Patch "KVM: selftests: Make hyperv_clock selftest more stable" has been added to the 5.15-stable tree

[Sasha Levin <sashal@xxxxxxxxxx>]

This is a note to let you know that I've just added the patch titled

    KVM: selftests: Make hyperv_clock selftest more stable

to the 5.15-stable tree which can be found at:
    http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
     kvm-selftests-make-hyperv_clock-selftest-more-stable.patch
and it can be found in the queue-5.15 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.



commit 751876953b7f9686f73d96d8686f1ae06bf84d54
Author: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
Date:   Wed Jun 1 16:43:22 2022 +0200

    KVM: selftests: Make hyperv_clock selftest more stable
    
    [ Upstream commit eae260be3a0111a28fe95923e117a55dddec0384 ]
    
    hyperv_clock doesn't always give a stable test result, especially with
    AMD CPUs. The test compares Hyper-V MSR clocksource (acquired either
    with rdmsr() from within the guest or KVM_GET_MSRS from the host)
    against rdtsc(). To increase the accuracy, increase the measured delay
    (done with nop loop) by two orders of magnitude and take the mean rdtsc()
    value before and after rdmsr()/KVM_GET_MSRS.
    
    Reported-by: Maxim Levitsky <mlevitsk@xxxxxxxxxx>
    Signed-off-by: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
    Reviewed-by: Maxim Levitsky <mlevitsk@xxxxxxxxxx>
    Tested-by: Maxim Levitsky <mlevitsk@xxxxxxxxxx>
    Message-Id: <20220601144322.1968742-1-vkuznets@xxxxxxxxxx>
    Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>
    Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>

diff --git a/tools/testing/selftests/kvm/x86_64/hyperv_clock.c b/tools/testing/selftests/kvm/x86_64/hyperv_clock.c
index e0b2bb1339b1..3330fb183c68 100644
--- a/tools/testing/selftests/kvm/x86_64/hyperv_clock.c
+++ b/tools/testing/selftests/kvm/x86_64/hyperv_clock.c
@@ -44,7 +44,7 @@ static inline void nop_loop(void)
 {
 	int i;
 
-	for (i = 0; i < 1000000; i++)
+	for (i = 0; i < 100000000; i++)
 		asm volatile("nop");
 }
 
@@ -56,12 +56,14 @@ static inline void check_tsc_msr_rdtsc(void)
 	tsc_freq = rdmsr(HV_X64_MSR_TSC_FREQUENCY);
 	GUEST_ASSERT(tsc_freq > 0);
 
-	/* First, check MSR-based clocksource */
+	/* For increased accuracy, take mean rdtsc() before and afrer rdmsr() */
 	r1 = rdtsc();
 	t1 = rdmsr(HV_X64_MSR_TIME_REF_COUNT);
+	r1 = (r1 + rdtsc()) / 2;
 	nop_loop();
 	r2 = rdtsc();
 	t2 = rdmsr(HV_X64_MSR_TIME_REF_COUNT);
+	r2 = (r2 + rdtsc()) / 2;
 
 	GUEST_ASSERT(r2 > r1 && t2 > t1);
 
@@ -181,12 +183,14 @@ static void host_check_tsc_msr_rdtsc(struct kvm_vm *vm)
 	tsc_freq = vcpu_get_msr(vm, VCPU_ID, HV_X64_MSR_TSC_FREQUENCY);
 	TEST_ASSERT(tsc_freq > 0, "TSC frequency must be nonzero");
 
-	/* First, check MSR-based clocksource */
+	/* For increased accuracy, take mean rdtsc() before and afrer ioctl */
 	r1 = rdtsc();
 	t1 = vcpu_get_msr(vm, VCPU_ID, HV_X64_MSR_TIME_REF_COUNT);
+	r1 = (r1 + rdtsc()) / 2;
 	nop_loop();
 	r2 = rdtsc();
 	t2 = vcpu_get_msr(vm, VCPU_ID, HV_X64_MSR_TIME_REF_COUNT);
+	r2 = (r2 + rdtsc()) / 2;
 
 	TEST_ASSERT(t2 > t1, "Time reference MSR is not monotonic (%ld <= %ld)", t1, t2);