KVM: x86/mmu: nVMX: 5-level paging fixes and enabling
v2 cover KVM: x86/mmu: nVMX: 5-level paging fixes and enabling
Two fixes for 5-level paging bugs with a 100% fatality rate, a patch to enable 5-level EPT in L1, and additional clean up on top (mostly renames of functions/variables that caused me no end of confusion when trying to figure out what was broken).
Tested fixed kernels at L0, L1 and L2, with most combinations of EPT, shadow paging, 4-level and 5-level. EPT kvm-unit-tests runs clean in L0. Patches for kvm-unit-tests incoming to play nice with 5-level nested EPT.
Ideally patches 1 and 2 would get into 5.6, 5-level paging is quite broken without them.
v2:
- Increase the nested EPT array sizes to accomodate 5-level paging in the patch that adds support for 5-level nested EPT, not in the bug fix for 5-level shadow paging.
Sean Christopherson (7):
KVM: nVMX: Use correct root level for nested EPT shadow page tables
KVM: x86/mmu: Fix struct guest_walker arrays for 5-level paging
KVM: nVMX: Allow L1 to use 5-level page walks for nested EPT
KVM: nVMX: Rename nested_ept_get_cr3() to nested_ept_get_eptp()
KVM: nVMX: Rename EPTP validity helper and associated variables
KVM: x86/mmu: Rename kvm_mmu->get_cr3() to ->get_guest_cr3_or_eptp()
KVM: nVMX: Drop unnecessary check on ept caps for execute-only
arch/x86/include/asm/kvm_host.h | 2 +- arch/x86/include/asm/vmx.h | 12 +++++++ arch/x86/kvm/mmu/mmu.c | 35 ++++++++++---------- arch/x86/kvm/mmu/paging_tmpl.h | 6 ++-- arch/x86/kvm/svm.c | 10 +++--- arch/x86/kvm/vmx/nested.c | 58 ++++++++++++++++++++------------- arch/x86/kvm/vmx/nested.h | 4 +-- arch/x86/kvm/vmx/vmx.c | 2 ++ arch/x86/kvm/x86.c | 2 +- 9 files changed, 79 insertions(+), 52 deletions(-)
Signature · 1 line
2.24.1
PATCH v2 1/7 KVM: nVMX: Use correct root level for nested EPT shadow page tables
Hardcode the EPT page-walk level for L2 to be 4 levels, as KVM's MMU currently also hardcodes the page walk level for nested EPT to be 4 levels. The L2 guest is all but guaranteed to soft hang on its first instruction when L1 is using EPT, as KVM will construct 4-level page tables and then tell hardware to use 5-level page tables.
--- arch/x86/kvm/vmx/vmx.c | 3 +++ 1 file changed, 3 insertions(+)
arch/x86/kvm/vmx/vmx.c +3 -0
@@ -2947,6 +2947,9 @@ void vmx_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)static int get_ept_level (struct kvm_vcpu *vcpu) { +/* Nested EPT currently only supports 4-level walks. */ +if (is_guest_mode (vcpu) &&nested_cpu_has_ept (get_vmcs12 (vcpu))) +return 4 ;if (cpu_has_vmx_ept_5levels () && (cpuid_maxphyaddr (vcpu) >48 ))return 5 ;return 4 ;
Signature · 1 line
2.24.1
PATCH v2 2/7 KVM: x86/mmu: Fix struct guest_walker arrays for 5-level paging
In reply to Sean ChristophersonDefine PT_MAX_FULL_LEVELS as PT64_ROOT_MAX_LEVEL, i.e. 5, to fix shadow paging for 5-level guest page tables. PT_MAX_FULL_LEVELS is used to size the arrays that track guest pages table information, i.e. using a "max levels" of 4 causes KVM to access garbage beyond the end of an array when querying state for level 5 entries. E.g. FNAME(gpte_changed) will read garbage and most likely return %true for a level 5 entry, soft-hanging the guest because FNAME(fetch) will restart the guest instead of creating SPTEs because it thinks the guest PTE has changed.
Note, KVM doesn't yet support 5-level nested EPT, so PT_MAX_FULL_LEVELS gets to stay "4" for the PTTYPE_EPT case.
--- arch/x86/kvm/mmu/paging_tmpl.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kvm/mmu/paging_tmpl.h +1 -1
@@ -33,7 +33,7 @@#define PT_GUEST_ACCESSED_SHIFT PT_ACCESSED_SHIFT#define PT_HAVE_ACCESSED_DIRTY (mmu) true#ifdef CONFIG_X86_64 -#define PT_MAX_FULL_LEVELS 4 +#define PT_MAX_FULL_LEVELS PT64_ROOT_MAX_LEVEL#define CMPXCHG cmpxchg#else #define CMPXCHG cmpxchg64
Signature · 1 line
2.24.1
PATCH v2 3/7 KVM: nVMX: Allow L1 to use 5-level page walks for nested EPT
In reply to Sean ChristophersonAdd support for 5-level nested EPT, and advertise said support in the EPT capabilities MSR. KVM's MMU can already handle 5-level legacy page tables, there's no reason to force an L1 VMM to use shadow paging if it wants to employ 5-level page tables.
--- arch/x86/include/asm/vmx.h | 12 ++++++++++++ arch/x86/kvm/mmu/mmu.c | 11 ++++++----- arch/x86/kvm/mmu/paging_tmpl.h | 2 +- arch/x86/kvm/vmx/nested.c | 21 +++++++++++++++++---- arch/x86/kvm/vmx/vmx.c | 3 +-- 5 files changed, 37 insertions(+), 12 deletions(-)
arch/x86/include/asm/vmx.h +12 -0
@@ -500,6 +500,18 @@ enum vmcs_field { VMX_EPT_EXECUTABLE_MASK)#define VMX_EPT_MT_MASK (7ull << VMX_EPT_MT_EPTE_SHIFT) +static inline u8 vmx_eptp_page_walk_level (u64 eptp) +{ +u64 encoded_level = eptp &VMX_EPTP_PWL_MASK ; + +if (encoded_level ==VMX_EPTP_PWL_5 ) +return 5 ; + +/* @eptp must be pre-validated by the caller. */ +WARN_ON_ONCE (encoded_level !=VMX_EPTP_PWL_4 ); +return 4 ; +} +/* The mask to use to trigger an EPT Misconfiguration in order to track MMIO */ #define VMX_EPT_MISCONFIG_WX_VALUE (VMX_EPT_WRITABLE_MASK | \ VMX_EPT_EXECUTABLE_MASK)
arch/x86/kvm/mmu/mmu.c +6 -5
@@ -5012,14 +5012,14 @@ EXPORT_SYMBOL_GPL(kvm_init_shadow_mmu);static union kvm_mmu_role kvm_calc_shadow_ept_root_page_role (struct kvm_vcpu *vcpu,bool accessed_dirty, -bool execonly) +bool execonly,u8 level) {union kvm_mmu_role role = {0 };/* SMM flag is inherited from root_mmu */ role.base.smm = vcpu->arch.root_mmu.mmu_role.base.smm; - role.base.level =PT64_ROOT_4LEVEL ; + role.base.level = level; role.base.gpte_is_8_bytes = true; role.base.direct = false; role.base.ad_disabled = !accessed_dirty;
@@ -5043,9 +5043,10 @@ void kvm_init_shadow_ept_mmu(struct kvm_vcpu *vcpu, bool execonly, bool accessed_dirty, gpa_t new_eptp) {struct kvm_mmu *context = vcpu->arch.mmu; +u8 level =vmx_eptp_page_walk_level (new_eptp);union kvm_mmu_role new_role =kvm_calc_shadow_ept_root_page_role (vcpu, accessed_dirty, - execonly); + execonly, level);__kvm_mmu_new_cr3 (vcpu, new_eptp, new_role.base, false);
@@ -5053,7 +5054,7 @@ void kvm_init_shadow_ept_mmu(struct kvm_vcpu *vcpu, bool execonly,if (new_role.as_u64 == context->mmu_role.as_u64)return ; - context->shadow_root_level =PT64_ROOT_4LEVEL ; + context->shadow_root_level = level; context->nx = true; context->ept_ad = accessed_dirty;
@@ -5062,7 +5063,7 @@ void kvm_init_shadow_ept_mmu(struct kvm_vcpu *vcpu, bool execonly, context->sync_page = ept_sync_page; context->invlpg = ept_invlpg; context->update_pte = ept_update_pte; - context->root_level =PT64_ROOT_4LEVEL ; + context->root_level = level; context->direct_map = false; context->mmu_role.as_u64 = new_role.as_u64;
arch/x86/kvm/mmu/paging_tmpl.h +1 -1
@@ -66,7 +66,7 @@#define PT_GUEST_ACCESSED_SHIFT 8#define PT_HAVE_ACCESSED_DIRTY (mmu) ((mmu)->ept_ad)#define CMPXCHG cmpxchg64 -#define PT_MAX_FULL_LEVELS 4 +#define PT_MAX_FULL_LEVELS PT64_ROOT_MAX_LEVEL#else #error Invalid PTTYPE value#endif
arch/x86/kvm/vmx/nested.c +17 -4
@@ -2581,9 +2581,19 @@ static bool valid_ept_address(struct kvm_vcpu *vcpu, u64 address)return false; } -/* only 4 levels page-walk length are valid */ -if (CC ((address &VMX_EPTP_PWL_MASK ) !=VMX_EPTP_PWL_4 )) +/* Page-walk levels validity. */ +switch (address &VMX_EPTP_PWL_MASK ) { +case VMX_EPTP_PWL_5 : +if (CC (!(vmx->nested.msrs.ept_caps &VMX_EPT_PAGE_WALK_5_BIT ))) +return false; +break ; +case VMX_EPTP_PWL_4 : +if (CC (!(vmx->nested.msrs.ept_caps &VMX_EPT_PAGE_WALK_4_BIT ))) +return false; +break ; +default :return false; + }/* Reserved bits should not be set */ if (CC (address >> maxphyaddr || ((address >>7 ) &0x1f )))
@@ -6057,8 +6067,11 @@ void nested_vmx_setup_ctls_msrs(struct nested_vmx_msrs *msrs, u32 ept_caps,/* nested EPT: emulate EPT also to L1 */ msrs->secondary_ctls_high |=SECONDARY_EXEC_ENABLE_EPT ; - msrs->ept_caps =VMX_EPT_PAGE_WALK_4_BIT | -VMX_EPTP_WB_BIT |VMX_EPT_INVEPT_BIT ; + msrs->ept_caps = +VMX_EPT_PAGE_WALK_4_BIT | +VMX_EPT_PAGE_WALK_5_BIT | +VMX_EPTP_WB_BIT | +VMX_EPT_INVEPT_BIT ;if (cpu_has_vmx_ept_execute_only ()) msrs->ept_caps |=VMX_EPT_EXECUTE_ONLY_BIT ;
arch/x86/kvm/vmx/vmx.c +1 -2
@@ -2947,9 +2947,8 @@ void vmx_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)static int get_ept_level (struct kvm_vcpu *vcpu) { -/* Nested EPT currently only supports 4-level walks. */ if (is_guest_mode (vcpu) &&nested_cpu_has_ept (get_vmcs12 (vcpu))) -return 4 ; +return vmx_eptp_page_walk_level (nested_ept_get_cr3 (vcpu));if (cpu_has_vmx_ept_5levels () && (cpuid_maxphyaddr (vcpu) >48 ))return 5 ;return 4 ;
Signature · 1 line
2.24.1
PATCH v2 4/7 KVM: nVMX: Rename nested_ept_get_cr3() to nested_ept_get_eptp()
In reply to Sean ChristophersonRename the accessor for vmcs12.EPTP to use "eptp" instead of "cr3". The accessor has no relation to cr3 whatsoever, other than it being assigned to the also poorly named kvm_mmu->get_cr3() hook.
No functional change intended.
--- arch/x86/kvm/vmx/nested.c | 4 ++-- arch/x86/kvm/vmx/nested.h | 4 ++-- arch/x86/kvm/vmx/vmx.c | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-)
arch/x86/kvm/vmx/nested.c +2 -2
@@ -353,9 +353,9 @@ static void nested_ept_init_mmu_context(struct kvm_vcpu *vcpu)to_vmx (vcpu)->nested.msrs.ept_caps &VMX_EPT_EXECUTE_ONLY_BIT ,nested_ept_ad_enabled (vcpu), - nested_ept_get_cr3(vcpu)); + nested_ept_get_eptp(vcpu)); vcpu->arch.mmu->set_cr3 = vmx_set_cr3; - vcpu->arch.mmu->get_cr3 = nested_ept_get_cr3; + vcpu->arch.mmu->get_cr3 = nested_ept_get_eptp; vcpu->arch.mmu->inject_page_fault = nested_ept_inject_page_fault; vcpu->arch.mmu->get_pdptr = kvm_pdptr_read;
arch/x86/kvm/vmx/nested.h +2 -2
@@ -59,7 +59,7 @@ static inline int vmx_has_valid_vmcs12(struct kvm_vcpu *vcpu) vmx->nested.hv_evmcs; } -static inline unsigned long nested_ept_get_cr3 (struct kvm_vcpu *vcpu) +static inline unsigned long nested_ept_get_eptp (struct kvm_vcpu *vcpu) {/* return the page table to be shadowed - in our case, EPT12 */ return get_vmcs12 (vcpu)->ept_pointer;
@@ -67,7 +67,7 @@ static inline unsigned long nested_ept_get_cr3(struct kvm_vcpu *vcpu)static inline bool nested_ept_ad_enabled (struct kvm_vcpu *vcpu) { -return nested_ept_get_cr3 (vcpu) &VMX_EPTP_AD_ENABLE_BIT ; +return nested_ept_get_eptp (vcpu) &VMX_EPTP_AD_ENABLE_BIT ; }/*
arch/x86/kvm/vmx/vmx.c +1 -1
@@ -2948,7 +2948,7 @@ void vmx_set_cr0(struct kvm_vcpu *vcpu, unsigned long cr0)static int get_ept_level (struct kvm_vcpu *vcpu) {if (is_guest_mode (vcpu) &&nested_cpu_has_ept (get_vmcs12 (vcpu))) -return vmx_eptp_page_walk_level (nested_ept_get_cr3 (vcpu)); +return vmx_eptp_page_walk_level (nested_ept_get_eptp (vcpu));if (cpu_has_vmx_ept_5levels () && (cpuid_maxphyaddr (vcpu) >48 ))return 5 ;return 4 ;
Signature · 1 line
2.24.1
PATCH v2 5/7 KVM: nVMX: Rename EPTP validity helper and associated variables
In reply to Sean ChristophersonRename valid_ept_address() to nested_vmx_check_eptp() to follow the nVMX nomenclature and to reflect that the function now checks a lot more than just the address contained in the EPTP. Rename address to new_eptp in associated code.
No functional change intended.
--- arch/x86/kvm/vmx/nested.c | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-)
arch/x86/kvm/vmx/nested.c +12 -12
@@ -2562,13 +2562,13 @@ static int nested_vmx_check_nmi_controls(struct vmcs12 *vmcs12)return 0 ; } -static bool valid_ept_address (struct kvm_vcpu *vcpu,u64 address) +static bool nested_vmx_check_eptp (struct kvm_vcpu *vcpu,u64 new_eptp) {struct vcpu_vmx *vmx =to_vmx (vcpu);int maxphyaddr =cpuid_maxphyaddr (vcpu);/* Check for memory type validity */ -switch (address &VMX_EPTP_MT_MASK ) { +switch (new_eptp &VMX_EPTP_MT_MASK ) {case VMX_EPTP_MT_UC :if (CC (!(vmx->nested.msrs.ept_caps &VMX_EPTP_UC_BIT )))return false;
@@ -2582,7 +2582,7 @@ static bool valid_ept_address(struct kvm_vcpu *vcpu, u64 address) }/* Page-walk levels validity. */ -switch (address &VMX_EPTP_PWL_MASK ) { +switch (new_eptp &VMX_EPTP_PWL_MASK ) {case VMX_EPTP_PWL_5 :if (CC (!(vmx->nested.msrs.ept_caps &VMX_EPT_PAGE_WALK_5_BIT )))return false;
@@ -2596,11 +2596,11 @@ static bool valid_ept_address(struct kvm_vcpu *vcpu, u64 address) }/* Reserved bits should not be set */ -if (CC (address >> maxphyaddr || ((address >>7 ) &0x1f ))) +if (CC (new_eptp >> maxphyaddr || ((new_eptp >>7 ) &0x1f )))return false;/* AD, if set, should be supported */ -if (address &VMX_EPTP_AD_ENABLE_BIT ) { +if (new_eptp &VMX_EPTP_AD_ENABLE_BIT ) {if (CC (!(vmx->nested.msrs.ept_caps &VMX_EPT_AD_BIT )))return false; }
@@ -2649,7 +2649,7 @@ static int nested_check_vm_execution_controls(struct kvm_vcpu *vcpu,return -EINVAL ;if (nested_cpu_has_ept (vmcs12) && -CC (!valid_ept_address (vcpu, vmcs12->ept_pointer))) +CC (!nested_vmx_check_eptp (vcpu, vmcs12->ept_pointer)))return -EINVAL ;if (nested_cpu_has_vmfunc (vmcs12)) {
@@ -5188,7 +5188,7 @@ static int nested_vmx_eptp_switching(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12) {u32 index =kvm_rcx_read (vcpu); -u64 address; +u64 new_eptp;bool accessed_dirty;struct kvm_mmu *mmu = vcpu->arch.walk_mmu;
@@ -5201,23 +5201,23 @@ static int nested_vmx_eptp_switching(struct kvm_vcpu *vcpu,if (kvm_vcpu_read_guest_page (vcpu, vmcs12->eptp_list_address >>PAGE_SHIFT , - &address, index *8 ,8 )) + &new_eptp, index *8 ,8 ))return 1 ; - accessed_dirty = !!(address &VMX_EPTP_AD_ENABLE_BIT ); + accessed_dirty = !!(new_eptp &VMX_EPTP_AD_ENABLE_BIT );/* * If the (L2) guest does a vmfunc to the currently * active ept pointer, we don't have to do anything else */ -if (vmcs12->ept_pointer != address) { -if (!valid_ept_address (vcpu, address)) +if (vmcs12->ept_pointer != new_eptp) { +if (!nested_vmx_check_eptp (vcpu, new_eptp))return 1 ;kvm_mmu_unload (vcpu); mmu->ept_ad = accessed_dirty; mmu->mmu_role.base.ad_disabled = !accessed_dirty; - vmcs12->ept_pointer = address; + vmcs12->ept_pointer = new_eptp;/* * TODO: Check what's the correct approach in case * mmu reload fails. Currently, we just let the next
Signature · 1 line
2.24.1
PATCH v2 6/7 KVM: x86/mmu: Rename kvm_mmu->get_cr3() to ->get_guest_cr3_or_eptp()
In reply to Sean ChristophersonRename kvm_mmu->get_cr3() to call out that it is retrieving a guest value, as opposed to kvm_mmu->set_cr3(), which sets a host value, and to note that it will return L1's EPTP when nested EPT is in use. Hopefully the new name will also make it more obvious that L1's nested_cr3 is returned in SVM's nested NPT case.
No functional change intended.
--- arch/x86/include/asm/kvm_host.h | 2 +- arch/x86/kvm/mmu/mmu.c | 24 ++++++++++++------------ arch/x86/kvm/mmu/paging_tmpl.h | 2 +- arch/x86/kvm/svm.c | 10 +++++----- arch/x86/kvm/vmx/nested.c | 8 ++++---- arch/x86/kvm/x86.c | 2 +- 6 files changed, 24 insertions(+), 24 deletions(-)
arch/x86/include/asm/kvm_host.h +1 -1
@@ -383,7 +383,7 @@ struct kvm_mmu_root_info {*/ struct kvm_mmu {void (*set_cr3)(struct kvm_vcpu *vcpu,unsigned long root); -unsigned long (*get_cr3)(struct kvm_vcpu *vcpu); +unsigned long (*get_guest_cr3_or_eptp)(struct kvm_vcpu *vcpu);u64 (*get_pdptr)(struct kvm_vcpu *vcpu,int index);int (*page_fault)(struct kvm_vcpu *vcpu,gpa_t cr2_or_gpa,u32 err,bool prefault);
arch/x86/kvm/mmu/mmu.c +12 -12
@@ -3731,7 +3731,7 @@ static int mmu_alloc_direct_roots(struct kvm_vcpu *vcpu) vcpu->arch.mmu->root_hpa =__pa (vcpu->arch.mmu->pae_root); }else BUG (); - vcpu->arch.mmu->root_cr3 = vcpu->arch.mmu->get_cr3 (vcpu); + vcpu->arch.mmu->root_cr3 = vcpu->arch.mmu->get_guest_cr3_or_eptp (vcpu);return 0 ; }
@@ -3743,7 +3743,7 @@ static int mmu_alloc_shadow_roots(struct kvm_vcpu *vcpu)gfn_t root_gfn, root_cr3;int i; - root_cr3 = vcpu->arch.mmu->get_cr3 (vcpu); + root_cr3 = vcpu->arch.mmu->get_guest_cr3_or_eptp (vcpu); root_gfn = root_cr3 >>PAGE_SHIFT ;if (mmu_check_root (vcpu, root_gfn))
@@ -4080,7 +4080,7 @@ static int kvm_arch_setup_async_pf(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa, arch.token = (vcpu->arch.apf.id++ <<12 ) | vcpu->vcpu_id; arch.gfn = gfn; arch.direct_map = vcpu->arch.mmu->direct_map; - arch.cr3 = vcpu->arch.mmu->get_cr3 (vcpu); + arch.cr3 = vcpu->arch.mmu->get_guest_cr3_or_eptp (vcpu);return kvm_setup_async_pf (vcpu, cr2_or_gpa,kvm_vcpu_gfn_to_hva (vcpu, gfn), &arch);
@@ -4932,7 +4932,7 @@ static void init_kvm_tdp_mmu(struct kvm_vcpu *vcpu) context->shadow_root_level = kvm_x86_ops->get_tdp_level (vcpu); context->direct_map = true; context->set_cr3 = kvm_x86_ops->set_tdp_cr3; - context->get_cr3 = get_cr3; + context->get_guest_cr3_or_eptp = get_cr3; context->get_pdptr = kvm_pdptr_read; context->inject_page_fault = kvm_inject_page_fault;
@@ -5080,10 +5080,10 @@ static void init_kvm_softmmu(struct kvm_vcpu *vcpu)struct kvm_mmu *context = vcpu->arch.mmu;kvm_init_shadow_mmu (vcpu); - context->set_cr3 = kvm_x86_ops->set_cr3; - context->get_cr3 = get_cr3; - context->get_pdptr = kvm_pdptr_read; - context->inject_page_fault = kvm_inject_page_fault; + context->set_cr3 = kvm_x86_ops->set_cr3; + context->get_guest_cr3_or_eptp = get_cr3; + context->get_pdptr = kvm_pdptr_read; + context->inject_page_fault = kvm_inject_page_fault; }static void init_kvm_nested_mmu (struct kvm_vcpu *vcpu)
@@ -5095,10 +5095,10 @@ static void init_kvm_nested_mmu(struct kvm_vcpu *vcpu)if (new_role.as_u64 == g_context->mmu_role.as_u64)return ; - g_context->mmu_role.as_u64 = new_role.as_u64; - g_context->get_cr3 = get_cr3; - g_context->get_pdptr = kvm_pdptr_read; - g_context->inject_page_fault = kvm_inject_page_fault; + g_context->mmu_role.as_u64 = new_role.as_u64; + g_context->get_guest_cr3_or_eptp = get_cr3; + g_context->get_pdptr = kvm_pdptr_read; + g_context->inject_page_fault = kvm_inject_page_fault;/* * Note that arch.mmu->gva_to_gpa translates l2_gpa to l1_gpa using
arch/x86/kvm/mmu/paging_tmpl.h +1 -1
@@ -333,7 +333,7 @@ static int FNAME(walk_addr_generic)(struct guest_walker *walker,trace_kvm_mmu_pagetable_walk (addr, access); retry_walk: walker->level = mmu->root_level; - pte = mmu->get_cr3 (vcpu); + pte = mmu->get_guest_cr3_or_eptp (vcpu); have_ad =PT_HAVE_ACCESSED_DIRTY (mmu);#if PTTYPE ==64
arch/x86/kvm/svm.c +5 -5
@@ -3026,11 +3026,11 @@ static void nested_svm_init_mmu_context(struct kvm_vcpu *vcpu) vcpu->arch.mmu = &vcpu->arch.guest_mmu;kvm_init_shadow_mmu (vcpu); - vcpu->arch.mmu->set_cr3 = nested_svm_set_tdp_cr3; - vcpu->arch.mmu->get_cr3 = nested_svm_get_tdp_cr3; - vcpu->arch.mmu->get_pdptr = nested_svm_get_tdp_pdptr; - vcpu->arch.mmu->inject_page_fault = nested_svm_inject_npf_exit; - vcpu->arch.mmu->shadow_root_level =get_npt_level (vcpu); + vcpu->arch.mmu->set_cr3 = nested_svm_set_tdp_cr3; + vcpu->arch.mmu->get_guest_cr3_or_eptp = nested_svm_get_tdp_cr3; + vcpu->arch.mmu->get_pdptr = nested_svm_get_tdp_pdptr; + vcpu->arch.mmu->inject_page_fault = nested_svm_inject_npf_exit; + vcpu->arch.mmu->shadow_root_level =get_npt_level (vcpu);reset_shadow_zero_bits_mask (vcpu, vcpu->arch.mmu); vcpu->arch.walk_mmu = &vcpu->arch.nested_mmu; }
arch/x86/kvm/vmx/nested.c +4 -4
@@ -354,10 +354,10 @@ static void nested_ept_init_mmu_context(struct kvm_vcpu *vcpu)VMX_EPT_EXECUTE_ONLY_BIT ,nested_ept_ad_enabled (vcpu), nested_ept_get_eptp(vcpu)); - vcpu->arch.mmu->set_cr3 = vmx_set_cr3; - vcpu->arch.mmu->get_cr3 = nested_ept_get_eptp; - vcpu->arch.mmu->inject_page_fault = nested_ept_inject_page_fault; - vcpu->arch.mmu->get_pdptr = kvm_pdptr_read; + vcpu->arch.mmu->set_cr3 = vmx_set_cr3; + vcpu->arch.mmu->get_guest_cr3_or_eptp = nested_ept_get_eptp; + vcpu->arch.mmu->inject_page_fault = nested_ept_inject_page_fault; + vcpu->arch.mmu->get_pdptr = kvm_pdptr_read; vcpu->arch.walk_mmu = &vcpu->arch.nested_mmu; }
arch/x86/kvm/x86.c +1 -1
@@ -10179,7 +10179,7 @@ void kvm_arch_async_page_ready(struct kvm_vcpu *vcpu, struct kvm_async_pf *work)return ;if (!vcpu->arch.mmu->direct_map && - work->arch.cr3 != vcpu->arch.mmu->get_cr3 (vcpu)) + work->arch.cr3 != vcpu->arch.mmu->get_guest_cr3_or_eptp (vcpu))return ; vcpu->arch.mmu->page_fault (vcpu, work->cr2_or_gpa,0 , true);
Signature · 1 line
2.24.1
Paolo Bonzini
maintainerRe: [PATCH v2 6/7] KVM: x86/mmu: Rename kvm_mmu->get_cr3() to ->get_guest_cr3_or_eptp()
Quoted 7 lines · Sean Christopherson
Rename
kvm_mmu->get_cr3()to call out that it is retrieving a guest value, as opposed tokvm_mmu->set_cr3(), which sets a host value, and to note that it will return L1's EPTP when nested EPT is in use. Hopefully the new name will also make it more obvious that L1'snested_cr3is returned in SVM's nested NPT case.No functional change intended.
Should we call it "get_pgd", since that is how Linux calls the top-level directory? I always get confused by PUD/PMD, but as long as we only keep one /p.d/ moniker it should be fine.
Paolo
Quoted 160 lines
Signed-off-by Sean Christopherson <[email protected]>--- arch/x86/include/asm/kvm_host.h | 2 +- arch/x86/kvm/mmu/mmu.c | 24 ++++++++++++------------ arch/x86/kvm/mmu/paging_tmpl.h | 2 +- arch/x86/kvm/svm.c | 10 +++++----- arch/x86/kvm/vmx/nested.c | 8 ++++---- arch/x86/kvm/x86.c | 2 +- 6 files changed, 24 insertions(+), 24 deletions(-)arch/x86/include/asm/kvm_host.h +1
-1@@ -383,7 +383,7 @@ struct kvm_mmu_root_info {*/ struct kvm_mmu {void (*set_cr3)(struct kvm_vcpu *vcpu,unsigned long root); -unsigned long (*get_cr3)(struct kvm_vcpu *vcpu); +unsigned long (*get_guest_cr3_or_eptp)(struct kvm_vcpu *vcpu);u64 (*get_pdptr)(struct kvm_vcpu *vcpu,int index);int (*page_fault)(struct kvm_vcpu *vcpu,gpa_t cr2_or_gpa,u32 err,bool prefault);arch/x86/kvm/mmu/mmu.c +12
-12@@ -3731,7 +3731,7 @@ static int mmu_alloc_direct_roots(struct kvm_vcpu *vcpu) vcpu->arch.mmu->root_hpa =__pa (vcpu->arch.mmu->pae_root); }else BUG (); - vcpu->arch.mmu->root_cr3 = vcpu->arch.mmu->get_cr3 (vcpu); + vcpu->arch.mmu->root_cr3 = vcpu->arch.mmu->get_guest_cr3_or_eptp (vcpu);return 0 ; }@@ -3743,7 +3743,7 @@ static int mmu_alloc_shadow_roots(struct kvm_vcpu *vcpu)gfn_t root_gfn, root_cr3;int i; - root_cr3 = vcpu->arch.mmu->get_cr3 (vcpu); + root_cr3 = vcpu->arch.mmu->get_guest_cr3_or_eptp (vcpu); root_gfn = root_cr3 >>PAGE_SHIFT ;if (mmu_check_root (vcpu, root_gfn))@@ -4080,7 +4080,7 @@ static int kvm_arch_setup_async_pf(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa, arch.token = (vcpu->arch.apf.id++ <<12 ) | vcpu->vcpu_id; arch.gfn = gfn; arch.direct_map = vcpu->arch.mmu->direct_map; - arch.cr3 = vcpu->arch.mmu->get_cr3 (vcpu); + arch.cr3 = vcpu->arch.mmu->get_guest_cr3_or_eptp (vcpu);return kvm_setup_async_pf (vcpu, cr2_or_gpa,kvm_vcpu_gfn_to_hva (vcpu, gfn), &arch);@@ -4932,7 +4932,7 @@ static void init_kvm_tdp_mmu(struct kvm_vcpu *vcpu) context->shadow_root_level = kvm_x86_ops->get_tdp_level (vcpu); context->direct_map = true; context->set_cr3 = kvm_x86_ops->set_tdp_cr3; - context->get_cr3 = get_cr3; + context->get_guest_cr3_or_eptp = get_cr3; context->get_pdptr = kvm_pdptr_read; context->inject_page_fault = kvm_inject_page_fault;@@ -5080,10 +5080,10 @@ static void init_kvm_softmmu(struct kvm_vcpu *vcpu)struct kvm_mmu *context = vcpu->arch.mmu;kvm_init_shadow_mmu (vcpu); - context->set_cr3 = kvm_x86_ops->set_cr3; - context->get_cr3 = get_cr3; - context->get_pdptr = kvm_pdptr_read; - context->inject_page_fault = kvm_inject_page_fault; + context->set_cr3 = kvm_x86_ops->set_cr3; + context->get_guest_cr3_or_eptp = get_cr3; + context->get_pdptr = kvm_pdptr_read; + context->inject_page_fault = kvm_inject_page_fault; }static void init_kvm_nested_mmu (struct kvm_vcpu *vcpu)@@ -5095,10 +5095,10 @@ static void init_kvm_nested_mmu(struct kvm_vcpu *vcpu)if (new_role.as_u64 == g_context->mmu_role.as_u64)return ; - g_context->mmu_role.as_u64 = new_role.as_u64; - g_context->get_cr3 = get_cr3; - g_context->get_pdptr = kvm_pdptr_read; - g_context->inject_page_fault = kvm_inject_page_fault; + g_context->mmu_role.as_u64 = new_role.as_u64; + g_context->get_guest_cr3_or_eptp = get_cr3; + g_context->get_pdptr = kvm_pdptr_read; + g_context->inject_page_fault = kvm_inject_page_fault;/* * Note that arch.mmu->gva_to_gpa translates l2_gpa to l1_gpa using arch/x86/kvm/mmu/paging_tmpl.h +1
-1@@ -333,7 +333,7 @@ static int FNAME(walk_addr_generic)(struct guest_walker *walker,trace_kvm_mmu_pagetable_walk (addr, access); retry_walk: walker->level = mmu->root_level; - pte = mmu->get_cr3 (vcpu); + pte = mmu->get_guest_cr3_or_eptp (vcpu); have_ad =PT_HAVE_ACCESSED_DIRTY (mmu);#if PTTYPE ==64 arch/x86/kvm/svm.c +5
-5@@ -3026,11 +3026,11 @@ static void nested_svm_init_mmu_context(struct kvm_vcpu *vcpu) vcpu->arch.mmu = &vcpu->arch.guest_mmu;kvm_init_shadow_mmu (vcpu); - vcpu->arch.mmu->set_cr3 = nested_svm_set_tdp_cr3; - vcpu->arch.mmu->get_cr3 = nested_svm_get_tdp_cr3; - vcpu->arch.mmu->get_pdptr = nested_svm_get_tdp_pdptr; - vcpu->arch.mmu->inject_page_fault = nested_svm_inject_npf_exit; - vcpu->arch.mmu->shadow_root_level =get_npt_level (vcpu); + vcpu->arch.mmu->set_cr3 = nested_svm_set_tdp_cr3; + vcpu->arch.mmu->get_guest_cr3_or_eptp = nested_svm_get_tdp_cr3; + vcpu->arch.mmu->get_pdptr = nested_svm_get_tdp_pdptr; + vcpu->arch.mmu->inject_page_fault = nested_svm_inject_npf_exit; + vcpu->arch.mmu->shadow_root_level =get_npt_level (vcpu);reset_shadow_zero_bits_mask (vcpu, vcpu->arch.mmu); vcpu->arch.walk_mmu = &vcpu->arch.nested_mmu; }arch/x86/kvm/vmx/nested.c +4
-4@@ -354,10 +354,10 @@ static void nested_ept_init_mmu_context(struct kvm_vcpu *vcpu)VMX_EPT_EXECUTE_ONLY_BIT ,nested_ept_ad_enabled (vcpu), nested_ept_get_eptp(vcpu)); - vcpu->arch.mmu->set_cr3 = vmx_set_cr3; - vcpu->arch.mmu->get_cr3 = nested_ept_get_eptp; - vcpu->arch.mmu->inject_page_fault = nested_ept_inject_page_fault; - vcpu->arch.mmu->get_pdptr = kvm_pdptr_read; + vcpu->arch.mmu->set_cr3 = vmx_set_cr3; + vcpu->arch.mmu->get_guest_cr3_or_eptp = nested_ept_get_eptp; + vcpu->arch.mmu->inject_page_fault = nested_ept_inject_page_fault; + vcpu->arch.mmu->get_pdptr = kvm_pdptr_read; vcpu->arch.walk_mmu = &vcpu->arch.nested_mmu; }arch/x86/kvm/x86.c +1
-1@@ -10179,7 +10179,7 @@ void kvm_arch_async_page_ready(struct kvm_vcpu *vcpu, struct kvm_async_pf *work)return ;if (!vcpu->arch.mmu->direct_map && - work->arch.cr3 != vcpu->arch.mmu->get_cr3 (vcpu)) + work->arch.cr3 != vcpu->arch.mmu->get_guest_cr3_or_eptp (vcpu))return ; vcpu->arch.mmu->page_fault (vcpu, work->cr2_or_gpa,0 , true);
Re: [PATCH v2 6/7] KVM: x86/mmu: Rename kvm_mmu->get_cr3() to ->get_guest_cr3_or_eptp()
Quoted 7 lines · Paolo Bonzini
Quoted 6 lines · Sean Christopherson
Rename
kvm_mmu->get_cr3()to call out that it is retrieving a guest value, as opposed tokvm_mmu->set_cr3(), which sets a host value, and to note that it will return L1's EPTP when nested EPT is in use. Hopefully the new name will also make it more obvious that L1'snested_cr3is returned in SVM's nested NPT case.
Quoted 5 lines
Quoted 1 line
No functional change intended.
Should we call it "
get_pgd", since that is how Linux calls the top-level directory? I always get confused by PUD/PMD, but as long as we only keep one /p.d/ moniker it should be fine.
Heh, I have the exact same sentiment. get_pgd() works for me.
Paolo Bonzini
maintainerRe: [PATCH v2 6/7] KVM: x86/mmu: Rename kvm_mmu->get_cr3() to ->get_guest_cr3_or_eptp()
Quoted 8 lines · Sean Christopherson
Quoted 7 lines · Paolo Bonzini
Quoted 6 lines · Sean Christopherson
Rename
kvm_mmu->get_cr3()to call out that it is retrieving a guest value, as opposed tokvm_mmu->set_cr3(), which sets a host value, and to note that it will return L1's EPTP when nested EPT is in use. Hopefully the new name will also make it more obvious that L1'snested_cr3is returned in SVM's nested NPT case.
Quoted 7 lines
Quoted 5 lines
Quoted 1 line
No functional change intended.
Should we call it "
get_pgd", since that is how Linux calls the top-level directory? I always get confused by PUD/PMD, but as long as we only keep one /p.d/ moniker it should be fine.Heh, I have the exact same sentiment.
get_pgd()works for me.
Ok, I'll post a patch that uses get_guest_pgd() as soon as I open kvm/next for 5.7 material.
Paolo
Re: [PATCH v2 6/7] KVM: x86/mmu: Rename kvm_mmu->get_cr3() to ->get_guest_cr3_or_eptp()
Quoted 11 lines · Paolo Bonzini
Quoted 10 lines · Sean Christopherson
Quoted 9 lines · Paolo Bonzini
Quoted 7 lines · Sean Christopherson
Rename
kvm_mmu->get_cr3()to call out that it is retrieving a guest value, as opposed tokvm_mmu->set_cr3(), which sets a host value, and to note that it will return L1's EPTP when nested EPT is in use. Hopefully the new name will also make it more obvious that L1'snested_cr3is returned in SVM's nested NPT case.No functional change intended.
Quoted 8 lines
Quoted 5 lines
Quoted 3 lines
Should we call it "
get_pgd", since that is how Linux calls the top-level directory? I always get confused by PUD/PMD, but as long as we only keep one /p.d/ moniker it should be fine.Heh, I have the exact same sentiment.
get_pgd()works for me.Ok, I'll post a patch that uses
get_guest_pgd()as soon as I open kvm/next for 5.7 material.
I need to resend the 5-level nested EPT support, I'll include this change. Should I also include patches 4, 5 and 7 when I send v3 of that series? Your earlier mail said they were queued for 5.6, but AFAICT only patches 1 and 2 made it into 5.6 (which is not a big deal at all).
Quoted 12 lines · Paolo Bonzini
Quoted 9 lines · Sean Christopherson
Sean Christopherson (7):
KVM: nVMX: Use correct root level for nested EPT shadow page tables
KVM: x86/mmu: Fix structguest_walkerarrays for 5-level paging
KVM: nVMX: Allow L1 to use 5-level page walks for nested EPT
KVM: nVMX: Renamenested_ept_get_cr3()tonested_ept_get_eptp()
KVM: nVMX: Rename EPTP validity helper and associated variables
KVM: x86/mmu: Renamekvm_mmu->get_cr3()to ->get_guest_cr3_or_eptp()
KVM: nVMX: Drop unnecessary check on ept caps for execute-onlyQueued 1-2-4-5-7 (for 5.6), thanks!
PATCH v2 7/7 KVM: nVMX: Drop unnecessary check on ept caps for execute-only
In reply to Sean ChristophersonDrop the call to cpu_has_vmx_ept_execute_only() when calculating which EPT capabilities will be exposed to L1 for nested EPT. The resulting configuration is immediately sanitized by the passed in @ept_caps, and except for the call from vmx_check_processor_compat(), @ept_caps is the capabilities that are queried by cpu_has_vmx_ept_execute_only(). For vmx_check_processor_compat(), KVM *wants* to ignore vmx_capability.ept so that a divergence in EPT capabilities between CPUs is detected.
--- arch/x86/kvm/vmx/nested.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-)
arch/x86/kvm/vmx/nested.c +3 -4
@@ -6071,10 +6071,9 @@ void nested_vmx_setup_ctls_msrs(struct nested_vmx_msrs *msrs, u32 ept_caps,VMX_EPT_PAGE_WALK_4_BIT |VMX_EPT_PAGE_WALK_5_BIT |VMX_EPTP_WB_BIT | -VMX_EPT_INVEPT_BIT ; -if (cpu_has_vmx_ept_execute_only ()) - msrs->ept_caps |= -VMX_EPT_EXECUTE_ONLY_BIT ; +VMX_EPT_INVEPT_BIT | +VMX_EPT_EXECUTE_ONLY_BIT ; + msrs->ept_caps &= ept_caps; msrs->ept_caps |=VMX_EPT_EXTENT_GLOBAL_BIT | VMX_EPT_EXTENT_CONTEXT_BIT | VMX_EPT_2MB_PAGE_BIT |
Signature · 1 line
2.24.1
Quoted 26 lines · Sean Christopherson
Two fixes for 5-level paging bugs with a 100% fatality rate, a patch to enable 5-level EPT in L1, and additional clean up on top (mostly renames of functions/variables that caused me no end of confusion when trying to figure out what was broken).
Tested fixed kernels at L0, L1 and L2, with most combinations of EPT, shadow paging, 4-level and 5-level. EPT kvm-unit-tests runs clean in L0. Patches for kvm-unit-tests incoming to play nice with 5-level nested EPT.
Ideally patches 1 and 2 would get into 5.6, 5-level paging is quite broken without them.
v2:
- Increase the nested EPT array sizes to accomodate 5-level paging in the patch that adds support for 5-level nested EPT, not in the bug fix for 5-level shadow paging.Sean Christopherson (7):
KVM: nVMX: Use correct root level for nested EPT shadow page tables
KVM: x86/mmu: Fix structguest_walkerarrays for 5-level paging
KVM: nVMX: Allow L1 to use 5-level page walks for nested EPT
KVM: nVMX: Renamenested_ept_get_cr3()tonested_ept_get_eptp()
KVM: nVMX: Rename EPTP validity helper and associated variables
KVM: x86/mmu: Renamekvm_mmu->get_cr3()to ->get_guest_cr3_or_eptp()
KVM: nVMX: Drop unnecessary check on ept caps for execute-only
Quoted 11 lines
arch/x86/include/asm/kvm_host.h | 2 +- arch/x86/include/asm/vmx.h | 12 +++++++ arch/x86/kvm/mmu/mmu.c | 35 ++++++++++---------- arch/x86/kvm/mmu/paging_tmpl.h | 6 ++-- arch/x86/kvm/svm.c | 10 +++--- arch/x86/kvm/vmx/nested.c | 58 ++++++++++++++++++++------------- arch/x86/kvm/vmx/nested.h | 4 +-- arch/x86/kvm/vmx/vmx.c | 2 ++ arch/x86/kvm/x86.c | 2 +- 9 files changed, 79 insertions(+), 52 deletions(-)
Queued 1-2-4-5-7 (for 5.6), thanks!
Paolo