#include #include #include "luffa-hash-2way.h" #include #if defined(__AVX2__) #include "simd-utils.h" #define uint32 uint32_t /* initial values of chaining variables */ static const uint32_t IV[40] __attribute((aligned(64))) = { 0xdbf78465,0x4eaa6fb4,0x44b051e0,0x6d251e69, 0xdef610bb,0xee058139,0x90152df4,0x6e292011, 0xde099fa3,0x70eee9a0,0xd9d2f256,0xc3b44b95, 0x746cd581,0xcf1ccf0e,0x8fc944b3,0x5d9b0557, 0xad659c05,0x04016ce5,0x5dba5781,0xf7efc89d, 0x8b264ae7,0x24aa230a,0x666d1836,0x0306194f, 0x204b1f67,0xe571f7d7,0x36d79cce,0x858075d5, 0x7cde72ce,0x14bcb808,0x57e9e923,0x35870c6a, 0xaffb4363,0xc825b7c7,0x5ec41e22,0x6c68e9be, 0x03e86cea,0xb07224cc,0x0fc688f1,0xf5df3999 }; /* Round Constants */ static const uint32_t CNS_INIT[128] __attribute((aligned(64))) = { 0xb213afa5,0xfc20d9d2,0xb6de10ed,0x303994a6, 0xe028c9bf,0xe25e72c1,0x01685f3d,0xe0337818, 0xc84ebe95,0x34552e25,0x70f47aae,0xc0e65299, 0x44756f91,0xe623bb72,0x05a17cf4,0x441ba90d, 0x4e608a22,0x7ad8818f,0x0707a3d4,0x6cc33a12, 0x7e8fce32,0x5c58a4a4,0xbd09caca,0x7f34d442, 0x56d858fe,0x8438764a,0x1c1e8f51,0xdc56983e, 0x956548be,0x1e38e2e7,0xf4272b28,0x9389217f, 0x343b138f,0xbb6de032,0x707a3d45,0x1e00108f, 0xfe191be2,0x78e38b9d,0x144ae5cc,0xe5a8bce6, 0xd0ec4e3d,0xedb780c8,0xaeb28562,0x7800423d, 0x3cb226e5,0x27586719,0xfaa7ae2b,0x5274baf4, 0x2ceb4882,0xd9847356,0xbaca1589,0x8f5b7882, 0x5944a28e,0x36eda57f,0x2e48f1c1,0x26889ba7, 0xb3ad2208,0xa2c78434,0x40a46f3e,0x96e1db12, 0xa1c4c355,0x703aace7,0xb923c704,0x9a226e9d, 0x00000000,0x00000000,0x00000000,0xf0d2e9e3, 0x00000000,0x00000000,0x00000000,0x5090d577, 0x00000000,0x00000000,0x00000000,0xac11d7fa, 0x00000000,0x00000000,0x00000000,0x2d1925ab, 0x00000000,0x00000000,0x00000000,0x1bcb66f2, 0x00000000,0x00000000,0x00000000,0xb46496ac, 0x00000000,0x00000000,0x00000000,0x6f2d9bc9, 0x00000000,0x00000000,0x00000000,0xd1925ab0, 0x00000000,0x00000000,0x00000000,0x78602649, 0x00000000,0x00000000,0x00000000,0x29131ab6, 0x00000000,0x00000000,0x00000000,0x8edae952, 0x00000000,0x00000000,0x00000000,0x0fc053c3, 0x00000000,0x00000000,0x00000000,0x3b6ba548, 0x00000000,0x00000000,0x00000000,0x3f014f0c, 0x00000000,0x00000000,0x00000000,0xedae9520, 0x00000000,0x00000000,0x00000000,0xfc053c31 }; #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__) #define cns4w(i) mm512_bcast_m128( ( (__m128i*)CNS_INIT)[i] ) #define ADD_CONSTANT4W( a, b, c0, c1 ) \ a = _mm512_xor_si512( a, c0 ); \ b = _mm512_xor_si512( b, c1 ); #define MULT24W( a0, a1 ) \ { \ __m512i b = _mm512_xor_si512( a0, \ _mm512_maskz_shuffle_epi32( 0xbbbb, a1, 0x10 ) ); \ a0 = _mm512_alignr_epi8( a1, b, 4 ); \ a1 = _mm512_alignr_epi8( b, a1, 4 ); \ } #define SUBCRUMB4W( a0, a1, a2, a3 ) \ { \ __m512i t = a0; \ a0 = mm512_xoror( a3, a0, a1 ); \ a2 = _mm512_xor_si512( a2, a3 ); \ a1 = _mm512_ternarylogic_epi64( a1, a3, t, 0x87 ); /* a1 xnor (a3 & t) */ \ a3 = mm512_xorand( a2, a3, t ); \ a2 = mm512_xorand( a1, a2, a0); \ a1 = _mm512_or_si512( a1, a3 ); \ a3 = _mm512_xor_si512( a3, a2 ); \ t = _mm512_xor_si512( t, a1 ); \ a2 = _mm512_and_si512( a2, a1 ); \ a1 = mm512_xnor( a1, a0 ); \ a0 = t; \ } #define MIXWORD4W( a, b ) \ b = _mm512_xor_si512( a, b ); \ a = _mm512_xor_si512( b, _mm512_rol_epi32( a, 2 ) ); \ b = _mm512_xor_si512( a, _mm512_rol_epi32( b, 14 ) ); \ a = _mm512_xor_si512( b, _mm512_rol_epi32( a, 10 ) ); \ b = _mm512_rol_epi32( b, 1 ); #define STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, c0, c1 ) \ SUBCRUMB4W( x0, x1, x2, x3 ); \ SUBCRUMB4W( x5, x6, x7, x4 ); \ MIXWORD4W( x0, x4 ); \ MIXWORD4W( x1, x5 ); \ MIXWORD4W( x2, x6 ); \ MIXWORD4W( x3, x7 ); \ ADD_CONSTANT4W( x0, x4, c0, c1 ); #define STEP_PART24W( a0, a1, t0, t1, c0, c1 ) \ t0 = _mm512_shuffle_epi32( a1, 147 ); \ a1 = _mm512_unpacklo_epi32( t0, a0 ); \ t0 = _mm512_unpackhi_epi32( t0, a0 ); \ t1 = _mm512_shuffle_epi32( t0, 78 ); \ a0 = _mm512_shuffle_epi32( a1, 78 ); \ SUBCRUMB4W( t1, t0, a0, a1 ); \ t0 = _mm512_unpacklo_epi32( t0, t1 ); \ a1 = _mm512_unpacklo_epi32( a1, a0 ); \ a0 = _mm512_unpackhi_epi64( a1, t0 ); \ a1 = _mm512_unpacklo_epi64( a1, t0 ); \ a1 = _mm512_shuffle_epi32( a1, 57 ); \ MIXWORD4W( a0, a1 ); \ ADD_CONSTANT4W( a0, a1, c0, c1 ); #define NMLTOM10244W(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3)\ s1 = _mm512_unpackhi_epi32( r3, r2 ); \ q1 = _mm512_unpackhi_epi32( p3, p2 ); \ s3 = _mm512_unpacklo_epi32( r3, r2 ); \ q3 = _mm512_unpacklo_epi32( p3, p2 ); \ r3 = _mm512_unpackhi_epi32( r1, r0 ); \ r1 = _mm512_unpacklo_epi32( r1, r0 ); \ p3 = _mm512_unpackhi_epi32( p1, p0 ); \ p1 = _mm512_unpacklo_epi32( p1, p0 ); \ s0 = _mm512_unpackhi_epi64( s1, r3 ); \ q0 = _mm512_unpackhi_epi64( q1 ,p3 ); \ s1 = _mm512_unpacklo_epi64( s1, r3 ); \ q1 = _mm512_unpacklo_epi64( q1, p3 ); \ s2 = _mm512_unpackhi_epi64( s3, r1 ); \ q2 = _mm512_unpackhi_epi64( q3, p1 ); \ s3 = _mm512_unpacklo_epi64( s3, r1 ); \ q3 = _mm512_unpacklo_epi64( q3, p1 ); #define MIXTON10244W(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3)\ NMLTOM10244W(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3); void rnd512_4way( luffa_4way_context *state, const __m512i *msg ) { __m512i t0, t1; __m512i *chainv = state->chainv; __m512i x0, x1, x2, x3, x4, x5, x6, x7; t0 = mm512_xor3( chainv[0], chainv[2], chainv[4] ); t1 = mm512_xor3( chainv[1], chainv[3], chainv[5] ); t0 = mm512_xor3( t0, chainv[6], chainv[8] ); t1 = mm512_xor3( t1, chainv[7], chainv[9] ); MULT24W( t0, t1 ); chainv[0] = _mm512_xor_si512( chainv[0], t0 ); chainv[1] = _mm512_xor_si512( chainv[1], t1 ); chainv[2] = _mm512_xor_si512( chainv[2], t0 ); chainv[3] = _mm512_xor_si512( chainv[3], t1 ); chainv[4] = _mm512_xor_si512( chainv[4], t0 ); chainv[5] = _mm512_xor_si512( chainv[5], t1 ); chainv[6] = _mm512_xor_si512( chainv[6], t0 ); chainv[7] = _mm512_xor_si512( chainv[7], t1 ); chainv[8] = _mm512_xor_si512( chainv[8], t0 ); chainv[9] = _mm512_xor_si512( chainv[9], t1 ); t0 = chainv[0]; t1 = chainv[1]; MULT24W( chainv[0], chainv[1] ); chainv[0] = _mm512_xor_si512( chainv[0], chainv[2] ); chainv[1] = _mm512_xor_si512( chainv[1], chainv[3] ); MULT24W( chainv[2], chainv[3] ); chainv[2] = _mm512_xor_si512(chainv[2], chainv[4]); chainv[3] = _mm512_xor_si512(chainv[3], chainv[5]); MULT24W( chainv[4], chainv[5] ); chainv[4] = _mm512_xor_si512(chainv[4], chainv[6]); chainv[5] = _mm512_xor_si512(chainv[5], chainv[7]); MULT24W( chainv[6], chainv[7] ); chainv[6] = _mm512_xor_si512(chainv[6], chainv[8]); chainv[7] = _mm512_xor_si512(chainv[7], chainv[9]); MULT24W( chainv[8], chainv[9] ); t0 = chainv[8] = _mm512_xor_si512( chainv[8], t0 ); t1 = chainv[9] = _mm512_xor_si512( chainv[9], t1 ); MULT24W( chainv[8], chainv[9] ); chainv[8] = _mm512_xor_si512( chainv[8], chainv[6] ); chainv[9] = _mm512_xor_si512( chainv[9], chainv[7] ); MULT24W( chainv[6], chainv[7] ); chainv[6] = _mm512_xor_si512( chainv[6], chainv[4] ); chainv[7] = _mm512_xor_si512( chainv[7], chainv[5] ); MULT24W( chainv[4], chainv[5] ); chainv[4] = _mm512_xor_si512( chainv[4], chainv[2] ); chainv[5] = _mm512_xor_si512( chainv[5], chainv[3] ); MULT24W( chainv[2], chainv[3] ); chainv[2] = _mm512_xor_si512( chainv[2], chainv[0] ); chainv[3] = _mm512_xor_si512( chainv[3], chainv[1] ); MULT24W( chainv[0], chainv[1] ); chainv[0] = _mm512_xor_si512( chainv[0], t0 ); chainv[1] = _mm512_xor_si512( chainv[1], t1 ); if ( msg ) { __m512i msg0, msg1; msg0 = _mm512_shuffle_epi32( msg[0], 27 ); msg1 = _mm512_shuffle_epi32( msg[1], 27 ); chainv[0] = _mm512_xor_si512( chainv[0], msg0 ); chainv[1] = _mm512_xor_si512( chainv[1], msg1 ); MULT24W( msg0, msg1 ); chainv[2] = _mm512_xor_si512( chainv[2], msg0 ); chainv[3] = _mm512_xor_si512( chainv[3], msg1 ); MULT24W( msg0, msg1 ); chainv[4] = _mm512_xor_si512( chainv[4], msg0 ); chainv[5] = _mm512_xor_si512( chainv[5], msg1 ); MULT24W( msg0, msg1 ); chainv[6] = _mm512_xor_si512( chainv[6], msg0 ); chainv[7] = _mm512_xor_si512( chainv[7], msg1 ); MULT24W( msg0, msg1); chainv[8] = _mm512_xor_si512( chainv[8], msg0 ); chainv[9] = _mm512_xor_si512( chainv[9], msg1 ); } chainv[3] = _mm512_rol_epi32( chainv[3], 1 ); chainv[5] = _mm512_rol_epi32( chainv[5], 2 ); chainv[7] = _mm512_rol_epi32( chainv[7], 3 ); chainv[9] = _mm512_rol_epi32( chainv[9], 4 ); NMLTOM10244W( chainv[0], chainv[2], chainv[4], chainv[6], x0, x1, x2, x3, chainv[1], chainv[3], chainv[5], chainv[7], x4, x5, x6, x7 ); STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, cns4w( 0), cns4w( 1) ); STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, cns4w( 2), cns4w( 3) ); STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, cns4w( 4), cns4w( 5) ); STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, cns4w( 6), cns4w( 7) ); STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, cns4w( 8), cns4w( 9) ); STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, cns4w(10), cns4w(11) ); STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, cns4w(12), cns4w(13) ); STEP_PART4W( x0, x1, x2, x3, x4, x5, x6, x7, cns4w(14), cns4w(15) ); MIXTON10244W( x0, x1, x2, x3, chainv[0], chainv[2], chainv[4], chainv[6], x4, x5, x6, x7, chainv[1], chainv[3], chainv[5], chainv[7] ); /* Process last 256-bit block */ STEP_PART24W( chainv[8], chainv[9], t0, t1, cns4w(16), cns4w(17) ); STEP_PART24W( chainv[8], chainv[9], t0, t1, cns4w(18), cns4w(19) ); STEP_PART24W( chainv[8], chainv[9], t0, t1, cns4w(20), cns4w(21) ); STEP_PART24W( chainv[8], chainv[9], t0, t1, cns4w(22), cns4w(23) ); STEP_PART24W( chainv[8], chainv[9], t0, t1, cns4w(24), cns4w(25) ); STEP_PART24W( chainv[8], chainv[9], t0, t1, cns4w(26), cns4w(27) ); STEP_PART24W( chainv[8], chainv[9], t0, t1, cns4w(28), cns4w(29) ); STEP_PART24W( chainv[8], chainv[9], t0, t1, cns4w(30), cns4w(31) ); } void finalization512_4way( luffa_4way_context *state, uint32 *b ) { uint32_t hash[8*4] __attribute((aligned(128))); __m512i* chainv = state->chainv; __m512i t[2]; const __m512i shuff_bswap32 = mm512_bcast_m128( _mm_set_epi64x( 0x0c0d0e0f08090a0b, 0x0405060700010203 ) ); /*---- blank round with m=0 ----*/ rnd512_4way( state, NULL ); t[0] = mm512_xor3( chainv[0], chainv[2], chainv[4] ); t[1] = mm512_xor3( chainv[1], chainv[3], chainv[5] ); t[0] = mm512_xor3( t[0], chainv[6], chainv[8] ); t[1] = mm512_xor3( t[1], chainv[7], chainv[9] ); t[0] = _mm512_shuffle_epi32( t[0], 27 ); t[1] = _mm512_shuffle_epi32( t[1], 27 ); _mm512_store_si512( (__m512i*)&hash[ 0], t[0] ); _mm512_store_si512( (__m512i*)&hash[16], t[1] ); casti_m512i( b,0 ) = _mm512_shuffle_epi8( casti_m512i( hash,0 ), shuff_bswap32 ); casti_m512i( b,1 ) = _mm512_shuffle_epi8( casti_m512i( hash,1 ), shuff_bswap32 ); rnd512_4way( state, NULL ); t[0] = mm512_xor3( chainv[0], chainv[2], chainv[4] ); t[1] = mm512_xor3( chainv[1], chainv[3], chainv[5] ); t[0] = mm512_xor3( t[0], chainv[6], chainv[8] ); t[1] = mm512_xor3( t[1], chainv[7], chainv[9] ); t[0] = _mm512_shuffle_epi32( t[0], 27 ); t[1] = _mm512_shuffle_epi32( t[1], 27 ); _mm512_store_si512( (__m512i*)&hash[ 0], t[0] ); _mm512_store_si512( (__m512i*)&hash[16], t[1] ); casti_m512i( b,2 ) = _mm512_shuffle_epi8( casti_m512i( hash,0 ), shuff_bswap32 ); casti_m512i( b,3 ) = _mm512_shuffle_epi8( casti_m512i( hash,1 ), shuff_bswap32 ); } int luffa_4way_init( luffa_4way_context *state, int hashbitlen ) { state->hashbitlen = hashbitlen; __m128i *iv = (__m128i*)IV; state->chainv[0] = mm512_bcast_m128( iv[0] ); state->chainv[1] = mm512_bcast_m128( iv[1] ); state->chainv[2] = mm512_bcast_m128( iv[2] ); state->chainv[3] = mm512_bcast_m128( iv[3] ); state->chainv[4] = mm512_bcast_m128( iv[4] ); state->chainv[5] = mm512_bcast_m128( iv[5] ); state->chainv[6] = mm512_bcast_m128( iv[6] ); state->chainv[7] = mm512_bcast_m128( iv[7] ); state->chainv[8] = mm512_bcast_m128( iv[8] ); state->chainv[9] = mm512_bcast_m128( iv[9] ); ((__m512i*)state->buffer)[0] = m512_zero; ((__m512i*)state->buffer)[1] = m512_zero; return 0; } int luffa512_4way_init( luffa_4way_context *state ) { return luffa_4way_init( state, 512 ); } // Do not call luffa_update_close after having called luffa_update. // Once luffa_update has been called only call luffa_update or luffa_close. int luffa_4way_update( luffa_4way_context *state, const void *data, size_t len ) { __m512i *vdata = (__m512i*)data; __m512i *buffer = (__m512i*)state->buffer; __m512i msg[2]; int i; int blocks = (int)len >> 5; const __m512i shuff_bswap32 = mm512_bcast_m128( _mm_set_epi64x( 0x0c0d0e0f08090a0b, 0x0405060700010203 ) ); state->rembytes = (int)len & 0x1F; // full blocks for ( i = 0; i < blocks; i++, vdata+=2 ) { msg[0] = _mm512_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = _mm512_shuffle_epi8( vdata[ 1 ], shuff_bswap32 ); rnd512_4way( state, msg ); } // 16 byte partial block exists for 80 byte len // store in buffer for transform in final for midstate to work if ( state->rembytes ) { // remaining data bytes buffer[0] = _mm512_shuffle_epi8( vdata[0], shuff_bswap32 ); buffer[1] = mm512_bcast128lo_64( 0x0000000080000000 ); } return 0; } /* int luffa512_4way_update( luffa_4way_context *state, const void *data, size_t len ) { return luffa_4way_update( state, data, len ); } */ int luffa_4way_close( luffa_4way_context *state, void *hashval ) { __m512i *buffer = (__m512i*)state->buffer; __m512i msg[2]; // transform pad block if ( state->rembytes ) // not empty, data is in buffer rnd512_4way( state, buffer ); else { // empty pad block, constant data msg[0] = mm512_bcast128lo_64( 0x0000000080000000 ); msg[1] = m512_zero; rnd512_4way( state, msg ); } finalization512_4way( state, (uint32*)hashval ); if ( state->hashbitlen > 512 ) finalization512_4way( state, (uint32*)( hashval+32 ) ); return 0; } /* int luffa512_4way_close( luffa_4way_context *state, void *hashval ) { return luffa_4way_close( state, hashval ); } */ int luffa512_4way_full( luffa_4way_context *state, void *output, const void *data, size_t inlen ) { state->hashbitlen = 512; __m128i *iv = (__m128i*)IV; state->chainv[0] = mm512_bcast_m128( iv[0] ); state->chainv[1] = mm512_bcast_m128( iv[1] ); state->chainv[2] = mm512_bcast_m128( iv[2] ); state->chainv[3] = mm512_bcast_m128( iv[3] ); state->chainv[4] = mm512_bcast_m128( iv[4] ); state->chainv[5] = mm512_bcast_m128( iv[5] ); state->chainv[6] = mm512_bcast_m128( iv[6] ); state->chainv[7] = mm512_bcast_m128( iv[7] ); state->chainv[8] = mm512_bcast_m128( iv[8] ); state->chainv[9] = mm512_bcast_m128( iv[9] ); ((__m512i*)state->buffer)[0] = m512_zero; ((__m512i*)state->buffer)[1] = m512_zero; const __m512i *vdata = (__m512i*)data; __m512i msg[2]; int i; const int blocks = (int)( inlen >> 5 ); const __m512i shuff_bswap32 = mm512_bcast_m128( _mm_set_epi64x( 0x0c0d0e0f08090a0b, 0x0405060700010203 ) ); state->rembytes = inlen & 0x1F; // full blocks for ( i = 0; i < blocks; i++, vdata+=2 ) { msg[0] = _mm512_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = _mm512_shuffle_epi8( vdata[ 1 ], shuff_bswap32 ); rnd512_4way( state, msg ); } // 16 byte partial block exists for 80 byte len if ( state->rembytes ) { // padding of partial block msg[0] = _mm512_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = mm512_bcast128lo_64( 0x0000000080000000 ); rnd512_4way( state, msg ); } else { // empty pad block msg[0] = mm512_bcast128lo_64( 0x0000000080000000 ); msg[1] = m512_zero; rnd512_4way( state, msg ); } finalization512_4way( state, (uint32*)output ); if ( state->hashbitlen > 512 ) finalization512_4way( state, (uint32*)( output+64 ) ); return 0; } int luffa_4way_update_close( luffa_4way_context *state, void *output, const void *data, size_t inlen ) { // Optimized for integrals of 16 bytes, good for 64 and 80 byte len const __m512i *vdata = (__m512i*)data; __m512i msg[2]; int i; const int blocks = (int)( inlen >> 5 ); const __m512i shuff_bswap32 = mm512_bcast_m128( _mm_set_epi64x( 0x0c0d0e0f08090a0b, 0x0405060700010203 ) ); state->rembytes = inlen & 0x1F; // full blocks for ( i = 0; i < blocks; i++, vdata+=2 ) { msg[0] = _mm512_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = _mm512_shuffle_epi8( vdata[ 1 ], shuff_bswap32 ); rnd512_4way( state, msg ); } // 16 byte partial block exists for 80 byte len if ( state->rembytes ) { // padding of partial block msg[0] = _mm512_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = mm512_bcast128lo_64( 0x0000000080000000 ); rnd512_4way( state, msg ); } else { // empty pad block msg[0] = mm512_bcast128lo_64( 0x0000000080000000 ); msg[1] = m512_zero; rnd512_4way( state, msg ); } finalization512_4way( state, (uint32*)output ); if ( state->hashbitlen > 512 ) finalization512_4way( state, (uint32*)( output+64 ) ); return 0; } #endif // AVX512 #define cns(i) mm256_bcast_m128( ( (__m128i*)CNS_INIT)[i] ) #define ADD_CONSTANT( a, b, c0, c1 ) \ a = _mm256_xor_si256( a, c0 ); \ b = _mm256_xor_si256( b, c1 ); //TODO Enable for AVX10_256, not used with AVX512 or AVX10_512 #if defined(__AVX512VL__) #define MULT2( a0, a1 ) \ { \ __m256i b = _mm256_xor_si256( a0, \ _mm256_maskz_shuffle_epi32( 0xbb, a1, 0x10 ) ); \ a0 = _mm256_alignr_epi8( a1, b, 4 ); \ a1 = _mm256_alignr_epi8( b, a1, 4 ); \ } #define SUBCRUMB( a0, a1, a2, a3 ) \ { \ __m256i t = a0; \ a0 = mm256_xoror( a3, a0, a1 ); \ a2 = _mm256_xor_si256( a2, a3 ); \ a1 = _mm256_ternarylogic_epi64( a1, a3, t, 0x87 ); /* a1 xnor (a3 & t) */ \ a3 = mm256_xorand( a2, a3, t ); \ a2 = mm256_xorand( a1, a2, a0); \ a1 = _mm256_or_si256( a1, a3 ); \ a3 = _mm256_xor_si256( a3, a2 ); \ t = _mm256_xor_si256( t, a1 ); \ a2 = _mm256_and_si256( a2, a1 ); \ a1 = mm256_xnor( a1, a0 ); \ a0 = t; \ } #else #define MULT2( a0, a1 ) \ { \ __m256i b = _mm256_xor_si256( a0, _mm256_shuffle_epi32( \ _mm256_blend_epi32( a1, m256_zero, 0xee ), 0x10 ) ); \ a0 = _mm256_alignr_epi8( a1, b, 4 ); \ a1 = _mm256_alignr_epi8( b, a1, 4 ); \ } #define SUBCRUMB( a0, a1, a2, a3 ) \ { \ __m256i t = a0; \ a0 = _mm256_or_si256( a0, a1 ); \ a2 = _mm256_xor_si256( a2, a3 ); \ a1 = mm256_not( a1 ); \ a0 = _mm256_xor_si256( a0, a3 ); \ a3 = _mm256_and_si256( a3, t ); \ a1 = _mm256_xor_si256( a1, a3 ); \ a3 = _mm256_xor_si256( a3, a2 ); \ a2 = _mm256_and_si256( a2, a0 ); \ a0 = mm256_not( a0 ); \ a2 = _mm256_xor_si256( a2, a1 ); \ a1 = _mm256_or_si256( a1, a3 ); \ t = _mm256_xor_si256( t, a1 ); \ a3 = _mm256_xor_si256( a3, a2 ); \ a2 = _mm256_and_si256( a2, a1 ); \ a1 = _mm256_xor_si256( a1, a0 ); \ a0 = t; \ } #endif #define MIXWORD( a, b ) \ b = _mm256_xor_si256( a, b ); \ a = _mm256_xor_si256( b, mm256_rol_32( a, 2 ) ); \ b = _mm256_xor_si256( a, mm256_rol_32( b, 14 ) ); \ a = _mm256_xor_si256( b, mm256_rol_32( a, 10 ) ); \ b = mm256_rol_32( b, 1 ); #define STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, c0, c1 ) \ SUBCRUMB( x0, x1, x2, x3 ); \ SUBCRUMB( x5, x6, x7, x4 ); \ MIXWORD( x0, x4 ); \ MIXWORD( x1, x5 ); \ MIXWORD( x2, x6 ); \ MIXWORD( x3, x7 ); \ ADD_CONSTANT( x0, x4, c0, c1 ); #define STEP_PART2( a0, a1, t0, t1, c0, c1 ) \ t0 = _mm256_shuffle_epi32( a1, 147 ); \ a1 = _mm256_unpacklo_epi32( t0, a0 ); \ t0 = _mm256_unpackhi_epi32( t0, a0 ); \ t1 = _mm256_shuffle_epi32( t0, 78 ); \ a0 = _mm256_shuffle_epi32( a1, 78 ); \ SUBCRUMB( t1, t0, a0, a1 ); \ t0 = _mm256_unpacklo_epi32( t0, t1 ); \ a1 = _mm256_unpacklo_epi32( a1, a0 ); \ a0 = _mm256_unpackhi_epi64( a1, t0 ); \ a1 = _mm256_unpacklo_epi64( a1, t0 ); \ a1 = _mm256_shuffle_epi32( a1, 57 ); \ MIXWORD( a0, a1 ); \ ADD_CONSTANT( a0, a1, c0, c1 ); #define NMLTOM1024(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3)\ s1 = _mm256_unpackhi_epi32( r3, r2 ); \ q1 = _mm256_unpackhi_epi32( p3, p2 ); \ s3 = _mm256_unpacklo_epi32( r3, r2 ); \ q3 = _mm256_unpacklo_epi32( p3, p2 ); \ r3 = _mm256_unpackhi_epi32( r1, r0 ); \ r1 = _mm256_unpacklo_epi32( r1, r0 ); \ p3 = _mm256_unpackhi_epi32( p1, p0 ); \ p1 = _mm256_unpacklo_epi32( p1, p0 ); \ s0 = _mm256_unpackhi_epi64( s1, r3 ); \ q0 = _mm256_unpackhi_epi64( q1 ,p3 ); \ s1 = _mm256_unpacklo_epi64( s1, r3 ); \ q1 = _mm256_unpacklo_epi64( q1, p3 ); \ s2 = _mm256_unpackhi_epi64( s3, r1 ); \ q2 = _mm256_unpackhi_epi64( q3, p1 ); \ s3 = _mm256_unpacklo_epi64( s3, r1 ); \ q3 = _mm256_unpacklo_epi64( q3, p1 ); #define MIXTON1024(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3)\ NMLTOM1024(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3); /***************************************************/ /* Round function */ /* state: hash context */ void rnd512_2way( luffa_2way_context *state, const __m256i *msg ) { __m256i t0, t1; __m256i *chainv = state->chainv; __m256i x0, x1, x2, x3, x4, x5, x6, x7; t0 = mm256_xor3( chainv[0], chainv[2], chainv[4] ); t1 = mm256_xor3( chainv[1], chainv[3], chainv[5] ); t0 = mm256_xor3( t0, chainv[6], chainv[8] ); t1 = mm256_xor3( t1, chainv[7], chainv[9] ); MULT2( t0, t1 ); chainv[0] = _mm256_xor_si256( chainv[0], t0 ); chainv[1] = _mm256_xor_si256( chainv[1], t1 ); chainv[2] = _mm256_xor_si256( chainv[2], t0 ); chainv[3] = _mm256_xor_si256( chainv[3], t1 ); chainv[4] = _mm256_xor_si256( chainv[4], t0 ); chainv[5] = _mm256_xor_si256( chainv[5], t1 ); chainv[6] = _mm256_xor_si256( chainv[6], t0 ); chainv[7] = _mm256_xor_si256( chainv[7], t1 ); chainv[8] = _mm256_xor_si256( chainv[8], t0 ); chainv[9] = _mm256_xor_si256( chainv[9], t1 ); t0 = chainv[0]; t1 = chainv[1]; MULT2( chainv[0], chainv[1] ); chainv[0] = _mm256_xor_si256( chainv[0], chainv[2] ); chainv[1] = _mm256_xor_si256( chainv[1], chainv[3] ); MULT2( chainv[2], chainv[3] ); chainv[2] = _mm256_xor_si256(chainv[2], chainv[4]); chainv[3] = _mm256_xor_si256(chainv[3], chainv[5]); MULT2( chainv[4], chainv[5] ); chainv[4] = _mm256_xor_si256(chainv[4], chainv[6]); chainv[5] = _mm256_xor_si256(chainv[5], chainv[7]); MULT2( chainv[6], chainv[7] ); chainv[6] = _mm256_xor_si256(chainv[6], chainv[8]); chainv[7] = _mm256_xor_si256(chainv[7], chainv[9]); MULT2( chainv[8], chainv[9] ); t0 = chainv[8] = _mm256_xor_si256( chainv[8], t0 ); t1 = chainv[9] = _mm256_xor_si256( chainv[9], t1 ); MULT2( chainv[8], chainv[9] ); chainv[8] = _mm256_xor_si256( chainv[8], chainv[6] ); chainv[9] = _mm256_xor_si256( chainv[9], chainv[7] ); MULT2( chainv[6], chainv[7] ); chainv[6] = _mm256_xor_si256( chainv[6], chainv[4] ); chainv[7] = _mm256_xor_si256( chainv[7], chainv[5] ); MULT2( chainv[4], chainv[5] ); chainv[4] = _mm256_xor_si256( chainv[4], chainv[2] ); chainv[5] = _mm256_xor_si256( chainv[5], chainv[3] ); MULT2( chainv[2], chainv[3] ); chainv[2] = _mm256_xor_si256( chainv[2], chainv[0] ); chainv[3] = _mm256_xor_si256( chainv[3], chainv[1] ); MULT2( chainv[0], chainv[1] ); chainv[0] = _mm256_xor_si256( chainv[0], t0 ); chainv[1] = _mm256_xor_si256( chainv[1], t1 ); if ( msg ) { __m256i msg0, msg1; msg0 = _mm256_shuffle_epi32( msg[0], 27 ); msg1 = _mm256_shuffle_epi32( msg[1], 27 ); chainv[0] = _mm256_xor_si256( chainv[0], msg0 ); chainv[1] = _mm256_xor_si256( chainv[1], msg1 ); MULT2( msg0, msg1 ); chainv[2] = _mm256_xor_si256( chainv[2], msg0 ); chainv[3] = _mm256_xor_si256( chainv[3], msg1 ); MULT2( msg0, msg1 ); chainv[4] = _mm256_xor_si256( chainv[4], msg0 ); chainv[5] = _mm256_xor_si256( chainv[5], msg1 ); MULT2( msg0, msg1 ); chainv[6] = _mm256_xor_si256( chainv[6], msg0 ); chainv[7] = _mm256_xor_si256( chainv[7], msg1 ); MULT2( msg0, msg1 ); chainv[8] = _mm256_xor_si256( chainv[8], msg0 ); chainv[9] = _mm256_xor_si256( chainv[9], msg1 ); } chainv[3] = mm256_rol_32( chainv[3], 1 ); chainv[5] = mm256_rol_32( chainv[5], 2 ); chainv[7] = mm256_rol_32( chainv[7], 3 ); chainv[9] = mm256_rol_32( chainv[9], 4 ); NMLTOM1024( chainv[0], chainv[2], chainv[4], chainv[6], x0, x1, x2, x3, chainv[1], chainv[3], chainv[5], chainv[7], x4, x5, x6, x7 ); STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 0), cns( 1) ); STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 2), cns( 3) ); STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 4), cns( 5) ); STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 6), cns( 7) ); STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 8), cns( 9) ); STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns(10), cns(11) ); STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns(12), cns(13) ); STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns(14), cns(15) ); MIXTON1024( x0, x1, x2, x3, chainv[0], chainv[2], chainv[4], chainv[6], x4, x5, x6, x7, chainv[1], chainv[3], chainv[5], chainv[7]); /* Process last 256-bit block */ STEP_PART2( chainv[8], chainv[9], t0, t1, cns(16), cns(17) ); STEP_PART2( chainv[8], chainv[9], t0, t1, cns(18), cns(19) ); STEP_PART2( chainv[8], chainv[9], t0, t1, cns(20), cns(21) ); STEP_PART2( chainv[8], chainv[9], t0, t1, cns(22), cns(23) ); STEP_PART2( chainv[8], chainv[9], t0, t1, cns(24), cns(25) ); STEP_PART2( chainv[8], chainv[9], t0, t1, cns(26), cns(27) ); STEP_PART2( chainv[8], chainv[9], t0, t1, cns(28), cns(29) ); STEP_PART2( chainv[8], chainv[9], t0, t1, cns(30), cns(31) ); } /***************************************************/ /* Finalization function */ /* state: hash context */ /* b[8]: hash values */ void finalization512_2way( luffa_2way_context *state, uint32 *b ) { uint32 hash[8*2] __attribute((aligned(64))); __m256i* chainv = state->chainv; __m256i t0, t1; const __m256i shuff_bswap32 = mm256_set2_64( 0x0c0d0e0f08090a0b, 0x0405060700010203 ); /*---- blank round with m=0 ----*/ rnd512_2way( state, NULL ); t0 = mm256_xor3( chainv[0], chainv[2], chainv[4] ); t1 = mm256_xor3( chainv[1], chainv[3], chainv[5] ); t0 = mm256_xor3( t0, chainv[6], chainv[8] ); t1 = mm256_xor3( t1, chainv[7], chainv[9] ); t0 = _mm256_shuffle_epi32( t0, 27 ); t1 = _mm256_shuffle_epi32( t1, 27 ); _mm256_store_si256( (__m256i*)&hash[0], t0 ); _mm256_store_si256( (__m256i*)&hash[8], t1 ); casti_m256i( b, 0 ) = _mm256_shuffle_epi8( casti_m256i( hash, 0 ), shuff_bswap32 ); casti_m256i( b, 1 ) = _mm256_shuffle_epi8( casti_m256i( hash, 1 ), shuff_bswap32 ); rnd512_2way( state, NULL ); t0 = mm256_xor3( chainv[0], chainv[2], chainv[4] ); t1 = mm256_xor3( chainv[1], chainv[3], chainv[5] ); t0 = mm256_xor3( t0, chainv[6], chainv[8] ); t1 = mm256_xor3( t1, chainv[7], chainv[9] ); t0 = _mm256_shuffle_epi32( t0, 27 ); t1 = _mm256_shuffle_epi32( t1, 27 ); _mm256_store_si256( (__m256i*)&hash[0], t0 ); _mm256_store_si256( (__m256i*)&hash[8], t1 ); casti_m256i( b, 2 ) = _mm256_shuffle_epi8( casti_m256i( hash, 0 ), shuff_bswap32 ); casti_m256i( b, 3 ) = _mm256_shuffle_epi8( casti_m256i( hash, 1 ), shuff_bswap32 ); } int luffa_2way_init( luffa_2way_context *state, int hashbitlen ) { state->hashbitlen = hashbitlen; __m128i *iv = (__m128i*)IV; state->chainv[0] = mm256_bcast_m128( iv[0] ); state->chainv[1] = mm256_bcast_m128( iv[1] ); state->chainv[2] = mm256_bcast_m128( iv[2] ); state->chainv[3] = mm256_bcast_m128( iv[3] ); state->chainv[4] = mm256_bcast_m128( iv[4] ); state->chainv[5] = mm256_bcast_m128( iv[5] ); state->chainv[6] = mm256_bcast_m128( iv[6] ); state->chainv[7] = mm256_bcast_m128( iv[7] ); state->chainv[8] = mm256_bcast_m128( iv[8] ); state->chainv[9] = mm256_bcast_m128( iv[9] ); ((__m256i*)state->buffer)[0] = m256_zero; ((__m256i*)state->buffer)[1] = m256_zero; return 0; } // Do not call luffa_update_close after having called luffa_update. // Once luffa_update has been called only call luffa_update or luffa_close. int luffa_2way_update( luffa_2way_context *state, const void *data, size_t len ) { __m256i *vdata = (__m256i*)data; __m256i *buffer = (__m256i*)state->buffer; __m256i msg[2]; int i; int blocks = (int)len >> 5; const __m256i shuff_bswap32 = mm256_set2_64( 0x0c0d0e0f08090a0b, 0x0405060700010203 ); state-> rembytes = (int)len & 0x1F; // full blocks for ( i = 0; i < blocks; i++, vdata+=2 ) { msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = _mm256_shuffle_epi8( vdata[ 1 ], shuff_bswap32 ); rnd512_2way( state, msg ); } // 16 byte partial block exists for 80 byte len // store in buffer for transform in final for midstate to work if ( state->rembytes ) { // remaining data bytes buffer[0] = _mm256_shuffle_epi8( vdata[0], shuff_bswap32 ); buffer[1] = mm256_bcast128lo_64( 0x0000000080000000 ); } return 0; } int luffa_2way_close( luffa_2way_context *state, void *hashval ) { __m256i *buffer = (__m256i*)state->buffer; __m256i msg[2]; // transform pad block if ( state->rembytes ) // not empty, data is in buffer rnd512_2way( state, buffer ); else { // empty pad block, constant data msg[0] = mm256_bcast128lo_64( 0x0000000080000000 ); msg[1] = m256_zero; rnd512_2way( state, msg ); } finalization512_2way( state, (uint32*)hashval ); if ( state->hashbitlen > 512 ) finalization512_2way( state, (uint32*)( hashval+32 ) ); return 0; } int luffa512_2way_full( luffa_2way_context *state, void *output, const void *data, size_t inlen ) { state->hashbitlen = 512; __m128i *iv = (__m128i*)IV; state->chainv[0] = mm256_bcast_m128( iv[0] ); state->chainv[1] = mm256_bcast_m128( iv[1] ); state->chainv[2] = mm256_bcast_m128( iv[2] ); state->chainv[3] = mm256_bcast_m128( iv[3] ); state->chainv[4] = mm256_bcast_m128( iv[4] ); state->chainv[5] = mm256_bcast_m128( iv[5] ); state->chainv[6] = mm256_bcast_m128( iv[6] ); state->chainv[7] = mm256_bcast_m128( iv[7] ); state->chainv[8] = mm256_bcast_m128( iv[8] ); state->chainv[9] = mm256_bcast_m128( iv[9] ); ((__m256i*)state->buffer)[0] = m256_zero; ((__m256i*)state->buffer)[1] = m256_zero; const __m256i *vdata = (__m256i*)data; __m256i msg[2]; int i; const int blocks = (int)( inlen >> 5 ); const __m256i shuff_bswap32 = mm256_set2_64( 0x0c0d0e0f08090a0b, 0x0405060700010203 ); state->rembytes = inlen & 0x1F; // full blocks for ( i = 0; i < blocks; i++, vdata+=2 ) { msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = _mm256_shuffle_epi8( vdata[ 1 ], shuff_bswap32 ); rnd512_2way( state, msg ); } // 16 byte partial block exists for 80 byte len if ( state->rembytes ) { // padding of partial block msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = mm256_bcast128lo_64( 0x0000000080000000 ); rnd512_2way( state, msg ); } else { // empty pad block msg[0] = mm256_bcast128lo_64( 0x0000000080000000 ); msg[1] = m256_zero; rnd512_2way( state, msg ); } finalization512_2way( state, (uint32*)output ); if ( state->hashbitlen > 512 ) finalization512_2way( state, (uint32*)( output+32 ) ); return 0; } int luffa_2way_update_close( luffa_2way_context *state, void *output, const void *data, size_t inlen ) { // Optimized for integrals of 16 bytes, good for 64 and 80 byte len const __m256i *vdata = (__m256i*)data; __m256i msg[2]; int i; const int blocks = (int)( inlen >> 5 ); const __m256i shuff_bswap32 = mm256_set2_64( 0x0c0d0e0f08090a0b, 0x0405060700010203 ); state->rembytes = inlen & 0x1F; // full blocks for ( i = 0; i < blocks; i++, vdata+=2 ) { msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = _mm256_shuffle_epi8( vdata[ 1 ], shuff_bswap32 ); rnd512_2way( state, msg ); } // 16 byte partial block exists for 80 byte len if ( state->rembytes ) { // padding of partial block msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 ); msg[1] = mm256_bcast128lo_64( 0x0000000080000000 ); rnd512_2way( state, msg ); } else { // empty pad block msg[0] = mm256_bcast128lo_64( 0x0000000080000000 ); msg[1] = m256_zero; rnd512_2way( state, msg ); } finalization512_2way( state, (uint32*)output ); if ( state->hashbitlen > 512 ) finalization512_2way( state, (uint32*)( output+32 ) ); return 0; } #endif