On 21.10.2021 11:57, Jan Beulich wrote: > In the course of reading the response to v1 (patch 1 only) I realized > that not only that patch needs further adjustment, but that also > further changes are needed (and there's likely yet more amiss). > > 1: x86/IOMMU: mark IOMMU / intremap not in use when ACPI tables are missing > 2: x86/APIC: avoid iommu_supports_x2apic() on error path > 3: AMD/IOMMU: iommu_enable vs iommu_intremap
Ian, while we further discuss / refine patch 3, the first two have the needed R-b, but will now need you release-ack aiui. Andrew, did you perhaps have a chance to actually try v2 of patch 1? It works for me when suitably configuring the BIOS on my Skylake, so I wouldn't feel uncertain in committing without a Tested-by, but it would feel even better if I had one. Thanks, Jan
