On 2025-02-12 2:03 pm, Jason Gunthorpe wrote:
On Wed, Feb 12, 2025 at 01:56:55PM +0000, Robin Murphy wrote:
On 2025-02-07 2:46 pm, Jason Gunthorpe wrote:
virtio-iommu has a mode where the IDENTITY domain is actually a paging
domain with an identity mapping covering some of the system address
space manually created.
To support this add a new domain_alloc_identity() op that accepts
the struct device so that virtio can allocate and fully finalize a
paging domain to return.
Oh, I'd already managed to forget this idea - this could be convenient for
DART to implement per-instance identity domain support as well.
If we are going to broaden this someday then I suggest the core code should
implement it entirely buy asking the driver for a paging domain and
then mapping all of system memory into it. There is nothing special
here that needs to be in driver code.
Sorry, that was just in terms of having a domain_alloc_identity() op
that's allowed to fail, being potentially neater than otherwise having
to register distinct per-instance ops with and without the static
identity_domain. I'm inclined to agree that if anyone really is
desperate for identity-mapped paging domains in the absence of true
identity support, then indeed we probably should handle that in core
code much like we do for improvised blocking domains.
Thanks,
Robin.
