From: Oliver Upton <oupton@google.com>
To: kvm@vger.kernel.org
Cc: Paolo Bonzini <pbonzini@redhat.com>,
Sean Christopherson <seanjc@google.com>,
Vitaly Kuznetsov <vkuznets@redhat.com>,
Wanpeng Li <wanpengli@tencent.com>,
Jim Mattson <jmattson@google.com>, Joerg Roedel <joro@8bytes.org>,
Oliver Upton <oupton@google.com>
Subject: [PATCH v2 5/7] selftests: KVM: Add test for PERF_GLOBAL_CTRL VMX control MSR bits
Date: Fri, 4 Feb 2022 20:47:03 +0000 [thread overview]
Message-ID: <20220204204705.3538240-6-oupton@google.com> (raw)
In-Reply-To: <20220204204705.3538240-1-oupton@google.com>
Test that the default behavior of KVM is to ignore userspace MSR writes
and conditionally expose the "load IA32_PERF_GLOBAL_CTRL" bits in the
VMX control MSRs if the guest CPUID exposes a supporting vPMU.
Additionally, test that when the corresponding quirk is disabled,
userspace can still clear these bits regardless of what is exposed in
CPUID.
Signed-off-by: Oliver Upton <oupton@google.com>
---
tools/testing/selftests/kvm/.gitignore | 1 +
tools/testing/selftests/kvm/Makefile | 1 +
.../kvm/x86_64/vmx_control_msrs_test.c | 113 ++++++++++++++++++
3 files changed, 115 insertions(+)
create mode 100644 tools/testing/selftests/kvm/x86_64/vmx_control_msrs_test.c
diff --git a/tools/testing/selftests/kvm/.gitignore b/tools/testing/selftests/kvm/.gitignore
index dce7de7755e6..044aef3a8574 100644
--- a/tools/testing/selftests/kvm/.gitignore
+++ b/tools/testing/selftests/kvm/.gitignore
@@ -36,6 +36,7 @@
/x86_64/userspace_io_test
/x86_64/userspace_msr_exit_test
/x86_64/vmx_apic_access_test
+/x86_64/vmx_control_msrs_test
/x86_64/vmx_close_while_nested_test
/x86_64/vmx_dirty_log_test
/x86_64/vmx_exception_with_invalid_guest_state
diff --git a/tools/testing/selftests/kvm/Makefile b/tools/testing/selftests/kvm/Makefile
index 0e4926bc9a58..88b99d9de373 100644
--- a/tools/testing/selftests/kvm/Makefile
+++ b/tools/testing/selftests/kvm/Makefile
@@ -68,6 +68,7 @@ TEST_GEN_PROGS_x86_64 += x86_64/sync_regs_test
TEST_GEN_PROGS_x86_64 += x86_64/userspace_io_test
TEST_GEN_PROGS_x86_64 += x86_64/userspace_msr_exit_test
TEST_GEN_PROGS_x86_64 += x86_64/vmx_apic_access_test
+TEST_GEN_PROGS_x86_64 += x86_64/vmx_control_msrs_test
TEST_GEN_PROGS_x86_64 += x86_64/vmx_close_while_nested_test
TEST_GEN_PROGS_x86_64 += x86_64/vmx_dirty_log_test
TEST_GEN_PROGS_x86_64 += x86_64/vmx_exception_with_invalid_guest_state
diff --git a/tools/testing/selftests/kvm/x86_64/vmx_control_msrs_test.c b/tools/testing/selftests/kvm/x86_64/vmx_control_msrs_test.c
new file mode 100644
index 000000000000..ac5fdeb50eee
--- /dev/null
+++ b/tools/testing/selftests/kvm/x86_64/vmx_control_msrs_test.c
@@ -0,0 +1,113 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * VMX control MSR test
+ *
+ * Copyright (C) 2022 Google LLC.
+ *
+ * Tests for KVM ownership of bits in the VMX entry/exit control MSRs. Checks
+ * that KVM will set owned bits where appropriate, and will not if
+ * KVM_X86_QUIRK_TWEAK_VMX_CTRL_MSRS is disabled.
+ */
+
+#include "kvm_util.h"
+#include "vmx.h"
+
+#define VCPU_ID 0
+
+static void get_vmx_control_msr(struct kvm_vm *vm, uint32_t msr_index,
+ uint32_t *low, uint32_t *high)
+{
+ uint64_t val;
+
+ val = vcpu_get_msr(vm, VCPU_ID, msr_index);
+ *low = val;
+ *high = val >> 32;
+}
+
+static void set_vmx_control_msr(struct kvm_vm *vm, uint32_t msr_index,
+ uint32_t low, uint32_t high)
+{
+ uint64_t val = (((uint64_t) high) << 32) | low;
+
+ vcpu_set_msr(vm, VCPU_ID, msr_index, val);
+}
+
+static void test_vmx_control_msr(struct kvm_vm *vm, uint32_t msr_index, uint32_t set,
+ uint32_t clear, uint32_t exp_set, uint32_t exp_clear)
+{
+ uint32_t low, high;
+
+ get_vmx_control_msr(vm, msr_index, &low, &high);
+
+ high &= ~clear;
+ high |= set;
+
+ set_vmx_control_msr(vm, msr_index, low, high);
+
+ get_vmx_control_msr(vm, msr_index, &low, &high);
+ ASSERT_EQ(high & exp_set, exp_set);
+ ASSERT_EQ(~high & exp_clear, exp_clear);
+}
+
+static void load_perf_global_ctrl_test(struct kvm_vm *vm)
+{
+ uint32_t entry_low, entry_high, exit_low, exit_high;
+ struct kvm_enable_cap cap = {0};
+
+ get_vmx_control_msr(vm, MSR_IA32_VMX_TRUE_ENTRY_CTLS, &entry_low, &entry_high);
+ get_vmx_control_msr(vm, MSR_IA32_VMX_TRUE_EXIT_CTLS, &exit_low, &exit_high);
+
+ if (!(entry_high & VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL) ||
+ !(exit_high & VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL)) {
+ print_skip("\"load IA32_PERF_GLOBAL_CTRL\" VM-{Entry,Exit} controls not supported");
+ return;
+ }
+
+ /*
+ * Test that KVM will set these bits regardless of userspace if the
+ * guest CPUID exposes a supporting vPMU.
+ */
+ test_vmx_control_msr(vm, MSR_IA32_VMX_TRUE_ENTRY_CTLS, 0,
+ VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL,
+ VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL,
+ 0);
+ test_vmx_control_msr(vm, MSR_IA32_VMX_TRUE_EXIT_CTLS, 0,
+ VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL,
+ VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL,
+ 0);
+
+ /*
+ * Disable the quirk, giving userspace control of the VMX capability
+ * MSRs.
+ */
+ cap.cap = KVM_CAP_DISABLE_QUIRKS;
+ cap.args[0] = KVM_X86_QUIRK_TWEAK_VMX_CTRL_MSRS;
+ vm_enable_cap(vm, &cap);
+
+ /*
+ * Test that userspace can clear these bits, even if it exposes a vPMU
+ * that supports IA32_PERF_GLOBAL_CTRL.
+ */
+ test_vmx_control_msr(vm, MSR_IA32_VMX_TRUE_ENTRY_CTLS, 0,
+ VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL,
+ 0,
+ VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL);
+ test_vmx_control_msr(vm, MSR_IA32_VMX_TRUE_EXIT_CTLS, 0,
+ VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL,
+ 0,
+ VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL);
+}
+
+int main(void)
+{
+ struct kvm_vm *vm;
+
+ nested_vmx_check_supported();
+
+ /* No need to run a guest for these tests */
+ vm = vm_create_default(VCPU_ID, 0, NULL);
+
+ load_perf_global_ctrl_test(vm);
+
+ kvm_vm_free(vm);
+}
--
2.35.0.263.gb82422642f-goog
next prev parent reply other threads:[~2022-02-04 20:47 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-02-04 20:46 [PATCH v2 0/7] VMX: nVMX: VMX control MSR fixes Oliver Upton
2022-02-04 20:46 ` [PATCH v2 1/7] KVM: nVMX: Keep KVM updates to BNDCFGS ctrl bits across MSR write Oliver Upton
2022-02-07 17:21 ` Paolo Bonzini
2022-02-07 18:13 ` Sean Christopherson
2022-02-07 18:22 ` Oliver Upton
2022-02-07 18:27 ` Paolo Bonzini
2022-02-07 18:34 ` Sean Christopherson
2022-02-07 18:52 ` Oliver Upton
2022-02-04 20:47 ` [PATCH v2 2/7] KVM: nVMX: Keep KVM updates to PERF_GLOBAL_CTRL " Oliver Upton
2022-02-07 16:33 ` Paolo Bonzini
2022-02-04 20:47 ` [PATCH v2 3/7] KVM: nVMX: Roll all entry/exit ctl updates into a single helper Oliver Upton
2022-02-05 7:43 ` kernel test robot
2022-02-05 19:41 ` Oliver Upton
2022-02-07 17:56 ` Sean Christopherson
2022-02-04 20:47 ` [PATCH v2 4/7] KVM: nVMX: Add a quirk for KVM tweaks to VMX control MSRs Oliver Upton
2022-02-07 18:06 ` Sean Christopherson
2022-02-09 1:50 ` Oliver Upton
2022-02-09 20:23 ` Sean Christopherson
2022-02-04 20:47 ` Oliver Upton [this message]
2022-02-04 20:47 ` [PATCH v2 6/7] selftests: KVM: Add test for BNDCFGS VMX control MSR bits Oliver Upton
2022-02-07 16:42 ` Paolo Bonzini
2022-02-04 20:47 ` [PATCH v2 7/7] KVM: VMX: Use local pointer to vcpu_vmx in vmx_vcpu_after_set_cpuid() Oliver Upton
2022-02-07 16:42 ` Paolo Bonzini
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20220204204705.3538240-6-oupton@google.com \
--to=oupton@google.com \
--cc=jmattson@google.com \
--cc=joro@8bytes.org \
--cc=kvm@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=seanjc@google.com \
--cc=vkuznets@redhat.com \
--cc=wanpengli@tencent.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox