On Wed, Sep 30, 2020 at 02:49:41PM -0700, Raj, Ashok wrote: > One of the parameters it has is the "supported system page-sizes" which is > usually there in the SRIOV properties. So it needed a place holder for > that. No idea why this would be a PCI cap. It is certainly not something so universal it needs standardizing. There are many ways a device can manage a BAR to match a required protection granularity. > When we provision an entire PCI device that is IMS capable. The guest > driver does know it can update the IMS entries directly without going to > the host. But in order to do remapping we need something like how we manage > PASID allocation from guest, so an IRTE entry can be allocated and the host > driver can write the proper values for IMS. Look at the architecture we ended up with. You need to make pci_subdevice_msi_create_irq_domain() fail if the platform can't provide the functionality. Working around that with PCI caps is pretty gross. Jason