#include <stdint.h>
#include <string.h>
#include <avr/pgmspace.h>
+#include "memxor.h"
#include "bmw_small.h"
static
uint32_t bmw_small_expand1(uint8_t j, const uint32_t* q, const void* m, const void* h){
uint32_t(*s[])(uint32_t) = {bmw_small_s1, bmw_small_s2, bmw_small_s3, bmw_small_s0};
- uint32_t r=0;
+ uint32_t r;
uint8_t i;
/* r = 0x05555555*(j+16); */
- for(i=0; i<16; ++i){
- r += s[i%4](q[j+i]);
- }
+
#if TWEAK
- r += ( ROTL32(((uint32_t*)m)[j&0xf], ((j+0)&0xf)+1 )
+ r = ( ROTL32(((uint32_t*)m)[j&0xf], ((j+0)&0xf)+1 )
+ ROTL32(((uint32_t*)m)[(j+3)&0xf], ((j+3)&0xf)+1 )
- ROTL32(((uint32_t*)m)[(j+10)&0xf], ((j+10)&0xf)+1 )
+ pgm_read_dword(k_lut+j)
) ^ ((uint32_t*)h)[(j+7)&0xf];
#else
- r += pgm_read_dword(k_lut+j);
+ r = pgm_read_dword(k_lut+j);
r += ((uint32_t*)m)[j&0xf];
r += ((uint32_t*)m)[(j+3)&0xf];
r -= ((uint32_t*)m)[(j+10)&0xf];
#endif
+ for(i=0; i<16; ++i){
+ r += s[i%4](q[j+i]);
+ }
return r;
}
#if F0_HACK==2
/* to understand this implementation take a look at f0-opt-table.txt */
-uint16_t hack_table[5] PROGMEM = { 0x0311, 0xDDB3, 0x2A79, 0x07AA, 0x51C2 };
-uint8_t offset_table[5] PROGMEM = { 4+16, 6+16, 9+16, 12+16, 13+16 };
+static uint16_t hack_table[5] PROGMEM = { 0x0311, 0xDDB3, 0x2A79, 0x07AA, 0x51C2 };
+static uint8_t offset_table[5] PROGMEM = { 4+16, 6+16, 9+16, 12+16, 13+16 };
static
void bmw_small_f0(uint32_t* q, uint32_t* h, const void* m){
}
static
-void bmw_small_f2(uint32_t* h, const uint32_t* q, const void* m){
+void bmw_small_f2(uint32_t* h, uint32_t* q, const void* m){
uint32_t xl=0, xh;
uint8_t i;
for(i=16;i<24;++i){
h[8+i] ^= xh ^ q[24+i];
h[8+i] += ROTL32(h[(4+i)%8],i+9);
}
+/*
h[ 8] += SHL32(xl, 8) ^ q[23] ^ q[ 8];
h[ 9] += SHR32(xl, 6) ^ q[16] ^ q[ 9];
h[10] += SHL32(xl, 6) ^ q[17] ^ q[10];
h[13] += SHR32(xl, 4) ^ q[20] ^ q[13];
h[14] += SHR32(xl, 7) ^ q[21] ^ q[14];
h[15] += SHR32(xl, 2) ^ q[22] ^ q[15];
+*/
+ memxor(q+9, q+16, 7*4);
+ q[8] ^= q[23];
+ h[ 8] += SHL32(xl, 8) ^ q[ 8];
+ h[ 9] += SHR32(xl, 6) ^ q[ 9];
+ h[10] += SHL32(xl, 6) ^ q[10];
+ h[11] += SHL32(xl, 4) ^ q[11];
+ h[12] += SHR32(xl, 3) ^ q[12];
+ h[13] += SHR32(xl, 4) ^ q[13];
+ h[14] += SHR32(xl, 7) ^ q[14];
+ h[15] += SHR32(xl, 2) ^ q[15];
+
}
void bmw_small_nextBlock(bmw_small_ctx_t* ctx, const void* block){