target-sparc/vis_helper.c

   1 /*
   2  * VIS op helpers
   3  *
   4  *  Copyright (c) 2003-2005 Fabrice Bellard
   5  *
   6  * This library is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2 of the License, or (at your option) any later version.
  10  *
  11  * This library is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with this library; if not, see <http://www.gnu.org/licenses/>.
  18  */
  19
  20 #include "cpu.h"
  21 #include "helper.h"
  22
  23 /* This function uses non-native bit order */
  24 #define GET_FIELD(X, FROM, TO)                                  \
  25     ((X) >> (63 - (TO)) & ((1ULL << ((TO) - (FROM) + 1)) - 1))
  26
  27 /* This function uses the order in the manuals, i.e. bit 0 is 2^0 */
  28 #define GET_FIELD_SP(X, FROM, TO)               \
  29     GET_FIELD(X, 63 - (TO), 63 - (FROM))
  30
  31 target_ulong helper_array8(CPUState *env, target_ulong pixel_addr,
  32                            target_ulong cubesize)
  33 {
  34     return (GET_FIELD_SP(pixel_addr, 60, 63) << (17 + 2 * cubesize)) |
  35         (GET_FIELD_SP(pixel_addr, 39, 39 + cubesize - 1) << (17 + cubesize)) |
  36         (GET_FIELD_SP(pixel_addr, 17 + cubesize - 1, 17) << 17) |
  37         (GET_FIELD_SP(pixel_addr, 56, 59) << 13) |
  38         (GET_FIELD_SP(pixel_addr, 35, 38) << 9) |
  39         (GET_FIELD_SP(pixel_addr, 13, 16) << 5) |
  40         (((pixel_addr >> 55) & 1) << 4) |
  41         (GET_FIELD_SP(pixel_addr, 33, 34) << 2) |
  42         GET_FIELD_SP(pixel_addr, 11, 12);
  43 }
  44
  45 target_ulong helper_alignaddr(CPUState *env, target_ulong addr,
  46                               target_ulong offset)
  47 {
  48     uint64_t tmp;
  49
  50     tmp = addr + offset;
  51     env->gsr &= ~7ULL;
  52     env->gsr |= tmp & 7ULL;
  53     return tmp & ~7ULL;
  54 }
  55
  56 uint64_t helper_faligndata(CPUState *env, uint64_t src1, uint64_t src2)
  57 {
  58     uint64_t tmp;
  59
  60     tmp = src1 << ((env->gsr & 7) * 8);
  61     /* on many architectures a shift of 64 does nothing */
  62     if ((env->gsr & 7) != 0) {
  63         tmp |= src2 >> (64 - (env->gsr & 7) * 8);
  64     }
  65     return tmp;
  66 }
  67
  68 #ifdef HOST_WORDS_BIGENDIAN
  69 #define VIS_B64(n) b[7 - (n)]
  70 #define VIS_W64(n) w[3 - (n)]
  71 #define VIS_SW64(n) sw[3 - (n)]
  72 #define VIS_L64(n) l[1 - (n)]
  73 #define VIS_B32(n) b[3 - (n)]
  74 #define VIS_W32(n) w[1 - (n)]
  75 #else
  76 #define VIS_B64(n) b[n]
  77 #define VIS_W64(n) w[n]
  78 #define VIS_SW64(n) sw[n]
  79 #define VIS_L64(n) l[n]
  80 #define VIS_B32(n) b[n]
  81 #define VIS_W32(n) w[n]
  82 #endif
  83
  84 typedef union {
  85     uint8_t b[8];
  86     uint16_t w[4];
  87     int16_t sw[4];
  88     uint32_t l[2];
  89     uint64_t ll;
  90     float64 d;
  91 } VIS64;
  92
  93 typedef union {
  94     uint8_t b[4];
  95     uint16_t w[2];
  96     uint32_t l;
  97     float32 f;
  98 } VIS32;
  99
 100 uint64_t helper_fpmerge(CPUState *env, uint64_t src1, uint64_t src2)
 101 {
 102     VIS64 s, d;
 103
 104     s.ll = src1;
 105     d.ll = src2;
 106
 107     /* Reverse calculation order to handle overlap */
 108     d.VIS_B64(7) = s.VIS_B64(3);
 109     d.VIS_B64(6) = d.VIS_B64(3);
 110     d.VIS_B64(5) = s.VIS_B64(2);
 111     d.VIS_B64(4) = d.VIS_B64(2);
 112     d.VIS_B64(3) = s.VIS_B64(1);
 113     d.VIS_B64(2) = d.VIS_B64(1);
 114     d.VIS_B64(1) = s.VIS_B64(0);
 115     /* d.VIS_B64(0) = d.VIS_B64(0); */
 116
 117     return d.ll;
 118 }
 119
 120 uint64_t helper_fmul8x16(CPUState *env, uint64_t src1, uint64_t src2)
 121 {
 122     VIS64 s, d;
 123     uint32_t tmp;
 124
 125     s.ll = src1;
 126     d.ll = src2;
 127
 128 #define PMUL(r)                                                 \
 129     tmp = (int32_t)d.VIS_SW64(r) * (int32_t)s.VIS_B64(r);       \
 130     if ((tmp & 0xff) > 0x7f) {                                  \
 131         tmp += 0x100;                                           \
 132     }                                                           \
 133     d.VIS_W64(r) = tmp >> 8;
 134
 135     PMUL(0);
 136     PMUL(1);
 137     PMUL(2);
 138     PMUL(3);
 139 #undef PMUL
 140
 141     return d.ll;
 142 }
 143
 144 uint64_t helper_fmul8x16al(CPUState *env, uint64_t src1, uint64_t src2)
 145 {
 146     VIS64 s, d;
 147     uint32_t tmp;
 148
 149     s.ll = src1;
 150     d.ll = src2;
 151
 152 #define PMUL(r)                                                 \
 153     tmp = (int32_t)d.VIS_SW64(1) * (int32_t)s.VIS_B64(r);       \
 154     if ((tmp & 0xff) > 0x7f) {                                  \
 155         tmp += 0x100;                                           \
 156     }                                                           \
 157     d.VIS_W64(r) = tmp >> 8;
 158
 159     PMUL(0);
 160     PMUL(1);
 161     PMUL(2);
 162     PMUL(3);
 163 #undef PMUL
 164
 165     return d.ll;
 166 }
 167
 168 uint64_t helper_fmul8x16au(CPUState *env, uint64_t src1, uint64_t src2)
 169 {
 170     VIS64 s, d;
 171     uint32_t tmp;
 172
 173     s.ll = src1;
 174     d.ll = src2;
 175
 176 #define PMUL(r)                                                 \
 177     tmp = (int32_t)d.VIS_SW64(0) * (int32_t)s.VIS_B64(r);       \
 178     if ((tmp & 0xff) > 0x7f) {                                  \
 179         tmp += 0x100;                                           \
 180     }                                                           \
 181     d.VIS_W64(r) = tmp >> 8;
 182
 183     PMUL(0);
 184     PMUL(1);
 185     PMUL(2);
 186     PMUL(3);
 187 #undef PMUL
 188
 189     return d.ll;
 190 }
 191
 192 uint64_t helper_fmul8sux16(CPUState *env, uint64_t src1, uint64_t src2)
 193 {
 194     VIS64 s, d;
 195     uint32_t tmp;
 196
 197     s.ll = src1;
 198     d.ll = src2;
 199
 200 #define PMUL(r)                                                         \
 201     tmp = (int32_t)d.VIS_SW64(r) * ((int32_t)s.VIS_SW64(r) >> 8);       \
 202     if ((tmp & 0xff) > 0x7f) {                                          \
 203         tmp += 0x100;                                                   \
 204     }                                                                   \
 205     d.VIS_W64(r) = tmp >> 8;
 206
 207     PMUL(0);
 208     PMUL(1);
 209     PMUL(2);
 210     PMUL(3);
 211 #undef PMUL
 212
 213     return d.ll;
 214 }
 215
 216 uint64_t helper_fmul8ulx16(CPUState *env, uint64_t src1, uint64_t src2)
 217 {
 218     VIS64 s, d;
 219     uint32_t tmp;
 220
 221     s.ll = src1;
 222     d.ll = src2;
 223
 224 #define PMUL(r)                                                         \
 225     tmp = (int32_t)d.VIS_SW64(r) * ((uint32_t)s.VIS_B64(r * 2));        \
 226     if ((tmp & 0xff) > 0x7f) {                                          \
 227         tmp += 0x100;                                                   \
 228     }                                                                   \
 229     d.VIS_W64(r) = tmp >> 8;
 230
 231     PMUL(0);
 232     PMUL(1);
 233     PMUL(2);
 234     PMUL(3);
 235 #undef PMUL
 236
 237     return d.ll;
 238 }
 239
 240 uint64_t helper_fmuld8sux16(CPUState *env, uint64_t src1, uint64_t src2)
 241 {
 242     VIS64 s, d;
 243     uint32_t tmp;
 244
 245     s.ll = src1;
 246     d.ll = src2;
 247
 248 #define PMUL(r)                                                         \
 249     tmp = (int32_t)d.VIS_SW64(r) * ((int32_t)s.VIS_SW64(r) >> 8);       \
 250     if ((tmp & 0xff) > 0x7f) {                                          \
 251         tmp += 0x100;                                                   \
 252     }                                                                   \
 253     d.VIS_L64(r) = tmp;
 254
 255     /* Reverse calculation order to handle overlap */
 256     PMUL(1);
 257     PMUL(0);
 258 #undef PMUL
 259
 260     return d.ll;
 261 }
 262
 263 uint64_t helper_fmuld8ulx16(CPUState *env, uint64_t src1, uint64_t src2)
 264 {
 265     VIS64 s, d;
 266     uint32_t tmp;
 267
 268     s.ll = src1;
 269     d.ll = src2;
 270
 271 #define PMUL(r)                                                         \
 272     tmp = (int32_t)d.VIS_SW64(r) * ((uint32_t)s.VIS_B64(r * 2));        \
 273     if ((tmp & 0xff) > 0x7f) {                                          \
 274         tmp += 0x100;                                                   \
 275     }                                                                   \
 276     d.VIS_L64(r) = tmp;
 277
 278     /* Reverse calculation order to handle overlap */
 279     PMUL(1);
 280     PMUL(0);
 281 #undef PMUL
 282
 283     return d.ll;
 284 }
 285
 286 uint64_t helper_fexpand(CPUState *env, uint64_t src1, uint64_t src2)
 287 {
 288     VIS32 s;
 289     VIS64 d;
 290
 291     s.l = (uint32_t)src1;
 292     d.ll = src2;
 293     d.VIS_W64(0) = s.VIS_B32(0) << 4;
 294     d.VIS_W64(1) = s.VIS_B32(1) << 4;
 295     d.VIS_W64(2) = s.VIS_B32(2) << 4;
 296     d.VIS_W64(3) = s.VIS_B32(3) << 4;
 297
 298     return d.ll;
 299 }
 300
 301 #define VIS_HELPER(name, F)                             \
 302     uint64_t name##16(CPUState *env, uint64_t src1, uint64_t src2) \
 303     {                                                   \
 304         VIS64 s, d;                                     \
 305                                                         \
 306         s.ll = src1;                                    \
 307         d.ll = src2;                                    \
 308                                                         \
 309         d.VIS_W64(0) = F(d.VIS_W64(0), s.VIS_W64(0));   \
 310         d.VIS_W64(1) = F(d.VIS_W64(1), s.VIS_W64(1));   \
 311         d.VIS_W64(2) = F(d.VIS_W64(2), s.VIS_W64(2));   \
 312         d.VIS_W64(3) = F(d.VIS_W64(3), s.VIS_W64(3));   \
 313                                                         \
 314         return d.ll;                                    \
 315     }                                                   \
 316                                                         \
 317     uint32_t name##16s(CPUState *env, uint32_t src1,    \
 318                        uint32_t src2)                   \
 319     {                                                   \
 320         VIS32 s, d;                                     \
 321                                                         \
 322         s.l = src1;                                     \
 323         d.l = src2;                                     \
 324                                                         \
 325         d.VIS_W32(0) = F(d.VIS_W32(0), s.VIS_W32(0));   \
 326         d.VIS_W32(1) = F(d.VIS_W32(1), s.VIS_W32(1));   \
 327                                                         \
 328         return d.l;                                     \
 329     }                                                   \
 330                                                         \
 331     uint64_t name##32(CPUState *env, uint64_t src1, uint64_t src2) \
 332     {                                                   \
 333         VIS64 s, d;                                     \
 334                                                         \
 335         s.ll = src1;                                    \
 336         d.ll = src2;                                    \
 337                                                         \
 338         d.VIS_L64(0) = F(d.VIS_L64(0), s.VIS_L64(0));   \
 339         d.VIS_L64(1) = F(d.VIS_L64(1), s.VIS_L64(1));   \
 340                                                         \
 341         return d.ll;                                    \
 342     }                                                   \
 343                                                         \
 344     uint32_t name##32s(CPUState *env, uint32_t src1,    \
 345                        uint32_t src2)                   \
 346     {                                                   \
 347         VIS32 s, d;                                     \
 348                                                         \
 349         s.l = src1;                                     \
 350         d.l = src2;                                     \
 351                                                         \
 352         d.l = F(d.l, s.l);                              \
 353                                                         \
 354         return d.l;                                     \
 355     }
 356
 357 #define FADD(a, b) ((a) + (b))
 358 #define FSUB(a, b) ((a) - (b))
 359 VIS_HELPER(helper_fpadd, FADD)
 360 VIS_HELPER(helper_fpsub, FSUB)
 361
 362 #define VIS_CMPHELPER(name, F)                                    \
 363     uint64_t name##16(CPUState *env, uint64_t src1, uint64_t src2) \
 364     {                                                             \
 365         VIS64 s, d;                                               \
 366                                                                   \
 367         s.ll = src1;                                              \
 368         d.ll = src2;                                              \
 369                                                                   \
 370         d.VIS_W64(0) = F(s.VIS_W64(0), d.VIS_W64(0)) ? 1 : 0;     \
 371         d.VIS_W64(0) |= F(s.VIS_W64(1), d.VIS_W64(1)) ? 2 : 0;    \
 372         d.VIS_W64(0) |= F(s.VIS_W64(2), d.VIS_W64(2)) ? 4 : 0;    \
 373         d.VIS_W64(0) |= F(s.VIS_W64(3), d.VIS_W64(3)) ? 8 : 0;    \
 374         d.VIS_W64(1) = d.VIS_W64(2) = d.VIS_W64(3) = 0;           \
 375                                                                   \
 376         return d.ll;                                              \
 377     }                                                             \
 378                                                                   \
 379     uint64_t name##32(CPUState *env, uint64_t src1, uint64_t src2) \
 380     {                                                             \
 381         VIS64 s, d;                                               \
 382                                                                   \
 383         s.ll = src1;                                              \
 384         d.ll = src2;                                              \
 385                                                                   \
 386         d.VIS_L64(0) = F(s.VIS_L64(0), d.VIS_L64(0)) ? 1 : 0;     \
 387         d.VIS_L64(0) |= F(s.VIS_L64(1), d.VIS_L64(1)) ? 2 : 0;    \
 388         d.VIS_L64(1) = 0;                                         \
 389                                                                   \
 390         return d.ll;                                              \
 391     }
 392
 393 #define FCMPGT(a, b) ((a) > (b))
 394 #define FCMPEQ(a, b) ((a) == (b))
 395 #define FCMPLE(a, b) ((a) <= (b))
 396 #define FCMPNE(a, b) ((a) != (b))
 397
 398 VIS_CMPHELPER(helper_fcmpgt, FCMPGT)
 399 VIS_CMPHELPER(helper_fcmpeq, FCMPEQ)
 400 VIS_CMPHELPER(helper_fcmple, FCMPLE)
 401 VIS_CMPHELPER(helper_fcmpne, FCMPNE)