Comment 8 for bug 1907421

Regarding comment #6:
(Didn't noticed that the 5.9 series is also closed now - nevermind ...)
The commit ccd59dce1a21 is indeed the missing puzzle piece.
I was just surprised that just one commit fixes the significant delta,
but since it contains a huge chunk of refactoring work:
"vfio/type1: Refactor vfio_iommu_type1_ioctl()"
this explains it.
I just tried to cherry pick this one (ccd59dce1a21) and the initial one on top (7d6e1329652e) and everything applied cleanly.

With that I can do a manual kernel SRU to groovy, so that we can fix the potential upgrade regression situation.

Thanks for letting me know the missing one ...

And reading comment #7 I think we can (well, actually need) to ask you for the verification tests once the kernel fixes are in, since you have the right hw and a custom qemu.