X86InstrFPStack.td 39.5 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
//===- X86InstrFPStack.td - FPU Instruction Set ------------*- tablegen -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file describes the X86 x87 FPU instruction set, defining the
// instructions, and properties of the instructions which are needed for code
// generation, machine code emission, and analysis.
//
//===----------------------------------------------------------------------===//

//===----------------------------------------------------------------------===//
// FPStack specific DAG Nodes.
//===----------------------------------------------------------------------===//

def SDTX86Fld       : SDTypeProfile<1, 1, [SDTCisFP<0>,
                                           SDTCisPtrTy<1>]>;
def SDTX86Fst       : SDTypeProfile<0, 2, [SDTCisFP<0>,
                                           SDTCisPtrTy<1>]>;
def SDTX86Fild      : SDTypeProfile<1, 1, [SDTCisFP<0>, SDTCisPtrTy<1>]>;
def SDTX86Fist      : SDTypeProfile<0, 2, [SDTCisFP<0>, SDTCisPtrTy<1>]>;

def SDTX86CwdStore  : SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>;

def X86fld          : SDNode<"X86ISD::FLD", SDTX86Fld,
                             [SDNPHasChain, SDNPMayLoad, SDNPMemOperand]>;
def X86fst          : SDNode<"X86ISD::FST", SDTX86Fst,
                             [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;
def X86fild         : SDNode<"X86ISD::FILD", SDTX86Fild,
                             [SDNPHasChain, SDNPMayLoad, SDNPMemOperand]>;
def X86fist         : SDNode<"X86ISD::FIST", SDTX86Fist,
                             [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;
def X86fp_to_mem : SDNode<"X86ISD::FP_TO_INT_IN_MEM", SDTX86Fst,
                          [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;
def X86fp_cwd_get16 : SDNode<"X86ISD::FNSTCW16m",          SDTX86CwdStore,
                             [SDNPHasChain, SDNPMayStore, SDNPSideEffect,
                              SDNPMemOperand]>;

def X86fstf32 : PatFrag<(ops node:$val, node:$ptr),
                        (X86fst node:$val, node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::f32;
}]>;
def X86fstf64 : PatFrag<(ops node:$val, node:$ptr),
                        (X86fst node:$val, node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::f64;
}]>;
def X86fstf80 : PatFrag<(ops node:$val, node:$ptr),
                        (X86fst node:$val, node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::f80;
}]>;

def X86fldf32 : PatFrag<(ops node:$ptr), (X86fld node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::f32;
}]>;
def X86fldf64 : PatFrag<(ops node:$ptr), (X86fld node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::f64;
}]>;
def X86fldf80 : PatFrag<(ops node:$ptr), (X86fld node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::f80;
}]>;

def X86fild16 : PatFrag<(ops node:$ptr), (X86fild node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def X86fild32 : PatFrag<(ops node:$ptr), (X86fild node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;
def X86fild64 : PatFrag<(ops node:$ptr), (X86fild node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

def X86fist32 : PatFrag<(ops node:$val, node:$ptr),
                        (X86fist node:$val, node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;

def X86fist64 : PatFrag<(ops node:$val, node:$ptr),
                        (X86fist node:$val, node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

def X86fp_to_i16mem : PatFrag<(ops node:$val, node:$ptr),
                              (X86fp_to_mem node:$val, node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def X86fp_to_i32mem : PatFrag<(ops node:$val, node:$ptr),
                              (X86fp_to_mem node:$val, node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;
def X86fp_to_i64mem : PatFrag<(ops node:$val, node:$ptr),
                              (X86fp_to_mem node:$val, node:$ptr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

//===----------------------------------------------------------------------===//
// FPStack pattern fragments
//===----------------------------------------------------------------------===//

def fpimm0 : FPImmLeaf<fAny, [{
  return Imm.isExactlyValue(+0.0);
}]>;

def fpimmneg0 : FPImmLeaf<fAny, [{
  return Imm.isExactlyValue(-0.0);
}]>;

def fpimm1 : FPImmLeaf<fAny, [{
  return Imm.isExactlyValue(+1.0);
}]>;

def fpimmneg1 : FPImmLeaf<fAny, [{
  return Imm.isExactlyValue(-1.0);
}]>;

// Some 'special' instructions - expanded after instruction selection.
// Clobbers EFLAGS due to OR instruction used internally.
// FIXME: Can we model this in SelectionDAG?
let usesCustomInserter = 1, hasNoSchedulingInfo = 1, Defs = [EFLAGS] in {
  def FP32_TO_INT16_IN_MEM : PseudoI<(outs), (ins i16mem:$dst, RFP32:$src),
                              [(X86fp_to_i16mem RFP32:$src, addr:$dst)]>;
  def FP32_TO_INT32_IN_MEM : PseudoI<(outs), (ins i32mem:$dst, RFP32:$src),
                              [(X86fp_to_i32mem RFP32:$src, addr:$dst)]>;
  def FP32_TO_INT64_IN_MEM : PseudoI<(outs), (ins i64mem:$dst, RFP32:$src),
                              [(X86fp_to_i64mem RFP32:$src, addr:$dst)]>;
  def FP64_TO_INT16_IN_MEM : PseudoI<(outs), (ins i16mem:$dst, RFP64:$src),
                              [(X86fp_to_i16mem RFP64:$src, addr:$dst)]>;
  def FP64_TO_INT32_IN_MEM : PseudoI<(outs), (ins i32mem:$dst, RFP64:$src),
                              [(X86fp_to_i32mem RFP64:$src, addr:$dst)]>;
  def FP64_TO_INT64_IN_MEM : PseudoI<(outs), (ins i64mem:$dst, RFP64:$src),
                              [(X86fp_to_i64mem RFP64:$src, addr:$dst)]>;
  def FP80_TO_INT16_IN_MEM : PseudoI<(outs), (ins i16mem:$dst, RFP80:$src),
                              [(X86fp_to_i16mem RFP80:$src, addr:$dst)]>;
  def FP80_TO_INT32_IN_MEM : PseudoI<(outs), (ins i32mem:$dst, RFP80:$src),
                              [(X86fp_to_i32mem RFP80:$src, addr:$dst)]>;
  def FP80_TO_INT64_IN_MEM : PseudoI<(outs), (ins i64mem:$dst, RFP80:$src),
                              [(X86fp_to_i64mem RFP80:$src, addr:$dst)]>;
}

// All FP Stack operations are represented with four instructions here.  The
// first three instructions, generated by the instruction selector, use "RFP32"
// "RFP64" or "RFP80" registers: traditional register files to reference 32-bit,
// 64-bit or 80-bit floating point values.  These sizes apply to the values,
// not the registers, which are always 80 bits; RFP32, RFP64 and RFP80 can be
// copied to each other without losing information.  These instructions are all
// pseudo instructions and use the "_Fp" suffix.
// In some cases there are additional variants with a mixture of different
// register sizes.
// The second instruction is defined with FPI, which is the actual instruction
// emitted by the assembler.  These use "RST" registers, although frequently
// the actual register(s) used are implicit.  These are always 80 bits.
// The FP stackifier pass converts one to the other after register allocation
// occurs.
//
// Note that the FpI instruction should have instruction selection info (e.g.
// a pattern) and the FPI instruction should have emission info (e.g. opcode
// encoding and asm printing info).

// FpIf32, FpIf64 - Floating Point Pseudo Instruction template.
// f32 instructions can use SSE1 and are predicated on FPStackf32 == !SSE1.
// f64 instructions can use SSE2 and are predicated on FPStackf64 == !SSE2.
// f80 instructions cannot use SSE and use neither of these.
class FpIf32<dag outs, dag ins, FPFormat fp, list<dag> pattern> :
             FpI_<outs, ins, fp, pattern>, Requires<[FPStackf32]>;
class FpIf64<dag outs, dag ins, FPFormat fp, list<dag> pattern> :
             FpI_<outs, ins, fp, pattern>, Requires<[FPStackf64]>;

// Factoring for arithmetic.
multiclass FPBinary_rr<SDNode OpNode> {
// Register op register -> register
// These are separated out because they have no reversed form.
def _Fp32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, RFP32:$src2), TwoArgFP,
                [(set RFP32:$dst, (OpNode RFP32:$src1, RFP32:$src2))]>;
def _Fp64 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, RFP64:$src2), TwoArgFP,
                [(set RFP64:$dst, (OpNode RFP64:$src1, RFP64:$src2))]>;
def _Fp80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, RFP80:$src2), TwoArgFP,
                [(set RFP80:$dst, (OpNode RFP80:$src1, RFP80:$src2))]>;
}
// The FopST0 series are not included here because of the irregularities
// in where the 'r' goes in assembly output.
// These instructions cannot address 80-bit memory.
multiclass FPBinary<SDNode OpNode, Format fp, string asmstring,
                    bit Forward = 1> {
// ST(0) = ST(0) + [mem]
def _Fp32m  : FpIf32<(outs RFP32:$dst),
                     (ins RFP32:$src1, f32mem:$src2), OneArgFPRW,
                  [!if(Forward,
                       (set RFP32:$dst,
                        (OpNode RFP32:$src1, (loadf32 addr:$src2))),
                       (set RFP32:$dst,
                        (OpNode (loadf32 addr:$src2), RFP32:$src1)))]>;
def _Fp64m  : FpIf64<(outs RFP64:$dst),
                     (ins RFP64:$src1, f64mem:$src2), OneArgFPRW,
                  [!if(Forward,
                       (set RFP64:$dst,
                        (OpNode RFP64:$src1, (loadf64 addr:$src2))),
                       (set RFP64:$dst,
                        (OpNode (loadf64 addr:$src2), RFP64:$src1)))]>;
def _Fp64m32: FpIf64<(outs RFP64:$dst),
                     (ins RFP64:$src1, f32mem:$src2), OneArgFPRW,
                  [!if(Forward,
                       (set RFP64:$dst,
                        (OpNode RFP64:$src1, (f64 (extloadf32 addr:$src2)))),
                       (set RFP64:$dst,
                        (OpNode (f64 (extloadf32 addr:$src2)), RFP64:$src1)))]>;
def _Fp80m32: FpI_<(outs RFP80:$dst),
                   (ins RFP80:$src1, f32mem:$src2), OneArgFPRW,
                  [!if(Forward,
                       (set RFP80:$dst,
                        (OpNode RFP80:$src1, (f80 (extloadf32 addr:$src2)))),
                       (set RFP80:$dst,
                        (OpNode (f80 (extloadf32 addr:$src2)), RFP80:$src1)))]>;
def _Fp80m64: FpI_<(outs RFP80:$dst),
                   (ins RFP80:$src1, f64mem:$src2), OneArgFPRW,
                  [!if(Forward,
                       (set RFP80:$dst,
                        (OpNode RFP80:$src1, (f80 (extloadf64 addr:$src2)))),
                       (set RFP80:$dst,
                        (OpNode (f80 (extloadf64 addr:$src2)), RFP80:$src1)))]>;
let mayLoad = 1 in
def _F32m  : FPI<0xD8, fp, (outs), (ins f32mem:$src),
                 !strconcat("f", asmstring, "{s}\t$src")>;
let mayLoad = 1 in
def _F64m  : FPI<0xDC, fp, (outs), (ins f64mem:$src),
                 !strconcat("f", asmstring, "{l}\t$src")>;
// ST(0) = ST(0) + [memint]
def _FpI16m32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, i16mem:$src2),
                       OneArgFPRW,
                       [!if(Forward,
                            (set RFP32:$dst,
                             (OpNode RFP32:$src1, (X86fild16 addr:$src2))),
                            (set RFP32:$dst,
                             (OpNode (X86fild16 addr:$src2), RFP32:$src1)))]>;
def _FpI32m32 : FpIf32<(outs RFP32:$dst), (ins RFP32:$src1, i32mem:$src2),
                       OneArgFPRW,
                       [!if(Forward,
                            (set RFP32:$dst,
                             (OpNode RFP32:$src1, (X86fild32 addr:$src2))),
                            (set RFP32:$dst,
                             (OpNode (X86fild32 addr:$src2), RFP32:$src1)))]>;
def _FpI16m64 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, i16mem:$src2),
                       OneArgFPRW,
                       [!if(Forward,
                            (set RFP64:$dst,
                             (OpNode RFP64:$src1, (X86fild16 addr:$src2))),
                            (set RFP64:$dst,
                             (OpNode (X86fild16 addr:$src2), RFP64:$src1)))]>;
def _FpI32m64 : FpIf64<(outs RFP64:$dst), (ins RFP64:$src1, i32mem:$src2),
                       OneArgFPRW,
                       [!if(Forward,
                            (set RFP64:$dst,
                             (OpNode RFP64:$src1, (X86fild32 addr:$src2))),
                            (set RFP64:$dst,
                             (OpNode (X86fild32 addr:$src2), RFP64:$src1)))]>;
def _FpI16m80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, i16mem:$src2),
                     OneArgFPRW,
                     [!if(Forward,
                          (set RFP80:$dst,
                           (OpNode RFP80:$src1, (X86fild16 addr:$src2))),
                          (set RFP80:$dst,
                           (OpNode (X86fild16 addr:$src2), RFP80:$src1)))]>;
def _FpI32m80 : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, i32mem:$src2),
                     OneArgFPRW,
                     [!if(Forward,
                          (set RFP80:$dst,
                           (OpNode RFP80:$src1, (X86fild32 addr:$src2))),
                          (set RFP80:$dst,
                           (OpNode (X86fild32 addr:$src2), RFP80:$src1)))]>;
let mayLoad = 1 in
def _FI16m  : FPI<0xDE, fp, (outs), (ins i16mem:$src),
                  !strconcat("fi", asmstring, "{s}\t$src")>;
let mayLoad = 1 in
def _FI32m  : FPI<0xDA, fp, (outs), (ins i32mem:$src),
                  !strconcat("fi", asmstring, "{l}\t$src")>;
}

let Uses = [FPCW], mayRaiseFPException = 1 in {
// FPBinary_rr just defines pseudo-instructions, no need to set a scheduling
// resources.
let hasNoSchedulingInfo = 1 in {
defm ADD : FPBinary_rr<any_fadd>;
defm SUB : FPBinary_rr<any_fsub>;
defm MUL : FPBinary_rr<any_fmul>;
defm DIV : FPBinary_rr<any_fdiv>;
}

// Sets the scheduling resources for the actual NAME#_F<size>m definitions.
let SchedRW = [WriteFAddLd] in {
defm ADD : FPBinary<any_fadd, MRM0m, "add">;
defm SUB : FPBinary<any_fsub, MRM4m, "sub">;
defm SUBR: FPBinary<any_fsub ,MRM5m, "subr", 0>;
}

let SchedRW = [WriteFMulLd] in {
defm MUL : FPBinary<any_fmul, MRM1m, "mul">;
}

let SchedRW = [WriteFDivLd] in {
defm DIV : FPBinary<any_fdiv, MRM6m, "div">;
defm DIVR: FPBinary<any_fdiv, MRM7m, "divr", 0>;
}
} // Uses = [FPCW], mayRaiseFPException = 1

class FPST0rInst<Format fp, string asm>
  : FPI<0xD8, fp, (outs), (ins RSTi:$op), asm>;
class FPrST0Inst<Format fp, string asm>
  : FPI<0xDC, fp, (outs), (ins RSTi:$op), asm>;
class FPrST0PInst<Format fp, string asm>
  : FPI<0xDE, fp, (outs), (ins RSTi:$op), asm>;

// NOTE: GAS and apparently all other AT&T style assemblers have a broken notion
// of some of the 'reverse' forms of the fsub and fdiv instructions.  As such,
// we have to put some 'r's in and take them out of weird places.
let SchedRW = [WriteFAdd], Uses = [FPCW], mayRaiseFPException = 1 in {
def ADD_FST0r   : FPST0rInst <MRM0r, "fadd\t{$op, %st|st, $op}">;
def ADD_FrST0   : FPrST0Inst <MRM0r, "fadd\t{%st, $op|$op, st}">;
def ADD_FPrST0  : FPrST0PInst<MRM0r, "faddp\t{%st, $op|$op, st}">;
def SUBR_FST0r  : FPST0rInst <MRM5r, "fsubr\t{$op, %st|st, $op}">;
def SUB_FrST0   : FPrST0Inst <MRM5r, "fsub{r}\t{%st, $op|$op, st}">;
def SUB_FPrST0  : FPrST0PInst<MRM5r, "fsub{r}p\t{%st, $op|$op, st}">;
def SUB_FST0r   : FPST0rInst <MRM4r, "fsub\t{$op, %st|st, $op}">;
def SUBR_FrST0  : FPrST0Inst <MRM4r, "fsub{|r}\t{%st, $op|$op, st}">;
def SUBR_FPrST0 : FPrST0PInst<MRM4r, "fsub{|r}p\t{%st, $op|$op, st}">;
} // SchedRW
let SchedRW = [WriteFCom], Uses = [FPCW], mayRaiseFPException = 1 in {
def COM_FST0r   : FPST0rInst <MRM2r, "fcom\t$op">;
def COMP_FST0r  : FPST0rInst <MRM3r, "fcomp\t$op">;
} // SchedRW
let SchedRW = [WriteFMul], Uses = [FPCW], mayRaiseFPException = 1 in {
def MUL_FST0r   : FPST0rInst <MRM1r, "fmul\t{$op, %st|st, $op}">;
def MUL_FrST0   : FPrST0Inst <MRM1r, "fmul\t{%st, $op|$op, st}">;
def MUL_FPrST0  : FPrST0PInst<MRM1r, "fmulp\t{%st, $op|$op, st}">;
} // SchedRW
let SchedRW = [WriteFDiv], Uses = [FPCW], mayRaiseFPException = 1 in {
def DIVR_FST0r  : FPST0rInst <MRM7r, "fdivr\t{$op, %st|st, $op}">;
def DIV_FrST0   : FPrST0Inst <MRM7r, "fdiv{r}\t{%st, $op|$op, st}">;
def DIV_FPrST0  : FPrST0PInst<MRM7r, "fdiv{r}p\t{%st, $op|$op, st}">;
def DIV_FST0r   : FPST0rInst <MRM6r, "fdiv\t{$op, %st|st, $op}">;
def DIVR_FrST0  : FPrST0Inst <MRM6r, "fdiv{|r}\t{%st, $op|$op, st}">;
def DIVR_FPrST0 : FPrST0PInst<MRM6r, "fdiv{|r}p\t{%st, $op|$op, st}">;
} // SchedRW

// Unary operations.
multiclass FPUnary<SDNode OpNode, Format fp, string asmstring> {
def _Fp32  : FpIf32<(outs RFP32:$dst), (ins RFP32:$src), OneArgFPRW,
                 [(set RFP32:$dst, (OpNode RFP32:$src))]>;
def _Fp64  : FpIf64<(outs RFP64:$dst), (ins RFP64:$src), OneArgFPRW,
                 [(set RFP64:$dst, (OpNode RFP64:$src))]>;
def _Fp80  : FpI_<(outs RFP80:$dst), (ins RFP80:$src), OneArgFPRW,
                 [(set RFP80:$dst, (OpNode RFP80:$src))]>;
def _F     : FPI<0xD9, fp, (outs), (ins), asmstring>;
}

let SchedRW = [WriteFSign] in {
defm CHS : FPUnary<fneg, MRM_E0, "fchs">;
defm ABS : FPUnary<fabs, MRM_E1, "fabs">;
}

let Uses = [FPCW], mayRaiseFPException = 1 in {
let SchedRW = [WriteFSqrt80] in
defm SQRT: FPUnary<any_fsqrt,MRM_FA, "fsqrt">;

let SchedRW = [WriteFCom] in {
let hasSideEffects = 0 in {
def TST_Fp32  : FpIf32<(outs), (ins RFP32:$src), OneArgFP, []>;
def TST_Fp64  : FpIf64<(outs), (ins RFP64:$src), OneArgFP, []>;
def TST_Fp80  : FpI_<(outs), (ins RFP80:$src), OneArgFP, []>;
} // hasSideEffects

def TST_F  : FPI<0xD9, MRM_E4, (outs), (ins), "ftst">;
} // SchedRW
} // Uses = [FPCW], mayRaiseFPException = 1

// Versions of FP instructions that take a single memory operand.  Added for the
//   disassembler; remove as they are included with patterns elsewhere.
let SchedRW = [WriteFComLd], Uses = [FPCW], mayRaiseFPException = 1,
    mayLoad = 1 in {
def FCOM32m  : FPI<0xD8, MRM2m, (outs), (ins f32mem:$src), "fcom{s}\t$src">;
def FCOMP32m : FPI<0xD8, MRM3m, (outs), (ins f32mem:$src), "fcomp{s}\t$src">;

def FCOM64m  : FPI<0xDC, MRM2m, (outs), (ins f64mem:$src), "fcom{l}\t$src">;
def FCOMP64m : FPI<0xDC, MRM3m, (outs), (ins f64mem:$src), "fcomp{l}\t$src">;

def FICOM16m : FPI<0xDE, MRM2m, (outs), (ins i16mem:$src), "ficom{s}\t$src">;
def FICOMP16m: FPI<0xDE, MRM3m, (outs), (ins i16mem:$src), "ficomp{s}\t$src">;

def FICOM32m : FPI<0xDA, MRM2m, (outs), (ins i32mem:$src), "ficom{l}\t$src">;
def FICOMP32m: FPI<0xDA, MRM3m, (outs), (ins i32mem:$src), "ficomp{l}\t$src">;
} // SchedRW

let SchedRW = [WriteMicrocoded] in {
let Defs = [FPSW, FPCW], mayLoad = 1 in {
def FLDENVm  : FPI<0xD9, MRM4m, (outs), (ins f32mem:$src), "fldenv\t$src">;
def FRSTORm  : FPI<0xDD, MRM4m, (outs), (ins f32mem:$dst), "frstor\t$dst">;
}

let Defs = [FPSW, FPCW], Uses = [FPSW, FPCW], mayStore = 1 in {
def FSTENVm  : FPI<0xD9, MRM6m, (outs), (ins f32mem:$dst), "fnstenv\t$dst">;
def FSAVEm   : FPI<0xDD, MRM6m, (outs), (ins f32mem:$dst), "fnsave\t$dst">;
}

let Uses = [FPSW], mayStore = 1 in
def FNSTSWm  : FPI<0xDD, MRM7m, (outs), (ins i16mem:$dst), "fnstsw\t$dst">;

let mayLoad = 1 in
def FBLDm    : FPI<0xDF, MRM4m, (outs), (ins f80mem:$src), "fbld\t$src">;
let Uses = [FPCW] ,mayRaiseFPException = 1, mayStore = 1 in
def FBSTPm   : FPI<0xDF, MRM6m, (outs), (ins f80mem:$dst), "fbstp\t$dst">;
} // SchedRW

// Floating point cmovs.
class FpIf32CMov<dag outs, dag ins, FPFormat fp, list<dag> pattern> :
  FpI_<outs, ins, fp, pattern>, Requires<[FPStackf32, HasCMov]>;
class FpIf64CMov<dag outs, dag ins, FPFormat fp, list<dag> pattern> :
  FpI_<outs, ins, fp, pattern>, Requires<[FPStackf64, HasCMov]>;

multiclass FPCMov<PatLeaf cc> {
  def _Fp32  : FpIf32CMov<(outs RFP32:$dst), (ins RFP32:$src1, RFP32:$src2),
                       CondMovFP,
                     [(set RFP32:$dst, (X86cmov RFP32:$src1, RFP32:$src2,
                                        cc, EFLAGS))]>;
  def _Fp64  : FpIf64CMov<(outs RFP64:$dst), (ins RFP64:$src1, RFP64:$src2),
                       CondMovFP,
                     [(set RFP64:$dst, (X86cmov RFP64:$src1, RFP64:$src2,
                                        cc, EFLAGS))]>;
  def _Fp80  : FpI_<(outs RFP80:$dst), (ins RFP80:$src1, RFP80:$src2),
                     CondMovFP,
                     [(set RFP80:$dst, (X86cmov RFP80:$src1, RFP80:$src2,
                                        cc, EFLAGS))]>,
                                        Requires<[HasCMov]>;
}

let SchedRW = [WriteFCMOV] in {
let Uses = [EFLAGS], Constraints = "$src1 = $dst" in {
defm CMOVB  : FPCMov<X86_COND_B>;
defm CMOVBE : FPCMov<X86_COND_BE>;
defm CMOVE  : FPCMov<X86_COND_E>;
defm CMOVP  : FPCMov<X86_COND_P>;
defm CMOVNB : FPCMov<X86_COND_AE>;
defm CMOVNBE: FPCMov<X86_COND_A>;
defm CMOVNE : FPCMov<X86_COND_NE>;
defm CMOVNP : FPCMov<X86_COND_NP>;
} // Uses = [EFLAGS], Constraints = "$src1 = $dst"

let Predicates = [HasCMov] in {
// These are not factored because there's no clean way to pass DA/DB.
def CMOVB_F  : FPI<0xDA, MRM0r, (outs), (ins RSTi:$op),
                  "fcmovb\t{$op, %st|st, $op}">;
def CMOVBE_F : FPI<0xDA, MRM2r, (outs), (ins RSTi:$op),
                  "fcmovbe\t{$op, %st|st, $op}">;
def CMOVE_F  : FPI<0xDA, MRM1r, (outs), (ins RSTi:$op),
                  "fcmove\t{$op, %st|st, $op}">;
def CMOVP_F  : FPI<0xDA, MRM3r, (outs), (ins RSTi:$op),
                  "fcmovu\t{$op, %st|st, $op}">;
def CMOVNB_F : FPI<0xDB, MRM0r, (outs), (ins RSTi:$op),
                  "fcmovnb\t{$op, %st|st, $op}">;
def CMOVNBE_F: FPI<0xDB, MRM2r, (outs), (ins RSTi:$op),
                  "fcmovnbe\t{$op, %st|st, $op}">;
def CMOVNE_F : FPI<0xDB, MRM1r, (outs), (ins RSTi:$op),
                  "fcmovne\t{$op, %st|st, $op}">;
def CMOVNP_F : FPI<0xDB, MRM3r, (outs), (ins RSTi:$op),
                  "fcmovnu\t{$op, %st|st, $op}">;
} // Predicates = [HasCMov]
} // SchedRW

let mayRaiseFPException = 1 in {
// Floating point loads & stores.
let SchedRW = [WriteLoad], Uses = [FPCW] in {
let canFoldAsLoad = 1 in {
def LD_Fp32m   : FpIf32<(outs RFP32:$dst), (ins f32mem:$src), ZeroArgFP,
                  [(set RFP32:$dst, (loadf32 addr:$src))]>;
def LD_Fp64m : FpIf64<(outs RFP64:$dst), (ins f64mem:$src), ZeroArgFP,
                  [(set RFP64:$dst, (loadf64 addr:$src))]>;
def LD_Fp80m   : FpI_<(outs RFP80:$dst), (ins f80mem:$src), ZeroArgFP,
                  [(set RFP80:$dst, (loadf80 addr:$src))]>;
} // canFoldAsLoad
def LD_Fp32m64 : FpIf64<(outs RFP64:$dst), (ins f32mem:$src), ZeroArgFP,
                  [(set RFP64:$dst, (f64 (extloadf32 addr:$src)))]>;
def LD_Fp64m80 : FpI_<(outs RFP80:$dst), (ins f64mem:$src), ZeroArgFP,
                  [(set RFP80:$dst, (f80 (extloadf64 addr:$src)))]>;
def LD_Fp32m80 : FpI_<(outs RFP80:$dst), (ins f32mem:$src), ZeroArgFP,
                  [(set RFP80:$dst, (f80 (extloadf32 addr:$src)))]>;
let mayRaiseFPException = 0 in {
def ILD_Fp16m32: FpIf32<(outs RFP32:$dst), (ins i16mem:$src), ZeroArgFP,
                  [(set RFP32:$dst, (X86fild16 addr:$src))]>;
def ILD_Fp32m32: FpIf32<(outs RFP32:$dst), (ins i32mem:$src), ZeroArgFP,
                  [(set RFP32:$dst, (X86fild32 addr:$src))]>;
def ILD_Fp64m32: FpIf32<(outs RFP32:$dst), (ins i64mem:$src), ZeroArgFP,
                  [(set RFP32:$dst, (X86fild64 addr:$src))]>;
def ILD_Fp16m64: FpIf64<(outs RFP64:$dst), (ins i16mem:$src), ZeroArgFP,
                  [(set RFP64:$dst, (X86fild16 addr:$src))]>;
def ILD_Fp32m64: FpIf64<(outs RFP64:$dst), (ins i32mem:$src), ZeroArgFP,
                  [(set RFP64:$dst, (X86fild32 addr:$src))]>;
def ILD_Fp64m64: FpIf64<(outs RFP64:$dst), (ins i64mem:$src), ZeroArgFP,
                  [(set RFP64:$dst, (X86fild64 addr:$src))]>;
def ILD_Fp16m80: FpI_<(outs RFP80:$dst), (ins i16mem:$src), ZeroArgFP,
                  [(set RFP80:$dst, (X86fild16 addr:$src))]>;
def ILD_Fp32m80: FpI_<(outs RFP80:$dst), (ins i32mem:$src), ZeroArgFP,
                  [(set RFP80:$dst, (X86fild32 addr:$src))]>;
def ILD_Fp64m80: FpI_<(outs RFP80:$dst), (ins i64mem:$src), ZeroArgFP,
                  [(set RFP80:$dst, (X86fild64 addr:$src))]>;
} // mayRaiseFPException = 0
} // SchedRW

let SchedRW = [WriteStore], Uses = [FPCW] in {
def ST_Fp32m   : FpIf32<(outs), (ins f32mem:$op, RFP32:$src), OneArgFP,
                  [(store RFP32:$src, addr:$op)]>;
def ST_Fp64m32 : FpIf64<(outs), (ins f32mem:$op, RFP64:$src), OneArgFP,
                  [(truncstoref32 RFP64:$src, addr:$op)]>;
def ST_Fp64m   : FpIf64<(outs), (ins f64mem:$op, RFP64:$src), OneArgFP,
                  [(store RFP64:$src, addr:$op)]>;
def ST_Fp80m32 : FpI_<(outs), (ins f32mem:$op, RFP80:$src), OneArgFP,
                  [(truncstoref32 RFP80:$src, addr:$op)]>;
def ST_Fp80m64 : FpI_<(outs), (ins f64mem:$op, RFP80:$src), OneArgFP,
                  [(truncstoref64 RFP80:$src, addr:$op)]>;
// FST does not support 80-bit memory target; FSTP must be used.

let mayStore = 1, hasSideEffects = 0 in {
def ST_FpP32m    : FpIf32<(outs), (ins f32mem:$op, RFP32:$src), OneArgFP, []>;
def ST_FpP64m32  : FpIf64<(outs), (ins f32mem:$op, RFP64:$src), OneArgFP, []>;
def ST_FpP64m    : FpIf64<(outs), (ins f64mem:$op, RFP64:$src), OneArgFP, []>;
def ST_FpP80m32  : FpI_<(outs), (ins f32mem:$op, RFP80:$src), OneArgFP, []>;
def ST_FpP80m64  : FpI_<(outs), (ins f64mem:$op, RFP80:$src), OneArgFP, []>;
} // mayStore

def ST_FpP80m    : FpI_<(outs), (ins f80mem:$op, RFP80:$src), OneArgFP,
                    [(store RFP80:$src, addr:$op)]>;

let mayStore = 1, hasSideEffects = 0 in {
def IST_Fp16m32  : FpIf32<(outs), (ins i16mem:$op, RFP32:$src), OneArgFP, []>;
def IST_Fp32m32  : FpIf32<(outs), (ins i32mem:$op, RFP32:$src), OneArgFP,
                          [(X86fist32 RFP32:$src, addr:$op)]>;
def IST_Fp64m32  : FpIf32<(outs), (ins i64mem:$op, RFP32:$src), OneArgFP,
                          [(X86fist64 RFP32:$src, addr:$op)]>;
def IST_Fp16m64  : FpIf64<(outs), (ins i16mem:$op, RFP64:$src), OneArgFP, []>;
def IST_Fp32m64  : FpIf64<(outs), (ins i32mem:$op, RFP64:$src), OneArgFP,
                          [(X86fist32 RFP64:$src, addr:$op)]>;
def IST_Fp64m64  : FpIf64<(outs), (ins i64mem:$op, RFP64:$src), OneArgFP,
                          [(X86fist64 RFP64:$src, addr:$op)]>;
def IST_Fp16m80  : FpI_<(outs), (ins i16mem:$op, RFP80:$src), OneArgFP, []>;
def IST_Fp32m80  : FpI_<(outs), (ins i32mem:$op, RFP80:$src), OneArgFP,
                        [(X86fist32 RFP80:$src, addr:$op)]>;
def IST_Fp64m80  : FpI_<(outs), (ins i64mem:$op, RFP80:$src), OneArgFP,
                        [(X86fist64 RFP80:$src, addr:$op)]>;
} // mayStore
} // SchedRW, Uses = [FPCW]

let mayLoad = 1, SchedRW = [WriteLoad], Uses = [FPCW] in {
def LD_F32m   : FPI<0xD9, MRM0m, (outs), (ins f32mem:$src), "fld{s}\t$src">;
def LD_F64m   : FPI<0xDD, MRM0m, (outs), (ins f64mem:$src), "fld{l}\t$src">;
def LD_F80m   : FPI<0xDB, MRM5m, (outs), (ins f80mem:$src), "fld{t}\t$src">;
let mayRaiseFPException = 0 in {
def ILD_F16m  : FPI<0xDF, MRM0m, (outs), (ins i16mem:$src), "fild{s}\t$src">;
def ILD_F32m  : FPI<0xDB, MRM0m, (outs), (ins i32mem:$src), "fild{l}\t$src">;
def ILD_F64m  : FPI<0xDF, MRM5m, (outs), (ins i64mem:$src), "fild{ll}\t$src">;
}
}
let mayStore = 1, SchedRW = [WriteStore], Uses = [FPCW] in {
def ST_F32m   : FPI<0xD9, MRM2m, (outs), (ins f32mem:$dst), "fst{s}\t$dst">;
def ST_F64m   : FPI<0xDD, MRM2m, (outs), (ins f64mem:$dst), "fst{l}\t$dst">;
def ST_FP32m  : FPI<0xD9, MRM3m, (outs), (ins f32mem:$dst), "fstp{s}\t$dst">;
def ST_FP64m  : FPI<0xDD, MRM3m, (outs), (ins f64mem:$dst), "fstp{l}\t$dst">;
def ST_FP80m  : FPI<0xDB, MRM7m, (outs), (ins f80mem:$dst), "fstp{t}\t$dst">;
def IST_F16m  : FPI<0xDF, MRM2m, (outs), (ins i16mem:$dst), "fist{s}\t$dst">;
def IST_F32m  : FPI<0xDB, MRM2m, (outs), (ins i32mem:$dst), "fist{l}\t$dst">;
def IST_FP16m : FPI<0xDF, MRM3m, (outs), (ins i16mem:$dst), "fistp{s}\t$dst">;
def IST_FP32m : FPI<0xDB, MRM3m, (outs), (ins i32mem:$dst), "fistp{l}\t$dst">;
def IST_FP64m : FPI<0xDF, MRM7m, (outs), (ins i64mem:$dst), "fistp{ll}\t$dst">;
}

// FISTTP requires SSE3 even though it's a FPStack op.
let Predicates = [HasSSE3], SchedRW = [WriteStore], Uses = [FPCW] in {
def ISTT_Fp16m32 : FpI_<(outs), (ins i16mem:$op, RFP32:$src), OneArgFP,
                    [(X86fp_to_i16mem RFP32:$src, addr:$op)]>;
def ISTT_Fp32m32 : FpI_<(outs), (ins i32mem:$op, RFP32:$src), OneArgFP,
                    [(X86fp_to_i32mem RFP32:$src, addr:$op)]>;
def ISTT_Fp64m32 : FpI_<(outs), (ins i64mem:$op, RFP32:$src), OneArgFP,
                    [(X86fp_to_i64mem RFP32:$src, addr:$op)]>;
def ISTT_Fp16m64 : FpI_<(outs), (ins i16mem:$op, RFP64:$src), OneArgFP,
                    [(X86fp_to_i16mem RFP64:$src, addr:$op)]>;
def ISTT_Fp32m64 : FpI_<(outs), (ins i32mem:$op, RFP64:$src), OneArgFP,
                    [(X86fp_to_i32mem RFP64:$src, addr:$op)]>;
def ISTT_Fp64m64 : FpI_<(outs), (ins i64mem:$op, RFP64:$src), OneArgFP,
                    [(X86fp_to_i64mem RFP64:$src, addr:$op)]>;
def ISTT_Fp16m80 : FpI_<(outs), (ins i16mem:$op, RFP80:$src), OneArgFP,
                    [(X86fp_to_i16mem RFP80:$src, addr:$op)]>;
def ISTT_Fp32m80 : FpI_<(outs), (ins i32mem:$op, RFP80:$src), OneArgFP,
                    [(X86fp_to_i32mem RFP80:$src, addr:$op)]>;
def ISTT_Fp64m80 : FpI_<(outs), (ins i64mem:$op, RFP80:$src), OneArgFP,
                    [(X86fp_to_i64mem RFP80:$src, addr:$op)]>;
} // Predicates = [HasSSE3]

let mayStore = 1, SchedRW = [WriteStore], Uses = [FPCW] in {
def ISTT_FP16m : FPI<0xDF, MRM1m, (outs), (ins i16mem:$dst), "fisttp{s}\t$dst">;
def ISTT_FP32m : FPI<0xDB, MRM1m, (outs), (ins i32mem:$dst), "fisttp{l}\t$dst">;
def ISTT_FP64m : FPI<0xDD, MRM1m, (outs), (ins i64mem:$dst), "fisttp{ll}\t$dst">;
}

// FP Stack manipulation instructions.
let SchedRW = [WriteMove], Uses = [FPCW] in {
def LD_Frr   : FPI<0xD9, MRM0r, (outs), (ins RSTi:$op), "fld\t$op">;
def ST_Frr   : FPI<0xDD, MRM2r, (outs), (ins RSTi:$op), "fst\t$op">;
def ST_FPrr  : FPI<0xDD, MRM3r, (outs), (ins RSTi:$op), "fstp\t$op">;
let mayRaiseFPException = 0 in
def XCH_F    : FPI<0xD9, MRM1r, (outs), (ins RSTi:$op), "fxch\t$op">;
}

// Floating point constant loads.
let SchedRW = [WriteZero], Uses = [FPCW] in {
def LD_Fp032 : FpIf32<(outs RFP32:$dst), (ins), ZeroArgFP,
                [(set RFP32:$dst, fpimm0)]>;
def LD_Fp132 : FpIf32<(outs RFP32:$dst), (ins), ZeroArgFP,
                [(set RFP32:$dst, fpimm1)]>;
def LD_Fp064 : FpIf64<(outs RFP64:$dst), (ins), ZeroArgFP,
                [(set RFP64:$dst, fpimm0)]>;
def LD_Fp164 : FpIf64<(outs RFP64:$dst), (ins), ZeroArgFP,
                [(set RFP64:$dst, fpimm1)]>;
def LD_Fp080 : FpI_<(outs RFP80:$dst), (ins), ZeroArgFP,
                [(set RFP80:$dst, fpimm0)]>;
def LD_Fp180 : FpI_<(outs RFP80:$dst), (ins), ZeroArgFP,
                [(set RFP80:$dst, fpimm1)]>;
}

let SchedRW = [WriteFLD0], Uses = [FPCW], mayRaiseFPException = 0 in
def LD_F0 : FPI<0xD9, MRM_EE, (outs), (ins), "fldz">;

let SchedRW = [WriteFLD1], Uses = [FPCW], mayRaiseFPException = 0 in
def LD_F1 : FPI<0xD9, MRM_E8, (outs), (ins), "fld1">;

let SchedRW = [WriteFLDC], Defs = [FPSW], Uses = [FPCW], mayRaiseFPException = 0 in {
def FLDL2T : I<0xD9, MRM_E9, (outs), (ins), "fldl2t", []>;
def FLDL2E : I<0xD9, MRM_EA, (outs), (ins), "fldl2e", []>;
def FLDPI : I<0xD9, MRM_EB, (outs), (ins), "fldpi", []>;
def FLDLG2 : I<0xD9, MRM_EC, (outs), (ins), "fldlg2", []>;
def FLDLN2 : I<0xD9, MRM_ED, (outs), (ins), "fldln2", []>;
} // SchedRW

// Floating point compares.
let SchedRW = [WriteFCom], Uses = [FPCW], hasSideEffects = 0 in {
def UCOM_Fpr32 : FpIf32<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP, []>;
def UCOM_Fpr64 : FpIf64<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP, []>;
def UCOM_Fpr80 : FpI_  <(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP, []>;
def COM_Fpr32  : FpIf32<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP, []>;
def COM_Fpr64  : FpIf64<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP, []>;
def COM_Fpr80  : FpI_  <(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP, []>;
} // SchedRW
} // mayRaiseFPException = 1

let SchedRW = [WriteFCom], mayRaiseFPException = 1 in {
// CC = ST(0) cmp ST(i)
let Defs = [EFLAGS, FPSW], Uses = [FPCW] in {
def UCOM_FpIr32: FpI_<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP,
                  [(set EFLAGS, (X86any_fcmp RFP32:$lhs, RFP32:$rhs))]>,
                  Requires<[FPStackf32, HasCMov]>;
def UCOM_FpIr64: FpI_<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP,
                  [(set EFLAGS, (X86any_fcmp RFP64:$lhs, RFP64:$rhs))]>,
                  Requires<[FPStackf64, HasCMov]>;
def UCOM_FpIr80: FpI_<(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP,
                  [(set EFLAGS, (X86any_fcmp RFP80:$lhs, RFP80:$rhs))]>,
                  Requires<[HasCMov]>;
def COM_FpIr32: FpI_<(outs), (ins RFP32:$lhs, RFP32:$rhs), CompareFP,
                  [(set EFLAGS, (X86strict_fcmps RFP32:$lhs, RFP32:$rhs))]>,
                  Requires<[FPStackf32, HasCMov]>;
def COM_FpIr64: FpI_<(outs), (ins RFP64:$lhs, RFP64:$rhs), CompareFP,
                  [(set EFLAGS, (X86strict_fcmps RFP64:$lhs, RFP64:$rhs))]>,
                  Requires<[FPStackf64, HasCMov]>;
def COM_FpIr80: FpI_<(outs), (ins RFP80:$lhs, RFP80:$rhs), CompareFP,
                  [(set EFLAGS, (X86strict_fcmps RFP80:$lhs, RFP80:$rhs))]>,
                  Requires<[HasCMov]>;
}

let Uses = [ST0, FPCW] in {
def UCOM_Fr    : FPI<0xDD, MRM4r,    // FPSW = cmp ST(0) with ST(i)
                    (outs), (ins RSTi:$reg), "fucom\t$reg">;
def UCOM_FPr   : FPI<0xDD, MRM5r,    // FPSW = cmp ST(0) with ST(i), pop
                    (outs), (ins RSTi:$reg), "fucomp\t$reg">;
def UCOM_FPPr  : FPI<0xDA, MRM_E9,       // cmp ST(0) with ST(1), pop, pop
                    (outs), (ins), "fucompp">;
}

let Defs = [EFLAGS, FPSW], Uses = [ST0, FPCW] in {
def UCOM_FIr   : FPI<0xDB, MRM5r,     // CC = cmp ST(0) with ST(i)
                    (outs), (ins RSTi:$reg), "fucomi\t{$reg, %st|st, $reg}">;
def UCOM_FIPr  : FPI<0xDF, MRM5r,     // CC = cmp ST(0) with ST(i), pop
                    (outs), (ins RSTi:$reg), "fucompi\t{$reg, %st|st, $reg}">;

def COM_FIr : FPI<0xDB, MRM6r, (outs), (ins RSTi:$reg),
                  "fcomi\t{$reg, %st|st, $reg}">;
def COM_FIPr : FPI<0xDF, MRM6r, (outs), (ins RSTi:$reg),
                   "fcompi\t{$reg, %st|st, $reg}">;
}
} // SchedRW

// Floating point flag ops.
let SchedRW = [WriteALU] in {
let Defs = [AX, FPSW], Uses = [FPSW], hasSideEffects = 0 in
def FNSTSW16r : I<0xDF, MRM_E0,                  // AX = fp flags
                  (outs), (ins), "fnstsw\t{%ax|ax}", []>;
let Defs = [FPSW], Uses = [FPCW] in
def FNSTCW16m : I<0xD9, MRM7m,                   // [mem16] = X87 control world
                  (outs), (ins i16mem:$dst), "fnstcw\t$dst",
                  [(X86fp_cwd_get16 addr:$dst)]>;
} // SchedRW
let Defs = [FPSW,FPCW], mayLoad = 1 in
def FLDCW16m  : I<0xD9, MRM5m,                   // X87 control world = [mem16]
                  (outs), (ins i16mem:$dst), "fldcw\t$dst", []>,
                Sched<[WriteLoad]>;

// FPU control instructions
let SchedRW = [WriteMicrocoded] in {
def FFREE : FPI<0xDD, MRM0r, (outs), (ins RSTi:$reg), "ffree\t$reg">;
def FFREEP : FPI<0xDF, MRM0r, (outs), (ins RSTi:$reg), "ffreep\t$reg">;

let Defs = [FPSW, FPCW] in
def FNINIT : I<0xDB, MRM_E3, (outs), (ins), "fninit", []>;
// Clear exceptions
let Defs = [FPSW] in
def FNCLEX : I<0xDB, MRM_E2, (outs), (ins), "fnclex", []>;
} // SchedRW

// Operand-less floating-point instructions for the disassembler.
let Defs = [FPSW] in
def FNOP : I<0xD9, MRM_D0, (outs), (ins), "fnop", []>, Sched<[WriteNop]>;

let SchedRW = [WriteMicrocoded] in {
let Defs = [FPSW] in {
def WAIT : I<0x9B, RawFrm, (outs), (ins), "wait", []>;
def FXAM : I<0xD9, MRM_E5, (outs), (ins), "fxam", []>;
def FDECSTP : I<0xD9, MRM_F6, (outs), (ins), "fdecstp", []>;
def FINCSTP : I<0xD9, MRM_F7, (outs), (ins), "fincstp", []>;
let Uses = [FPCW], mayRaiseFPException = 1 in {
def F2XM1 : I<0xD9, MRM_F0, (outs), (ins), "f2xm1", []>;
def FYL2X : I<0xD9, MRM_F1, (outs), (ins), "fyl2x", []>;
def FPTAN : I<0xD9, MRM_F2, (outs), (ins), "fptan", []>;
def FPATAN : I<0xD9, MRM_F3, (outs), (ins), "fpatan", []>;
def FXTRACT : I<0xD9, MRM_F4, (outs), (ins), "fxtract", []>;
def FPREM1 : I<0xD9, MRM_F5, (outs), (ins), "fprem1", []>;
def FPREM : I<0xD9, MRM_F8, (outs), (ins), "fprem", []>;
def FYL2XP1 : I<0xD9, MRM_F9, (outs), (ins), "fyl2xp1", []>;
def FSIN : I<0xD9, MRM_FE, (outs), (ins), "fsin", []>;
def FCOS : I<0xD9, MRM_FF, (outs), (ins), "fcos", []>;
def FSINCOS : I<0xD9, MRM_FB, (outs), (ins), "fsincos", []>;
def FRNDINT : I<0xD9, MRM_FC, (outs), (ins), "frndint", []>;
def FSCALE : I<0xD9, MRM_FD, (outs), (ins), "fscale", []>;
def FCOMPP : I<0xDE, MRM_D9, (outs), (ins), "fcompp", []>;
} // Uses = [FPCW], mayRaiseFPException = 1
} // Defs = [FPSW]

let Uses = [FPSW, FPCW] in {
def FXSAVE : I<0xAE, MRM0m, (outs), (ins opaquemem:$dst),
             "fxsave\t$dst", [(int_x86_fxsave addr:$dst)]>, PS,
             Requires<[HasFXSR]>;
def FXSAVE64 : RI<0xAE, MRM0m, (outs), (ins opaquemem:$dst),
               "fxsave64\t$dst", [(int_x86_fxsave64 addr:$dst)]>,
               PS, Requires<[HasFXSR, In64BitMode]>;
} // Uses = [FPSW, FPCW]

let Defs = [FPSW, FPCW] in {
def FXRSTOR : I<0xAE, MRM1m, (outs), (ins opaquemem:$src),
              "fxrstor\t$src", [(int_x86_fxrstor addr:$src)]>,
              PS, Requires<[HasFXSR]>;
def FXRSTOR64 : RI<0xAE, MRM1m, (outs), (ins opaquemem:$src),
                "fxrstor64\t$src", [(int_x86_fxrstor64 addr:$src)]>,
                PS, Requires<[HasFXSR, In64BitMode]>;
} // Defs = [FPSW, FPCW]
} // SchedRW

//===----------------------------------------------------------------------===//
// Non-Instruction Patterns
//===----------------------------------------------------------------------===//

// Required for RET of f32 / f64 / f80 values.
def : Pat<(X86fldf32 addr:$src), (LD_Fp32m addr:$src)>;
def : Pat<(X86fldf32 addr:$src), (LD_Fp32m64 addr:$src)>;
def : Pat<(X86fldf64 addr:$src), (LD_Fp64m addr:$src)>;
def : Pat<(X86fldf32 addr:$src), (LD_Fp32m80 addr:$src)>;
def : Pat<(X86fldf64 addr:$src), (LD_Fp64m80 addr:$src)>;
def : Pat<(X86fldf80 addr:$src), (LD_Fp80m addr:$src)>;

// Required for CALL which return f32 / f64 / f80 values.
def : Pat<(X86fstf32 RFP32:$src, addr:$op), (ST_Fp32m addr:$op, RFP32:$src)>;
def : Pat<(X86fstf32 RFP64:$src, addr:$op), (ST_Fp64m32 addr:$op, RFP64:$src)>;
def : Pat<(X86fstf64 RFP64:$src, addr:$op), (ST_Fp64m addr:$op, RFP64:$src)>;
def : Pat<(X86fstf32 RFP80:$src, addr:$op), (ST_Fp80m32 addr:$op, RFP80:$src)>;
def : Pat<(X86fstf64 RFP80:$src, addr:$op), (ST_Fp80m64 addr:$op, RFP80:$src)>;
def : Pat<(X86fstf80 RFP80:$src, addr:$op), (ST_FpP80m addr:$op, RFP80:$src)>;

// Floating point constant -0.0 and -1.0
def : Pat<(f32 fpimmneg0), (CHS_Fp32 (LD_Fp032))>, Requires<[FPStackf32]>;
def : Pat<(f32 fpimmneg1), (CHS_Fp32 (LD_Fp132))>, Requires<[FPStackf32]>;
def : Pat<(f64 fpimmneg0), (CHS_Fp64 (LD_Fp064))>, Requires<[FPStackf64]>;
def : Pat<(f64 fpimmneg1), (CHS_Fp64 (LD_Fp164))>, Requires<[FPStackf64]>;
def : Pat<(f80 fpimmneg0), (CHS_Fp80 (LD_Fp080))>;
def : Pat<(f80 fpimmneg1), (CHS_Fp80 (LD_Fp180))>;

// FP extensions map onto simple pseudo-value conversions if they are to/from
// the FP stack.
def : Pat<(f64 (any_fpextend RFP32:$src)), (COPY_TO_REGCLASS RFP32:$src, RFP64)>,
          Requires<[FPStackf32]>;
def : Pat<(f80 (any_fpextend RFP32:$src)), (COPY_TO_REGCLASS RFP32:$src, RFP80)>,
           Requires<[FPStackf32]>;
def : Pat<(f80 (any_fpextend RFP64:$src)), (COPY_TO_REGCLASS RFP64:$src, RFP80)>,
           Requires<[FPStackf64]>;

// FP truncations map onto simple pseudo-value conversions if they are to/from
// the FP stack.  We have validated that only value-preserving truncations make
// it through isel.
def : Pat<(f32 (any_fpround RFP64:$src)), (COPY_TO_REGCLASS RFP64:$src, RFP32)>,
          Requires<[FPStackf32]>;
def : Pat<(f32 (any_fpround RFP80:$src)), (COPY_TO_REGCLASS RFP80:$src, RFP32)>,
           Requires<[FPStackf32]>;
def : Pat<(f64 (any_fpround RFP80:$src)), (COPY_TO_REGCLASS RFP80:$src, RFP64)>,
           Requires<[FPStackf64]>;