Re: [PATCH v5 2/4] selftests: KVM: Remove unused x86_register enum

Re: [PATCH v5 2/4] selftests: KVM: Remove unused x86_register enum

[linmiaohe]

Eric Auger <eric.auger@redhat.com> writes:
>x86_register enum is not used. Its presence incites us to enumerate GPRs in the same order in other looming structs. So let's remove it.
>
>Signed-off-by: Eric Auger <eric.auger@redhat.com>
>Suggested-by: Vitaly Kuznetsov <vkuznets@redhat.com>

Make sense for me. Thanks.

Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>

[PATCH v5 0/4] selftests: KVM: AMD Nested SVM test infrastructure

[Eric Auger]

Add the basic infrastructure needed to test AMD nested SVM.
Also add a first basic vmcall test.

Best regards

Eric

This series can be found at:
https://github.com/eauger/linux/tree/v5.5-amd-svm-v5

History:
v4 -> v5:
- Added "selftests: KVM: Remove unused x86_register enum"
- reorder GPRs within gpr64_regs
- removed vmcb_hva and save_area_hva from svm_test_data
- remove the naming for vmcb_gpa in run_guest

v3 -> v4:
- gpr64_regs struct just contains 64b mode GPRs ordered
  as in x86_register
- cleanup in run_guest (vm* instructions) and reduce
  clubber list.
- add some comments

v2 -> v3:
- Took into account Vitaly's comment:
  - added "selftests: KVM: Replace get_gdt/idt_base() by
    get_gdt/idt()"
  - svm.h now is a copy of arch/x86/include/asm/svm.h
  - avoid duplicates

v1 -> v2:
- split into 2 patches
- remove the infrastructure to run low-level sub-tests and only
  keep vmmcall's one.
- move struct regs into processor.h
- force vmcb_gpa into rax in run_guest()

Eric Auger (4):
  selftests: KVM: Replace get_[gdt | idt]_base() by get_[gdt | idt]()
  selftests: KVM: Remove unused x86_register enum
  selftests: KVM: AMD Nested test infrastructure
  selftests: KVM: SVM: Add vmcall test

 tools/testing/selftests/kvm/Makefile          |   3 +-
 .../selftests/kvm/include/x86_64/processor.h  |  44 +--
 .../selftests/kvm/include/x86_64/svm.h        | 297 ++++++++++++++++++
 .../selftests/kvm/include/x86_64/svm_util.h   |  36 +++
 tools/testing/selftests/kvm/lib/x86_64/svm.c  | 159 ++++++++++
 tools/testing/selftests/kvm/lib/x86_64/vmx.c  |   6 +-
 .../selftests/kvm/x86_64/svm_vmcall_test.c    |  79 +++++
 7 files changed, 598 insertions(+), 26 deletions(-)
 create mode 100644 tools/testing/selftests/kvm/include/x86_64/svm.h
 create mode 100644 tools/testing/selftests/kvm/include/x86_64/svm_util.h
 create mode 100644 tools/testing/selftests/kvm/lib/x86_64/svm.c
 create mode 100644 tools/testing/selftests/kvm/x86_64/svm_vmcall_test.c

-- 
2.20.1

[PATCH v5 2/4] selftests: KVM: Remove unused x86_register enum

[Eric Auger]

x86_register enum is not used. Its presence incites us
to enumerate GPRs in the same order in other looming
structs. So let's remove it.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Suggested-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
 .../selftests/kvm/include/x86_64/processor.h  | 20 -------------------
 1 file changed, 20 deletions(-)

diff --git a/tools/testing/selftests/kvm/include/x86_64/processor.h b/tools/testing/selftests/kvm/include/x86_64/processor.h
index 6f7fffaea2e8..e48dac5c29e8 100644
--- a/tools/testing/selftests/kvm/include/x86_64/processor.h
+++ b/tools/testing/selftests/kvm/include/x86_64/processor.h
@@ -36,26 +36,6 @@
 #define X86_CR4_SMAP		(1ul << 21)
 #define X86_CR4_PKE		(1ul << 22)
 
-/* The enum values match the intruction encoding of each register */
-enum x86_register {
-	RAX = 0,
-	RCX,
-	RDX,
-	RBX,
-	RSP,
-	RBP,
-	RSI,
-	RDI,
-	R8,
-	R9,
-	R10,
-	R11,
-	R12,
-	R13,
-	R14,
-	R15,
-};
-
 struct desc64 {
 	uint16_t limit0;
 	uint16_t base0;
-- 
2.20.1

Re: [PATCH v5 2/4] selftests: KVM: Remove unused x86_register enum

[Vitaly Kuznetsov]

Eric Auger <eric.auger@redhat.com> writes:

> x86_register enum is not used. Its presence incites us
> to enumerate GPRs in the same order in other looming
> structs. So let's remove it.
>
> Signed-off-by: Eric Auger <eric.auger@redhat.com>
> Suggested-by: Vitaly Kuznetsov <vkuznets@redhat.com>
> ---
>  .../selftests/kvm/include/x86_64/processor.h  | 20 -------------------
>  1 file changed, 20 deletions(-)
>
> diff --git a/tools/testing/selftests/kvm/include/x86_64/processor.h b/tools/testing/selftests/kvm/include/x86_64/processor.h
> index 6f7fffaea2e8..e48dac5c29e8 100644
> --- a/tools/testing/selftests/kvm/include/x86_64/processor.h
> +++ b/tools/testing/selftests/kvm/include/x86_64/processor.h
> @@ -36,26 +36,6 @@
>  #define X86_CR4_SMAP		(1ul << 21)
>  #define X86_CR4_PKE		(1ul << 22)
>  
> -/* The enum values match the intruction encoding of each register */
> -enum x86_register {
> -	RAX = 0,
> -	RCX,
> -	RDX,
> -	RBX,
> -	RSP,
> -	RBP,
> -	RSI,
> -	RDI,
> -	R8,
> -	R9,
> -	R10,
> -	R11,
> -	R12,
> -	R13,
> -	R14,
> -	R15,
> -};
> -
>  struct desc64 {
>  	uint16_t limit0;
>  	uint16_t base0;

Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>

-- 
Vitaly

Re: [PATCH v5 2/4] selftests: KVM: Remove unused x86_register enum

[Krish Sadhukhan]

On 2/7/20 6:27 AM, Eric Auger wrote:
> x86_register enum is not used. Its presence incites us
> to enumerate GPRs in the same order in other looming
> structs. So let's remove it.
>
> Signed-off-by: Eric Auger <eric.auger@redhat.com>
> Suggested-by: Vitaly Kuznetsov <vkuznets@redhat.com>
> ---
>   .../selftests/kvm/include/x86_64/processor.h  | 20 -------------------
>   1 file changed, 20 deletions(-)
>
> diff --git a/tools/testing/selftests/kvm/include/x86_64/processor.h b/tools/testing/selftests/kvm/include/x86_64/processor.h
> index 6f7fffaea2e8..e48dac5c29e8 100644
> --- a/tools/testing/selftests/kvm/include/x86_64/processor.h
> +++ b/tools/testing/selftests/kvm/include/x86_64/processor.h
> @@ -36,26 +36,6 @@
>   #define X86_CR4_SMAP		(1ul << 21)
>   #define X86_CR4_PKE		(1ul << 22)
>   
> -/* The enum values match the intruction encoding of each register */
> -enum x86_register {
> -	RAX = 0,
> -	RCX,
> -	RDX,
> -	RBX,
> -	RSP,
> -	RBP,
> -	RSI,
> -	RDI,
> -	R8,
> -	R9,
> -	R10,
> -	R11,
> -	R12,
> -	R13,
> -	R14,
> -	R15,
> -};
> -
>   struct desc64 {
>   	uint16_t limit0;
>   	uint16_t base0;
Reviewed-by: Krish Sadhukhan <krish.sadhukhan@oracle.com>