/** * Author......: See docs/credits.txt * License.....: MIT */ #define NEW_SIMD_CODE #ifdef KERNEL_STATIC #include M2S(INCLUDE_PATH/inc_vendor.h) #include M2S(INCLUDE_PATH/inc_types.h) #include M2S(INCLUDE_PATH/inc_platform.cl) #include M2S(INCLUDE_PATH/inc_common.cl) #include M2S(INCLUDE_PATH/inc_rp_optimized.h) #include M2S(INCLUDE_PATH/inc_rp_optimized.cl) #include M2S(INCLUDE_PATH/inc_simd.cl) #include M2S(INCLUDE_PATH/inc_hash_md4.cl) #endif KERNEL_FQ void m00900_m04 (KERN_ATTR_RULES ()) { /** * modifier */ const u64 lid = get_local_id (0); /** * base */ const u64 gid = get_global_id (0); if (gid >= GID_CNT) return; 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_optimized (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; u32x a = MD4M_A; u32x b = MD4M_B; u32x c = MD4M_C; u32x d = MD4M_D; MD4_STEP (MD4_Fo, a, b, c, d, w0[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w0[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w0[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w0[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w1[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w1[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w1[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w1[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w2[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w2[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w2[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w2[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w3[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w3[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w3[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w3[3], MD4C00, MD4S03); MD4_STEP (MD4_Go, a, b, c, d, w0[0], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[0], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[0], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[0], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[1], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[1], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[1], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[1], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[2], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[2], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[2], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[2], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[3], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[3], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[3], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[3], MD4C01, MD4S13); MD4_STEP (MD4_H , a, b, c, d, w0[0], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[0], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[0], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[0], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[2], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[2], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[2], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[2], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[1], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[1], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[1], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[1], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[3], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[3], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[3], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[3], MD4C02, MD4S23); COMPARE_M_SIMD (a, d, c, b); } } KERNEL_FQ void m00900_m08 (KERN_ATTR_RULES ()) { } KERNEL_FQ void m00900_m16 (KERN_ATTR_RULES ()) { } KERNEL_FQ void m00900_s04 (KERN_ATTR_RULES ()) { /** * modifier */ const u64 lid = get_local_id (0); /** * base */ const u64 gid = get_global_id (0); if (gid >= GID_CNT) return; 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_HOST].digest_buf[DGST_R0], digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R1], digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R2], digests_buf[DIGESTS_OFFSET_HOST].digest_buf[DGST_R3] }; /** * 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_optimized (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; u32x a = MD4M_A; u32x b = MD4M_B; u32x c = MD4M_C; u32x d = MD4M_D; MD4_STEP (MD4_Fo, a, b, c, d, w0[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w0[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w0[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w0[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w1[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w1[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w1[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w1[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w2[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w2[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w2[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w2[3], MD4C00, MD4S03); MD4_STEP (MD4_Fo, a, b, c, d, w3[0], MD4C00, MD4S00); MD4_STEP (MD4_Fo, d, a, b, c, w3[1], MD4C00, MD4S01); MD4_STEP (MD4_Fo, c, d, a, b, w3[2], MD4C00, MD4S02); MD4_STEP (MD4_Fo, b, c, d, a, w3[3], MD4C00, MD4S03); MD4_STEP (MD4_Go, a, b, c, d, w0[0], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[0], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[0], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[0], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[1], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[1], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[1], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[1], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[2], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[2], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[2], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[2], MD4C01, MD4S13); MD4_STEP (MD4_Go, a, b, c, d, w0[3], MD4C01, MD4S10); MD4_STEP (MD4_Go, d, a, b, c, w1[3], MD4C01, MD4S11); MD4_STEP (MD4_Go, c, d, a, b, w2[3], MD4C01, MD4S12); MD4_STEP (MD4_Go, b, c, d, a, w3[3], MD4C01, MD4S13); MD4_STEP (MD4_H , a, b, c, d, w0[0], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[0], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[0], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[0], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[2], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[2], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[2], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[2], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[1], MD4C02, MD4S20); MD4_STEP (MD4_H , d, a, b, c, w2[1], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[1], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[1], MD4C02, MD4S23); MD4_STEP (MD4_H , a, b, c, d, w0[3], MD4C02, MD4S20); if (MATCHES_NONE_VS (a, search[0])) continue; MD4_STEP (MD4_H , d, a, b, c, w2[3], MD4C02, MD4S21); MD4_STEP (MD4_H , c, d, a, b, w1[3], MD4C02, MD4S22); MD4_STEP (MD4_H , b, c, d, a, w3[3], MD4C02, MD4S23); COMPARE_S_SIMD (a, d, c, b); } } KERNEL_FQ void m00900_s08 (KERN_ATTR_RULES ()) { } KERNEL_FQ void m00900_s16 (KERN_ATTR_RULES ()) { }