Re: [PATCH v4 01/12] KVM: x86/mmu: helpers to return if KVM honors guest MTRRs

[Yan Zhao]

On Mon, Oct 09, 2023 at 02:27:16PM -0700, Sean Christopherson wrote:
> On Mon, Oct 09, 2023, Sean Christopherson wrote:
> > On Sat, Oct 07, 2023, Like Xu wrote:
> > > On 14/7/2023 2:50 pm, Yan Zhao wrote:
> > > > diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
> > > > index 92d5a1924fc1..38bd449226f6 100644
> > > > --- a/arch/x86/kvm/mmu.h
> > > > +++ b/arch/x86/kvm/mmu.h
> > > > @@ -235,6 +235,13 @@ static inline u8 permission_fault(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
> > > >   	return -(u32)fault & errcode;
> > > >   }
> > > > +bool __kvm_mmu_honors_guest_mtrrs(struct kvm *kvm, bool vm_has_noncoherent_dma);
> > > > +
> > > > +static inline bool kvm_mmu_honors_guest_mtrrs(struct kvm *kvm)
> > > > +{
> > > > +	return __kvm_mmu_honors_guest_mtrrs(kvm, kvm_arch_has_noncoherent_dma(kvm));
> > > > +}
> > > > +
> > > >   void kvm_zap_gfn_range(struct kvm *kvm, gfn_t gfn_start, gfn_t gfn_end);
> > > >   int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu);
> > > > diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> > > > index 1e5db621241f..b4f89f015c37 100644
> > > > --- a/arch/x86/kvm/mmu/mmu.c
> > > > +++ b/arch/x86/kvm/mmu/mmu.c
> > > > @@ -4516,6 +4516,21 @@ static int kvm_tdp_mmu_page_fault(struct kvm_vcpu *vcpu,
> > > >   }
> > > >   #endif
> > > > +bool __kvm_mmu_honors_guest_mtrrs(struct kvm *kvm, bool vm_has_noncoherent_dma)
> > > 
> > > According to the motivation provided in the comment, the function will no
> > > longer need to be passed the parameter "struct kvm *kvm" but will rely on
> > > the global parameters (plus vm_has_noncoherent_dma), removing "*kvm" ?
> > 
> > Yeah, I'll fixup the commit to drop @kvm from the inner helper.  Thanks!
> 
> Gah, and I gave more bad advice when I suggested this idea.  There's no need to
> explicitly check tdp_enabled, as shadow_memtype_mask is set to zero if TDP is
> disabled.  And that must be the case, e.g. make_spte() would generate a corrupt
> shadow_memtype_mask were non-zero on Intel with shadow paging.
> 
> Yan, can you take a look at what I ended up with (see below) to make sure it
> looks sane/acceptable to you?
yes, tested working on my side.
I think why we added the checking of tdp_enabled was because of the existing check
in patch 3. As noncoherent DMAs checking is not on hot paths, the previous double
checking is also good :)

BTW, as param "kvm" is now removed from the helper, better to remove the word
"second" in comment in patch 4, i.e.

-        * So, specify the second parameter as true here to indicate
-        * non-coherent DMAs are/were involved and TDP zap might be
-        * necessary.
+        * So, specify the parameter as true here to indicate non-coherent
+        * DMAs are/were involved and TDP zap might be necessary.

Sorry and thanks a lot for helps on this series!