From: Ștefan Șicleru <ssicleru@xxxxxxxxxxxxxxx>
The introspection tool can use #VE to reduce the number of VM-exits
caused by SPT violations for some guests.
Signed-off-by: Ștefan Șicleru <ssicleru@xxxxxxxxxxxxxxx>
Signed-off-by: Adalbert Lazăr <alazar@xxxxxxxxxxxxxxx>
---
Documentation/virt/kvm/kvmi.rst | 63 +++++++++++++++++++
arch/x86/include/uapi/asm/kvmi.h | 8 +++
arch/x86/kvm/kvmi.c | 19 ++++++
include/uapi/linux/kvmi.h | 2 +
.../testing/selftests/kvm/x86_64/kvmi_test.c | 52 +++++++++++++++
virt/kvm/introspection/kvmi_int.h | 3 +
virt/kvm/introspection/kvmi_msg.c | 30 +++++++++
7 files changed, 177 insertions(+)
diff --git a/Documentation/virt/kvm/kvmi.rst b/Documentation/virt/kvm/kvmi.rst
index caa51fccc463..c50c40638d46 100644
--- a/Documentation/virt/kvm/kvmi.rst
+++ b/Documentation/virt/kvm/kvmi.rst
@@ -1230,6 +1230,69 @@ is terminated.
* -KVM_EINVAL - padding is not zero
* -KVM_EINVAL - the selected EPT view is not valid
+29. KVMI_VCPU_SET_VE_INFO
+-------------------------
+
+:Architecture: x86
+:Versions: >= 1
+:Parameters:
+
+::
+
+ struct kvmi_vcpu_hdr;
+ struct kvmi_vcpu_set_ve_info {
+ __u64 gpa;
+ __u8 trigger_vmexit;
+ __u8 padding1;
+ __u16 padding2;
+ __u32 padding3;
+ };
+
+:Returns:
+
+::
+
+ struct kvmi_error_code;
+
+Configures the guest physical address for the #VE info page and enables
+the #VE mechanism. If ``trigger_vmexit`` is true, any virtualization
+exception will trigger a vm-exit. Otherwise, the exception is delivered
+using gate descriptor 20 from the Interrupt Descriptor Table (IDT).
+
+:Errors:
+
+* -KVM_EINVAL - the selected vCPU is invalid
+* -KVM_EINVAL - one of the specified GPAs is invalid
+* -KVM_EOPNOTSUPP - the hardware does not support #VE
+* -KVM_EINVAL - padding is not zero
+* -KVM_EAGAIN - the selected vCPU can't be introspected yet
+
+30. KVMI_VCPU_DISABLE_VE
+------------------------
+
+:Architecture: x86
+:Versions: >= 1
+:Parameters:
+
+::
+
+ struct kvmi_vcpu_hdr;
+
+:Returns:
+
+::
+
+ struct kvmi_error_code;
+
+Disables the #VE mechanism. All EPT violations will trigger a vm-exit,
+regardless of the corresponding spte 63rd bit (SVE) for the GPA that
+triggered the EPT violation within a specific EPT view.
+
+:Errors:
+
+* -KVM_EINVAL - the selected vCPU is invalid
+* -KVM_EAGAIN - the selected vCPU can't be introspected yet
+
Events
======
diff --git a/arch/x86/include/uapi/asm/kvmi.h b/arch/x86/include/uapi/asm/kvmi.h
index 56992dacfb69..d925e6d49f50 100644
--- a/arch/x86/include/uapi/asm/kvmi.h
+++ b/arch/x86/include/uapi/asm/kvmi.h
@@ -174,4 +174,12 @@ struct kvmi_vcpu_control_ept_view {
__u32 padding2;
};
+struct kvmi_vcpu_set_ve_info {
+ __u64 gpa;
+ __u8 trigger_vmexit;
+ __u8 padding1;
+ __u16 padding2;
+ __u32 padding3;
+};
+
#endif /* _UAPI_ASM_X86_KVMI_H */
diff --git a/arch/x86/kvm/kvmi.c b/arch/x86/kvm/kvmi.c
index 3e8c83623703..e101ac390809 100644
--- a/arch/x86/kvm/kvmi.c
+++ b/arch/x86/kvm/kvmi.c
@@ -1464,3 +1464,22 @@ int kvmi_arch_cmd_control_ept_view(struct kvm_vcpu *vcpu, u16 view,
return kvm_x86_ops.control_ept_view(vcpu, view, visible);
}
+
+int kvmi_arch_cmd_set_ve_info(struct kvm_vcpu *vcpu, u64 gpa,
+ bool trigger_vmexit)
+{
+ unsigned long ve_info = (unsigned long) gpa;
+
+ if (!kvm_x86_ops.set_ve_info)
+ return -KVM_EINVAL;
+
+ return kvm_x86_ops.set_ve_info(vcpu, ve_info, trigger_vmexit);
+}
+
+int kvmi_arch_cmd_disable_ve(struct kvm_vcpu *vcpu)
+{
+ if (!kvm_x86_ops.disable_ve)
+ return 0;
+
+ return kvm_x86_ops.disable_ve(vcpu);
+}
diff --git a/include/uapi/linux/kvmi.h b/include/uapi/linux/kvmi.h
index 505a865cd115..a17cd1fa16d0 100644
--- a/include/uapi/linux/kvmi.h
+++ b/include/uapi/linux/kvmi.h
@@ -52,6 +52,8 @@ enum {
KVMI_VCPU_GET_EPT_VIEW = 26,
KVMI_VCPU_SET_EPT_VIEW = 27,
KVMI_VCPU_CONTROL_EPT_VIEW = 28,
+ KVMI_VCPU_SET_VE_INFO = 29,
+ KVMI_VCPU_DISABLE_VE = 30,
KVMI_NUM_MESSAGES
};
diff --git a/tools/testing/selftests/kvm/x86_64/kvmi_test.c b/tools/testing/selftests/kvm/x86_64/kvmi_test.c
index 4e099cbfcf4e..a3ea22f546ec 100644
--- a/tools/testing/selftests/kvm/x86_64/kvmi_test.c
+++ b/tools/testing/selftests/kvm/x86_64/kvmi_test.c
@@ -35,6 +35,10 @@ static vm_vaddr_t test_gva;
static void *test_hva;
static vm_paddr_t test_gpa;
+static vm_vaddr_t test_ve_info_gva;
+static void *test_ve_info_hva;
+static vm_paddr_t test_ve_info_gpa;
+
static uint8_t test_write_pattern;
static int page_size;
@@ -2258,6 +2262,43 @@ static void test_cmd_vcpu_vmfunc(struct kvm_vm *vm)
test_guest_switch_to_visible_view(vm);
}
+static void enable_ve(struct kvm_vm *vm)
+{
+ struct {
+ struct kvmi_msg_hdr hdr;
+ struct kvmi_vcpu_hdr vcpu_hdr;
+ struct kvmi_vcpu_set_ve_info cmd;
+ } req = {};
+
+ req.cmd.gpa = test_ve_info_gpa;
+ req.cmd.trigger_vmexit = 1;
+
+ test_vcpu0_command(vm, KVMI_VCPU_SET_VE_INFO, &req.hdr,
+ sizeof(req), NULL, 0);
+}
+
+static void disable_ve(struct kvm_vm *vm)
+{
+ struct {
+ struct kvmi_msg_hdr hdr;
+ struct kvmi_vcpu_hdr vcpu_hdr;
+ } req = {};
+
+ test_vcpu0_command(vm, KVMI_VCPU_DISABLE_VE, &req.hdr,
+ sizeof(req), NULL, 0);
+}
+
+static void test_virtualization_exceptions(struct kvm_vm *vm)
+{
+ if (!features.ve) {
+ print_skip("#VE not supported");
+ return;
+ }
+
+ enable_ve(vm);
+ disable_ve(vm);
+}
+
static void test_introspection(struct kvm_vm *vm)
{
srandom(time(0));
@@ -2297,6 +2338,7 @@ static void test_introspection(struct kvm_vm *vm)
test_cmd_vcpu_get_ept_view(vm);
test_cmd_vcpu_set_ept_view(vm);
test_cmd_vcpu_vmfunc(vm);
+ test_virtualization_exceptions(vm);
unhook_introspection(vm);
}
@@ -2311,6 +2353,16 @@ static void setup_test_pages(struct kvm_vm *vm)
memset(test_hva, 0, page_size);
test_gpa = addr_gva2gpa(vm, test_gva);
+
+ /* Allocate #VE info page */
+ test_ve_info_gva = vm_vaddr_alloc(vm, page_size, KVM_UTIL_MIN_VADDR,
+ 0, 0);
+ sync_global_to_guest(vm, test_ve_info_gva);
+
+ test_ve_info_hva = addr_gva2hva(vm, test_ve_info_gva);
+ memset(test_ve_info_hva, 0, page_size);
+
+ test_ve_info_gpa = addr_gva2gpa(vm, test_ve_info_gva);
}
int main(int argc, char *argv[])
diff --git a/virt/kvm/introspection/kvmi_int.h b/virt/kvm/introspection/kvmi_int.h
index fc6dbd3a6472..a0062fbde49e 100644
--- a/virt/kvm/introspection/kvmi_int.h
+++ b/virt/kvm/introspection/kvmi_int.h
@@ -151,5 +151,8 @@ u16 kvmi_arch_cmd_get_ept_view(struct kvm_vcpu *vcpu);
int kvmi_arch_cmd_set_ept_view(struct kvm_vcpu *vcpu, u16 view);
int kvmi_arch_cmd_control_ept_view(struct kvm_vcpu *vcpu, u16 view,
bool visible);
+int kvmi_arch_cmd_set_ve_info(struct kvm_vcpu *vcpu, u64 gpa,
+ bool trigger_vmexit);
+int kvmi_arch_cmd_disable_ve(struct kvm_vcpu *vcpu);
#endif
diff --git a/virt/kvm/introspection/kvmi_msg.c b/virt/kvm/introspection/kvmi_msg.c
index 696857f6d008..664b78d545c3 100644
--- a/virt/kvm/introspection/kvmi_msg.c
+++ b/virt/kvm/introspection/kvmi_msg.c
@@ -711,6 +711,34 @@ static int handle_vcpu_control_ept_view(const struct kvmi_vcpu_msg_job *job,
return kvmi_msg_vcpu_reply(job, msg, ec, NULL, 0);
}
+static int handle_vcpu_set_ve_info(const struct kvmi_vcpu_msg_job *job,
+ const struct kvmi_msg_hdr *msg,
+ const void *_req)
+{
+ const struct kvmi_vcpu_set_ve_info *req = _req;
+ bool trigger_vmexit = !!req->trigger_vmexit;
+ int ec;
+
+ if (req->padding1 || req->padding2 || req->padding3)
+ ec = -KVM_EINVAL;
+ else
+ ec = kvmi_arch_cmd_set_ve_info(job->vcpu, req->gpa,
+ trigger_vmexit);
+
+ return kvmi_msg_vcpu_reply(job, msg, ec, NULL, 0);
+}
+
+static int handle_vcpu_disable_ve(const struct kvmi_vcpu_msg_job *job,
+ const struct kvmi_msg_hdr *msg,
+ const void *req)
+{
+ int ec;
+
+ ec = kvmi_arch_cmd_disable_ve(job->vcpu);
+
+ return kvmi_msg_vcpu_reply(job, msg, ec, NULL, 0);
+}
+
/*
* These functions are executed from the vCPU thread. The receiving thread
* passes the messages using a newly allocated 'struct kvmi_vcpu_msg_job'
@@ -725,6 +753,7 @@ static int(*const msg_vcpu[])(const struct kvmi_vcpu_msg_job *,
[KVMI_VCPU_CONTROL_EVENTS] = handle_vcpu_control_events,
[KVMI_VCPU_CONTROL_MSR] = handle_vcpu_control_msr,
[KVMI_VCPU_CONTROL_SINGLESTEP] = handle_vcpu_control_singlestep,
+ [KVMI_VCPU_DISABLE_VE] = handle_vcpu_disable_ve,
[KVMI_VCPU_GET_CPUID] = handle_vcpu_get_cpuid,
[KVMI_VCPU_GET_EPT_VIEW] = handle_vcpu_get_ept_view,
[KVMI_VCPU_GET_INFO] = handle_vcpu_get_info,
@@ -736,6 +765,7 @@ static int(*const msg_vcpu[])(const struct kvmi_vcpu_msg_job *,
[KVMI_VCPU_SET_EPT_VIEW] = handle_vcpu_set_ept_view,
[KVMI_VCPU_SET_REGISTERS] = handle_vcpu_set_registers,
[KVMI_VCPU_SET_XSAVE] = handle_vcpu_set_xsave,
+ [KVMI_VCPU_SET_VE_INFO] = handle_vcpu_set_ve_info,
[KVMI_VCPU_TRANSLATE_GVA] = handle_vcpu_translate_gva,
};
