diff --git a/lib/libc/string/arch/arm/memcpy.S b/lib/libc/string/arch/arm/memcpy.S
new file mode 100644
index 0000000..3b7816d
--- /dev/null
+++ b/lib/libc/string/arch/arm/memcpy.S
@@ -0,0 +1,177 @@
+/*
+ * Copyright (c) 2008 Travis Geiselbrecht
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files
+ * (the "Software"), to deal in the Software without restriction,
+ * including without limitation the rights to use, copy, modify, merge,
+ * publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so,
+ * subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+ * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include <asm.h>
+#include <arch/arm/cores.h>
+
+.text
+.align 2
+
+/* void bcopy(const void *src, void *dest, size_t n); */
+FUNCTION(bcopy)
+	// swap args for bcopy
+	mov	r12, r0
+	mov	r0, r1
+	mov r1, r12
+
+/* void *memcpy(void *dest, const void *src, size_t n); */
+FUNCTION(memmove)
+FUNCTION(memcpy)
+	// check for zero length copy or the same pointer
+	cmp		r2, #0
+	cmpne	r1, r0
+	bxeq	lr
+
+	// save a few registers for use and the return code (input dst)
+	stmfd	sp!, {r0, r4, r5, lr}
+
+	// check for forwards overlap (src > dst, distance < len)
+	subs	r3, r0, r1
+	cmpgt	r2, r3
+	bgt		.L_forwardoverlap
+
+	// check for a short copy len.
+	// 20 bytes is enough so that if a 16 byte alignment needs to happen there is at least a 
+	//   wordwise copy worth of work to be done.
+	cmp		r2, #(16+4)
+	blt		.L_bytewise
+
+	// see if they are similarly aligned on 4 byte boundaries
+	eor		r3, r0, r1
+	tst		r3, #3
+	bne		.L_bytewise		// dissimilarly aligned, nothing we can do (for now)
+
+	// check for 16 byte alignment on dst.
+	// this will also catch src being not 4 byte aligned, since it is similarly 4 byte 
+	//   aligned with dst at this point.
+	tst		r0, #15
+	bne		.L_not16bytealigned
+
+	// check to see if we have at least 32 bytes of data to copy.
+	// if not, just revert to wordwise copy
+	cmp		r2, #32
+	blt		.L_wordwise
+
+.L_bigcopy:
+	// copy 32 bytes at a time. src & dst need to be at least 4 byte aligned, 
+	// and we need at least 32 bytes remaining to copy
+
+	// save r6-r7 for use in the big copy
+	stmfd	sp!, {r6-r7}
+
+	sub		r2, r2, #32		// subtract an extra 32 to the len so we can avoid an extra compare
+
+.L_bigcopy_loop:
+	ldmia	r1!, {r4, r5, r6, r7}
+	stmia	r0!, {r4, r5, r6, r7}
+	ldmia	r1!, {r4, r5, r6, r7}
+	subs	r2, r2, #32
+	stmia	r0!, {r4, r5, r6, r7}
+	bge		.L_bigcopy_loop
+
+	// restore r6-r7
+	ldmfd	sp!, {r6-r7}
+
+	// see if we are done
+	adds	r2, r2, #32
+	beq		.L_done
+
+	// less then 4 bytes left?
+	cmp		r2, #4
+	blt		.L_bytewise
+
+.L_wordwise:
+	// copy 4 bytes at a time.
+	// src & dst are guaranteed to be word aligned, and at least 4 bytes are left to copy.
+	subs	r2, r2, #4
+
+.L_wordwise_loop:
+	ldr		r3, [r1], #4
+	subs	r2, r2, #4
+	str		r3, [r0], #4
+	bge		.L_wordwise_loop
+
+	// correct the remaining len and test for completion
+	adds	r2, r2, #4	
+	beq		.L_done
+
+.L_bytewise:
+	// simple bytewise copy
+	ldrb	r3, [r1], #1
+	subs	r2, r2, #1
+	strb	r3, [r0], #1
+	bgt		.L_bytewise
+
+.L_done:
+	// load dst for return and restore r4,r5
+#if ARM_ARCH_LEVEL >= 5
+	ldmfd	sp!, {r0, r4, r5, pc}
+#else
+	ldmfd	sp!, {r0, r4, r5, lr}
+	bx		lr
+#endif
+
+.L_not16bytealigned:
+	// dst is not 16 byte aligned, so we will copy up to 15 bytes to get it aligned.
+	// src is guaranteed to be similarly word aligned with dst.
+
+	// set the condition flags based on the alignment.
+	lsl		r12, r0, #28
+	rsb		r12, r12, #0
+	msr		CPSR_f, r12				// move into NZCV fields in CPSR
+
+	// move as many bytes as necessary to get the dst aligned
+	ldrvsb	r3, [r1], #1			// V set
+	ldrcsh	r4, [r1], #2			// C set
+	ldreq	r5, [r1], #4			// Z set
+
+	strvsb	r3, [r0], #1
+	strcsh	r4, [r0], #2
+	streq	r5, [r0], #4
+
+	ldmmiia	r1!, {r3-r4}			// N set
+	stmmiia	r0!, {r3-r4}
+
+	// fix the remaining len
+	sub		r2, r2, r12, lsr #28
+
+	// test to see what we should do now
+	cmp		r2, #32
+	bge		.L_bigcopy
+	b		.L_wordwise
+	
+	// src and dest overlap 'forwards' or dst > src
+.L_forwardoverlap:
+
+	// do a bytewise reverse copy for now
+	add		r1, r1, r2
+	add		r0, r0, r2
+
+.L_bytewisereverse:
+	// simple bytewise reverse copy
+	ldrb	r3, [r1], #-1
+	subs	r2, r2, #1
+	strb	r3, [r0], #-1
+	bgt		.L_bytewisereverse
+
+	b		.L_done
+
diff --git a/lib/libc/string/arch/arm/memset.S b/lib/libc/string/arch/arm/memset.S
new file mode 100644
index 0000000..b31d053
--- /dev/null
+++ b/lib/libc/string/arch/arm/memset.S
@@ -0,0 +1,113 @@
+/*
+ * Copyright (c) 2008 Travis Geiselbrecht
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files
+ * (the "Software"), to deal in the Software without restriction,
+ * including without limitation the rights to use, copy, modify, merge,
+ * publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so,
+ * subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+ * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include <asm.h>
+#include <arch/arm/cores.h>
+
+.text
+.align 2
+
+/* void bzero(void *s, size_t n); */
+FUNCTION(bzero)
+	mov		r2, r1
+	mov		r1, #0
+
+/* void *memset(void *s, int c, size_t n); */
+FUNCTION(memset)
+	// check for zero length
+	cmp		r2, #0
+	bxeq	lr
+
+	// save the original pointer
+	mov		r12, r0
+
+	// short memsets aren't worth optimizing
+	cmp		r2, #(32 + 16)
+	blt		.L_bytewise
+
+	// fill a 32 bit register with the 8 bit value
+	and		r1, r1, #0xff
+	orr		r1, r1, r1, lsl #8
+	orr		r1, r1, r1, lsl #16
+
+	// check for 16 byte alignment
+	tst		r0, #15
+	bne		.L_not16bytealigned
+
+.L_bigset:
+	// dump some registers to make space for our values
+	stmfd	sp!, { r4-r5 }
+	
+	// fill a bunch of registers with the set value
+	mov		r3, r1
+	mov		r4, r1
+	mov		r5, r1
+
+	// prepare the count register so we can avoid an extra compare
+	sub 	r2, r2, #32
+
+	// 32 bytes at a time
+.L_bigset_loop:
+	stmia	r0!, { r1, r3, r4, r5 }
+	subs	r2, r2, #32
+	stmia	r0!, { r1, r3, r4, r5 }
+	bge		.L_bigset_loop
+
+	// restore our dumped registers
+	ldmfd	sp!, { r4-r5 }
+
+	// see if we're done
+	adds	r2, r2, #32
+	beq		.L_done
+
+.L_bytewise:
+	// bytewise memset
+	subs	r2, r2, #1
+	strb	r1, [r0], #1
+	bgt		.L_bytewise
+
+.L_done:
+	// restore the base pointer as return value
+	mov		r0, r12
+	bx		lr
+
+.L_not16bytealigned:
+	// dst is not 16 byte aligned, so we will set up to 15 bytes to get it aligned.
+
+	// set the condition flags based on the alignment.
+	lsl     r3, r0, #28
+	rsb     r3, r3, #0
+	msr     CPSR_f, r3             // move into NZCV fields in CPSR
+
+	// move as many bytes as necessary to get the dst aligned
+	strvsb  r1, [r0], #1			// V set
+	strcsh  r1, [r0], #2			// C set
+	streq   r1, [r0], #4			// Z set
+	strmi   r1, [r0], #4			// N set
+	strmi   r1, [r0], #4			// N set
+
+	// fix the remaining len
+	sub     r2, r2, r3, lsr #28
+
+	// do the large memset
+	b       .L_bigset
+
diff --git a/lib/libc/string/arch/arm/rules.mk b/lib/libc/string/arch/arm/rules.mk
new file mode 100644
index 0000000..89a68f4
--- /dev/null
+++ b/lib/libc/string/arch/arm/rules.mk
@@ -0,0 +1,11 @@
+LOCAL_DIR := $(GET_LOCAL_DIR)
+
+ASM_STRING_OPS := bcopy bzero memcpy memmove memset
+
+OBJS += \
+	$(LOCAL_DIR)/memcpy.o \
+	$(LOCAL_DIR)/memset.o
+
+# filter out the C implementation
+C_STRING_OPS := $(filter-out $(ASM_STRING_OPS),$(C_STRING_OPS))
+