Re: [PATCH v14 22/22] KVM: selftest: Add tests for KVM_{GET,SET}_ONE_REG

[Sean Christopherson <seanjc@google.com>]

On Tue, Sep 09, 2025, Chao Gao wrote:
> Add tests for newly added KVM_{GET,SET}_ONE_REG support for x86. Verify the
> new ioctls can read and write real MSRs and synthetic MSRs.
> 
> Signed-off-by: Chao Gao <chao.gao@intel.com>
> ---
>  tools/arch/x86/include/uapi/asm/kvm.h         | 29 ++++++++++++++++++
>  tools/testing/selftests/kvm/Makefile.kvm      |  1 +
>  .../selftests/kvm/x86/get_set_one_reg.c       | 30 +++++++++++++++++++
>  3 files changed, 60 insertions(+)
>  create mode 100644 tools/testing/selftests/kvm/x86/get_set_one_reg.c
> 
> diff --git a/tools/arch/x86/include/uapi/asm/kvm.h b/tools/arch/x86/include/uapi/asm/kvm.h
> index 6f3499507c5e..59ac0b46ebcc 100644
> --- a/tools/arch/x86/include/uapi/asm/kvm.h
> +++ b/tools/arch/x86/include/uapi/asm/kvm.h

Don't copy KVM headers to tools/, KVM selftests don't actually use them (i.e.
copying them is confusing/misleadling).  The copied headers are mainly used by
tools/perf, and they run a script to synchronize everything.

> diff --git a/tools/testing/selftests/kvm/Makefile.kvm b/tools/testing/selftests/kvm/Makefile.kvm
> index f6fe7a07a0a2..9a375d5faf1c 100644
> --- a/tools/testing/selftests/kvm/Makefile.kvm
> +++ b/tools/testing/selftests/kvm/Makefile.kvm
> @@ -136,6 +136,7 @@ TEST_GEN_PROGS_x86 += x86/max_vcpuid_cap_test
>  TEST_GEN_PROGS_x86 += x86/triple_fault_event_test
>  TEST_GEN_PROGS_x86 += x86/recalc_apic_map_test
>  TEST_GEN_PROGS_x86 += x86/aperfmperf_test
> +TEST_GEN_PROGS_x86 += x86/get_set_one_reg
>  TEST_GEN_PROGS_x86 += access_tracking_perf_test
>  TEST_GEN_PROGS_x86 += coalesced_io_test
>  TEST_GEN_PROGS_x86 += dirty_log_perf_test
> diff --git a/tools/testing/selftests/kvm/x86/get_set_one_reg.c b/tools/testing/selftests/kvm/x86/get_set_one_reg.c
> new file mode 100644
> index 000000000000..8a4dbc812214
> --- /dev/null
> +++ b/tools/testing/selftests/kvm/x86/get_set_one_reg.c
> @@ -0,0 +1,30 @@
> +// SPDX-License-Identifier: GPL-2.0
> +#include <fcntl.h>
> +#include <stdint.h>
> +#include <sys/ioctl.h>
> +
> +#include "test_util.h"
> +#include "kvm_util.h"
> +#include "processor.h"
> +
> +int main(int argc, char *argv[])
> +{
> +	struct kvm_vcpu *vcpu;
> +	struct kvm_vm *vm;
> +	u64 data;
> +
> +	TEST_REQUIRE(kvm_has_cap(KVM_CAP_ONE_REG));
> +
> +	vm = vm_create_with_one_vcpu(&vcpu, NULL);
> +
> +	TEST_ASSERT_EQ(__vcpu_get_reg(vcpu, KVM_X86_REG_MSR(MSR_EFER), &data), 0);
> +	TEST_ASSERT_EQ(__vcpu_set_reg(vcpu, KVM_X86_REG_MSR(MSR_EFER), data), 0);
> +
> +	if (kvm_cpu_has(X86_FEATURE_SHSTK)) {
> +		TEST_ASSERT_EQ(__vcpu_get_reg(vcpu, KVM_X86_REG_KVM(KVM_REG_GUEST_SSP), &data), 0);
> +		TEST_ASSERT_EQ(__vcpu_set_reg(vcpu, KVM_X86_REG_KVM(KVM_REG_GUEST_SSP), data), 0);

This isn't a very useful test, nor is it extensible.  I finally bit the bullet
and created an MSR test to mostly replace KUT's msr.c, and to add coverage for
KVM_{G,S}ET_ONE_REG and KVM_GET_REG_LIST.