KVM: x86: hyper-v: make KVM_GET_SUPPORTED_HV_CPUID more useful
Subsystems (6)
Vitaly Kuznetsov
maintainerv3 cover KVM: x86: hyper-v: make KVM_GET_SUPPORTED_HV_CPUID more useful
Changes since v2:
- Keep vCPU version of the ioctl intact but make it 'deprecated' in api.rst [Paolo Bonzini]
- First two patches of v2 series already made it to kvm/queue
QEMU series using the feature:
https://lists.gnu.org/archive/html/qemu-devel/2020-09/msg02017.html
Original description:
KVM_GET_SUPPORTED_HV_CPUID was initially implemented as a vCPU ioctl but this is not very useful when VMM is just trying to query which Hyper-V features are supported by the host prior to creating VM/vCPUs. The data in KVM_GET_SUPPORTED_HV_CPUID is mostly static with a few exceptions but it seems we can change this. Add support for KVM_GET_SUPPORTED_HV_CPUID as a system ioctl as well.
QEMU specific description:
In some cases QEMU needs to collect the information about which Hyper-V features are supported by KVM and pass it up the stack. For non-hyper-v features this is done with system-wide KVM_GET_SUPPORTED_CPUID/ KVM_GET_MSRS ioctls but Hyper-V specific features don't get in the output (as Hyper-V CPUIDs intersect with KVM's). In QEMU, CPU feature expansion happens before any KVM vcpus are created so KVM_GET_SUPPORTED_HV_CPUID can't be used in its current shape.
Vitaly Kuznetsov (2):
KVM: x86: hyper-v: allow KVM_GET_SUPPORTED_HV_CPUID as a system ioctl
KVM: selftests: test KVM_GET_SUPPORTED_HV_CPUID as a system ioctl
Documentation/virt/kvm/api.rst | 16 ++-- arch/x86/kvm/hyperv.c | 6 +- arch/x86/kvm/hyperv.h | 4 +- arch/x86/kvm/vmx/evmcs.c | 3 +- arch/x86/kvm/x86.c | 45 ++++++---- include/uapi/linux/kvm.h | 3 +- .../testing/selftests/kvm/include/kvm_util.h | 2 + tools/testing/selftests/kvm/lib/kvm_util.c | 26 ++++++ .../selftests/kvm/x86_64/hyperv_cpuid.c | 87 +++++++++++-------- 9 files changed, 123 insertions(+), 69 deletions(-)
Signature · 1 line
2.25.4
Vitaly Kuznetsov
maintainerPATCH v3 1/2 KVM: x86: hyper-v: allow KVM_GET_SUPPORTED_HV_CPUID as a system ioctl
KVM_GET_SUPPORTED_HV_CPUID is a vCPU ioctl but its output is now independent from vCPU and in some cases VMMs may want to use it as a system ioctl instead. In particular, QEMU doesn CPU feature expansion before any vCPU gets created so KVM_GET_SUPPORTED_HV_CPUID can't be used.
Convert KVM_GET_SUPPORTED_HV_CPUID to 'dual' system/vCPU ioctl with the same meaning.
--- Documentation/virt/kvm/api.rst | 16 ++++++++---- arch/x86/kvm/hyperv.c | 6 ++--- arch/x86/kvm/hyperv.h | 4 +-- arch/x86/kvm/vmx/evmcs.c | 3 +-- arch/x86/kvm/x86.c | 45 ++++++++++++++++++++-------------- include/uapi/linux/kvm.h | 3 ++- 6 files changed, 46 insertions(+), 31 deletions(-)
Documentation/virt/kvm/api.rst +11 -5
@@ -4455,9 +4455,9 @@ that KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2 is present. 4.118 KVM_GET_SUPPORTED_HV_CPUID -------------------------------- -:Capability: KVM_CAP_HYPERV_CPUID +:Capability: KVM_CAP_HYPERV_CPUID (vcpu), KVM_CAP_SYS_HYPERV_CPUID (system) :Architectures: x86 -:Type: vcpu ioctl +:Type: system ioctl, vcpu ioctl :Parameters: struct kvm_cpuid2 (in/out) :Returns: 0 on success, -1 on error
@@ -4502,9 +4502,6 @@ Currently, the following list of CPUID leaves are returned: - HYPERV_CPUID_SYNDBG_INTERFACE - HYPERV_CPUID_SYNDBG_PLATFORM_CAPABILITIES -HYPERV_CPUID_NESTED_FEATURES leaf is only exposed when Enlightened VMCS was -enabled on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS). - Userspace invokes KVM_GET_SUPPORTED_HV_CPUID by passing a kvm_cpuid2 structure with the 'nent' field indicating the number of entries in the variable-size array 'entries'. If the number of entries is too low to describe all Hyper-V
@@ -4515,6 +4512,15 @@ number of valid entries in the 'entries' array, which is then filled. 'index' and 'flags' fields in 'struct kvm_cpuid_entry2' are currently reserved, userspace should not expect to get any particular value there. +Note, vcpu version of KVM_GET_SUPPORTED_HV_CPUID is currently deprecated. Unlike +system ioctl which exposes all supported feature bits unconditionally, vcpu +version has the following quirks: +- HYPERV_CPUID_NESTED_FEATURES leaf and HV_X64_ENLIGHTENED_VMCS_RECOMMENDED + feature bit are only exposed when Enlightened VMCS was previously enabled + on the corresponding vCPU (KVM_CAP_HYPERV_ENLIGHTENED_VMCS). +- HV_STIMER_DIRECT_MODE_AVAILABLE bit is only exposed with in-kernel LAPIC. + (presumes KVM_CREATE_IRQCHIP has already been called). + 4.119 KVM_ARM_VCPU_FINALIZE ---------------------------
arch/x86/kvm/hyperv.c +3 -3
@@ -1951,8 +1951,8 @@ int kvm_vm_ioctl_hv_eventfd(struct kvm *kvm, struct kvm_hyperv_eventfd *args)return kvm_hv_eventfd_assign (kvm, args->conn_id, args->fd); } -int kvm_vcpu_ioctl_get_hv_cpuid (struct kvm_vcpu *vcpu,struct kvm_cpuid2 *cpuid, -struct kvm_cpuid_entry2 __user *entries) +int kvm_get_hv_cpuid (struct kvm_vcpu *vcpu,struct kvm_cpuid2 *cpuid, +struct kvm_cpuid_entry2 __user *entries) {uint16_t evmcs_ver =0 ;struct kvm_cpuid_entry2 cpuid_entries[] = {
@@ -2037,7 +2037,7 @@ int kvm_vcpu_ioctl_get_hv_cpuid(struct kvm_vcpu *vcpu, struct kvm_cpuid2 *cpuid,* Direct Synthetic timers only make sense with in-kernel * LAPIC */ -if (lapic_in_kernel (vcpu)) +if (!vcpu ||lapic_in_kernel (vcpu)) ent->edx |=HV_STIMER_DIRECT_MODE_AVAILABLE ;break ;
arch/x86/kvm/hyperv.h +2 -2
@@ -126,7 +126,7 @@ void kvm_hv_setup_tsc_page(struct kvm *kvm,void kvm_hv_init_vm (struct kvm *kvm);void kvm_hv_destroy_vm (struct kvm *kvm);int kvm_vm_ioctl_hv_eventfd (struct kvm *kvm,struct kvm_hyperv_eventfd *args); -int kvm_vcpu_ioctl_get_hv_cpuid (struct kvm_vcpu *vcpu,struct kvm_cpuid2 *cpuid, -struct kvm_cpuid_entry2 __user *entries); +int kvm_get_hv_cpuid (struct kvm_vcpu *vcpu,struct kvm_cpuid2 *cpuid, +struct kvm_cpuid_entry2 __user *entries);#endif
arch/x86/kvm/vmx/evmcs.c +1 -2
@@ -327,7 +327,6 @@ bool nested_enlightened_vmentry(struct kvm_vcpu *vcpu, u64 *evmcs_gpa)uint16_t nested_get_evmcs_version (struct kvm_vcpu *vcpu) { -struct vcpu_vmx *vmx =to_vmx (vcpu);/* * vmcs_version represents the range of supported Enlightened VMCS * versions: lower 8 bits is the minimal version, higher 8 bits is the
@@ -335,7 +334,7 @@ uint16_t nested_get_evmcs_version(struct kvm_vcpu *vcpu)* KVM_EVMCS_VERSION. */ if (kvm_cpu_cap_get (X86_FEATURE_VMX ) && - vmx->nested.enlightened_vmcs_enabled) + (!vcpu ||to_vmx (vcpu)->nested.enlightened_vmcs_enabled))return (KVM_EVMCS_VERSION <<8 ) |1 ;return 0 ;
arch/x86/kvm/x86.c +27 -18
70 lines not on this page · raw message
include/uapi/linux/kvm.h +2 -1
@@ -1052,6 +1052,7 @@ struct kvm_ppc_resize_hpt {#define KVM_CAP_STEAL_TIME 187#define KVM_CAP_X86_USER_SPACE_MSR 188#define KVM_CAP_X86_MSR_FILTER 189 +#define KVM_CAP_SYS_HYPERV_CPUID 190#ifdef KVM_CAP_IRQ_ROUTING
@@ -1510,7 +1511,7 @@ struct kvm_enc_region {/* Available with KVM_CAP_MANUAL_DIRTY_LOG_PROTECT_2 */ #define KVM_CLEAR_DIRTY_LOG _IOWR(KVMIO, 0xc0, struct kvm_clear_dirty_log) -/* Available with KVM_CAP_HYPERV_CPUID */ +/* Available with KVM_CAP_HYPERV_CPUID (vcpu) / KVM_CAP_SYS_HYPERV_CPUID (system) */ #define KVM_GET_SUPPORTED_HV_CPUID _IOWR(KVMIO, 0xc1, struct kvm_cpuid2)/* Available with KVM_CAP_ARM_SVE */
Signature · 1 line
2.25.4
Vitaly Kuznetsov
maintainerPATCH v3 2/2 KVM: selftests: test KVM_GET_SUPPORTED_HV_CPUID as a system ioctl
In reply to Vitaly KuznetsovKVM_GET_SUPPORTED_HV_CPUID is now supported as both vCPU and VM ioctl, test that.
--- .../testing/selftests/kvm/include/kvm_util.h | 2 + tools/testing/selftests/kvm/lib/kvm_util.c | 26 ++++++ .../selftests/kvm/x86_64/hyperv_cpuid.c | 87 +++++++++++-------- 3 files changed, 77 insertions(+), 38 deletions(-)
tools/testing/selftests/kvm/include/kvm_util.h +2 -0
@@ -112,6 +112,8 @@ void vcpu_ioctl(struct kvm_vm *vm, uint32_t vcpuid, unsigned long ioctl,int _vcpu_ioctl (struct kvm_vm *vm,uint32_t vcpuid,unsigned long ioctl,void *arg);void vm_ioctl (struct kvm_vm *vm,unsigned long ioctl,void *arg); +void kvm_ioctl (struct kvm_vm *vm,unsigned long ioctl,void *arg); +int _kvm_ioctl (struct kvm_vm *vm,unsigned long ioctl,void *arg);void vm_mem_region_set_flags (struct kvm_vm *vm,uint32_t slot,uint32_t flags);void vm_mem_region_move (struct kvm_vm *vm,uint32_t slot,uint64_t new_gpa);void vm_mem_region_delete (struct kvm_vm *vm,uint32_t slot);
tools/testing/selftests/kvm/lib/kvm_util.c +26 -0
@@ -1518,6 +1518,32 @@ void vm_ioctl(struct kvm_vm *vm, unsigned long cmd, void *arg) cmd, ret, errno, strerror(errno)); } +/* +* KVM system ioctl +* +* Input Args: +* vm - Virtual Machine +* cmd - Ioctl number +* arg - Argument to pass to the ioctl +* +* Return: None +* +* Issues an arbitrary ioctl on a KVM fd. +*/ +void kvm_ioctl (struct kvm_vm *vm,unsigned long cmd,void *arg) +{ +int ret; + + ret =ioctl (vm->kvm_fd, cmd, arg); +TEST_ASSERT (ret ==0 ,"KVM ioctl %lu failed, rc: %i errno: %i (%s)" , + cmd, ret, errno,strerror (errno)); +} + +int _kvm_ioctl (struct kvm_vm *vm,unsigned long cmd,void *arg) +{ +return ioctl (vm->kvm_fd, cmd, arg); +} +/* * VM Dump *
tools/testing/selftests/kvm/x86_64/hyperv_cpuid.c +49 -38
154 lines not on this page · raw message
Signature · 1 line
2.25.4
Vitaly Kuznetsov
maintainerQuoted 30 lines · Vitaly Kuznetsov
Changes since v2:
- Keep vCPU version of the ioctl intact but make it 'deprecated' inapi.rst[Paolo Bonzini]
- First two patches of v2 series already made it to kvm/queueQEMU series using the feature:
https://lists.gnu.org/archive/html/qemu-devel/2020-09/msg02017.htmlOriginal description:
KVM_GET_SUPPORTED_HV_CPUIDwas initially implemented as a vCPU ioctl but this is not very useful when VMM is just trying to query which Hyper-V features are supported by the host prior to creating VM/vCPUs. The data inKVM_GET_SUPPORTED_HV_CPUIDis mostly static with a few exceptions but it seems we can change this. Add support forKVM_GET_SUPPORTED_HV_CPUIDas a system ioctl as well.QEMU specific description:
In some cases QEMU needs to collect the information about which Hyper-V features are supported by KVM and pass it up the stack. For non-hyper-v features this is done with system-wide KVM_GET_SUPPORTED_CPUID/KVM_GET_MSRSioctls but Hyper-V specific features don't get in the output (as Hyper-V CPUIDs intersect with KVM's). In QEMU, CPU feature expansion happens before any KVM vcpus are created soKVM_GET_SUPPORTED_HV_CPUIDcan't be used in its current shape.Vitaly Kuznetsov (2):
KVM: x86: hyper-v: allowKVM_GET_SUPPORTED_HV_CPUIDas a system ioctl
KVM: selftests: testKVM_GET_SUPPORTED_HV_CPUIDas a system ioctl
Quoted 10 lines
Documentation/virt/kvm/api.rst | 16 ++-- arch/x86/kvm/hyperv.c | 6 +- arch/x86/kvm/hyperv.h | 4 +- arch/x86/kvm/vmx/evmcs.c | 3 +- arch/x86/kvm/x86.c | 45 ++++++---- include/uapi/linux/kvm.h | 3 +- .../testing/selftests/kvm/include/kvm_util.h | 2 + tools/testing/selftests/kvm/lib/kvm_util.c | 26 ++++++ .../selftests/kvm/x86_64/hyperv_cpuid.c | 87 +++++++++++-------- 9 files changed, 123 insertions(+), 69 deletions(-)
Ping)
Still hoping this can be picked up for 5.10.
the latest QEMU patchset was posted last Friday:
https://lists.gnu.org/archive/html/qemu-devel/2020-10/msg02443.html
Signature · 1 line
Vitaly
Quoted 30 lines · Vitaly Kuznetsov
Changes since v2:
- Keep vCPU version of the ioctl intact but make it 'deprecated' inapi.rst[Paolo Bonzini]
- First two patches of v2 series already made it to kvm/queueQEMU series using the feature:
https://lists.gnu.org/archive/html/qemu-devel/2020-09/msg02017.htmlOriginal description:
KVM_GET_SUPPORTED_HV_CPUIDwas initially implemented as a vCPU ioctl but this is not very useful when VMM is just trying to query which Hyper-V features are supported by the host prior to creating VM/vCPUs. The data inKVM_GET_SUPPORTED_HV_CPUIDis mostly static with a few exceptions but it seems we can change this. Add support forKVM_GET_SUPPORTED_HV_CPUIDas a system ioctl as well.QEMU specific description:
In some cases QEMU needs to collect the information about which Hyper-V features are supported by KVM and pass it up the stack. For non-hyper-v features this is done with system-wide KVM_GET_SUPPORTED_CPUID/KVM_GET_MSRSioctls but Hyper-V specific features don't get in the output (as Hyper-V CPUIDs intersect with KVM's). In QEMU, CPU feature expansion happens before any KVM vcpus are created soKVM_GET_SUPPORTED_HV_CPUIDcan't be used in its current shape.Vitaly Kuznetsov (2):
KVM: x86: hyper-v: allowKVM_GET_SUPPORTED_HV_CPUIDas a system ioctl
KVM: selftests: testKVM_GET_SUPPORTED_HV_CPUIDas a system ioctl
Quoted 11 lines
Documentation/virt/kvm/api.rst | 16 ++-- arch/x86/kvm/hyperv.c | 6 +- arch/x86/kvm/hyperv.h | 4 +- arch/x86/kvm/vmx/evmcs.c | 3 +- arch/x86/kvm/x86.c | 45 ++++++---- include/uapi/linux/kvm.h | 3 +- .../testing/selftests/kvm/include/kvm_util.h | 2 + tools/testing/selftests/kvm/lib/kvm_util.c | 26 ++++++ .../selftests/kvm/x86_64/hyperv_cpuid.c | 87 +++++++++++-------- 9 files changed, 123 insertions(+), 69 deletions(-)
Queued, thanks.
Paolo