- markus@cvs.openbsd.org 2000/12/06 23:10:39

     [rijndael.c]
     unexpand(1)
   - markus@cvs.openbsd.org 2000/12/06 23:05:43
     [cipher.c cipher.h rijndael.c rijndael.h rijndael_boxes.h]
     new rijndael implementation. fixes endian bugs

1 files changed, 294 insertions, 476 deletions

diff --git a/rijndael.c b/rijndael.c
index 96373803..92a39762 100644
--- a/rijndael.c
+++ b/rijndael.c
@@ -1,493 +1,311 @@
-/*	$OpenBSD: rijndael.c,v 1.2 2000/10/15 14:14:01 markus Exp $	*/
-
-/* This is an independent implementation of the encryption algorithm:   */
-/*                                                                      */
-/*         RIJNDAEL by Joan Daemen and Vincent Rijmen                   */
-/*                                                                      */
-/* which is a candidate algorithm in the Advanced Encryption Standard   */
-/* programme of the US National Institute of Standards and Technology.  */
-/*                                                                      */
-/* Copyright in this implementation is held by Dr B R Gladman but I     */
-/* hereby give permission for its free direct or derivative use subject */
-/* to acknowledgment of its origin and compliance with any conditions   */
-/* that the originators of the algorithm place on its exploitation.     */
-/*                                                                      */
-/* Dr Brian Gladman (gladman@seven77.demon.co.uk) 14th January 1999     */
-
-/* Timing data for Rijndael (rijndael.c)
-
-Algorithm: rijndael (rijndael.c)
-
-128 bit key:
-Key Setup:    305/1389 cycles (encrypt/decrypt)
-Encrypt:       374 cycles =    68.4 mbits/sec
-Decrypt:       352 cycles =    72.7 mbits/sec
-Mean:          363 cycles =    70.5 mbits/sec
-
-192 bit key:
-Key Setup:    277/1595 cycles (encrypt/decrypt)
-Encrypt:       439 cycles =    58.3 mbits/sec
-Decrypt:       425 cycles =    60.2 mbits/sec
-Mean:          432 cycles =    59.3 mbits/sec
-
-256 bit key:
-Key Setup:    374/1960 cycles (encrypt/decrypt)
-Encrypt:       502 cycles =    51.0 mbits/sec
-Decrypt:       498 cycles =    51.4 mbits/sec
-Mean:          500 cycles =    51.2 mbits/sec
-
-*/
+/*
+ * rijndael-alg-fst.c   v2.4   April '2000
+ * rijndael-alg-api.c   v2.4   April '2000
+ *
+ * Optimised ANSI C code
+ *
+ * authors: v1.0: Antoon Bosselaers
+ *          v2.0: Vincent Rijmen, K.U.Leuven
+ *          v2.3: Paulo Barreto
+ *          v2.4: Vincent Rijmen, K.U.Leuven
+ *
+ * This code is placed in the public domain.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <assert.h>
 
 #include "config.h"
 #include "rijndael.h"
+#include "rijndael_boxes.h"
 
-void gen_tabs	__P((void));
-
-/* 3. Basic macros for speeding up generic operations               */
-
-/* Circular rotate of 32 bit values                                 */
-
-#define rotr(x,n)   (((x) >> ((int)(n))) | ((x) << (32 - (int)(n))))
-#define rotl(x,n)   (((x) << ((int)(n))) | ((x) >> (32 - (int)(n))))
-
-/* Invert byte order in a 32 bit variable                           */
-
-#define bswap(x)    (rotl(x, 8) & 0x00ff00ff | rotr(x, 8) & 0xff00ff00)
-
-/* Extract byte from a 32 bit quantity (little endian notation)     */ 
-
-#define byte(x,n)   ((u1byte)((x) >> (8 * n)))
-
-#if BYTE_ORDER != LITTLE_ENDIAN
-#define BLOCK_SWAP
-#endif
-
-/* For inverting byte order in input/output 32 bit words if needed  */
-
-#ifdef  BLOCK_SWAP
-#define BYTE_SWAP
-#define WORD_SWAP
-#endif
-
-#ifdef  BYTE_SWAP
-#define io_swap(x)  bswap(x)
-#else
-#define io_swap(x)  (x)
-#endif
-
-/* For inverting the byte order of input/output blocks if needed    */
-
-#ifdef  WORD_SWAP
-
-#define get_block(x)                            \
-    ((u4byte*)(x))[0] = io_swap(in_blk[3]);     \
-    ((u4byte*)(x))[1] = io_swap(in_blk[2]);     \
-    ((u4byte*)(x))[2] = io_swap(in_blk[1]);     \
-    ((u4byte*)(x))[3] = io_swap(in_blk[0])
-
-#define put_block(x)                            \
-    out_blk[3] = io_swap(((u4byte*)(x))[0]);    \
-    out_blk[2] = io_swap(((u4byte*)(x))[1]);    \
-    out_blk[1] = io_swap(((u4byte*)(x))[2]);    \
-    out_blk[0] = io_swap(((u4byte*)(x))[3])
-
-#define get_key(x,len)                          \
-    ((u4byte*)(x))[4] = ((u4byte*)(x))[5] =     \
-    ((u4byte*)(x))[6] = ((u4byte*)(x))[7] = 0;  \
-    switch((((len) + 63) / 64)) {               \
-    case 2:                                     \
-    ((u4byte*)(x))[0] = io_swap(in_key[3]);     \
-    ((u4byte*)(x))[1] = io_swap(in_key[2]);     \
-    ((u4byte*)(x))[2] = io_swap(in_key[1]);     \
-    ((u4byte*)(x))[3] = io_swap(in_key[0]);     \
-    break;                                      \
-    case 3:                                     \
-    ((u4byte*)(x))[0] = io_swap(in_key[5]);     \
-    ((u4byte*)(x))[1] = io_swap(in_key[4]);     \
-    ((u4byte*)(x))[2] = io_swap(in_key[3]);     \
-    ((u4byte*)(x))[3] = io_swap(in_key[2]);     \
-    ((u4byte*)(x))[4] = io_swap(in_key[1]);     \
-    ((u4byte*)(x))[5] = io_swap(in_key[0]);     \
-    break;                                      \
-    case 4:                                     \
-    ((u4byte*)(x))[0] = io_swap(in_key[7]);     \
-    ((u4byte*)(x))[1] = io_swap(in_key[6]);     \
-    ((u4byte*)(x))[2] = io_swap(in_key[5]);     \
-    ((u4byte*)(x))[3] = io_swap(in_key[4]);     \
-    ((u4byte*)(x))[4] = io_swap(in_key[3]);     \
-    ((u4byte*)(x))[5] = io_swap(in_key[2]);     \
-    ((u4byte*)(x))[6] = io_swap(in_key[1]);     \
-    ((u4byte*)(x))[7] = io_swap(in_key[0]);     \
-    }
-
-#else
-
-#define get_block(x)                            \
-    ((u4byte*)(x))[0] = io_swap(in_blk[0]);     \
-    ((u4byte*)(x))[1] = io_swap(in_blk[1]);     \
-    ((u4byte*)(x))[2] = io_swap(in_blk[2]);     \
-    ((u4byte*)(x))[3] = io_swap(in_blk[3])
-
-#define put_block(x)                            \
-    out_blk[0] = io_swap(((u4byte*)(x))[0]);    \
-    out_blk[1] = io_swap(((u4byte*)(x))[1]);    \
-    out_blk[2] = io_swap(((u4byte*)(x))[2]);    \
-    out_blk[3] = io_swap(((u4byte*)(x))[3])
-
-#define get_key(x,len)                          \
-    ((u4byte*)(x))[4] = ((u4byte*)(x))[5] =     \
-    ((u4byte*)(x))[6] = ((u4byte*)(x))[7] = 0;  \
-    switch((((len) + 63) / 64)) {               \
-    case 4:                                     \
-    ((u4byte*)(x))[6] = io_swap(in_key[6]);     \
-    ((u4byte*)(x))[7] = io_swap(in_key[7]);     \
-    case 3:                                     \
-    ((u4byte*)(x))[4] = io_swap(in_key[4]);     \
-    ((u4byte*)(x))[5] = io_swap(in_key[5]);     \
-    case 2:                                     \
-    ((u4byte*)(x))[0] = io_swap(in_key[0]);     \
-    ((u4byte*)(x))[1] = io_swap(in_key[1]);     \
-    ((u4byte*)(x))[2] = io_swap(in_key[2]);     \
-    ((u4byte*)(x))[3] = io_swap(in_key[3]);     \
-    }
-
-#endif
-
-#define LARGE_TABLES
-
-u1byte  pow_tab[256];
-u1byte  log_tab[256];
-u1byte  sbx_tab[256];
-u1byte  isb_tab[256];
-u4byte  rco_tab[ 10];
-u4byte  ft_tab[4][256];
-u4byte  it_tab[4][256];
-
-#ifdef  LARGE_TABLES
-  u4byte  fl_tab[4][256];
-  u4byte  il_tab[4][256];
-#endif
-
-u4byte  tab_gen = 0;
-
-#define ff_mult(a,b)    (a && b ? pow_tab[(log_tab[a] + log_tab[b]) % 255] : 0)
-
-#define f_rn(bo, bi, n, k)                          \
-    bo[n] =  ft_tab[0][byte(bi[n],0)] ^             \
-             ft_tab[1][byte(bi[(n + 1) & 3],1)] ^   \
-             ft_tab[2][byte(bi[(n + 2) & 3],2)] ^   \
-             ft_tab[3][byte(bi[(n + 3) & 3],3)] ^ *(k + n)
-
-#define i_rn(bo, bi, n, k)                          \
-    bo[n] =  it_tab[0][byte(bi[n],0)] ^             \
-             it_tab[1][byte(bi[(n + 3) & 3],1)] ^   \
-             it_tab[2][byte(bi[(n + 2) & 3],2)] ^   \
-             it_tab[3][byte(bi[(n + 1) & 3],3)] ^ *(k + n)
-
-#ifdef LARGE_TABLES
-
-#define ls_box(x)                \
-    ( fl_tab[0][byte(x, 0)] ^    \
-      fl_tab[1][byte(x, 1)] ^    \
-      fl_tab[2][byte(x, 2)] ^    \
-      fl_tab[3][byte(x, 3)] )
-
-#define f_rl(bo, bi, n, k)                          \
-    bo[n] =  fl_tab[0][byte(bi[n],0)] ^             \
-             fl_tab[1][byte(bi[(n + 1) & 3],1)] ^   \
-             fl_tab[2][byte(bi[(n + 2) & 3],2)] ^   \
-             fl_tab[3][byte(bi[(n + 3) & 3],3)] ^ *(k + n)
-
-#define i_rl(bo, bi, n, k)                          \
-    bo[n] =  il_tab[0][byte(bi[n],0)] ^             \
-             il_tab[1][byte(bi[(n + 3) & 3],1)] ^   \
-             il_tab[2][byte(bi[(n + 2) & 3],2)] ^   \
-             il_tab[3][byte(bi[(n + 1) & 3],3)] ^ *(k + n)
-
-#else
-
-#define ls_box(x)                            \
-    ((u4byte)sbx_tab[byte(x, 0)] <<  0) ^    \
-    ((u4byte)sbx_tab[byte(x, 1)] <<  8) ^    \
-    ((u4byte)sbx_tab[byte(x, 2)] << 16) ^    \
-    ((u4byte)sbx_tab[byte(x, 3)] << 24)
-
-#define f_rl(bo, bi, n, k)                                      \
-    bo[n] = (u4byte)sbx_tab[byte(bi[n],0)] ^                    \
-        rotl(((u4byte)sbx_tab[byte(bi[(n + 1) & 3],1)]),  8) ^  \
-        rotl(((u4byte)sbx_tab[byte(bi[(n + 2) & 3],2)]), 16) ^  \
-        rotl(((u4byte)sbx_tab[byte(bi[(n + 3) & 3],3)]), 24) ^ *(k + n)
-
-#define i_rl(bo, bi, n, k)                                      \
-    bo[n] = (u4byte)isb_tab[byte(bi[n],0)] ^                    \
-        rotl(((u4byte)isb_tab[byte(bi[(n + 3) & 3],1)]),  8) ^  \
-        rotl(((u4byte)isb_tab[byte(bi[(n + 2) & 3],2)]), 16) ^  \
-        rotl(((u4byte)isb_tab[byte(bi[(n + 1) & 3],3)]), 24) ^ *(k + n)
-
-#endif
-
-void
-gen_tabs(void)
+int
+rijndael_keysched(u_int8_t k[RIJNDAEL_MAXKC][4],
+    u_int8_t W[RIJNDAEL_MAXROUNDS+1][4][4], int ROUNDS)
 {
-	u4byte  i, t;
-	u1byte  p, q;
-
-	/* log and power tables for GF(2**8) finite field with  */
-	/* 0x11b as modular polynomial - the simplest prmitive  */
-	/* root is 0x11, used here to generate the tables       */
-
-	for(i = 0,p = 1; i < 256; ++i) {
-		pow_tab[i] = (u1byte)p; log_tab[p] = (u1byte)i;
-
-		p = p ^ (p << 1) ^ (p & 0x80 ? 0x01b : 0);
-	}
-
-	log_tab[1] = 0; p = 1;
-
-	for(i = 0; i < 10; ++i) {
-		rco_tab[i] = p; 
-
-		p = (p << 1) ^ (p & 0x80 ? 0x1b : 0);
+	/* Calculate the necessary round keys
+	 * The number of calculations depends on keyBits and blockBits
+	 */ 
+	int j, r, t, rconpointer = 0;
+	u_int8_t tk[RIJNDAEL_MAXKC][4];
+	int KC = ROUNDS - 6;
+
+	for (j = KC-1; j >= 0; j--) {
+		*((u_int32_t*)tk[j]) = *((u_int32_t*)k[j]);
 	}
-
-	/* note that the affine byte transformation matrix in   */
-	/* rijndael specification is in big endian format with  */
-	/* bit 0 as the most significant bit. In the remainder  */
-	/* of the specification the bits are numbered from the  */
-	/* least significant end of a byte.                     */
-
-	for(i = 0; i < 256; ++i) {
-		p = (i ? pow_tab[255 - log_tab[i]] : 0); q = p; 
-		q = (q >> 7) | (q << 1); p ^= q; 
-		q = (q >> 7) | (q << 1); p ^= q; 
-		q = (q >> 7) | (q << 1); p ^= q; 
-		q = (q >> 7) | (q << 1); p ^= q ^ 0x63; 
-		sbx_tab[i] = (u1byte)p; isb_tab[p] = (u1byte)i;
-	}
-
-	for(i = 0; i < 256; ++i) {
-		p = sbx_tab[i]; 
-
-#ifdef  LARGE_TABLES        
-        
-		t = p; fl_tab[0][i] = t;
-		fl_tab[1][i] = rotl(t,  8);
-		fl_tab[2][i] = rotl(t, 16);
-		fl_tab[3][i] = rotl(t, 24);
-#endif
-		t = ((u4byte)ff_mult(2, p)) |
-			((u4byte)p <<  8) |
-			((u4byte)p << 16) |
-			((u4byte)ff_mult(3, p) << 24);
-        
-		ft_tab[0][i] = t;
-		ft_tab[1][i] = rotl(t,  8);
-		ft_tab[2][i] = rotl(t, 16);
-		ft_tab[3][i] = rotl(t, 24);
-
-		p = isb_tab[i]; 
-
-#ifdef  LARGE_TABLES        
-        
-		t = p; il_tab[0][i] = t; 
-		il_tab[1][i] = rotl(t,  8); 
-		il_tab[2][i] = rotl(t, 16); 
-		il_tab[3][i] = rotl(t, 24);
-#endif 
-		t = ((u4byte)ff_mult(14, p)) |
-			((u4byte)ff_mult( 9, p) <<  8) |
-			((u4byte)ff_mult(13, p) << 16) |
-			((u4byte)ff_mult(11, p) << 24);
-        
-		it_tab[0][i] = t; 
-		it_tab[1][i] = rotl(t,  8); 
-		it_tab[2][i] = rotl(t, 16); 
-		it_tab[3][i] = rotl(t, 24); 
-	}
-
-	tab_gen = 1;
-}
-
-#define star_x(x) (((x) & 0x7f7f7f7f) << 1) ^ ((((x) & 0x80808080) >> 7) * 0x1b)
-
-#define imix_col(y,x)       \
-    u   = star_x(x);        \
-    v   = star_x(u);        \
-    w   = star_x(v);        \
-    t   = w ^ (x);          \
-   (y)  = u ^ v ^ w;        \
-   (y) ^= rotr(u ^ t,  8) ^ \
-          rotr(v ^ t, 16) ^ \
-          rotr(t,24)
-
-/* initialise the key schedule from the user supplied key   */
-
-#define loop4(i)                                    \
-{   t = ls_box(rotr(t,  8)) ^ rco_tab[i];           \
-    t ^= e_key[4 * i];     e_key[4 * i + 4] = t;    \
-    t ^= e_key[4 * i + 1]; e_key[4 * i + 5] = t;    \
-    t ^= e_key[4 * i + 2]; e_key[4 * i + 6] = t;    \
-    t ^= e_key[4 * i + 3]; e_key[4 * i + 7] = t;    \
-}
-
-#define loop6(i)                                    \
-{   t = ls_box(rotr(t,  8)) ^ rco_tab[i];           \
-    t ^= e_key[6 * i];     e_key[6 * i + 6] = t;    \
-    t ^= e_key[6 * i + 1]; e_key[6 * i + 7] = t;    \
-    t ^= e_key[6 * i + 2]; e_key[6 * i + 8] = t;    \
-    t ^= e_key[6 * i + 3]; e_key[6 * i + 9] = t;    \
-    t ^= e_key[6 * i + 4]; e_key[6 * i + 10] = t;   \
-    t ^= e_key[6 * i + 5]; e_key[6 * i + 11] = t;   \
-}
-
-#define loop8(i)                                    \
-{   t = ls_box(rotr(t,  8)) ^ rco_tab[i];           \
-    t ^= e_key[8 * i];     e_key[8 * i + 8] = t;    \
-    t ^= e_key[8 * i + 1]; e_key[8 * i + 9] = t;    \
-    t ^= e_key[8 * i + 2]; e_key[8 * i + 10] = t;   \
-    t ^= e_key[8 * i + 3]; e_key[8 * i + 11] = t;   \
-    t  = e_key[8 * i + 4] ^ ls_box(t);              \
-    e_key[8 * i + 12] = t;                          \
-    t ^= e_key[8 * i + 5]; e_key[8 * i + 13] = t;   \
-    t ^= e_key[8 * i + 6]; e_key[8 * i + 14] = t;   \
-    t ^= e_key[8 * i + 7]; e_key[8 * i + 15] = t;   \
-}
-
-rijndael_ctx *
-rijndael_set_key(rijndael_ctx *ctx, const u4byte *in_key, const u4byte key_len,
-		 int encrypt)
-{  
-	u4byte  i, t, u, v, w;
-	u4byte *e_key = ctx->e_key;
-	u4byte *d_key = ctx->d_key;
-
-	ctx->decrypt = !encrypt;
-
-	if(!tab_gen)
-		gen_tabs();
-
-	ctx->k_len = (key_len + 31) / 32;
-
-	e_key[0] = in_key[0]; e_key[1] = in_key[1];
-	e_key[2] = in_key[2]; e_key[3] = in_key[3];
-	
-	switch(ctx->k_len) {
-        case 4: t = e_key[3];
-                for(i = 0; i < 10; ++i) 
-			loop4(i);
-                break;
-
-        case 6: e_key[4] = in_key[4]; t = e_key[5] = in_key[5];
-                for(i = 0; i < 8; ++i) 
-			loop6(i);
-                break;
-
-        case 8: e_key[4] = in_key[4]; e_key[5] = in_key[5];
-                e_key[6] = in_key[6]; t = e_key[7] = in_key[7];
-                for(i = 0; i < 7; ++i) 
-			loop8(i);
-                break;
-	}
-
-	if (!encrypt) {
-		d_key[0] = e_key[0]; d_key[1] = e_key[1];
-		d_key[2] = e_key[2]; d_key[3] = e_key[3];
-
-		for(i = 4; i < 4 * ctx->k_len + 24; ++i) {
-			imix_col(d_key[i], e_key[i]);
+	r = 0;
+	t = 0;
+	/* copy values into round key array */
+	for (j = 0; (j < KC) && (r < ROUNDS + 1); ) {
+		for (; (j < KC) && (t < 4); j++, t++) {
+			*((u_int32_t*)W[r][t]) = *((u_int32_t*)tk[j]);
+		}
+		if (t == 4) {
+			r++;
+			t = 0;
 		}
 	}
-
-	return ctx;
+		
+	while (r < ROUNDS + 1) { /* while not enough round key material calculated */
+		/* calculate new values */
+		tk[0][0] ^= S[tk[KC-1][1]];
+		tk[0][1] ^= S[tk[KC-1][2]];
+		tk[0][2] ^= S[tk[KC-1][3]];
+		tk[0][3] ^= S[tk[KC-1][0]];
+		tk[0][0] ^= rcon[rconpointer++];
+
+		if (KC != 8) {
+			for (j = 1; j < KC; j++) {
+				*((u_int32_t*)tk[j]) ^= *((u_int32_t*)tk[j-1]);
+			}
+		} else {
+			for (j = 1; j < KC/2; j++) {
+				*((u_int32_t*)tk[j]) ^= *((u_int32_t*)tk[j-1]);
+			}
+			tk[KC/2][0] ^= S[tk[KC/2 - 1][0]];
+			tk[KC/2][1] ^= S[tk[KC/2 - 1][1]];
+			tk[KC/2][2] ^= S[tk[KC/2 - 1][2]];
+			tk[KC/2][3] ^= S[tk[KC/2 - 1][3]];
+			for (j = KC/2 + 1; j < KC; j++) {
+				*((u_int32_t*)tk[j]) ^= *((u_int32_t*)tk[j-1]);
+			}
+		}
+		/* copy values into round key array */
+		for (j = 0; (j < KC) && (r < ROUNDS + 1); ) {
+			for (; (j < KC) && (t < 4); j++, t++) {
+				*((u_int32_t*)W[r][t]) = *((u_int32_t*)tk[j]);
+			}
+			if (t == 4) {
+				r++;
+				t = 0;
+			}
+		}
+	}		
+	return 0;
 }
 
-/* encrypt a block of text  */
-
-#define f_nround(bo, bi, k) \
-    f_rn(bo, bi, 0, k);     \
-    f_rn(bo, bi, 1, k);     \
-    f_rn(bo, bi, 2, k);     \
-    f_rn(bo, bi, 3, k);     \
-    k += 4
-
-#define f_lround(bo, bi, k) \
-    f_rl(bo, bi, 0, k);     \
-    f_rl(bo, bi, 1, k);     \
-    f_rl(bo, bi, 2, k);     \
-    f_rl(bo, bi, 3, k)
-
-void
-rijndael_encrypt(rijndael_ctx *ctx, const u4byte *in_blk, u4byte *out_blk)
-{   
-	u4byte k_len = ctx->k_len;
-	u4byte *e_key = ctx->e_key;
-	u4byte  b0[4], b1[4], *kp;
-
-	b0[0] = in_blk[0] ^ e_key[0]; b0[1] = in_blk[1] ^ e_key[1];
-	b0[2] = in_blk[2] ^ e_key[2]; b0[3] = in_blk[3] ^ e_key[3];
-
-	kp = e_key + 4;
-
-	if(k_len > 6) {
-		f_nround(b1, b0, kp); f_nround(b0, b1, kp);
+int
+rijndael_key_enc_to_dec(u_int8_t W[RIJNDAEL_MAXROUNDS+1][4][4], int ROUNDS)
+{
+	int r;
+	u_int8_t *w;
+
+	for (r = 1; r < ROUNDS; r++) {
+		w = W[r][0];
+		*((u_int32_t*)w) = *((u_int32_t*)U1[w[0]])
+				 ^ *((u_int32_t*)U2[w[1]])
+				 ^ *((u_int32_t*)U3[w[2]])
+				 ^ *((u_int32_t*)U4[w[3]]);
+
+		w = W[r][1];
+		*((u_int32_t*)w) = *((u_int32_t*)U1[w[0]])
+				 ^ *((u_int32_t*)U2[w[1]])
+				 ^ *((u_int32_t*)U3[w[2]])
+				 ^ *((u_int32_t*)U4[w[3]]);
+
+		w = W[r][2];
+		*((u_int32_t*)w) = *((u_int32_t*)U1[w[0]])
+				 ^ *((u_int32_t*)U2[w[1]])
+				 ^ *((u_int32_t*)U3[w[2]])
+				 ^ *((u_int32_t*)U4[w[3]]);
+
+		w = W[r][3];
+		*((u_int32_t*)w) = *((u_int32_t*)U1[w[0]])
+				 ^ *((u_int32_t*)U2[w[1]])
+				 ^ *((u_int32_t*)U3[w[2]])
+				 ^ *((u_int32_t*)U4[w[3]]);
 	}
-
-	if(k_len > 4) {
-		f_nround(b1, b0, kp); f_nround(b0, b1, kp);
+	return 0;
+}	
+
+/**
+ * Encrypt a single block. 
+ */
+int
+rijndael_encrypt(rijndael_key *key, u_int8_t a[16], u_int8_t b[16])
+{
+	u_int8_t (*rk)[4][4] = key->keySched;
+	int ROUNDS = key->ROUNDS;
+	int r;
+	u_int8_t temp[4][4];
+
+	*((u_int32_t*)temp[0]) = *((u_int32_t*)(a   )) ^ *((u_int32_t*)rk[0][0]);
+	*((u_int32_t*)temp[1]) = *((u_int32_t*)(a+ 4)) ^ *((u_int32_t*)rk[0][1]);
+	*((u_int32_t*)temp[2]) = *((u_int32_t*)(a+ 8)) ^ *((u_int32_t*)rk[0][2]);
+	*((u_int32_t*)temp[3]) = *((u_int32_t*)(a+12)) ^ *((u_int32_t*)rk[0][3]);
+	*((u_int32_t*)(b    )) = *((u_int32_t*)T1[temp[0][0]])
+			       ^ *((u_int32_t*)T2[temp[1][1]])
+			       ^ *((u_int32_t*)T3[temp[2][2]]) 
+			       ^ *((u_int32_t*)T4[temp[3][3]]);
+	*((u_int32_t*)(b + 4)) = *((u_int32_t*)T1[temp[1][0]])
+			       ^ *((u_int32_t*)T2[temp[2][1]])
+			       ^ *((u_int32_t*)T3[temp[3][2]]) 
+			       ^ *((u_int32_t*)T4[temp[0][3]]);
+	*((u_int32_t*)(b + 8)) = *((u_int32_t*)T1[temp[2][0]])
+			       ^ *((u_int32_t*)T2[temp[3][1]])
+			       ^ *((u_int32_t*)T3[temp[0][2]]) 
+			       ^ *((u_int32_t*)T4[temp[1][3]]);
+	*((u_int32_t*)(b +12)) = *((u_int32_t*)T1[temp[3][0]])
+			       ^ *((u_int32_t*)T2[temp[0][1]])
+			       ^ *((u_int32_t*)T3[temp[1][2]]) 
+			       ^ *((u_int32_t*)T4[temp[2][3]]);
+	for (r = 1; r < ROUNDS-1; r++) {
+		*((u_int32_t*)temp[0]) = *((u_int32_t*)(b   )) ^ *((u_int32_t*)rk[r][0]);
+		*((u_int32_t*)temp[1]) = *((u_int32_t*)(b+ 4)) ^ *((u_int32_t*)rk[r][1]);
+		*((u_int32_t*)temp[2]) = *((u_int32_t*)(b+ 8)) ^ *((u_int32_t*)rk[r][2]);
+		*((u_int32_t*)temp[3]) = *((u_int32_t*)(b+12)) ^ *((u_int32_t*)rk[r][3]);
+
+		*((u_int32_t*)(b    )) = *((u_int32_t*)T1[temp[0][0]])
+				       ^ *((u_int32_t*)T2[temp[1][1]])
+				       ^ *((u_int32_t*)T3[temp[2][2]]) 
+				       ^ *((u_int32_t*)T4[temp[3][3]]);
+		*((u_int32_t*)(b + 4)) = *((u_int32_t*)T1[temp[1][0]])
+				       ^ *((u_int32_t*)T2[temp[2][1]])
+				       ^ *((u_int32_t*)T3[temp[3][2]]) 
+				       ^ *((u_int32_t*)T4[temp[0][3]]);
+		*((u_int32_t*)(b + 8)) = *((u_int32_t*)T1[temp[2][0]])
+				       ^ *((u_int32_t*)T2[temp[3][1]])
+				       ^ *((u_int32_t*)T3[temp[0][2]]) 
+				       ^ *((u_int32_t*)T4[temp[1][3]]);
+		*((u_int32_t*)(b +12)) = *((u_int32_t*)T1[temp[3][0]])
+				       ^ *((u_int32_t*)T2[temp[0][1]])
+				       ^ *((u_int32_t*)T3[temp[1][2]]) 
+				       ^ *((u_int32_t*)T4[temp[2][3]]);
 	}
-
-	f_nround(b1, b0, kp); f_nround(b0, b1, kp);
-	f_nround(b1, b0, kp); f_nround(b0, b1, kp);
-	f_nround(b1, b0, kp); f_nround(b0, b1, kp);
-	f_nround(b1, b0, kp); f_nround(b0, b1, kp);
-	f_nround(b1, b0, kp); f_lround(b0, b1, kp);
-
-	out_blk[0] = b0[0]; out_blk[1] = b0[1];
-	out_blk[2] = b0[2]; out_blk[3] = b0[3];
+	/* last round is special */   
+	*((u_int32_t*)temp[0]) = *((u_int32_t*)(b   )) ^ *((u_int32_t*)rk[ROUNDS-1][0]);
+	*((u_int32_t*)temp[1]) = *((u_int32_t*)(b+ 4)) ^ *((u_int32_t*)rk[ROUNDS-1][1]);
+	*((u_int32_t*)temp[2]) = *((u_int32_t*)(b+ 8)) ^ *((u_int32_t*)rk[ROUNDS-1][2]);
+	*((u_int32_t*)temp[3]) = *((u_int32_t*)(b+12)) ^ *((u_int32_t*)rk[ROUNDS-1][3]);
+	b[ 0] = T1[temp[0][0]][1];
+	b[ 1] = T1[temp[1][1]][1];
+	b[ 2] = T1[temp[2][2]][1];
+	b[ 3] = T1[temp[3][3]][1];
+	b[ 4] = T1[temp[1][0]][1];
+	b[ 5] = T1[temp[2][1]][1];
+	b[ 6] = T1[temp[3][2]][1];
+	b[ 7] = T1[temp[0][3]][1];
+	b[ 8] = T1[temp[2][0]][1];
+	b[ 9] = T1[temp[3][1]][1];
+	b[10] = T1[temp[0][2]][1];
+	b[11] = T1[temp[1][3]][1];
+	b[12] = T1[temp[3][0]][1];
+	b[13] = T1[temp[0][1]][1];
+	b[14] = T1[temp[1][2]][1];
+	b[15] = T1[temp[2][3]][1];
+	*((u_int32_t*)(b   )) ^= *((u_int32_t*)rk[ROUNDS][0]);
+	*((u_int32_t*)(b+ 4)) ^= *((u_int32_t*)rk[ROUNDS][1]);
+	*((u_int32_t*)(b+ 8)) ^= *((u_int32_t*)rk[ROUNDS][2]);
+	*((u_int32_t*)(b+12)) ^= *((u_int32_t*)rk[ROUNDS][3]);
+
+	return 0;
 }
 
-/* decrypt a block of text  */
-
-#define i_nround(bo, bi, k) \
-    i_rn(bo, bi, 0, k);     \
-    i_rn(bo, bi, 1, k);     \
-    i_rn(bo, bi, 2, k);     \
-    i_rn(bo, bi, 3, k);     \
-    k -= 4
-
-#define i_lround(bo, bi, k) \
-    i_rl(bo, bi, 0, k);     \
-    i_rl(bo, bi, 1, k);     \
-    i_rl(bo, bi, 2, k);     \
-    i_rl(bo, bi, 3, k)
-
-void
-rijndael_decrypt(rijndael_ctx *ctx, const u4byte *in_blk, u4byte *out_blk)
-{  
-	u4byte  b0[4], b1[4], *kp;
-	u4byte k_len = ctx->k_len;
-	u4byte *e_key = ctx->e_key;
-	u4byte *d_key = ctx->d_key;
-
-	b0[0] = in_blk[0] ^ e_key[4 * k_len + 24]; b0[1] = in_blk[1] ^ e_key[4 * k_len + 25];
-	b0[2] = in_blk[2] ^ e_key[4 * k_len + 26]; b0[3] = in_blk[3] ^ e_key[4 * k_len + 27];
-
-	kp = d_key + 4 * (k_len + 5);
-
-	if(k_len > 6) {
-		i_nround(b1, b0, kp); i_nround(b0, b1, kp);
-	}
-
-	if(k_len > 4) {
-		i_nround(b1, b0, kp); i_nround(b0, b1, kp);
+/**
+ * Decrypt a single block.
+ */
+int
+rijndael_decrypt(rijndael_key *key, u_int8_t a[16], u_int8_t b[16])
+{
+	u_int8_t (*rk)[4][4] = key->keySched;
+	int ROUNDS = key->ROUNDS;
+	int r;
+	u_int8_t temp[4][4];
+	
+	*((u_int32_t*)temp[0]) = *((u_int32_t*)(a   )) ^ *((u_int32_t*)rk[ROUNDS][0]);
+	*((u_int32_t*)temp[1]) = *((u_int32_t*)(a+ 4)) ^ *((u_int32_t*)rk[ROUNDS][1]);
+	*((u_int32_t*)temp[2]) = *((u_int32_t*)(a+ 8)) ^ *((u_int32_t*)rk[ROUNDS][2]);
+	*((u_int32_t*)temp[3]) = *((u_int32_t*)(a+12)) ^ *((u_int32_t*)rk[ROUNDS][3]);
+
+	*((u_int32_t*)(b   )) = *((u_int32_t*)T5[temp[0][0]])
+			      ^ *((u_int32_t*)T6[temp[3][1]])
+			      ^ *((u_int32_t*)T7[temp[2][2]]) 
+			      ^ *((u_int32_t*)T8[temp[1][3]]);
+	*((u_int32_t*)(b+ 4)) = *((u_int32_t*)T5[temp[1][0]])
+			      ^ *((u_int32_t*)T6[temp[0][1]])
+			      ^ *((u_int32_t*)T7[temp[3][2]]) 
+			      ^ *((u_int32_t*)T8[temp[2][3]]);
+	*((u_int32_t*)(b+ 8)) = *((u_int32_t*)T5[temp[2][0]])
+			      ^ *((u_int32_t*)T6[temp[1][1]])
+			      ^ *((u_int32_t*)T7[temp[0][2]]) 
+			      ^ *((u_int32_t*)T8[temp[3][3]]);
+	*((u_int32_t*)(b+12)) = *((u_int32_t*)T5[temp[3][0]])
+			      ^ *((u_int32_t*)T6[temp[2][1]])
+			      ^ *((u_int32_t*)T7[temp[1][2]]) 
+			      ^ *((u_int32_t*)T8[temp[0][3]]);
+	for (r = ROUNDS-1; r > 1; r--) {
+		*((u_int32_t*)temp[0]) = *((u_int32_t*)(b   )) ^ *((u_int32_t*)rk[r][0]);
+		*((u_int32_t*)temp[1]) = *((u_int32_t*)(b+ 4)) ^ *((u_int32_t*)rk[r][1]);
+		*((u_int32_t*)temp[2]) = *((u_int32_t*)(b+ 8)) ^ *((u_int32_t*)rk[r][2]);
+		*((u_int32_t*)temp[3]) = *((u_int32_t*)(b+12)) ^ *((u_int32_t*)rk[r][3]);
+		*((u_int32_t*)(b   )) = *((u_int32_t*)T5[temp[0][0]])
+				      ^ *((u_int32_t*)T6[temp[3][1]])
+				      ^ *((u_int32_t*)T7[temp[2][2]]) 
+				      ^ *((u_int32_t*)T8[temp[1][3]]);
+		*((u_int32_t*)(b+ 4)) = *((u_int32_t*)T5[temp[1][0]])
+				      ^ *((u_int32_t*)T6[temp[0][1]])
+				      ^ *((u_int32_t*)T7[temp[3][2]]) 
+				      ^ *((u_int32_t*)T8[temp[2][3]]);
+		*((u_int32_t*)(b+ 8)) = *((u_int32_t*)T5[temp[2][0]])
+				      ^ *((u_int32_t*)T6[temp[1][1]])
+				      ^ *((u_int32_t*)T7[temp[0][2]]) 
+				      ^ *((u_int32_t*)T8[temp[3][3]]);
+		*((u_int32_t*)(b+12)) = *((u_int32_t*)T5[temp[3][0]])
+				      ^ *((u_int32_t*)T6[temp[2][1]])
+				      ^ *((u_int32_t*)T7[temp[1][2]]) 
+				      ^ *((u_int32_t*)T8[temp[0][3]]);
 	}
+	/* last round is special */   
+	*((u_int32_t*)temp[0]) = *((u_int32_t*)(b   )) ^ *((u_int32_t*)rk[1][0]);
+	*((u_int32_t*)temp[1]) = *((u_int32_t*)(b+ 4)) ^ *((u_int32_t*)rk[1][1]);
+	*((u_int32_t*)temp[2]) = *((u_int32_t*)(b+ 8)) ^ *((u_int32_t*)rk[1][2]);
+	*((u_int32_t*)temp[3]) = *((u_int32_t*)(b+12)) ^ *((u_int32_t*)rk[1][3]);
+	b[ 0] = S5[temp[0][0]];
+	b[ 1] = S5[temp[3][1]];
+	b[ 2] = S5[temp[2][2]];
+	b[ 3] = S5[temp[1][3]];
+	b[ 4] = S5[temp[1][0]];
+	b[ 5] = S5[temp[0][1]];
+	b[ 6] = S5[temp[3][2]];
+	b[ 7] = S5[temp[2][3]];
+	b[ 8] = S5[temp[2][0]];
+	b[ 9] = S5[temp[1][1]];
+	b[10] = S5[temp[0][2]];
+	b[11] = S5[temp[3][3]];
+	b[12] = S5[temp[3][0]];
+	b[13] = S5[temp[2][1]];
+	b[14] = S5[temp[1][2]];
+	b[15] = S5[temp[0][3]];
+	*((u_int32_t*)(b   )) ^= *((u_int32_t*)rk[0][0]);
+	*((u_int32_t*)(b+ 4)) ^= *((u_int32_t*)rk[0][1]);
+	*((u_int32_t*)(b+ 8)) ^= *((u_int32_t*)rk[0][2]);
+	*((u_int32_t*)(b+12)) ^= *((u_int32_t*)rk[0][3]);
+
+	return 0;
+}
 
-	i_nround(b1, b0, kp); i_nround(b0, b1, kp);
-	i_nround(b1, b0, kp); i_nround(b0, b1, kp);
-	i_nround(b1, b0, kp); i_nround(b0, b1, kp);
-	i_nround(b1, b0, kp); i_nround(b0, b1, kp);
-	i_nround(b1, b0, kp); i_lround(b0, b1, kp);
-
-	out_blk[0] = b0[0]; out_blk[1] = b0[1];
-	out_blk[2] = b0[2]; out_blk[3] = b0[3];
+int
+rijndael_makekey(rijndael_key *key, int direction, int keyLen, u_int8_t *keyMaterial)
+{
+	u_int8_t k[RIJNDAEL_MAXKC][4];
+	int i;
+	
+	if (key == NULL)
+		return -1;
+	if ((direction != RIJNDAEL_ENCRYPT) && (direction != RIJNDAEL_DECRYPT))
+		return -1;
+	if ((keyLen != 128) && (keyLen != 192) && (keyLen != 256))
+		return -1;
+
+	key->ROUNDS = keyLen/32 + 6;
+
+	/* initialize key schedule: */
+	for (i = 0; i < keyLen/8; i++)
+		k[i >> 2][i & 3] = (u_int8_t)keyMaterial[i]; 
+
+	rijndael_keysched(k, key->keySched, key->ROUNDS);
+	if (direction == RIJNDAEL_DECRYPT)
+		rijndael_key_enc_to_dec(key->keySched, key->ROUNDS);
+	return 0;
 }