big bug fixed, still some problems with flow control

[avr-crypto-lib.git] / aes_dec.c

diff --git a/aes_dec.c b/aes_dec.c
index 3b546411aba18aa7ae082a8aad7565ee9939213c..8c476de0dbe70d3105eda6412befb75f334e78ba 100644 (file)
--- a/aes_dec.c
+++ b/aes_dec.c
@@ -1,7 +1,7 @@
 /* aes.c */
 /*
-    This file is part of the Crypto-avr-lib/microcrypt-lib.
-    Copyright (C) 2008  Daniel Otte (daniel.otte@rub.de)
+    This file is part of the AVR-Crypto-Lib.
+    Copyright (C) 2008, 2009  Daniel Otte (daniel.otte@rub.de)
 
     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
@@ -34,41 +34,66 @@ void aes_invshiftrow(void* data, uint8_t shift){
        memcpy(data, tmp, 4);
 }
 
+void aes_invshiftcol(void* data, uint8_t shift){
+       uint8_t tmp[4];
+       tmp[0] = ((uint8_t*)data)[ 0];
+       tmp[1] = ((uint8_t*)data)[ 4];
+       tmp[2] = ((uint8_t*)data)[ 8];
+       tmp[3] = ((uint8_t*)data)[12];
+       ((uint8_t*)data)[ 0] = tmp[(4-shift+0)&3];
+       ((uint8_t*)data)[ 4] = tmp[(4-shift+1)&3];
+       ((uint8_t*)data)[ 8] = tmp[(4-shift+2)&3];
+       ((uint8_t*)data)[12] = tmp[(4-shift+3)&3];
+}
 static
 void aes_dec_round(aes_cipher_state_t* state, const aes_roundkey_t* k){
        uint8_t tmp[16];
        uint8_t i;
+       uint8_t t,u,v,w;
        /* keyAdd */
        for(i=0; i<16; ++i){
                tmp[i] = state->s[i] ^ k->ks[i];
        }
        /* mixColums */
        for(i=0; i<4; ++i){
-               state->s[4*0+i] =
-                         gf256mul(0xe, tmp[4*0+i], 0x1b)
-                       ^ gf256mul(0xb, tmp[4*1+i], 0x1b)
-                       ^ gf256mul(0xd, tmp[4*2+i], 0x1b)
-                       ^ gf256mul(0x9, tmp[4*3+i], 0x1b);
-               state->s[4*1+i] =
-                         gf256mul(0x9, tmp[4*0+i], 0x1b)
-                       ^ gf256mul(0xe, tmp[4*1+i], 0x1b)
-                       ^ gf256mul(0xb, tmp[4*2+i], 0x1b)
-                       ^ gf256mul(0xd, tmp[4*3+i], 0x1b);
-               state->s[4*2+i] =
-                         gf256mul(0xd, tmp[4*0+i], 0x1b)
-                       ^ gf256mul(0x9, tmp[4*1+i], 0x1b)
-                       ^ gf256mul(0xe, tmp[4*2+i], 0x1b)
-                       ^ gf256mul(0xb, tmp[4*3+i], 0x1b);
-               state->s[4*3+i] =
-                         gf256mul(0xb, tmp[4*0+i], 0x1b)
-                       ^ gf256mul(0xd, tmp[4*1+i], 0x1b)
-                       ^ gf256mul(0x9, tmp[4*2+i], 0x1b)
-                       ^ gf256mul(0xe, tmp[4*3+i], 0x1b);
+               t = tmp[4*i+3] ^ tmp[4*i+2];
+               u = tmp[4*i+1] ^ tmp[4*i+0];
+               v = t ^ u;
+               v = gf256mul(0x09, v, 0x1b);
+               w = v ^ gf256mul(0x04, tmp[4*i+2] ^ tmp[4*i+0], 0x1b);
+               v = v ^ gf256mul(0x04, tmp[4*i+3] ^ tmp[4*i+1], 0x1b);
+               state->s[4*i+3] = tmp[4*i+3] ^ v ^ gf256mul(0x02, tmp[4*i+0] ^ tmp[4*i+3], 0x1b);
+               state->s[4*i+2] = tmp[4*i+2] ^ w ^ gf256mul(0x02, t, 0x1b);
+               state->s[4*i+1] = tmp[4*i+1] ^ v ^ gf256mul(0x02, tmp[4*i+2] ^ tmp[4*i+1], 0x1b);
+               state->s[4*i+0] = tmp[4*i+0] ^ w ^ gf256mul(0x02, u, 0x1b);
+               
+               /*
+               state->s[4*i+0] =
+                         gf256mul(0xe, tmp[4*i+0], 0x1b)
+                       ^ gf256mul(0xb, tmp[4*i+1], 0x1b)
+                       ^ gf256mul(0xd, tmp[4*i+2], 0x1b)
+                       ^ gf256mul(0x9, tmp[4*i+3], 0x1b);
+               state->s[4*i+1] =
+                         gf256mul(0x9, tmp[4*i+0], 0x1b)
+                       ^ gf256mul(0xe, tmp[4*i+1], 0x1b)
+                       ^ gf256mul(0xb, tmp[4*i+2], 0x1b)
+                       ^ gf256mul(0xd, tmp[4*i+3], 0x1b);
+               state->s[4*i+2] =
+                         gf256mul(0xd, tmp[4*i+0], 0x1b)
+                       ^ gf256mul(0x9, tmp[4*i+1], 0x1b)
+                       ^ gf256mul(0xe, tmp[4*i+2], 0x1b)
+                       ^ gf256mul(0xb, tmp[4*i+3], 0x1b);
+               state->s[4*i+3] =
+                         gf256mul(0xb, tmp[4*i+0], 0x1b)
+                       ^ gf256mul(0xd, tmp[4*i+1], 0x1b)
+                       ^ gf256mul(0x9, tmp[4*i+2], 0x1b)
+                       ^ gf256mul(0xe, tmp[4*i+3], 0x1b);
+               */
        }       
        /* shiftRows */
-       aes_invshiftrow(state->s+4, 1);
-       aes_invshiftrow(state->s+8, 2);
-       aes_invshiftrow(state->s+12, 3);                
+       aes_invshiftcol(state->s+1, 1);
+       aes_invshiftcol(state->s+2, 2);
+       aes_invshiftcol(state->s+3, 3);         
        /* subBytes */
        for(i=0; i<16; ++i){
                state->s[i] = pgm_read_byte(aes_invsbox+state->s[i]);
@@ -84,9 +109,9 @@ void aes_dec_firstround(aes_cipher_state_t* state, const aes_roundkey_t* k){
                state->s[i] ^= k->ks[i];
        }
        /* shiftRows */
-       aes_invshiftrow(state->s+4, 1);
-       aes_invshiftrow(state->s+8, 2);
-       aes_invshiftrow(state->s+12, 3);                
+       aes_invshiftcol(state->s+1, 1);
+       aes_invshiftcol(state->s+2, 2);
+       aes_invshiftcol(state->s+3, 3);         
        /* subBytes */
        for(i=0; i<16; ++i){
                state->s[i] = pgm_read_byte(aes_invsbox+state->s[i]);