Use TMP1 and TMP2, so we don't need to pass

them as arguments.
(AES_FINAL_ROUND): New macro.

Rev: src/nettle/sparc/machine.m4:1.3

sparc/machine.m4

@@ -4,53 +4,89 @@ 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)
+C Used as temporaries by the AES macros
+define(<TMP1>, <%o0>)
+define(<TMP2>, <%o1>)
+
+C AES_LOAD(i, src, key, res)
 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
+	ldub	[$2 + eval(4*$1 + 1)], TMP1
+	ldub	[$2 + eval(4*$1 + 2)], TMP2
+	sll	TMP1, 8, TMP1
 	
-	or	$4, $5, $4	! U
-	ldub	[$2 + eval(4*$1+3)], $5
-	sll	$6, 16, $6
-	or	$4, $6, $4
+	or	$4, TMP1, $4
+	ldub	[$2 + eval(4*$1+3)], TMP1
+	sll	TMP2, 16, TMP2
+	or	$4, TMP2, $4
 	
-	sll	$5, 24, $5
-	! Get subkey
-	ld	[$3 + eval(4*$1)], $6
-	or	$4, $5, $4
-	xor	$4, $6, $4>)dnl
+	sll	TMP1, 24, TMP1
+	C	Get subkey
+	ld	[$3 + eval(4*$1)], TMP2
+	or	$4, TMP1, $4
+	xor	$4, TMP2, $4>)dnl
 
-C AES_ROUND(i, T, a, b, c, d, key, res, t)
+C AES_ROUND(i, T, a, b, c, d, key, res)
 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
+	and	$3, 0xff, TMP1
+	sll	TMP1, 2, TMP1
+	add	TMP1, AES_TABLE0, TMP1
+	ld	[$2 + TMP1], $8
+
+	srl	$4, 6, TMP1
+	and	TMP1, 0x3fc, TMP1
+	add	TMP1, AES_TABLE1, TMP1
+	ld	[$2 + TMP1], TMP1
+	xor	$8, TMP1, $8
+
+	srl	$5, 14, TMP1
+	and	TMP1, 0x3fc, TMP1
+	add	TMP1, AES_TABLE2, TMP1
+	ld	[$2 + TMP1], TMP1
+	xor	$8, TMP1, $8
+
+	srl	$4, 22, TMP1
+	and	TMP1, 0x3fc, TMP1
+	add	TMP1, AES_TABLE3, TMP1
+	ld	[$2 + TMP1], TMP1
+	xor	$8, TMP1, $8
+
+	ld	[$7 + eval(4*$1)], TMP1
+	xor	$8, TMP1, $8>)dnl
+
+C AES_FINAL_ROUND(i, T, a, b, c, d, key, dst)
+C Compute one word in the final round function. Output is converted to
+C octets and stored at dst. Relies on AES_SBOX being zero.
+define(<AES_FINAL_ROUND>, <
+	C	Load subkey
+	ld	[$7 + eval(4*$1)], TMP1
+
+	and	$3, 0xff, TMP2
+	ldub	[T + TMP2], TMP2
+	xor	TMP1, TMP2, TMP2
+	stb	[$8 + eval(4*i)]
+	
+	srl	$4, 8, TMP2
+	and	TMP2, 0xff, TMP2
+	ldub	[T + TMP2], TMP2
+	srl	TMP1, 8, TMP1
+	xor	TMP1, TMP2, TMP2
+	stb	[$8 + eval(4*i + 1)]
+
+	srl	$5, 16, TMP2
+	and	TMP2, 0xff, TMP2
+	ldub	[T + TMP2], TMP2
+	srl	TMP1, 8, TMP1
+	xor	TMP1, TMP2, TMP2
+	stb	[$8 + eval(4*i + 2)]
+
+	srl	$6, 24, TMP2
+	ldub	[T + TMP2], TMP2
+	srl	TMP1, 8, TMP1
+	xor	TMP1, TMP2, TMP2
+	stb	[$8 + eval(4*i + 1)]>)