| Message ID | 20220128005208.4008533-10-seanjc@google.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | KVM: x86: Fill *_x86_ops via kvm-x86-ops.h | expand |
Sean Christopherson <seanjc@google.com> writes: > Uninline and export Hyper-V's hv_track_root_tdp(), which is (somewhat > indirectly) the last remaining reference to kvm_x86_ops from vendor > modules, i.e. will allow unexporting kvm_x86_ops. Reloading the TDP PGD > isn't the fastest of paths, hv_track_root_tdp() isn't exactly tiny, and > disallowing vendor code from accessing kvm_x86_ops provides nice-to-have > encapsulation of common x86 code (and of Hyper-V code for that > matter). We can add a static branch for "kvm_x86_ops.tlb_remote_flush == hv_remote_flush_tlb" condition and check it in vendor modules prior to calling into hv_track_root_tdp() but I seriously doubt it'll bring us noticable performance gain. > > No functional change intended. > > Signed-off-by: Sean Christopherson <seanjc@google.com> > --- > arch/x86/kvm/kvm_onhyperv.c | 14 ++++++++++++++ > arch/x86/kvm/kvm_onhyperv.h | 14 +------------- > 2 files changed, 15 insertions(+), 13 deletions(-) > > diff --git a/arch/x86/kvm/kvm_onhyperv.c b/arch/x86/kvm/kvm_onhyperv.c > index b469f45e3fe4..ee4f696a0782 100644 > --- a/arch/x86/kvm/kvm_onhyperv.c > +++ b/arch/x86/kvm/kvm_onhyperv.c > @@ -92,3 +92,17 @@ int hv_remote_flush_tlb(struct kvm *kvm) > return hv_remote_flush_tlb_with_range(kvm, NULL); > } > EXPORT_SYMBOL_GPL(hv_remote_flush_tlb); > + > +void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp) > +{ > + struct kvm_arch *kvm_arch = &vcpu->kvm->arch; > + > + if (kvm_x86_ops.tlb_remote_flush == hv_remote_flush_tlb) { > + spin_lock(&kvm_arch->hv_root_tdp_lock); > + vcpu->arch.hv_root_tdp = root_tdp; > + if (root_tdp != kvm_arch->hv_root_tdp) > + kvm_arch->hv_root_tdp = INVALID_PAGE; > + spin_unlock(&kvm_arch->hv_root_tdp_lock); > + } > +} > +EXPORT_SYMBOL_GPL(hv_track_root_tdp); > diff --git a/arch/x86/kvm/kvm_onhyperv.h b/arch/x86/kvm/kvm_onhyperv.h > index 1c67abf2eba9..287e98ef9df3 100644 > --- a/arch/x86/kvm/kvm_onhyperv.h > +++ b/arch/x86/kvm/kvm_onhyperv.h > @@ -10,19 +10,7 @@ > int hv_remote_flush_tlb_with_range(struct kvm *kvm, > struct kvm_tlb_range *range); > int hv_remote_flush_tlb(struct kvm *kvm); > - > -static inline void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp) > -{ > - struct kvm_arch *kvm_arch = &vcpu->kvm->arch; > - > - if (kvm_x86_ops.tlb_remote_flush == hv_remote_flush_tlb) { > - spin_lock(&kvm_arch->hv_root_tdp_lock); > - vcpu->arch.hv_root_tdp = root_tdp; > - if (root_tdp != kvm_arch->hv_root_tdp) > - kvm_arch->hv_root_tdp = INVALID_PAGE; > - spin_unlock(&kvm_arch->hv_root_tdp_lock); > - } > -} > +void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp); > #else /* !CONFIG_HYPERV */ > static inline void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp) > { Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
diff --git a/arch/x86/kvm/kvm_onhyperv.c b/arch/x86/kvm/kvm_onhyperv.c index b469f45e3fe4..ee4f696a0782 100644 --- a/arch/x86/kvm/kvm_onhyperv.c +++ b/arch/x86/kvm/kvm_onhyperv.c @@ -92,3 +92,17 @@ int hv_remote_flush_tlb(struct kvm *kvm) return hv_remote_flush_tlb_with_range(kvm, NULL); } EXPORT_SYMBOL_GPL(hv_remote_flush_tlb); + +void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp) +{ + struct kvm_arch *kvm_arch = &vcpu->kvm->arch; + + if (kvm_x86_ops.tlb_remote_flush == hv_remote_flush_tlb) { + spin_lock(&kvm_arch->hv_root_tdp_lock); + vcpu->arch.hv_root_tdp = root_tdp; + if (root_tdp != kvm_arch->hv_root_tdp) + kvm_arch->hv_root_tdp = INVALID_PAGE; + spin_unlock(&kvm_arch->hv_root_tdp_lock); + } +} +EXPORT_SYMBOL_GPL(hv_track_root_tdp); diff --git a/arch/x86/kvm/kvm_onhyperv.h b/arch/x86/kvm/kvm_onhyperv.h index 1c67abf2eba9..287e98ef9df3 100644 --- a/arch/x86/kvm/kvm_onhyperv.h +++ b/arch/x86/kvm/kvm_onhyperv.h @@ -10,19 +10,7 @@ int hv_remote_flush_tlb_with_range(struct kvm *kvm, struct kvm_tlb_range *range); int hv_remote_flush_tlb(struct kvm *kvm); - -static inline void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp) -{ - struct kvm_arch *kvm_arch = &vcpu->kvm->arch; - - if (kvm_x86_ops.tlb_remote_flush == hv_remote_flush_tlb) { - spin_lock(&kvm_arch->hv_root_tdp_lock); - vcpu->arch.hv_root_tdp = root_tdp; - if (root_tdp != kvm_arch->hv_root_tdp) - kvm_arch->hv_root_tdp = INVALID_PAGE; - spin_unlock(&kvm_arch->hv_root_tdp_lock); - } -} +void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp); #else /* !CONFIG_HYPERV */ static inline void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp) {
Uninline and export Hyper-V's hv_track_root_tdp(), which is (somewhat indirectly) the last remaining reference to kvm_x86_ops from vendor modules, i.e. will allow unexporting kvm_x86_ops. Reloading the TDP PGD isn't the fastest of paths, hv_track_root_tdp() isn't exactly tiny, and disallowing vendor code from accessing kvm_x86_ops provides nice-to-have encapsulation of common x86 code (and of Hyper-V code for that matter). No functional change intended. Signed-off-by: Sean Christopherson <seanjc@google.com> --- arch/x86/kvm/kvm_onhyperv.c | 14 ++++++++++++++ arch/x86/kvm/kvm_onhyperv.h | 14 +------------- 2 files changed, 15 insertions(+), 13 deletions(-)