On Sat, Jun 18, 2022 at 11:18:17PM -0700, Christoph Hellwig wrote:

...

...

There is a bunch of code an comments in the iommu type1 code that suggest we can pin memory that is not page backed.

Would you mind explaining the use case for pinning memory that isn't page backed? And do we have such use case so far?

Sorry, I should have deleted that sentence. I wrote it before spending some more time to dig through the code and all the locked memory has page backing. There just seem to be a lot of checks left inbetween if a pfn is page backed, mostly due to the pfn based calling convetions.

OK. We'd be safe to move on then. Thanks for the clarification.

...

I can do that. I tried once, but there were just too much changes inside type1 code that felt like a chain reaction. If we plan to eventually replace with IOMMUFD implementations, these changes in type1 might not be necessary, I thought.

To make sure we keep full compatibility I suspect the final iommufd implementation has to be gradutally created from the existing code anyway.

Hmm. I think Jason can give some insight. Meanwhile, I will try to add a patch to type1 code, in case we'd end up with what you suspected.