Re: [PATCH 17/22] KVM: x86/mmu: pull struct kvm_pagewalk out of struct kvm_mmu

[Paolo Bonzini]

On Tue, May 19, 2026 at 1:12 AM Kishen Maloor <kishen.maloor@xxxxxxxxx> wrote:
> +1, and I wonder if it helps to name the walkers after what they walk, to shape one's conceptual model. E.g. cr3_walk (instead of cpu_walk) for the walker over CR3-rooted tables of the currently executing guest (L1's or L2's CR3), and nested_tdp_walk for the walker over L1's nested EPT/NPT when L0 services L2. tdp_walk alone felt a bit vague — it read as "the TDP MMU's walker," whereas this specifically applies to the nested TDP path.

You are influenced by the old names though. If you stop thinking of
kvm_mmu and kvm_pagewalk as related in any way, it helps. kvm_pagewalk
exists independent of the MMU, except inasmuch a (shadow) kvm_mmu has
a kvm_pagewalk that it uses to walk the page tables it shadows.

Paolo