Added some details on checking IOMMU for PCI passthrough

[alfh]

Add info on how to check if IOMMU is enabled on the current running kernel, and how to add the kernel parameters. Also add some info related to iommu_groups, and hint on iommu=pt for performance.

I used this to enable PCI passthrough of USB controller, which worked, even if I did not have anything in the /sys/kernel/iommu_groups.

Before submitting the pull request, you must agree with the following statements by checking both boxes with a 'x'.

• [ x] "I accept that my contribution is placed under the CC BY-SA 2.0 license [1]."
• [ x] "My contribution complies with the Developer Certificate of Origin [2]."

[1] https://creativecommons.org/licenses/by-sa/2.0/
[2] https://xcp-ng.org/docs/contributing.html#developer-certificate-of-origin-dco

[andSmv]

The dom0 kernel doesn't have to access the platform IOMMU(s). The IOMMU need to be enabled on the platform in the firmware (BIOS/UEFI), but it is managed by XEN and in totally transparent manner to the VMs. So this is true for PCI passthrough needs. However, there are some patches in XCP-ng, which under some circumstances enable such an access for the VMs. But this can be veiwed as a very basic paravartualized IOMMU (and which is not pluged to Linux kernel IOMMU framework) . This pv IOMMU is not discovered through Intel DMAR (or AMD IVRS) ACPI tables. To be honest, this facility is almost never used. So, in any circumstances, you don't have to manipulate dom0 kernel command line IOMMU-related parameters.

[stormi]

Marking as "changes requested" so that we don't merge by mistake.