Loading...
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 | /* * This file is subject to the terms and conditions of the GNU General Public * License. See the file "COPYING" in the main directory of this archive * for more details. * * Unified implementation of memcpy, memmove and the __copy_user backend. * * Copyright (C) 1998, 1999, 2000 Ralf Baechle * Copyright (C) 1999, 2000 Silicon Graphics, Inc. * * For __rmemcpy and memmove an exception is always a kernel bug, therefore * they're not protected. In order to keep the exception fixup routine * simple all memory accesses in __copy_user to src rsp. dst are stricly * incremental. The fixup routine depends on $at not being changed. */ #include <asm/asm.h> #include <asm/offset.h> #include <asm/regdef.h> /* * The fixup routine for copy_to_user depends on copying strictly in * increasing order. Gas expands the ulw/usw macros in the wrong order for * little endian machines, so we cannot depend on them. */ #ifdef __MIPSEB__ #define uswL swl #define uswU swr #define ulwL lwl #define ulwU lwr #define usdL sdl #define usdU sdr #define uldL ldl #define uldU ldr #endif #ifdef __MIPSEL__ #define uswL swr #define uswU swl #define ulwL lwr #define ulwU lwl #define usdL sdr #define usdU sdl #define uldL ldr #define uldU ldl #endif #define EX(insn,reg,addr,handler) \ 9: insn reg, addr; \ .section __ex_table,"a"; \ PTR 9b, handler; \ .previous #define UEX(insn,reg,addr,handler) \ 9: insn ## L reg, addr; \ 10: insn ## U reg, 3 + addr; \ .section __ex_table,"a"; \ PTR 9b, handler; \ PTR 10b, handler; \ .previous #define UEXD(insn,reg,addr,handler) \ 9: insn ## L reg, addr; \ 10: insn ## U reg, 7 + addr; \ .section __ex_table,"a"; \ PTR 9b, handler; \ PTR 10b, handler; \ .previous /* ascending order, destination aligned */ #define MOVE_BIGGERCHUNK(src, dst, offset, t0, t1, t2, t3) \ EX(ld, t0, (offset + 0x00)(src), l_fixup); \ EX(ld, t1, (offset + 0x08)(src), l_fixup); \ EX(ld, t2, (offset + 0x10)(src), l_fixup); \ EX(ld, t3, (offset + 0x18)(src), l_fixup); \ EX(sd, t0, (offset + 0x00)(dst), s_fixup); \ EX(sd, t1, (offset + 0x08)(dst), s_fixup); \ EX(sd, t2, (offset + 0x10)(dst), s_fixup); \ EX(sd, t3, (offset + 0x18)(dst), s_fixup); \ EX(ld, t0, (offset + 0x20)(src), l_fixup); \ EX(ld, t1, (offset + 0x28)(src), l_fixup); \ EX(ld, t2, (offset + 0x30)(src), l_fixup); \ EX(ld, t3, (offset + 0x38)(src), l_fixup); \ EX(sd, t0, (offset + 0x20)(dst), s_fixup); \ EX(sd, t1, (offset + 0x28)(dst), s_fixup); \ EX(sd, t2, (offset + 0x30)(dst), s_fixup); \ EX(sd, t3, (offset + 0x38)(dst), s_fixup) /* ascending order, destination aligned */ #define MOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \ EX(lw, t0, (offset + 0x00)(src), l_fixup); \ EX(lw, t1, (offset + 0x04)(src), l_fixup); \ EX(lw, t2, (offset + 0x08)(src), l_fixup); \ EX(lw, t3, (offset + 0x0c)(src), l_fixup); \ EX(sw, t0, (offset + 0x00)(dst), s_fixup); \ EX(sw, t1, (offset + 0x04)(dst), s_fixup); \ EX(sw, t2, (offset + 0x08)(dst), s_fixup); \ EX(sw, t3, (offset + 0x0c)(dst), s_fixup); \ EX(lw, t0, (offset + 0x10)(src), l_fixup); \ EX(lw, t1, (offset + 0x14)(src), l_fixup); \ EX(lw, t2, (offset + 0x18)(src), l_fixup); \ EX(lw, t3, (offset + 0x1c)(src), l_fixup); \ EX(sw, t0, (offset + 0x10)(dst), s_fixup); \ EX(sw, t1, (offset + 0x14)(dst), s_fixup); \ EX(sw, t2, (offset + 0x18)(dst), s_fixup); \ EX(sw, t3, (offset + 0x1c)(dst), s_fixup) /* ascending order, destination unaligned */ #define UMOVE_BIGGERCHUNK(src, dst, offset, t0, t1, t2, t3) \ EX(ld, t0, (offset + 0x00)(src), l_fixup); \ EX(ld, t1, (offset + 0x08)(src), l_fixup); \ EX(ld, t2, (offset + 0x10)(src), l_fixup); \ EX(ld, t3, (offset + 0x18)(src), l_fixup); \ UEXD(usd, t0, (offset + 0x00)(dst), s_fixup); \ UEXD(usd, t1, (offset + 0x08)(dst), s_fixup); \ UEXD(usd, t2, (offset + 0x10)(dst), s_fixup); \ UEXD(usd, t3, (offset + 0x18)(dst), s_fixup); \ EX(ld, t0, (offset + 0x20)(src), l_fixup); \ EX(ld, t1, (offset + 0x28)(src), l_fixup); \ EX(ld, t2, (offset + 0x30)(src), l_fixup); \ EX(ld, t3, (offset + 0x38)(src), l_fixup); \ UEXD(usd, t0, (offset + 0x20)(dst), s_fixup); \ UEXD(usd, t1, (offset + 0x28)(dst), s_fixup); \ UEXD(usd, t2, (offset + 0x30)(dst), s_fixup); \ UEXD(usd, t3, (offset + 0x38)(dst), s_fixup) /* ascending order, destination unaligned */ #define UMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \ EX(lw, t0, (offset + 0x00)(src), l_fixup); \ EX(lw, t1, (offset + 0x04)(src), l_fixup); \ EX(lw, t2, (offset + 0x08)(src), l_fixup); \ EX(lw, t3, (offset + 0x0c)(src), l_fixup); \ UEX(usw, t0, (offset + 0x00)(dst), s_fixup); \ UEX(usw, t1, (offset + 0x04)(dst), s_fixup); \ UEX(usw, t2, (offset + 0x08)(dst), s_fixup); \ UEX(usw, t3, (offset + 0x0c)(dst), s_fixup); \ EX(lw, t0, (offset + 0x10)(src), l_fixup); \ EX(lw, t1, (offset + 0x14)(src), l_fixup); \ EX(lw, t2, (offset + 0x18)(src), l_fixup); \ EX(lw, t3, (offset + 0x1c)(src), l_fixup); \ UEX(usw, t0, (offset + 0x10)(dst), s_fixup); \ UEX(usw, t1, (offset + 0x14)(dst), s_fixup); \ UEX(usw, t2, (offset + 0x18)(dst), s_fixup); \ UEX(usw, t3, (offset + 0x1c)(dst), s_fixup) .text .set noreorder .set noat .align 5 LEAF(memcpy) /* a0=dst a1=src a2=len */ move v0, a0 /* return value */ __memcpy: FEXPORT(__copy_user) xor ta0, a0, a1 andi ta0, ta0, 0x7 move t3, a0 beqz ta0, can_align sltiu t8, a2, 0x8 b memcpy_u_src # bad alignment move ta2, a2 can_align: bnez t8, small_memcpy # < 8 bytes to copy move ta2, a2 beqz a2, out andi t8, a1, 0x1 hword_align: beqz t8, word_align andi t8, a1, 0x2 EX(lb, ta0, (a1), l_fixup) dsubu a2, a2, 0x1 EX(sb, ta0, (a0), s_fixup) daddu a1, a1, 0x1 daddu a0, a0, 0x1 andi t8, a1, 0x2 word_align: beqz t8, dword_align sltiu t8, a2, 56 EX(lh, ta0, (a1), l_fixup) dsubu a2, a2, 0x2 EX(sh, ta0, (a0), s_fixup) sltiu t8, a2, 56 daddu a0, a0, 0x2 daddu a1, a1, 0x2 dword_align: bnez t8, do_end_words move t8, a2 andi t8, a1, 0x4 beqz t8, qword_align andi t8, a1, 0x8 EX(lw, ta0, 0x00(a1), l_fixup) dsubu a2, a2, 0x4 EX(sw, ta0, 0x00(a0), s_fixup) daddu a1, a1, 0x4 daddu a0, a0, 0x4 andi t8, a1, 0x8 qword_align: beqz t8, oword_align andi t8, a1, 0x10 EX(lw, ta0, 0x00(a1), l_fixup) EX(lw, ta1, 0x04(a1), l_fixup) dsubu a2, a2, 0x8 EX(sw, ta0, 0x00(a0), s_fixup) EX(sw, ta1, 0x04(a0), s_fixup) daddu a1, a1, 0x8 andi t8, a1, 0x10 daddu a0, a0, 0x8 oword_align: beqz t8, begin_movement srl t8, a2, 0x7 EX(lw, ta3, 0x00(a1), l_fixup) EX(lw, t0, 0x04(a1), l_fixup) EX(lw, ta0, 0x08(a1), l_fixup) EX(lw, ta1, 0x0c(a1), l_fixup) EX(sw, ta3, 0x00(a0), s_fixup) EX(sw, t0, 0x04(a0), s_fixup) EX(sw, ta0, 0x08(a0), s_fixup) EX(sw, ta1, 0x0c(a0), s_fixup) dsubu a2, a2, 0x10 daddu a1, a1, 0x10 srl t8, a2, 0x7 daddu a0, a0, 0x10 begin_movement: beqz t8, 0f andi ta2, a2, 0x40 move_128bytes: #if (_MIPS_ISA == _MIPS_ISA_MIPS4) pref 0, 2*128(a0) pref 1, 2*128(a1) #endif MOVE_BIGGERCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) MOVE_BIGGERCHUNK(a1, a0, 0x40, ta0, ta1, ta3, t0) dsubu t8, t8, 0x01 daddu a1, a1, 0x80 bnez t8, move_128bytes daddu a0, a0, 0x80 0: beqz ta2, 1f andi ta2, a2, 0x20 move_64bytes: MOVE_BIGGERCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) daddu a1, a1, 0x40 daddu a0, a0, 0x40 1: beqz ta2, do_end_words andi t8, a2, 0x1c move_32bytes: MOVE_BIGCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) andi t8, a2, 0x1c daddu a1, a1, 0x20 daddu a0, a0, 0x20 do_end_words: beqz t8, maybe_end_cruft srl t8, t8, 0x2 end_words: EX(lw, ta0, (a1), l_fixup) dsubu t8, t8, 0x1 EX(sw, ta0, (a0), s_fixup) daddu a1, a1, 0x4 bnez t8, end_words daddu a0, a0, 0x4 maybe_end_cruft: andi ta2, a2, 0x3 small_memcpy: beqz ta2, out move a2, ta2 end_bytes: EX(lb, ta0, (a1), l_fixup) dsubu a2, a2, 0x1 EX(sb, ta0, (a0), s_fixup) daddu a1, a1, 0x1 bnez a2, end_bytes daddu a0, a0, 0x1 out: jr ra move a2, zero /* ------------------------------------------------------------------------- */ /* Bad, bad. At least try to align the source */ memcpy_u_src: bnez t8, small_memcpy # < 8 bytes? move ta2, a2 daddiu ta0, a1, 7 # ta0: how much to align ori ta0, 7 xori ta0, 7 dsubu ta0, a1 UEXD(uld, ta1, 0(a1), l_fixup) # dword alignment UEXD(usd, ta1, 0(a0), s_fixup) daddu a1, ta0 # src daddu a0, ta0 # dst dsubu a2, ta0 # len sltiu t8, a2, 56 bnez t8, u_do_end_words andi t8, a2, 0x3c andi t8, a1, 8 # now qword aligned? u_qword_align: beqz t8, u_oword_align andi t8, a1, 0x10 EX(ld, ta0, 0x00(a1), l_fixup) dsubu a2, a2, 0x8 UEXD(usd, ta0, 0x00(a0), s_fixup) daddu a1, a1, 0x8 andi t8, a1, 0x10 daddu a0, a0, 0x8 u_oword_align: beqz t8, u_begin_movement srl t8, a2, 0x7 EX(lw, ta3, 0x08(a1), l_fixup) EX(lw, t0, 0x0c(a1), l_fixup) EX(lw, ta0, 0x00(a1), l_fixup) EX(lw, ta1, 0x04(a1), l_fixup) UEX(usw, ta3, 0x08(a0), s_fixup) UEX(usw, t0, 0x0c(a0), s_fixup) UEX(usw, ta0, 0x00(a0), s_fixup) UEX(usw, ta1, 0x04(a0), s_fixup) dsubu a2, a2, 0x10 daddu a1, a1, 0x10 srl t8, a2, 0x7 daddu a0, a0, 0x10 u_begin_movement: beqz t8, 0f andi ta2, a2, 0x40 u_move_128bytes: UMOVE_BIGGERCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) UMOVE_BIGGERCHUNK(a1, a0, 0x40, ta0, ta1, ta3, t0) dsubu t8, t8, 0x01 daddu a1, a1, 0x80 bnez t8, u_move_128bytes daddu a0, a0, 0x80 0: beqz ta2, 1f andi ta2, a2, 0x20 u_move_64bytes: UMOVE_BIGGERCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) daddu a1, a1, 0x40 daddu a0, a0, 0x40 1: beqz ta2, u_do_end_words andi t8, a2, 0x1c u_move_32bytes: UMOVE_BIGCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) andi t8, a2, 0x1c daddu a1, a1, 0x20 daddu a0, a0, 0x20 u_do_end_words: beqz t8, u_maybe_end_cruft srl t8, t8, 0x2 u_end_words: EX(lw, ta0, 0x00(a1), l_fixup) dsubu t8, t8, 0x1 UEX(usw, ta0, 0x00(a0), s_fixup) daddu a1, a1, 0x4 bnez t8, u_end_words daddu a0, a0, 0x4 u_maybe_end_cruft: andi ta2, a2, 0x3 u_cannot_optimize: beqz ta2, out move a2, ta2 u_end_bytes: EX(lb, ta0, (a1), l_fixup) dsubu a2, a2, 0x1 EX(sb, ta0, (a0), s_fixup) daddu a1, a1, 0x1 bnez a2, u_end_bytes daddu a0, a0, 0x1 jr ra move a2, zero END(memcpy) /* descending order, destination aligned */ #define RMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \ lw t0, (offset + 0x10)(src); \ lw t1, (offset + 0x14)(src); \ lw t2, (offset + 0x18)(src); \ lw t3, (offset + 0x1c)(src); \ sw t0, (offset + 0x10)(dst); \ sw t1, (offset + 0x14)(dst); \ sw t2, (offset + 0x18)(dst); \ sw t3, (offset + 0x1c)(dst); \ lw t0, (offset + 0x00)(src); \ lw t1, (offset + 0x04)(src); \ lw t2, (offset + 0x08)(src); \ lw t3, (offset + 0x0c)(src); \ sw t0, (offset + 0x00)(dst); \ sw t1, (offset + 0x04)(dst); \ sw t2, (offset + 0x08)(dst); \ sw t3, (offset + 0x0c)(dst) /* descending order, destination ununaligned */ #define RUMOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3) \ lw t0, (offset + 0x10)(src); \ lw t1, (offset + 0x14)(src); \ lw t2, (offset + 0x18)(src); \ lw t3, (offset + 0x1c)(src); \ usw t0, (offset + 0x10)(dst); \ usw t1, (offset + 0x14)(dst); \ usw t2, (offset + 0x18)(dst); \ usw t3, (offset + 0x1c)(dst); \ lw t0, (offset + 0x00)(src); \ lw t1, (offset + 0x04)(src); \ lw t2, (offset + 0x08)(src); \ lw t3, (offset + 0x0c)(src); \ usw t0, (offset + 0x00)(dst); \ usw t1, (offset + 0x04)(dst); \ usw t2, (offset + 0x08)(dst); \ usw t3, (offset + 0x0c)(dst) .align 5 LEAF(memmove) sltu ta0, a0, a1 # dst < src -> memcpy bnez ta0, memcpy daddu v0, a0, a2 sltu ta0, v0, a1 # dst + len < src -> non- bnez ta0, __memcpy # overlapping, can use memcpy move v0, a0 /* return value */ beqz a2, r_out END(memmove) LEAF(__rmemcpy) /* a0=dst a1=src a2=len */ daddu a0, a2 # dst = dst + len daddu a1, a2 # src = src + len #if 0 /* Horror fix */ xor ta0, a0, a1 andi ta0, ta0, 0x3 move t3, a0 beqz ta0, r_can_align sltiu t8, a2, 0x8 b r_memcpy_u_src # bad alignment move ta2, a2 r_can_align: bnez t8, r_small_memcpy # < 8 bytes to copy move ta2, a2 beqz a2, r_out andi t8, a1, 0x1 r_hword_align: beqz t8, r_word_align andi t8, a1, 0x2 lb ta0, -1(a1) dsubu a2, a2, 0x1 sb ta0, -1(a0) dsubu a1, a1, 0x1 dsubu a0, a0, 0x1 andi t8, a1, 0x2 r_word_align: beqz t8, r_dword_align sltiu t8, a2, 56 lh ta0, -2(a1) dsubu a2, a2, 0x2 sh ta0, -2(a0) sltiu t8, a2, 56 dsubu a0, a0, 0x2 dsubu a1, a1, 0x2 r_dword_align: bnez t8, r_do_end_words move t8, a2 andi t8, a1, 0x4 beqz t8, r_qword_align andi t8, a1, 0x8 lw ta0, -4(a1) dsubu a2, a2, 0x4 sw ta0, -4(a0) dsubu a1, a1, 0x4 dsubu a0, a0, 0x4 andi t8, a1, 0x8 r_qword_align: beqz t8, r_oword_align andi t8, a1, 0x10 dsubu a1, a1, 0x8 lw ta0, 0x04(a1) lw ta1, 0x00(a1) dsubu a0, a0, 0x8 sw ta0, 0x04(a0) sw ta1, 0x00(a0) dsubu a2, a2, 0x8 andi t8, a1, 0x10 r_oword_align: beqz t8, r_begin_movement srl t8, a2, 0x7 dsubu a1, a1, 0x10 lw ta3, 0x08(a1) # assumes subblock ordering lw t0, 0x0c(a1) lw ta0, 0x00(a1) lw ta1, 0x04(a1) dsubu a0, a0, 0x10 sw ta3, 0x08(a0) sw t0, 0x0c(a0) sw ta0, 0x00(a0) sw ta1, 0x04(a0) dsubu a2, a2, 0x10 srl t8, a2, 0x7 r_begin_movement: beqz t8, 0f andi ta2, a2, 0x40 r_move_128bytes: RMOVE_BIGCHUNK(a1, a0, -0x80, ta0, ta1, ta3, t0) RMOVE_BIGCHUNK(a1, a0, -0x60, ta0, ta1, ta3, t0) RMOVE_BIGCHUNK(a1, a0, -0x40, ta0, ta1, ta3, t0) RMOVE_BIGCHUNK(a1, a0, -0x20, ta0, ta1, ta3, t0) dsubu t8, t8, 0x01 dsubu a1, a1, 0x80 bnez t8, r_move_128bytes dsubu a0, a0, 0x80 0: beqz ta2, 1f andi ta2, a2, 0x20 r_move_64bytes: dsubu a1, a1, 0x40 dsubu a0, a0, 0x40 RMOVE_BIGCHUNK(a1, a0, 0x20, ta0, ta1, ta3, t0) RMOVE_BIGCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) 1: beqz ta2, r_do_end_words andi t8, a2, 0x1c r_move_32bytes: dsubu a1, a1, 0x20 dsubu a0, a0, 0x20 RMOVE_BIGCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) andi t8, a2, 0x1c r_do_end_words: beqz t8, r_maybe_end_cruft srl t8, t8, 0x2 r_end_words: lw ta0, -4(a1) dsubu t8, t8, 0x1 sw ta0, -4(a0) dsubu a1, a1, 0x4 bnez t8, r_end_words dsubu a0, a0, 0x4 r_maybe_end_cruft: andi ta2, a2, 0x3 r_small_memcpy: beqz ta2, r_out move a2, ta2 #endif /* Horror fix */ r_end_bytes: lb ta0, -1(a1) dsubu a2, a2, 0x1 sb ta0, -1(a0) dsubu a1, a1, 0x1 bnez a2, r_end_bytes dsubu a0, a0, 0x1 r_out: jr ra move a2, zero #if 0 /* Horror fix */ /* ------------------------------------------------------------------------- */ /* Bad, bad. At least try to align the source */ r_memcpy_u_src: bnez t8, r_small_memcpy # < 8 bytes? move ta2, a2 andi ta0, a1, 7 # ta0: how much to align ulw ta1, -8(a1) # dword alignment ulw ta2, -4(a1) usw ta1, -8(a0) usw ta2, -4(a0) dsubu a1, ta0 # src dsubu a0, ta0 # dst dsubu a2, ta0 # len sltiu t8, a2, 56 bnez t8, ru_do_end_words andi t8, a2, 0x3c andi t8, a1, 8 # now qword aligned? ru_qword_align: beqz t8, ru_oword_align andi t8, a1, 0x10 dsubu a1, a1, 0x8 lw ta0, 0x00(a1) lw ta1, 0x04(a1) dsubu a0, a0, 0x8 usw ta0, 0x00(a0) usw ta1, 0x04(a0) dsubu a2, a2, 0x8 andi t8, a1, 0x10 ru_oword_align: beqz t8, ru_begin_movement srl t8, a2, 0x7 dsubu a1, a1, 0x10 lw ta3, 0x08(a1) # assumes subblock ordering lw t0, 0x0c(a1) lw ta0, 0x00(a1) lw ta1, 0x04(a1) dsubu a0, a0, 0x10 usw ta3, 0x08(a0) usw t0, 0x0c(a0) usw ta0, 0x00(a0) usw ta1, 0x04(a0) dsubu a2, a2, 0x10 srl t8, a2, 0x7 ru_begin_movement: beqz t8, 0f andi ta2, a2, 0x40 ru_move_128bytes: RUMOVE_BIGCHUNK(a1, a0, -0x80, ta0, ta1, ta3, t0) RUMOVE_BIGCHUNK(a1, a0, -0x60, ta0, ta1, ta3, t0) RUMOVE_BIGCHUNK(a1, a0, -0x40, ta0, ta1, ta3, t0) RUMOVE_BIGCHUNK(a1, a0, -0x20, ta0, ta1, ta3, t0) dsubu t8, t8, 0x01 dsubu a1, a1, 0x80 bnez t8, ru_move_128bytes dsubu a0, a0, 0x80 0: beqz ta2, 1f andi ta2, a2, 0x20 ru_move_64bytes: dsubu a1, a1, 0x40 dsubu a0, a0, 0x40 RUMOVE_BIGCHUNK(a1, a0, 0x20, ta0, ta1, ta3, t0) RUMOVE_BIGCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) 1: beqz ta2, ru_do_end_words andi t8, a2, 0x1c ru_move_32bytes: dsubu a1, a1, 0x20 dsubu a0, a0, 0x20 RUMOVE_BIGCHUNK(a1, a0, 0x00, ta0, ta1, ta3, t0) andi t8, a2, 0x1c ru_do_end_words: beqz t8, ru_maybe_end_cruft srl t8, t8, 0x2 ru_end_words: lw ta0, -4(a1) usw ta0, -4(a0) dsubu t8, t8, 0x1 dsubu a1, a1, 0x4 bnez t8, ru_end_words dsubu a0, a0, 0x4 ru_maybe_end_cruft: andi ta2, a2, 0x3 ru_cannot_optimize: beqz ta2, r_out move a2, ta2 ru_end_bytes: lb ta0, -1(a1) dsubu a2, a2, 0x1 sb ta0, -1(a0) dsubu a1, a1, 0x1 bnez a2, ru_end_bytes dsubu a0, a0, 0x1 jr ra move a2, zero #endif /* Horror fix */ END(__rmemcpy) l_fixup: # clear the rest of the buffer ld ta0, THREAD_BUADDR($28) nop dsubu a2, AT, ta0 # a2 bytes to go daddu a0, ta0 # compute start address in a1 dsubu a0, a1 j __bzero move a1, zero s_fixup: jr ra nop |