
then.

Best regards,
baolu