Index: head/sys/amd64/amd64/machdep.c
===================================================================
--- head/sys/amd64/amd64/machdep.c
+++ head/sys/amd64/amd64/machdep.c
@@ -131,6 +131,7 @@
 #include <machine/trap.h>
 #include <machine/tss.h>
 #include <x86/ucode.h>
+#include <x86/ifunc.h>
 #ifdef SMP
 #include <machine/smp.h>
 #endif
@@ -2661,3 +2662,34 @@
 }
 
 #endif /* KDB */
+
+#undef memset
+#undef memmove
+#undef memcpy
+
+void	*memset_std(void *buf, int c, size_t len);
+void	*memset_erms(void *buf, int c, size_t len);
+DEFINE_IFUNC(, void *, memset, (void *, int, size_t), static)
+{
+
+	return ((cpu_stdext_feature & CPUID_STDEXT_ERMS) != 0 ?
+		memset_erms : memset_std);
+}
+
+void    *memmove_std(void * _Nonnull dst, const void * _Nonnull src, size_t len);
+void    *memmove_erms(void * _Nonnull dst, const void * _Nonnull src, size_t len);
+DEFINE_IFUNC(, void *, memmove, (void * _Nonnull, const void * _Nonnull, size_t), static)
+{
+
+	return ((cpu_stdext_feature & CPUID_STDEXT_ERMS) != 0 ?
+		memmove_erms : memmove_std);
+}
+
+void    *memcpy_std(void * _Nonnull dst, const void * _Nonnull src, size_t len);
+void    *memcpy_erms(void * _Nonnull dst, const void * _Nonnull src, size_t len);
+DEFINE_IFUNC(, void *, memcpy, (void * _Nonnull, const void * _Nonnull, size_t), static)
+{
+
+	return ((cpu_stdext_feature & CPUID_STDEXT_ERMS) != 0 ?
+		memcpy_erms : memcpy_std);
+}
Index: head/sys/amd64/amd64/support.S
===================================================================
--- head/sys/amd64/amd64/support.S
+++ head/sys/amd64/amd64/support.S
@@ -96,7 +96,7 @@
  * Adapted from bcopy written by:
  *  ws@tools.de     (Wolfgang Solfrank, TooLs GmbH) +49-228-985800
  */
-ENTRY(memmove)
+ENTRY(memmove_std)
 	PUSH_FRAME_POINTER
 	movq	%rdi,%r9
 	movq	%rdx,%rcx
@@ -142,15 +142,45 @@
 	movq	%r9,%rax
 	POP_FRAME_POINTER
 	ret
-END(memmove)
+END(memmove_std)
 
+ENTRY(memmove_erms)
+	PUSH_FRAME_POINTER
+	movq	%rdi,%r9
+	movq	%rdx,%rcx
+
+	movq	%rdi,%rax
+	subq	%rsi,%rax
+	cmpq	%rcx,%rax			/* overlapping && src < dst? */
+	jb	1f
+
+	rep
+	movsb
+	movq	%r9,%rax
+	POP_FRAME_POINTER
+	ret
+
+1:
+	addq	%rcx,%rdi			/* copy backwards */
+	addq	%rcx,%rsi
+	decq	%rdi
+	decq	%rsi
+	std
+	rep
+	movsb
+	cld
+	movq	%r9,%rax
+	POP_FRAME_POINTER
+	ret
+END(memmove_erms)
+
 /*
  * memcpy(dst, src, len)
  *        rdi, rsi, rdx
  *
  * Note: memcpy does not support overlapping copies
  */
-ENTRY(memcpy)
+ENTRY(memcpy_std)
 	PUSH_FRAME_POINTER
 	movq	%rdi,%rax
 	movq	%rdx,%rcx
@@ -167,13 +197,23 @@
 	movsb
 	POP_FRAME_POINTER
 	ret
-END(memcpy)
+END(memcpy_std)
 
+ENTRY(memcpy_erms)
+	PUSH_FRAME_POINTER
+	movq	%rdi,%rax
+	movq	%rdx,%rcx
+	rep
+	movsb
+	POP_FRAME_POINTER
+	ret
+END(memcpy_erms)
+
 /*
  * memset(dst, c,   len)
  *        rdi, rsi, rdx
  */
-ENTRY(memset)
+ENTRY(memset_std)
 	PUSH_FRAME_POINTER
 	movq	%rdi,%r9
 	movq	%rdx,%rcx
@@ -195,7 +235,19 @@
 	movq	%r9,%rax
 	POP_FRAME_POINTER
 	ret
-END(memset)
+END(memset_std)
+
+ENTRY(memset_erms)
+	PUSH_FRAME_POINTER
+	movq	%rdi,%r9
+	movq	%rdx,%rcx
+	movb	%sil,%al
+	rep
+	stosb
+	movq	%r9,%rax
+	POP_FRAME_POINTER
+	ret
+END(memset_erms)
 
 /* fillw(pat, base, cnt) */
 /*       %rdi,%rsi, %rdx */