#define ROTL32(a,n) (((a)<<(n))|((a)>>(32-(n))))
static
-void quaterround(uint32_t* a, uint32_t* b, uint32_t* c, uint32_t* d){
+void quaterround(uint32_t *a, uint32_t *b, uint32_t *c, uint32_t *d){
*b ^= ROTL32(*a + *d, 7);
*c ^= ROTL32(*b + *a, 9);
*d ^= ROTL32(*c + *b, 13);
}
static
-void rowround(uint32_t* a){
+void rowround(uint32_t *a){
quaterround(a+ 0, a+ 1, a+ 2, a+ 3);
quaterround(a+ 5, a+ 6, a+ 7, a+ 4);
quaterround(a+10, a+11, a+ 8, a+ 9);
}
static
-void columnround(uint32_t* a){
+void columnround(uint32_t *a){
quaterround(a+ 0, a+ 4, a+ 8, a+12);
quaterround(a+ 5, a+ 9, a+13, a+ 1);
quaterround(a+10, a+14, a+ 2, a+ 6);
}
static
-void doubleround(uint32_t* a){
+void doubleround(uint32_t *a){
columnround(a);
rowround(a);
}
-void salsa20_hash(uint32_t* a){
+void salsa20_hash(uint32_t *a){
uint8_t i;
uint32_t b[16];
memcpy(b, a, 64);
const uint8_t sigma[] PROGMEM = {'e','x','p','a','n','d',' ','3','2','-','b','y','t','e',' ','k'};
const uint8_t theta[] PROGMEM = {'e','x','p','a','n','d',' ','1','6','-','b','y','t','e',' ','k'};
-void salsa_k32(uint32_t* dest, const uint32_t* k, const uint32_t* n){
+void salsa_k32(uint32_t *dest, const uint32_t *k, const uint32_t *n){
memcpy_P(dest+ 0, sigma+ 0, 4);
memcpy( dest+ 4, k+ 0, 16);
memcpy_P(dest+20, sigma+ 4, 4);
salsa20_hash(dest);
}
-void salsa_k16(uint32_t* dest, const uint32_t* k, const uint32_t* n){
+void salsa_k16(uint32_t *dest, const uint32_t *k, const uint32_t *n){
memcpy_P(dest+ 0, theta+ 0, 4);
memcpy( dest+ 4, k+ 0, 16);
memcpy_P(dest+20, theta+ 4, 4);
salsa20_hash(dest);
}
-void salsa20_genBlock256(void* dest, const void* k, const void* iv, uint64_t i){
+void salsa20_genBlock256(void *dest, const void *k, const void *iv, uint64_t i){
uint32_t n[4];
memcpy(n, iv, 8);
memcpy(n+8, &i, 8);
salsa_k32((uint32_t*)dest, (uint32_t*)k, n);
}
-void salsa20_genBlock128(void* dest, const void* k, const void* iv, uint64_t i){
+void salsa20_genBlock128(void *dest, const void *k, const void *iv, uint64_t i){
uint32_t n[4];
memcpy(n, iv, 8);
memcpy(n+8, &i, 8);
}
-void salsa20_init(void* key, uint16_t keylength_b, void* iv, salsa20_ctx_t* ctx){
+void salsa20_init(void *key, uint16_t keylength_b, void *iv, salsa20_ctx_t *ctx){
if(keylength_b==256){
- memcpy_P((ctx->a+ 0), sigma+ 0, 4);
- memcpy_P((ctx->a+20), sigma+ 4, 4);
- memcpy_P((ctx->a+40), sigma+ 8, 4);
- memcpy( (ctx->a+44), (uint8_t*)key+16, 16);
- memcpy_P((ctx->a+60), sigma+12, 4);
+ memcpy_P((ctx->a.v8+ 0), sigma+ 0, 4);
+ memcpy_P((ctx->a.v8+20), sigma+ 4, 4);
+ memcpy_P((ctx->a.v8+40), sigma+ 8, 4);
+ memcpy( (ctx->a.v8+44), (uint8_t*)key+16, 16);
+ memcpy_P((ctx->a.v8+60), sigma+12, 4);
}else{
- memcpy_P((ctx->a+ 0), theta+ 0, 4);
- memcpy_P((ctx->a+20), theta+ 4, 4);
- memcpy_P((ctx->a+40), theta+ 8, 4);
- memcpy( (ctx->a+44), (uint8_t*)key+ 0, 16);
- memcpy_P((ctx->a+60), theta+12, 4);
+ memcpy_P((ctx->a.v8+ 0), theta+ 0, 4);
+ memcpy_P((ctx->a.v8+20), theta+ 4, 4);
+ memcpy_P((ctx->a.v8+40), theta+ 8, 4);
+ memcpy( (ctx->a.v8+44), (uint8_t*)key+ 0, 16);
+ memcpy_P((ctx->a.v8+60), theta+12, 4);
}
- memcpy( (ctx->a+ 4), key, 16);
- memset( (ctx->a+24), 0, 16);
+ memcpy( (ctx->a.v8+ 4), key, 16);
+ memset( (ctx->a.v8+24), 0, 16);
if(iv){
- memcpy( (ctx->a+24), iv, 8);
+ memcpy( (ctx->a.v8+24), iv, 8);
}
ctx->buffer_idx=64;
}
-uint8_t salsa20_gen(salsa20_ctx_t* ctx){
+uint8_t salsa20_gen(salsa20_ctx_t *ctx){
if(ctx->buffer_idx==64){
- memcpy(ctx->buffer, ctx->a, 64);
+ memcpy(ctx->buffer, ctx->a.v8, 64);
salsa20_hash((uint32_t*)(ctx->buffer));
- *((uint64_t*)(ctx->a+32)) += 1;
+ ctx->a.v64[4] += 1;
ctx->buffer_idx = 0;
}
return ctx->buffer[ctx->buffer_idx++];