On Fri, Apr 08, 2022 at 06:33:23AM -0500, Suravee Suthikulpanit wrote:
commit 4a204f7895878363ca8211f50ec610408c8c70aa upstream.
Expand KVM's mask for the AVIC host physical ID to the full 12 bits defined by the architecture. The number of bits consumed by hardware is model specific, e.g. early CPUs ignored bits 11:8, but there is no way for KVM to enumerate the "true" size. So, KVM must allow using all bits, else it risks rejecting completely legal x2APIC IDs on newer CPUs.
This means KVM relies on hardware to not assign x2APIC IDs that exceed the "true" width of the field, but presumably hardware is smart enough to tie the width to the max x2APIC ID. KVM also relies on hardware to support at least 8 bits, as the legacy xAPIC ID is writable by software. But, those assumptions are unavoidable due to the lack of any way to enumerate the "true" width.
Note: This patch has been modified from the upstream commit due to the conflict caused by the commit 391503528257 ("KVM: x86: SVM: move avic definitions from AMD's spec to svm.h")
Please apply this patch to the following kernel tree:
- 4.14-stable
- 4.9-stable
- 4.19-stable
- 5.4-stable
- 5.10-stable
- 5.15-stable
- 5.16-stable
It only applied to 5.15 and 5.16, we need backports that work (and obviously have been tested) to apply this to older kernel trees. Please fix up and send if you want this in older trees.
thanks,
greg k-h