[CRYPTO]: Add serpent algorithm.

crypto/Kconfig

@@ -82,6 +82,18 @@ config CRYPTO_TWOFISH
 	  See also:
 	  http://www.counterpane.com/twofish.html
 
+config CRYPTO_SERPENT
+	tristate "Serpent cipher algorithm"
+	depends on CRYPTO
+	help
+	  Serpent cipher algorithm, by Anderson, Biham & Knudsen.
+
+	  Keys are allowed to be from 0 to 256 bits in length, in steps
+	  of 8 bits.
+
+	  See also:
+	  http://www.cl.cam.ac.uk/~rja14/serpent.html
+
 config CRYPTO_TEST
 	tristate "Testing module"
 	depends on CRYPTO

crypto/Makefile

@@ -17,6 +17,7 @@ obj-$(CONFIG_CRYPTO_SHA256) += sha256.o
 obj-$(CONFIG_CRYPTO_DES) += des.o
 obj-$(CONFIG_CRYPTO_BLOWFISH) += blowfish.o
 obj-$(CONFIG_CRYPTO_TWOFISH) += twofish.o
+obj-$(CONFIG_CRYPTO_SERPENT) += serpent.o
 
 obj-$(CONFIG_CRYPTO_TEST) += tcrypt.o
 

crypto/serpent.c

+/*
+ * Cryptographic API.
+ *
+ * Serpent Cipher Algorithm.
+ *
+ * Copyright (C) 2002 Dag Arne Osvik <osvik@ii.uib.no>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <asm/byteorder.h>
+#include <linux/crypto.h>
+
+/* Key is padded to the maximum of 256 bits before round key generation.
+ * Any key length <= 256 bits (32 bytes) is allowed by the algorithm.
+ */
+
+#define SERPENT_MIN_KEY_SIZE		  0
+#define SERPENT_MAX_KEY_SIZE		 32
+#define SERPENT_EXPKEY_WORDS		132
+#define SERPENT_BLOCK_SIZE		 16
+
+#define PHI 0x9e3779b9UL
+#define ROL(x,r) ((x) = ((x) << (r)) | ((x) >> (32-(r))))
+#define ROR(x,r) ((x) = ((x) >> (r)) | ((x) << (32-(r))))
+
+#define keyiter(a,b,c,d,i,j) \
+        b ^= d; b ^= c; b ^= a; b ^= PHI ^ i; ROL(b,11); k[j] = b;
+
+#define loadkeys(x0,x1,x2,x3,i) \
+	x0=k[i]; x1=k[i+1]; x2=k[i+2]; x3=k[i+3];
+
+#define storekeys(x0,x1,x2,x3,i) \
+	k[i]=x0; k[i+1]=x1; k[i+2]=x2; k[i+3]=x3;
+
+#define K(x0,x1,x2,x3,i)				\
+	x3 ^= k[4*(i)+3];        x2 ^= k[4*(i)+2];	\
+	x1 ^= k[4*(i)+1];        x0 ^= k[4*(i)+0];
+
+#define LK(x0,x1,x2,x3,x4,i)				\
+					ROL(x0,13);	\
+	ROL(x2,3);	x1 ^= x0;	x4  = x0 << 3;	\
+	x3 ^= x2;	x1 ^= x2;			\
+	ROL(x1,1);	x3 ^= x4;			\
+	ROL(x3,7);	x4  = x1;			\
+	x0 ^= x1;	x4 <<= 7;	x2 ^= x3;	\
+	x0 ^= x3;	x2 ^= x4;	x3 ^= k[4*i+3];	\
+	x1 ^= k[4*i+1];	ROL(x0,5);	ROL(x2,22);	\
+	x0 ^= k[4*i+0];	x2 ^= k[4*i+2];
+
+#define KL(x0,x1,x2,x3,x4,i)				\
+	x0 ^= k[4*i+0];	x1 ^= k[4*i+1];	x2 ^= k[4*i+2];	\
+	x3 ^= k[4*i+3];	ROR(x0,5);	ROR(x2,22);	\
+	x4 =  x1;	x2 ^= x3;	x0 ^= x3;	\
+	x4 <<= 7;	x0 ^= x1;	ROR(x1,1);	\
+	x2 ^= x4;	ROR(x3,7);	x4 = x0 << 3;	\
+	x1 ^= x0;	x3 ^= x4;	ROR(x0,13);	\
+	x1 ^= x2;	x3 ^= x2;	ROR(x2,3);
+
+#define S0(x0,x1,x2,x3,x4)				\
+					x4  = x3;	\
+	x3 |= x0;	x0 ^= x4;	x4 ^= x2;	\
+	x4 =~ x4;	x3 ^= x1;	x1 &= x0;	\
+	x1 ^= x4;	x2 ^= x0;	x0 ^= x3;	\
+	x4 |= x0;	x0 ^= x2;	x2 &= x1;	\
+	x3 ^= x2;	x1 =~ x1;	x2 ^= x4;	\
+	x1 ^= x2;
+
+#define S1(x0,x1,x2,x3,x4)				\
+					x4  = x1;	\
+	x1 ^= x0;	x0 ^= x3;	x3 =~ x3;	\
+	x4 &= x1;	x0 |= x1;	x3 ^= x2;	\
+	x0 ^= x3;	x1 ^= x3;	x3 ^= x4;	\
+	x1 |= x4;	x4 ^= x2;	x2 &= x0;	\
+	x2 ^= x1;	x1 |= x0;	x0 =~ x0;	\
+	x0 ^= x2;	x4 ^= x1;
+
+#define S2(x0,x1,x2,x3,x4)				\
+					x3 =~ x3;	\
+	x1 ^= x0;	x4  = x0;	x0 &= x2;	\
+	x0 ^= x3;	x3 |= x4;	x2 ^= x1;	\
+	x3 ^= x1;	x1 &= x0;	x0 ^= x2;	\
+	x2 &= x3;	x3 |= x1;	x0 =~ x0;	\
+	x3 ^= x0;	x4 ^= x0;	x0 ^= x2;	\
+	x1 |= x2;
+
+#define S3(x0,x1,x2,x3,x4)				\
+					x4  = x1;	\
+	x1 ^= x3;	x3 |= x0;	x4 &= x0;	\
+	x0 ^= x2;	x2 ^= x1;	x1 &= x3;	\
+	x2 ^= x3;	x0 |= x4;	x4 ^= x3;	\
+	x1 ^= x0;	x0 &= x3;	x3 &= x4;	\
+	x3 ^= x2;	x4 |= x1;	x2 &= x1;	\
+	x4 ^= x3;	x0 ^= x3;	x3 ^= x2;
+
+#define S4(x0,x1,x2,x3,x4)				\
+					x4  = x3;	\
+	x3 &= x0;	x0 ^= x4;			\
+	x3 ^= x2;	x2 |= x4;	x0 ^= x1;	\
+	x4 ^= x3;	x2 |= x0;			\
+	x2 ^= x1;	x1 &= x0;			\
+	x1 ^= x4;	x4 &= x2;	x2 ^= x3;	\
+	x4 ^= x0;	x3 |= x1;	x1 =~ x1;	\
+	x3 ^= x0;
+
+#define S5(x0,x1,x2,x3,x4)				\
+	x4  = x1;	x1 |= x0;			\
+	x2 ^= x1;	x3 =~ x3;	x4 ^= x0;	\
+	x0 ^= x2;	x1 &= x4;	x4 |= x3;	\
+	x4 ^= x0;	x0 &= x3;	x1 ^= x3;	\
+	x3 ^= x2;	x0 ^= x1;	x2 &= x4;	\
+	x1 ^= x2;	x2 &= x0;			\
+	x3 ^= x2;
+
+#define S6(x0,x1,x2,x3,x4)				\
+					x4  = x1;	\
+	x3 ^= x0;	x1 ^= x2;	x2 ^= x0;	\
+	x0 &= x3;	x1 |= x3;	x4 =~ x4;	\
+	x0 ^= x1;	x1 ^= x2;			\
+	x3 ^= x4;	x4 ^= x0;	x2 &= x0;	\
+	x4 ^= x1;	x2 ^= x3;	x3 &= x1;	\
+	x3 ^= x0;	x1 ^= x2;
+
+#define S7(x0,x1,x2,x3,x4)				\
+					x1 =~ x1;	\
+	x4  = x1;	x0 =~ x0;	x1 &= x2;	\
+	x1 ^= x3;	x3 |= x4;	x4 ^= x2;	\
+	x2 ^= x3;	x3 ^= x0;	x0 |= x1;	\
+	x2 &= x0;	x0 ^= x4;	x4 ^= x3;	\
+	x3 &= x0;	x4 ^= x1;			\
+	x2 ^= x4;	x3 ^= x1;	x4 |= x0;	\
+	x4 ^= x1;
+
+#define SI0(x0,x1,x2,x3,x4)				\
+			x4  = x3;	x1 ^= x0;	\
+	x3 |= x1;	x4 ^= x1;	x0 =~ x0;	\
+	x2 ^= x3;	x3 ^= x0;	x0 &= x1;	\
+	x0 ^= x2;	x2 &= x3;	x3 ^= x4;	\
+	x2 ^= x3;	x1 ^= x3;	x3 &= x0;	\
+	x1 ^= x0;	x0 ^= x2;	x4 ^= x3;
+
+#define SI1(x0,x1,x2,x3,x4)				\
+	x1 ^= x3;	x4  = x0;			\
+	x0 ^= x2;	x2 =~ x2;	x4 |= x1;	\
+	x4 ^= x3;	x3 &= x1;	x1 ^= x2;	\
+	x2 &= x4;	x4 ^= x1;	x1 |= x3;	\
+	x3 ^= x0;	x2 ^= x0;	x0 |= x4;	\
+	x2 ^= x4;	x1 ^= x0;			\
+	x4 ^= x1;
+
+#define SI2(x0,x1,x2,x3,x4)				\
+	x2 ^= x1;	x4  = x3;	x3 =~ x3;	\
+	x3 |= x2;	x2 ^= x4;	x4 ^= x0;	\
+	x3 ^= x1;	x1 |= x2;	x2 ^= x0;	\
+	x1 ^= x4;	x4 |= x3;	x2 ^= x3;	\
+	x4 ^= x2;	x2 &= x1;			\
+	x2 ^= x3;	x3 ^= x4;	x4 ^= x0;
+
+#define SI3(x0,x1,x2,x3,x4)				\
+					x2 ^= x1;	\
+	x4  = x1;	x1 &= x2;			\
+	x1 ^= x0;	x0 |= x4;	x4 ^= x3;	\
+	x0 ^= x3;	x3 |= x1;	x1 ^= x2;	\
+	x1 ^= x3;	x0 ^= x2;	x2 ^= x3;	\
+	x3 &= x1;	x1 ^= x0;	x0 &= x2;	\
+	x4 ^= x3;	x3 ^= x0;	x0 ^= x1;
+
+#define SI4(x0,x1,x2,x3,x4)				\
+	x2 ^= x3;	x4  = x0;	x0 &= x1;	\
+	x0 ^= x2;	x2 |= x3;	x4 =~ x4;	\
+	x1 ^= x0;	x0 ^= x2;	x2 &= x4;	\
+	x2 ^= x0;	x0 |= x4;			\
+	x0 ^= x3;	x3 &= x2;			\
+	x4 ^= x3;	x3 ^= x1;	x1 &= x0;	\
+	x4 ^= x1;	x0 ^= x3;
+
+#define SI5(x0,x1,x2,x3,x4)				\
+			x4  = x1;	x1 |= x2;	\
+	x2 ^= x4;	x1 ^= x3;	x3 &= x4;	\
+	x2 ^= x3;	x3 |= x0;	x0 =~ x0;	\
+	x3 ^= x2;	x2 |= x0;	x4 ^= x1;	\
+	x2 ^= x4;	x4 &= x0;	x0 ^= x1;	\
+	x1 ^= x3;	x0 &= x2;	x2 ^= x3;	\
+	x0 ^= x2;	x2 ^= x4;	x4 ^= x3;
+
+#define SI6(x0,x1,x2,x3,x4)				\
+			x0 ^= x2;			\
+	x4  = x0;	x0 &= x3;	x2 ^= x3;	\
+	x0 ^= x2;	x3 ^= x1;	x2 |= x4;	\
+	x2 ^= x3;	x3 &= x0;	x0 =~ x0;	\
+	x3 ^= x1;	x1 &= x2;	x4 ^= x0;	\
+	x3 ^= x4;	x4 ^= x2;	x0 ^= x1;	\
+	x2 ^= x0;
+
+#define SI7(x0,x1,x2,x3,x4)				\
+	x4  = x3;	x3 &= x0;	x0 ^= x2;	\
+	x2 |= x4;	x4 ^= x1;	x0 =~ x0;	\
+	x1 |= x3;	x4 ^= x0;	x0 &= x2;	\
+	x0 ^= x1;	x1 &= x2;	x3 ^= x2;	\
+	x4 ^= x3;	x2 &= x3;	x3 |= x0;	\
+	x1 ^= x4;	x3 ^= x4;	x4 &= x0;	\
+	x4 ^= x2;
+
+struct serpent_ctx {
+	u8 iv[SERPENT_BLOCK_SIZE];
+	u32 expkey[SERPENT_EXPKEY_WORDS];
+};
+
+static int setkey(void *ctx, const u8 *key, unsigned int keylen, u32 *flags)
+{
+	u32 *k = ((struct serpent_ctx *)ctx)->expkey;
+	u8  *k8 = (u8 *)k;
+	u32 r0,r1,r2,r3,r4;
+	int i;
+
+	if ((keylen < SERPENT_MIN_KEY_SIZE)
+			|| (keylen > SERPENT_MAX_KEY_SIZE))
+	{
+		*flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+
+	/* Copy key, add padding */
+
+	for (i = 0; i < keylen; ++i)
+		k8[i] = key[i];
+	if (i < SERPENT_MAX_KEY_SIZE)
+		k8[i++] = 1;
+	while (i < SERPENT_MAX_KEY_SIZE)
+		k8[i++] = 0;
+
+	/* Expand key using polynomial */
+
+	r0 = le32_to_cpu(k[3]);
+	r1 = le32_to_cpu(k[4]);
+	r2 = le32_to_cpu(k[5]);
+	r3 = le32_to_cpu(k[6]);
+	r4 = le32_to_cpu(k[7]);
+
+	keyiter(le32_to_cpu(k[0]),r0,r4,r2,0,0);
+	keyiter(le32_to_cpu(k[1]),r1,r0,r3,1,1);
+	keyiter(le32_to_cpu(k[2]),r2,r1,r4,2,2);
+	keyiter(le32_to_cpu(k[3]),r3,r2,r0,3,3);
+	keyiter(le32_to_cpu(k[4]),r4,r3,r1,4,4);
+	keyiter(le32_to_cpu(k[5]),r0,r4,r2,5,5);
+	keyiter(le32_to_cpu(k[6]),r1,r0,r3,6,6);
+	keyiter(le32_to_cpu(k[7]),r2,r1,r4,7,7);
+
+	keyiter(k[  0],r3,r2,r0,  8,  8); keyiter(k[  1],r4,r3,r1,  9,  9);
+	keyiter(k[  2],r0,r4,r2, 10, 10); keyiter(k[  3],r1,r0,r3, 11, 11);
+	keyiter(k[  4],r2,r1,r4, 12, 12); keyiter(k[  5],r3,r2,r0, 13, 13);
+	keyiter(k[  6],r4,r3,r1, 14, 14); keyiter(k[  7],r0,r4,r2, 15, 15);
+	keyiter(k[  8],r1,r0,r3, 16, 16); keyiter(k[  9],r2,r1,r4, 17, 17);
+	keyiter(k[ 10],r3,r2,r0, 18, 18); keyiter(k[ 11],r4,r3,r1, 19, 19);
+	keyiter(k[ 12],r0,r4,r2, 20, 20); keyiter(k[ 13],r1,r0,r3, 21, 21);
+	keyiter(k[ 14],r2,r1,r4, 22, 22); keyiter(k[ 15],r3,r2,r0, 23, 23);
+	keyiter(k[ 16],r4,r3,r1, 24, 24); keyiter(k[ 17],r0,r4,r2, 25, 25);
+	keyiter(k[ 18],r1,r0,r3, 26, 26); keyiter(k[ 19],r2,r1,r4, 27, 27);
+	keyiter(k[ 20],r3,r2,r0, 28, 28); keyiter(k[ 21],r4,r3,r1, 29, 29);
+	keyiter(k[ 22],r0,r4,r2, 30, 30); keyiter(k[ 23],r1,r0,r3, 31, 31);
+
+	k += 50;
+
+	keyiter(k[-26],r2,r1,r4, 32,-18); keyiter(k[-25],r3,r2,r0, 33,-17);
+	keyiter(k[-24],r4,r3,r1, 34,-16); keyiter(k[-23],r0,r4,r2, 35,-15);
+	keyiter(k[-22],r1,r0,r3, 36,-14); keyiter(k[-21],r2,r1,r4, 37,-13);
+	keyiter(k[-20],r3,r2,r0, 38,-12); keyiter(k[-19],r4,r3,r1, 39,-11);
+	keyiter(k[-18],r0,r4,r2, 40,-10); keyiter(k[-17],r1,r0,r3, 41, -9);
+	keyiter(k[-16],r2,r1,r4, 42, -8); keyiter(k[-15],r3,r2,r0, 43, -7);
+	keyiter(k[-14],r4,r3,r1, 44, -6); keyiter(k[-13],r0,r4,r2, 45, -5);
+	keyiter(k[-12],r1,r0,r3, 46, -4); keyiter(k[-11],r2,r1,r4, 47, -3);
+	keyiter(k[-10],r3,r2,r0, 48, -2); keyiter(k[ -9],r4,r3,r1, 49, -1);
+	keyiter(k[ -8],r0,r4,r2, 50,  0); keyiter(k[ -7],r1,r0,r3, 51,  1);
+	keyiter(k[ -6],r2,r1,r4, 52,  2); keyiter(k[ -5],r3,r2,r0, 53,  3);
+	keyiter(k[ -4],r4,r3,r1, 54,  4); keyiter(k[ -3],r0,r4,r2, 55,  5);
+	keyiter(k[ -2],r1,r0,r3, 56,  6); keyiter(k[ -1],r2,r1,r4, 57,  7);
+	keyiter(k[  0],r3,r2,r0, 58,  8); keyiter(k[  1],r4,r3,r1, 59,  9);
+	keyiter(k[  2],r0,r4,r2, 60, 10); keyiter(k[  3],r1,r0,r3, 61, 11);
+	keyiter(k[  4],r2,r1,r4, 62, 12); keyiter(k[  5],r3,r2,r0, 63, 13);
+	keyiter(k[  6],r4,r3,r1, 64, 14); keyiter(k[  7],r0,r4,r2, 65, 15);
+	keyiter(k[  8],r1,r0,r3, 66, 16); keyiter(k[  9],r2,r1,r4, 67, 17);
+	keyiter(k[ 10],r3,r2,r0, 68, 18); keyiter(k[ 11],r4,r3,r1, 69, 19);
+	keyiter(k[ 12],r0,r4,r2, 70, 20); keyiter(k[ 13],r1,r0,r3, 71, 21);
+	keyiter(k[ 14],r2,r1,r4, 72, 22); keyiter(k[ 15],r3,r2,r0, 73, 23);
+	keyiter(k[ 16],r4,r3,r1, 74, 24); keyiter(k[ 17],r0,r4,r2, 75, 25);
+	keyiter(k[ 18],r1,r0,r3, 76, 26); keyiter(k[ 19],r2,r1,r4, 77, 27);
+	keyiter(k[ 20],r3,r2,r0, 78, 28); keyiter(k[ 21],r4,r3,r1, 79, 29);
+	keyiter(k[ 22],r0,r4,r2, 80, 30); keyiter(k[ 23],r1,r0,r3, 81, 31);
+
+	k += 50;
+
+	keyiter(k[-26],r2,r1,r4, 82,-18); keyiter(k[-25],r3,r2,r0, 83,-17);
+	keyiter(k[-24],r4,r3,r1, 84,-16); keyiter(k[-23],r0,r4,r2, 85,-15);
+	keyiter(k[-22],r1,r0,r3, 86,-14); keyiter(k[-21],r2,r1,r4, 87,-13);
+	keyiter(k[-20],r3,r2,r0, 88,-12); keyiter(k[-19],r4,r3,r1, 89,-11);
+	keyiter(k[-18],r0,r4,r2, 90,-10); keyiter(k[-17],r1,r0,r3, 91, -9);
+	keyiter(k[-16],r2,r1,r4, 92, -8); keyiter(k[-15],r3,r2,r0, 93, -7);
+	keyiter(k[-14],r4,r3,r1, 94, -6); keyiter(k[-13],r0,r4,r2, 95, -5);
+	keyiter(k[-12],r1,r0,r3, 96, -4); keyiter(k[-11],r2,r1,r4, 97, -3);
+	keyiter(k[-10],r3,r2,r0, 98, -2); keyiter(k[ -9],r4,r3,r1, 99, -1);
+	keyiter(k[ -8],r0,r4,r2,100,  0); keyiter(k[ -7],r1,r0,r3,101,  1);
+	keyiter(k[ -6],r2,r1,r4,102,  2); keyiter(k[ -5],r3,r2,r0,103,  3);
+	keyiter(k[ -4],r4,r3,r1,104,  4); keyiter(k[ -3],r0,r4,r2,105,  5);
+	keyiter(k[ -2],r1,r0,r3,106,  6); keyiter(k[ -1],r2,r1,r4,107,  7);
+	keyiter(k[  0],r3,r2,r0,108,  8); keyiter(k[  1],r4,r3,r1,109,  9);
+	keyiter(k[  2],r0,r4,r2,110, 10); keyiter(k[  3],r1,r0,r3,111, 11);
+	keyiter(k[  4],r2,r1,r4,112, 12); keyiter(k[  5],r3,r2,r0,113, 13);
+	keyiter(k[  6],r4,r3,r1,114, 14); keyiter(k[  7],r0,r4,r2,115, 15);
+	keyiter(k[  8],r1,r0,r3,116, 16); keyiter(k[  9],r2,r1,r4,117, 17);
+	keyiter(k[ 10],r3,r2,r0,118, 18); keyiter(k[ 11],r4,r3,r1,119, 19);
+	keyiter(k[ 12],r0,r4,r2,120, 20); keyiter(k[ 13],r1,r0,r3,121, 21);
+	keyiter(k[ 14],r2,r1,r4,122, 22); keyiter(k[ 15],r3,r2,r0,123, 23);
+	keyiter(k[ 16],r4,r3,r1,124, 24); keyiter(k[ 17],r0,r4,r2,125, 25);
+	keyiter(k[ 18],r1,r0,r3,126, 26); keyiter(k[ 19],r2,r1,r4,127, 27);
+	keyiter(k[ 20],r3,r2,r0,128, 28); keyiter(k[ 21],r4,r3,r1,129, 29);
+	keyiter(k[ 22],r0,r4,r2,130, 30); keyiter(k[ 23],r1,r0,r3,131, 31);
+
+	/* Apply S-boxes */
+
+	S3(r3,r4,r0,r1,r2); storekeys(r1,r2,r4,r3, 28); loadkeys(r1,r2,r4,r3, 24);
+	S4(r1,r2,r4,r3,r0); storekeys(r2,r4,r3,r0, 24); loadkeys(r2,r4,r3,r0, 20);
+	S5(r2,r4,r3,r0,r1); storekeys(r1,r2,r4,r0, 20); loadkeys(r1,r2,r4,r0, 16);
+	S6(r1,r2,r4,r0,r3); storekeys(r4,r3,r2,r0, 16); loadkeys(r4,r3,r2,r0, 12);
+	S7(r4,r3,r2,r0,r1); storekeys(r1,r2,r0,r4, 12); loadkeys(r1,r2,r0,r4,  8);
+	S0(r1,r2,r0,r4,r3); storekeys(r0,r2,r4,r1,  8); loadkeys(r0,r2,r4,r1,  4);
+	S1(r0,r2,r4,r1,r3); storekeys(r3,r4,r1,r0,  4); loadkeys(r3,r4,r1,r0,  0);
+	S2(r3,r4,r1,r0,r2); storekeys(r2,r4,r3,r0,  0); loadkeys(r2,r4,r3,r0, -4);
+	S3(r2,r4,r3,r0,r1); storekeys(r0,r1,r4,r2, -4); loadkeys(r0,r1,r4,r2, -8);
+	S4(r0,r1,r4,r2,r3); storekeys(r1,r4,r2,r3, -8); loadkeys(r1,r4,r2,r3,-12);
+	S5(r1,r4,r2,r3,r0); storekeys(r0,r1,r4,r3,-12); loadkeys(r0,r1,r4,r3,-16);
+	S6(r0,r1,r4,r3,r2); storekeys(r4,r2,r1,r3,-16); loadkeys(r4,r2,r1,r3,-20);
+	S7(r4,r2,r1,r3,r0); storekeys(r0,r1,r3,r4,-20); loadkeys(r0,r1,r3,r4,-24);
+	S0(r0,r1,r3,r4,r2); storekeys(r3,r1,r4,r0,-24); loadkeys(r3,r1,r4,r0,-28);
+	k -= 50;
+	S1(r3,r1,r4,r0,r2); storekeys(r2,r4,r0,r3, 22); loadkeys(r2,r4,r0,r3, 18);
+	S2(r2,r4,r0,r3,r1); storekeys(r1,r4,r2,r3, 18); loadkeys(r1,r4,r2,r3, 14);
+	S3(r1,r4,r2,r3,r0); storekeys(r3,r0,r4,r1, 14); loadkeys(r3,r0,r4,r1, 10);
+	S4(r3,r0,r4,r1,r2); storekeys(r0,r4,r1,r2, 10); loadkeys(r0,r4,r1,r2,  6);
+	S5(r0,r4,r1,r2,r3); storekeys(r3,r0,r4,r2,  6); loadkeys(r3,r0,r4,r2,  2);
+	S6(r3,r0,r4,r2,r1); storekeys(r4,r1,r0,r2,  2); loadkeys(r4,r1,r0,r2, -2);
+	S7(r4,r1,r0,r2,r3); storekeys(r3,r0,r2,r4, -2); loadkeys(r3,r0,r2,r4, -6);
+	S0(r3,r0,r2,r4,r1); storekeys(r2,r0,r4,r3, -6); loadkeys(r2,r0,r4,r3,-10);
+	S1(r2,r0,r4,r3,r1); storekeys(r1,r4,r3,r2,-10); loadkeys(r1,r4,r3,r2,-14);
+	S2(r1,r4,r3,r2,r0); storekeys(r0,r4,r1,r2,-14); loadkeys(r0,r4,r1,r2,-18);
+	S3(r0,r4,r1,r2,r3); storekeys(r2,r3,r4,r0,-18); loadkeys(r2,r3,r4,r0,-22);
+	k -= 50;
+	S4(r2,r3,r4,r0,r1); storekeys(r3,r4,r0,r1, 28); loadkeys(r3,r4,r0,r1, 24);
+	S5(r3,r4,r0,r1,r2); storekeys(r2,r3,r4,r1, 24); loadkeys(r2,r3,r4,r1, 20);
+	S6(r2,r3,r4,r1,r0); storekeys(r4,r0,r3,r1, 20); loadkeys(r4,r0,r3,r1, 16);
+	S7(r4,r0,r3,r1,r2); storekeys(r2,r3,r1,r4, 16); loadkeys(r2,r3,r1,r4, 12);
+	S0(r2,r3,r1,r4,r0); storekeys(r1,r3,r4,r2, 12); loadkeys(r1,r3,r4,r2,  8);
+	S1(r1,r3,r4,r2,r0); storekeys(r0,r4,r2,r1,  8); loadkeys(r0,r4,r2,r1,  4);
+	S2(r0,r4,r2,r1,r3); storekeys(r3,r4,r0,r1,  4); loadkeys(r3,r4,r0,r1,  0);
+	S3(r3,r4,r0,r1,r2); storekeys(r1,r2,r4,r3,  0);
+
+	return 0;
+}
+
+static void encrypt(void *ctx, u8 *dst, const u8 *src)
+{
+	const u32
+		*k = ((struct serpent_ctx *)ctx)->expkey,
+		*s = (const u32 *)src;
+	u32	*d = (u32 *)dst,
+		r0, r1, r2, r3, r4;
+
+/*
+ * Note: The conversions between u8* and u32* might cause trouble
+ * on architectures with stricter alignment rules than x86
+ */
+
+	r0 = le32_to_cpu(s[0]);
+	r1 = le32_to_cpu(s[1]);
+	r2 = le32_to_cpu(s[2]);
+	r3 = le32_to_cpu(s[3]);
+
+				 K(r0,r1,r2,r3,0);
+	S0(r0,r1,r2,r3,r4);	LK(r2,r1,r3,r0,r4,1);
+	S1(r2,r1,r3,r0,r4);	LK(r4,r3,r0,r2,r1,2);
+	S2(r4,r3,r0,r2,r1);	LK(r1,r3,r4,r2,r0,3);
+	S3(r1,r3,r4,r2,r0);	LK(r2,r0,r3,r1,r4,4);
+	S4(r2,r0,r3,r1,r4);	LK(r0,r3,r1,r4,r2,5);
+	S5(r0,r3,r1,r4,r2);	LK(r2,r0,r3,r4,r1,6);
+	S6(r2,r0,r3,r4,r1);	LK(r3,r1,r0,r4,r2,7);
+	S7(r3,r1,r0,r4,r2);	LK(r2,r0,r4,r3,r1,8);
+	S0(r2,r0,r4,r3,r1);	LK(r4,r0,r3,r2,r1,9);
+	S1(r4,r0,r3,r2,r1);	LK(r1,r3,r2,r4,r0,10);
+	S2(r1,r3,r2,r4,r0);	LK(r0,r3,r1,r4,r2,11);
+	S3(r0,r3,r1,r4,r2);	LK(r4,r2,r3,r0,r1,12);
+	S4(r4,r2,r3,r0,r1);	LK(r2,r3,r0,r1,r4,13);
+	S5(r2,r3,r0,r1,r4);	LK(r4,r2,r3,r1,r0,14);
+	S6(r4,r2,r3,r1,r0);	LK(r3,r0,r2,r1,r4,15);
+	S7(r3,r0,r2,r1,r4);	LK(r4,r2,r1,r3,r0,16);
+	S0(r4,r2,r1,r3,r0);	LK(r1,r2,r3,r4,r0,17);
+	S1(r1,r2,r3,r4,r0);	LK(r0,r3,r4,r1,r2,18);
+	S2(r0,r3,r4,r1,r2);	LK(r2,r3,r0,r1,r4,19);
+	S3(r2,r3,r0,r1,r4);	LK(r1,r4,r3,r2,r0,20);
+	S4(r1,r4,r3,r2,r0);	LK(r4,r3,r2,r0,r1,21);
+	S5(r4,r3,r2,r0,r1);	LK(r1,r4,r3,r0,r2,22);
+	S6(r1,r4,r3,r0,r2);	LK(r3,r2,r4,r0,r1,23);
+	S7(r3,r2,r4,r0,r1);	LK(r1,r4,r0,r3,r2,24);
+	S0(r1,r4,r0,r3,r2);	LK(r0,r4,r3,r1,r2,25);
+	S1(r0,r4,r3,r1,r2);	LK(r2,r3,r1,r0,r4,26);
+	S2(r2,r3,r1,r0,r4);	LK(r4,r3,r2,r0,r1,27);
+	S3(r4,r3,r2,r0,r1);	LK(r0,r1,r3,r4,r2,28);
+	S4(r0,r1,r3,r4,r2);	LK(r1,r3,r4,r2,r0,29);
+	S5(r1,r3,r4,r2,r0);	LK(r0,r1,r3,r2,r4,30);
+	S6(r0,r1,r3,r2,r4);	LK(r3,r4,r1,r2,r0,31);
+	S7(r3,r4,r1,r2,r0);	 K(r0,r1,r2,r3,32);
+
+	d[0] = cpu_to_le32(r0);
+	d[1] = cpu_to_le32(r1);
+	d[2] = cpu_to_le32(r2);
+	d[3] = cpu_to_le32(r3);
+}
+
+static void decrypt(void *ctx, u8 *dst, const u8 *src)
+{
+	const u32
+		*k = ((struct serpent_ctx *)ctx)->expkey,
+		*s = (const u32 *)src;
+	u32	*d = (u32 *)dst,
+		r0, r1, r2, r3, r4;
+
+	r0 = le32_to_cpu(s[0]);
+	r1 = le32_to_cpu(s[1]);
+	r2 = le32_to_cpu(s[2]);
+	r3 = le32_to_cpu(s[3]);
+
+				K(r0,r1,r2,r3,32);
+	SI7(r0,r1,r2,r3,r4);	KL(r1,r3,r0,r4,r2,31);
+	SI6(r1,r3,r0,r4,r2);	KL(r0,r2,r4,r1,r3,30);
+	SI5(r0,r2,r4,r1,r3);	KL(r2,r3,r0,r4,r1,29);
+	SI4(r2,r3,r0,r4,r1);	KL(r2,r0,r1,r4,r3,28);
+	SI3(r2,r0,r1,r4,r3);	KL(r1,r2,r3,r4,r0,27);
+	SI2(r1,r2,r3,r4,r0);	KL(r2,r0,r4,r3,r1,26);
+	SI1(r2,r0,r4,r3,r1);	KL(r1,r0,r4,r3,r2,25);
+	SI0(r1,r0,r4,r3,r2);	KL(r4,r2,r0,r1,r3,24);
+	SI7(r4,r2,r0,r1,r3);	KL(r2,r1,r4,r3,r0,23);
+	SI6(r2,r1,r4,r3,r0);	KL(r4,r0,r3,r2,r1,22);
+	SI5(r4,r0,r3,r2,r1);	KL(r0,r1,r4,r3,r2,21);
+	SI4(r0,r1,r4,r3,r2);	KL(r0,r4,r2,r3,r1,20);
+	SI3(r0,r4,r2,r3,r1);	KL(r2,r0,r1,r3,r4,19);
+	SI2(r2,r0,r1,r3,r4);	KL(r0,r4,r3,r1,r2,18);
+	SI1(r0,r4,r3,r1,r2);	KL(r2,r4,r3,r1,r0,17);
+	SI0(r2,r4,r3,r1,r0);	KL(r3,r0,r4,r2,r1,16);
+	SI7(r3,r0,r4,r2,r1);	KL(r0,r2,r3,r1,r4,15);
+	SI6(r0,r2,r3,r1,r4);	KL(r3,r4,r1,r0,r2,14);
+	SI5(r3,r4,r1,r0,r2);	KL(r4,r2,r3,r1,r0,13);
+	SI4(r4,r2,r3,r1,r0);	KL(r4,r3,r0,r1,r2,12);
+	SI3(r4,r3,r0,r1,r2);	KL(r0,r4,r2,r1,r3,11);
+	SI2(r0,r4,r2,r1,r3);	KL(r4,r3,r1,r2,r0,10);
+	SI1(r4,r3,r1,r2,r0);	KL(r0,r3,r1,r2,r4,9);
+	SI0(r0,r3,r1,r2,r4);	KL(r1,r4,r3,r0,r2,8);
+	SI7(r1,r4,r3,r0,r2);	KL(r4,r0,r1,r2,r3,7);
+	SI6(r4,r0,r1,r2,r3);	KL(r1,r3,r2,r4,r0,6);
+	SI5(r1,r3,r2,r4,r0);	KL(r3,r0,r1,r2,r4,5);
+	SI4(r3,r0,r1,r2,r4);	KL(r3,r1,r4,r2,r0,4);
+	SI3(r3,r1,r4,r2,r0);	KL(r4,r3,r0,r2,r1,3);
+	SI2(r4,r3,r0,r2,r1);	KL(r3,r1,r2,r0,r4,2);
+	SI1(r3,r1,r2,r0,r4);	KL(r4,r1,r2,r0,r3,1);
+	SI0(r4,r1,r2,r0,r3);	K(r2,r3,r1,r4,0);
+
+	d[0] = cpu_to_le32(r2);
+	d[1] = cpu_to_le32(r3);
+	d[2] = cpu_to_le32(r1);
+	d[3] = cpu_to_le32(r4);
+}
+
+static struct crypto_alg serpent_alg = {
+	.cra_name		=	"serpent",
+	.cra_flags		=	CRYPTO_ALG_TYPE_CIPHER,
+	.cra_blocksize		=	SERPENT_BLOCK_SIZE,
+	.cra_ctxsize		=	sizeof(struct serpent_ctx),
+	.cra_module		=	THIS_MODULE,
+	.cra_list		=	LIST_HEAD_INIT(serpent_alg.cra_list),
+	.cra_u			=	{ .cipher = {
+	.cia_min_keysize	=	SERPENT_MIN_KEY_SIZE,
+	.cia_max_keysize	=	SERPENT_MAX_KEY_SIZE,
+	.cia_ivsize		=	SERPENT_BLOCK_SIZE,
+	.cia_setkey   		= 	setkey,
+	.cia_encrypt 		=	encrypt,
+	.cia_decrypt  		=	decrypt } }
+};
+
+static int __init init(void)
+{
+	return crypto_register_alg(&serpent_alg);
+}
+
+static void __exit fini(void)
+{
+	crypto_unregister_alg(&serpent_alg);
+}
+
+module_init(init);
+module_exit(fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Serpent Cipher Algorithm");
+MODULE_AUTHOR("Dag Arne Osvik <osvik@ii.uib.no>");

crypto/tcrypt.c

@@ -48,7 +48,7 @@ static char *tvmem;
 
 static char *check[] = {
 	"des", "md5", "des3_ede", "rot13", "sha1", "sha256", "blowfish",
-	"twofish",
+	"twofish", "serpent",
 	 NULL
 };
 
@@ -1884,6 +1884,105 @@ test_twofish(void)
 	crypto_free_tfm(tfm);
 }
 
+void
+test_serpent(void)
+{
+	unsigned int ret, i, tsize;
+	u8 *p, *q, *key;
+	struct crypto_tfm *tfm;
+	struct serpent_tv *serp_tv;
+	struct scatterlist sg[1];
+
+	printk("\ntesting serpent encryption\n");
+
+	tfm = crypto_alloc_tfm("serpent", 0);
+	if (tfm == NULL) {
+		printk("failed to load transform for serpent (default ecb)\n");
+		return;
+	}
+
+	tsize = sizeof (serpent_enc_tv_template);
+	if (tsize > TVMEMSIZE) {
+		printk("template (%u) too big for tvmem (%u)\n", tsize,
+		       TVMEMSIZE);
+		return;
+	}
+
+	memcpy(tvmem, serpent_enc_tv_template, tsize);
+	serp_tv = (void *) tvmem;
+	for (i = 0; i < SERPENT_ENC_TEST_VECTORS; i++) {
+		printk("test %u (%d bit key):\n", i + 1, serp_tv[i].keylen * 8);
+		key = serp_tv[i].key;
+
+		ret = crypto_cipher_setkey(tfm, key, serp_tv[i].keylen);
+		if (ret) {
+			printk("setkey() failed flags=%x\n", tfm->crt_flags);
+
+			if (!serp_tv[i].fail)
+				goto out;
+		}
+
+		p = serp_tv[i].plaintext;
+		sg[0].page = virt_to_page(p);
+		sg[0].offset = ((long) p & ~PAGE_MASK);
+		sg[0].length = sizeof(serp_tv[i].plaintext);
+		ret = crypto_cipher_encrypt(tfm, sg, 1);
+		if (ret) {
+			printk("encrypt() failed flags=%x\n", tfm->crt_flags);
+			goto out;
+		}
+
+		q = kmap(sg[0].page) + sg[0].offset;
+		hexdump(q, sizeof(serp_tv[i].result));
+
+		printk("%s\n", memcmp(q, serp_tv[i].result,
+			sizeof(serp_tv[i].result)) ? "fail" : "pass");
+	}
+
+	printk("\ntesting serpent decryption\n");
+
+	tsize = sizeof (serpent_dec_tv_template);
+	if (tsize > TVMEMSIZE) {
+		printk("template (%u) too big for tvmem (%u)\n", tsize,
+		       TVMEMSIZE);
+		return;
+	}
+
+	memcpy(tvmem, serpent_dec_tv_template, tsize);
+	serp_tv = (void *) tvmem;
+	for (i = 0; i < SERPENT_DEC_TEST_VECTORS; i++) {
+		printk("test %u (%d bit key):\n", i + 1, serp_tv[i].keylen * 8);
+		key = serp_tv[i].key;
+
+		ret = crypto_cipher_setkey(tfm, key, serp_tv[i].keylen);
+		if (ret) {
+			printk("setkey() failed flags=%x\n", tfm->crt_flags);
+
+			if (!serp_tv[i].fail)
+				goto out;
+		}
+
+		p = serp_tv[i].plaintext;
+		sg[0].page = virt_to_page(p);
+		sg[0].offset = ((long) p & ~PAGE_MASK);
+		sg[0].length = sizeof(serp_tv[i].plaintext);
+		ret = crypto_cipher_decrypt(tfm, sg, 1);
+		if (ret) {
+			printk("decrypt() failed flags=%x\n", tfm->crt_flags);
+			goto out;
+		}
+
+		q = kmap(sg[0].page) + sg[0].offset;
+		hexdump(q, sizeof(serp_tv[i].result));
+
+		printk("%s\n", memcmp(q, serp_tv[i].result,
+			sizeof(serp_tv[i].result)) ? "fail" : "pass");
+	}
+
+out:
+	crypto_free_tfm(tfm);
+}
+
 static void
 test_available(void)
 {
@@ -1911,6 +2010,7 @@ do_test(void)
 		test_sha256();
 		test_blowfish();
 		test_twofish();
+		test_serpent();
 #ifdef CONFIG_CRYPTO_HMAC
 		test_hmac_md5();
 		test_hmac_sha1();
@@ -1950,6 +2050,10 @@ do_test(void)
 		test_twofish();
 		break;
 
+	case 9:
+		test_serpent();
+		break;
+
 #ifdef CONFIG_CRYPTO_HMAC
 	case 100:
 		test_hmac_md5();

crypto/tcrypt.h

@@ -1384,4 +1384,100 @@ struct tf_tv tf_cbc_dec_tv_template[] = {
 	},
 };
 
+/*
+ * Serpent test vectors.  These are backwards because Serpent writes
+ * octect sequences in right-to-left mode.
+ */
+#define SERPENT_ENC_TEST_VECTORS	4
+#define SERPENT_DEC_TEST_VECTORS	4
+
+struct serpent_tv {
+	unsigned int keylen, fail;
+	u8 key[32], plaintext[16], result[16];
+};
+
+struct serpent_tv serpent_enc_tv_template[] = 
+{
+	{
+		0, 0,
+		{ 0 },
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f },
+		{ 0x12, 0x07, 0xfc, 0xce, 0x9b, 0xd0, 0xd6, 0x47,
+		  0x6a, 0xe9, 0x8f, 0xbe, 0xd1, 0x43, 0xa0, 0xe2 }
+	},
+	{
+		16, 0,
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f },
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f },
+		{ 0x4c, 0x7d, 0x8a, 0x32, 0x80, 0x72, 0xa2, 0x2c,
+		  0x82, 0x3e, 0x4a, 0x1f, 0x3a, 0xcd, 0xa1, 0x6d }
+	},
+	{
+		32, 0,
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
+		  0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
+		  0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f },
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f },
+		{ 0xde, 0x26, 0x9f, 0xf8, 0x33, 0xe4, 0x32, 0xb8,
+		  0x5b, 0x2e, 0x88, 0xd2, 0x70, 0x1c, 0xe7, 0x5c }
+	},
+	{
+		16, 0,
+		{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+		  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80 },
+		{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+		  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 },
+		{ 0xdd, 0xd2, 0x6b, 0x98, 0xa5, 0xff, 0xd8, 0x2c,
+		  0x05, 0x34, 0x5a, 0x9d, 0xad, 0xbf, 0xaf, 0x49}
+	}
+};
+
+struct serpent_tv serpent_dec_tv_template[] = 
+{
+	{
+		0, 0,
+		{ 0 },
+		{ 0x12, 0x07, 0xfc, 0xce, 0x9b, 0xd0, 0xd6, 0x47,
+		  0x6a, 0xe9, 0x8f, 0xbe, 0xd1, 0x43, 0xa0, 0xe2 },
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f },
+
+	},
+	{
+		16, 0,
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f },
+		{ 0x4c, 0x7d, 0x8a, 0x32, 0x80, 0x72, 0xa2, 0x2c,
+		  0x82, 0x3e, 0x4a, 0x1f, 0x3a, 0xcd, 0xa1, 0x6d },
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f },
+	},
+	{
+		32, 0,
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
+		  0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
+		  0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f },
+
+		{ 0xde, 0x26, 0x9f, 0xf8, 0x33, 0xe4, 0x32, 0xb8,
+		  0x5b, 0x2e, 0x88, 0xd2, 0x70, 0x1c, 0xe7, 0x5c },
+		{ 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+		  0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f },
+	},
+	{
+		16, 0,
+		{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+		  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80 },
+		{ 0xdd, 0xd2, 0x6b, 0x98, 0xa5, 0xff, 0xd8, 0x2c,
+		  0x05, 0x34, 0x5a, 0x9d, 0xad, 0xbf, 0xaf, 0x49},
+		{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+		  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 },  
+	}
+};
+
 #endif	/* _CRYPTO_TCRYPT_H */