diff --git a/sparc/machine.m4 b/sparc/machine.m4
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..7d22a2d5c5285fdafd02d07d4d626d83237c6f35 100644
--- a/sparc/machine.m4
+++ b/sparc/machine.m4
@@ -0,0 +1,56 @@
+C FIXME: Do we need an OFFSET macro? Or is it fine to use expressions such as [%i1 + 0]
+
+C FIXME: How much can we rely on the assembler to be able to
+C understand arithmetic expressions? Mayby we don't need to use m4
+C eval.
+
+C AES_LOAD(i, src, key, res, t1, t2)
+define(<AES_LOAD>, <
+	ldub	[$2 + eval(4*$1)], $4
+	ldub	[$2 + eval(4*$1 + 1)], $5
+	ldub	[$2 + eval(4*$1 + 2)], $6
+	sll	$5, 8, $5
+	
+	or	$4, $5, $4	! U
+	ldub	[$2 + eval(4*$1+3)], $5
+	sll	$6, 16, $6
+	or	$4, $6, $4
+	
+	sll	$5, 24, $5
+	! Get subkey
+	ld	[$3 + eval(4*$1)], $6
+	or	$4, $5, $4
+	xor	$4, $6, $4>)dnl
+
+C AES_ROUND(i, T, a, b, c, d, key, res, t)
+C Computes one word of the AES round
+C FIXME: Could use registers pointing directly to the four tables
+C FIXME: Needs better instruction scheduling, and perhaps more temporaries
+C Alternatively, we can use a single table and some rotations
+define(<AES_ROUND>, <
+	and	$3, 0xff, $9
+	sll	$9, 2, $9
+	add	$9, AES_TABLE0, $9
+	ld	[$2 + $9], $8
+
+	srl	$4, 6, $9
+	and	$9, 0x3fc, $9
+	add	$9, AES_TABLE1, $9
+	ld	[$2 + $9], $9
+	xor	$9, $8
+
+	srl	$5, 14, $9
+	and	$9, 0x3fc, $9
+	add	$9, AES_TABLE2, $9
+	ld	[$2 + $9], $9
+	xor	$9, $8
+
+	srl	$4, 22, $9
+	and	$9, 0x3fc, $9
+	add	$9, AES_TABLE3, $9
+	ld	[$2 + $9], $9
+	xor	$9, $8
+
+	ld	[$7 + eval(4*$1)], $9
+	xor	$9, $8>)dnl
+