From fe38379d0d0eebefa97cd89bacc26820f369f22f Mon Sep 17 00:00:00 2001 From: jsteube Date: Mon, 17 Jul 2017 08:29:46 +0200 Subject: [PATCH] Add -L kernel for SHA224 --- OpenCL/inc_hash_sha224.cl | 1882 +++++++++++++++++++++++++++++++++++++ OpenCL/m01300_a0-pure.cl | 134 +++ OpenCL/m01300_a1-pure.cl | 110 +++ OpenCL/m01300_a3-pure.cl | 144 +++ 4 files changed, 2270 insertions(+) create mode 100644 OpenCL/inc_hash_sha224.cl create mode 100644 OpenCL/m01300_a0-pure.cl create mode 100644 OpenCL/m01300_a1-pure.cl create mode 100644 OpenCL/m01300_a3-pure.cl diff --git a/OpenCL/inc_hash_sha224.cl b/OpenCL/inc_hash_sha224.cl new file mode 100644 index 000000000..538d8a515 --- /dev/null +++ b/OpenCL/inc_hash_sha224.cl @@ -0,0 +1,1882 @@ + +// important notes on this: +// input buf unused bytes needs to be set to zero +// input buf needs to be in algorithm native byte order (md5 = LE, sha256 = BE, etc) +// input buf needs to be 64 byte aligned when using md5_update() + +__constant u32a k_sha256[64] = +{ + SHA256C00, SHA256C01, SHA256C02, SHA256C03, + SHA256C04, SHA256C05, SHA256C06, SHA256C07, + SHA256C08, SHA256C09, SHA256C0a, SHA256C0b, + SHA256C0c, SHA256C0d, SHA256C0e, SHA256C0f, + SHA256C10, SHA256C11, SHA256C12, SHA256C13, + SHA256C14, SHA256C15, SHA256C16, SHA256C17, + SHA256C18, SHA256C19, SHA256C1a, SHA256C1b, + SHA256C1c, SHA256C1d, SHA256C1e, SHA256C1f, + SHA256C20, SHA256C21, SHA256C22, SHA256C23, + SHA256C24, SHA256C25, SHA256C26, SHA256C27, + SHA256C28, SHA256C29, SHA256C2a, SHA256C2b, + SHA256C2c, SHA256C2d, SHA256C2e, SHA256C2f, + SHA256C30, SHA256C31, SHA256C32, SHA256C33, + SHA256C34, SHA256C35, SHA256C36, SHA256C37, + SHA256C38, SHA256C39, SHA256C3a, SHA256C3b, + SHA256C3c, SHA256C3d, SHA256C3e, SHA256C3f, +}; + +typedef struct sha224_ctx +{ + u32 h[8]; + + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int len; + +} sha224_ctx_t; + +void sha224_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[8]) +{ + u32 a = digest[0]; + u32 b = digest[1]; + u32 c = digest[2]; + u32 d = digest[3]; + u32 e = digest[4]; + u32 f = digest[5]; + u32 g = digest[6]; + u32 h = digest[7]; + + u32 w0_t = w0[0]; + u32 w1_t = w0[1]; + u32 w2_t = w0[2]; + u32 w3_t = w0[3]; + u32 w4_t = w1[0]; + u32 w5_t = w1[1]; + u32 w6_t = w1[2]; + u32 w7_t = w1[3]; + u32 w8_t = w2[0]; + u32 w9_t = w2[1]; + u32 wa_t = w2[2]; + u32 wb_t = w2[3]; + u32 wc_t = w3[0]; + u32 wd_t = w3[1]; + u32 we_t = w3[2]; + u32 wf_t = w3[3]; + + #define ROUND_EXPAND_S() \ + { \ + w0_t = SHA256_EXPAND_S (we_t, w9_t, w1_t, w0_t); \ + w1_t = SHA256_EXPAND_S (wf_t, wa_t, w2_t, w1_t); \ + w2_t = SHA256_EXPAND_S (w0_t, wb_t, w3_t, w2_t); \ + w3_t = SHA256_EXPAND_S (w1_t, wc_t, w4_t, w3_t); \ + w4_t = SHA256_EXPAND_S (w2_t, wd_t, w5_t, w4_t); \ + w5_t = SHA256_EXPAND_S (w3_t, we_t, w6_t, w5_t); \ + w6_t = SHA256_EXPAND_S (w4_t, wf_t, w7_t, w6_t); \ + w7_t = SHA256_EXPAND_S (w5_t, w0_t, w8_t, w7_t); \ + w8_t = SHA256_EXPAND_S (w6_t, w1_t, w9_t, w8_t); \ + w9_t = SHA256_EXPAND_S (w7_t, w2_t, wa_t, w9_t); \ + wa_t = SHA256_EXPAND_S (w8_t, w3_t, wb_t, wa_t); \ + wb_t = SHA256_EXPAND_S (w9_t, w4_t, wc_t, wb_t); \ + wc_t = SHA256_EXPAND_S (wa_t, w5_t, wd_t, wc_t); \ + wd_t = SHA256_EXPAND_S (wb_t, w6_t, we_t, wd_t); \ + we_t = SHA256_EXPAND_S (wc_t, w7_t, wf_t, we_t); \ + wf_t = SHA256_EXPAND_S (wd_t, w8_t, w0_t, wf_t); \ + } + + #define ROUND_STEP_S(i) \ + { \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i + 0]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i + 1]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i + 2]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i + 3]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i + 4]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i + 5]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i + 6]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i + 7]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i + 8]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i + 9]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \ + SHA256_STEP_S (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \ + } + + ROUND_STEP_S (0); + + #ifdef _unroll + #pragma unroll + #endif + for (int i = 16; i < 64; i += 16) + { + ROUND_EXPAND_S (); ROUND_STEP_S (i); + } + + #undef ROUND_EXPAND_S + #undef ROUND_STEP_S + + digest[0] += a; + digest[1] += b; + digest[2] += c; + digest[3] += d; + digest[4] += e; + digest[5] += f; + digest[6] += g; + digest[7] += h; +} + +void sha224_init (sha224_ctx_t *ctx) +{ + ctx->h[0] = SHA224M_A; + ctx->h[1] = SHA224M_B; + ctx->h[2] = SHA224M_C; + ctx->h[3] = SHA224M_D; + ctx->h[4] = SHA224M_E; + ctx->h[5] = SHA224M_F; + ctx->h[6] = SHA224M_G; + ctx->h[7] = SHA224M_H; + + ctx->w0[0] = 0; + ctx->w0[1] = 0; + ctx->w0[2] = 0; + ctx->w0[3] = 0; + ctx->w1[0] = 0; + ctx->w1[1] = 0; + ctx->w1[2] = 0; + ctx->w1[3] = 0; + ctx->w2[0] = 0; + ctx->w2[1] = 0; + ctx->w2[2] = 0; + ctx->w2[3] = 0; + ctx->w3[0] = 0; + ctx->w3[1] = 0; + ctx->w3[2] = 0; + ctx->w3[3] = 0; + + ctx->len = 0; +} + +void sha224_update_64 (sha224_ctx_t *ctx, u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const int len) +{ + const int pos = ctx->len & 63; + + ctx->len += len; + + if ((pos + len) < 64) + { + switch_buffer_by_offset_be_S (w0, w1, w2, w3, pos); + + ctx->w0[0] |= w0[0]; + ctx->w0[1] |= w0[1]; + ctx->w0[2] |= w0[2]; + ctx->w0[3] |= w0[3]; + ctx->w1[0] |= w1[0]; + ctx->w1[1] |= w1[1]; + ctx->w1[2] |= w1[2]; + ctx->w1[3] |= w1[3]; + ctx->w2[0] |= w2[0]; + ctx->w2[1] |= w2[1]; + ctx->w2[2] |= w2[2]; + ctx->w2[3] |= w2[3]; + ctx->w3[0] |= w3[0]; + ctx->w3[1] |= w3[1]; + ctx->w3[2] |= w3[2]; + ctx->w3[3] |= w3[3]; + } + else + { + u32 c0[4] = { 0 }; + u32 c1[4] = { 0 }; + u32 c2[4] = { 0 }; + u32 c3[4] = { 0 }; + + switch_buffer_by_offset_carry_be_S (w0, w1, w2, w3, c0, c1, c2, c3, pos); + + ctx->w0[0] |= w0[0]; + ctx->w0[1] |= w0[1]; + ctx->w0[2] |= w0[2]; + ctx->w0[3] |= w0[3]; + ctx->w1[0] |= w1[0]; + ctx->w1[1] |= w1[1]; + ctx->w1[2] |= w1[2]; + ctx->w1[3] |= w1[3]; + ctx->w2[0] |= w2[0]; + ctx->w2[1] |= w2[1]; + ctx->w2[2] |= w2[2]; + ctx->w2[3] |= w2[3]; + ctx->w3[0] |= w3[0]; + ctx->w3[1] |= w3[1]; + ctx->w3[2] |= w3[2]; + ctx->w3[3] |= w3[3]; + + sha224_transform (ctx->w0, ctx->w1, ctx->w2, ctx->w3, ctx->h); + + ctx->w0[0] = c0[0]; + ctx->w0[1] = c0[1]; + ctx->w0[2] = c0[2]; + ctx->w0[3] = c0[3]; + ctx->w1[0] = c1[0]; + ctx->w1[1] = c1[1]; + ctx->w1[2] = c1[2]; + ctx->w1[3] = c1[3]; + ctx->w2[0] = c2[0]; + ctx->w2[1] = c2[1]; + ctx->w2[2] = c2[2]; + ctx->w2[3] = c2[3]; + ctx->w3[0] = c3[0]; + ctx->w3[1] = c3[1]; + ctx->w3[2] = c3[2]; + ctx->w3[3] = c3[3]; + } +} + +void sha224_update (sha224_ctx_t *ctx, const u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 64; pos1 += 64, pos4 += 16) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + sha224_update_64 (ctx, w0, w1, w2, w3, 64); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + sha224_update_64 (ctx, w0, w1, w2, w3, len - pos1); +} + +void sha224_update_swap (sha224_ctx_t *ctx, const u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 64; pos1 += 64, pos4 += 16) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); + + sha224_update_64 (ctx, w0, w1, w2, w3, 64); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); + + sha224_update_64 (ctx, w0, w1, w2, w3, len - pos1); +} + +void sha224_update_utf16le (sha224_ctx_t *ctx, const u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 32; pos1 += 32, pos4 += 8) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le_S (w1, w2, w3); + make_utf16le_S (w0, w0, w1); + + sha224_update_64 (ctx, w0, w1, w2, w3, 32 * 2); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le_S (w1, w2, w3); + make_utf16le_S (w0, w0, w1); + + sha224_update_64 (ctx, w0, w1, w2, w3, (len - pos1) * 2); +} + +void sha224_update_utf16le_swap (sha224_ctx_t *ctx, const u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 32; pos1 += 32, pos4 += 8) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le_S (w1, w2, w3); + make_utf16le_S (w0, w0, w1); + + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); + + sha224_update_64 (ctx, w0, w1, w2, w3, 32 * 2); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le_S (w1, w2, w3); + make_utf16le_S (w0, w0, w1); + + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); + + sha224_update_64 (ctx, w0, w1, w2, w3, (len - pos1) * 2); +} + +void sha224_update_global (sha224_ctx_t *ctx, const __global u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 64; pos1 += 64, pos4 += 16) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + sha224_update_64 (ctx, w0, w1, w2, w3, 64); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + sha224_update_64 (ctx, w0, w1, w2, w3, len - pos1); +} + +void sha224_update_global_swap (sha224_ctx_t *ctx, const __global u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 64; pos1 += 64, pos4 += 16) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); + + sha224_update_64 (ctx, w0, w1, w2, w3, 64); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); + + sha224_update_64 (ctx, w0, w1, w2, w3, len - pos1); +} + +void sha224_update_global_utf16le (sha224_ctx_t *ctx, const __global u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 32; pos1 += 32, pos4 += 8) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le_S (w1, w2, w3); + make_utf16le_S (w0, w0, w1); + + sha224_update_64 (ctx, w0, w1, w2, w3, 32 * 2); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le_S (w1, w2, w3); + make_utf16le_S (w0, w0, w1); + + sha224_update_64 (ctx, w0, w1, w2, w3, (len - pos1) * 2); +} + +void sha224_update_global_utf16le_swap (sha224_ctx_t *ctx, const __global u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 32; pos1 += 32, pos4 += 8) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le_S (w1, w2, w3); + make_utf16le_S (w0, w0, w1); + + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); + + sha224_update_64 (ctx, w0, w1, w2, w3, 32 * 2); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le_S (w1, w2, w3); + make_utf16le_S (w0, w0, w1); + + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); + + sha224_update_64 (ctx, w0, w1, w2, w3, (len - pos1) * 2); +} + +void sha224_final (sha224_ctx_t *ctx) +{ + int pos = ctx->len & 63; + + append_0x80_4x4_S (ctx->w0, ctx->w1, ctx->w2, ctx->w3, pos ^ 3); + + if (pos >= 56) + { + sha224_transform (ctx->w0, ctx->w1, ctx->w2, ctx->w3, ctx->h); + + ctx->w0[0] = 0; + ctx->w0[1] = 0; + ctx->w0[2] = 0; + ctx->w0[3] = 0; + ctx->w1[0] = 0; + ctx->w1[1] = 0; + ctx->w1[2] = 0; + ctx->w1[3] = 0; + ctx->w2[0] = 0; + ctx->w2[1] = 0; + ctx->w2[2] = 0; + ctx->w2[3] = 0; + ctx->w3[0] = 0; + ctx->w3[1] = 0; + ctx->w3[2] = 0; + ctx->w3[3] = 0; + } + + ctx->w3[2] = 0; + ctx->w3[3] = ctx->len * 8; + + sha224_transform (ctx->w0, ctx->w1, ctx->w2, ctx->w3, ctx->h); +} + +// sha224_hmac + +typedef struct sha224_hmac_ctx +{ + sha224_ctx_t ipad; + sha224_ctx_t opad; + +} sha224_hmac_ctx_t; + +void sha224_hmac_init_64 (sha224_hmac_ctx_t *ctx, const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4]) +{ + u32 t0[4]; + u32 t1[4]; + u32 t2[4]; + u32 t3[4]; + + // ipad + + t0[0] = w0[0] ^ 0x36363636; + t0[1] = w0[1] ^ 0x36363636; + t0[2] = w0[2] ^ 0x36363636; + t0[3] = w0[3] ^ 0x36363636; + t1[0] = w1[0] ^ 0x36363636; + t1[1] = w1[1] ^ 0x36363636; + t1[2] = w1[2] ^ 0x36363636; + t1[3] = w1[3] ^ 0x36363636; + t2[0] = w2[0] ^ 0x36363636; + t2[1] = w2[1] ^ 0x36363636; + t2[2] = w2[2] ^ 0x36363636; + t2[3] = w2[3] ^ 0x36363636; + t3[0] = w3[0] ^ 0x36363636; + t3[1] = w3[1] ^ 0x36363636; + t3[2] = w3[2] ^ 0x36363636; + t3[3] = w3[3] ^ 0x36363636; + + sha224_init (&ctx->ipad); + + sha224_update_64 (&ctx->ipad, t0, t1, t2, t3, 64); + + // opad + + t0[0] = w0[0] ^ 0x5c5c5c5c; + t0[1] = w0[1] ^ 0x5c5c5c5c; + t0[2] = w0[2] ^ 0x5c5c5c5c; + t0[3] = w0[3] ^ 0x5c5c5c5c; + t1[0] = w1[0] ^ 0x5c5c5c5c; + t1[1] = w1[1] ^ 0x5c5c5c5c; + t1[2] = w1[2] ^ 0x5c5c5c5c; + t1[3] = w1[3] ^ 0x5c5c5c5c; + t2[0] = w2[0] ^ 0x5c5c5c5c; + t2[1] = w2[1] ^ 0x5c5c5c5c; + t2[2] = w2[2] ^ 0x5c5c5c5c; + t2[3] = w2[3] ^ 0x5c5c5c5c; + t3[0] = w3[0] ^ 0x5c5c5c5c; + t3[1] = w3[1] ^ 0x5c5c5c5c; + t3[2] = w3[2] ^ 0x5c5c5c5c; + t3[3] = w3[3] ^ 0x5c5c5c5c; + + sha224_init (&ctx->opad); + + sha224_update_64 (&ctx->opad, t0, t1, t2, t3, 64); +} + +void sha224_hmac_init (sha224_hmac_ctx_t *ctx, const u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + if (len > 64) + { + sha224_ctx_t tmp; + + sha224_init (&tmp); + + sha224_update (&tmp, w, len); + + sha224_final (&tmp); + + w0[0] = tmp.h[0]; + w0[1] = tmp.h[1]; + w0[2] = tmp.h[2]; + w0[3] = tmp.h[3]; + w1[0] = tmp.h[4]; + w1[1] = tmp.h[5]; + w1[2] = tmp.h[6]; + w1[3] = tmp.h[7]; + w2[0] = 0; + w2[1] = 0; + w2[2] = 0; + w2[3] = 0; + w3[0] = 0; + w3[1] = 0; + w3[2] = 0; + w3[3] = 0; + } + else + { + w0[0] = w[ 0]; + w0[1] = w[ 1]; + w0[2] = w[ 2]; + w0[3] = w[ 3]; + w1[0] = w[ 4]; + w1[1] = w[ 5]; + w1[2] = w[ 6]; + w1[3] = w[ 7]; + w2[0] = w[ 8]; + w2[1] = w[ 9]; + w2[2] = w[10]; + w2[3] = w[11]; + w3[0] = w[12]; + w3[1] = w[13]; + w3[2] = w[14]; + w3[3] = w[15]; + } + + sha224_hmac_init_64 (ctx, w0, w1, w2, w3); +} + +void sha224_hmac_init_swap (sha224_hmac_ctx_t *ctx, const u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + if (len > 64) + { + sha224_ctx_t tmp; + + sha224_init (&tmp); + + sha224_update_swap (&tmp, w, len); + + sha224_final (&tmp); + + w0[0] = tmp.h[0]; + w0[1] = tmp.h[1]; + w0[2] = tmp.h[2]; + w0[3] = tmp.h[3]; + w1[0] = tmp.h[4]; + w1[1] = tmp.h[5]; + w1[2] = tmp.h[6]; + w1[3] = tmp.h[7]; + w2[0] = 0; + w2[1] = 0; + w2[2] = 0; + w2[3] = 0; + w3[0] = 0; + w3[1] = 0; + w3[2] = 0; + w3[3] = 0; + } + else + { + w0[0] = swap32_S (w[ 0]); + w0[1] = swap32_S (w[ 1]); + w0[2] = swap32_S (w[ 2]); + w0[3] = swap32_S (w[ 3]); + w1[0] = swap32_S (w[ 4]); + w1[1] = swap32_S (w[ 5]); + w1[2] = swap32_S (w[ 6]); + w1[3] = swap32_S (w[ 7]); + w2[0] = swap32_S (w[ 8]); + w2[1] = swap32_S (w[ 9]); + w2[2] = swap32_S (w[10]); + w2[3] = swap32_S (w[11]); + w3[0] = swap32_S (w[12]); + w3[1] = swap32_S (w[13]); + w3[2] = swap32_S (w[14]); + w3[3] = swap32_S (w[15]); + } + + sha224_hmac_init_64 (ctx, w0, w1, w2, w3); +} + +void sha224_hmac_init_global (sha224_hmac_ctx_t *ctx, __global const u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + if (len > 64) + { + sha224_ctx_t tmp; + + sha224_init (&tmp); + + sha224_update_global (&tmp, w, len); + + sha224_final (&tmp); + + w0[0] = tmp.h[0]; + w0[1] = tmp.h[1]; + w0[2] = tmp.h[2]; + w0[3] = tmp.h[3]; + w1[0] = tmp.h[4]; + w1[1] = tmp.h[5]; + w1[2] = tmp.h[6]; + w1[3] = tmp.h[7]; + w2[0] = 0; + w2[1] = 0; + w2[2] = 0; + w2[3] = 0; + w3[0] = 0; + w3[1] = 0; + w3[2] = 0; + w3[3] = 0; + } + else + { + w0[0] = w[ 0]; + w0[1] = w[ 1]; + w0[2] = w[ 2]; + w0[3] = w[ 3]; + w1[0] = w[ 4]; + w1[1] = w[ 5]; + w1[2] = w[ 6]; + w1[3] = w[ 7]; + w2[0] = w[ 8]; + w2[1] = w[ 9]; + w2[2] = w[10]; + w2[3] = w[11]; + w3[0] = w[12]; + w3[1] = w[13]; + w3[2] = w[14]; + w3[3] = w[15]; + } + + sha224_hmac_init_64 (ctx, w0, w1, w2, w3); +} + +void sha224_hmac_init_global_swap (sha224_hmac_ctx_t *ctx, __global const u32 *w, const int len) +{ + u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; + + if (len > 64) + { + sha224_ctx_t tmp; + + sha224_init (&tmp); + + sha224_update_global_swap (&tmp, w, len); + + sha224_final (&tmp); + + w0[0] = tmp.h[0]; + w0[1] = tmp.h[1]; + w0[2] = tmp.h[2]; + w0[3] = tmp.h[3]; + w1[0] = tmp.h[4]; + w1[1] = tmp.h[5]; + w1[2] = tmp.h[6]; + w1[3] = tmp.h[7]; + w2[0] = 0; + w2[1] = 0; + w2[2] = 0; + w2[3] = 0; + w3[0] = 0; + w3[1] = 0; + w3[2] = 0; + w3[3] = 0; + } + else + { + w0[0] = swap32_S (w[ 0]); + w0[1] = swap32_S (w[ 1]); + w0[2] = swap32_S (w[ 2]); + w0[3] = swap32_S (w[ 3]); + w1[0] = swap32_S (w[ 4]); + w1[1] = swap32_S (w[ 5]); + w1[2] = swap32_S (w[ 6]); + w1[3] = swap32_S (w[ 7]); + w2[0] = swap32_S (w[ 8]); + w2[1] = swap32_S (w[ 9]); + w2[2] = swap32_S (w[10]); + w2[3] = swap32_S (w[11]); + w3[0] = swap32_S (w[12]); + w3[1] = swap32_S (w[13]); + w3[2] = swap32_S (w[14]); + w3[3] = swap32_S (w[15]); + } + + sha224_hmac_init_64 (ctx, w0, w1, w2, w3); +} + +void sha224_hmac_update_64 (sha224_hmac_ctx_t *ctx, u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], const int len) +{ + sha224_update_64 (&ctx->ipad, w0, w1, w2, w3, len); +} + +void sha224_hmac_update (sha224_hmac_ctx_t *ctx, const u32 *w, const int len) +{ + sha224_update (&ctx->ipad, w, len); +} + +void sha224_hmac_update_swap (sha224_hmac_ctx_t *ctx, const u32 *w, const int len) +{ + sha224_update_swap (&ctx->ipad, w, len); +} + +void sha224_hmac_update_utf16le (sha224_hmac_ctx_t *ctx, const u32 *w, const int len) +{ + sha224_update_utf16le (&ctx->ipad, w, len); +} + +void sha224_hmac_update_utf16le_swap (sha224_hmac_ctx_t *ctx, const u32 *w, const int len) +{ + sha224_update_utf16le_swap (&ctx->ipad, w, len); +} + +void sha224_hmac_update_global (sha224_hmac_ctx_t *ctx, const __global u32 *w, const int len) +{ + sha224_update_global (&ctx->ipad, w, len); +} + +void sha224_hmac_update_global_swap (sha224_hmac_ctx_t *ctx, const __global u32 *w, const int len) +{ + sha224_update_global_swap (&ctx->ipad, w, len); +} + +void sha224_hmac_update_global_utf16le (sha224_hmac_ctx_t *ctx, const __global u32 *w, const int len) +{ + sha224_update_global_utf16le (&ctx->ipad, w, len); +} + +void sha224_hmac_update_global_utf16le_swap (sha224_hmac_ctx_t *ctx, const __global u32 *w, const int len) +{ + sha224_update_global_utf16le_swap (&ctx->ipad, w, len); +} + +void sha224_hmac_final (sha224_hmac_ctx_t *ctx) +{ + sha224_final (&ctx->ipad); + + u32 t0[4]; + u32 t1[4]; + u32 t2[4]; + u32 t3[4]; + + t0[0] = ctx->ipad.h[0]; + t0[1] = ctx->ipad.h[1]; + t0[2] = ctx->ipad.h[2]; + t0[3] = ctx->ipad.h[3]; + t1[0] = ctx->ipad.h[4]; + t1[1] = ctx->ipad.h[5]; + t1[2] = ctx->ipad.h[6]; + t1[3] = ctx->ipad.h[7]; + t2[0] = 0; + t2[1] = 0; + t2[2] = 0; + t2[3] = 0; + t3[0] = 0; + t3[1] = 0; + t3[2] = 0; + t3[3] = 0; + + sha224_update_64 (&ctx->opad, t0, t1, t2, t3, 32); + + sha224_final (&ctx->opad); +} + +// while input buf can be a vector datatype, the length of the different elements can not + +typedef struct sha224_ctx_vector +{ + u32x h[8]; + + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; + + int len; + +} sha224_ctx_vector_t; + +void sha224_transform_vector (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[8]) +{ + u32x a = digest[0]; + u32x b = digest[1]; + u32x c = digest[2]; + u32x d = digest[3]; + u32x e = digest[4]; + u32x f = digest[5]; + u32x g = digest[6]; + u32x h = digest[7]; + + u32x w0_t = w0[0]; + u32x w1_t = w0[1]; + u32x w2_t = w0[2]; + u32x w3_t = w0[3]; + u32x w4_t = w1[0]; + u32x w5_t = w1[1]; + u32x w6_t = w1[2]; + u32x w7_t = w1[3]; + u32x w8_t = w2[0]; + u32x w9_t = w2[1]; + u32x wa_t = w2[2]; + u32x wb_t = w2[3]; + u32x wc_t = w3[0]; + u32x wd_t = w3[1]; + u32x we_t = w3[2]; + u32x wf_t = w3[3]; + + #define ROUND_EXPAND() \ + { \ + w0_t = SHA256_EXPAND (we_t, w9_t, w1_t, w0_t); \ + w1_t = SHA256_EXPAND (wf_t, wa_t, w2_t, w1_t); \ + w2_t = SHA256_EXPAND (w0_t, wb_t, w3_t, w2_t); \ + w3_t = SHA256_EXPAND (w1_t, wc_t, w4_t, w3_t); \ + w4_t = SHA256_EXPAND (w2_t, wd_t, w5_t, w4_t); \ + w5_t = SHA256_EXPAND (w3_t, we_t, w6_t, w5_t); \ + w6_t = SHA256_EXPAND (w4_t, wf_t, w7_t, w6_t); \ + w7_t = SHA256_EXPAND (w5_t, w0_t, w8_t, w7_t); \ + w8_t = SHA256_EXPAND (w6_t, w1_t, w9_t, w8_t); \ + w9_t = SHA256_EXPAND (w7_t, w2_t, wa_t, w9_t); \ + wa_t = SHA256_EXPAND (w8_t, w3_t, wb_t, wa_t); \ + wb_t = SHA256_EXPAND (w9_t, w4_t, wc_t, wb_t); \ + wc_t = SHA256_EXPAND (wa_t, w5_t, wd_t, wc_t); \ + wd_t = SHA256_EXPAND (wb_t, w6_t, we_t, wd_t); \ + we_t = SHA256_EXPAND (wc_t, w7_t, wf_t, we_t); \ + wf_t = SHA256_EXPAND (wd_t, w8_t, w0_t, wf_t); \ + } + + #define ROUND_STEP(i) \ + { \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i + 0]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i + 1]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i + 2]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i + 3]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i + 4]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i + 5]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i + 6]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i + 7]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i + 8]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i + 9]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \ + } + + ROUND_STEP (0); + + #ifdef _unroll + #pragma unroll + #endif + for (int i = 16; i < 64; i += 16) + { + ROUND_EXPAND (); ROUND_STEP (i); + } + + #undef ROUND_EXPAND + #undef ROUND_STEP + + digest[0] += a; + digest[1] += b; + digest[2] += c; + digest[3] += d; + digest[4] += e; + digest[5] += f; + digest[6] += g; + digest[7] += h; +} + +void sha224_init_vector (sha224_ctx_vector_t *ctx) +{ + ctx->h[0] = SHA224M_A; + ctx->h[1] = SHA224M_B; + ctx->h[2] = SHA224M_C; + ctx->h[3] = SHA224M_D; + ctx->h[4] = SHA224M_E; + ctx->h[5] = SHA224M_F; + ctx->h[6] = SHA224M_G; + ctx->h[7] = SHA224M_H; + + ctx->w0[0] = 0; + ctx->w0[1] = 0; + ctx->w0[2] = 0; + ctx->w0[3] = 0; + ctx->w1[0] = 0; + ctx->w1[1] = 0; + ctx->w1[2] = 0; + ctx->w1[3] = 0; + ctx->w2[0] = 0; + ctx->w2[1] = 0; + ctx->w2[2] = 0; + ctx->w2[3] = 0; + ctx->w3[0] = 0; + ctx->w3[1] = 0; + ctx->w3[2] = 0; + ctx->w3[3] = 0; + + ctx->len = 0; +} + +void sha224_init_vector_from_scalar (sha224_ctx_vector_t *ctx, sha224_ctx_t *ctx0) +{ + ctx->h[0] = ctx0->h[0]; + ctx->h[1] = ctx0->h[1]; + ctx->h[2] = ctx0->h[2]; + ctx->h[3] = ctx0->h[3]; + ctx->h[4] = ctx0->h[4]; + ctx->h[5] = ctx0->h[5]; + ctx->h[6] = ctx0->h[6]; + ctx->h[7] = ctx0->h[7]; + + ctx->w0[0] = ctx0->w0[0]; + ctx->w0[1] = ctx0->w0[1]; + ctx->w0[2] = ctx0->w0[2]; + ctx->w0[3] = ctx0->w0[3]; + ctx->w1[0] = ctx0->w1[0]; + ctx->w1[1] = ctx0->w1[1]; + ctx->w1[2] = ctx0->w1[2]; + ctx->w1[3] = ctx0->w1[3]; + ctx->w2[0] = ctx0->w2[0]; + ctx->w2[1] = ctx0->w2[1]; + ctx->w2[2] = ctx0->w2[2]; + ctx->w2[3] = ctx0->w2[3]; + ctx->w3[0] = ctx0->w3[0]; + ctx->w3[1] = ctx0->w3[1]; + ctx->w3[2] = ctx0->w3[2]; + ctx->w3[3] = ctx0->w3[3]; + + ctx->len = ctx0->len; +} + +void sha224_update_vector_64 (sha224_ctx_vector_t *ctx, u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], const int len) +{ + const int pos = ctx->len & 63; + + ctx->len += len; + + if ((pos + len) < 64) + { + switch_buffer_by_offset_be (w0, w1, w2, w3, pos); + + ctx->w0[0] |= w0[0]; + ctx->w0[1] |= w0[1]; + ctx->w0[2] |= w0[2]; + ctx->w0[3] |= w0[3]; + ctx->w1[0] |= w1[0]; + ctx->w1[1] |= w1[1]; + ctx->w1[2] |= w1[2]; + ctx->w1[3] |= w1[3]; + ctx->w2[0] |= w2[0]; + ctx->w2[1] |= w2[1]; + ctx->w2[2] |= w2[2]; + ctx->w2[3] |= w2[3]; + ctx->w3[0] |= w3[0]; + ctx->w3[1] |= w3[1]; + ctx->w3[2] |= w3[2]; + ctx->w3[3] |= w3[3]; + } + else + { + u32x c0[4] = { 0 }; + u32x c1[4] = { 0 }; + u32x c2[4] = { 0 }; + u32x c3[4] = { 0 }; + + switch_buffer_by_offset_carry_be (w0, w1, w2, w3, c0, c1, c2, c3, pos); + + ctx->w0[0] |= w0[0]; + ctx->w0[1] |= w0[1]; + ctx->w0[2] |= w0[2]; + ctx->w0[3] |= w0[3]; + ctx->w1[0] |= w1[0]; + ctx->w1[1] |= w1[1]; + ctx->w1[2] |= w1[2]; + ctx->w1[3] |= w1[3]; + ctx->w2[0] |= w2[0]; + ctx->w2[1] |= w2[1]; + ctx->w2[2] |= w2[2]; + ctx->w2[3] |= w2[3]; + ctx->w3[0] |= w3[0]; + ctx->w3[1] |= w3[1]; + ctx->w3[2] |= w3[2]; + ctx->w3[3] |= w3[3]; + + sha224_transform_vector (ctx->w0, ctx->w1, ctx->w2, ctx->w3, ctx->h); + + ctx->w0[0] = c0[0]; + ctx->w0[1] = c0[1]; + ctx->w0[2] = c0[2]; + ctx->w0[3] = c0[3]; + ctx->w1[0] = c1[0]; + ctx->w1[1] = c1[1]; + ctx->w1[2] = c1[2]; + ctx->w1[3] = c1[3]; + ctx->w2[0] = c2[0]; + ctx->w2[1] = c2[1]; + ctx->w2[2] = c2[2]; + ctx->w2[3] = c2[3]; + ctx->w3[0] = c3[0]; + ctx->w3[1] = c3[1]; + ctx->w3[2] = c3[2]; + ctx->w3[3] = c3[3]; + } +} + +void sha224_update_vector (sha224_ctx_vector_t *ctx, const u32x *w, const int len) +{ + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 64; pos1 += 64, pos4 += 16) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, 64); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, len - pos1); +} + +void sha224_update_vector_swap (sha224_ctx_vector_t *ctx, const u32x *w, const int len) +{ + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 64; pos1 += 64, pos4 += 16) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + w0[0] = swap32 (w0[0]); + w0[1] = swap32 (w0[1]); + w0[2] = swap32 (w0[2]); + w0[3] = swap32 (w0[3]); + w1[0] = swap32 (w1[0]); + w1[1] = swap32 (w1[1]); + w1[2] = swap32 (w1[2]); + w1[3] = swap32 (w1[3]); + w2[0] = swap32 (w2[0]); + w2[1] = swap32 (w2[1]); + w2[2] = swap32 (w2[2]); + w2[3] = swap32 (w2[3]); + w3[0] = swap32 (w3[0]); + w3[1] = swap32 (w3[1]); + w3[2] = swap32 (w3[2]); + w3[3] = swap32 (w3[3]); + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, 64); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + w2[0] = w[pos4 + 8]; + w2[1] = w[pos4 + 9]; + w2[2] = w[pos4 + 10]; + w2[3] = w[pos4 + 11]; + w3[0] = w[pos4 + 12]; + w3[1] = w[pos4 + 13]; + w3[2] = w[pos4 + 14]; + w3[3] = w[pos4 + 15]; + + w0[0] = swap32 (w0[0]); + w0[1] = swap32 (w0[1]); + w0[2] = swap32 (w0[2]); + w0[3] = swap32 (w0[3]); + w1[0] = swap32 (w1[0]); + w1[1] = swap32 (w1[1]); + w1[2] = swap32 (w1[2]); + w1[3] = swap32 (w1[3]); + w2[0] = swap32 (w2[0]); + w2[1] = swap32 (w2[1]); + w2[2] = swap32 (w2[2]); + w2[3] = swap32 (w2[3]); + w3[0] = swap32 (w3[0]); + w3[1] = swap32 (w3[1]); + w3[2] = swap32 (w3[2]); + w3[3] = swap32 (w3[3]); + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, len - pos1); +} + +void sha224_update_vector_utf16le (sha224_ctx_vector_t *ctx, const u32x *w, const int len) +{ + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 32; pos1 += 32, pos4 += 8) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le (w1, w2, w3); + make_utf16le (w0, w0, w1); + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, 32 * 2); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le (w1, w2, w3); + make_utf16le (w0, w0, w1); + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, (len - pos1) * 2); +} + +void sha224_update_vector_utf16le_swap (sha224_ctx_vector_t *ctx, const u32x *w, const int len) +{ + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 32; pos1 += 32, pos4 += 8) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le (w1, w2, w3); + make_utf16le (w0, w0, w1); + + w0[0] = swap32 (w0[0]); + w0[1] = swap32 (w0[1]); + w0[2] = swap32 (w0[2]); + w0[3] = swap32 (w0[3]); + w1[0] = swap32 (w1[0]); + w1[1] = swap32 (w1[1]); + w1[2] = swap32 (w1[2]); + w1[3] = swap32 (w1[3]); + w2[0] = swap32 (w2[0]); + w2[1] = swap32 (w2[1]); + w2[2] = swap32 (w2[2]); + w2[3] = swap32 (w2[3]); + w3[0] = swap32 (w3[0]); + w3[1] = swap32 (w3[1]); + w3[2] = swap32 (w3[2]); + w3[3] = swap32 (w3[3]); + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, 32 * 2); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16le (w1, w2, w3); + make_utf16le (w0, w0, w1); + + w0[0] = swap32 (w0[0]); + w0[1] = swap32 (w0[1]); + w0[2] = swap32 (w0[2]); + w0[3] = swap32 (w0[3]); + w1[0] = swap32 (w1[0]); + w1[1] = swap32 (w1[1]); + w1[2] = swap32 (w1[2]); + w1[3] = swap32 (w1[3]); + w2[0] = swap32 (w2[0]); + w2[1] = swap32 (w2[1]); + w2[2] = swap32 (w2[2]); + w2[3] = swap32 (w2[3]); + w3[0] = swap32 (w3[0]); + w3[1] = swap32 (w3[1]); + w3[2] = swap32 (w3[2]); + w3[3] = swap32 (w3[3]); + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, (len - pos1) * 2); +} + +void sha224_update_vector_utf16beN (sha224_ctx_vector_t *ctx, const u32x *w, const int len) +{ + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; + + int pos1; + int pos4; + + for (pos1 = 0, pos4 = 0; pos1 < len - 32; pos1 += 32, pos4 += 8) + { + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16beN (w1, w2, w3); + make_utf16beN (w0, w0, w1); + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, 32 * 2); + } + + w0[0] = w[pos4 + 0]; + w0[1] = w[pos4 + 1]; + w0[2] = w[pos4 + 2]; + w0[3] = w[pos4 + 3]; + w1[0] = w[pos4 + 4]; + w1[1] = w[pos4 + 5]; + w1[2] = w[pos4 + 6]; + w1[3] = w[pos4 + 7]; + + make_utf16beN (w1, w2, w3); + make_utf16beN (w0, w0, w1); + + sha224_update_vector_64 (ctx, w0, w1, w2, w3, (len - pos1) * 2); +} + +void sha224_final_vector (sha224_ctx_vector_t *ctx) +{ + int pos = ctx->len & 63; + + append_0x80_4x4 (ctx->w0, ctx->w1, ctx->w2, ctx->w3, pos ^ 3); + + if (pos >= 56) + { + sha224_transform_vector (ctx->w0, ctx->w1, ctx->w2, ctx->w3, ctx->h); + + ctx->w0[0] = 0; + ctx->w0[1] = 0; + ctx->w0[2] = 0; + ctx->w0[3] = 0; + ctx->w1[0] = 0; + ctx->w1[1] = 0; + ctx->w1[2] = 0; + ctx->w1[3] = 0; + ctx->w2[0] = 0; + ctx->w2[1] = 0; + ctx->w2[2] = 0; + ctx->w2[3] = 0; + ctx->w3[0] = 0; + ctx->w3[1] = 0; + ctx->w3[2] = 0; + ctx->w3[3] = 0; + } + + ctx->w3[2] = 0; + ctx->w3[3] = ctx->len * 8; + + sha224_transform_vector (ctx->w0, ctx->w1, ctx->w2, ctx->w3, ctx->h); +} + +// HMAC + Vector + +typedef struct sha224_hmac_ctx_vector +{ + sha224_ctx_vector_t ipad; + sha224_ctx_vector_t opad; + +} sha224_hmac_ctx_vector_t; + +void sha224_hmac_init_vector_64 (sha224_hmac_ctx_vector_t *ctx, const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4]) +{ + u32x t0[4]; + u32x t1[4]; + u32x t2[4]; + u32x t3[4]; + + // ipad + + t0[0] = w0[0] ^ 0x36363636; + t0[1] = w0[1] ^ 0x36363636; + t0[2] = w0[2] ^ 0x36363636; + t0[3] = w0[3] ^ 0x36363636; + t1[0] = w1[0] ^ 0x36363636; + t1[1] = w1[1] ^ 0x36363636; + t1[2] = w1[2] ^ 0x36363636; + t1[3] = w1[3] ^ 0x36363636; + t2[0] = w2[0] ^ 0x36363636; + t2[1] = w2[1] ^ 0x36363636; + t2[2] = w2[2] ^ 0x36363636; + t2[3] = w2[3] ^ 0x36363636; + t3[0] = w3[0] ^ 0x36363636; + t3[1] = w3[1] ^ 0x36363636; + t3[2] = w3[2] ^ 0x36363636; + t3[3] = w3[3] ^ 0x36363636; + + sha224_init_vector (&ctx->ipad); + + sha224_update_vector_64 (&ctx->ipad, t0, t1, t2, t3, 64); + + // opad + + t0[0] = w0[0] ^ 0x5c5c5c5c; + t0[1] = w0[1] ^ 0x5c5c5c5c; + t0[2] = w0[2] ^ 0x5c5c5c5c; + t0[3] = w0[3] ^ 0x5c5c5c5c; + t1[0] = w1[0] ^ 0x5c5c5c5c; + t1[1] = w1[1] ^ 0x5c5c5c5c; + t1[2] = w1[2] ^ 0x5c5c5c5c; + t1[3] = w1[3] ^ 0x5c5c5c5c; + t2[0] = w2[0] ^ 0x5c5c5c5c; + t2[1] = w2[1] ^ 0x5c5c5c5c; + t2[2] = w2[2] ^ 0x5c5c5c5c; + t2[3] = w2[3] ^ 0x5c5c5c5c; + t3[0] = w3[0] ^ 0x5c5c5c5c; + t3[1] = w3[1] ^ 0x5c5c5c5c; + t3[2] = w3[2] ^ 0x5c5c5c5c; + t3[3] = w3[3] ^ 0x5c5c5c5c; + + sha224_init_vector (&ctx->opad); + + sha224_update_vector_64 (&ctx->opad, t0, t1, t2, t3, 64); +} + +void sha224_hmac_init_vector (sha224_hmac_ctx_vector_t *ctx, const u32x *w, const int len) +{ + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; + + if (len > 64) + { + sha224_ctx_vector_t tmp; + + sha224_init_vector (&tmp); + + sha224_update_vector (&tmp, w, len); + + sha224_final_vector (&tmp); + + w0[0] = tmp.h[0]; + w0[1] = tmp.h[1]; + w0[2] = tmp.h[2]; + w0[3] = tmp.h[3]; + w1[0] = tmp.h[4]; + w1[1] = tmp.h[5]; + w1[2] = tmp.h[6]; + w1[3] = tmp.h[7]; + w2[0] = 0; + w2[1] = 0; + w2[2] = 0; + w2[3] = 0; + w3[0] = 0; + w3[1] = 0; + w3[2] = 0; + w3[3] = 0; + } + else + { + w0[0] = w[ 0]; + w0[1] = w[ 1]; + w0[2] = w[ 2]; + w0[3] = w[ 3]; + w1[0] = w[ 4]; + w1[1] = w[ 5]; + w1[2] = w[ 6]; + w1[3] = w[ 7]; + w2[0] = w[ 8]; + w2[1] = w[ 9]; + w2[2] = w[10]; + w2[3] = w[11]; + w3[0] = w[12]; + w3[1] = w[13]; + w3[2] = w[14]; + w3[3] = w[15]; + } + + sha224_hmac_init_vector_64 (ctx, w0, w1, w2, w3); +} + +void sha224_hmac_update_vector_64 (sha224_hmac_ctx_vector_t *ctx, u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], const int len) +{ + sha224_update_vector_64 (&ctx->ipad, w0, w1, w2, w3, len); +} + +void sha224_hmac_update_vector (sha224_hmac_ctx_vector_t *ctx, const u32x *w, const int len) +{ + sha224_update_vector (&ctx->ipad, w, len); +} + +void sha224_hmac_final_vector (sha224_hmac_ctx_vector_t *ctx) +{ + sha224_final_vector (&ctx->ipad); + + u32x t0[4]; + u32x t1[4]; + u32x t2[4]; + u32x t3[4]; + + t0[0] = ctx->ipad.h[0]; + t0[1] = ctx->ipad.h[1]; + t0[2] = ctx->ipad.h[2]; + t0[3] = ctx->ipad.h[3]; + t1[0] = ctx->ipad.h[4]; + t1[1] = ctx->ipad.h[5]; + t1[2] = ctx->ipad.h[6]; + t1[3] = ctx->ipad.h[7]; + t2[0] = 0; + t2[1] = 0; + t2[2] = 0; + t2[3] = 0; + t3[0] = 0; + t3[1] = 0; + t3[2] = 0; + t3[3] = 0; + + sha224_update_vector_64 (&ctx->opad, t0, t1, t2, t3, 32); + + sha224_final_vector (&ctx->opad); +} diff --git a/OpenCL/m01300_a0-pure.cl b/OpenCL/m01300_a0-pure.cl new file mode 100644 index 000000000..7607c95f8 --- /dev/null +++ b/OpenCL/m01300_a0-pure.cl @@ -0,0 +1,134 @@ +/** + * Author......: See docs/credits.txt + * License.....: MIT + */ + +//#define NEW_SIMD_CODE + +#include "inc_vendor.cl" +#include "inc_hash_constants.h" +#include "inc_hash_functions.cl" +#include "inc_types.cl" +#include "inc_common.cl" +#include "inc_rp.h" +#include "inc_rp.cl" +#include "inc_scalar.cl" +#include "inc_hash_sha224.cl" + +__kernel void m01300_mxx (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) +{ + /** + * modifier + */ + + const u32 lid = get_local_id (0); + const u32 gid = get_global_id (0); + + if (gid >= gid_max) return; + + /** + * base + */ + + const u32 pw_len = pws[gid].pw_len; + + const u32 pw_lenv = ceil ((float) pw_len / 4); + + u32 w[64] = { 0 }; + + for (int idx = 0; idx < pw_lenv; idx++) + { + w[idx] = pws[gid].i[idx]; + + barrier (CLK_GLOBAL_MEM_FENCE); + } + + /** + * loop + */ + + for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) + { + // todo: add rules engine + + sha224_ctx_t ctx; + + sha224_init (&ctx); + + sha224_update_swap (&ctx, w, pw_len); + + sha224_final (&ctx); + + const u32 r0 = ctx.h[DGST_R0]; + const u32 r1 = ctx.h[DGST_R1]; + const u32 r2 = ctx.h[DGST_R2]; + const u32 r3 = ctx.h[DGST_R3]; + + COMPARE_M_SCALAR (r0, r1, r2, r3); + } +} + +__kernel void m01300_sxx (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) +{ + /** + * modifier + */ + + const u32 lid = get_local_id (0); + const u32 gid = get_global_id (0); + + if (gid >= gid_max) return; + + /** + * digest + */ + + const u32 search[4] = + { + digests_buf[digests_offset].digest_buf[DGST_R0], + digests_buf[digests_offset].digest_buf[DGST_R1], + digests_buf[digests_offset].digest_buf[DGST_R2], + digests_buf[digests_offset].digest_buf[DGST_R3] + }; + + /** + * base + */ + + const u32 pw_len = pws[gid].pw_len; + + const u32 pw_lenv = ceil ((float) pw_len / 4); + + u32 w[64] = { 0 }; + + for (int idx = 0; idx < pw_lenv; idx++) + { + w[idx] = pws[gid].i[idx]; + + barrier (CLK_GLOBAL_MEM_FENCE); + } + + /** + * loop + */ + + for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) + { + // todo: add rules engine + + sha224_ctx_t ctx; + + sha224_init (&ctx); + + sha224_update_swap (&ctx, w, pw_len); + + sha224_final (&ctx); + + const u32 r0 = ctx.h[DGST_R0]; + const u32 r1 = ctx.h[DGST_R1]; + const u32 r2 = ctx.h[DGST_R2]; + const u32 r3 = ctx.h[DGST_R3]; + + COMPARE_S_SCALAR (r0, r1, r2, r3); + } +} diff --git a/OpenCL/m01300_a1-pure.cl b/OpenCL/m01300_a1-pure.cl new file mode 100644 index 000000000..3f48fc9e2 --- /dev/null +++ b/OpenCL/m01300_a1-pure.cl @@ -0,0 +1,110 @@ +/** + * Author......: See docs/credits.txt + * License.....: MIT + */ + +//#define NEW_SIMD_CODE + +#include "inc_vendor.cl" +#include "inc_hash_constants.h" +#include "inc_hash_functions.cl" +#include "inc_types.cl" +#include "inc_common.cl" +#include "inc_scalar.cl" +#include "inc_hash_sha224.cl" + +__kernel void m01300_mxx (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) +{ + /** + * modifier + */ + + const u32 lid = get_local_id (0); + const u32 gid = get_global_id (0); + + if (gid >= gid_max) return; + + /** + * base + */ + + sha224_ctx_t ctx0; + + sha224_init (&ctx0); + + sha224_update_global_swap (&ctx0, pws[gid].i, pws[gid].pw_len); + + /** + * loop + */ + + for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) + { + sha224_ctx_t ctx = ctx0; + + sha224_update_global_swap (&ctx, combs_buf[il_pos].i, combs_buf[il_pos].pw_len); + + sha224_final (&ctx); + + const u32 r0 = ctx.h[DGST_R0]; + const u32 r1 = ctx.h[DGST_R1]; + const u32 r2 = ctx.h[DGST_R2]; + const u32 r3 = ctx.h[DGST_R3]; + + COMPARE_M_SCALAR (r0, r1, r2, r3); + } +} + +__kernel void m01300_sxx (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) +{ + /** + * modifier + */ + + const u32 lid = get_local_id (0); + const u32 gid = get_global_id (0); + + if (gid >= gid_max) return; + + /** + * digest + */ + + const u32 search[4] = + { + digests_buf[digests_offset].digest_buf[DGST_R0], + digests_buf[digests_offset].digest_buf[DGST_R1], + digests_buf[digests_offset].digest_buf[DGST_R2], + digests_buf[digests_offset].digest_buf[DGST_R3] + }; + + /** + * base + */ + + sha224_ctx_t ctx0; + + sha224_init (&ctx0); + + sha224_update_global_swap (&ctx0, pws[gid].i, pws[gid].pw_len); + + /** + * loop + */ + + for (u32 il_pos = 0; il_pos < il_cnt; il_pos++) + { + sha224_ctx_t ctx = ctx0; + + sha224_update_global_swap (&ctx, combs_buf[il_pos].i, combs_buf[il_pos].pw_len); + + sha224_final (&ctx); + + const u32 r0 = ctx.h[DGST_R0]; + const u32 r1 = ctx.h[DGST_R1]; + const u32 r2 = ctx.h[DGST_R2]; + const u32 r3 = ctx.h[DGST_R3]; + + COMPARE_S_SCALAR (r0, r1, r2, r3); + } +} diff --git a/OpenCL/m01300_a3-pure.cl b/OpenCL/m01300_a3-pure.cl new file mode 100644 index 000000000..cbb1203a8 --- /dev/null +++ b/OpenCL/m01300_a3-pure.cl @@ -0,0 +1,144 @@ +/** + * Author......: See docs/credits.txt + * License.....: MIT + */ + +#define NEW_SIMD_CODE + +#include "inc_vendor.cl" +#include "inc_hash_constants.h" +#include "inc_hash_functions.cl" +#include "inc_types.cl" +#include "inc_common.cl" +#include "inc_simd.cl" +#include "inc_hash_sha224.cl" + +__kernel void m01300_mxx (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __constant const u32x *words_buf_r, __global void *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) +{ + /** + * modifier + */ + + const u32 lid = get_local_id (0); + const u32 gid = get_global_id (0); + + if (gid >= gid_max) return; + + /** + * base + */ + + const u32 pw_len = pws[gid].pw_len; + + const u32 pw_lenv = ceil ((float) pw_len / 4); + + u32x w[64] = { 0 }; + + for (int idx = 0; idx < pw_lenv; idx++) + { + w[idx] = pws[gid].i[idx]; + + barrier (CLK_GLOBAL_MEM_FENCE); + } + + /** + * loop + */ + + u32x w0l = w[0]; + + for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) + { + const u32x w0r = words_buf_r[il_pos / VECT_SIZE]; + + const u32x w0 = w0l | w0r; + + w[0] = w0; + + sha224_ctx_vector_t ctx; + + sha224_init_vector (&ctx); + + sha224_update_vector (&ctx, w, pw_len); + + sha224_final_vector (&ctx); + + const u32x r0 = ctx.h[DGST_R0]; + const u32x r1 = ctx.h[DGST_R1]; + const u32x r2 = ctx.h[DGST_R2]; + const u32x r3 = ctx.h[DGST_R3]; + + COMPARE_M_SIMD (r0, r1, r2, r3); + } +} + +__kernel void m01300_sxx (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __constant const u32x *words_buf_r, __global void *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max) +{ + /** + * modifier + */ + + const u32 lid = get_local_id (0); + const u32 gid = get_global_id (0); + + if (gid >= gid_max) return; + + /** + * digest + */ + + const u32 search[4] = + { + digests_buf[digests_offset].digest_buf[DGST_R0], + digests_buf[digests_offset].digest_buf[DGST_R1], + digests_buf[digests_offset].digest_buf[DGST_R2], + digests_buf[digests_offset].digest_buf[DGST_R3] + }; + + /** + * base + */ + + const u32 pw_len = pws[gid].pw_len; + + const u32 pw_lenv = ceil ((float) pw_len / 4); + + u32x w[64] = { 0 }; + + for (int idx = 0; idx < pw_lenv; idx++) + { + w[idx] = pws[gid].i[idx]; + + barrier (CLK_GLOBAL_MEM_FENCE); + } + + /** + * loop + */ + + u32x w0l = w[0]; + + for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) + { + const u32x w0r = words_buf_r[il_pos / VECT_SIZE]; + + const u32x w0 = w0l | w0r; + + w[0] = w0; + + sha224_ctx_vector_t ctx; + + sha224_init_vector (&ctx); + + sha224_update_vector (&ctx, w, pw_len); + + sha224_final_vector (&ctx); + + const u32x r0 = ctx.h[DGST_R0]; + const u32x r1 = ctx.h[DGST_R1]; + const u32x r2 = ctx.h[DGST_R2]; + const u32x r3 = ctx.h[DGST_R3]; + + COMPARE_S_SIMD (r0, r1, r2, r3); + } +}