| Message ID | 20220622151728.13622-1-pdurrant@amazon.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v4] KVM: x86/xen: Update Xen CPUID Leaf 4 (tsc info) sub-leaves, if present | expand |
On Wed, Jun 22, 2022, Paul Durrant wrote: > The scaling information in subleaf 1 should match the values set by KVM in > the 'vcpu_info' sub-structure 'time_info' (a.k.a. pvclock_vcpu_time_info) > which is shared with the guest, but is not directly available to the VMM. > The offset values are not set since a TSC offset is already applied. > The TSC frequency should also be set in sub-leaf 2. > > Cache pointers to the sub-leaves when CPUID is updated by the VMM and > populate the relevant information prior to entering the guest. All of my comments about the code still apply. https://lore.kernel.org/all/YrMqtHzNSean+qkh@google.com > Signed-off-by: Paul Durrant <pdurrant@amazon.com> > --- > Cc: David Woodhouse <dwmw2@infradead.org> Cc: can go in the changelog, it's helpful info to carry with the commit as it documents who was made aware of the patch, e.g. show who may have had a cance to object/review. > > v2: > - Make sure sub-leaf pointers are NULLed if the time leaf is removed > > v3: > - Add leaf limit check in kvm_xen_set_cpuid() > > v4: > - Update commit comment Please start with the most recent verison and work backardwards, that way reviewers can quickly see the delta for _this_ version. I.e. v4: v3: v2: v1:
> -----Original Message----- > From: Sean Christopherson <seanjc@google.com> > Sent: 27 June 2022 16:19 > To: Durrant, Paul <pdurrant@amazon.co.uk> > Cc: x86@kernel.org; kvm@vger.kernel.org; linux-kernel@vger.kernel.org; David Woodhouse > <dwmw2@infradead.org>; Paolo Bonzini <pbonzini@redhat.com>; Vitaly Kuznetsov <vkuznets@redhat.com>; > Wanpeng Li <wanpengli@tencent.com>; Jim Mattson <jmattson@google.com>; Joerg Roedel <joro@8bytes.org>; > Thomas Gleixner <tglx@linutronix.de>; Ingo Molnar <mingo@redhat.com>; Borislav Petkov <bp@alien8.de>; > Dave Hansen <dave.hansen@linux.intel.com>; H. Peter Anvin <hpa@zytor.com> > Subject: RE: [EXTERNAL][PATCH v4] KVM: x86/xen: Update Xen CPUID Leaf 4 (tsc info) sub-leaves, if > present > > CAUTION: This email originated from outside of the organization. Do not click links or open > attachments unless you can confirm the sender and know the content is safe. > > > > On Wed, Jun 22, 2022, Paul Durrant wrote: > > The scaling information in subleaf 1 should match the values set by KVM in > > the 'vcpu_info' sub-structure 'time_info' (a.k.a. pvclock_vcpu_time_info) > > which is shared with the guest, but is not directly available to the VMM. > > The offset values are not set since a TSC offset is already applied. > > The TSC frequency should also be set in sub-leaf 2. > > > > Cache pointers to the sub-leaves when CPUID is updated by the VMM and > > populate the relevant information prior to entering the guest. > > All of my comments about the code still apply. > > https://lore.kernel.org/all/YrMqtHzNSean+qkh@google.com > Apologies. Not sure how I missed them; I'll send a response shortly. > > Signed-off-by: Paul Durrant <pdurrant@amazon.com> > > --- > > Cc: David Woodhouse <dwmw2@infradead.org> > > Cc: can go in the changelog, it's helpful info to carry with the commit as it > documents who was made aware of the patch, e.g. show who may have had a cance to > object/review. > > > > > v2: > > - Make sure sub-leaf pointers are NULLed if the time leaf is removed > > > > v3: > > - Add leaf limit check in kvm_xen_set_cpuid() > > > > v4: > > - Update commit comment > > Please start with the most recent verison and work backardwards, that way reviewers > can quickly see the delta for _this_ version. I.e. > > > v4: > > v3: > > v2: > > v1: Ok. Paul
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 1038ccb7056a..f77a4940542f 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -638,6 +638,8 @@ struct kvm_vcpu_xen { struct hrtimer timer; int poll_evtchn; struct timer_list poll_timer; + struct kvm_cpuid_entry2 *tsc_info_1; + struct kvm_cpuid_entry2 *tsc_info_2; }; struct kvm_vcpu_arch { diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c index d47222ab8e6e..eb6cd88c974a 100644 --- a/arch/x86/kvm/cpuid.c +++ b/arch/x86/kvm/cpuid.c @@ -25,6 +25,7 @@ #include "mmu.h" #include "trace.h" #include "pmu.h" +#include "xen.h" /* * Unlike "struct cpuinfo_x86.x86_capability", kvm_cpu_caps doesn't need to be @@ -310,6 +311,7 @@ static void kvm_vcpu_after_set_cpuid(struct kvm_vcpu *vcpu) __cr4_reserved_bits(guest_cpuid_has, vcpu); kvm_hv_set_cpuid(vcpu); + kvm_xen_set_cpuid(vcpu); /* Invoke the vendor callback only after the above state is updated. */ static_call(kvm_x86_vcpu_after_set_cpuid)(vcpu); diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 00e23dc518e0..8b45f9975e45 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -3123,6 +3123,7 @@ static int kvm_guest_time_update(struct kvm_vcpu *v) if (vcpu->xen.vcpu_time_info_cache.active) kvm_setup_guest_pvclock(v, &vcpu->xen.vcpu_time_info_cache, 0); kvm_hv_setup_tsc_page(v->kvm, &vcpu->hv_clock); + kvm_xen_setup_tsc_info(v); return 0; } diff --git a/arch/x86/kvm/xen.c b/arch/x86/kvm/xen.c index 610beba35907..08e65ec6c793 100644 --- a/arch/x86/kvm/xen.c +++ b/arch/x86/kvm/xen.c @@ -10,6 +10,9 @@ #include "xen.h" #include "hyperv.h" #include "lapic.h" +#include "cpuid.h" + +#include <asm/xen/cpuid.h> #include <linux/eventfd.h> #include <linux/kvm_host.h> @@ -1855,3 +1858,49 @@ void kvm_xen_destroy_vm(struct kvm *kvm) if (kvm->arch.xen_hvm_config.msr) static_branch_slow_dec_deferred(&kvm_xen_enabled); } + +void kvm_xen_set_cpuid(struct kvm_vcpu *vcpu) +{ + u32 base = 0; + u32 limit; + u32 function; + + vcpu->arch.xen.tsc_info_1 = NULL; + vcpu->arch.xen.tsc_info_2 = NULL; + + for_each_possible_hypervisor_cpuid_base(function) { + struct kvm_cpuid_entry2 *entry = kvm_find_cpuid_entry(vcpu, function, 0); + + if (entry && + entry->ebx == XEN_CPUID_SIGNATURE_EBX && + entry->ecx == XEN_CPUID_SIGNATURE_ECX && + entry->edx == XEN_CPUID_SIGNATURE_EDX) { + base = function; + limit = entry->eax; + break; + } + } + if (!base) + return; + + function = base | XEN_CPUID_LEAF(3); + if (function > limit) + return; + + vcpu->arch.xen.tsc_info_1 = kvm_find_cpuid_entry(vcpu, function, 1); + vcpu->arch.xen.tsc_info_2 = kvm_find_cpuid_entry(vcpu, function, 2); +} + +void kvm_xen_setup_tsc_info(struct kvm_vcpu *vcpu) +{ + struct kvm_cpuid_entry2 *entry = vcpu->arch.xen.tsc_info_1; + + if (entry) { + entry->ecx = vcpu->arch.hv_clock.tsc_to_system_mul; + entry->edx = vcpu->arch.hv_clock.tsc_shift; + } + + entry = vcpu->arch.xen.tsc_info_2; + if (entry) + entry->eax = vcpu->arch.hw_tsc_khz; +} diff --git a/arch/x86/kvm/xen.h b/arch/x86/kvm/xen.h index 532a535a9e99..1afb663318a9 100644 --- a/arch/x86/kvm/xen.h +++ b/arch/x86/kvm/xen.h @@ -32,6 +32,8 @@ int kvm_xen_set_evtchn_fast(struct kvm_xen_evtchn *xe, int kvm_xen_setup_evtchn(struct kvm *kvm, struct kvm_kernel_irq_routing_entry *e, const struct kvm_irq_routing_entry *ue); +void kvm_xen_set_cpuid(struct kvm_vcpu *vcpu); +void kvm_xen_setup_tsc_info(struct kvm_vcpu *vcpu); static inline bool kvm_xen_msr_enabled(struct kvm *kvm) { @@ -135,6 +137,14 @@ static inline bool kvm_xen_timer_enabled(struct kvm_vcpu *vcpu) { return false; } + +static inline void kvm_xen_set_cpuid(struct kvm_vcpu *vcpu) +{ +} + +static inline void kvm_xen_setup_tsc_info(struct kvm_vcpu *vcpu) +{ +} #endif int kvm_xen_hypercall(struct kvm_vcpu *vcpu);
The scaling information in subleaf 1 should match the values set by KVM in the 'vcpu_info' sub-structure 'time_info' (a.k.a. pvclock_vcpu_time_info) which is shared with the guest, but is not directly available to the VMM. The offset values are not set since a TSC offset is already applied. The TSC frequency should also be set in sub-leaf 2. Cache pointers to the sub-leaves when CPUID is updated by the VMM and populate the relevant information prior to entering the guest. Signed-off-by: Paul Durrant <pdurrant@amazon.com> --- Cc: David Woodhouse <dwmw2@infradead.org> v2: - Make sure sub-leaf pointers are NULLed if the time leaf is removed v3: - Add leaf limit check in kvm_xen_set_cpuid() v4: - Update commit comment --- arch/x86/include/asm/kvm_host.h | 2 ++ arch/x86/kvm/cpuid.c | 2 ++ arch/x86/kvm/x86.c | 1 + arch/x86/kvm/xen.c | 49 +++++++++++++++++++++++++++++++++ arch/x86/kvm/xen.h | 10 +++++++ 5 files changed, 64 insertions(+)