Change AES code to be strict-aliasing-safe.

Makes it give correct results with e.g. gcc 4.4.
For unknown reasons the generate asm code also changes
on e.g. gcc 4.3, making the code a bit larger but also
a bit faster.

Originally committed as revision 23896 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavutil/aes.c

@@ -23,11 +23,18 @@
 #include "common.h"
 #include "aes.h"
 
+typedef union {
+    uint64_t u64[2];
+    uint32_t u32[4];
+    uint8_t u8x4[4][4];
+    uint8_t u8[16];
+} av_aes_block;
+
 typedef struct AVAES{
     // Note: round_key[16] is accessed in the init code, but this only
     // overwrites state, which does not matter (see also r7471).
-    uint8_t round_key[15][4][4];
-    uint8_t state[2][4][4];
+    av_aes_block round_key[15];
+    av_aes_block state[2];
     int rounds;
 }AVAES;
 
@@ -47,18 +54,18 @@ static uint32_t enc_multbl[4][256];
 static uint32_t dec_multbl[4][256];
 #endif
 
-static inline void addkey(uint64_t dst[2], const uint64_t src[2], const uint64_t round_key[2]){
-    dst[0] = src[0] ^ round_key[0];
-    dst[1] = src[1] ^ round_key[1];
+static inline void addkey(av_aes_block *dst, const av_aes_block *src, const av_aes_block *round_key){
+    dst->u64[0] = src->u64[0] ^ round_key->u64[0];
+    dst->u64[1] = src->u64[1] ^ round_key->u64[1];
 }
 
-static void subshift(uint8_t s0[2][16], int s, const uint8_t *box){
-    uint8_t (*s1)[16]= s0[0] - s;
-    uint8_t (*s3)[16]= s0[0] + s;
-    s0[0][0]=box[s0[1][ 0]]; s0[0][ 4]=box[s0[1][ 4]]; s0[0][ 8]=box[s0[1][ 8]]; s0[0][12]=box[s0[1][12]];
-    s1[0][3]=box[s1[1][ 7]]; s1[0][ 7]=box[s1[1][11]]; s1[0][11]=box[s1[1][15]]; s1[0][15]=box[s1[1][ 3]];
-    s0[0][2]=box[s0[1][10]]; s0[0][10]=box[s0[1][ 2]]; s0[0][ 6]=box[s0[1][14]]; s0[0][14]=box[s0[1][ 6]];
-    s3[0][1]=box[s3[1][13]]; s3[0][13]=box[s3[1][ 9]]; s3[0][ 9]=box[s3[1][ 5]]; s3[0][ 5]=box[s3[1][ 1]];
+static void subshift(av_aes_block s0[2], int s, const uint8_t *box){
+    av_aes_block *s1= (av_aes_block *)(s0[0].u8 - s);
+    av_aes_block *s3= (av_aes_block *)(s0[0].u8 + s);
+    s0[0].u8[0]=box[s0[1].u8[ 0]]; s0[0].u8[ 4]=box[s0[1].u8[ 4]]; s0[0].u8[ 8]=box[s0[1].u8[ 8]]; s0[0].u8[12]=box[s0[1].u8[12]];
+    s1[0].u8[3]=box[s1[1].u8[ 7]]; s1[0].u8[ 7]=box[s1[1].u8[11]]; s1[0].u8[11]=box[s1[1].u8[15]]; s1[0].u8[15]=box[s1[1].u8[ 3]];
+    s0[0].u8[2]=box[s0[1].u8[10]]; s0[0].u8[10]=box[s0[1].u8[ 2]]; s0[0].u8[ 6]=box[s0[1].u8[14]]; s0[0].u8[14]=box[s0[1].u8[ 6]];
+    s3[0].u8[1]=box[s3[1].u8[13]]; s3[0].u8[13]=box[s3[1].u8[ 9]]; s3[0].u8[ 9]=box[s3[1].u8[ 5]]; s3[0].u8[ 5]=box[s3[1].u8[ 1]];
 }
 
 static inline int mix_core(uint32_t multbl[][256], int a, int b, int c, int d){
@@ -70,11 +77,11 @@ static inline int mix_core(uint32_t multbl[][256], int a, int b, int c, int d){
 #endif
 }
 
-static inline void mix(uint8_t state[2][4][4], uint32_t multbl[][256], int s1, int s3){
-    ((uint32_t *)(state))[0] = mix_core(multbl, state[1][0][0], state[1][s1  ][1], state[1][2][2], state[1][s3  ][3]);
-    ((uint32_t *)(state))[1] = mix_core(multbl, state[1][1][0], state[1][s3-1][1], state[1][3][2], state[1][s1-1][3]);
-    ((uint32_t *)(state))[2] = mix_core(multbl, state[1][2][0], state[1][s3  ][1], state[1][0][2], state[1][s1  ][3]);
-    ((uint32_t *)(state))[3] = mix_core(multbl, state[1][3][0], state[1][s1-1][1], state[1][1][2], state[1][s3-1][3]);
+static inline void mix(av_aes_block state[2], uint32_t multbl[][256], int s1, int s3){
+    state[0].u32[0] = mix_core(multbl, state[1].u8x4[0][0], state[1].u8x4[s1  ][1], state[1].u8x4[2][2], state[1].u8x4[s3  ][3]);
+    state[0].u32[1] = mix_core(multbl, state[1].u8x4[1][0], state[1].u8x4[s3-1][1], state[1].u8x4[3][2], state[1].u8x4[s1-1][3]);
+    state[0].u32[2] = mix_core(multbl, state[1].u8x4[2][0], state[1].u8x4[s3  ][1], state[1].u8x4[0][2], state[1].u8x4[s1  ][3]);
+    state[0].u32[3] = mix_core(multbl, state[1].u8x4[3][0], state[1].u8x4[s1-1][1], state[1].u8x4[1][2], state[1].u8x4[s3-1][3]);
 }
 
 static inline void crypt(AVAES *a, int s, const uint8_t *sbox, uint32_t multbl[][256]){
@@ -82,29 +89,32 @@ static inline void crypt(AVAES *a, int s, const uint8_t *sbox, uint32_t multbl[]
 
     for(r=a->rounds-1; r>0; r--){
         mix(a->state, multbl, 3-s, 1+s);
-        addkey(a->state[1], a->state[0], a->round_key[r]);
+        addkey(&a->state[1], &a->state[0], &a->round_key[r]);
     }
-    subshift(a->state[0][0], s, sbox);
+    subshift(&a->state[0], s, sbox);
 }
 
-void av_aes_crypt(AVAES *a, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt){
+void av_aes_crypt(AVAES *a, uint8_t *dst_, const uint8_t *src_, int count, uint8_t *iv_, int decrypt){
+    av_aes_block *dst = (av_aes_block *)dst_;
+    const av_aes_block *src = (const av_aes_block *)src_;
+    av_aes_block *iv = (av_aes_block *)iv_;
     while(count--){
-        addkey(a->state[1], src, a->round_key[a->rounds]);
+        addkey(&a->state[1], src, &a->round_key[a->rounds]);
         if(decrypt) {
             crypt(a, 0, inv_sbox, dec_multbl);
             if(iv){
-                addkey(a->state[0], a->state[0], iv);
+                addkey(&a->state[0], &a->state[0], iv);
                 memcpy(iv, src, 16);
             }
-            addkey(dst, a->state[0], a->round_key[0]);
+            addkey(dst, &a->state[0], &a->round_key[0]);
         }else{
-            if(iv) addkey(a->state[1], a->state[1], iv);
+            if(iv) addkey(&a->state[1], &a->state[1], iv);
             crypt(a, 2,     sbox, enc_multbl);
-            addkey(dst, a->state[0], a->round_key[0]);
+            addkey(dst, &a->state[0], &a->round_key[0]);
             if(iv) memcpy(iv, dst, 16);
         }
-        src+=16;
-        dst+=16;
+        src++;
+        dst++;
     }
 }
 
@@ -158,7 +168,7 @@ int av_aes_init(AVAES *a, const uint8_t *key, int key_bits, int decrypt) {
     memcpy(tk, key, KC*4);
 
     for(t= 0; t < (rounds+1)*16;) {
-        memcpy(a->round_key[0][0]+t, tk, KC*4);
+        memcpy(a->round_key[0].u8+t, tk, KC*4);
         t+= KC*4;
 
         for(i = 0; i < 4; i++)
@@ -175,16 +185,16 @@ int av_aes_init(AVAES *a, const uint8_t *key, int key_bits, int decrypt) {
 
     if(decrypt){
         for(i=1; i<rounds; i++){
-            uint8_t tmp[3][16];
-            memcpy(tmp[2], a->round_key[i][0], 16);
+            av_aes_block tmp[3];
+            memcpy(&tmp[2], &a->round_key[i], 16);
             subshift(&tmp[1], 0, sbox);
             mix(tmp, dec_multbl, 1, 3);
-            memcpy(a->round_key[i][0], tmp[0], 16);
+            memcpy(&a->round_key[i], &tmp[0], 16);
         }
     }else{
         for(i=0; i<(rounds+1)>>1; i++){
             for(j=0; j<16; j++)
-                FFSWAP(int, a->round_key[i][0][j], a->round_key[rounds-i][0][j]);
+                FFSWAP(int, a->round_key[i].u8[j], a->round_key[rounds-i].u8[j]);
         }
     }