Download cstack/src/cs_sha256.c from Snapkitty/flash-attention-rtl: direct link, hf CLI and curl.
- Browser
- Download file 5.38 kB
-
https://huggingface.co/Snapkitty/flash-attention-rtl/resolve/main/cstack/src/cs_sha256.c
- Command line
-
hf download hf://Snapkitty/flash-attention-rtl/cstack/src/cs_sha256.c
-
curl -L -o cs_sha256.c https://huggingface.co/Snapkitty/flash-attention-rtl/resolve/main/cstack/src/cs_sha256.c
5.38 kB
| 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 uint32_t rotr(uint32_t x, unsigned n) { return (x >> n) | (x << (32 - n)); } | |
| static void compress(uint32_t h[8], const uint8_t blk[64]) | |
| { | |
| uint32_t w[64]; | |
| for (int i = 0; i < 16; i++) | |
| w[i] = ((uint32_t)blk[4 * i] << 24) | ((uint32_t)blk[4 * i + 1] << 16) | | |
| ((uint32_t)blk[4 * i + 2] << 8) | (uint32_t)blk[4 * i + 3]; | |
| for (int i = 16; i < 64; i++) { | |
| uint32_t s0 = rotr(w[i - 15], 7) ^ rotr(w[i - 15], 18) ^ (w[i - 15] >> 3); | |
| uint32_t s1 = rotr(w[i - 2], 17) ^ rotr(w[i - 2], 19) ^ (w[i - 2] >> 10); | |
| w[i] = w[i - 16] + s0 + w[i - 7] + s1; | |
| } | |
| uint32_t a = h[0], b = h[1], c = h[2], d = h[3], e = h[4], f = h[5], g = h[6], hh = h[7]; | |
| for (int i = 0; i < 64; i++) { | |
| uint32_t S1 = rotr(e, 6) ^ rotr(e, 11) ^ rotr(e, 25); | |
| uint32_t ch = (e & f) ^ (~e & g); | |
| uint32_t t1 = hh + S1 + ch + K[i] + w[i]; | |
| uint32_t S0 = rotr(a, 2) ^ rotr(a, 13) ^ rotr(a, 22); | |
| uint32_t mj = (a & b) ^ (a & c) ^ (b & c); | |
| uint32_t t2 = S0 + mj; | |
| hh = g; g = f; f = e; e = d + t1; d = c; c = b; b = a; a = t1 + t2; | |
| } | |
| h[0] += a; h[1] += b; h[2] += c; h[3] += d; h[4] += e; h[5] += f; h[6] += g; h[7] += hh; | |
| } | |
| void cs_sha256_init(cs_sha256_ctx *c) | |
| { | |
| static const uint32_t iv[8] = {0x6a09e667, 0xbb67ae85, 0x3c6ef372, 0xa54ff53a, | |
| 0x510e527f, 0x9b05688c, 0x1f83d9ab, 0x5be0cd19}; | |
| memcpy(c->h, iv, sizeof iv); | |
| c->total = 0; | |
| c->n = 0; | |
| } | |
| void cs_sha256_update(cs_sha256_ctx *c, const void *data, size_t len) | |
| { | |
| const uint8_t *p = data; | |
| c->total += len; | |
| while (len > 0) { | |
| size_t take = 64 - c->n; | |
| if (take > len) take = len; | |
| memcpy(c->buf + c->n, p, take); | |
| c->n += take; p += take; len -= take; | |
| if (c->n == 64) { compress(c->h, c->buf); c->n = 0; } | |
| } | |
| } | |
| void cs_sha256_update_u64(cs_sha256_ctx *c, uint64_t v) | |
| { | |
| uint8_t b[8]; | |
| for (int i = 0; i < 8; i++) b[i] = (uint8_t)(v >> (56 - 8 * i)); | |
| cs_sha256_update(c, b, 8); | |
| } | |
| void cs_sha256_final(cs_sha256_ctx *c, uint8_t out[CS_SHA256_LEN]) | |
| { | |
| uint64_t bits = c->total * 8; | |
| uint8_t pad = 0x80; | |
| cs_sha256_update(c, &pad, 1); | |
| pad = 0; | |
| while (c->n != 56) cs_sha256_update(c, &pad, 1); | |
| uint8_t lenb[8]; | |
| for (int i = 0; i < 8; i++) lenb[i] = (uint8_t)(bits >> (56 - 8 * i)); | |
| cs_sha256_update(c, lenb, 8); | |
| for (int i = 0; i < 8; i++) { | |
| out[4 * i] = (uint8_t)(c->h[i] >> 24); | |
| out[4 * i + 1] = (uint8_t)(c->h[i] >> 16); | |
| out[4 * i + 2] = (uint8_t)(c->h[i] >> 8); | |
| out[4 * i + 3] = (uint8_t)c->h[i]; | |
| } | |
| memset(c, 0, sizeof *c); | |
| } | |
| void cs_sha256(const void *data, size_t len, uint8_t out[CS_SHA256_LEN]) | |
| { | |
| cs_sha256_ctx c; | |
| cs_sha256_init(&c); | |
| cs_sha256_update(&c, data, len); | |
| cs_sha256_final(&c, out); | |
| } | |
| void cs_hmac_sha256(const void *key, size_t klen, const void *msg, size_t mlen, | |
| uint8_t out[CS_SHA256_LEN]) | |
| { | |
| uint8_t k0[64] = {0}, ipad[64], opad[64], inner[CS_SHA256_LEN]; | |
| if (klen > 64) cs_sha256(key, klen, k0); | |
| else if (klen > 0) memcpy(k0, key, klen); | |
| for (int i = 0; i < 64; i++) { ipad[i] = k0[i] ^ 0x36; opad[i] = k0[i] ^ 0x5c; } | |
| cs_sha256_ctx c; | |
| cs_sha256_init(&c); | |
| cs_sha256_update(&c, ipad, 64); | |
| cs_sha256_update(&c, msg, mlen); | |
| cs_sha256_final(&c, inner); | |
| cs_sha256_init(&c); | |
| cs_sha256_update(&c, opad, 64); | |
| cs_sha256_update(&c, inner, CS_SHA256_LEN); | |
| cs_sha256_final(&c, out); | |
| } | |
| int cs_ct_equal(const void *a, const void *b, size_t n) | |
| { | |
| const volatile uint8_t *x = a, *y = b; | |
| uint8_t d = 0; | |
| for (size_t i = 0; i < n; i++) d |= (uint8_t)(x[i] ^ y[i]); | |
| return d == 0; | |
| } | |
| void cs_hex(const uint8_t *in, size_t n, char *out) | |
| { | |
| static const char *digits = "0123456789abcdef"; | |
| for (size_t i = 0; i < n; i++) { | |
| out[2 * i] = digits[in[i] >> 4]; | |
| out[2 * i + 1] = digits[in[i] & 15]; | |
| } | |
| out[2 * n] = '\0'; | |
| } | |
| static int nib(char c) | |
| { | |
| if (c >= '0' && c <= '9') return c - '0'; | |
| if (c >= 'a' && c <= 'f') return c - 'a' + 10; | |
| if (c >= 'A' && c <= 'F') return c - 'A' + 10; | |
| return -1; | |
| } | |
| int cs_unhex(const char *hex, uint8_t *out, size_t n) | |
| { | |
| for (size_t i = 0; i < n; i++) { | |
| int hi = nib(hex[2 * i]); | |
| if (hi < 0) return -1; | |
| int lo = nib(hex[2 * i + 1]); | |
| if (lo < 0) return -1; | |
| out[i] = (uint8_t)((hi << 4) | lo); | |
| } | |
| return hex[2 * n] == '\0' ? 0 : -1; | |
| } | |