/* dvdisaster: Additional error correction for optical media.
* Copyright (C) 2004-2015 Carsten Gnoerlich.
*
* Email: carsten@dvdisaster.org -or- cgnoerlich@fsfe.org
* Project homepage: http://www.dvdisaster.org
*
* This file is part of dvdisaster.
*
* dvdisaster is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* dvdisaster is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with dvdisaster. If not, see .
*/
#include "dvdisaster.h"
/***
*** Reed-Solomon encoding
***/
/* Portable (non-SSE2) version.
* Using 32bit operands seems to be a good choice for the lowest
* common denominator between the non-SSE2 systems.
*/
#ifdef HAVE_BIG_ENDIAN
#define SHIFT_LEFT <<
#define SHIFT_RIGHT >>
#else
#define SHIFT_LEFT >>
#define SHIFT_RIGHT <<
#endif /* HAVE_BIG_ENDIAN */
static void encode_next_layer_portable(ReedSolomonTables *rt, unsigned char *data, unsigned char *parity, guint64 layer_size, int shift)
{ gint32 *gf_index_of = rt->gfTables->indexOf;
gint32 *enc_alpha_to = rt->gfTables->encAlphaTo;
gint32 *rs_gpoly = rt->gpoly;
int nroots = rt->nroots;
int nroots_aligned = (nroots+15)&~15;
int nroots_aligned32 = (nroots+3)&~3;
int nroots_full = nroots_aligned32>>2;
int i,j;
for(i=0; ibLut[feedback]+(offset&~3)));
/* Process lut in 32 bit steps */
switch(byte_offset)
{ case 0:
for(j=nroots_full; j; j--)
*par_idx++ ^= *e_lut++;
break;
case 1:
{ for(j=nroots_full; j; j--)
{ guint32 span = *e_lut SHIFT_LEFT 8;
e_lut++;
span |= *e_lut SHIFT_RIGHT 24;
*par_idx++ ^= span;
}
}
break;
case 2:
{ for(j=nroots_full; j; j--)
{ guint32 span = *e_lut SHIFT_LEFT 16;
e_lut++;
span |= *e_lut SHIFT_RIGHT 16;
*par_idx++ ^= span;
}
}
break;
case 3:
{ for(j=nroots_full; j; j--)
{ guint32 span = *e_lut SHIFT_LEFT 24;
e_lut++;
span |= *e_lut SHIFT_RIGHT 8;
*par_idx++ ^= span;
}
}
break;
}
parity[shift] = enc_alpha_to[feedback + rs_gpoly[0]];
}
else /* zero feedback term */
parity[shift] = 0;
parity += nroots_aligned;
}
}
/* 64bit integer (non-SSE2) version.
* May perform better on systems which have shared FPU/SSE2 units
* between several cores.
*/
static void encode_next_layer_64bit(ReedSolomonTables *rt, unsigned char *data, unsigned char *parity, guint64 layer_size, int shift)
{ gint32 *gf_index_of = rt->gfTables->indexOf;
gint32 *enc_alpha_to = rt->gfTables->encAlphaTo;
gint32 *rs_gpoly = rt->gpoly;
int nroots = rt->nroots;
int nroots_aligned = (nroots+15)&~15;
int nroots_aligned64 = (nroots+7)&~7;
int nroots_full = nroots_aligned64>>3;
int i,j;
for(i=0; ibLut[feedback]+(offset&~7)));
/* Process lut in 64 bit steps */
switch(byte_offset)
{ case 0:
for(j=nroots_full; j; j--)
*par_idx++ ^= *e_lut++;
break;
case 1:
{ for(j=nroots_full; j; j--)
{ guint64 span = *e_lut SHIFT_LEFT 8;
e_lut++;
span |= *e_lut SHIFT_RIGHT 56;
*par_idx++ ^= span;
}
}
break;
case 2:
{ for(j=nroots_full; j; j--)
{ guint64 span = *e_lut SHIFT_LEFT 16;
e_lut++;
span |= *e_lut SHIFT_RIGHT 48;
*par_idx++ ^= span;
}
}
break;
case 3:
{ for(j=nroots_full; j; j--)
{ guint64 span = *e_lut SHIFT_LEFT 24;
e_lut++;
span |= *e_lut SHIFT_RIGHT 40;
*par_idx++ ^= span;
}
}
break;
case 4:
{ for(j=nroots_full; j; j--)
{ guint64 span = *e_lut SHIFT_LEFT 32;
e_lut++;
span |= *e_lut SHIFT_RIGHT 32;
*par_idx++ ^= span;
}
}
break;
case 5:
{ for(j=nroots_full; j; j--)
{ guint64 span = *e_lut SHIFT_LEFT 40;
e_lut++;
span |= *e_lut SHIFT_RIGHT 24;
*par_idx++ ^= span;
}
}
break;
case 6:
{ for(j=nroots_full; j; j--)
{ guint64 span = *e_lut SHIFT_LEFT 48;
e_lut++;
span |= *e_lut SHIFT_RIGHT 16;
*par_idx++ ^= span;
}
}
break;
case 7:
{ for(j=nroots_full; j; j--)
{ guint64 span = *e_lut SHIFT_LEFT 56;
e_lut++;
span |= *e_lut SHIFT_RIGHT 8;
*par_idx++ ^= span;
}
}
break;
}
parity[shift] = enc_alpha_to[feedback + rs_gpoly[0]];
}
else /* zero feedback term */
parity[shift] = 0;
parity += nroots_aligned;
}
}
/*
* Dispatch upon availability of SSE2 intrinsics
*/
void encode_next_layer_sse2(ReedSolomonTables*, unsigned char*, unsigned char*, guint64, int);
void encode_next_layer_altivec(ReedSolomonTables*, unsigned char*, unsigned char*, guint64, int);
void EncodeNextLayer(ReedSolomonTables *rt, unsigned char *data, unsigned char *parity, guint64 layer_size, int shift)
{
switch(Closure->encodingAlgorithm)
{ case ENCODING_ALG_32BIT:
encode_next_layer_portable(rt, data, parity, layer_size, shift);
break;
case ENCODING_ALG_64BIT:
encode_next_layer_64bit(rt, data, parity, layer_size, shift);
break;
case ENCODING_ALG_SSE2:
encode_next_layer_sse2(rt, data, parity, layer_size, shift);
break;
case ENCODING_ALG_ALTIVEC:
encode_next_layer_altivec(rt, data, parity, layer_size, shift);
break;
case ENCODING_ALG_DEFAULT:
if(Closure->useSSE2)
encode_next_layer_sse2(rt, data, parity, layer_size, shift);
else if(Closure->useAltiVec)
encode_next_layer_altivec(rt, data, parity, layer_size, shift);
else
encode_next_layer_portable(rt, data, parity, layer_size, shift);
break;
}
}
/*
* Provide textual description for current encoder parameters
*/
void DescribeRSEncoder(char **algorithm, char **iostrategy)
{
switch(Closure->encodingAlgorithm)
{ case ENCODING_ALG_32BIT:
*algorithm="32bit";
break;
case ENCODING_ALG_64BIT:
*algorithm="64bit";
break;
case ENCODING_ALG_SSE2:
*algorithm="SSE2";
break;
case ENCODING_ALG_ALTIVEC:
*algorithm="AltiVec";
break;
case ENCODING_ALG_DEFAULT:
if(Closure->useSSE2)
*algorithm="SSE2";
else if(Closure->useAltiVec)
*algorithm="AltiVec";
else
*algorithm="64bit";
break;
}
if(Closure->encodingIOStrategy == IO_STRATEGY_MMAP)
*iostrategy="mmap";
else *iostrategy="read/write";
}