[PATCH 0/1] drop wrappers of SVM instructions

[PATCH 0/1] drop wrappers of SVM instructions

[Alexander M. Merritt]

Part of cleanup following toolchain baseline upgrade.

As the new toolchain baseline knows the SVM instructions, this series drops
macros with SVM instructions, and updates SVM wrappers to use real instructions
rather than bytecode.

Resolves: xen-project/xen#204

Alexander M. Merritt (1):
  xen/x86: drop wrappers of SVM instructions

 xen/arch/x86/hvm/svm/svm.h           | 14 +++-----------
 xen/arch/x86/include/asm/asm-defns.h | 12 ------------
 2 files changed, 3 insertions(+), 23 deletions(-)

-- 
2.45.2

[XEN PATCH 1/1] xen/x86: drop wrappers of SVM instructions

[Alexander M. Merritt]

The new toolchain baseline knows the SVM instructions.

Resolves: xen-project/xen#204

Signed-off-by: Alexander M. Merritt <alexander@edera.dev>
---
 xen/arch/x86/hvm/svm/svm.h           | 14 +++-----------
 xen/arch/x86/include/asm/asm-defns.h | 12 ------------
 2 files changed, 3 insertions(+), 23 deletions(-)

diff --git a/xen/arch/x86/hvm/svm/svm.h b/xen/arch/x86/hvm/svm/svm.h
index beeaaab4f7..5c5e91a1ac 100644
--- a/xen/arch/x86/hvm/svm/svm.h
+++ b/xen/arch/x86/hvm/svm/svm.h
@@ -23,25 +23,17 @@ void __update_guest_eip(struct cpu_user_regs *regs, unsigned int inst_len);
 
 static inline void svm_vmload_pa(paddr_t vmcb)
 {
-    asm volatile (
-        ".byte 0x0f,0x01,0xda" /* vmload */
-        : : "a" (vmcb) : "memory" );
+    asm volatile ( "vmload" : : "a" (vmcb) : "memory" );
 }
 
 static inline void svm_vmsave_pa(paddr_t vmcb)
 {
-    asm volatile (
-        ".byte 0x0f,0x01,0xdb" /* vmsave */
-        : : "a" (vmcb) : "memory" );
+    asm volatile ( "vmsave" : : "a" (vmcb) : "memory" );
 }
 
 static inline void svm_invlpga(unsigned long linear, uint32_t asid)
 {
-    asm volatile (
-        ".byte 0x0f,0x01,0xdf"
-        : /* output */
-        : /* input */
-        "a" (linear), "c" (asid) );
+    asm volatile ( "invlpga" : /* output */ : /* input */ "a" (linear), "c" (asid) );
 }
 
 /*
diff --git a/xen/arch/x86/include/asm/asm-defns.h b/xen/arch/x86/include/asm/asm-defns.h
index 32d6b44910..5a64b42495 100644
--- a/xen/arch/x86/include/asm/asm-defns.h
+++ b/xen/arch/x86/include/asm/asm-defns.h
@@ -10,18 +10,6 @@
 .endm
 #endif
 
-.macro vmrun
-    .byte 0x0f, 0x01, 0xd8
-.endm
-
-.macro stgi
-    .byte 0x0f, 0x01, 0xdc
-.endm
-
-.macro clgi
-    .byte 0x0f, 0x01, 0xdd
-.endm
-
 /*
  * Call a noreturn function.  This could be JMP, but CALL results in a more
  * helpful backtrace.  BUG is to catch functions which do decide to return...
-- 
2.45.2

Re: [XEN PATCH 1/1] xen/x86: drop wrappers of SVM instructions

[Andrew Cooper]

On 02/04/2025 3:54 pm, Alexander M. Merritt wrote:
> The new toolchain baseline knows the SVM instructions.
>
> Resolves: xen-project/xen#204
>
> Signed-off-by: Alexander M. Merritt <alexander@edera.dev>

Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>, with a couple of
notes.

> ---
>  xen/arch/x86/hvm/svm/svm.h           | 14 +++-----------
>  xen/arch/x86/include/asm/asm-defns.h | 12 ------------
>  2 files changed, 3 insertions(+), 23 deletions(-)
>
> diff --git a/xen/arch/x86/hvm/svm/svm.h b/xen/arch/x86/hvm/svm/svm.h
> index beeaaab4f7..5c5e91a1ac 100644
> --- a/xen/arch/x86/hvm/svm/svm.h
> +++ b/xen/arch/x86/hvm/svm/svm.h
> @@ -23,25 +23,17 @@ void __update_guest_eip(struct cpu_user_regs *regs, unsigned int inst_len);
>  
>  static inline void svm_vmload_pa(paddr_t vmcb)
>  {
> -    asm volatile (
> -        ".byte 0x0f,0x01,0xda" /* vmload */
> -        : : "a" (vmcb) : "memory" );
> +    asm volatile ( "vmload" : : "a" (vmcb) : "memory" );

We tend to prefer :: for empty sections.

>  }
>  
>  static inline void svm_vmsave_pa(paddr_t vmcb)
>  {
> -    asm volatile (
> -        ".byte 0x0f,0x01,0xdb" /* vmsave */
> -        : : "a" (vmcb) : "memory" );
> +    asm volatile ( "vmsave" : : "a" (vmcb) : "memory" );
>  }
>  
>  static inline void svm_invlpga(unsigned long linear, uint32_t asid)
>  {
> -    asm volatile (
> -        ".byte 0x0f,0x01,0xdf"
> -        : /* output */
> -        : /* input */
> -        "a" (linear), "c" (asid) );
> +    asm volatile ( "invlpga" : /* output */ : /* input */ "a" (linear), "c" (asid) );

And here, the output/input comments are useless and can be dropped.

I can fix both on commit if you're happy?

~Andrew

Re: [XEN PATCH 1/1] xen/x86: drop wrappers of SVM instructions

[Alexander Merritt]

On 2025-04-02 10:56, Andrew Cooper wrote:
> On 02/04/2025 3:54 pm, Alexander M. Merritt wrote:

> I can fix both on commit if you're happy?
> 
> ~Andrew

Noted for future updates. Yes please go ahead.

-Alex