The new toolchain baseline knows all of these instructions.

For the remaining uses of .byte for instructions, annotate the toolchain
minima.

No functional change.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Jan Beulich <JBeulich@suse.com>
CC: Roger Pau Monné <roger.pau@citrix.com>
---
 xen/arch/x86/arch.mk                   |  4 +++
 xen/arch/x86/include/asm/asm-defns.h   |  4 ---
 xen/arch/x86/include/asm/msr.h         |  2 ++
 xen/arch/x86/include/asm/prot-key.h    |  6 ++---
 xen/arch/x86/include/asm/xstate.h      |  3 +--
 xen/arch/x86/x86_emulate/0f01.c        |  2 +-
 xen/arch/x86/x86_emulate/x86_emulate.c | 34 ++++++++++++--------------
 7 files changed, 26 insertions(+), 29 deletions(-)

diff --git a/xen/arch/x86/arch.mk b/xen/arch/x86/arch.mk
index 0203138a819a..3d8d9bfe4916 100644
--- a/xen/arch/x86/arch.mk
+++ b/xen/arch/x86/arch.mk
@@ -16,7 +16,11 @@ CFLAGS-$(CONFIG_CC_IS_GCC) += -malign-data=abi
 $(call cc-options-add,CFLAGS,CC,$(EMBEDDED_EXTRA_CFLAGS))
 $(call cc-option-add,CFLAGS,CC,-Wnested-externs)
 $(call as-option-add,CFLAGS,CC,".equ \"x\"$(comma)1",-DHAVE_AS_QUOTED_SYM)
+
+# Binutils >= 2.31, Clang >= 7
 $(call as-option-add,CFLAGS,CC,"movdiri %rax$(comma)(%rax)",-DHAVE_AS_MOVDIR)
+
+# Binutils >= 2.33, Clang >= 9
 $(call as-option-add,CFLAGS,CC,"enqcmd (%rax)$(comma)%rax",-DHAVE_AS_ENQCMD)
 
 # Check to see whether the assembler supports the .nop directive.
diff --git a/xen/arch/x86/include/asm/asm-defns.h b/xen/arch/x86/include/asm/asm-defns.h
index 239dc3af096c..f6fe4596a852 100644
--- a/xen/arch/x86/include/asm/asm-defns.h
+++ b/xen/arch/x86/include/asm/asm-defns.h
@@ -1,9 +1,5 @@
 #include <asm/page-bits.h>
 
-.macro clzero
-    .byte 0x0f, 0x01, 0xfc
-.endm
-
 /* binutils >= 2.41 or LLVM >= 19 */
 .macro eretu
     .byte 0xf3, 0x0f, 0x01, 0xca
diff --git a/xen/arch/x86/include/asm/msr.h b/xen/arch/x86/include/asm/msr.h
index 941a7612f4ba..1377d156f4e1 100644
--- a/xen/arch/x86/include/asm/msr.h
+++ b/xen/arch/x86/include/asm/msr.h
@@ -63,6 +63,8 @@ static inline void wrmsrns(uint32_t msr, uint64_t val)
     /*
      * WRMSR is 2 bytes.  WRMSRNS is 3 bytes.  Pad WRMSR with a redundant CS
      * prefix to avoid a trailing NOP.
+     *
+     * Binutils >= 2.40, Clang >= 16
      */
     alternative_input(".byte 0x2e; wrmsr",
                       ".byte 0x0f,0x01,0xc6", X86_FEATURE_WRMSRNS,
diff --git a/xen/arch/x86/include/asm/prot-key.h b/xen/arch/x86/include/asm/prot-key.h
index 8fb15b5c32e9..1752756fd9c1 100644
--- a/xen/arch/x86/include/asm/prot-key.h
+++ b/xen/arch/x86/include/asm/prot-key.h
@@ -19,16 +19,14 @@ static inline uint32_t rdpkru(void)
 {
     uint32_t pkru;
 
-    asm volatile ( ".byte 0x0f,0x01,0xee"
-                   : "=a" (pkru) : "c" (0) : "dx" );
+    asm volatile ( "rdpkru" : "=a" (pkru) : "c" (0) : "dx" );
 
     return pkru;
 }
 
 static inline void wrpkru(uint32_t pkru)
 {
-    asm volatile ( ".byte 0x0f,0x01,0xef"
-                   :: "a" (pkru), "d" (0), "c" (0) );
+    asm volatile ( "wrpkru" :: "a" (pkru), "d" (0), "c" (0) );
 }
 
 /*
diff --git a/xen/arch/x86/include/asm/xstate.h b/xen/arch/x86/include/asm/xstate.h
index e3b9745543d7..9cfee1fa9c5a 100644
--- a/xen/arch/x86/include/asm/xstate.h
+++ b/xen/arch/x86/include/asm/xstate.h
@@ -120,8 +120,7 @@ static inline uint64_t xgetbv(unsigned int index)
     uint32_t lo, hi;
 
     ASSERT(index); /* get_xcr0() should be used instead. */
-    asm volatile ( ".byte 0x0f,0x01,0xd0" /* xgetbv */
-                   : "=a" (lo), "=d" (hi) : "c" (index) );
+    asm volatile ( "xgetbv" : "=a" (lo), "=d" (hi) : "c" (index) );
 
     return lo | ((uint64_t)hi << 32);
 }
diff --git a/xen/arch/x86/x86_emulate/0f01.c b/xen/arch/x86/x86_emulate/0f01.c
index 1ba99609d6fd..4791465fc83f 100644
--- a/xen/arch/x86/x86_emulate/0f01.c
+++ b/xen/arch/x86/x86_emulate/0f01.c
@@ -122,7 +122,7 @@ int x86emul_0f01(struct x86_emulate_state *s,
         {
         case vex_none: /* serialize */
             host_and_vcpu_must_have(serialize);
-            asm volatile ( ".byte 0x0f, 0x01, 0xe8" );
+            asm volatile ( ".byte 0x0f, 0x01, 0xe8" ); /* Binutils >= 2.34, Clang >= 11 */
             break;
         case vex_f2: /* xsusldtrk */
             vcpu_must_have(tsxldtrk);
diff --git a/xen/arch/x86/x86_emulate/x86_emulate.c b/xen/arch/x86/x86_emulate/x86_emulate.c
index d830aea430d4..a3b7142fde7e 100644
--- a/xen/arch/x86/x86_emulate/x86_emulate.c
+++ b/xen/arch/x86/x86_emulate/x86_emulate.c
@@ -4748,27 +4748,25 @@ x86_emulate(
                  */
                 if ( vex.l )
                 {
-                    /* vpxor %xmmN, %xmmN, %xmmN */
-                    asm volatile ( ".byte 0xc5,0xf9,0xef,0xc0" );
-                    asm volatile ( ".byte 0xc5,0xf1,0xef,0xc9" );
-                    asm volatile ( ".byte 0xc5,0xe9,0xef,0xd2" );
-                    asm volatile ( ".byte 0xc5,0xe1,0xef,0xdb" );
-                    asm volatile ( ".byte 0xc5,0xd9,0xef,0xe4" );
-                    asm volatile ( ".byte 0xc5,0xd1,0xef,0xed" );
-                    asm volatile ( ".byte 0xc5,0xc9,0xef,0xf6" );
-                    asm volatile ( ".byte 0xc5,0xc1,0xef,0xff" );
+                    asm volatile ( "vpxor %xmm0, %xmm0, %xmm0" );
+                    asm volatile ( "vpxor %xmm1, %xmm1, %xmm1" );
+                    asm volatile ( "vpxor %xmm2, %xmm2, %xmm2" );
+                    asm volatile ( "vpxor %xmm3, %xmm3, %xmm3" );
+                    asm volatile ( "vpxor %xmm4, %xmm4, %xmm4" );
+                    asm volatile ( "vpxor %xmm5, %xmm5, %xmm5" );
+                    asm volatile ( "vpxor %xmm6, %xmm6, %xmm6" );
+                    asm volatile ( "vpxor %xmm7, %xmm7, %xmm7" );
                 }
                 else
                 {
-                    /* vpor %xmmN, %xmmN, %xmmN */
-                    asm volatile ( ".byte 0xc5,0xf9,0xeb,0xc0" );
-                    asm volatile ( ".byte 0xc5,0xf1,0xeb,0xc9" );
-                    asm volatile ( ".byte 0xc5,0xe9,0xeb,0xd2" );
-                    asm volatile ( ".byte 0xc5,0xe1,0xeb,0xdb" );
-                    asm volatile ( ".byte 0xc5,0xd9,0xeb,0xe4" );
-                    asm volatile ( ".byte 0xc5,0xd1,0xeb,0xed" );
-                    asm volatile ( ".byte 0xc5,0xc9,0xeb,0xf6" );
-                    asm volatile ( ".byte 0xc5,0xc1,0xeb,0xff" );
+                    asm volatile ( "vpor %xmm0, %xmm0, %xmm0" );
+                    asm volatile ( "vpor %xmm1, %xmm1, %xmm1" );
+                    asm volatile ( "vpor %xmm2, %xmm2, %xmm2" );
+                    asm volatile ( "vpor %xmm3, %xmm3, %xmm3" );
+                    asm volatile ( "vpor %xmm4, %xmm4, %xmm4" );
+                    asm volatile ( "vpor %xmm5, %xmm5, %xmm5" );
+                    asm volatile ( "vpor %xmm6, %xmm6, %xmm6" );
+                    asm volatile ( "vpor %xmm7, %xmm7, %xmm7" );
                 }
 
                 ASSERT(!state->simd_size);
-- 
2.39.5

On 30.12.2025 14:54, Andrew Cooper wrote:
> --- a/xen/arch/x86/include/asm/asm-defns.h
> +++ b/xen/arch/x86/include/asm/asm-defns.h
> @@ -1,9 +1,5 @@
>  #include <asm/page-bits.h>
>  
> -.macro clzero
> -    .byte 0x0f, 0x01, 0xfc
> -.endm

This can't go away yet, as it became known to gas only in 2.26.

> --- a/xen/arch/x86/include/asm/prot-key.h
> +++ b/xen/arch/x86/include/asm/prot-key.h
> @@ -19,16 +19,14 @@ static inline uint32_t rdpkru(void)
>  {
>      uint32_t pkru;
>  
> -    asm volatile ( ".byte 0x0f,0x01,0xee"
> -                   : "=a" (pkru) : "c" (0) : "dx" );
> +    asm volatile ( "rdpkru" : "=a" (pkru) : "c" (0) : "dx" );
>  
>      return pkru;
>  }
>  
>  static inline void wrpkru(uint32_t pkru)
>  {
> -    asm volatile ( ".byte 0x0f,0x01,0xef"
> -                   :: "a" (pkru), "d" (0), "c" (0) );
> +    asm volatile ( "wrpkru" :: "a" (pkru), "d" (0), "c" (0) );
>  }

Same for both of these.

Jan