Re: [PATCH v2 3/4] iommu/hyperv: Add para-virtualized IOMMU support for Hyper-V guest

[Jacob Pan <jacob.pan@linux.microsoft.com>]

Hi Michael,

On Sat, 11 Jul 2026 18:31:15 +0000
Michael Kelley <mhklinux@outlook.com> wrote:

> One new thought:  Have you considered the hibernate/resume
> cycle? Does anything need to be done with the pvIOMMU to
> make it functional again after resume? I see that the Intel and
> AMD IOMMU drivers have suspend and resume functions. I
> don't know enough about the Hyper-V pvIOMMU to know if it
> might also need suspend and resume functions.

I don't think the Hyper-V pvIOMMU guest driver needs the same kind of
suspend/resume handling as a hardware IOMMU driver. Unlike VT-d or AMD
IOMMU, the guest driver does not own physical IOMMU registers, root
tables, command queues, or translation enable state that must be saved
and reprogrammed on resume.

For nested translation, the guest does own the stage-1 I/O page tables,
but those are normal guest memory. They survive S3 as system RAM. The
guest driver still needs to issue the normal pvIOMMU invalidations when
it changes S1 mappings, but suspend/resume by itself does not modify
the S1 page tables and should not require a special flush.

The important contract is on the Hyper-V side: Hyper-V owns translation
enable/disable and must prevent device DMA while translation state is
not valid.

Thanks,

Jacob