Re: [PATCH v2 0/2] Two enhancements to iommu_at[de]tach_device_pasid()

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

On 3/29/24 10:12 AM, Duan, Zhenzhong wrote:
> 
>> -----Original Message-----
>> From: Liu, Yi L<yi.l.liu@intel.com>
>> Subject: [PATCH v2 0/2] Two enhancements to
>> iommu_at[de]tach_device_pasid()
>>
>> There are minor mistakes in the iommu set_dev_pasid() and
>> remove_dev_pasid()
>> paths. The set_dev_pasid() path updates the group->pasid_array first, and
>> then call into remove_dev_pasid() in error handling when there are devices
>> within the group that failed to set_dev_pasid.
> Not related to this patch, just curious in which cases some of the devices
> In same group failed to set_dev_pasid while others succeed?

The failure cases could be checked in the set_dev_pasid implementation
of the individual iommu driver. For x86 platforms, which are PCI fabric-
based, there's no such case as PCI/PASID requires a singleton iommu
group.

Best regards,
baolu