Hi Juergen,
I'm working on a v3 of Dave's "x86/msr: Consolidate native/paravirt MSR
functions" series on top of yours, so I tried testing this patch.
While I'm here: you mentioned you'd send a V3 after rc1 if conflicts
turned up. Is that still the plan? I wasn't sure whether to base on v2
plus the respins of 4 and 13, or wait.
Most of the code this touches can't run on my box, so I compared compiled
output instead. Built patches 1-12, then 1-13, with the same .config and
compiler, and diffed "objdump -d" output for every .o in both trees. All
9396 objects came out identical. The 1-13 kernel boots fine under KVM.
Only tried the one config: x86_64, gcc 11.5.0,
CPU_SUP_{INTEL,AMD,ZHAOXIN}=y, HYPERV=y, KVM_INTEL=y, KVM_INTEL_TDX=y.
Tested-by: Shreshth Srivastava <[email protected]>
