On Wed, Jun 28, 2023 at 04:02:05PM -0700, Sean Christopherson wrote: > On Fri, Jun 16, 2023, Yan Zhao wrote: > > This series refines mmu zap caused by EPT memory type update when guest > > MTRRs are honored. > > ... > > > Yan Zhao (11): > > KVM: x86/mmu: helpers to return if KVM honors guest MTRRs > > KVM: x86/mmu: Use KVM honors guest MTRRs helper in > > kvm_tdp_page_fault() > > KVM: x86/mmu: Use KVM honors guest MTRRs helper when CR0.CD toggles > > KVM: x86/mmu: Use KVM honors guest MTRRs helper when update mtrr > > KVM: x86/mmu: zap KVM TDP when noncoherent DMA assignment starts/stops > > KVM: x86/mmu: move TDP zaps from guest MTRRs update to CR0.CD toggling > > KVM: VMX: drop IPAT in memtype when CD=1 for > > KVM_X86_QUIRK_CD_NW_CLEARED > > KVM: x86: move vmx code to get EPT memtype when CR0.CD=1 to x86 common > > code > > KVM: x86/mmu: serialize vCPUs to zap gfn when guest MTRRs are honored > > KVM: x86/mmu: fine-grained gfn zap when guest MTRRs are honored > > KVM: x86/mmu: split a single gfn zap range when guest MTRRs are > > honored > > I got through the easy patches, I'll circle back for the last few patches in a > few weeks (probably 3+ weeks at this point). Thanks for this heads-up. I addressed almost all the comments for v3 currently, except about where to get memtype for CR0.CD=1, and feel free to decline my new proposal in v4 as explained in another mail :) v4 is available here https://lore.kernel.org/all/20230714064656.20147-1-yan.y.zhao@xxxxxxxxx/ Please review the new version directly. Thanks!
