KVM: nVMX: Don't leak L1 MMIO regions to L2
PATCH 1/1 KVM: nVMX: Don't leak L1 MMIO regions to L2
If the "virtualize APIC accesses" VM-execution control is set in the VMCS, the APIC virtualization hardware is triggered when a page walk in VMX non-root mode terminates at a PTE wherein the address of the 4k page frame matches the APIC-access address specified in the VMCS. On hardware, the APIC-access address may be any valid 4k-aligned physical address.
KVM's nVMX implementation enforces the additional constraint that the APIC-access address specified in the vmcs12 must be backed by cacheable memory in L1. If not, L0 will simply clear the "virtualize APIC accesses" VM-execution control in the vmcs02.
The problem with this approach is that the L1 guest has arranged the vmcs12 EPT tables--or shadow page tables, if the "enable EPT" VM-execution control is clear in the vmcs12--so that the L2 guest physical address(es)--or L2 guest linear address(es)--that reference the L2 APIC map to the APIC-access address specified in the vmcs12. Without the "virtualize APIC accesses" VM-execution control in the vmcs02, the APIC accesses in the L2 guest will directly access the APIC-access page in L1.
When there is no mapping whatsoever for the APIC-access address in L1, the L2 VM just loses the intended APIC virtualization. However, when the APIC-access address is mapped to an MMIO region in L1, the L2 guest gets direct access to the L1 MMIO device. For example, if the APIC-access address specified in the vmcs12 is 0xfee00000, then L2 gets direct access to L1's APIC.
Since this vmcs12 configuration is something that KVM cannot faithfully emulate, the appropriate response is to exit to userspace with KVM_INTERNAL_ERROR_EMULATION.
---
v1: Same code as the RFC.
arch/x86/include/asm/kvm_host.h | 2 +- arch/x86/kvm/vmx/nested.c | 46 ++++++++++++++++++--------------- arch/x86/kvm/x86.c | 9 +++++-- 3 files changed, 33 insertions(+), 24 deletions(-)
arch/x86/include/asm/kvm_host.h +1 -1
@@ -1189,7 +1189,7 @@ struct kvm_x86_ops {int (*set_nested_state)(struct kvm_vcpu *vcpu,struct kvm_nested_state __user *user_kvm_nested_state,struct kvm_nested_state *kvm_state); -void (*get_vmcs12_pages)(struct kvm_vcpu *vcpu); +int (*get_vmcs12_pages)(struct kvm_vcpu *vcpu);int (*smi_allowed)(struct kvm_vcpu *vcpu);int (*pre_enter_smm)(struct kvm_vcpu *vcpu,char *smstate);
arch/x86/kvm/vmx/nested.c +25 -21
110 lines not on this page · raw message
arch/x86/kvm/x86.c +7 -2
@@ -7931,8 +7931,13 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)bool req_immediate_exit = false;if (kvm_request_pending (vcpu)) { -if (kvm_check_request (KVM_REQ_GET_VMCS12_PAGES , vcpu)) - kvm_x86_ops->get_vmcs12_pages (vcpu); +if (kvm_check_request (KVM_REQ_GET_VMCS12_PAGES , vcpu)) { + r = kvm_x86_ops->get_vmcs12_pages (vcpu); +if (unlikely (r)) { + r =0 ; + goto out; + } + }if (kvm_check_request (KVM_REQ_MMU_RELOAD , vcpu))kvm_mmu_unload (vcpu);if (kvm_check_request (KVM_REQ_MIGRATE_TIMER , vcpu))
Signature · 1 line
2.23.0.581.g78d2f28ef7-goog
Paolo Bonzini
maintainerAs usual, nothing to say about the behavior, just about the code...
Quoted 6 lines · Jim Mattson
+ * Returns:
+ * 0 - success, i.e. proceed with actual VMEnter+ * -EFAULT - consistency check VMExit + * -EINVAL - consistency check VMFail + * -ENOTSUPP - kvm internal error*/
... the error codes do not mean much here. Can you define an enum instead?
(I also thought about passing the exit reason, where bit 31 could be used to distinguish VMX instruction failure from an entry failure VMexit, which sounds cleaner if you just look at the prototype but becomes messy fairly quickly because you have to pass back the exit qualification too. The from_vmentry argument could become u32 *p_exit_qual and be NULL if not called from VMentry, but it doesn't seem worthwhile at all).
Thanks,
Paolo
Quoted 80 lines
int
nested_vmx_enter_non_root_mode(struct kvm_vcpu *vcpu, bool from_vmentry){arch/x86/kvm/vmx/nested.c +12
-7@@ -3045,6 +3044,7 @@ int nested_vmx_enter_non_root_mode(struct kvm_vcpu *vcpu, bool from_vmentry)bool evaluate_pending_interrupts;u32 exit_reason =EXIT_REASON_INVALID_STATE ;u32 exit_qual; +int r; evaluate_pending_interrupts =exec_controls_get (vmx) & (CPU_BASED_VIRTUAL_INTR_PENDING |CPU_BASED_VIRTUAL_NMI_PENDING );@@ -3081,11 +3081,13 @@ int nested_vmx_enter_non_root_mode(struct kvm_vcpu *vcpu, bool from_vmentry)prepare_vmcs02_early (vmx, vmcs12);if (from_vmentry) { -nested_get_vmcs12_pages (vcpu); + r =nested_get_vmcs12_pages (vcpu); +if (unlikely (r)) +return r;if (nested_vmx_check_vmentry_hw (vcpu)) {vmx_switch_vmcs (vcpu, &vmx->vmcs01); -return -1 ; +return -EINVAL ; }if (nested_vmx_check_guest_state (vcpu, vmcs12, &exit_qual))@@ -3165,14 +3167,14 @@ int nested_vmx_enter_non_root_mode(struct kvm_vcpu *vcpu, bool from_vmentry)vmx_switch_vmcs (vcpu, &vmx->vmcs01);if (!from_vmentry) -return 1 ; +return -EFAULT ;load_vmcs12_host_state (vcpu, vmcs12); vmcs12->vm_exit_reason = exit_reason |VMX_EXIT_REASONS_FAILED_VMENTRY ; vmcs12->exit_qualification = exit_qual;if (enable_shadow_vmcs || vmx->nested.hv_evmcs) vmx->nested.need_vmcs12_to_shadow_sync = true; -return 1 ; +return -EFAULT ; }/* @@ -3246,11 +3248,13 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch) vmx->nested.nested_run_pending =1 ; ret =nested_vmx_enter_non_root_mode (vcpu, true); vmx->nested.nested_run_pending = !ret; -if (ret >0 ) -return 1 ; -else if (ret) +if (ret == -EINVAL )return nested_vmx_failValid (vcpu,VMXERR_ENTRY_INVALID_CONTROL_FIELD ); +else if (ret == -ENOTSUPP ) +return 0 ; +else if (ret) +return 1 ;/* Hide L1D cache contents from the nested guest. */ vmx->vcpu.arch.l1tf_flush_l1d = true;arch/x86/kvm/x86.c +7
-2@@ -7931,8 +7931,13 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)bool req_immediate_exit = false;if (kvm_request_pending (vcpu)) { -if (kvm_check_request (KVM_REQ_GET_VMCS12_PAGES , vcpu)) - kvm_x86_ops->get_vmcs12_pages (vcpu); +if (kvm_check_request (KVM_REQ_GET_VMCS12_PAGES , vcpu)) { + r = kvm_x86_ops->get_vmcs12_pages (vcpu); +if (unlikely (r)) { + r =0 ; + goto out; + } + }if (kvm_check_request (KVM_REQ_MMU_RELOAD , vcpu))kvm_mmu_unload (vcpu);if (kvm_check_request (KVM_REQ_MIGRATE_TIMER , vcpu))
Quoted 11 lines · Paolo Bonzini
As usual, nothing to say about the behavior, just about the code...
Quoted 6 lines · Jim Mattson
+ * Returns:
+ * 0 - success, i.e. proceed with actual VMEnter+ * -EFAULT - consistency check VMExit + * -EINVAL - consistency check VMFail + * -ENOTSUPP - kvm internal error*/
... the error codes do not mean much here. Can you define an enum instead?
Agreed. Liran also suggested an enum.
Quoted 7 lines
(I also thought about passing the exit reason, where bit 31 could be used to distinguish VMX instruction failure from an entry failure VMexit, which sounds cleaner if you just look at the prototype but becomes messy fairly quickly because you have to pass back the exit qualification too. The
from_vmentryargument could become u32*p_exit_qualand be NULL if not called from VMentry, but it doesn't seem worthwhile at all).
Ya. I also tried (and failed) to find a clever solution that didn't require a multi-state return value. As much as I dislike returning an enum, it's the lesser of all evils.