[PATCH v2 0/4] KVM: x86: selftests: Add Hygon CPUs support and fix failures

Zhiquan Li posted 4 patches 1 month, 2 weeks ago
.../testing/selftests/kvm/include/x86/processor.h |  7 +++++++
tools/testing/selftests/kvm/lib/x86/processor.c   | 15 +++++++++++----
.../selftests/kvm/x86/fix_hypercall_test.c        |  2 +-
tools/testing/selftests/kvm/x86/msrs_test.c       |  4 ++--
.../selftests/kvm/x86/pmu_event_filter_test.c     |  3 ++-
.../testing/selftests/kvm/x86/xapic_state_test.c  |  2 +-
6 files changed, 24 insertions(+), 9 deletions(-)
[PATCH v2 0/4] KVM: x86: selftests: Add Hygon CPUs support and fix failures
Posted by Zhiquan Li 1 month, 2 weeks ago
This series to add support for Hygon CPUs and fix 11 KVM selftest failures
on Hygon architecture.

Patch 1 add CPU vendor detection for Hygon and add a global variable
"host_cpu_is_hygon" to identify if the test is running on a Hygon CPU.
It is the prerequisite for the following fixes.

Patch 2 add a flag to identify AMD compatible CPU and figure out the
compatible cases, so that Hygon CPUs can re-use them.
Following test failures on Hygon platform can be fixed by this patch:
- access_tracking_perf_test
- demand_paging_test
- dirty_log_perf_test
- dirty_log_test
- kvm_page_table_test
- memslot_modification_stress_test
- pre_fault_memory_test
- x86/dirty_log_page_splitting_test
- x86/fix_hypercall_test

Patch 3 fix x86/pmu_event_filter_test failure by allowing the tests for
Hygon CPUs.

Patch 4 fix x86/msrs_test failure while writing the MSR_TSC_AUX reserved
bits without RDPID support.
Sean has made a perfect solution for the issue and provided the patch.
It has been verified on Intel, AMD and Hygon platforms, no regression.

---

V1: https://lore.kernel.org/kvm/20260209041305.64906-1-zhiquan_li@163.com/T/#t

Changes since V1:
- Rebased to kvm-x86/next.
- Followed Sean's suggestion, added a flag to identify AMD compatible test
  cases, then v1/patch 2 and v1/patch 3 can be combined to v2/patch 2.
- Followed Sean's suggestion, simplified patch 4, that is v2/patch 3 now.
- Sean provided the v2/patch 4 for the issue reported by v1/patch5, I
  replaced my SoB with "Reported-by" tag.

---

Sean Christopherson (1):
  KVM: selftests: Fix reserved value WRMSR testcase for multi-feature
    MSRs

Zhiquan Li (3):
  KVM: x86: selftests: Add CPU vendor detection for Hygon
  KVM: x86: selftests: Add a flag to identify AMD compatible test cases
  KVM: x86: selftests: Allow the PMU event filter test for Hygon

 .../testing/selftests/kvm/include/x86/processor.h |  7 +++++++
 tools/testing/selftests/kvm/lib/x86/processor.c   | 15 +++++++++++----
 .../selftests/kvm/x86/fix_hypercall_test.c        |  2 +-
 tools/testing/selftests/kvm/x86/msrs_test.c       |  4 ++--
 .../selftests/kvm/x86/pmu_event_filter_test.c     |  3 ++-
 .../testing/selftests/kvm/x86/xapic_state_test.c  |  2 +-
 6 files changed, 24 insertions(+), 9 deletions(-)


base-commit: e944fe2c09f405a2e2d147145c9b470084bc4c9a
-- 
2.43.0
Re: [PATCH v2 0/4] KVM: x86: selftests: Add Hygon CPUs support and fix failures
Posted by Sean Christopherson 3 weeks, 6 days ago
On Thu, 12 Feb 2026 18:38:37 +0800, Zhiquan Li wrote:
> This series to add support for Hygon CPUs and fix 11 KVM selftest failures
> on Hygon architecture.
> 
> Patch 1 add CPU vendor detection for Hygon and add a global variable
> "host_cpu_is_hygon" to identify if the test is running on a Hygon CPU.
> It is the prerequisite for the following fixes.
> 
> [...]

Applied to kvm-x86 selftests, thanks!

[1/4] KVM: x86: selftests: Add CPU vendor detection for Hygon
      https://github.com/kvm-x86/linux/commit/0c96c47d4345
[2/4] KVM: x86: selftests: Add a flag to identify AMD compatible test cases
      https://github.com/kvm-x86/linux/commit/53b2869231d3
[3/4] KVM: x86: selftests: Allow the PMU event filter test for Hygon
      https://github.com/kvm-x86/linux/commit/6b8b11ba4715
[4/4] KVM: selftests: Fix reserved value WRMSR testcase for multi-feature MSRs
      https://github.com/kvm-x86/linux/commit/9396cc1e282a

--
https://github.com/kvm-x86/linux/tree/next
Re: [PATCH v2 0/4] KVM: x86: selftests: Add Hygon CPUs support and fix failures
Posted by Zhiquan Li 1 month ago
On 2/12/26 18:38, Zhiquan Li wrote:
> This series to add support for Hygon CPUs and fix 11 KVM selftest failures
> on Hygon architecture.
> 
> Patch 1 add CPU vendor detection for Hygon and add a global variable
> "host_cpu_is_hygon" to identify if the test is running on a Hygon CPU.
> It is the prerequisite for the following fixes.
> 
> Patch 2 add a flag to identify AMD compatible CPU and figure out the
> compatible cases, so that Hygon CPUs can re-use them.
> Following test failures on Hygon platform can be fixed by this patch:
> - access_tracking_perf_test
> - demand_paging_test
> - dirty_log_perf_test
> - dirty_log_test
> - kvm_page_table_test
> - memslot_modification_stress_test
> - pre_fault_memory_test
> - x86/dirty_log_page_splitting_test
> - x86/fix_hypercall_test
> 
> Patch 3 fix x86/pmu_event_filter_test failure by allowing the tests for
> Hygon CPUs.
> 
> Patch 4 fix x86/msrs_test failure while writing the MSR_TSC_AUX reserved
> bits without RDPID support.
> Sean has made a perfect solution for the issue and provided the patch.
> It has been verified on Intel, AMD and Hygon platforms, no regression.
> 

Kindly ping for any review on these.

Best Regards,
Zhiquan

> ---
> 
> V1: https://lore.kernel.org/kvm/20260209041305.64906-1-zhiquan_li@163.com/T/#t
> 
> Changes since V1:
> - Rebased to kvm-x86/next.
> - Followed Sean's suggestion, added a flag to identify AMD compatible test
>   cases, then v1/patch 2 and v1/patch 3 can be combined to v2/patch 2.
> - Followed Sean's suggestion, simplified patch 4, that is v2/patch 3 now.
> - Sean provided the v2/patch 4 for the issue reported by v1/patch5, I
>   replaced my SoB with "Reported-by" tag.
> 
> ---
> 
> Sean Christopherson (1):
>   KVM: selftests: Fix reserved value WRMSR testcase for multi-feature
>     MSRs
> 
> Zhiquan Li (3):
>   KVM: x86: selftests: Add CPU vendor detection for Hygon
>   KVM: x86: selftests: Add a flag to identify AMD compatible test cases
>   KVM: x86: selftests: Allow the PMU event filter test for Hygon
> 
>  .../testing/selftests/kvm/include/x86/processor.h |  7 +++++++
>  tools/testing/selftests/kvm/lib/x86/processor.c   | 15 +++++++++++----
>  .../selftests/kvm/x86/fix_hypercall_test.c        |  2 +-
>  tools/testing/selftests/kvm/x86/msrs_test.c       |  4 ++--
>  .../selftests/kvm/x86/pmu_event_filter_test.c     |  3 ++-
>  .../testing/selftests/kvm/x86/xapic_state_test.c  |  2 +-
>  6 files changed, 24 insertions(+), 9 deletions(-)
> 
> 
> base-commit: e944fe2c09f405a2e2d147145c9b470084bc4c9a
Re: [PATCH v2 0/4] KVM: x86: selftests: Add Hygon CPUs support and fix failures
Posted by Sean Christopherson 1 month ago
On Tue, Feb 24, 2026, Zhiquan Li wrote:
> 
> On 2/12/26 18:38, Zhiquan Li wrote:
> > This series to add support for Hygon CPUs and fix 11 KVM selftest failures
> > on Hygon architecture.
> > 
> > Patch 1 add CPU vendor detection for Hygon and add a global variable
> > "host_cpu_is_hygon" to identify if the test is running on a Hygon CPU.
> > It is the prerequisite for the following fixes.
> > 
> > Patch 2 add a flag to identify AMD compatible CPU and figure out the
> > compatible cases, so that Hygon CPUs can re-use them.
> > Following test failures on Hygon platform can be fixed by this patch:
> > - access_tracking_perf_test
> > - demand_paging_test
> > - dirty_log_perf_test
> > - dirty_log_test
> > - kvm_page_table_test
> > - memslot_modification_stress_test
> > - pre_fault_memory_test
> > - x86/dirty_log_page_splitting_test
> > - x86/fix_hypercall_test
> > 
> > Patch 3 fix x86/pmu_event_filter_test failure by allowing the tests for
> > Hygon CPUs.
> > 
> > Patch 4 fix x86/msrs_test failure while writing the MSR_TSC_AUX reserved
> > bits without RDPID support.
> > Sean has made a perfect solution for the issue and provided the patch.
> > It has been verified on Intel, AMD and Hygon platforms, no regression.
> > 
> 
> Kindly ping for any review on these.

They're in my queue, but I typically don't apply anything except urgent fixes
until after -rc2.