Minimize aes_core.c footprint when AES_[en|de]crypt is implemented in

assembler.

1 files changed, 200 insertions, 1 deletions

diff --git a/crypto/aes/aes_core.c b/crypto/aes/aes_core.c
index 3a80e18b0a..a7ec54f4da 100644
--- a/crypto/aes/aes_core.c
+++ b/crypto/aes/aes_core.c
@@ -39,6 +39,7 @@
 #include <openssl/aes.h>
 #include "aes_locl.h"
 
+#ifndef AES_ASM
 /*
 Te0[x] = S [x].[02, 01, 01, 03];
 Te1[x] = S [x].[03, 02, 01, 01];
@@ -773,7 +774,6 @@ int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
 	return 0;
 }
 
-#ifndef AES_ASM
 /*
  * Encrypt a single block
  * in and out can overlap
@@ -1156,4 +1156,203 @@ void AES_decrypt(const unsigned char *in, unsigned char *out,
 	PUTU32(out + 12, s3);
 }
 
+#else /* AES_ASM */
+
+static const u8 Te4[256] = {
+    0x63U, 0x7cU, 0x77U, 0x7bU, 0xf2U, 0x6bU, 0x6fU, 0xc5U,
+    0x30U, 0x01U, 0x67U, 0x2bU, 0xfeU, 0xd7U, 0xabU, 0x76U,
+    0xcaU, 0x82U, 0xc9U, 0x7dU, 0xfaU, 0x59U, 0x47U, 0xf0U,
+    0xadU, 0xd4U, 0xa2U, 0xafU, 0x9cU, 0xa4U, 0x72U, 0xc0U,
+    0xb7U, 0xfdU, 0x93U, 0x26U, 0x36U, 0x3fU, 0xf7U, 0xccU,
+    0x34U, 0xa5U, 0xe5U, 0xf1U, 0x71U, 0xd8U, 0x31U, 0x15U,
+    0x04U, 0xc7U, 0x23U, 0xc3U, 0x18U, 0x96U, 0x05U, 0x9aU,
+    0x07U, 0x12U, 0x80U, 0xe2U, 0xebU, 0x27U, 0xb2U, 0x75U,
+    0x09U, 0x83U, 0x2cU, 0x1aU, 0x1bU, 0x6eU, 0x5aU, 0xa0U,
+    0x52U, 0x3bU, 0xd6U, 0xb3U, 0x29U, 0xe3U, 0x2fU, 0x84U,
+    0x53U, 0xd1U, 0x00U, 0xedU, 0x20U, 0xfcU, 0xb1U, 0x5bU,
+    0x6aU, 0xcbU, 0xbeU, 0x39U, 0x4aU, 0x4cU, 0x58U, 0xcfU,
+    0xd0U, 0xefU, 0xaaU, 0xfbU, 0x43U, 0x4dU, 0x33U, 0x85U,
+    0x45U, 0xf9U, 0x02U, 0x7fU, 0x50U, 0x3cU, 0x9fU, 0xa8U,
+    0x51U, 0xa3U, 0x40U, 0x8fU, 0x92U, 0x9dU, 0x38U, 0xf5U,
+    0xbcU, 0xb6U, 0xdaU, 0x21U, 0x10U, 0xffU, 0xf3U, 0xd2U,
+    0xcdU, 0x0cU, 0x13U, 0xecU, 0x5fU, 0x97U, 0x44U, 0x17U,
+    0xc4U, 0xa7U, 0x7eU, 0x3dU, 0x64U, 0x5dU, 0x19U, 0x73U,
+    0x60U, 0x81U, 0x4fU, 0xdcU, 0x22U, 0x2aU, 0x90U, 0x88U,
+    0x46U, 0xeeU, 0xb8U, 0x14U, 0xdeU, 0x5eU, 0x0bU, 0xdbU,
+    0xe0U, 0x32U, 0x3aU, 0x0aU, 0x49U, 0x06U, 0x24U, 0x5cU,
+    0xc2U, 0xd3U, 0xacU, 0x62U, 0x91U, 0x95U, 0xe4U, 0x79U,
+    0xe7U, 0xc8U, 0x37U, 0x6dU, 0x8dU, 0xd5U, 0x4eU, 0xa9U,
+    0x6cU, 0x56U, 0xf4U, 0xeaU, 0x65U, 0x7aU, 0xaeU, 0x08U,
+    0xbaU, 0x78U, 0x25U, 0x2eU, 0x1cU, 0xa6U, 0xb4U, 0xc6U,
+    0xe8U, 0xddU, 0x74U, 0x1fU, 0x4bU, 0xbdU, 0x8bU, 0x8aU,
+    0x70U, 0x3eU, 0xb5U, 0x66U, 0x48U, 0x03U, 0xf6U, 0x0eU,
+    0x61U, 0x35U, 0x57U, 0xb9U, 0x86U, 0xc1U, 0x1dU, 0x9eU,
+    0xe1U, 0xf8U, 0x98U, 0x11U, 0x69U, 0xd9U, 0x8eU, 0x94U,
+    0x9bU, 0x1eU, 0x87U, 0xe9U, 0xceU, 0x55U, 0x28U, 0xdfU,
+    0x8cU, 0xa1U, 0x89U, 0x0dU, 0xbfU, 0xe6U, 0x42U, 0x68U,
+    0x41U, 0x99U, 0x2dU, 0x0fU, 0xb0U, 0x54U, 0xbbU, 0x16U
+};
+static const u32 rcon[] = {
+	0x01000000, 0x02000000, 0x04000000, 0x08000000,
+	0x10000000, 0x20000000, 0x40000000, 0x80000000,
+	0x1B000000, 0x36000000, /* for 128-bit blocks, Rijndael never uses more than 10 rcon values */
+};
+
+/**
+ * Expand the cipher key into the encryption key schedule.
+ */
+int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
+			AES_KEY *key) {
+	u32 *rk;
+   	int i = 0;
+	u32 temp;
+
+	if (!userKey || !key)
+		return -1;
+	if (bits != 128 && bits != 192 && bits != 256)
+		return -2;
+
+	rk = key->rd_key;
+
+	if (bits==128)
+		key->rounds = 10;
+	else if (bits==192)
+		key->rounds = 12;
+	else
+		key->rounds = 14;
+
+	rk[0] = GETU32(userKey     );
+	rk[1] = GETU32(userKey +  4);
+	rk[2] = GETU32(userKey +  8);
+	rk[3] = GETU32(userKey + 12);
+	if (bits == 128) {
+		while (1) {
+			temp  = rk[3];
+			rk[4] = rk[0] ^
+				(Te4[(temp >> 16) & 0xff] << 24) ^
+				(Te4[(temp >>  8) & 0xff] << 16) ^
+				(Te4[(temp      ) & 0xff] << 8) ^
+				(Te4[(temp >> 24)       ]) ^
+				rcon[i];
+			rk[5] = rk[1] ^ rk[4];
+			rk[6] = rk[2] ^ rk[5];
+			rk[7] = rk[3] ^ rk[6];
+			if (++i == 10) {
+				return 0;
+			}
+			rk += 4;
+		}
+	}
+	rk[4] = GETU32(userKey + 16);
+	rk[5] = GETU32(userKey + 20);
+	if (bits == 192) {
+		while (1) {
+			temp = rk[ 5];
+			rk[ 6] = rk[ 0] ^
+				(Te4[(temp >> 16) & 0xff] << 24) ^
+				(Te4[(temp >>  8) & 0xff] << 16) ^
+				(Te4[(temp      ) & 0xff] << 8) ^
+				(Te4[(temp >> 24)       ]) ^
+				rcon[i];
+			rk[ 7] = rk[ 1] ^ rk[ 6];
+			rk[ 8] = rk[ 2] ^ rk[ 7];
+			rk[ 9] = rk[ 3] ^ rk[ 8];
+			if (++i == 8) {
+				return 0;
+			}
+			rk[10] = rk[ 4] ^ rk[ 9];
+			rk[11] = rk[ 5] ^ rk[10];
+			rk += 6;
+		}
+	}
+	rk[6] = GETU32(userKey + 24);
+	rk[7] = GETU32(userKey + 28);
+	if (bits == 256) {
+		while (1) {
+			temp = rk[ 7];
+			rk[ 8] = rk[ 0] ^
+				(Te4[(temp >> 16) & 0xff] << 24) ^
+				(Te4[(temp >>  8) & 0xff] << 16) ^
+				(Te4[(temp      ) & 0xff] << 8) ^
+				(Te4[(temp >> 24)       ]) ^
+				rcon[i];
+			rk[ 9] = rk[ 1] ^ rk[ 8];
+			rk[10] = rk[ 2] ^ rk[ 9];
+			rk[11] = rk[ 3] ^ rk[10];
+			if (++i == 7) {
+				return 0;
+			}
+			temp = rk[11];
+			rk[12] = rk[ 4] ^
+				(Te4[(temp >> 24)       ] << 24) ^
+				(Te4[(temp >> 16) & 0xff] << 16) ^
+				(Te4[(temp >>  8) & 0xff] << 8) ^
+				(Te4[(temp      ) & 0xff]);
+			rk[13] = rk[ 5] ^ rk[12];
+			rk[14] = rk[ 6] ^ rk[13];
+			rk[15] = rk[ 7] ^ rk[14];
+
+			rk += 8;
+        	}
+	}
+	return 0;
+}
+
+/**
+ * Expand the cipher key into the decryption key schedule.
+ */
+int AES_set_decrypt_key(const unsigned char *userKey, const int bits,
+			 AES_KEY *key) {
+
+        u32 *rk;
+	int i, j, status;
+	u32 temp;
+
+	/* first, start with an encryption schedule */
+	status = AES_set_encrypt_key(userKey, bits, key);
+	if (status < 0)
+		return status;
+
+	rk = key->rd_key;
+
+	/* invert the order of the round keys: */
+	for (i = 0, j = 4*(key->rounds); i < j; i += 4, j -= 4) {
+		temp = rk[i    ]; rk[i    ] = rk[j    ]; rk[j    ] = temp;
+		temp = rk[i + 1]; rk[i + 1] = rk[j + 1]; rk[j + 1] = temp;
+		temp = rk[i + 2]; rk[i + 2] = rk[j + 2]; rk[j + 2] = temp;
+		temp = rk[i + 3]; rk[i + 3] = rk[j + 3]; rk[j + 3] = temp;
+	}
+	/* apply the inverse MixColumn transform to all round keys but the first and the last: */
+	for (i = 1; i < (key->rounds); i++) {
+		rk += 4;
+		for (j = 0; j < 4; j++) {
+			u32 tp1, tp2, tp4, tp8, tp9, tpb, tpd, tpe, m;
+
+			tp1 = rk[j];
+			m = tp1 & 0x80808080;
+			tp2 = ((tp1 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp2 & 0x80808080;
+			tp4 = ((tp2 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			m = tp4 & 0x80808080;
+			tp8 = ((tp4 & 0x7f7f7f7f) << 1) ^
+				((m - (m >> 7)) & 0x1b1b1b1b);
+			tp9 = tp8 ^ tp1;
+			tpb = tp9 ^ tp2;
+			tpd = tp9 ^ tp4;
+			tpe = tp8 ^ tp4 ^ tp2;
+#if defined(ROTATE)
+			rk[j] = tpe ^ ROTATE(tpd,16) ^
+				ROTATE(tp9,24) ^ ROTATE(tpb,8);
+#else
+			rk[j] = tpe ^ (tpd >> 16) ^ (tpd << 16) ^ 
+				(tp9 >> 8) ^ (tp9 << 24) ^
+				(tpb >> 24) ^ (tpb << 8);
+#endif
+		}
+	}
+	return 0;
+}
+
 #endif /* AES_ASM */