Related
I want to write my own png reader using little help from other libraries. So far, I can read back images with only one IDAT chunk perfectly well. However, when I need to concatenate multiple chunks, the file ends up corrupted. Here is an example of one:
I have a lot of code, so I'm going to try and get it to as 'minimum reproducable example' as I can get, but this along with the fact I do not know specifically what part is causing it, is going to make it a bit longer than I'm sure some of you might like, and I'm sorry.
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "zlib.h"
unsigned char IEND_CHUNK[4] = {'I', 'E', 'N', 'D'}; //Indicates End Of File
unsigned char IDAT_CHUNK[4] = {'I', 'D', 'A', 'T'}; //Holds Pixel Information
typedef struct Chunk Chunk;
typedef struct IHDR IHDR;
struct Chunk {
int length;
unsigned char type[4];
unsigned char *data;
unsigned char crc[4];
};
struct IHDR {
int width;
int height;
int bitd;
int colort;
int compm;
int filterm;
int interlacem;
int channels;
int bytesPerPixel;
};
int bytesToInt(unsigned char a, unsigned char b, unsigned char c, unsigned char d)
{
return (int) a << 24 | b << 16 | c << 8 | d;
}
int compType(unsigned char *a, unsigned char b[], int len)
{
int equal = 1;
for(int i = 0; i < len; i++) equal = a[i] == b[i] ? equal : 0;
return equal;
}
Chunk* getChunksFromBytes(unsigned char* bytes)
{
int next_seg = 7;
int chunks = 0;
long int size = 1;
Chunk* temp = (Chunk*) malloc(sizeof(Chunk));
while(1){
size += sizeof(Chunk);
temp = realloc(temp, size);
temp[chunks].length = bytesToInt(bytes[next_seg+1], bytes[next_seg+2],
bytes[next_seg+3], bytes[next_seg+4]);
temp[chunks].type[0] = bytes[next_seg+5];
temp[chunks].type[1] = bytes[next_seg+6];
temp[chunks].type[2] = bytes[next_seg+7];
temp[chunks].type[3] = bytes[next_seg+8];
temp[chunks].data = malloc(temp[chunks].length);
if(temp[chunks].length > 0){
memcpy(temp[chunks].data, bytes+next_seg+9, temp[chunks].length);
}
else temp[chunks].data = NULL;
temp[chunks].crc[0] = bytes[next_seg+temp[chunks].length+9];
temp[chunks].crc[1] = bytes[next_seg+temp[chunks].length+10];
temp[chunks].crc[2] = bytes[next_seg+temp[chunks].length+11];
temp[chunks].crc[3] = bytes[next_seg+temp[chunks].length+12];
if(compType(temp[chunks].type, IEND_CHUNK, 4)) break;
next_seg+=temp[chunks].length+12;
chunks++;
}
return temp;
}
IHDR getIHDRFromChunks(Chunk* chunks)
{
IHDR img_info;
int channels[7] = {1, 0, 3, 1, 2, 0, 4};
img_info.width = bytesToInt(chunks[0].data[0], chunks[0].data[1],
chunks[0].data[2], chunks[0].data[3]);
img_info.height = bytesToInt(chunks[0].data[4], chunks[0].data[5],
chunks[0].data[6], chunks[0].data[7]);
img_info.bitd = (int) chunks[0].data[8];
img_info.colort = (int) chunks[0].data[9];
img_info.compm = (int) chunks[0].data[10];
img_info.filterm = (int) chunks[0].data[11];
img_info.interlacem = (int) chunks[0].data[12];
img_info.channels = channels[img_info.colort];
img_info.bytesPerPixel = (int) img_info.channels * (img_info.bitd / 8);
return img_info;
}
unsigned char* getImgFromChunks(Chunk* chunks)
{
int count = 0;
IHDR ihdr_data = getIHDRFromChunks(chunks);
uLongf compressed_size = 0;
uLongf uncompressed_size = (ihdr_data.width*ihdr_data.height*ihdr_data.bytesPerPixel) + ihdr_data.height + 1;
unsigned char* compressed_idat = (unsigned char*) malloc(sizeof(unsigned char)*4);
unsigned char* uncompressed_idat = (unsigned char*) malloc(sizeof(unsigned char)*uncompressed_size);
while(1){
if(compType(chunks[count].type, IEND_CHUNK, 4)) break;
else if (compType(chunks[count].type, IDAT_CHUNK, 4)){
compressed_size += sizeof(unsigned char)*chunks[count].length;
compressed_idat = realloc(compressed_idat, compressed_size);
memcpy(compressed_idat + compressed_size - chunks[count].length, chunks[count].data, chunks[count].length);
}
count++;
}
int ret = uncompress(uncompressed_idat, &uncompressed_size, compressed_idat, compressed_size);
free(compressed_idat);
return ret == 0 ? uncompressed_idat : '\0';
}
int PaethPredictor(int a, int b, int c)
{
int pa = abs(b - c);
int pb = abs(a - c);
int pc = abs(a + b - (2*c));
if(pa <= pb && pa <= pc) return a;
else if(pb <= pc) return b;
return c;
}
int* getPixelsFromImg(unsigned char* img, IHDR ihdr_info)
{
int* pixels = (int*) malloc(sizeof(int)*ihdr_info.width*ihdr_info.height*ihdr_info.bytesPerPixel);
int filter_type;
int i = 0;
int current_pixel = 0;
int stride = ihdr_info.width * ihdr_info.bytesPerPixel;
int filt_a;
int filt_b;
int filt_c;
for(int r = 0; r < ihdr_info.height; r++){
filter_type = img[i];
i++;
for(int c = 0; c < stride; c++){
filt_a = c >= ihdr_info.bytesPerPixel ? pixels[r * stride + c - ihdr_info.bytesPerPixel] : 0;
filt_b = r > 0 ? pixels[(r - 1) * stride + c] : 0;
filt_c = r > 0 && c >= ihdr_info.bytesPerPixel ? pixels[(r - 1) * stride + c - ihdr_info.bytesPerPixel] : 0;
switch(filter_type){
case 0:
pixels[current_pixel] = img[i] & 0xff;
break;
case 1:
pixels[current_pixel] = (img[i] + filt_a) & 0xff;
break;
case 2:
pixels[current_pixel] = (img[i] + filt_b) & 0xff;
break;
case 3:
pixels[current_pixel] = (img[i] + filt_a + filt_c) & 0xff;
break;
case 4:
pixels[current_pixel] = (img[i] + PaethPredictor(filt_a, filt_b, filt_c)) & 0xff;
break;
}
current_pixel++;
i++;
}
}
return pixels;
}
These are the combinations of funtcions i use to achieve this. I first use a separate function which I know is not the problem and have therefore left out to get the bytes of the png file. I then call getChunksFromBytes() to gather the chunks of the png file, and I separate the IHDR from these chunks into a separate struct. Finally, I call getImgFromChunks() to uncompress the IDAT data, and getPixelsFromImg() to unfilter the uncompressed values.
First off, bytes per pixel is not a thing since pixels can be bits packed into bytes. So your calculation of uncompressed_size in getImgFromChunks() is wrong. It should be:
uLongf uncompressed_size = ihdr_data.height *
(1 + ((ihdr_data.bitd * ihdr_data.channels * (uLongf)ihdr_data.width + 7) >> 3));
Given that bits per pixel is not a thing, you need to rethink how you are doing getPixelsFromImg(). So there's no point in digging further here until you do that.
There is nothing wrong with your processing of multiple IDAT chunks. You may have misidentified the salient difference of the png file that is giving you issues.
I've come across this answer.
But I hit 2 issues, I'm not sure how to adapt it to pack 64 bits (see below), and I can't get my head around unpacking them.
Here is what I have for packing:
const int i = 1;
#define is_bigendian() ((*(char *) &i) == 0)
#define MAGIC (is_bigendian() ? 0x0102040810204080 : 0x8040201008040201)
inline uint8_t pack8b(bool *a) {
uint64_t t = *((uint64_t *) a);
return (MAGIC * t >> 56) & 0xFF;
}
uint32_t pack32b(bool *a) {
return (pack8b(a + 0) << 24) | (pack8b(a + 8) << 16) |
(pack8b(a + 16) << 8) | (pack8b(a + 24) << 0);
}
uint64_t pack64b(bool *a) {
return ((uint64_t) pack32b(a) << 32) | pack32b(a + 32);
}
EDIT: Here is what I have so far that works like a charm, i'm open to suggestion for performance improvement:
uint64_t MAGIC() {
static const int i = 1;
// Takes of little / big endian
bool bigEndian = ((*(char *) &i) == 0);
return bigEndian ? 0x0102040810204080 : 0x8040201008040201;
}
uint8_t pack8b(bool *a) {
uint64_t t = *((uint64_t *) a);
return (MAGIC() * t >> 56) & 0xFF;
}
uint16_t pack16b(bool *a) {
return ((uint16_t) pack8b(a + 0) << 8) | (pack8b(a + 8) << 0);
}
uint32_t pack32b(bool *a) {
return ((uint32_t) pack16b(a + 0) << 16) | (pack16b(a + 16) << 0);
}
uint64_t pack64b(bool *a) {
return ((uint64_t) pack32b(a + 0) << 32) | (pack32b(a + 32) << 0);
}
void unpack8b(bool *a, uint8_t v) {
uint64_t mask = 0x8080808080808080ULL;
*((uint64_t *) a) = ((MAGIC() * v) & mask) >> 7;
}
void unpack16b(bool *a, uint16_t v) {
unpack8b(a + 0, v >> 8);
unpack8b(a + 8, v >> 0);
}
void unpack32b(bool *a, uint32_t v) {
unpack16b(a + 0, v >> 16);
unpack16b(a + 16, v >> 0);
}
void unpack64b(bool *a, uint64_t v) {
unpack32b(a + 0, v >> 32);
unpack32b(a + 32, v >> 0);
}
Your solution works for me I tested it with the gcc on tio: Try it online!
The simplest function for unpacking the uint64_t would be:
void unpack64b(uint64_t num, bool* bit_field) {
for (int i = 63; 0 <= i; i--) {
bit_field[i] = num & 1;
num = num >> 1;
}
}
If bit_field is initialized to all zeros then early exiting is possible:
void unpack64b(uint64_t num, bool* bit_field) {
for (int i = 63; num > 0; i--) {
bit_field[i] = num & 1;
num = num >> 1;
}
}
To the packing for 32 and 64 bit, shouldn't this also be aware of the endian encoding?
I can't get my head around unpacking them.
For starters, concentrate on just iterating the bits, do a simplest method that works. For me that was just simple array assignments with bit-and mask:
#include <stdbool.h>
#include <stdint.h>
#include <string.h>
#include <assert.h>
#include <stdio.h>
#include <stdlib.h>
#include <limits.h>
static const int i = 1;
uint64_t MAGIC(void) {
return ((*(const char *)&i) == 0) ? 0x0102040810204080ull : 0x8040201008040201ull;
}
void unpack8b(bool a[8], uint8_t v) {
if (((*(const char *)&i) == 0)) {
a[0] = v & 0x01;
a[1] = v & 0x02;
a[2] = v & 0x04;
a[3] = v & 0x08;
a[4] = v & 0x10;
a[5] = v & 0x20;
a[6] = v & 0x40;
a[7] = v & 0x80;
} else {
a[7] = v & 0x01;
a[6] = v & 0x02;
a[5] = v & 0x04;
a[4] = v & 0x08;
a[3] = v & 0x10;
a[2] = v & 0x20;
a[1] = v & 0x40;
a[0] = v & 0x80;
}
}
void unpack16b(bool a[16], uint16_t v) {
unpack8b(&a[0], v >> 8);
unpack8b(&a[8], v);
}
void unpack32b(bool a[32], uint32_t v) {
unpack16b(&a[0], v >> 16);
unpack16b(&a[16], v);
}
void unpack64b(bool a[64], uint64_t v) {
unpack32b(&a[0], v >> 32);
unpack32b(&a[32], v);
}
uint8_t pack8b(bool a[8]) {
static_assert(sizeof(bool) == 1, "");
static_assert(sizeof(uint64_t) == 8, "");
static_assert(CHAR_BIT == 8, "");
uint64_t t;
memcpy(&t, a, sizeof(t));
return (MAGIC() * t >> 56) & 0xFF;
}
uint16_t pack16b(bool a[16]) {
return pack8b(&a[0]) << 8 | pack8b(&a[8]);
}
uint32_t pack32b(bool a[32]) {
return (uint32_t)pack16b(&a[0]) << 16 | pack16b(&a[16]);
}
uint64_t pack64b(bool a[64]) {
return ((uint64_t)pack32b(&a[0]) << 32) | pack32b(&a[32]);
}
int main() {
_Alignas(uint64_t) bool a[64];
for (int i = 0; i < 64; ++i) {
a[i] = rand() % 2;
}
for (int i = 0; i < 64; ++i) printf("%d%s", a[i], !((i+1)%8)?" ":"");
printf("\n");
uint64_t v = pack64b(a);
printf("%llx\n", v);
memset(a, -1, sizeof(a));
unpack64b(a, v);
for (int i = 0; i < 64; ++i) printf("%d%s", a[i], !((i+1)%8)?" ":"");
printf("\n");
}
Outputs:
10111100 11010110 00001011 00011110 00111010 11110100 10101001 00011101
bcd60b1e3af4a91d
10111100 11010110 00001011 00011110 00111010 11110100 10101001 00011101
Note that when int on a platform has 16 bits, the (pack8b(a + 0) << 24) in your code is undefined behavior. Cast it to type with more bits then 24 (uint32_t)pack8b(a + 0) << 24 before shifting to be extra safe.
I have a problem and hope you can help me.
I have a function written in C that returns hash a value. My
headache is when I execute the program from another tool it takes a lot of time to run, probably because inside my function I run a command that hashes my value in SHA256, so I would like to know if there is another way to do it, maybe a function or something like that.
Here is what I have:
const char *EncryptSHA256 (char *Arg1) {
char command[128];
char result[512];
//I want to replace from here
snprintf(command, sizeof command, "echo -n %s | sha256sum | cut -c1-64",Arg1);
FILE *fpipe;
if (0 == (fpipe = (FILE*)popen(command, "r"))) {
perror("popen() failed.");
exit(1);
}
fread(result, 1, 512, fpipe);
pclose(fpipe);
const char *sha256 = &result[0];
//to here
return sha256;
}
Your code has undefined behavior because you return a pointer to result, a local array with automatic storage. Reading from this array by the caller has undefined behavior.
You should at least make result static so its contents remain readable after EncryptSHA256 returns to its caller.
Regarding the inefficiency of the method, here is a public domain implementation of SHA256 that you can use directly inside your program:
/* public domain sha256 implementation based on fips180-3 */
#include <stddef.h>
#include <stdint.h>
#include <string.h>
/* Public API */
struct sha256 {
uint64_t len; /* processed message length */
uint32_t h[8]; /* hash state */
uint8_t buf[64]; /* message block buffer */
};
/* reset state */
void sha256_init(struct sha256 *s);
/* process message */
void sha256_update(struct sha256 *s, const void *m, size_t len);
/* get message digest */
/* state is ruined after sum, keep a copy if multiple sum is needed */
/* part of the message might be left in s, zero it if secrecy is needed */
void sha256_sum(struct sha256 *s, uint8_t md[32]);
/* Implementation */
static uint32_t ror(uint32_t n, int k) {
return (n >> k) | (n << (32 - k));
}
#define Ch(x,y,z) (z ^ (x & (y ^ z)))
#define Maj(x,y,z) ((x & y) | (z & (x | y)))
#define S0(x) (ror(x,2) ^ ror(x,13) ^ ror(x,22))
#define S1(x) (ror(x,6) ^ ror(x,11) ^ ror(x,25))
#define R0(x) (ror(x,7) ^ ror(x,18) ^ (x>>3))
#define R1(x) (ror(x,17) ^ ror(x,19) ^ (x>>10))
static const uint32_t K[64] = {
0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5,
0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5,
0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3,
0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174,
0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc,
0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da,
0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7,
0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967,
0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13,
0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85,
0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3,
0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070,
0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5,
0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3,
0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208,
0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
};
static void processblock(struct sha256 *s, const uint8_t *buf) {
uint32_t W[64], t1, t2, a, b, c, d, e, f, g, h;
int i;
for (i = 0; i < 16; i++) {
W[i] = (uint32_t)buf[4 * i + 0] << 24;
W[i] |= (uint32_t)buf[4 * i + 1] << 16;
W[i] |= (uint32_t)buf[4 * i + 2] << 8;
W[i] |= buf[4 * i + 3];
}
for (; i < 64; i++)
W[i] = R1(W[i-2]) + W[i-7] + R0(W[i-15]) + W[i-16];
a = s->h[0];
b = s->h[1];
c = s->h[2];
d = s->h[3];
e = s->h[4];
f = s->h[5];
g = s->h[6];
h = s->h[7];
#define ROUND(a,b,c,d,e,f,g,h,i) \
t1 = h + S1(e) + Ch(e,f,g) + K[i] + W[i]; \
t2 = S0(a) + Maj(a,b,c); \
d += t1; \
h = t1 + t2;
for (i = 0; i < 64; ) {
ROUND(a, b, c, d, e, f, g, h, i); i++;
ROUND(h, a, b, c, d, e, f, g, i); i++;
ROUND(g, h, a, b, c, d, e, f, i); i++;
ROUND(f, g, h, a, b, c, d, e, i); i++;
ROUND(e, f, g, h, a, b, c, d, i); i++;
ROUND(d, e, f, g, h, a, b, c, i); i++;
ROUND(c, d, e, f, g, h, a, b, i); i++;
ROUND(b, c, d, e, f, g, h, a, i); i++;
}
#undef ROUND
s->h[0] += a;
s->h[1] += b;
s->h[2] += c;
s->h[3] += d;
s->h[4] += e;
s->h[5] += f;
s->h[6] += g;
s->h[7] += h;
}
static void pad(struct sha256 *s) {
unsigned r = s->len % 64;
s->buf[r++] = 0x80;
if (r > 56) {
memset(s->buf + r, 0, 64 - r);
r = 0;
processblock(s, s->buf);
}
memset(s->buf + r, 0, 56 - r);
s->len *= 8;
s->buf[56] = s->len >> 56;
s->buf[57] = s->len >> 48;
s->buf[58] = s->len >> 40;
s->buf[59] = s->len >> 32;
s->buf[60] = s->len >> 24;
s->buf[61] = s->len >> 16;
s->buf[62] = s->len >> 8;
s->buf[63] = s->len;
processblock(s, s->buf);
}
void sha256_init(struct sha256 *s) {
s->len = 0;
s->h[0] = 0x6a09e667;
s->h[1] = 0xbb67ae85;
s->h[2] = 0x3c6ef372;
s->h[3] = 0xa54ff53a;
s->h[4] = 0x510e527f;
s->h[5] = 0x9b05688c;
s->h[6] = 0x1f83d9ab;
s->h[7] = 0x5be0cd19;
}
void sha256_sum(struct sha256 *s, uint8_t md[20]) {
int i;
pad(s);
for (i = 0; i < 8; i++) {
md[4 * i + 0] = s->h[i] >> 24;
md[4 * i + 1] = s->h[i] >> 16;
md[4 * i + 2] = s->h[i] >> 8;
md[4 * i + 3] = s->h[i];
}
}
void sha256_update(struct sha256 *s, const void *m, unsigned long len) {
const uint8_t *p = m;
unsigned r = s->len % 64;
s->len += len;
if (r) {
if (len < 64 - r) {
memcpy(s->buf + r, p, len);
return;
}
memcpy(s->buf + r, p, 64 - r);
len -= 64 - r;
p += 64 - r;
processblock(s, s->buf);
}
for (; len >= 64; len -= 64, p += 64)
processblock(s, p);
memcpy(s->buf, p, len);
}
You would change your function to this:
const char *EncryptSHA256(char *Arg1) {
struct sha256 s;
unsigned char md[32];
static char result[65];
sha256_init(&s);
sha256_update(&s, Arg1, strlen(Arg1));
sha256_sum(&s, md);
for (int i = 0; i < 32; i++) {
sprintf(result + i * 2, "%02x", md[i]);
}
return result;
}
You could also change the API to pass an array of 32 unsigned characters to get the binary form if it is more convenient.
I've implemented sha2 256 based on the rfc spec found at: https://www.rfc-editor.org/rfc/rfc4634.
My implementation:
#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <errno.h>
uint32_t K[] = {
0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5,
0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5,
0xd807aa98,0x12835b01,0x243185be,0x550c7dc3,
0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174,
0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc,
0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da,
0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7,
0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967,
0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13,
0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85,
0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3,
0xd192e819,0xd6990624,0xf40e3585,0x106aa070,
0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5,
0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3,
0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208,
0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2,
};
uint32_t CH(uint32_t x, uint32_t y, uint32_t z) {
uint32_t answ = (x & y) ^ (~x & z);
return answ;
}
uint32_t MAJ(uint32_t x, uint32_t y, uint32_t z) {
uint32_t answ = (x & y) ^ (x & z) ^ (y & z);
return answ;
}
uint32_t ROTL(uint32_t x, short n) {
return (x << n) | (x>>(32 - n));
}
uint32_t ROTR(uint32_t x, short n) {
return (x >> n) | (x<<(32 - n));
}
uint32_t BIGS0(uint32_t x) {
return ROTR(x,2) ^ ROTR(x,13) ^ ROTR(x,22);
}
uint32_t BIGS1(uint32_t x) {
return ROTR(x,6) ^ ROTR(x,11) ^ ROTR(x,25);
}
uint32_t SSIG0(uint32_t x) {
return ROTR(x,7) ^ ROTR(x,18) ^ (x >> 3);
}
uint32_t SSIG1(uint32_t x) {
return ROTR(x,17) ^ ROTR(x,19) ^ (x >> 10);
}
uint32_t toInt(uint8_t *t) {
return (t[3] << 24) | (t[2] << 16) | (t[1] << 8) | t[0];
}
void process(uint32_t *block) {
uint32_t H[] = {
0x6a09e667,
0xbb67ae85,
0x3c6ef372,
0xa54ff53a,
0x510e527f,
0x9b05688c,
0x1f83d9ab,
0x5be0cd19
};
uint32_t *W = (uint32_t *) malloc (sizeof(uint32_t*) * 64);
for (int t = 0; t < 16; t++) {
W[t] = block[t];
}
for (int t = 16; t < 64; t++) {
W[t] = SSIG1(W[t-2]) + W[t-7] + SSIG0(t-15) + W[t-16];
}
uint32_t a = H[0];
uint32_t b = H[1];
uint32_t c = H[2];
uint32_t d = H[3];
uint32_t e = H[4];
uint32_t f = H[5];
uint32_t g = H[6];
uint32_t h = H[7];
uint32_t T1 = 0;
uint32_t T2 = 0;
for(int t =0; t < 64 ; t++) {
T1 = h + BIGS1(e) + CH(e,f,g) + K[t] + W[t];
T2 = BIGS0(a) + MAJ(a,b,c);
h = g;
g = f;
f = e;
e = d + T1;
d = c;
c = b;
b = a;
a = T1 + T2;
}
H[0] = a + H[0];
H[1] = b + H[1];
H[2] = c + H[2];
H[3] = d + H[3];
H[4] = e + H[4];
H[5] = f + H[5];
H[6] = g + H[6];
H[7] = h + H[7];
for (int j = 0; j < 8; j++ ) {
printf("%08x", H[j] );
}
free(W);
}
int main(void)
{
uint32_t block[] = {
0x61626380,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000000,
0x00000018,
};//abc in 512 bit padded block
process(block);
return 0;
}
For now I'm only doing one 512 message block and my testing message is 'abc'. Doing all the preprocessing as required results in a padded message block as follows:
0x61626380 0x00000000 0x00000000 0x00000000
0x00000000 0x00000000 0x00000000 0x00000000
0x00000000 0x00000000 0x00000000 0x00000000
0x00000000 0x00000000 0x00000000 0x00000018
After feeding it through my implementation I get the hash as: 4b9cc43100a30340dbc8f2328e2c80a91fdbd7b8cd20962d1b64e31283c4b99d
Where as the correct hash is:
ba7816bf8f01cfea414140de5dae2223b00361a396177a9cb410ff61f20015ad
Tracing my output and comparing it to the example provided at http://csrc.nist.gov/publications/fips/fips180-2/fips180-2withchangenotice.pdf page 34. My implementation seems to go wrong at iteration 16 (thus t=16) when the variable e needs to be calculated.
Any help would be appreciated.
This line:
W[t] = SSIG1(W[t-2]) + W[t-7] + SSIG0(t-15) + W[t-16];
Should be
W[t] = SSIG1(W[t-2]) + W[t-7] + SSIG0(W[t-15]) + W[t-16];
The problem kicks in on iteration 16 because that's when you first refer to a value buffered using that assignment (earlier values were just copies of the initial block).
http://en.wikipedia.org/wiki/Multiply-with-carry#Complementary-multiply-with-carry_RNGs
has some code in C
I'm trying to compile it, but when I do I get this error
c:\program files (x86)\codeblocks\mingw\bin..\lib\gcc\mingw32\4.4.1......\libmingw32.a(main.o):main.c|| undefined reference to `WinMain#16'|
Here's the code
I hope to figure out how to implement it and use it instead of booster's mersenne twister.
As someone pointed out, I was missing my main function, but now with it, I'm still having an issue getting it to run.
#include <stdint.h>
#define PHI 0x9e3779b9
static uint32_t Q[4096], c = 362436;
int main
{
init_rand(6);
int c = rand_cmwc();
};
void init_rand(uint32_t x)
{
int i;
Q[0] = x;
Q[1] = x + PHI;
Q[2] = x + PHI + PHI;
for (i = 3; i < 4096; i++)
Q[i] = Q[i - 3] ^ Q[i - 2] ^ PHI ^ i;
}
uint32_t rand_cmwc(void)
{
uint64_t t, a = 18782LL;
static uint32_t i = 4095;
uint32_t x, r = 0xfffffffe;
i = (i + 1) & 4095;
t = a * Q[i] + c;
c = (t >> 32);
x = t + c;
if (x < c) {
x++;
c++;
}
return (Q[i] = r - x);
}
Okay, I figured it out.
I was missing my main function.
I then implemented the code as a class, in hopes to make it thread safe (allowing for multiple seeds)
#include <iostream>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <time.h>
#define PHI 0x9e3779b9
using namespace std;
static uint32_t Q[4096], c = 362436;
class cmwc
{
public:
void init_rand(uint32_t x)
{
int i;
Q[0] = x;
Q[1] = x + PHI;
Q[2] = x + PHI + PHI;
for (i = 3; i < 4096; i++)
Q[i] = Q[i - 3] ^ Q[i - 2] ^ PHI ^ i;
}
uint32_t rand_cmwc(void)
{
uint64_t t, a = 18782LL;
static uint32_t i = 4095;
uint32_t x, r = 0xfffffffe;
i = (i + 1) & 4095;
t = a * Q[i] + c;
c = (t >> 32);
x = t + c;
if (x < c) {
x++;
c++;
}
return (Q[i] = r - x);
}
uint32_t random_num;
};
int main()
{
cmwc rng1;
rng1.init_rand(time(NULL));
rng1.random_num = 1 + rng1.rand_cmwc()%6;
//int r = rng1.rand_cmwc();
cout << "die: " << rng1.random_num << endl;
return 0;
}