Re: [PATCH v6 20/22] KVM: x86/mmu: Refactor drop_large_spte()

2022-06-22 Thread Paolo Bonzini
On 6/22/22 18:13, Paolo Bonzini wrote: Even better, drop_large_spte() is always called right before kvm_mmu_get_child_sp(), so: Actually, we can even include the call from eager page splitting if __link_shadow_page() is the one that takes care of dropping the large SPTE: From

Re: [PATCH v6 20/22] KVM: x86/mmu: Refactor drop_large_spte()

2022-06-22 Thread Paolo Bonzini
On 6/17/22 19:11, Sean Christopherson wrote: since the shortlog is already a somewhat vague "do a refactor", I vote to opportunistically: - rename drop_large_spte() to drop_spte_if_huge() - rename __drop_large_spte() to drop_huge_spte() - move "if (!is_large_pte(*sptep))" to