diff --git a/sys/arm64/arm64/identcpu.c b/sys/arm64/arm64/identcpu.c
--- a/sys/arm64/arm64/identcpu.c
+++ b/sys/arm64/arm64/identcpu.c
@@ -2675,14 +2675,15 @@
 
 	if (cpu == 0) {
 		/* Create a user visible cpu description with safe values */
-		memset(&user_cpu_desc, 0, sizeof(user_cpu_desc));
+		memset_early(&user_cpu_desc, 0, sizeof(user_cpu_desc));
 		/* Safe values for these registers */
 		user_cpu_desc.id_aa64pfr0 = ID_AA64PFR0_AdvSIMD_NONE |
 		    ID_AA64PFR0_FP_NONE | ID_AA64PFR0_EL1_64 |
 		    ID_AA64PFR0_EL0_64;
 		user_cpu_desc.id_aa64dfr0 = ID_AA64DFR0_DebugVer_8;
 		/* Create the Linux user visible cpu description */
-		memcpy(&l_user_cpu_desc, &user_cpu_desc, sizeof(user_cpu_desc));
+		memcpy_early(&l_user_cpu_desc, &user_cpu_desc,
+		    sizeof(user_cpu_desc));
 	}
 
 	desc = get_cpu_desc(cpu);
diff --git a/sys/arm64/arm64/machdep.c b/sys/arm64/arm64/machdep.c
--- a/sys/arm64/arm64/machdep.c
+++ b/sys/arm64/arm64/machdep.c
@@ -80,6 +80,7 @@
 #include <machine/cpu_feat.h>
 #include <machine/debug_monitor.h>
 #include <machine/hypervisor.h>
+#include <machine/ifunc.h>
 #include <machine/kdb.h>
 #include <machine/machdep.h>
 #include <machine/metadata.h>
@@ -809,6 +810,9 @@
 
 	update_special_regs(0);
 
+	sched_instance_select();
+	link_elf_ireloc();
+
 	/* Set the pcpu data, this is needed by pmap_bootstrap */
 	pcpup = &pcpu0;
 	pcpu_init(pcpup, 0, sizeof(struct pcpu));
@@ -825,8 +829,6 @@
 	PCPU_SET(curthread, &thread0);
 	PCPU_SET(midr, get_midr());
 
-	sched_instance_select();
-	link_elf_ireloc();
 #ifdef FDT
 	try_load_dtb();
 #endif
@@ -1078,3 +1080,35 @@
 		db_printf("show vtop <virt_addr>\n");
 }
 #endif
+
+#undef memset
+#undef memmove
+#undef memcpy
+
+void	*memset_std(void *buf, int c, size_t len);
+void	*memset_mops(void *buf, int c, size_t len);
+void    *memmove_std(void * _Nonnull dst, const void * _Nonnull src,
+	    size_t len);
+void    *memmove_mops(void * _Nonnull dst, const void * _Nonnull src,
+	    size_t len);
+void    *memcpy_std(void * _Nonnull dst, const void * _Nonnull src,
+	    size_t len);
+void    *memcpy_mops(void * _Nonnull dst, const void * _Nonnull src,
+	    size_t len);
+
+DEFINE_IFUNC(, void *, memset, (void *, int, size_t))
+{
+	return ((elf_hwcap2 & HWCAP2_MOPS) != 0 ? memset_mops : memset_std);
+}
+
+DEFINE_IFUNC(, void *, memmove, (void * _Nonnull, const void * _Nonnull,
+    size_t))
+{
+	return ((elf_hwcap2 & HWCAP2_MOPS) != 0 ? memmove_mops : memmove_std);
+}
+
+DEFINE_IFUNC(, void *, memcpy, (void * _Nonnull, const void * _Nonnull,
+    size_t))
+{
+	return ((elf_hwcap2 & HWCAP2_MOPS) != 0 ? memcpy_mops : memcpy_std);
+}
diff --git a/sys/arm64/arm64/machdep_boot.c b/sys/arm64/arm64/machdep_boot.c
--- a/sys/arm64/arm64/machdep_boot.c
+++ b/sys/arm64/arm64/machdep_boot.c
@@ -115,7 +115,7 @@
 		PRELOAD_PUSH_VALUE(uint32_t, MODINFO_METADATA | MODINFOMD_DTBP);
 		PRELOAD_PUSH_VALUE(uint32_t, sizeof(uint64_t));
 		PRELOAD_PUSH_VALUE(uint64_t, (uint64_t)lastaddr);
-		memmove((void *)lastaddr, dtb_ptr, dtb_size);
+		memmove_early((void *)lastaddr, dtb_ptr, dtb_size);
 		lastaddr += dtb_size;
 		lastaddr = roundup(lastaddr, sizeof(int));
 	}
diff --git a/sys/arm64/arm64/memcpy.S b/sys/arm64/arm64/memcpy.S
--- a/sys/arm64/arm64/memcpy.S
+++ b/sys/arm64/arm64/memcpy.S
@@ -57,8 +57,8 @@
    The loop tail is handled by always copying 64 bytes from the end.
 */
 
-EENTRY(memmove)
-ENTRY(memcpy)
+EENTRY(memmove_std)
+ENTRY(memcpy_std)
 	add	srcend, src, count
 	add	dstend, dstin, count
 	cmp	count, 128
@@ -239,7 +239,23 @@
 	stp	B_l, B_h, [dstin, 16]
 	stp	C_l, C_h, [dstin]
 	ret
-END(memcpy)
-EEND(memmove)
+END(memcpy_std)
+EEND(memmove_std)
+
+ENTRY(memcpy_mops)
+	mov	x3, x0
+	.inst	0x19010443	/* cpyfp   [x3]!, [x1]!, x2!  */
+	.inst	0x19410443	/* cpyfm   [x3]!, [x1]!, x2!  */
+	.inst	0x19810443	/* cpyfe   [x3]!, [x1]!, x2!  */
+	ret
+END(memcpy_mops)
+
+ENTRY(memmove_mops)
+	mov	x3, x0
+	.inst	0x1d010443	/* cpyp    [x3]!, [x1]!, x2!  */
+	.inst	0x1d410443	/* cpym    [x3]!, [x1]!, x2!  */
+	.inst	0x1d810443	/* cpye    [x3]!, [x1]!, x2!  */
+	ret
+END(memmove_mops)
 
 GNU_PROPERTY_AARCH64_FEATURE_1_NOTE(GNU_PROPERTY_AARCH64_FEATURE_1_VAL)
diff --git a/sys/arm64/arm64/memset.S b/sys/arm64/arm64/memset.S
--- a/sys/arm64/arm64/memset.S
+++ b/sys/arm64/arm64/memset.S
@@ -51,7 +51,7 @@
 #define dst		x8
 #define tmp3w		w9
 
-ENTRY(memset)
+ENTRY(memset_std)
 
 	mov	dst, dstin		/* Preserve return value.  */
 	ands	A_lw, val, #255
@@ -196,6 +196,14 @@
 	ands	count, count, zva_bits_x
 	b.ne	.Ltail_maybe_long
 	ret
-END(memset)
+END(memset_std)
+
+ENTRY(memset_mops)
+	mov     x3, x0
+	.inst   0x19c10443	/* setp    [x3]!, x2!, x1  */
+	.inst   0x19c14443	/* setm    [x3]!, x2!, x1  */
+	.inst   0x19c18443	/* sete    [x3]!, x2!, x1  */
+	ret
+END(memset_mops)
 
 GNU_PROPERTY_AARCH64_FEATURE_1_NOTE(GNU_PROPERTY_AARCH64_FEATURE_1_VAL)
diff --git a/sys/arm64/include/cpu.h b/sys/arm64/include/cpu.h
--- a/sys/arm64/include/cpu.h
+++ b/sys/arm64/include/cpu.h
@@ -328,7 +328,12 @@
 ADDRESS_TRANSLATE_FUNC(s1e1w)
 
 #endif /* !__ASSEMBLER__ */
-#endif
+
+#define MEMSET_EARLY_FUNC	memset_std
+#define MEMCPY_EARLY_FUNC	memcpy_std
+#define MEMMOVE_EARLY_FUNC	memmove_std
+
+#endif /* _KERNEL */
 
 #endif /* !_MACHINE_CPU_H_ */