/** * Author......: See docs/credits.txt * License.....: MIT */ #define NEW_SIMD_CODE #ifdef KERNEL_STATIC #include "inc_vendor.h" #include "inc_types.h" #include "inc_common.cl" #include "inc_rp_optimized.h" #include "inc_rp_optimized.cl" #include "inc_simd.cl" #include "inc_hash_md5.cl" #include "inc_hash_sha1.cl" #endif #if VECT_SIZE == 1 #define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i)]) #elif VECT_SIZE == 2 #define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1]) #elif VECT_SIZE == 4 #define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3]) #elif VECT_SIZE == 8 #define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7]) #elif VECT_SIZE == 16 #define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7], l_bin2asc[(i).s8], l_bin2asc[(i).s9], l_bin2asc[(i).sa], l_bin2asc[(i).sb], l_bin2asc[(i).sc], l_bin2asc[(i).sd], l_bin2asc[(i).se], l_bin2asc[(i).sf]) #endif KERNEL_FQ void m04700_m04 (KERN_ATTR_RULES ()) { /** * modifier */ const u64 gid = get_global_id (0); const u64 lid = get_local_id (0); const u64 lsz = get_local_size (0); /** * bin2asc table */ LOCAL_AS u32 l_bin2asc[256]; for (u32 i = lid; i < 256; i += lsz) { const u32 i0 = (i >> 0) & 15; const u32 i1 = (i >> 4) & 15; l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'a' - 10 + i0) << 0 | ((i1 < 10) ? '0' + i1 : 'a' - 10 + i1) << 8; } barrier (CLK_LOCAL_MEM_FENCE); if (gid >= gid_max) return; /** * base */ u32 pw_buf0[4]; u32 pw_buf1[4]; pw_buf0[0] = pws[gid].i[0]; pw_buf0[1] = pws[gid].i[1]; pw_buf0[2] = pws[gid].i[2]; pw_buf0[3] = pws[gid].i[3]; pw_buf1[0] = pws[gid].i[4]; pw_buf1[1] = pws[gid].i[5]; pw_buf1[2] = pws[gid].i[6]; pw_buf1[3] = pws[gid].i[7]; const u32 pw_len = pws[gid].pw_len & 63; /** * loop */ for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) { u32x w0[4] = { 0 }; u32x w1[4] = { 0 }; u32x w2[4] = { 0 }; u32x w3[4] = { 0 }; const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1); append_0x80_2x4_VV (w0, w1, out_len); w3[2] = out_len * 8; w3[3] = 0; /** * md5 */ u32x a = MD5M_A; u32x b = MD5M_B; u32x c = MD5M_C; u32x d = MD5M_D; u32x e = 0; MD5_STEP (MD5_Fo, a, b, c, d, w0[0], MD5C00, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w0[1], MD5C01, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w0[2], MD5C02, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w0[3], MD5C03, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, w1[0], MD5C04, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w1[1], MD5C05, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w1[2], MD5C06, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w1[3], MD5C07, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, w2[0], MD5C08, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w2[1], MD5C09, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w2[2], MD5C0a, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w2[3], MD5C0b, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, w3[0], MD5C0c, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w3[1], MD5C0d, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w3[2], MD5C0e, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w3[3], MD5C0f, MD5S03); MD5_STEP (MD5_Go, a, b, c, d, w0[1], MD5C10, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w1[2], MD5C11, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w2[3], MD5C12, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w0[0], MD5C13, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, w1[1], MD5C14, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w2[2], MD5C15, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w3[3], MD5C16, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w1[0], MD5C17, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, w2[1], MD5C18, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w3[2], MD5C19, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w0[3], MD5C1a, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w2[0], MD5C1b, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, w3[1], MD5C1c, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w0[2], MD5C1d, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w1[3], MD5C1e, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w3[0], MD5C1f, MD5S13); u32x t; MD5_STEP (MD5_H1, a, b, c, d, w1[1], MD5C20, MD5S20); MD5_STEP (MD5_H2, d, a, b, c, w2[0], MD5C21, MD5S21); MD5_STEP (MD5_H1, c, d, a, b, w2[3], MD5C22, MD5S22); MD5_STEP (MD5_H2, b, c, d, a, w3[2], MD5C23, MD5S23); MD5_STEP (MD5_H1, a, b, c, d, w0[1], MD5C24, MD5S20); MD5_STEP (MD5_H2, d, a, b, c, w1[0], MD5C25, MD5S21); MD5_STEP (MD5_H1, c, d, a, b, w1[3], MD5C26, MD5S22); MD5_STEP (MD5_H2, b, c, d, a, w2[2], MD5C27, MD5S23); MD5_STEP (MD5_H1, a, b, c, d, w3[1], MD5C28, MD5S20); MD5_STEP (MD5_H2, d, a, b, c, w0[0], MD5C29, MD5S21); MD5_STEP (MD5_H1, c, d, a, b, w0[3], MD5C2a, MD5S22); MD5_STEP (MD5_H2, b, c, d, a, w1[2], MD5C2b, MD5S23); MD5_STEP (MD5_H1, a, b, c, d, w2[1], MD5C2c, MD5S20); MD5_STEP (MD5_H2, d, a, b, c, w3[0], MD5C2d, MD5S21); MD5_STEP (MD5_H1, c, d, a, b, w3[3], MD5C2e, MD5S22); MD5_STEP (MD5_H2, b, c, d, a, w0[2], MD5C2f, MD5S23); MD5_STEP (MD5_I , a, b, c, d, w0[0], MD5C30, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w1[3], MD5C31, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w3[2], MD5C32, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w1[1], MD5C33, MD5S33); MD5_STEP (MD5_I , a, b, c, d, w3[0], MD5C34, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w0[3], MD5C35, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w2[2], MD5C36, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w0[1], MD5C37, MD5S33); MD5_STEP (MD5_I , a, b, c, d, w2[0], MD5C38, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w3[3], MD5C39, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w1[2], MD5C3a, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w3[1], MD5C3b, MD5S33); MD5_STEP (MD5_I , a, b, c, d, w1[0], MD5C3c, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w2[3], MD5C3d, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w0[2], MD5C3e, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w2[1], MD5C3f, MD5S33); a += MD5M_A; b += MD5M_B; c += MD5M_C; d += MD5M_D; /* * sha1 */ u32x w0_t = uint_to_hex_lower8_le ((a >> 8) & 255) << 0 | uint_to_hex_lower8_le ((a >> 0) & 255) << 16; u32x w1_t = uint_to_hex_lower8_le ((a >> 24) & 255) << 0 | uint_to_hex_lower8_le ((a >> 16) & 255) << 16; u32x w2_t = uint_to_hex_lower8_le ((b >> 8) & 255) << 0 | uint_to_hex_lower8_le ((b >> 0) & 255) << 16; u32x w3_t = uint_to_hex_lower8_le ((b >> 24) & 255) << 0 | uint_to_hex_lower8_le ((b >> 16) & 255) << 16; u32x w4_t = uint_to_hex_lower8_le ((c >> 8) & 255) << 0 | uint_to_hex_lower8_le ((c >> 0) & 255) << 16; u32x w5_t = uint_to_hex_lower8_le ((c >> 24) & 255) << 0 | uint_to_hex_lower8_le ((c >> 16) & 255) << 16; u32x w6_t = uint_to_hex_lower8_le ((d >> 8) & 255) << 0 | uint_to_hex_lower8_le ((d >> 0) & 255) << 16; u32x w7_t = uint_to_hex_lower8_le ((d >> 24) & 255) << 0 | uint_to_hex_lower8_le ((d >> 16) & 255) << 16; u32x w8_t = 0x80000000; u32x w9_t = 0; u32x wa_t = 0; u32x wb_t = 0; u32x wc_t = 0; u32x wd_t = 0; u32x we_t = 0; u32x wf_t = 32 * 8; a = SHA1M_A; b = SHA1M_B; c = SHA1M_C; d = SHA1M_D; e = SHA1M_E; #undef K #define K SHA1C00 SHA1_STEP (SHA1_F0o, a, b, c, d, e, w0_t); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w1_t); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w2_t); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w3_t); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w4_t); SHA1_STEP (SHA1_F0o, a, b, c, d, e, w5_t); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w6_t); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w7_t); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w8_t); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w9_t); SHA1_STEP (SHA1_F0o, a, b, c, d, e, wa_t); SHA1_STEP (SHA1_F0o, e, a, b, c, d, wb_t); SHA1_STEP (SHA1_F0o, d, e, a, b, c, wc_t); SHA1_STEP (SHA1_F0o, c, d, e, a, b, wd_t); SHA1_STEP (SHA1_F0o, b, c, d, e, a, we_t); SHA1_STEP (SHA1_F0o, a, b, c, d, e, wf_t); w0_t = hc_rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w0_t); w1_t = hc_rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w1_t); w2_t = hc_rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w2_t); w3_t = hc_rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w3_t); #undef K #define K SHA1C01 w4_t = hc_rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w4_t); w5_t = hc_rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w5_t); w6_t = hc_rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w6_t); w7_t = hc_rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w7_t); w8_t = hc_rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w8_t); w9_t = hc_rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w9_t); wa_t = hc_rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wa_t); wb_t = hc_rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wb_t); wc_t = hc_rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wc_t); wd_t = hc_rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wd_t); we_t = hc_rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, we_t); wf_t = hc_rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wf_t); w0_t = hc_rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w0_t); w1_t = hc_rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w1_t); w2_t = hc_rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w2_t); w3_t = hc_rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w3_t); w4_t = hc_rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w4_t); w5_t = hc_rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w5_t); w6_t = hc_rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w6_t); w7_t = hc_rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w7_t); #undef K #define K SHA1C02 w8_t = hc_rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w8_t); w9_t = hc_rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w9_t); wa_t = hc_rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wa_t); wb_t = hc_rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wb_t); wc_t = hc_rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wc_t); wd_t = hc_rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, wd_t); we_t = hc_rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, we_t); wf_t = hc_rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wf_t); w0_t = hc_rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w0_t); w1_t = hc_rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w1_t); w2_t = hc_rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w2_t); w3_t = hc_rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w3_t); w4_t = hc_rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w4_t); w5_t = hc_rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w5_t); w6_t = hc_rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w6_t); w7_t = hc_rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w7_t); w8_t = hc_rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w8_t); w9_t = hc_rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w9_t); wa_t = hc_rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wa_t); wb_t = hc_rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wb_t); #undef K #define K SHA1C03 wc_t = hc_rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wc_t); wd_t = hc_rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wd_t); we_t = hc_rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, we_t); wf_t = hc_rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wf_t); w0_t = hc_rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w0_t); w1_t = hc_rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w1_t); w2_t = hc_rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w2_t); w3_t = hc_rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w3_t); w4_t = hc_rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w4_t); w5_t = hc_rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w5_t); w6_t = hc_rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w6_t); w7_t = hc_rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w7_t); w8_t = hc_rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w8_t); w9_t = hc_rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w9_t); wa_t = hc_rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa_t); wb_t = hc_rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wb_t); wc_t = hc_rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wc_t); wd_t = hc_rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wd_t); we_t = hc_rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, we_t); wf_t = hc_rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wf_t); COMPARE_M_SIMD (d, e, c, b); } } KERNEL_FQ void m04700_m08 (KERN_ATTR_RULES ()) { } KERNEL_FQ void m04700_m16 (KERN_ATTR_RULES ()) { } KERNEL_FQ void m04700_s04 (KERN_ATTR_RULES ()) { /** * modifier */ const u64 gid = get_global_id (0); const u64 lid = get_local_id (0); const u64 lsz = get_local_size (0); /** * bin2asc table */ LOCAL_AS u32 l_bin2asc[256]; for (u32 i = lid; i < 256; i += lsz) { const u32 i0 = (i >> 0) & 15; const u32 i1 = (i >> 4) & 15; l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'a' - 10 + i0) << 0 | ((i1 < 10) ? '0' + i1 : 'a' - 10 + i1) << 8; } barrier (CLK_LOCAL_MEM_FENCE); if (gid >= gid_max) return; /** * base */ u32 pw_buf0[4]; u32 pw_buf1[4]; pw_buf0[0] = pws[gid].i[0]; pw_buf0[1] = pws[gid].i[1]; pw_buf0[2] = pws[gid].i[2]; pw_buf0[3] = pws[gid].i[3]; pw_buf1[0] = pws[gid].i[4]; pw_buf1[1] = pws[gid].i[5]; pw_buf1[2] = pws[gid].i[6]; pw_buf1[3] = pws[gid].i[7]; const u32 pw_len = pws[gid].pw_len & 63; /** * 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] }; /** * reverse */ const u32 e_rev = hc_rotl32_S (search[1], 2u); /** * loop */ for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) { u32x w0[4] = { 0 }; u32x w1[4] = { 0 }; u32x w2[4] = { 0 }; u32x w3[4] = { 0 }; const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1); append_0x80_2x4_VV (w0, w1, out_len); w3[2] = out_len * 8; w3[3] = 0; /** * md5 */ u32x a = MD5M_A; u32x b = MD5M_B; u32x c = MD5M_C; u32x d = MD5M_D; u32x e = 0; MD5_STEP (MD5_Fo, a, b, c, d, w0[0], MD5C00, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w0[1], MD5C01, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w0[2], MD5C02, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w0[3], MD5C03, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, w1[0], MD5C04, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w1[1], MD5C05, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w1[2], MD5C06, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w1[3], MD5C07, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, w2[0], MD5C08, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w2[1], MD5C09, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w2[2], MD5C0a, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w2[3], MD5C0b, MD5S03); MD5_STEP (MD5_Fo, a, b, c, d, w3[0], MD5C0c, MD5S00); MD5_STEP (MD5_Fo, d, a, b, c, w3[1], MD5C0d, MD5S01); MD5_STEP (MD5_Fo, c, d, a, b, w3[2], MD5C0e, MD5S02); MD5_STEP (MD5_Fo, b, c, d, a, w3[3], MD5C0f, MD5S03); MD5_STEP (MD5_Go, a, b, c, d, w0[1], MD5C10, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w1[2], MD5C11, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w2[3], MD5C12, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w0[0], MD5C13, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, w1[1], MD5C14, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w2[2], MD5C15, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w3[3], MD5C16, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w1[0], MD5C17, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, w2[1], MD5C18, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w3[2], MD5C19, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w0[3], MD5C1a, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w2[0], MD5C1b, MD5S13); MD5_STEP (MD5_Go, a, b, c, d, w3[1], MD5C1c, MD5S10); MD5_STEP (MD5_Go, d, a, b, c, w0[2], MD5C1d, MD5S11); MD5_STEP (MD5_Go, c, d, a, b, w1[3], MD5C1e, MD5S12); MD5_STEP (MD5_Go, b, c, d, a, w3[0], MD5C1f, MD5S13); u32x t; MD5_STEP (MD5_H1, a, b, c, d, w1[1], MD5C20, MD5S20); MD5_STEP (MD5_H2, d, a, b, c, w2[0], MD5C21, MD5S21); MD5_STEP (MD5_H1, c, d, a, b, w2[3], MD5C22, MD5S22); MD5_STEP (MD5_H2, b, c, d, a, w3[2], MD5C23, MD5S23); MD5_STEP (MD5_H1, a, b, c, d, w0[1], MD5C24, MD5S20); MD5_STEP (MD5_H2, d, a, b, c, w1[0], MD5C25, MD5S21); MD5_STEP (MD5_H1, c, d, a, b, w1[3], MD5C26, MD5S22); MD5_STEP (MD5_H2, b, c, d, a, w2[2], MD5C27, MD5S23); MD5_STEP (MD5_H1, a, b, c, d, w3[1], MD5C28, MD5S20); MD5_STEP (MD5_H2, d, a, b, c, w0[0], MD5C29, MD5S21); MD5_STEP (MD5_H1, c, d, a, b, w0[3], MD5C2a, MD5S22); MD5_STEP (MD5_H2, b, c, d, a, w1[2], MD5C2b, MD5S23); MD5_STEP (MD5_H1, a, b, c, d, w2[1], MD5C2c, MD5S20); MD5_STEP (MD5_H2, d, a, b, c, w3[0], MD5C2d, MD5S21); MD5_STEP (MD5_H1, c, d, a, b, w3[3], MD5C2e, MD5S22); MD5_STEP (MD5_H2, b, c, d, a, w0[2], MD5C2f, MD5S23); MD5_STEP (MD5_I , a, b, c, d, w0[0], MD5C30, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w1[3], MD5C31, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w3[2], MD5C32, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w1[1], MD5C33, MD5S33); MD5_STEP (MD5_I , a, b, c, d, w3[0], MD5C34, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w0[3], MD5C35, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w2[2], MD5C36, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w0[1], MD5C37, MD5S33); MD5_STEP (MD5_I , a, b, c, d, w2[0], MD5C38, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w3[3], MD5C39, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w1[2], MD5C3a, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w3[1], MD5C3b, MD5S33); MD5_STEP (MD5_I , a, b, c, d, w1[0], MD5C3c, MD5S30); MD5_STEP (MD5_I , d, a, b, c, w2[3], MD5C3d, MD5S31); MD5_STEP (MD5_I , c, d, a, b, w0[2], MD5C3e, MD5S32); MD5_STEP (MD5_I , b, c, d, a, w2[1], MD5C3f, MD5S33); a += MD5M_A; b += MD5M_B; c += MD5M_C; d += MD5M_D; /* * sha1 */ u32x w0_t = uint_to_hex_lower8_le ((a >> 8) & 255) << 0 | uint_to_hex_lower8_le ((a >> 0) & 255) << 16; u32x w1_t = uint_to_hex_lower8_le ((a >> 24) & 255) << 0 | uint_to_hex_lower8_le ((a >> 16) & 255) << 16; u32x w2_t = uint_to_hex_lower8_le ((b >> 8) & 255) << 0 | uint_to_hex_lower8_le ((b >> 0) & 255) << 16; u32x w3_t = uint_to_hex_lower8_le ((b >> 24) & 255) << 0 | uint_to_hex_lower8_le ((b >> 16) & 255) << 16; u32x w4_t = uint_to_hex_lower8_le ((c >> 8) & 255) << 0 | uint_to_hex_lower8_le ((c >> 0) & 255) << 16; u32x w5_t = uint_to_hex_lower8_le ((c >> 24) & 255) << 0 | uint_to_hex_lower8_le ((c >> 16) & 255) << 16; u32x w6_t = uint_to_hex_lower8_le ((d >> 8) & 255) << 0 | uint_to_hex_lower8_le ((d >> 0) & 255) << 16; u32x w7_t = uint_to_hex_lower8_le ((d >> 24) & 255) << 0 | uint_to_hex_lower8_le ((d >> 16) & 255) << 16; u32x w8_t = 0x80000000; u32x w9_t = 0; u32x wa_t = 0; u32x wb_t = 0; u32x wc_t = 0; u32x wd_t = 0; u32x we_t = 0; u32x wf_t = 32 * 8; a = SHA1M_A; b = SHA1M_B; c = SHA1M_C; d = SHA1M_D; e = SHA1M_E; #undef K #define K SHA1C00 SHA1_STEP (SHA1_F0o, a, b, c, d, e, w0_t); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w1_t); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w2_t); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w3_t); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w4_t); SHA1_STEP (SHA1_F0o, a, b, c, d, e, w5_t); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w6_t); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w7_t); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w8_t); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w9_t); SHA1_STEP (SHA1_F0o, a, b, c, d, e, wa_t); SHA1_STEP (SHA1_F0o, e, a, b, c, d, wb_t); SHA1_STEP (SHA1_F0o, d, e, a, b, c, wc_t); SHA1_STEP (SHA1_F0o, c, d, e, a, b, wd_t); SHA1_STEP (SHA1_F0o, b, c, d, e, a, we_t); SHA1_STEP (SHA1_F0o, a, b, c, d, e, wf_t); w0_t = hc_rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, e, a, b, c, d, w0_t); w1_t = hc_rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, d, e, a, b, c, w1_t); w2_t = hc_rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, c, d, e, a, b, w2_t); w3_t = hc_rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, b, c, d, e, a, w3_t); #undef K #define K SHA1C01 w4_t = hc_rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w4_t); w5_t = hc_rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w5_t); w6_t = hc_rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w6_t); w7_t = hc_rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w7_t); w8_t = hc_rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w8_t); w9_t = hc_rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w9_t); wa_t = hc_rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wa_t); wb_t = hc_rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wb_t); wc_t = hc_rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wc_t); wd_t = hc_rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wd_t); we_t = hc_rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, we_t); wf_t = hc_rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wf_t); w0_t = hc_rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w0_t); w1_t = hc_rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w1_t); w2_t = hc_rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w2_t); w3_t = hc_rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w3_t); w4_t = hc_rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w4_t); w5_t = hc_rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w5_t); w6_t = hc_rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w6_t); w7_t = hc_rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w7_t); #undef K #define K SHA1C02 w8_t = hc_rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w8_t); w9_t = hc_rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w9_t); wa_t = hc_rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wa_t); wb_t = hc_rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wb_t); wc_t = hc_rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wc_t); wd_t = hc_rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, wd_t); we_t = hc_rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, we_t); wf_t = hc_rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, wf_t); w0_t = hc_rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w0_t); w1_t = hc_rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w1_t); w2_t = hc_rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w2_t); w3_t = hc_rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w3_t); w4_t = hc_rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w4_t); w5_t = hc_rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, w5_t); w6_t = hc_rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, w6_t); w7_t = hc_rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, a, b, c, d, e, w7_t); w8_t = hc_rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, e, a, b, c, d, w8_t); w9_t = hc_rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, d, e, a, b, c, w9_t); wa_t = hc_rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, c, d, e, a, b, wa_t); wb_t = hc_rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, b, c, d, e, a, wb_t); #undef K #define K SHA1C03 wc_t = hc_rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wc_t); wd_t = hc_rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wd_t); we_t = hc_rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, we_t); wf_t = hc_rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, wf_t); w0_t = hc_rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w0_t); w1_t = hc_rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w1_t); w2_t = hc_rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w2_t); w3_t = hc_rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w3_t); w4_t = hc_rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w4_t); w5_t = hc_rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, w5_t); w6_t = hc_rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, w6_t); w7_t = hc_rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, w7_t); w8_t = hc_rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, w8_t); w9_t = hc_rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, w9_t); wa_t = hc_rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wa_t); wb_t = hc_rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, a, b, c, d, e, wb_t); if (MATCHES_NONE_VS (e, e_rev)) continue; wc_t = hc_rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, e, a, b, c, d, wc_t); wd_t = hc_rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, d, e, a, b, c, wd_t); we_t = hc_rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, c, d, e, a, b, we_t); wf_t = hc_rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, b, c, d, e, a, wf_t); COMPARE_S_SIMD (d, e, c, b); } } KERNEL_FQ void m04700_s08 (KERN_ATTR_RULES ()) { } KERNEL_FQ void m04700_s16 (KERN_ATTR_RULES ()) { }