On 11/06/21 18:45, Sean Christopherson wrote:
Along which path though? I would have naively expected those to be driven
only by the context->root_level.
The functional code is driven by context->root_level, but if KVM doesn't include
the level in the mmu_role then it will fail to update context->root_level when
L2 changes from 32-bit PAE to 64-bit. If all the CR0/CR4/EFER bits remain the
same, only the level will differ. Without this patch, role.level is always '0'
for the nested MMU.
Yes the problem is the
if (new_role.as_u64 == g_context->mmu_role.as_u64)
return;
and the patch lets you preserve the optimization instead of dropping it.
I was wondering if I was missing something else because of the "when
walking the guest page tables" remark.
Paolo
