v3.8.4.1

v3.8.4
2025-09-17 23:44:27 +00:00 · 2018-03-22 14:28:03 -04:00 · 2018-03-18 12:51:03 -04:00
20 changed files with 1028 additions and 848 deletions
--- a/Makefile.am
+++ b/Makefile.am
@@ -68,6 +68,7 @@ cpuminer_SOURCES = \
  algo/cryptonight/cryptonight.c\
  algo/cubehash/sph_cubehash.c \
  algo/cubehash/sse2/cubehash_sse2.c\
+  algo/cubehash/cube-hash-2way.c \
  algo/echo/sph_echo.c \
  algo/echo/aes_ni/hash.c\
  algo/gost/sph_gost.c \
@@ -242,7 +243,7 @@ cpuminer_SOURCES = \
  algo/x17/hmq1725.c \
  algo/yescrypt/yescrypt.c \
  algo/yescrypt/sha256_Y.c \
-  algo/yescrypt/yescrypt-simd.c
+  algo/yescrypt/yescrypt-best.c

 disable_flags =

--- a/README.md
+++ b/README.md
@@ -28,11 +28,12 @@ performance.
 ARM CPUs are not supported.

 2. 64 bit Linux OS. Ubuntu and Fedora based distributions, including Mint and
-Centos are known to work and have all dependencies in their repositories.
-Others may work but may require more effort.
+Centos, are known to work and have all dependencies in their repositories.
+Others may work but may require more effort. Older versions such as Centos 6
+don't work due to missing features. 
 64 bit Windows OS is supported with mingw_w64 and msys or pre-built binaries.

-MacOS, OSx is not supported.
+MacOS, OSx and Android are not supported.

 3. Stratum pool. Some algos may work wallet mining using getwork or GBT. YMMV.

@@ -110,6 +111,7 @@ Supported Algorithms
                          yescrypt     Globalboost-Y (BSTY)
                          yescryptr8   BitZeny (ZNY)
                          yescryptr16  Yenten (YTN)
+                          yescryptr32  WAVI
                          zr5          Ziftr

 Errata
--- a/14
+++ b/14
@@ -1,4 +1,4 @@
-cpuminer-opt now supports HW SHA acceleration available on AMD Ryzen CPUs.
+puminer-opt now supports HW SHA acceleration available on AMD Ryzen CPUs.
 This feature requires recent SW including GCC version 5 or higher and
 openssl version 1.1 or higher. It may also require using "-march=znver1"
 compile flag.
@@ -160,6 +160,18 @@ Support for even older x86_64 without AES_NI or SSE2 is not availble.
 Change Log
 ----------

+v3.8.4.1
+
+Fixed sha256t low difficulty rejects.
+Fixed compile error on CPUs with AVX512.
+
+v3.8.4
+
+Added yescryptr32 algo for WAVI coin.
+Added URL to API data.
+Improved detection of __int128 support (linux only)
+Compile support for CPUs without SSSE3 (no binary support)
+
 v3.8.3.3

 Integrated getblocktemplate with algo_gate.
--- a/algo-gate-api.c
+++ b/algo-gate-api.c
@@ -227,6 +227,7 @@ bool register_algo_gate( int algo, algo_gate_t *gate )
     case ALGO_YESCRYPT:     register_yescrypt_algo    ( gate ); break;
     case ALGO_YESCRYPTR8:   register_yescryptr8_algo  ( gate ); break;
     case ALGO_YESCRYPTR16:  register_yescryptr16_algo ( gate ); break;
+     case ALGO_YESCRYPTR32:  register_yescryptr32_algo ( gate ); break;
     case ALGO_ZR5:          register_zr5_algo         ( gate ); break;
    default:
        applog(LOG_ERR,"FAIL: algo_gate registration failed, unknown algo %s.\n", algo_names[opt_algo] );
--- a/algo/cubehash/cube-hash-2way.c
+++ b/algo/cubehash/cube-hash-2way.c
@@ -0,0 +1,205 @@
+#if defined(__AVX2__)
+
+#include <stdbool.h>
+#include <unistd.h>
+#include <memory.h>
+#include "cube-hash-2way.h"
+
+// 2x128
+
+static void transform_2way( cube_2way_context *sp )
+{
+    int r;
+    const int rounds = sp->rounds;
+
+    __m256i x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3;
+
+    x0 = _mm256_load_si256( (__m256i*)sp->h     );
+    x1 = _mm256_load_si256( (__m256i*)sp->h + 1 );
+    x2 = _mm256_load_si256( (__m256i*)sp->h + 2 );
+    x3 = _mm256_load_si256( (__m256i*)sp->h + 3 );
+    x4 = _mm256_load_si256( (__m256i*)sp->h + 4 );
+    x5 = _mm256_load_si256( (__m256i*)sp->h + 5 );
+    x6 = _mm256_load_si256( (__m256i*)sp->h + 6 );
+    x7 = _mm256_load_si256( (__m256i*)sp->h + 7 );
+
+    for ( r = 0; r < rounds; ++r )
+    {
+        x4 = _mm256_add_epi32( x0, x4 );
+        x5 = _mm256_add_epi32( x1, x5 );
+        x6 = _mm256_add_epi32( x2, x6 );
+        x7 = _mm256_add_epi32( x3, x7 );
+        y0 = x2;
+        y1 = x3;
+        y2 = x0;
+        y3 = x1;
+        x0 = _mm256_xor_si256( _mm256_slli_epi32( y0,  7 ),
+                               _mm256_srli_epi32( y0, 25 ) );
+        x1 = _mm256_xor_si256( _mm256_slli_epi32( y1,  7 ),
+                               _mm256_srli_epi32( y1, 25 ) );
+        x2 = _mm256_xor_si256( _mm256_slli_epi32( y2,  7 ),
+                               _mm256_srli_epi32( y2, 25 ) );
+        x3 = _mm256_xor_si256( _mm256_slli_epi32( y3,  7 ),
+                               _mm256_srli_epi32( y3, 25 ) );
+        x0 = _mm256_xor_si256( x0, x4 );
+        x1 = _mm256_xor_si256( x1, x5 );
+        x2 = _mm256_xor_si256( x2, x6 );
+        x3 = _mm256_xor_si256( x3, x7 );
+        x4 = mm256_swap128_64( x4 );
+        x5 = mm256_swap128_64( x5 );
+        x6 = mm256_swap128_64( x6 );
+        x7 = mm256_swap128_64( x7 );
+        x4 = _mm256_add_epi32( x0, x4 );
+        x5 = _mm256_add_epi32( x1, x5 );
+        x6 = _mm256_add_epi32( x2, x6 );
+        x7 = _mm256_add_epi32( x3, x7 );
+        y0 = x1;
+        y1 = x0;
+        y2 = x3;
+        y3 = x2;
+        x0 = _mm256_xor_si256( _mm256_slli_epi32( y0, 11 ),
+                               _mm256_srli_epi32( y0, 21 ) );
+        x1 = _mm256_xor_si256( _mm256_slli_epi32( y1, 11 ),
+                               _mm256_srli_epi32( y1, 21 ) );
+        x2 = _mm256_xor_si256( _mm256_slli_epi32( y2, 11 ),
+                               _mm256_srli_epi32( y2, 21 ) );
+        x3 = _mm256_xor_si256( _mm256_slli_epi32( y3, 11 ),
+                               _mm256_srli_epi32( y3, 21 ) );
+        x0 = _mm256_xor_si256( x0, x4 );
+        x1 = _mm256_xor_si256( x1, x5 );
+        x2 = _mm256_xor_si256( x2, x6 );
+        x3 = _mm256_xor_si256( x3, x7 );
+        x4 = mm256_swap64_32( x4 );
+        x5 = mm256_swap64_32( x5 );
+        x6 = mm256_swap64_32( x6 );
+        x7 = mm256_swap64_32( x7 );
+    }
+
+    _mm256_store_si256( (__m256i*)sp->h,     x0 );
+    _mm256_store_si256( (__m256i*)sp->h + 1, x1 );
+    _mm256_store_si256( (__m256i*)sp->h + 2, x2 );
+    _mm256_store_si256( (__m256i*)sp->h + 3, x3 );
+    _mm256_store_si256( (__m256i*)sp->h + 4, x4 );
+    _mm256_store_si256( (__m256i*)sp->h + 5, x5 );
+    _mm256_store_si256( (__m256i*)sp->h + 6, x6 );
+    _mm256_store_si256( (__m256i*)sp->h + 7, x7 );
+
+}
+
+cube_2way_context cube_2way_ctx_cache __attribute__ ((aligned (64)));
+
+int cube_2way_reinit( cube_2way_context *sp )
+{
+   memcpy( sp, &cube_2way_ctx_cache, sizeof(cube_2way_context) );
+   return 0;
+
+}
+
+int cube_2way_init( cube_2way_context *sp, int hashbitlen, int rounds,
+                       int blockbytes )
+{
+    int i;
+
+    // all sizes of __m128i
+    cube_2way_ctx_cache.hashlen   = hashbitlen/128;
+    cube_2way_ctx_cache.blocksize = blockbytes/16;
+    cube_2way_ctx_cache.rounds    = rounds;
+    cube_2way_ctx_cache.pos       = 0;
+
+    for ( i = 0; i < 8; ++i )
+       cube_2way_ctx_cache.h[i] = m256_zero;
+
+    cube_2way_ctx_cache.h[0] = _mm256_set_epi32(
+                                   0, rounds, blockbytes, hashbitlen / 8,
+                                   0, rounds, blockbytes, hashbitlen / 8 );
+
+    for ( i = 0; i < 10; ++i )
+       transform_2way( &cube_2way_ctx_cache );
+
+    memcpy( sp, &cube_2way_ctx_cache, sizeof(cube_2way_context) );
+    return 0;
+}
+
+
+int cube_2way_update( cube_2way_context *sp, const void *data, size_t size )
+{
+    const int len = size / 16;
+    const __m256i *in = (__m256i*)data;
+    int i;
+
+    // It is assumed data is aligned to 256 bits and is a multiple of 128 bits.
+    // Current usage sata is either 64 or 80 bytes.
+
+    for ( i = 0; i < len; i++ )
+    {
+        sp->h[ sp->pos ] = _mm256_xor_si256( sp->h[ sp->pos ], in[i] );
+        sp->pos++;
+        if ( sp->pos == sp->blocksize )
+        {
+           transform_2way( sp );
+           sp->pos = 0;
+        }
+    }
+
+    return 0;
+}
+
+int cube_2way_close( cube_2way_context *sp, void *output )
+{
+    __m256i *hash = (__m256i*)output;
+    int i;
+
+    // pos is zero for 64 byte data, 1 for 80 byte data.
+    sp->h[ sp->pos ] = _mm256_xor_si256( sp->h[ sp->pos ],
+                    _mm256_set_epi8( 0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0x80,
+                                     0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0x80 ) );
+    transform_2way( sp );
+
+    sp->h[7] = _mm256_xor_si256( sp->h[7], _mm256_set_epi32( 1,0,0,0,
+                                                             1,0,0,0 ) );
+    for ( i = 0; i < 10; ++i )
+       transform_2way( &cube_2way_ctx_cache );
+
+    for ( i = 0; i < sp->hashlen; i++ )
+       hash[i] = sp->h[i];
+
+    return 0;
+}
+
+int cube_2way_update_close( cube_2way_context *sp, void *output,
+                               const void *data, size_t size )
+{
+    const int len = size / 16;
+    const __m256i *in = (__m256i*)data;
+    __m256i *hash = (__m256i*)output;
+    int i;
+
+    for ( i = 0; i < len; i++ )
+    {
+        sp->h[ sp->pos ] = _mm256_xor_si256( sp->h[ sp->pos ], in[i] );
+        sp->pos++;
+        if ( sp->pos == sp->blocksize )
+        {
+           transform_2way( sp );
+           sp->pos = 0;
+        }
+    }
+
+    // pos is zero for 64 byte data, 1 for 80 byte data.
+    sp->h[ sp->pos ] = _mm256_xor_si256( sp->h[ sp->pos ],
+                    _mm256_set_epi8( 0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0x80,
+                                     0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0x80 ) );
+    transform_2way( sp );
+
+    sp->h[7] = _mm256_xor_si256( sp->h[7], _mm256_set_epi32( 1,0,0,0,
+                                                             1,0,0,0 ) );
+    for ( i = 0; i < 10; ++i )
+       transform_2way( &cube_2way_ctx_cache );
+
+    for ( i = 0; i < sp->hashlen; i++ )
+       hash[i] = sp->h[i];
+
+    return 0;
+}
+
+#endif
--- a/algo/cubehash/cube-hash-2way.h
+++ b/algo/cubehash/cube-hash-2way.h
@@ -0,0 +1,36 @@
+#ifndef CUBE_HASH_2WAY_H__
+#define CUBE_HASH_2WAY_H__
+
+#if defined(__AVX2__)
+
+#include <stdint.h>
+#include "avxdefs.h"
+
+// 2x128, 2 way parallel SSE2
+
+struct _cube_2way_context
+{
+    int hashlen;           // __m128i
+    int rounds;
+    int blocksize;         // __m128i
+    int pos;               // number of __m128i read into x from current block
+    __m256i h[8] __attribute__ ((aligned (64)));
+};
+
+typedef struct _cube_2way_context cube_2way_context;
+
+int cube_2way_init( cube_2way_context* sp, int hashbitlen, int rounds,
+                       int blockbytes );
+// reinitialize context with same parameters, much faster.
+int cube_2way_reinit( cube_2way_context *sp );
+
+int cube_2way_update( cube_2way_context *sp, const void *data, size_t size );
+
+int cube_2way_close( cube_2way_context *sp, void *output );
+
+int cube_2way_update_close( cube_2way_context *sp, void *output,
+                            const void *data, size_t size );
+
+
+#endif
+#endif
--- a/algo/hodl/hodl-gate.c
+++ b/algo/hodl/hodl-gate.c
@@ -76,7 +76,6 @@ char* hodl_malloc_txs_request( struct work *work )
  return req;
 }

-
 void hodl_build_block_header( struct work* g_work, uint32_t version,
                              uint32_t *prevhash, uint32_t *merkle_tree,
                              uint32_t ntime, uint32_t nbits )
@@ -88,16 +87,16 @@ void hodl_build_block_header( struct work* g_work, uint32_t version,

   if ( have_stratum )
      for ( i = 0; i < 8; i++ )
-         g_work->data[1 + i] = le32dec( prevhash + i );
+         g_work->data[ 1+i ] = le32dec( prevhash + i );
   else
      for (i = 0; i < 8; i++)
         g_work->data[ 8-i ] = le32dec( prevhash + i );

   for ( i = 0; i < 8; i++ )
-      g_work->data[9 + i] = be32dec(  merkle_tree + i );
+      g_work->data[ 9+i ] = be32dec( merkle_tree + i );

-   g_work->data[ algo_gate.ntime_index ] =  ntime;
-   g_work->data[ algo_gate.nbits_index ] =  nbits;
+   g_work->data[ algo_gate.ntime_index ] = ntime;
+   g_work->data[ algo_gate.nbits_index ] = nbits;
   g_work->data[22] = 0x80000000;
   g_work->data[31] = 0x00000280;
 }
@@ -194,8 +193,13 @@ bool register_hodl_algo( algo_gate_t* gate )
  applog( LOG_ERR, "Only CPUs with AES are supported, use legacy version.");
  return false;
 #endif
+//  if ( TOTAL_CHUNKS % opt_n_threads )
+//  {
+//     applog(LOG_ERR,"Thread count must be power of 2.");
+//     return false;
+//  }
  pthread_barrier_init( &hodl_barrier, NULL, opt_n_threads );
-  gate->optimizations         = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT;
+  gate->optimizations         = AES_OPT | AVX_OPT | AVX2_OPT;
  gate->scanhash              = (void*)&hodl_scanhash;
  gate->get_new_work          = (void*)&hodl_get_new_work;
  gate->longpoll_rpc_call     = (void*)&hodl_longpoll_rpc_call;
--- a/algo/hodl/hodl-wolf.c
+++ b/algo/hodl/hodl-wolf.c
@@ -10,23 +10,26 @@

 #ifndef NO_AES_NI               

-void GenerateGarbageCore(CacheEntry *Garbage, int ThreadID, int ThreadCount, void *MidHash)
+void GenerateGarbageCore( CacheEntry *Garbage, int ThreadID, int ThreadCount,
+     void *MidHash )
 {
-#ifdef __AVX__
-    uint64_t* TempBufs[SHA512_PARALLEL_N] ;
-    uint64_t* desination[SHA512_PARALLEL_N];
+    const int Chunk = TOTAL_CHUNKS / ThreadCount;
+    const uint32_t StartChunk = ThreadID * Chunk;
+    const uint32_t EndChunk   = StartChunk + Chunk;

-    for ( int i=0; i<SHA512_PARALLEL_N; ++i )
+#ifdef __AVX__
+    uint64_t* TempBufs[ SHA512_PARALLEL_N ] ;
+    uint64_t* desination[ SHA512_PARALLEL_N ];
+
+    for ( int i=0; i < SHA512_PARALLEL_N; ++i )
    {
-        TempBufs[i] = (uint64_t*)malloc(32);
-        memcpy(TempBufs[i], MidHash, 32);
+        TempBufs[i] = (uint64_t*)malloc( 32 );
+        memcpy( TempBufs[i], MidHash, 32 );
    }

-    uint32_t StartChunk = ThreadID * (TOTAL_CHUNKS / ThreadCount);
-    for ( uint32_t i = StartChunk;
-          i < StartChunk + (TOTAL_CHUNKS / ThreadCount); i+= SHA512_PARALLEL_N )
+    for ( uint32_t i = StartChunk; i < EndChunk; i += SHA512_PARALLEL_N )
    {
-        for ( int j=0; j<SHA512_PARALLEL_N; ++j )
+        for ( int j = 0; j < SHA512_PARALLEL_N; ++j )
        {
            ( (uint32_t*)TempBufs[j] )[0] = i + j;
            desination[j] = (uint64_t*)( (uint8_t *)Garbage + ( (i+j)
@@ -35,15 +38,13 @@ void GenerateGarbageCore(CacheEntry *Garbage, int ThreadID, int ThreadCount, voi
        sha512Compute32b_parallel( TempBufs, desination );
    }

-    for ( int i=0; i<SHA512_PARALLEL_N; ++i )
+    for ( int i = 0; i < SHA512_PARALLEL_N; ++i )
        free( TempBufs[i] );
 #else
    uint32_t TempBuf[8];
    memcpy( TempBuf, MidHash, 32 );

-    uint32_t StartChunk = ThreadID * (TOTAL_CHUNKS / ThreadCount);
-    for ( uint32_t i = StartChunk;
-          i < StartChunk + (TOTAL_CHUNKS / ThreadCount); ++i )
+    for ( uint32_t i = StartChunk; i < EndChunk; ++i )
    {
        TempBuf[0] = i;
        SHA512( ( uint8_t *)TempBuf, 32,
--- a/algo/lyra2/sponge.h
+++ b/algo/lyra2/sponge.h
@@ -55,23 +55,23 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
 // returns void, updates all args
 #define G_4X64(a,b,c,d) \
   a = _mm256_add_epi64( a, b ); \
-   d = mm256_rotr_64( _mm256_xor_si256( d, a), 32 ); \
+   d = mm256_ror_64( _mm256_xor_si256( d, a), 32 ); \
   c = _mm256_add_epi64( c, d ); \
-   b = mm256_rotr_64( _mm256_xor_si256( b, c ), 24 ); \
+   b = mm256_ror_64( _mm256_xor_si256( b, c ), 24 ); \
   a = _mm256_add_epi64( a, b ); \
-   d = mm256_rotr_64( _mm256_xor_si256( d, a ), 16 ); \
+   d = mm256_ror_64( _mm256_xor_si256( d, a ), 16 ); \
   c = _mm256_add_epi64( c, d ); \
-   b = mm256_rotr_64( _mm256_xor_si256( b, c ), 63 );
+   b = mm256_ror_64( _mm256_xor_si256( b, c ), 63 );

 #define LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
   G_4X64( s0, s1, s2, s3 ); \
-   s1 = mm256_rotr256_1x64( s1); \
+   s1 = mm256_ror256_1x64( s1); \
   s2 = mm256_swap_128( s2 ); \
-   s3 = mm256_rotl256_1x64( s3 ); \
+   s3 = mm256_rol256_1x64( s3 ); \
   G_4X64( s0, s1, s2, s3 ); \
-   s1 = mm256_rotl256_1x64( s1 ); \
+   s1 = mm256_rol256_1x64( s1 ); \
   s2 = mm256_swap_128( s2 ); \
-   s3 = mm256_rotr256_1x64( s3 );
+   s3 = mm256_ror256_1x64( s3 );

 #define LYRA_12_ROUNDS_AVX2( s0, s1, s2, s3 ) \
   LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
@@ -94,25 +94,25 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
 // returns void, all args updated
 #define G_2X64(a,b,c,d) \
   a = _mm_add_epi64( a, b ); \
-   d = mm_rotr_64( _mm_xor_si128( d, a), 32 ); \
+   d = mm_ror_64( _mm_xor_si128( d, a), 32 ); \
   c = _mm_add_epi64( c, d ); \
-   b = mm_rotr_64( _mm_xor_si128( b, c ), 24 ); \
+   b = mm_ror_64( _mm_xor_si128( b, c ), 24 ); \
   a = _mm_add_epi64( a, b ); \
-   d = mm_rotr_64( _mm_xor_si128( d, a ), 16 ); \
+   d = mm_ror_64( _mm_xor_si128( d, a ), 16 ); \
   c = _mm_add_epi64( c, d ); \
-   b = mm_rotr_64( _mm_xor_si128( b, c ), 63 );
+   b = mm_ror_64( _mm_xor_si128( b, c ), 63 );

 #define LYRA_ROUND_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
   G_2X64( s0, s2, s4, s6 ); \
   G_2X64( s1, s3, s5, s7 ); \
-   mm_rotl256_1x64( s2, s3 ); \
+   mm_ror256_1x64( s2, s3 ); \
   mm_swap_128( s4, s5 ); \
-   mm_rotr256_1x64( s6, s7 ); \
+   mm_rol256_1x64( s6, s7 ); \
   G_2X64( s0, s2, s4, s6 ); \
   G_2X64( s1, s3, s5, s7 ); \
-   mm_rotr256_1x64( s2, s3 ); \
+   mm_rol256_1x64( s2, s3 ); \
   mm_swap_128( s4, s5 ); \
-   mm_rotl256_1x64( s6, s7 );
+   mm_ror256_1x64( s6, s7 );

 #define LYRA_12_ROUNDS_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
   LYRA_ROUND_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
--- a/algo/sha/sha256t.c
+++ b/algo/sha/sha256t.c
@@ -155,7 +155,7 @@ bool register_sha256t_algo( algo_gate_t* gate )
    gate->optimizations = SSE2_OPT | AVX_OPT | AVX2_OPT | SHA_OPT;
    gate->scanhash   = (void*)&scanhash_sha256t;
    gate->hash       = (void*)&sha256t_hash;
-    gate->set_target = (void*)&sha256t_set_target;
+//    gate->set_target = (void*)&sha256t_set_target;
    gate->get_max64  = (void*)&get_max64_0x3ffff;
    return true;
 }
--- a/algo/shavite/sph-shavite-aesni.c
+++ b/algo/shavite/sph-shavite-aesni.c
@@ -52,21 +52,6 @@ extern "C"{

 #define C32   SPH_C32

-/*
- * As of round 2 of the SHA-3 competition, the published reference
- * implementation and test vectors are wrong, because they use
- * big-endian AES tables while the internal decoding uses little-endian.
- * The code below follows the specification. To turn it into a code
- * which follows the reference implementation (the one called "BugFix"
- * on the SHAvite-3 web site, published on Nov 23rd, 2009), comment out
- * the code below (from the '#define AES_BIG_ENDIAN...' to the definition
- * of the AES_ROUND_NOKEY macro) and replace it with the version which
- * is commented out afterwards.
- */
-
-#define AES_BIG_ENDIAN   0
-#include "algo/sha/aes_helper.c"
-
 static const sph_u32 IV512[] = {
 	C32(0x72FCCDD8), C32(0x79CA4727), C32(0x128A077B), C32(0x40D55AEC),
 	C32(0xD1901A06), C32(0x430AE307), C32(0xB29F5CD1), C32(0xDF07FBFC),
@@ -74,210 +59,19 @@ static const sph_u32 IV512[] = {
 	C32(0xE275EADE), C32(0x502D9FCD), C32(0xB9357178), C32(0x022A4B9A)
 };

-// Return hi 128 bits with elements shifted one lane with vacated lane filled
-// with data rotated from lo.
 // Partially rotate elements in two 128 bit vectors as one 256 bit vector
 // and return the rotated high 128 bits.
-// Similar to mm_rotr256_1x32 but only a partial rotation as lo is not
-// completed. It's faster than a full rotation.
+#if defined(__SSSE3__)

-static inline __m128i mm_rotr256hi_1x32( __m128i hi, __m128i lo, int n )
-{   return _mm_or_si128( _mm_srli_si128( hi, n<<2 ),
-                        _mm_slli_si128( lo, 16 - (n<<2) ) );
-}
+#define mm_rotr256hi_1x32( hi, lo )  _mm_alignr_epi8( lo, hi, 4 )

-#define AES_ROUND_NOKEY(x0, x1, x2, x3)   do { \
-		sph_u32 t0 = (x0); \
-		sph_u32 t1 = (x1); \
-		sph_u32 t2 = (x2); \
-		sph_u32 t3 = (x3); \
-		AES_ROUND_NOKEY_LE(t0, t1, t2, t3, x0, x1, x2, x3); \
-	} while (0)
+#else  // SSE2

-  
-#define KEY_EXPAND_ELT(k0, k1, k2, k3)   do { \
-		sph_u32 kt; \
-		AES_ROUND_NOKEY(k1, k2, k3, k0); \
-		kt = (k0); \
-		(k0) = (k1); \
-		(k1) = (k2); \
-		(k2) = (k3); \
-		(k3) = kt; \
-	} while (0)
+#define mm_rotr256hi_1x32( hi, lo ) \
+   _mm_or_si128( _mm_srli_si128( hi,  4 ), \
+                 _mm_slli_si128( lo, 12 ) )

-
-#if SPH_SMALL_FOOTPRINT_SHAVITE
-
-/*
- * This function assumes that "msg" is aligned for 32-bit access.
- */
-static void
-c512(sph_shavite_big_context *sc, const void *msg)
-{
-	sph_u32 p0, p1, p2, p3, p4, p5, p6, p7;
-	sph_u32 p8, p9, pA, pB, pC, pD, pE, pF;
-	sph_u32 rk[448];
-	size_t u;
-	int r, s;
-
-#if SPH_LITTLE_ENDIAN
-	memcpy(rk, msg, 128);
-#else
-	for (u = 0; u < 32; u += 4) {
-		rk[u + 0] = sph_dec32le_aligned(
-			(const unsigned char *)msg + (u << 2) +  0);
-		rk[u + 1] = sph_dec32le_aligned(
-			(const unsigned char *)msg + (u << 2) +  4);
-		rk[u + 2] = sph_dec32le_aligned(
-			(const unsigned char *)msg + (u << 2) +  8);
-		rk[u + 3] = sph_dec32le_aligned(
-			(const unsigned char *)msg + (u << 2) + 12);
-	}
 #endif
-	u = 32;
-	for (;;) {
-		for (s = 0; s < 4; s ++) {
-			sph_u32 x0, x1, x2, x3;
-
-			x0 = rk[u - 31];
-			x1 = rk[u - 30];
-			x2 = rk[u - 29];
-			x3 = rk[u - 32];
-			AES_ROUND_NOKEY(x0, x1, x2, x3);
-			rk[u + 0] = x0 ^ rk[u - 4];
-			rk[u + 1] = x1 ^ rk[u - 3];
-			rk[u + 2] = x2 ^ rk[u - 2];
-			rk[u + 3] = x3 ^ rk[u - 1];
-			if (u == 32) {
-				rk[ 32] ^= sc->count0;
-				rk[ 33] ^= sc->count1;
-				rk[ 34] ^= sc->count2;
-				rk[ 35] ^= SPH_T32(~sc->count3);
-			} else if (u == 440) {
-				rk[440] ^= sc->count1;
-				rk[441] ^= sc->count0;
-				rk[442] ^= sc->count3;
-				rk[443] ^= SPH_T32(~sc->count2);
-			}
-			u += 4;
-
-			x0 = rk[u - 31];
-			x1 = rk[u - 30];
-			x2 = rk[u - 29];
-			x3 = rk[u - 32];
-			AES_ROUND_NOKEY(x0, x1, x2, x3);
-			rk[u + 0] = x0 ^ rk[u - 4];
-			rk[u + 1] = x1 ^ rk[u - 3];
-			rk[u + 2] = x2 ^ rk[u - 2];
-			rk[u + 3] = x3 ^ rk[u - 1];
-			if (u == 164) {
-				rk[164] ^= sc->count3;
-				rk[165] ^= sc->count2;
-				rk[166] ^= sc->count1;
-				rk[167] ^= SPH_T32(~sc->count0);
-			} else if (u == 316) {
-				rk[316] ^= sc->count2;
-				rk[317] ^= sc->count3;
-				rk[318] ^= sc->count0;
-				rk[319] ^= SPH_T32(~sc->count1);
-			}
-			u += 4;
-		}
-		if (u == 448)
-			break;
-		for (s = 0; s < 8; s ++) {
-			rk[u + 0] = rk[u - 32] ^ rk[u - 7];
-			rk[u + 1] = rk[u - 31] ^ rk[u - 6];
-			rk[u + 2] = rk[u - 30] ^ rk[u - 5];
-			rk[u + 3] = rk[u - 29] ^ rk[u - 4];
-			u += 4;
-		}
-	}
-
-	p0 = sc->h[0x0];
-	p1 = sc->h[0x1];
-	p2 = sc->h[0x2];
-	p3 = sc->h[0x3];
-	p4 = sc->h[0x4];
-	p5 = sc->h[0x5];
-	p6 = sc->h[0x6];
-	p7 = sc->h[0x7];
-	p8 = sc->h[0x8];
-	p9 = sc->h[0x9];
-	pA = sc->h[0xA];
-	pB = sc->h[0xB];
-	pC = sc->h[0xC];
-	pD = sc->h[0xD];
-	pE = sc->h[0xE];
-	pF = sc->h[0xF];
-	u = 0;
-	for (r = 0; r < 14; r ++) {
-#define C512_ELT(l0, l1, l2, l3, r0, r1, r2, r3)   do { \
-		sph_u32 x0, x1, x2, x3; \
-		x0 = r0 ^ rk[u ++]; \
-		x1 = r1 ^ rk[u ++]; \
-		x2 = r2 ^ rk[u ++]; \
-		x3 = r3 ^ rk[u ++]; \
-		AES_ROUND_NOKEY(x0, x1, x2, x3); \
-		x0 ^= rk[u ++]; \
-		x1 ^= rk[u ++]; \
-		x2 ^= rk[u ++]; \
-		x3 ^= rk[u ++]; \
-		AES_ROUND_NOKEY(x0, x1, x2, x3); \
-		x0 ^= rk[u ++]; \
-		x1 ^= rk[u ++]; \
-		x2 ^= rk[u ++]; \
-		x3 ^= rk[u ++]; \
-		AES_ROUND_NOKEY(x0, x1, x2, x3); \
-		x0 ^= rk[u ++]; \
-		x1 ^= rk[u ++]; \
-		x2 ^= rk[u ++]; \
-		x3 ^= rk[u ++]; \
-		AES_ROUND_NOKEY(x0, x1, x2, x3); \
-		l0 ^= x0; \
-		l1 ^= x1; \
-		l2 ^= x2; \
-		l3 ^= x3; \
-	} while (0)
-
-#define WROT(a, b, c, d)   do { \
-		sph_u32 t = d; \
-		d = c; \
-		c = b; \
-		b = a; \
-		a = t; \
-	} while (0)
-
-		C512_ELT(p0, p1, p2, p3, p4, p5, p6, p7);
-		C512_ELT(p8, p9, pA, pB, pC, pD, pE, pF);
-
-		WROT(p0, p4, p8, pC);
-		WROT(p1, p5, p9, pD);
-		WROT(p2, p6, pA, pE);
-		WROT(p3, p7, pB, pF);
-
-#undef C512_ELT
-#undef WROT
-	}
-	sc->h[0x0] ^= p0;
-	sc->h[0x1] ^= p1;
-	sc->h[0x2] ^= p2;
-	sc->h[0x3] ^= p3;
-	sc->h[0x4] ^= p4;
-	sc->h[0x5] ^= p5;
-	sc->h[0x6] ^= p6;
-	sc->h[0x7] ^= p7;
-	sc->h[0x8] ^= p8;
-	sc->h[0x9] ^= p9;
-	sc->h[0xA] ^= pA;
-	sc->h[0xB] ^= pB;
-	sc->h[0xC] ^= pC;
-	sc->h[0xD] ^= pD;
-	sc->h[0xE] ^= pE;
-	sc->h[0xF] ^= pF;
-}
-
-#else

 static void
 c512( sph_shavite_big_context *sc, const void *msg )
@@ -331,7 +125,7 @@ c512( sph_shavite_big_context *sc, const void *msg )
   for ( r = 0; r < 3; r ++ )
   {
      // round 1, 5, 9
-      k00 = mm_rotr_1x32( _mm_aesenc_si128( k00, m128_zero ) );
+      k00 = mm_ror_1x32( _mm_aesenc_si128( k00, m128_zero ) );
      k00 = _mm_xor_si128( k00, k13 ); 

      if ( r == 0 )
@@ -340,7 +134,7 @@ c512( sph_shavite_big_context *sc, const void *msg )

      x = _mm_xor_si128( p0, k00 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k01 = mm_rotr_1x32( _mm_aesenc_si128( k01, m128_zero ) );
+      k01 = mm_ror_1x32( _mm_aesenc_si128( k01, m128_zero ) );
      k01 = _mm_xor_si128( k01, k00 );

      if ( r == 1 )
@@ -349,33 +143,33 @@ c512( sph_shavite_big_context *sc, const void *msg )

      x = _mm_xor_si128( x, k01 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k02 = mm_rotr_1x32( _mm_aesenc_si128( k02, m128_zero ) );
+      k02 = mm_ror_1x32( _mm_aesenc_si128( k02, m128_zero ) );
      k02 = _mm_xor_si128( k02, k01 );

      x = _mm_xor_si128( x, k02 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k03 = mm_rotr_1x32( _mm_aesenc_si128( k03, m128_zero ) );
+      k03 = mm_ror_1x32( _mm_aesenc_si128( k03, m128_zero ) );
      k03 = _mm_xor_si128( k03, k02 );

      x = _mm_xor_si128( x, k03 );
      x = _mm_aesenc_si128( x, m128_zero );
      p3 = _mm_xor_si128( p3, x );
-      k10 = mm_rotr_1x32( _mm_aesenc_si128( k10, m128_zero ) );
+      k10 = mm_ror_1x32( _mm_aesenc_si128( k10, m128_zero ) );
      k10 = _mm_xor_si128( k10, k03 );

      x = _mm_xor_si128( p2, k10 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k11 = mm_rotr_1x32( _mm_aesenc_si128( k11, m128_zero ) );
+      k11 = mm_ror_1x32( _mm_aesenc_si128( k11, m128_zero ) );
      k11 = _mm_xor_si128( k11, k10 );

      x = _mm_xor_si128( x, k11 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k12 = mm_rotr_1x32( _mm_aesenc_si128( k12, m128_zero ) );
+      k12 = mm_ror_1x32( _mm_aesenc_si128( k12, m128_zero ) );
      k12 = _mm_xor_si128( k12, k11 );

      x = _mm_xor_si128( x, k12 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k13 = mm_rotr_1x32( _mm_aesenc_si128( k13, m128_zero ) );
+      k13 = mm_ror_1x32( _mm_aesenc_si128( k13, m128_zero ) );
      k13 = _mm_xor_si128( k13, k12 );

      if ( r == 2 )
@@ -388,80 +182,80 @@ c512( sph_shavite_big_context *sc, const void *msg )

      // round 2, 6, 10

-      k00 = _mm_xor_si128( k00, mm_rotr256hi_1x32( k12, k13, 1 ) );
+      k00 = _mm_xor_si128( k00, mm_rotr256hi_1x32( k12, k13 ) );
      x = _mm_xor_si128( p3, k00 );
      x = _mm_aesenc_si128( x, m128_zero );

-      k01 = _mm_xor_si128( k01, mm_rotr256hi_1x32( k13, k00, 1 ) );
+      k01 = _mm_xor_si128( k01, mm_rotr256hi_1x32( k13, k00 ) );
      x = _mm_xor_si128( x, k01 );
      x = _mm_aesenc_si128( x, m128_zero );

-      k02 = _mm_xor_si128( k02, mm_rotr256hi_1x32( k00, k01, 1 ) );
+      k02 = _mm_xor_si128( k02, mm_rotr256hi_1x32( k00, k01 ) );
      x = _mm_xor_si128( x, k02 );
      x = _mm_aesenc_si128( x, m128_zero );

-      k03 = _mm_xor_si128( k03, mm_rotr256hi_1x32( k01, k02, 1 ) );
+      k03 = _mm_xor_si128( k03, mm_rotr256hi_1x32( k01, k02 ) );
      x = _mm_xor_si128( x, k03 );
      x = _mm_aesenc_si128( x, m128_zero );

      p2 = _mm_xor_si128( p2, x );
-      k10 = _mm_xor_si128( k10, mm_rotr256hi_1x32( k02, k03, 1 ) );
+      k10 = _mm_xor_si128( k10, mm_rotr256hi_1x32( k02, k03 ) );
      x = _mm_xor_si128( p1, k10 );
      x = _mm_aesenc_si128( x, m128_zero );

-      k11 = _mm_xor_si128( k11, mm_rotr256hi_1x32( k03, k10, 1 ) );
+      k11 = _mm_xor_si128( k11, mm_rotr256hi_1x32( k03, k10 ) );
      x = _mm_xor_si128( x, k11 );
      x = _mm_aesenc_si128( x, m128_zero );

-      k12 = _mm_xor_si128( k12, mm_rotr256hi_1x32( k10, k11, 1 ) );
+      k12 = _mm_xor_si128( k12, mm_rotr256hi_1x32( k10, k11 ) );
      x = _mm_xor_si128( x, k12 );
      x = _mm_aesenc_si128( x, m128_zero );

-      k13 = _mm_xor_si128( k13, mm_rotr256hi_1x32( k11, k12, 1 ) );
+      k13 = _mm_xor_si128( k13, mm_rotr256hi_1x32( k11, k12 ) );
      x = _mm_xor_si128( x, k13 );
      x = _mm_aesenc_si128( x, m128_zero );
      p0 = _mm_xor_si128( p0, x );

      // round 3, 7, 11

-      k00 = mm_rotr_1x32( _mm_aesenc_si128( k00, m128_zero ) );
+      k00 = mm_ror_1x32( _mm_aesenc_si128( k00, m128_zero ) );
      k00 = _mm_xor_si128( k00, k13 );

      x = _mm_xor_si128( p2, k00 );
      x = _mm_aesenc_si128( x, m128_zero );

-      k01 = mm_rotr_1x32( _mm_aesenc_si128( k01, m128_zero ) );
+      k01 = mm_ror_1x32( _mm_aesenc_si128( k01, m128_zero ) );
      k01 = _mm_xor_si128( k01, k00 );

      x = _mm_xor_si128( x, k01 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k02 = mm_rotr_1x32( _mm_aesenc_si128( k02, m128_zero ) );
+      k02 = mm_ror_1x32( _mm_aesenc_si128( k02, m128_zero ) );
      k02 = _mm_xor_si128( k02, k01 );

      x = _mm_xor_si128( x, k02 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k03 = mm_rotr_1x32( _mm_aesenc_si128( k03, m128_zero ) );
+      k03 = mm_ror_1x32( _mm_aesenc_si128( k03, m128_zero ) );
      k03 = _mm_xor_si128( k03, k02 );

      x = _mm_xor_si128( x, k03 );
      x = _mm_aesenc_si128( x, m128_zero );
      p1 = _mm_xor_si128( p1, x );
-      k10 = mm_rotr_1x32( _mm_aesenc_si128( k10, m128_zero ) );
+      k10 = mm_ror_1x32( _mm_aesenc_si128( k10, m128_zero ) );
      k10 = _mm_xor_si128( k10, k03 );

      x = _mm_xor_si128( p0, k10 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k11 = mm_rotr_1x32( _mm_aesenc_si128( k11, m128_zero ) );
+      k11 = mm_ror_1x32( _mm_aesenc_si128( k11, m128_zero ) );
      k11 = _mm_xor_si128( k11, k10 );

      x = _mm_xor_si128( x, k11 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k12 = mm_rotr_1x32( _mm_aesenc_si128( k12, m128_zero ) );
+      k12 = mm_ror_1x32( _mm_aesenc_si128( k12, m128_zero ) );
      k12 = _mm_xor_si128( k12, k11 );

      x = _mm_xor_si128( x, k12 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k13 = mm_rotr_1x32( _mm_aesenc_si128( k13, m128_zero ) );
+      k13 = mm_ror_1x32( _mm_aesenc_si128( k13, m128_zero ) );
      k13 = _mm_xor_si128( k13, k12 );

      x = _mm_xor_si128( x, k13 );
@@ -470,36 +264,36 @@ c512( sph_shavite_big_context *sc, const void *msg )

      // round 4, 8, 12

-      k00 = _mm_xor_si128( k00, mm_rotr256hi_1x32( k12, k13, 1 ) );
+      k00 = _mm_xor_si128( k00, mm_rotr256hi_1x32( k12, k13 ) );

      x = _mm_xor_si128( p1, k00 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k01 = _mm_xor_si128( k01, mm_rotr256hi_1x32( k13, k00, 1 ) );
+      k01 = _mm_xor_si128( k01, mm_rotr256hi_1x32( k13, k00 ) );

      x = _mm_xor_si128( x, k01 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k02 = _mm_xor_si128( k02, mm_rotr256hi_1x32( k00, k01, 1 ) );
+      k02 = _mm_xor_si128( k02, mm_rotr256hi_1x32( k00, k01 ) );

      x = _mm_xor_si128( x, k02 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k03 = _mm_xor_si128( k03, mm_rotr256hi_1x32( k01, k02, 1 ) );
+      k03 = _mm_xor_si128( k03, mm_rotr256hi_1x32( k01, k02 ) );

      x = _mm_xor_si128( x, k03 );
      x = _mm_aesenc_si128( x, m128_zero );
      p0 = _mm_xor_si128( p0, x );
-      k10 = _mm_xor_si128( k10, mm_rotr256hi_1x32( k02, k03, 1 ) );
+      k10 = _mm_xor_si128( k10, mm_rotr256hi_1x32( k02, k03 ) );

      x = _mm_xor_si128( p3, k10 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k11 = _mm_xor_si128( k11, mm_rotr256hi_1x32( k03, k10, 1 ) );
+      k11 = _mm_xor_si128( k11, mm_rotr256hi_1x32( k03, k10 ) );

      x = _mm_xor_si128( x, k11 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k12 = _mm_xor_si128( k12, mm_rotr256hi_1x32( k10, k11, 1 ) );
+      k12 = _mm_xor_si128( k12, mm_rotr256hi_1x32( k10, k11 ) );

      x = _mm_xor_si128( x, k12 );
      x = _mm_aesenc_si128( x, m128_zero );
-      k13 = _mm_xor_si128( k13, mm_rotr256hi_1x32( k11, k12, 1 ) );
+      k13 = _mm_xor_si128( k13, mm_rotr256hi_1x32( k11, k12 ) );

      x = _mm_xor_si128( x, k13 );
      x = _mm_aesenc_si128( x, m128_zero );
@@ -508,44 +302,44 @@ c512( sph_shavite_big_context *sc, const void *msg )

   // round 13

-   k00 = mm_rotr_1x32( _mm_aesenc_si128( k00, m128_zero ) );
+   k00 = mm_ror_1x32( _mm_aesenc_si128( k00, m128_zero ) );
   k00 = _mm_xor_si128( k00, k13 );

   x = _mm_xor_si128( p0, k00 );
   x = _mm_aesenc_si128( x, m128_zero );
-   k01 = mm_rotr_1x32( _mm_aesenc_si128( k01, m128_zero ) ); 
+   k01 = mm_ror_1x32( _mm_aesenc_si128( k01, m128_zero ) ); 
   k01 = _mm_xor_si128( k01, k00 );

   x = _mm_xor_si128( x, k01 );
   x = _mm_aesenc_si128( x, m128_zero );
-   k02 = mm_rotr_1x32( _mm_aesenc_si128( k02, m128_zero ) );
+   k02 = mm_ror_1x32( _mm_aesenc_si128( k02, m128_zero ) );
   k02 = _mm_xor_si128( k02, k01 );

   x = _mm_xor_si128( x, k02 );
   x = _mm_aesenc_si128( x, m128_zero );
-   k03 = mm_rotr_1x32( _mm_aesenc_si128( k03, m128_zero ) );
+   k03 = mm_ror_1x32( _mm_aesenc_si128( k03, m128_zero ) );
   k03 = _mm_xor_si128( k03, k02 );

   x = _mm_xor_si128( x, k03 );
   x = _mm_aesenc_si128( x, m128_zero );
   p3 = _mm_xor_si128( p3, x );
-   k10 = mm_rotr_1x32( _mm_aesenc_si128( k10, m128_zero ) );
+   k10 = mm_ror_1x32( _mm_aesenc_si128( k10, m128_zero ) );
   k10 = _mm_xor_si128( k10, k03 );

   x = _mm_xor_si128( p2, k10 );
   x = _mm_aesenc_si128( x, m128_zero );
-   k11 = mm_rotr_1x32( _mm_aesenc_si128( k11, m128_zero ) );
+   k11 = mm_ror_1x32( _mm_aesenc_si128( k11, m128_zero ) );
   k11 = _mm_xor_si128( k11, k10 );

   x = _mm_xor_si128( x, k11 );
   x = _mm_aesenc_si128( x, m128_zero );
-   k12 = mm_rotr_1x32( _mm_aesenc_si128( k12, m128_zero ) );
+   k12 = mm_ror_1x32( _mm_aesenc_si128( k12, m128_zero ) );
   k12 = _mm_xor_si128( k12, _mm_xor_si128( k11, _mm_set_epi32(
               ~sc->count2, sc->count3, sc->count0, sc->count1 ) ) );

   x = _mm_xor_si128( x, k12 );
   x = _mm_aesenc_si128( x, m128_zero );
-   k13 = mm_rotr_1x32( _mm_aesenc_si128( k13, m128_zero ) );
+   k13 = mm_ror_1x32( _mm_aesenc_si128( k13, m128_zero ) );
   k13 = _mm_xor_si128( k13, k12 );

   x = _mm_xor_si128( x, k13 );
@@ -558,7 +352,6 @@ c512( sph_shavite_big_context *sc, const void *msg )
   h[3] = _mm_xor_si128( h[3], p1 );
 }

-#endif

 static void
 shavite_big_aesni_init( sph_shavite_big_context *sc, const sph_u32 *iv )
--- a/algo/yescrypt/yescrypt-simd.c
+++ b/algo/yescrypt/yescrypt-simd.c
@@ -1363,10 +1363,11 @@ yescrypt_kdf(const yescrypt_shared_t * shared, yescrypt_local_t * local,
 	   {
 		HMAC_SHA256_CTX ctx;
 		HMAC_SHA256_Init(&ctx, buf, buflen);
-                if ( client_key_hack ) // GlobalBoost-Y buggy yescrypt
-			HMAC_SHA256_Update(&ctx, salt, saltlen);
-                else // Proper yescrypt
-                        HMAC_SHA256_Update(&ctx, "Client Key", 10);
+                if ( yescrypt_client_key )
+                    HMAC_SHA256_Update( &ctx, (uint8_t*)yescrypt_client_key,
+                                        yescrypt_client_key_len );
+                else
+                    HMAC_SHA256_Update( &ctx, salt, saltlen );
 		HMAC_SHA256_Final(sha256, &ctx);
 	   }
 	   /* Compute StoredKey */
--- a/algo/yescrypt/yescrypt.c
+++ b/algo/yescrypt/yescrypt.c
@@ -25,7 +25,7 @@
 #include "compat.h"

 #include "yescrypt.h"
-
+#include "sha256_Y.h"
 #include "algo-gate-api.h"

 #define BYTES2CHARS(bytes) \
@@ -366,7 +366,8 @@ static int yescrypt_bsty(const uint8_t * passwd, size_t passwdlen,
 uint64_t YESCRYPT_N;
 uint32_t YESCRYPT_R;
 uint32_t YESCRYPT_P;
-bool client_key_hack;
+char *yescrypt_client_key = NULL;
+int yescrypt_client_key_len = 0;

 /* main hash 80 bytes input */
 void yescrypt_hash( const char *input, char *output, uint32_t len )
@@ -436,7 +437,8 @@ bool register_yescrypt_algo( algo_gate_t* gate )
 {
   yescrypt_gate_base( gate );
   gate->get_max64  = (void*)&yescrypt_get_max64;
-   client_key_hack = true;
+   yescrypt_client_key = NULL;
+   yescrypt_client_key_len = 0;
   YESCRYPT_N = 2048;
   YESCRYPT_R = 8;
   YESCRYPT_P = 1;
@@ -447,7 +449,8 @@ bool register_yescryptr8_algo( algo_gate_t* gate )
 {
   yescrypt_gate_base( gate );
   gate->get_max64  = (void*)&yescrypt_get_max64;
-   client_key_hack = false;
+   yescrypt_client_key = "Client Key";
+   yescrypt_client_key_len = 10;
   YESCRYPT_N = 2048;
   YESCRYPT_R = 8;
   YESCRYPT_P = 1;
@@ -458,10 +461,23 @@ bool register_yescryptr16_algo( algo_gate_t* gate )
 {
   yescrypt_gate_base( gate );
   gate->get_max64  = (void*)&yescryptr16_get_max64;
-   client_key_hack = false;
+   yescrypt_client_key = "Client Key";
+   yescrypt_client_key_len = 10;
   YESCRYPT_N = 4096;   
   YESCRYPT_R = 16;   
   YESCRYPT_P = 1;   
   return true;
 }

+bool register_yescryptr32_algo( algo_gate_t* gate )
+{
+   yescrypt_gate_base( gate );
+   gate->get_max64  = (void*)&yescryptr16_get_max64;
+   yescrypt_client_key = "WaviBanana";
+   yescrypt_client_key_len = 10;
+   YESCRYPT_N = 4096;
+   YESCRYPT_R = 32;
+   YESCRYPT_P = 1;
+   return true;
+}
+
--- a/algo/yescrypt/yescrypt.h
+++ b/algo/yescrypt/yescrypt.h
@@ -108,7 +108,8 @@ typedef enum {
 	__YESCRYPT_INIT_SHARED = 0x30000
 } yescrypt_flags_t;

-extern bool client_key_hack;    // true for GlobalBoost-Y
+extern char *yescrypt_client_key;
+extern int yescrypt_client_key_len;


 #define YESCRYPT_KNOWN_FLAGS \
--- a/api.c
+++ b/api.c
@@ -158,11 +158,12 @@ static char *getsummary( char *params )

   *buffer = '\0';
   sprintf( buffer, "NAME=%s;VER=%s;API=%s;"
-                    "ALGO=%s;CPUS=%d;HS=%.2f;KHS=%.2f;ACC=%d;REJ=%d;SOL=%d;"
+                    "ALGO=%s;CPUS=%d;URL=%s;"
+                    "HS=%.2f;KHS=%.2f;ACC=%d;REJ=%d;SOL=%d;"
                    "ACCMN=%.3f;DIFF=%s;TEMP=%.1f;FAN=%d;FREQ=%d;"
                    "UPTIME=%.0f;TS=%u|",
                    PACKAGE_NAME, PACKAGE_VERSION, APIVERSION,
-                    algo, opt_n_threads, hrate, hrate/1000.0,
+                    algo, opt_n_threads, short_url, hrate, hrate/1000.0,
                    accepted_count, rejected_count, solved_count,
                    accps, diff_str, cpu.cpu_temp, cpu.cpu_fan, cpu.cpu_clock,
                    uptime, (uint32_t) ts);
--- a/avxdefs.h
+++ b/avxdefs.h
--- a/20
+++ b/20
@@ -1,6 +1,6 @@
 #! /bin/sh
 # Guess values for system-dependent variables and create Makefiles.
-# Generated by GNU Autoconf 2.69 for cpuminer-opt 3.8.3.3.
+# Generated by GNU Autoconf 2.69 for cpuminer-opt 3.8.4.1.
 #
 #
 # Copyright (C) 1992-1996, 1998-2012 Free Software Foundation, Inc.
@@ -577,8 +577,8 @@ MAKEFLAGS=
 # Identity of this package.
 PACKAGE_NAME='cpuminer-opt'
 PACKAGE_TARNAME='cpuminer-opt'
-PACKAGE_VERSION='3.8.3.3'
-PACKAGE_STRING='cpuminer-opt 3.8.3.3'
+PACKAGE_VERSION='3.8.4.1'
+PACKAGE_STRING='cpuminer-opt 3.8.4.1'
 PACKAGE_BUGREPORT=''
 PACKAGE_URL=''

@@ -1321,7 +1321,7 @@ if test "$ac_init_help" = "long"; then
  # Omit some internal or obsolete options to make the list less imposing.
  # This message is too long to be a string in the A/UX 3.1 sh.
  cat <<_ACEOF
-\`configure' configures cpuminer-opt 3.8.3.3 to adapt to many kinds of systems.
+\`configure' configures cpuminer-opt 3.8.4.1 to adapt to many kinds of systems.

 Usage: $0 [OPTION]... [VAR=VALUE]...

@@ -1392,7 +1392,7 @@ fi

 if test -n "$ac_init_help"; then
  case $ac_init_help in
-     short | recursive ) echo "Configuration of cpuminer-opt 3.8.3.3:";;
+     short | recursive ) echo "Configuration of cpuminer-opt 3.8.4.1:";;
   esac
  cat <<\_ACEOF

@@ -1497,7 +1497,7 @@ fi
 test -n "$ac_init_help" && exit $ac_status
 if $ac_init_version; then
  cat <<\_ACEOF
-cpuminer-opt configure 3.8.3.3
+cpuminer-opt configure 3.8.4.1
 generated by GNU Autoconf 2.69

 Copyright (C) 2012 Free Software Foundation, Inc.
@@ -2000,7 +2000,7 @@ cat >config.log <<_ACEOF
 This file contains any messages produced by compilers while
 running configure, to aid debugging if configure makes a mistake.

-It was created by cpuminer-opt $as_me 3.8.3.3, which was
+It was created by cpuminer-opt $as_me 3.8.4.1, which was
 generated by GNU Autoconf 2.69.  Invocation command line was

  $ $0 $@
@@ -2981,7 +2981,7 @@ fi

 # Define the identity of the package.
 PACKAGE='cpuminer-opt'
- VERSION='3.8.3.3'
+ VERSION='3.8.4.1'


 cat >>confdefs.h <<_ACEOF
@@ -6677,7 +6677,7 @@ cat >>$CONFIG_STATUS <<\_ACEOF || ac_write_fail=1
 # report actual input values of CONFIG_FILES etc. instead of their
 # values after options handling.
 ac_log="
-This file was extended by cpuminer-opt $as_me 3.8.3.3, which was
+This file was extended by cpuminer-opt $as_me 3.8.4.1, which was
 generated by GNU Autoconf 2.69.  Invocation command line was

  CONFIG_FILES    = $CONFIG_FILES
@@ -6743,7 +6743,7 @@ _ACEOF
 cat >>$CONFIG_STATUS <<_ACEOF || ac_write_fail=1
 ac_cs_config="`$as_echo "$ac_configure_args" | sed 's/^ //; s/[\\""\`\$]/\\\\&/g'`"
 ac_cs_version="\\
-cpuminer-opt config.status 3.8.3.3
+cpuminer-opt config.status 3.8.4.1
 configured by $0, generated by GNU Autoconf 2.69,
  with options \\"\$ac_cs_config\\"

--- a/configure.ac
+++ b/configure.ac
@@ -1,4 +1,4 @@
-AC_INIT([cpuminer-opt], [3.8.3.3])
+AC_INIT([cpuminer-opt], [3.8.4.1])

 AC_PREREQ([2.59c])
 AC_CANONICAL_SYSTEM
--- a/cpu-miner.c
+++ b/cpu-miner.c
@@ -103,7 +103,7 @@ enum algos opt_algo = ALGO_NULL;
 int opt_scrypt_n = 0;
 int opt_pluck_n = 128;
 int opt_n_threads = 0;
-#ifdef __GNUC__
+#if ( __GNUC__ > 4 ) || ( ( __GNUC__ == 4 ) && ( __GNUC_MINOR__ >= 8 ) )
 __int128_t opt_affinity = -1LL;
 #else
 int64_t opt_affinity = -1LL;
@@ -200,20 +200,20 @@ static inline void drop_policy(void)
 #define pthread_setaffinity_np(tid,sz,s) {} /* only do process affinity */
 #endif

-#ifdef __GNUC__
+#if ( __GNUC__ > 4 ) || ( ( __GNUC__ == 4 ) && ( __GNUC_MINOR__ >= 8 ) )
 static void affine_to_cpu_mask( int id, unsigned __int128 mask )
 #else
 static void affine_to_cpu_mask( int id, unsigned long long mask )
 #endif
 {
   cpu_set_t set;
-   CPU_ZERO(&set);
+   CPU_ZERO( &set );
   uint8_t ncpus = (num_cpus > 256) ? 256 : num_cpus;       

   for ( uint8_t i = 0; i < ncpus; i++ ) 
   {
      // cpu mask
-#ifdef __GNUC__
+#if ( __GNUC__ > 4 ) || ( ( __GNUC__ == 4 ) && ( __GNUC_MINOR__ >= 8 ) )
      if( ( mask & ( (unsigned __int128)1ULL << i ) ) )  CPU_SET( i, &set );
 #else
      if( (ncpus > 64) || ( mask & (1ULL << i) ) )  CPU_SET( i, &set );
@@ -1792,7 +1792,7 @@ static void *miner_thread( void *userdata )
         if (opt_debug)
            applog( LOG_DEBUG, "Binding thread %d to cpu %d (mask %x)",
                   thr_id, thr_id % num_cpus, ( 1ULL << (thr_id % num_cpus) ) );
-#ifdef __GNUC__
+#if ( __GNUC__ > 4 ) || ( ( __GNUC__ == 4 ) && ( __GNUC_MINOR__ >= 8 ) )
         affine_to_cpu_mask( thr_id,
                             (unsigned __int128)1LL << (thr_id % num_cpus) );
 #else
--- a/miner.h
+++ b/miner.h
@@ -424,7 +424,7 @@ extern size_t rpc2_bloblen;
 extern uint32_t rpc2_target;
 extern char *rpc2_job_id;
 extern char *rpc_user;
-
+extern char *short_url;

 json_t *json_rpc2_call(CURL *curl, const char *url, const char *userpass, const char *rpc_req, int *curl_err, int flags);
 bool rpc2_login(CURL *curl);
@@ -553,6 +553,7 @@ enum algos {
        ALGO_YESCRYPT,
        ALGO_YESCRYPTR8,
        ALGO_YESCRYPTR16,
+        ALGO_YESCRYPTR32,
        ALGO_ZR5,
        ALGO_COUNT
 };
@@ -629,6 +630,7 @@ static const char* const algo_names[] = {
        "yescrypt",
        "yescryptr8",
        "yescryptr16",
+        "yescryptr32",
        "zr5",
        "\0"
 };
@@ -764,6 +766,7 @@ Options:\n\
                          yescrypt     Globlboost-Y (BSTY)\n\
                          yescryptr8   BitZeny (ZNY)\n\
                          yescryptr16  Yenten (YTN)\n\
+                          yescryptr32  WAVI\n\
                          zr5          Ziftr\n\
  -o, --url=URL         URL of mining server\n\
  -O, --userpass=U:P    username:password pair for mining server\n\
Author	SHA1	Message	Date
Jay D Dee	3363d61524	v3.8.4.1	2018-03-22 14:28:03 -04:00
Jay D Dee	20fe05054c	v3.8.4	2018-03-18 12:51:03 -04:00