1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/security/nss/lib/freebl/mpi/vis_64.il Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,997 @@
1.4 +!
1.5 +! This Source Code Form is subject to the terms of the Mozilla Public
1.6 +! License, v. 2.0. If a copy of the MPL was not distributed with this
1.7 +! file, You can obtain one at http://mozilla.org/MPL/2.0/.
1.8 +
1.9 +! This file is to be used in place of vis.il in 64-bit builds.
1.10 +
1.11 +!--------------------------------------------------------------------
1.12 +! Pure edge handling instructions
1.13 +!
1.14 +! int vis_edge8(void */*frs1*/, void */*frs2*/);
1.15 +!
1.16 + .inline vis_edge8,16
1.17 + edge8 %o0,%o1,%o0
1.18 + .end
1.19 +!
1.20 +! int vis_edge8l(void */*frs1*/, void */*frs2*/);
1.21 +!
1.22 + .inline vis_edge8l,16
1.23 + edge8l %o0,%o1,%o0
1.24 + .end
1.25 +!
1.26 +! int vis_edge16(void */*frs1*/, void */*frs2*/);
1.27 +!
1.28 + .inline vis_edge16,16
1.29 + edge16 %o0,%o1,%o0
1.30 + .end
1.31 +!
1.32 +! int vis_edge16l(void */*frs1*/, void */*frs2*/);
1.33 +!
1.34 + .inline vis_edge16l,16
1.35 + edge16l %o0,%o1,%o0
1.36 + .end
1.37 +!
1.38 +! int vis_edge32(void */*frs1*/, void */*frs2*/);
1.39 +!
1.40 + .inline vis_edge32,16
1.41 + edge32 %o0,%o1,%o0
1.42 + .end
1.43 +!
1.44 +! int vis_edge32l(void */*frs1*/, void */*frs2*/);
1.45 +!
1.46 + .inline vis_edge32l,16
1.47 + edge32l %o0,%o1,%o0
1.48 + .end
1.49 +
1.50 +!--------------------------------------------------------------------
1.51 +! Edge handling instructions with negative return values if cc set
1.52 +!
1.53 +! int vis_edge8cc(void */*frs1*/, void */*frs2*/);
1.54 +!
1.55 + .inline vis_edge8cc,16
1.56 + edge8 %o0,%o1,%o0
1.57 + mov 0,%o1
1.58 + movgu %xcc,-1024,%o1
1.59 + or %o1,%o0,%o0
1.60 + .end
1.61 +!
1.62 +! int vis_edge8lcc(void */*frs1*/, void */*frs2*/);
1.63 +!
1.64 + .inline vis_edge8lcc,16
1.65 + edge8l %o0,%o1,%o0
1.66 + mov 0,%o1
1.67 + movgu %xcc,-1024,%o1
1.68 + or %o1,%o0,%o0
1.69 + .end
1.70 +!
1.71 +! int vis_edge16cc(void */*frs1*/, void */*frs2*/);
1.72 +!
1.73 + .inline vis_edge16cc,16
1.74 + edge16 %o0,%o1,%o0
1.75 + mov 0,%o1
1.76 + movgu %xcc,-1024,%o1
1.77 + or %o1,%o0,%o0
1.78 + .end
1.79 +!
1.80 +! int vis_edge16lcc(void */*frs1*/, void */*frs2*/);
1.81 +!
1.82 + .inline vis_edge16lcc,16
1.83 + edge16l %o0,%o1,%o0
1.84 + mov 0,%o1
1.85 + movgu %xcc,-1024,%o1
1.86 + or %o1,%o0,%o0
1.87 + .end
1.88 +!
1.89 +! int vis_edge32cc(void */*frs1*/, void */*frs2*/);
1.90 +!
1.91 + .inline vis_edge32cc,16
1.92 + edge32 %o0,%o1,%o0
1.93 + mov 0,%o1
1.94 + movgu %xcc,-1024,%o1
1.95 + or %o1,%o0,%o0
1.96 + .end
1.97 +!
1.98 +! int vis_edge32lcc(void */*frs1*/, void */*frs2*/);
1.99 +!
1.100 + .inline vis_edge32lcc,16
1.101 + edge32l %o0,%o1,%o0
1.102 + mov 0,%o1
1.103 + movgu %xcc,-1024,%o1
1.104 + or %o1,%o0,%o0
1.105 + .end
1.106 +
1.107 +!--------------------------------------------------------------------
1.108 +! Alignment instructions
1.109 +!
1.110 +! void *vis_alignaddr(void */*rs1*/, int /*rs2*/);
1.111 +!
1.112 + .inline vis_alignaddr,12
1.113 + alignaddr %o0,%o1,%o0
1.114 + .end
1.115 +!
1.116 +! void *vis_alignaddrl(void */*rs1*/, int /*rs2*/);
1.117 +!
1.118 + .inline vis_alignaddrl,12
1.119 + alignaddrl %o0,%o1,%o0
1.120 + .end
1.121 +!
1.122 +! double vis_faligndata(double /*frs1*/, double /*frs2*/);
1.123 +!
1.124 + .inline vis_faligndata,16
1.125 + faligndata %f0,%f2,%f0
1.126 + .end
1.127 +
1.128 +!--------------------------------------------------------------------
1.129 +! Partitioned comparison instructions
1.130 +!
1.131 +! int vis_fcmple16(double /*frs1*/, double /*frs2*/);
1.132 +!
1.133 + .inline vis_fcmple16,16
1.134 + fcmple16 %f0,%f2,%o0
1.135 + .end
1.136 +!
1.137 +! int vis_fcmpne16(double /*frs1*/, double /*frs2*/);
1.138 +!
1.139 + .inline vis_fcmpne16,16
1.140 + fcmpne16 %f0,%f2,%o0
1.141 + .end
1.142 +!
1.143 +! int vis_fcmple32(double /*frs1*/, double /*frs2*/);
1.144 +!
1.145 + .inline vis_fcmple32,16
1.146 + fcmple32 %f0,%f2,%o0
1.147 + .end
1.148 +!
1.149 +! int vis_fcmpne32(double /*frs1*/, double /*frs2*/);
1.150 +!
1.151 + .inline vis_fcmpne32,16
1.152 + fcmpne32 %f0,%f2,%o0
1.153 + .end
1.154 +!
1.155 +! int vis_fcmpgt16(double /*frs1*/, double /*frs2*/);
1.156 +!
1.157 + .inline vis_fcmpgt16,16
1.158 + fcmpgt16 %f0,%f2,%o0
1.159 + .end
1.160 +!
1.161 +! int vis_fcmpeq16(double /*frs1*/, double /*frs2*/);
1.162 +!
1.163 + .inline vis_fcmpeq16,16
1.164 + fcmpeq16 %f0,%f2,%o0
1.165 + .end
1.166 +!
1.167 +! int vis_fcmpgt32(double /*frs1*/, double /*frs2*/);
1.168 +!
1.169 + .inline vis_fcmpgt32,16
1.170 + fcmpgt32 %f0,%f2,%o0
1.171 + .end
1.172 +!
1.173 +! int vis_fcmpeq32(double /*frs1*/, double /*frs2*/);
1.174 +!
1.175 + .inline vis_fcmpeq32,16
1.176 + fcmpeq32 %f0,%f2,%o0
1.177 + .end
1.178 +
1.179 +!--------------------------------------------------------------------
1.180 +! Partitioned arithmetic
1.181 +!
1.182 +! double vis_fmul8x16(float /*frs1*/, double /*frs2*/);
1.183 +!
1.184 + .inline vis_fmul8x16,12
1.185 + fmul8x16 %f1,%f2,%f0
1.186 + .end
1.187 +!
1.188 +! double vis_fmul8x16_dummy(float /*frs1*/, int /*dummy*/, double /*frs2*/);
1.189 +!
1.190 + .inline vis_fmul8x16_dummy,16
1.191 + fmul8x16 %f1,%f4,%f0
1.192 + .end
1.193 +!
1.194 +! double vis_fmul8x16au(float /*frs1*/, float /*frs2*/);
1.195 +!
1.196 + .inline vis_fmul8x16au,8
1.197 + fmul8x16au %f1,%f3,%f0
1.198 + .end
1.199 +!
1.200 +! double vis_fmul8x16al(float /*frs1*/, float /*frs2*/);
1.201 +!
1.202 + .inline vis_fmul8x16al,8
1.203 + fmul8x16al %f1,%f3,%f0
1.204 + .end
1.205 +!
1.206 +! double vis_fmul8sux16(double /*frs1*/, double /*frs2*/);
1.207 +!
1.208 + .inline vis_fmul8sux16,16
1.209 + fmul8sux16 %f0,%f2,%f0
1.210 + .end
1.211 +!
1.212 +! double vis_fmul8ulx16(double /*frs1*/, double /*frs2*/);
1.213 +!
1.214 + .inline vis_fmul8ulx16,16
1.215 + fmul8ulx16 %f0,%f2,%f0
1.216 + .end
1.217 +!
1.218 +! double vis_fmuld8sux16(float /*frs1*/, float /*frs2*/);
1.219 +!
1.220 + .inline vis_fmuld8sux16,8
1.221 + fmuld8sux16 %f1,%f3,%f0
1.222 + .end
1.223 +!
1.224 +! double vis_fmuld8ulx16(float /*frs1*/, float /*frs2*/);
1.225 +!
1.226 + .inline vis_fmuld8ulx16,8
1.227 + fmuld8ulx16 %f1,%f3,%f0
1.228 + .end
1.229 +!
1.230 +! double vis_fpadd16(double /*frs1*/, double /*frs2*/);
1.231 +!
1.232 + .inline vis_fpadd16,16
1.233 + fpadd16 %f0,%f2,%f0
1.234 + .end
1.235 +!
1.236 +! float vis_fpadd16s(float /*frs1*/, float /*frs2*/);
1.237 +!
1.238 + .inline vis_fpadd16s,8
1.239 + fpadd16s %f1,%f3,%f0
1.240 + .end
1.241 +!
1.242 +! double vis_fpadd32(double /*frs1*/, double /*frs2*/);
1.243 +!
1.244 + .inline vis_fpadd32,16
1.245 + fpadd32 %f0,%f2,%f0
1.246 + .end
1.247 +!
1.248 +! float vis_fpadd32s(float /*frs1*/, float /*frs2*/);
1.249 +!
1.250 + .inline vis_fpadd32s,8
1.251 + fpadd32s %f1,%f3,%f0
1.252 + .end
1.253 +!
1.254 +! double vis_fpsub16(double /*frs1*/, double /*frs2*/);
1.255 +!
1.256 + .inline vis_fpsub16,16
1.257 + fpsub16 %f0,%f2,%f0
1.258 + .end
1.259 +!
1.260 +! float vis_fpsub16s(float /*frs1*/, float /*frs2*/);
1.261 +!
1.262 + .inline vis_fpsub16s,8
1.263 + fpsub16s %f1,%f3,%f0
1.264 + .end
1.265 +!
1.266 +! double vis_fpsub32(double /*frs1*/, double /*frs2*/);
1.267 +!
1.268 + .inline vis_fpsub32,16
1.269 + fpsub32 %f0,%f2,%f0
1.270 + .end
1.271 +!
1.272 +! float vis_fpsub32s(float /*frs1*/, float /*frs2*/);
1.273 +!
1.274 + .inline vis_fpsub32s,8
1.275 + fpsub32s %f1,%f3,%f0
1.276 + .end
1.277 +
1.278 +!--------------------------------------------------------------------
1.279 +! Pixel packing
1.280 +!
1.281 +! float vis_fpack16(double /*frs2*/);
1.282 +!
1.283 + .inline vis_fpack16,8
1.284 + fpack16 %f0,%f0
1.285 + .end
1.286 +!
1.287 +! double vis_fpack16_pair(double /*frs2*/, double /*frs2*/);
1.288 +!
1.289 + .inline vis_fpack16_pair,16
1.290 + fpack16 %f0,%f0
1.291 + fpack16 %f2,%f1
1.292 + .end
1.293 +!
1.294 +! void vis_st2_fpack16(double, double, double *)
1.295 +!
1.296 + .inline vis_st2_fpack16,24
1.297 + fpack16 %f0,%f0
1.298 + fpack16 %f2,%f1
1.299 + st %f0,[%o2+0]
1.300 + st %f1,[%o2+4]
1.301 + .end
1.302 +!
1.303 +! void vis_std_fpack16(double, double, double *)
1.304 +!
1.305 + .inline vis_std_fpack16,24
1.306 + fpack16 %f0,%f0
1.307 + fpack16 %f2,%f1
1.308 + std %f0,[%o2]
1.309 + .end
1.310 +!
1.311 +! void vis_st2_fpackfix(double, double, double *)
1.312 +!
1.313 + .inline vis_st2_fpackfix,24
1.314 + fpackfix %f0,%f0
1.315 + fpackfix %f2,%f1
1.316 + st %f0,[%o2+0]
1.317 + st %f1,[%o2+4]
1.318 + .end
1.319 +!
1.320 +! double vis_fpack16_to_hi(double /*frs1*/, double /*frs2*/);
1.321 +!
1.322 + .inline vis_fpack16_to_hi,16
1.323 + fpack16 %f2,%f0
1.324 + .end
1.325 +
1.326 +! double vis_fpack16_to_lo(double /*frs1*/, double /*frs2*/);
1.327 +!
1.328 + .inline vis_fpack16_to_lo,16
1.329 + fpack16 %f2,%f3
1.330 + fmovs %f3,%f1 /* without this, optimizer goes wrong */
1.331 + .end
1.332 +
1.333 +!
1.334 +! double vis_fpack32(double /*frs1*/, double /*frs2*/);
1.335 +!
1.336 + .inline vis_fpack32,16
1.337 + fpack32 %f0,%f2,%f0
1.338 + .end
1.339 +!
1.340 +! float vis_fpackfix(double /*frs2*/);
1.341 +!
1.342 + .inline vis_fpackfix,8
1.343 + fpackfix %f0,%f0
1.344 + .end
1.345 +!
1.346 +! double vis_fpackfix_pair(double /*frs2*/, double /*frs2*/);
1.347 +!
1.348 + .inline vis_fpackfix_pair,16
1.349 + fpackfix %f0,%f0
1.350 + fpackfix %f2,%f1
1.351 + .end
1.352 +
1.353 +!--------------------------------------------------------------------
1.354 +! Motion estimation
1.355 +!
1.356 +! double vis_pxldist64(double accum /*frd*/, double pxls1 /*frs1*/,
1.357 +! double pxls2 /*frs2*/);
1.358 +!
1.359 + .inline vis_pxldist64,24
1.360 + pdist %f2,%f4,%f0
1.361 + .end
1.362 +
1.363 +!--------------------------------------------------------------------
1.364 +! Channel merging
1.365 +!
1.366 +! double vis_fpmerge(float /*frs1*/, float /*frs2*/);
1.367 +!
1.368 + .inline vis_fpmerge,8
1.369 + fpmerge %f1,%f3,%f0
1.370 + .end
1.371 +
1.372 +!--------------------------------------------------------------------
1.373 +! Pixel expansion
1.374 +!
1.375 +! double vis_fexpand(float /*frs2*/);
1.376 +!
1.377 + .inline vis_fexpand,4
1.378 + fexpand %f1,%f0
1.379 + .end
1.380 +
1.381 +! double vis_fexpand_hi(double /*frs2*/);
1.382 +!
1.383 + .inline vis_fexpand_hi,8
1.384 + fexpand %f0,%f0
1.385 + .end
1.386 +
1.387 +! double vis_fexpand_lo(double /*frs2*/);
1.388 +!
1.389 + .inline vis_fexpand_lo,8
1.390 + fexpand %f1,%f0
1.391 + .end
1.392 +
1.393 +!--------------------------------------------------------------------
1.394 +! Bitwise logical operations
1.395 +!
1.396 +! double vis_fnor(double /*frs1*/, double /*frs2*/);
1.397 +!
1.398 + .inline vis_fnor,16
1.399 + fnor %f0,%f2,%f0
1.400 + .end
1.401 +!
1.402 +! float vis_fnors(float /*frs1*/, float /*frs2*/);
1.403 +!
1.404 + .inline vis_fnors,8
1.405 + fnors %f1,%f3,%f0
1.406 + .end
1.407 +!
1.408 +! double vis_fandnot(double /*frs1*/, double /*frs2*/);
1.409 +!
1.410 + .inline vis_fandnot,16
1.411 + fandnot1 %f0,%f2,%f0
1.412 + .end
1.413 +!
1.414 +! float vis_fandnots(float /*frs1*/, float /*frs2*/);
1.415 +!
1.416 + .inline vis_fandnots,8
1.417 + fandnot1s %f1,%f3,%f0
1.418 + .end
1.419 +!
1.420 +! double vis_fnot(double /*frs1*/);
1.421 +!
1.422 + .inline vis_fnot,8
1.423 + fnot1 %f0,%f0
1.424 + .end
1.425 +!
1.426 +! float vis_fnots(float /*frs1*/);
1.427 +!
1.428 + .inline vis_fnots,4
1.429 + fnot1s %f1,%f0
1.430 + .end
1.431 +!
1.432 +! double vis_fxor(double /*frs1*/, double /*frs2*/);
1.433 +!
1.434 + .inline vis_fxor,16
1.435 + fxor %f0,%f2,%f0
1.436 + .end
1.437 +!
1.438 +! float vis_fxors(float /*frs1*/, float /*frs2*/);
1.439 +!
1.440 + .inline vis_fxors,8
1.441 + fxors %f1,%f3,%f0
1.442 + .end
1.443 +!
1.444 +! double vis_fnand(double /*frs1*/, double /*frs2*/);
1.445 +!
1.446 + .inline vis_fnand,16
1.447 + fnand %f0,%f2,%f0
1.448 + .end
1.449 +!
1.450 +! float vis_fnands(float /*frs1*/, float /*frs2*/);
1.451 +!
1.452 + .inline vis_fnands,8
1.453 + fnands %f1,%f3,%f0
1.454 + .end
1.455 +!
1.456 +! double vis_fand(double /*frs1*/, double /*frs2*/);
1.457 +!
1.458 + .inline vis_fand,16
1.459 + fand %f0,%f2,%f0
1.460 + .end
1.461 +!
1.462 +! float vis_fands(float /*frs1*/, float /*frs2*/);
1.463 +!
1.464 + .inline vis_fands,8
1.465 + fands %f1,%f3,%f0
1.466 + .end
1.467 +!
1.468 +! double vis_fxnor(double /*frs1*/, double /*frs2*/);
1.469 +!
1.470 + .inline vis_fxnor,16
1.471 + fxnor %f0,%f2,%f0
1.472 + .end
1.473 +!
1.474 +! float vis_fxnors(float /*frs1*/, float /*frs2*/);
1.475 +!
1.476 + .inline vis_fxnors,8
1.477 + fxnors %f1,%f3,%f0
1.478 + .end
1.479 +!
1.480 +! double vis_fsrc(double /*frs1*/);
1.481 +!
1.482 + .inline vis_fsrc,8
1.483 + fsrc1 %f0,%f0
1.484 + .end
1.485 +!
1.486 +! float vis_fsrcs(float /*frs1*/);
1.487 +!
1.488 + .inline vis_fsrcs,4
1.489 + fsrc1s %f1,%f0
1.490 + .end
1.491 +!
1.492 +! double vis_fornot(double /*frs1*/, double /*frs2*/);
1.493 +!
1.494 + .inline vis_fornot,16
1.495 + fornot1 %f0,%f2,%f0
1.496 + .end
1.497 +!
1.498 +! float vis_fornots(float /*frs1*/, float /*frs2*/);
1.499 +!
1.500 + .inline vis_fornots,8
1.501 + fornot1s %f1,%f3,%f0
1.502 + .end
1.503 +!
1.504 +! double vis_for(double /*frs1*/, double /*frs2*/);
1.505 +!
1.506 + .inline vis_for,16
1.507 + for %f0,%f2,%f0
1.508 + .end
1.509 +!
1.510 +! float vis_fors(float /*frs1*/, float /*frs2*/);
1.511 +!
1.512 + .inline vis_fors,8
1.513 + fors %f1,%f3,%f0
1.514 + .end
1.515 +!
1.516 +! double vis_fzero(/* void */)
1.517 +!
1.518 + .inline vis_fzero,0
1.519 + fzero %f0
1.520 + .end
1.521 +!
1.522 +! float vis_fzeros(/* void */)
1.523 +!
1.524 + .inline vis_fzeros,0
1.525 + fzeros %f0
1.526 + .end
1.527 +!
1.528 +! double vis_fone(/* void */)
1.529 +!
1.530 + .inline vis_fone,0
1.531 + fone %f0
1.532 + .end
1.533 +!
1.534 +! float vis_fones(/* void */)
1.535 +!
1.536 + .inline vis_fones,0
1.537 + fones %f0
1.538 + .end
1.539 +
1.540 +!--------------------------------------------------------------------
1.541 +! Partial store instructions
1.542 +!
1.543 +! vis_stdfa_ASI_PST8P(double frd, void *rs1, int rmask)
1.544 +!
1.545 + .inline vis_stdfa_ASI_PST8P,20
1.546 + stda %f0,[%o1]%o2,0xc0 ! ASI_PST8_P
1.547 + .end
1.548 +!
1.549 +! vis_stdfa_ASI_PST8PL(double frd, void *rs1, int rmask)
1.550 +!
1.551 + .inline vis_stdfa_ASI_PST8PL,20
1.552 + stda %f0,[%o1]%o2,0xc8 ! ASI_PST8_PL
1.553 + .end
1.554 +!
1.555 +! vis_stdfa_ASI_PST8P_int_pair(void *rs1, void *rs2, void *rs3, int rmask);
1.556 +!
1.557 + .inline vis_stdfa_ASI_PST8P_int_pair,28
1.558 + ld [%o0],%f4
1.559 + ld [%o1],%f5
1.560 + stda %f4,[%o2]%o3,0xc0 ! ASI_PST8_P
1.561 + .end
1.562 +!
1.563 +! vis_stdfa_ASI_PST8S(double frd, void *rs1, int rmask)
1.564 +!
1.565 + .inline vis_stdfa_ASI_PST8S,20
1.566 + stda %f0,[%o1]%o2,0xc1 ! ASI_PST8_S
1.567 + .end
1.568 +!
1.569 +! vis_stdfa_ASI_PST16P(double frd, void *rs1, int rmask)
1.570 +!
1.571 + .inline vis_stdfa_ASI_PST16P,20
1.572 + stda %f0,[%o1]%o2,0xc2 ! ASI_PST16_P
1.573 + .end
1.574 +!
1.575 +! vis_stdfa_ASI_PST16S(double frd, void *rs1, int rmask)
1.576 +!
1.577 + .inline vis_stdfa_ASI_PST16S,20
1.578 + stda %f0,[%o1]%o2,0xc3 ! ASI_PST16_S
1.579 + .end
1.580 +!
1.581 +! vis_stdfa_ASI_PST32P(double frd, void *rs1, int rmask)
1.582 +!
1.583 + .inline vis_stdfa_ASI_PST32P,20
1.584 + stda %f0,[%o1]%o2,0xc4 ! ASI_PST32_P
1.585 + .end
1.586 +!
1.587 +! vis_stdfa_ASI_PST32S(double frd, void *rs1, int rmask)
1.588 +!
1.589 + .inline vis_stdfa_ASI_PST32S,20
1.590 + stda %f0,[%o1]%o2,0xc5 ! ASI_PST32_S
1.591 + .end
1.592 +
1.593 +!--------------------------------------------------------------------
1.594 +! Short store instructions
1.595 +!
1.596 +! vis_stdfa_ASI_FL8P(double frd, void *rs1)
1.597 +!
1.598 + .inline vis_stdfa_ASI_FL8P,16
1.599 + stda %f0,[%o1]0xd0 ! ASI_FL8_P
1.600 + .end
1.601 +!
1.602 +! vis_stdfa_ASI_FL8P_index(double frd, void *rs1, long index)
1.603 +!
1.604 + .inline vis_stdfa_ASI_FL8P_index,24
1.605 + stda %f0,[%o1+%o2]0xd0 ! ASI_FL8_P
1.606 + .end
1.607 +!
1.608 +! vis_stdfa_ASI_FL8S(double frd, void *rs1)
1.609 +!
1.610 + .inline vis_stdfa_ASI_FL8S,16
1.611 + stda %f0,[%o1]0xd1 ! ASI_FL8_S
1.612 + .end
1.613 +!
1.614 +! vis_stdfa_ASI_FL16P(double frd, void *rs1)
1.615 +!
1.616 + .inline vis_stdfa_ASI_FL16P,16
1.617 + stda %f0,[%o1]0xd2 ! ASI_FL16_P
1.618 + .end
1.619 +!
1.620 +! vis_stdfa_ASI_FL16P_index(double frd, void *rs1, long index)
1.621 +!
1.622 + .inline vis_stdfa_ASI_FL16P_index,24
1.623 + stda %f0,[%o1+%o2]0xd2 ! ASI_FL16_P
1.624 + .end
1.625 +!
1.626 +! vis_stdfa_ASI_FL16S(double frd, void *rs1)
1.627 +!
1.628 + .inline vis_stdfa_ASI_FL16S,16
1.629 + stda %f0,[%o1]0xd3 ! ASI_FL16_S
1.630 + .end
1.631 +!
1.632 +! vis_stdfa_ASI_FL8PL(double frd, void *rs1)
1.633 +!
1.634 + .inline vis_stdfa_ASI_FL8PL,16
1.635 + stda %f0,[%o1]0xd8 ! ASI_FL8_PL
1.636 + .end
1.637 +!
1.638 +! vis_stdfa_ASI_FL8SL(double frd, void *rs1)
1.639 +!
1.640 + .inline vis_stdfa_ASI_FL8SL,16
1.641 + stda %f0,[%o1]0xd9 ! ASI_FL8_SL
1.642 + .end
1.643 +!
1.644 +! vis_stdfa_ASI_FL16PL(double frd, void *rs1)
1.645 +!
1.646 + .inline vis_stdfa_ASI_FL16PL,16
1.647 + stda %f0,[%o1]0xda ! ASI_FL16_PL
1.648 + .end
1.649 +!
1.650 +! vis_stdfa_ASI_FL16SL(double frd, void *rs1)
1.651 +!
1.652 + .inline vis_stdfa_ASI_FL16SL,16
1.653 + stda %f0,[%o1]0xdb ! ASI_FL16_SL
1.654 + .end
1.655 +
1.656 +!--------------------------------------------------------------------
1.657 +! Short load instructions
1.658 +!
1.659 +! double vis_lddfa_ASI_FL8P(void *rs1)
1.660 +!
1.661 + .inline vis_lddfa_ASI_FL8P,8
1.662 + ldda [%o0]0xd0,%f4 ! ASI_FL8_P
1.663 + fmovd %f4,%f0 ! Compiler can clean this up
1.664 + .end
1.665 +!
1.666 +! double vis_lddfa_ASI_FL8P_index(void *rs1, long index)
1.667 +!
1.668 + .inline vis_lddfa_ASI_FL8P_index,16
1.669 + ldda [%o0+%o1]0xd0,%f4
1.670 + fmovd %f4,%f0
1.671 + .end
1.672 +!
1.673 +! double vis_lddfa_ASI_FL8P_hi(void *rs1, unsigned int index)
1.674 +!
1.675 + .inline vis_lddfa_ASI_FL8P_hi,12
1.676 + sra %o1,16,%o1
1.677 + ldda [%o0+%o1]0xd0,%f4
1.678 + fmovd %f4,%f0
1.679 + .end
1.680 +!
1.681 +! double vis_lddfa_ASI_FL8P_lo(void *rs1, unsigned int index)
1.682 +!
1.683 + .inline vis_lddfa_ASI_FL8P_lo,12
1.684 + sll %o1,16,%o1
1.685 + sra %o1,16,%o1
1.686 + ldda [%o0+%o1]0xd0,%f4
1.687 + fmovd %f4,%f0
1.688 + .end
1.689 +!
1.690 +! double vis_lddfa_ASI_FL8S(void *rs1)
1.691 +!
1.692 + .inline vis_lddfa_ASI_FL8S,8
1.693 + ldda [%o0]0xd1,%f4 ! ASI_FL8_S
1.694 + fmovd %f4,%f0
1.695 + .end
1.696 +!
1.697 +! double vis_lddfa_ASI_FL16P(void *rs1)
1.698 +!
1.699 + .inline vis_lddfa_ASI_FL16P,8
1.700 + ldda [%o0]0xd2,%f4 ! ASI_FL16_P
1.701 + fmovd %f4,%f0
1.702 + .end
1.703 +!
1.704 +! double vis_lddfa_ASI_FL16P_index(void *rs1, long index)
1.705 +!
1.706 + .inline vis_lddfa_ASI_FL16P_index,16
1.707 + ldda [%o0+%o1]0xd2,%f4 ! ASI_FL16_P
1.708 + fmovd %f4,%f0
1.709 + .end
1.710 +!
1.711 +! double vis_lddfa_ASI_FL16S(void *rs1)
1.712 +!
1.713 + .inline vis_lddfa_ASI_FL16S,8
1.714 + ldda [%o0]0xd3,%f4 ! ASI_FL16_S
1.715 + fmovd %f4,%f0
1.716 + .end
1.717 +!
1.718 +! double vis_lddfa_ASI_FL8PL(void *rs1)
1.719 +!
1.720 + .inline vis_lddfa_ASI_FL8PL,8
1.721 + ldda [%o0]0xd8,%f4 ! ASI_FL8_PL
1.722 + fmovd %f4,%f0
1.723 + .end
1.724 +!
1.725 +! double vis_lddfa_ASI_FL8PL_index(void *rs1, long index)
1.726 +!
1.727 + .inline vis_lddfa_ASI_FL8PL_index,16
1.728 + ldda [%o0+%o1]0xd8,%f4 ! ASI_FL8_PL
1.729 + fmovd %f4,%f0
1.730 + .end
1.731 +!
1.732 +! double vis_lddfa_ASI_FL8SL(void *rs1)
1.733 +!
1.734 + .inline vis_lddfa_ASI_FL8SL,8
1.735 + ldda [%o0]0xd9,%f4 ! ASI_FL8_SL
1.736 + fmovd %f4,%f0
1.737 + .end
1.738 +!
1.739 +! double vis_lddfa_ASI_FL16PL(void *rs1)
1.740 +!
1.741 + .inline vis_lddfa_ASI_FL16PL,8
1.742 + ldda [%o0]0xda,%f4 ! ASI_FL16_PL
1.743 + fmovd %f4,%f0
1.744 + .end
1.745 +!
1.746 +! double vis_lddfa_ASI_FL16PL_index(void *rs1, long index)
1.747 +!
1.748 + .inline vis_lddfa_ASI_FL16PL_index,16
1.749 + ldda [%o0+%o1]0xda,%f4 ! ASI_FL16_PL
1.750 + fmovd %f4,%f0
1.751 + .end
1.752 +!
1.753 +! double vis_lddfa_ASI_FL16SL(void *rs1)
1.754 +!
1.755 + .inline vis_lddfa_ASI_FL16SL,8
1.756 + ldda [%o0]0xdb,%f4 ! ASI_FL16_SL
1.757 + fmovd %f4,%f0
1.758 + .end
1.759 +
1.760 +!--------------------------------------------------------------------
1.761 +! Graphics status register
1.762 +!
1.763 +! unsigned int vis_read_gsr(void)
1.764 +!
1.765 + .inline vis_read_gsr,0
1.766 + rd %gsr,%o0
1.767 + .end
1.768 +!
1.769 +! void vis_write_gsr(unsigned int /* GSR */)
1.770 +!
1.771 + .inline vis_write_gsr,4
1.772 + wr %g0,%o0,%gsr
1.773 + .end
1.774 +
1.775 +!--------------------------------------------------------------------
1.776 +! Voxel texture mapping
1.777 +!
1.778 +! unsigned long vis_array8(unsigned long long /*rs1 */, int /*rs2*/)
1.779 +!
1.780 + .inline vis_array8,12
1.781 + array8 %o0,%o1,%o0
1.782 + .end
1.783 +!
1.784 +! unsigned long vis_array16(unsigned long long /*rs1*/, int /*rs2*/)
1.785 +!
1.786 + .inline vis_array16,12
1.787 + array16 %o0,%o1,%o0
1.788 + .end
1.789 +!
1.790 +! unsigned long vis_array32(unsigned long long /*rs1*/, int /*rs2*/)
1.791 +!
1.792 + .inline vis_array32,12
1.793 + array32 %o0,%o1,%o0
1.794 + .end
1.795 +
1.796 +!--------------------------------------------------------------------
1.797 +! Register aliasing and type casts
1.798 +!
1.799 +! float vis_read_hi(double /* frs1 */);
1.800 +!
1.801 + .inline vis_read_hi,8
1.802 + fmovs %f0,%f0
1.803 + .end
1.804 +!
1.805 +! float vis_read_lo(double /* frs1 */);
1.806 +!
1.807 + .inline vis_read_lo,8
1.808 + fmovs %f1,%f0 ! %f0 = low word (frs1); return %f0;
1.809 + .end
1.810 +!
1.811 +! double vis_write_hi(double /* frs1 */, float /* frs2 */);
1.812 +!
1.813 + .inline vis_write_hi,12
1.814 + fmovs %f3,%f0 ! %f3 = float frs2; return %f0:f1;
1.815 + .end
1.816 +!
1.817 +! double vis_write_lo(double /* frs1 */, float /* frs2 */);
1.818 +!
1.819 + .inline vis_write_lo,12
1.820 + fmovs %f3,%f1 ! %f3 = float frs2; return %f0:f1;
1.821 + .end
1.822 +!
1.823 +! double vis_freg_pair(float /* frs1 */, float /* frs2 */);
1.824 +!
1.825 + .inline vis_freg_pair,8
1.826 + fmovs %f1,%f0 ! %f1 = float frs1; put in hi;
1.827 + fmovs %f3,%f1 ! %f3 = float frs2; put in lo; return %f0:f1;
1.828 + .end
1.829 +!
1.830 +! float vis_to_float(unsigned int /*value*/);
1.831 +!
1.832 + .inline vis_to_float,4
1.833 + st %o0,[%sp+2183]
1.834 + ld [%sp+2183],%f0
1.835 + .end
1.836 +!
1.837 +! double vis_to_double(unsigned int /*value1*/, unsigned int /*value2*/);
1.838 +!
1.839 + .inline vis_to_double,8
1.840 + st %o0,[%sp+2183]
1.841 + ld [%sp+2183],%f0
1.842 + st %o1,[%sp+2183]
1.843 + ld [%sp+2183],%f1
1.844 + .end
1.845 +!
1.846 +! double vis_to_double_dup(unsigned int /*value*/);
1.847 +!
1.848 + .inline vis_to_double_dup,4
1.849 + st %o0,[%sp+2183]
1.850 + ld [%sp+2183],%f1
1.851 + fmovs %f1,%f0 ! duplicate value
1.852 + .end
1.853 +!
1.854 +! double vis_ll_to_double(unsigned long long /*value*/);
1.855 +!
1.856 + .inline vis_ll_to_double,8
1.857 + stx %o0,[%sp+2183]
1.858 + ldd [%sp+2183],%f0
1.859 + .end
1.860 +
1.861 +!--------------------------------------------------------------------
1.862 +! Address space identifier (ASI) register
1.863 +!
1.864 +! unsigned int vis_read_asi(void)
1.865 +!
1.866 + .inline vis_read_asi,0
1.867 + rd %asi,%o0
1.868 + .end
1.869 +!
1.870 +! void vis_write_asi(unsigned int /* ASI */)
1.871 +!
1.872 + .inline vis_write_asi,4
1.873 + wr %g0,%o0,%asi
1.874 + .end
1.875 +
1.876 +!--------------------------------------------------------------------
1.877 +! Load/store from/into alternate space
1.878 +!
1.879 +! float vis_ldfa_ASI_REG(void *rs1)
1.880 +!
1.881 + .inline vis_ldfa_ASI_REG,8
1.882 + lda [%o0+0]%asi,%f4
1.883 + fmovs %f4,%f0 ! Compiler can clean this up
1.884 + .end
1.885 +!
1.886 +! float vis_ldfa_ASI_P(void *rs1)
1.887 +!
1.888 + .inline vis_ldfa_ASI_P,8
1.889 + lda [%o0]0x80,%f4 ! ASI_P
1.890 + fmovs %f4,%f0 ! Compiler can clean this up
1.891 + .end
1.892 +!
1.893 +! float vis_ldfa_ASI_PL(void *rs1)
1.894 +!
1.895 + .inline vis_ldfa_ASI_PL,8
1.896 + lda [%o0]0x88,%f4 ! ASI_PL
1.897 + fmovs %f4,%f0 ! Compiler can clean this up
1.898 + .end
1.899 +!
1.900 +! double vis_lddfa_ASI_REG(void *rs1)
1.901 +!
1.902 + .inline vis_lddfa_ASI_REG,8
1.903 + ldda [%o0+0]%asi,%f4
1.904 + fmovd %f4,%f0 ! Compiler can clean this up
1.905 + .end
1.906 +!
1.907 +! double vis_lddfa_ASI_P(void *rs1)
1.908 +!
1.909 + .inline vis_lddfa_ASI_P,8
1.910 + ldda [%o0]0x80,%f4 ! ASI_P
1.911 + fmovd %f4,%f0 ! Compiler can clean this up
1.912 + .end
1.913 +!
1.914 +! double vis_lddfa_ASI_PL(void *rs1)
1.915 +!
1.916 + .inline vis_lddfa_ASI_PL,8
1.917 + ldda [%o0]0x88,%f4 ! ASI_PL
1.918 + fmovd %f4,%f0 ! Compiler can clean this up
1.919 + .end
1.920 +!
1.921 +! vis_stfa_ASI_REG(float frs, void *rs1)
1.922 +!
1.923 + .inline vis_stfa_ASI_REG,12
1.924 + sta %f1,[%o1+0]%asi
1.925 + .end
1.926 +!
1.927 +! vis_stfa_ASI_P(float frs, void *rs1)
1.928 +!
1.929 + .inline vis_stfa_ASI_P,12
1.930 + sta %f1,[%o1]0x80 ! ASI_P
1.931 + .end
1.932 +!
1.933 +! vis_stfa_ASI_PL(float frs, void *rs1)
1.934 +!
1.935 + .inline vis_stfa_ASI_PL,12
1.936 + sta %f1,[%o1]0x88 ! ASI_PL
1.937 + .end
1.938 +!
1.939 +! vis_stdfa_ASI_REG(double frd, void *rs1)
1.940 +!
1.941 + .inline vis_stdfa_ASI_REG,16
1.942 + stda %f0,[%o1+0]%asi
1.943 + .end
1.944 +!
1.945 +! vis_stdfa_ASI_P(double frd, void *rs1)
1.946 +!
1.947 + .inline vis_stdfa_ASI_P,16
1.948 + stda %f0,[%o1]0x80 ! ASI_P
1.949 + .end
1.950 +!
1.951 +! vis_stdfa_ASI_PL(double frd, void *rs1)
1.952 +!
1.953 + .inline vis_stdfa_ASI_PL,16
1.954 + stda %f0,[%o1]0x88 ! ASI_PL
1.955 + .end
1.956 +!
1.957 +! unsigned short vis_lduha_ASI_REG(void *rs1)
1.958 +!
1.959 + .inline vis_lduha_ASI_REG,8
1.960 + lduha [%o0+0]%asi,%o0
1.961 + .end
1.962 +!
1.963 +! unsigned short vis_lduha_ASI_P(void *rs1)
1.964 +!
1.965 + .inline vis_lduha_ASI_P,8
1.966 + lduha [%o0]0x80,%o0 ! ASI_P
1.967 + .end
1.968 +!
1.969 +! unsigned short vis_lduha_ASI_PL(void *rs1)
1.970 +!
1.971 + .inline vis_lduha_ASI_PL,8
1.972 + lduha [%o0]0x88,%o0 ! ASI_PL
1.973 + .end
1.974 +!
1.975 +! unsigned short vis_lduha_ASI_P_index(void *rs1, long index)
1.976 +!
1.977 + .inline vis_lduha_ASI_P_index,16
1.978 + lduha [%o0+%o1]0x80,%o0 ! ASI_P
1.979 + .end
1.980 +!
1.981 +! unsigned short vis_lduha_ASI_PL_index(void *rs1, long index)
1.982 +!
1.983 + .inline vis_lduha_ASI_PL_index,16
1.984 + lduha [%o0+%o1]0x88,%o0 ! ASI_PL
1.985 + .end
1.986 +
1.987 +!--------------------------------------------------------------------
1.988 +! Prefetch
1.989 +!
1.990 +! void vis_prefetch_read(void * /*address*/);
1.991 +!
1.992 + .inline vis_prefetch_read,8
1.993 + prefetch [%o0+0],0
1.994 + .end
1.995 +!
1.996 +! void vis_prefetch_write(void * /*address*/);
1.997 +!
1.998 + .inline vis_prefetch_write,8
1.999 + prefetch [%o0+0],2
1.1000 + .end
