diff --git a/lib/libc/aarch32/memset.S b/lib/libc/aarch32/memset.S
new file mode 100644
index 0000000..880ba83
--- /dev/null
+++ b/lib/libc/aarch32/memset.S
@@ -0,0 +1,74 @@
+/*
+ * Copyright (c) 2020, Arm Limited. All rights reserved.
+ *
+ * SPDX-License-Identifier: BSD-3-Clause
+ */
+
+#include <asm_macros.S>
+
+	.syntax unified
+	.global	memset
+
+/* -----------------------------------------------------------------------
+ * void *memset(void *dst, int val, size_t count)
+ *
+ * Copy the value of 'val' (converted to an unsigned char) into
+ * each of the first 'count' characters of the object pointed to by 'dst'.
+ *
+ * Returns the value of 'dst'.
+ * -----------------------------------------------------------------------
+ */
+func memset
+	mov	r12, r0			/* keep r0 */
+	tst	r0, #3
+	beq	aligned			/* 4-bytes aligned */
+
+	/* Unaligned 'dst' */
+unaligned:
+	subs	r2, r2, #1
+	strbhs	r1, [r12], #1
+	bxls	lr			/* return if 0 */
+	tst	r12, #3
+	bne	unaligned		/* continue while unaligned */
+
+	/* 4-bytes aligned */
+aligned:bfi	r1, r1, #8, #8		/* propagate 'val' */
+	bfi	r1, r1, #16, #16
+
+	mov	r3, r1
+
+	cmp	r2, #16
+	blo	less_16			/* < 16 */
+
+	push	{r4, lr}
+	mov	r4, r1
+	mov	lr, r1
+
+write_32:
+	subs	r2, r2, #32
+	stmiahs	r12!, {r1, r3, r4, lr}
+	stmiahs	r12!, {r1, r3, r4, lr}
+	bhi	write_32		/* write 32 bytes in a loop */
+	popeq	{r4, pc}		/* return if 0 */
+	lsls	r2, r2, #28		/* C = r2[4]; N = r2[3]; Z = r2[3:0] */
+	stmiacs	r12!, {r1, r3, r4, lr}	/* write 16 bytes */
+	popeq	{r4, pc}		/* return if 16 */
+	stmiami	r12!, {r1, r3}		/* write 8 bytes */
+	lsls	r2, r2, #2		/* C = r2[2]; N = r2[1]; Z = r2[1:0] */
+	strcs	r1, [r12], #4		/* write 4 bytes */
+	popeq	{r4, pc}		/* return if 8 or 4 */
+	strhmi	r1, [r12], #2		/* write 2 bytes */
+	lsls	r2, r2, #1		/* N = Z = r2[0] */
+	strbmi	r1, [r12]		/* write 1 byte */
+	pop	{r4, pc}
+
+less_16:lsls	r2, r2, #29		/* C = r2[3]; N = r2[2]; Z = r2[2:0] */
+	stmiacs	r12!, {r1, r3}		/* write 8 bytes */
+	bxeq	lr			/* return if 8 */
+	strmi	r1, [r12], #4		/* write 4 bytes */
+	lsls	r2, r2, #2		/* C = r2[1]; N = Z = r2[0] */
+	strhcs	r1, [r12], #2		/* write 2 bytes */
+	strbmi	r1, [r12]		/* write 1 byte */
+	bx	lr
+
+endfunc memset