ceph/src/isa-l/erasure_code/ec_base.c

   1 /**********************************************************************
   2   Copyright(c) 2011-2015 Intel Corporation All rights reserved.
   3
   4   Redistribution and use in source and binary forms, with or without
   5   modification, are permitted provided that the following conditions
   6   are met:
   7     * Redistributions of source code must retain the above copyright
   8       notice, this list of conditions and the following disclaimer.
   9     * Redistributions in binary form must reproduce the above copyright
  10       notice, this list of conditions and the following disclaimer in
  11       the documentation and/or other materials provided with the
  12       distribution.
  13     * Neither the name of Intel Corporation nor the names of its
  14       contributors may be used to endorse or promote products derived
  15       from this software without specific prior written permission.
  16
  17   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  18   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  19   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
  20   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
  21   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
  22   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
  23   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
  24   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
  25   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  26   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
  27   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  28 **********************************************************************/
  29
  30 #include <limits.h>
  31 #include <string.h>             // for memset
  32 #include "erasure_code.h"
  33 #include "ec_base.h"            // for GF tables
  34 #include "types.h"
  35
  36 unsigned char gf_mul(unsigned char a, unsigned char b)
  37 {
  38 #ifndef GF_LARGE_TABLES
  39         int i;
  40
  41         if ((a == 0) || (b == 0))
  42                 return 0;
  43
  44         return gff_base[(i = gflog_base[a] + gflog_base[b]) > 254 ? i - 255 : i];
  45 #else
  46         return gf_mul_table_base[b * 256 + a];
  47 #endif
  48 }
  49
  50 unsigned char gf_inv(unsigned char a)
  51 {
  52 #ifndef GF_LARGE_TABLES
  53         if (a == 0)
  54                 return 0;
  55
  56         return gff_base[255 - gflog_base[a]];
  57 #else
  58         return gf_inv_table_base[a];
  59 #endif
  60 }
  61
  62 void gf_gen_rs_matrix(unsigned char *a, int m, int k)
  63 {
  64         int i, j;
  65         unsigned char p, gen = 1;
  66
  67         memset(a, 0, k * m);
  68         for (i = 0; i < k; i++)
  69                 a[k * i + i] = 1;
  70
  71         for (i = k; i < m; i++) {
  72                 p = 1;
  73                 for (j = 0; j < k; j++) {
  74                         a[k * i + j] = p;
  75                         p = gf_mul(p, gen);
  76                 }
  77                 gen = gf_mul(gen, 2);
  78         }
  79 }
  80
  81 void gf_gen_cauchy1_matrix(unsigned char *a, int m, int k)
  82 {
  83         int i, j;
  84         unsigned char *p;
  85
  86         // Identity matrix in high position
  87         memset(a, 0, k * m);
  88         for (i = 0; i < k; i++)
  89                 a[k * i + i] = 1;
  90
  91         // For the rest choose 1/(i + j) | i != j
  92         p = &a[k * k];
  93         for (i = k; i < m; i++)
  94                 for (j = 0; j < k; j++)
  95                         *p++ = gf_inv(i ^ j);
  96
  97 }
  98
  99 int gf_invert_matrix(unsigned char *in_mat, unsigned char *out_mat, const int n)
 100 {
 101         int i, j, k;
 102         unsigned char temp;
 103
 104         // Set out_mat[] to the identity matrix
 105         for (i = 0; i < n * n; i++)     // memset(out_mat, 0, n*n)
 106                 out_mat[i] = 0;
 107
 108         for (i = 0; i < n; i++)
 109                 out_mat[i * n + i] = 1;
 110
 111         // Inverse
 112         for (i = 0; i < n; i++) {
 113                 // Check for 0 in pivot element
 114                 if (in_mat[i * n + i] == 0) {
 115                         // Find a row with non-zero in current column and swap
 116                         for (j = i + 1; j < n; j++)
 117                                 if (in_mat[j * n + i])
 118                                         break;
 119
 120                         if (j == n)     // Couldn't find means it's singular
 121                                 return -1;
 122
 123                         for (k = 0; k < n; k++) {       // Swap rows i,j
 124                                 temp = in_mat[i * n + k];
 125                                 in_mat[i * n + k] = in_mat[j * n + k];
 126                                 in_mat[j * n + k] = temp;
 127
 128                                 temp = out_mat[i * n + k];
 129                                 out_mat[i * n + k] = out_mat[j * n + k];
 130                                 out_mat[j * n + k] = temp;
 131                         }
 132                 }
 133
 134                 temp = gf_inv(in_mat[i * n + i]);       // 1/pivot
 135                 for (j = 0; j < n; j++) {       // Scale row i by 1/pivot
 136                         in_mat[i * n + j] = gf_mul(in_mat[i * n + j], temp);
 137                         out_mat[i * n + j] = gf_mul(out_mat[i * n + j], temp);
 138                 }
 139
 140                 for (j = 0; j < n; j++) {
 141                         if (j == i)
 142                                 continue;
 143
 144                         temp = in_mat[j * n + i];
 145                         for (k = 0; k < n; k++) {
 146                                 out_mat[j * n + k] ^= gf_mul(temp, out_mat[i * n + k]);
 147                                 in_mat[j * n + k] ^= gf_mul(temp, in_mat[i * n + k]);
 148                         }
 149                 }
 150         }
 151         return 0;
 152 }
 153
 154 // Calculates const table gftbl in GF(2^8) from single input A
 155 // gftbl(A) = {A{00}, A{01}, A{02}, ... , A{0f} }, {A{00}, A{10}, A{20}, ... , A{f0} }
 156
 157 void gf_vect_mul_init(unsigned char c, unsigned char *tbl)
 158 {
 159         unsigned char c2 = (c << 1) ^ ((c & 0x80) ? 0x1d : 0);  //Mult by GF{2}
 160         unsigned char c4 = (c2 << 1) ^ ((c2 & 0x80) ? 0x1d : 0);        //Mult by GF{2}
 161         unsigned char c8 = (c4 << 1) ^ ((c4 & 0x80) ? 0x1d : 0);        //Mult by GF{2}
 162
 163 #if __WORDSIZE == 64 || _WIN64 || __x86_64__
 164         unsigned long long v1, v2, v4, v8, *t;
 165         unsigned long long v10, v20, v40, v80;
 166         unsigned char c17, c18, c20, c24;
 167
 168         t = (unsigned long long *)tbl;
 169
 170         v1 = c * 0x0100010001000100ull;
 171         v2 = c2 * 0x0101000001010000ull;
 172         v4 = c4 * 0x0101010100000000ull;
 173         v8 = c8 * 0x0101010101010101ull;
 174
 175         v4 = v1 ^ v2 ^ v4;
 176         t[0] = v4;
 177         t[1] = v8 ^ v4;
 178
 179         c17 = (c8 << 1) ^ ((c8 & 0x80) ? 0x1d : 0);     //Mult by GF{2}
 180         c18 = (c17 << 1) ^ ((c17 & 0x80) ? 0x1d : 0);   //Mult by GF{2}
 181         c20 = (c18 << 1) ^ ((c18 & 0x80) ? 0x1d : 0);   //Mult by GF{2}
 182         c24 = (c20 << 1) ^ ((c20 & 0x80) ? 0x1d : 0);   //Mult by GF{2}
 183
 184         v10 = c17 * 0x0100010001000100ull;
 185         v20 = c18 * 0x0101000001010000ull;
 186         v40 = c20 * 0x0101010100000000ull;
 187         v80 = c24 * 0x0101010101010101ull;
 188
 189         v40 = v10 ^ v20 ^ v40;
 190         t[2] = v40;
 191         t[3] = v80 ^ v40;
 192
 193 #else // 32-bit or other
 194         unsigned char c3, c5, c6, c7, c9, c10, c11, c12, c13, c14, c15;
 195         unsigned char c17, c18, c19, c20, c21, c22, c23, c24, c25, c26, c27, c28, c29, c30,
 196             c31;
 197
 198         c3 = c2 ^ c;
 199         c5 = c4 ^ c;
 200         c6 = c4 ^ c2;
 201         c7 = c4 ^ c3;
 202
 203         c9 = c8 ^ c;
 204         c10 = c8 ^ c2;
 205         c11 = c8 ^ c3;
 206         c12 = c8 ^ c4;
 207         c13 = c8 ^ c5;
 208         c14 = c8 ^ c6;
 209         c15 = c8 ^ c7;
 210
 211         tbl[0] = 0;
 212         tbl[1] = c;
 213         tbl[2] = c2;
 214         tbl[3] = c3;
 215         tbl[4] = c4;
 216         tbl[5] = c5;
 217         tbl[6] = c6;
 218         tbl[7] = c7;
 219         tbl[8] = c8;
 220         tbl[9] = c9;
 221         tbl[10] = c10;
 222         tbl[11] = c11;
 223         tbl[12] = c12;
 224         tbl[13] = c13;
 225         tbl[14] = c14;
 226         tbl[15] = c15;
 227
 228         c17 = (c8 << 1) ^ ((c8 & 0x80) ? 0x1d : 0);     //Mult by GF{2}
 229         c18 = (c17 << 1) ^ ((c17 & 0x80) ? 0x1d : 0);   //Mult by GF{2}
 230         c19 = c18 ^ c17;
 231         c20 = (c18 << 1) ^ ((c18 & 0x80) ? 0x1d : 0);   //Mult by GF{2}
 232         c21 = c20 ^ c17;
 233         c22 = c20 ^ c18;
 234         c23 = c20 ^ c19;
 235         c24 = (c20 << 1) ^ ((c20 & 0x80) ? 0x1d : 0);   //Mult by GF{2}
 236         c25 = c24 ^ c17;
 237         c26 = c24 ^ c18;
 238         c27 = c24 ^ c19;
 239         c28 = c24 ^ c20;
 240         c29 = c24 ^ c21;
 241         c30 = c24 ^ c22;
 242         c31 = c24 ^ c23;
 243
 244         tbl[16] = 0;
 245         tbl[17] = c17;
 246         tbl[18] = c18;
 247         tbl[19] = c19;
 248         tbl[20] = c20;
 249         tbl[21] = c21;
 250         tbl[22] = c22;
 251         tbl[23] = c23;
 252         tbl[24] = c24;
 253         tbl[25] = c25;
 254         tbl[26] = c26;
 255         tbl[27] = c27;
 256         tbl[28] = c28;
 257         tbl[29] = c29;
 258         tbl[30] = c30;
 259         tbl[31] = c31;
 260
 261 #endif //__WORDSIZE == 64 || _WIN64 || __x86_64__
 262 }
 263
 264 void gf_vect_dot_prod_base(int len, int vlen, unsigned char *v,
 265                            unsigned char **src, unsigned char *dest)
 266 {
 267         int i, j;
 268         unsigned char s;
 269         for (i = 0; i < len; i++) {
 270                 s = 0;
 271                 for (j = 0; j < vlen; j++)
 272                         s ^= gf_mul(src[j][i], v[j * 32 + 1]);
 273
 274                 dest[i] = s;
 275         }
 276 }
 277
 278 void gf_vect_mad_base(int len, int vec, int vec_i,
 279                       unsigned char *v, unsigned char *src, unsigned char *dest)
 280 {
 281         int i;
 282         unsigned char s;
 283         for (i = 0; i < len; i++) {
 284                 s = dest[i];
 285                 s ^= gf_mul(src[i], v[vec_i * 32 + 1]);
 286                 dest[i] = s;
 287         }
 288 }
 289
 290 void ec_encode_data_base(int len, int srcs, int dests, unsigned char *v,
 291                          unsigned char **src, unsigned char **dest)
 292 {
 293         int i, j, l;
 294         unsigned char s;
 295
 296         for (l = 0; l < dests; l++) {
 297                 for (i = 0; i < len; i++) {
 298                         s = 0;
 299                         for (j = 0; j < srcs; j++)
 300                                 s ^= gf_mul(src[j][i], v[j * 32 + l * srcs * 32 + 1]);
 301
 302                         dest[l][i] = s;
 303                 }
 304         }
 305 }
 306
 307 void ec_encode_data_update_base(int len, int k, int rows, int vec_i, unsigned char *v,
 308                                 unsigned char *data, unsigned char **dest)
 309 {
 310         int i, l;
 311         unsigned char s;
 312
 313         for (l = 0; l < rows; l++) {
 314                 for (i = 0; i < len; i++) {
 315                         s = dest[l][i];
 316                         s ^= gf_mul(data[i], v[vec_i * 32 + l * k * 32 + 1]);
 317
 318                         dest[l][i] = s;
 319                 }
 320         }
 321 }
 322
 323 void gf_vect_mul_base(int len, unsigned char *a, unsigned char *src, unsigned char *dest)
 324 {
 325         //2nd element of table array is ref value used to fill it in
 326         unsigned char c = a[1];
 327         while (len-- > 0)
 328                 *dest++ = gf_mul(c, *src++);
 329 }
 330
 331 struct slver {
 332         UINT16 snum;
 333         UINT8 ver;
 334         UINT8 core;
 335 };
 336
 337 // Version info
 338 struct slver gf_vect_mul_init_slver_00020035;
 339 struct slver gf_vect_mul_init_slver = { 0x0035, 0x02, 0x00 };
 340
 341 struct slver ec_encode_data_base_slver_00010135;
 342 struct slver ec_encode_data_base_slver = { 0x0135, 0x01, 0x00 };
 343
 344 struct slver gf_vect_mul_base_slver_00010136;
 345 struct slver gf_vect_mul_base_slver = { 0x0136, 0x01, 0x00 };
 346
 347 struct slver gf_vect_dot_prod_base_slver_00010137;
 348 struct slver gf_vect_dot_prod_base_slver = { 0x0137, 0x01, 0x00 };
 349
 350 struct slver gf_mul_slver_00000214;
 351 struct slver gf_mul_slver = { 0x0214, 0x00, 0x00 };
 352
 353 struct slver gf_invert_matrix_slver_00000215;
 354 struct slver gf_invert_matrix_slver = { 0x0215, 0x00, 0x00 };
 355
 356 struct slver gf_gen_rs_matrix_slver_00000216;
 357 struct slver gf_gen_rs_matrix_slver = { 0x0216, 0x00, 0x00 };
 358
 359 struct slver gf_gen_cauchy1_matrix_slver_00000217;
 360 struct slver gf_gen_cauchy1_matrix_slver = { 0x0217, 0x00, 0x00 };