1 files changed, 38 insertions, 0 deletions

diff --git a/arch/x86_64/lib/clear_page.S b/arch/x86_64/lib/clear_page.S
index 43d9fa136180..1f81b79b796c 100644
--- a/arch/x86_64/lib/clear_page.S
+++ b/arch/x86_64/lib/clear_page.S
@@ -5,8 +5,46 @@
         .globl clear_page
         .p2align 4
 clear_page:
+        xorl   %eax,%eax
+        movl   $4096/64,%ecx
+        .p2align 4
+.Lloop:
+        decl    %ecx
+#define PUT(x) movq %rax,x*8(%rdi)
+        movq %rax,(%rdi)
+        PUT(1)
+        PUT(2)
+        PUT(3)
+        PUT(4)
+        PUT(5)
+        PUT(6)
+        PUT(7)
+        leaq    64(%rdi),%rdi
+        jnz     .Lloop
+        nop
+        ret
+clear_page_end:
+
+        /* Some CPUs run faster using the string instructions.
+           It is also a lot simpler. Use this when possible */
+
+#include <asm/cpufeature.h>
+
+        .section .altinstructions,"a"
+        .align 8
+        .quad  clear_page
+        .quad  clear_page_c
+        .byte  X86_FEATURE_REP_GOOD
+        .byte  clear_page_end-clear_page
+        .byte  clear_page_c_end-clear_page_c
+        .previous
+
+        .section .altinstr_replacement,"ax"
+clear_page_c:
         movl $4096/8,%ecx
         xorl %eax,%eax
         rep 
         stosq
         ret
+clear_page_c_end:
+        .previous