insns.pl

   1 #!/usr/bin/perl
   2 ## --------------------------------------------------------------------------
   3 ##
   4 ##   Copyright 1996-2010 The NASM Authors - All Rights Reserved
   5 ##   See the file AUTHORS included with the NASM distribution for
   6 ##   the specific copyright holders.
   7 ##
   8 ##   Redistribution and use in source and binary forms, with or without
   9 ##   modification, are permitted provided that the following
  10 ##   conditions are met:
  11 ##
  12 ##   * Redistributions of source code must retain the above copyright
  13 ##     notice, this list of conditions and the following disclaimer.
  14 ##   * Redistributions in binary form must reproduce the above
  15 ##     copyright notice, this list of conditions and the following
  16 ##     disclaimer in the documentation and/or other materials provided
  17 ##     with the distribution.
  18 ##
  19 ##     THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
  20 ##     CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
  21 ##     INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
  22 ##     MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
  23 ##     DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
  24 ##     CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
  25 ##     SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
  26 ##     NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  27 ##     LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  28 ##     HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
  29 ##     CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
  30 ##     OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
  31 ##     EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  32 ##
  33 ## --------------------------------------------------------------------------
  34
  35 #
  36 # insns.pl
  37 #
  38 # Parse insns.dat and produce generated source code files
  39
  40 # Opcode prefixes which need their own opcode tables
  41 # LONGER PREFIXES FIRST!
  42 @disasm_prefixes = qw(0F24 0F25 0F38 0F3A 0F7A 0FA6 0FA7 0F);
  43
  44 # This should match MAX_OPERANDS from nasm.h
  45 $MAX_OPERANDS = 5;
  46
  47 # Add VEX/XOP prefixes
  48 @vex_class = ( 'vex', 'xop' );
  49 $vex_classes = scalar(@vex_class);
  50 @vexlist = ();
  51 %vexmap = ();
  52 for ($c = 0; $c < $vex_classes; $c++) {
  53     $vexmap{$vex_class[$c]} = $c;
  54     for ($m = 0; $m < 32; $m++) {
  55         for ($p = 0; $p < 4; $p++) {
  56             push(@vexlist, sprintf("%s%02X%01X", $vex_class[$c], $m, $p));
  57         }
  58     }
  59 }
  60 @disasm_prefixes = (@vexlist, @disasm_prefixes);
  61
  62 @bytecode_count = (0) x 256;
  63
  64 print STDERR "Reading insns.dat...\n";
  65
  66 @args   = ();
  67 undef $output;
  68 foreach $arg ( @ARGV ) {
  69     if ( $arg =~ /^\-/ ) {
  70         if  ( $arg =~ /^\-([abdin])$/ ) {
  71             $output = $1;
  72         } else {
  73             die "$0: Unknown option: ${arg}\n";
  74         }
  75     } else {
  76         push (@args, $arg);
  77     }
  78 }
  79
  80 $fname = "insns.dat" unless $fname = $args[0];
  81 open (F, $fname) || die "unable to open $fname";
  82
  83 %dinstables = ();
  84 @bytecode_list = ();
  85
  86 $line = 0;
  87 $insns = 0;
  88 while (<F>) {
  89     $line++;
  90     chomp;
  91     next if ( /^\s*(\;.*|)$/ );   # comments or blank lines
  92
  93     unless (/^\s*(\S+)\s+(\S+)\s+(\S+|\[.*\])\s+(\S+)\s*$/) {
  94         warn "line $line does not contain four fields\n";
  95         next;
  96     }
  97     @fields = ($1, $2, $3, $4);
  98     @field_list = ([@fields, 0]);
  99
 100     if ($fields[1] =~ /\*/) {
 101         # This instruction has relaxed form(s)
 102         if ($fields[2] !~ /^\[/) {
 103             warn "line $line has an * operand but uses raw bytecodes\n";
 104             next;
 105         }
 106
 107         $opmask = 0;
 108         @ops = split(/,/, $fields[1]);
 109         for ($oi = 0; $oi < scalar @ops; $oi++) {
 110             if ($ops[$oi] =~ /\*$/) {
 111                 if ($oi == 0) {
 112                     warn "line $line has a first operand with a *\n";
 113                     next;
 114                 }
 115                 $opmask |= 1 << $oi;
 116             }
 117         }
 118
 119         for ($oi = 1; $oi < (1 << scalar @ops); $oi++) {
 120             if (($oi & ~$opmask) == 0) {
 121                 my @xops = ();
 122                 my $omask = ~$oi;
 123                 for ($oj = 0; $oj < scalar(@ops); $oj++) {
 124                     if ($omask & 1) {
 125                         push(@xops, $ops[$oj]);
 126                     }
 127                     $omask >>= 1;
 128                 }
 129                 push(@field_list, [$fields[0], join(',', @xops),
 130                      $fields[2], $fields[3], $oi]);
 131             }
 132         }
 133     }
 134
 135     foreach $fptr (@field_list) {
 136         @fields = @$fptr;
 137         ($formatted, $nd) = format_insn(@fields);
 138         if ($formatted) {
 139             $insns++;
 140             $aname = "aa_$fields[0]";
 141             push @$aname, $formatted;
 142         }
 143         if ( $fields[0] =~ /cc$/ ) {
 144             # Conditional instruction
 145             $k_opcodes_cc{$fields[0]}++;
 146         } else {
 147             # Unconditional instruction
 148             $k_opcodes{$fields[0]}++;
 149         }
 150         if ($formatted && !$nd) {
 151             push @big, $formatted;
 152             my @sseq = startseq($fields[2], $fields[4]);
 153             foreach $i (@sseq) {
 154                 if (!defined($dinstables{$i})) {
 155                     $dinstables{$i} = [];
 156                 }
 157                 push(@{$dinstables{$i}}, $#big);
 158             }
 159         }
 160     }
 161 }
 162
 163 close F;
 164
 165 #
 166 # Generate the bytecode array.  At this point, @bytecode_list contains
 167 # the full set of bytecodes.
 168 #
 169
 170 # Sort by descending length
 171 @bytecode_list = sort { scalar(@$b) <=> scalar(@$a) } @bytecode_list;
 172 @bytecode_array = ();
 173 %bytecode_pos = ();
 174 $bytecode_next = 0;
 175 foreach $bl (@bytecode_list) {
 176     my $h = hexstr(@$bl);
 177     next if (defined($bytecode_pos{$h}));
 178
 179     push(@bytecode_array, $bl);
 180     while ($h ne '') {
 181         $bytecode_pos{$h} = $bytecode_next;
 182         $h = substr($h, 2);
 183         $bytecode_next++;
 184     }
 185 }
 186 undef @bytecode_list;
 187
 188 @opcodes    = sort keys(%k_opcodes);
 189 @opcodes_cc = sort keys(%k_opcodes_cc);
 190
 191 if ( !defined($output) || $output eq 'b') {
 192     print STDERR "Writing insnsb.c...\n";
 193
 194     open B, ">insnsb.c";
 195
 196     print B "/* This file auto-generated from insns.dat by insns.pl" .
 197         " - don't edit it */\n\n";
 198
 199     print B "#include \"nasm.h\"\n";
 200     print B "#include \"insns.h\"\n\n";
 201
 202     print B "const uint8_t nasm_bytecodes[$bytecode_next] = {\n";
 203
 204     $p = 0;
 205     foreach $bl (@bytecode_array) {
 206         printf B "    /* %5d */ ", $p;
 207         foreach $d (@$bl) {
 208             printf B "%#o,", $d;
 209             $p++;
 210         }
 211         printf B "\n";
 212     }
 213     print B "};\n";
 214
 215     print B "\n";
 216     print B "/*\n";
 217     print B " * Bytecode frequencies (including reuse):\n";
 218     print B " *\n";
 219     for ($i = 0; $i < 32; $i++) {
 220         print B " *";
 221         for ($j = 0; $j < 256; $j += 32) {
 222             print B " |" if ($j);
 223             printf B " %3o:%4d", $i+$j, $bytecode_count[$i+$j];
 224         }
 225         print B "\n";
 226     }
 227     print B " */\n";
 228
 229     close B;
 230 }
 231
 232 if ( !defined($output) || $output eq 'a' ) {
 233     print STDERR "Writing insnsa.c...\n";
 234
 235     open A, ">insnsa.c";
 236
 237     print A "/* This file auto-generated from insns.dat by insns.pl" .
 238         " - don't edit it */\n\n";
 239
 240     print A "#include \"nasm.h\"\n";
 241     print A "#include \"insns.h\"\n\n";
 242
 243     foreach $i (@opcodes, @opcodes_cc) {
 244         print A "static const struct itemplate instrux_${i}[] = {\n";
 245         $aname = "aa_$i";
 246         foreach $j (@$aname) {
 247             print A "    ", codesubst($j), "\n";
 248         }
 249         print A "    ITEMPLATE_END\n};\n\n";
 250     }
 251     print A "const struct itemplate * const nasm_instructions[] = {\n";
 252     foreach $i (@opcodes, @opcodes_cc) {
 253         print A "    instrux_${i},\n";
 254     }
 255     print A "};\n";
 256
 257     close A;
 258 }
 259
 260 if ( !defined($output) || $output eq 'd' ) {
 261     print STDERR "Writing insnsd.c...\n";
 262
 263     open D, ">insnsd.c";
 264
 265     print D "/* This file auto-generated from insns.dat by insns.pl" .
 266         " - don't edit it */\n\n";
 267
 268     print D "#include \"nasm.h\"\n";
 269     print D "#include \"insns.h\"\n\n";
 270
 271     print D "static const struct itemplate instrux[] = {\n";
 272     $n = 0;
 273     foreach $j (@big) {
 274         printf D "    /* %4d */ %s\n", $n++, codesubst($j);
 275     }
 276     print D "};\n";
 277
 278     foreach $h (sort(keys(%dinstables))) {
 279         next if ($h eq ''); # Skip pseudo-instructions
 280             print D "\nstatic const struct itemplate * const itable_${h}[] = {\n";
 281         foreach $j (@{$dinstables{$h}}) {
 282             print D "    instrux + $j,\n";
 283         }
 284         print D "};\n";
 285     }
 286
 287     @prefix_list = ();
 288     foreach $h (@disasm_prefixes, '') {
 289         for ($c = 0; $c < 256; $c++) {
 290             $nn = sprintf("%s%02X", $h, $c);
 291             if ($is_prefix{$nn} || defined($dinstables{$nn})) {
 292                 # At least one entry in this prefix table
 293                 push(@prefix_list, $h);
 294                 $is_prefix{$h} = 1;
 295                 last;
 296             }
 297         }
 298     }
 299
 300     foreach $h (@prefix_list) {
 301         print D "\n";
 302         print D "static " unless ($h eq '');
 303         print D "const struct disasm_index ";
 304         print D ($h eq '') ? 'itable' : "itable_$h";
 305         print D "[256] = {\n";
 306         for ($c = 0; $c < 256; $c++) {
 307             $nn = sprintf("%s%02X", $h, $c);
 308             if ($is_prefix{$nn}) {
 309                 die "$fname: ambiguous decoding of $nn\n"
 310                     if (defined($dinstables{$nn}));
 311                 printf D "    /* 0x%02x */ { itable_%s, -1 },\n", $c, $nn;
 312             } elsif (defined($dinstables{$nn})) {
 313                 printf D "    /* 0x%02x */ { itable_%s, %u },\n", $c,
 314                        $nn, scalar(@{$dinstables{$nn}});
 315             } else {
 316                 printf D "    /* 0x%02x */ { NULL, 0 },\n", $c;
 317             }
 318         }
 319         print D "};\n";
 320     }
 321
 322     printf D "\nconst struct disasm_index * const itable_vex[%d][32][4] =\n",
 323         $vex_classes;
 324     print D "{\n";
 325     for ($c = 0; $c < $vex_classes; $c++) {
 326         print D "    {\n";
 327         for ($m = 0; $m < 32; $m++) {
 328             print D "        { ";
 329             for ($p = 0; $p < 4; $p++) {
 330                 $vp = sprintf("%s%02X%01X", $vex_class[$c], $m, $p);
 331                 printf D "%-15s",
 332                        ($is_prefix{$vp} ? sprintf("itable_%s,", $vp) : 'NULL,');
 333             }
 334             print D "},\n";
 335         }
 336         print D "    },\n";
 337     }
 338     print D "};\n";
 339
 340     close D;
 341 }
 342
 343 if ( !defined($output) || $output eq 'i' ) {
 344     print STDERR "Writing insnsi.h...\n";
 345
 346     open I, ">insnsi.h";
 347
 348     print I "/* This file is auto-generated from insns.dat by insns.pl" .
 349         " - don't edit it */\n\n";
 350     print I "/* This file in included by nasm.h */\n\n";
 351
 352     print I "/* Instruction names */\n\n";
 353     print I "#ifndef NASM_INSNSI_H\n";
 354     print I "#define NASM_INSNSI_H 1\n\n";
 355     print I "enum opcode {\n";
 356     $maxlen = 0;
 357     foreach $i (@opcodes, @opcodes_cc) {
 358         print I "\tI_${i},\n";
 359         $len = length($i);
 360         $len++ if ( $i =~ /cc$/ ); # Condition codes can be 3 characters long
 361         $maxlen = $len if ( $len > $maxlen );
 362     }
 363     print I "\tI_none = -1\n";
 364     print I "};\n\n";
 365     print I "#define MAX_INSLEN ", $maxlen, "\n";
 366     print I "#define FIRST_COND_OPCODE I_", $opcodes_cc[0], "\n\n";
 367     print I "#endif /* NASM_INSNSI_H */\n";
 368
 369     close I;
 370 }
 371
 372 if ( !defined($output) || $output eq 'n' ) {
 373     print STDERR "Writing insnsn.c...\n";
 374
 375     open N, ">insnsn.c";
 376
 377     print N "/* This file is auto-generated from insns.dat by insns.pl" .
 378         " - don't edit it */\n\n";
 379     print N "#include \"tables.h\"\n\n";
 380
 381     print N "const char * const nasm_insn_names[] = {";
 382     $first = 1;
 383     foreach $i (@opcodes, @opcodes_cc) {
 384         print N "," if ( !$first );
 385         $first = 0;
 386         $ilower = $i;
 387         $ilower =~ s/cc$//;             # Remove conditional cc suffix
 388         $ilower =~ tr/A-Z/a-z/;         # Change to lower case (Perl 4 compatible)
 389         print N "\n\t\"${ilower}\"";
 390     }
 391     print N "\n};\n";
 392     close N;
 393 }
 394
 395 printf STDERR "Done: %d instructions\n", $insns;
 396
 397 # Count primary bytecodes, for statistics
 398 sub count_bytecodes(@) {
 399     my $skip = 0;
 400     foreach my $bc (@_) {
 401         if ($skip) {
 402             $skip--;
 403             next;
 404         }
 405         $bytecode_count[$bc]++;
 406         if ($bc >= 01 && $bc <= 04) {
 407             $skip = $bc;
 408         } elsif (($bc & ~03) == 010) {
 409             $skip = 1;
 410         } elsif (($bc & ~013) == 0144) {
 411             $skip = 1;
 412         } elsif ($bc == 0172) {
 413             $skip = 1;
 414         } elsif ($bc >= 0260 && $bc <= 0270) {
 415             $skip = 2;
 416         } elsif ($bc == 0330) {
 417             $skip = 1;
 418         }
 419     }
 420 }
 421
 422 sub format_insn($$$$$) {
 423     my ($opcode, $operands, $codes, $flags, $relax) = @_;
 424     my $num, $nd = 0;
 425     my @bytecode;
 426     my $op, @ops, $opp, @opx, @oppx;
 427
 428     return (undef, undef) if $operands eq "ignore";
 429
 430     # format the operands
 431     $operands =~ s/\*//g;
 432     $operands =~ s/:/|colon,/g;
 433     @ops = ();
 434     if ($operands ne 'void') {
 435         foreach $op (split(/,/, $operands)) {
 436             if ($op =~ /^\=([0-9]+)$/) {
 437                 $op = "same_as|$1";
 438             } else {
 439                 @opx = ();
 440                 foreach $opp (split(/\|/, $op)) {
 441                     @oppx = ();
 442                     if ($opp =~ /^(.*[^\d])(8|16|32|64|80|128|256)$/) {
 443                         my $ox = $1;
 444                         my $on = $2;
 445                         if ($ox !~ /^(sbyte|sdword|udword)$/) {
 446                             $opp = $ox;
 447                             push(@oppx, "bits$on");
 448                         }
 449                     }
 450                     $opp =~ s/^mem$/memory/;
 451                     $opp =~ s/^memory_offs$/mem_offs/;
 452                     $opp =~ s/^imm$/immediate/;
 453                     $opp =~ s/^([a-z]+)rm$/rm_$1/;
 454                     $opp =~ s/^rm$/rm_gpr/;
 455                     $opp =~ s/^reg$/reg_gpr/;
 456                     push(@opx, $opp, @oppx);
 457                 }
 458                 $op = join('|', @opx);
 459             }
 460             push(@ops, $op);
 461         }
 462     }
 463
 464     $num = scalar(@ops);
 465     while (scalar(@ops) < $MAX_OPERANDS) {
 466         push(@ops, '0');
 467     }
 468     $operands = join(',', @ops);
 469     $operands =~ tr/a-z/A-Z/;
 470
 471     # format the flags
 472     $flags =~ s/,/|IF_/g;
 473     $flags =~ s/(\|IF_ND|IF_ND\|)//, $nd = 1 if $flags =~ /IF_ND/;
 474     $flags = "IF_" . $flags;
 475
 476     @bytecode = (decodify($codes, $relax), 0);
 477     push(@bytecode_list, [@bytecode]);
 478     $codes = hexstr(@bytecode);
 479     count_bytecodes(@bytecode);
 480
 481     ("{I_$opcode, $num, {$operands}, \@\@CODES-$codes\@\@, $flags},", $nd);
 482 }
 483
 484 #
 485 # Look for @@CODES-xxx@@ sequences and replace them with the appropriate
 486 # offset into nasm_bytecodes
 487 #
 488 sub codesubst($) {
 489     my($s) = @_;
 490     my $n;
 491
 492     while ($s =~ /\@\@CODES-([0-9A-F]+)\@\@/) {
 493         my $pos = $bytecode_pos{$1};
 494         if (!defined($pos)) {
 495             die "$fname: no position assigned to byte code $1\n";
 496         }
 497         $s = $` . "nasm_bytecodes+${pos}" . "$'";
 498     }
 499     return $s;
 500 }
 501
 502 sub addprefix ($@) {
 503     my ($prefix, @list) = @_;
 504     my $x;
 505     my @l = ();
 506
 507     foreach $x (@list) {
 508         push(@l, sprintf("%s%02X", $prefix, $x));
 509     }
 510
 511     return @l;
 512 }
 513
 514 #
 515 # Turn a code string into a sequence of bytes
 516 #
 517 sub decodify($$) {
 518   # Although these are C-syntax strings, by convention they should have
 519   # only octal escapes (for directives) and hexadecimal escapes
 520   # (for verbatim bytes)
 521     my($codestr, $relax) = @_;
 522
 523     if ($codestr =~ /^\s*\[([^\]]*)\]\s*$/) {
 524         return byte_code_compile($1, $relax);
 525     }
 526
 527     my $c = $codestr;
 528     my @codes = ();
 529
 530     while ($c ne '') {
 531         if ($c =~ /^\\x([0-9a-f]+)(.*)$/i) {
 532             push(@codes, hex $1);
 533             $c = $2;
 534             next;
 535         } elsif ($c =~ /^\\([0-7]{1,3})(.*)$/) {
 536             push(@codes, oct $1);
 537             $c = $2;
 538             next;
 539         } else {
 540             die "$fname: unknown code format in \"$codestr\"\n";
 541         }
 542     }
 543
 544     return @codes;
 545 }
 546
 547 # Turn a numeric list into a hex string
 548 sub hexstr(@) {
 549     my $s = '';
 550     my $c;
 551
 552     foreach $c (@_) {
 553         $s .= sprintf("%02X", $c);
 554     }
 555     return $s;
 556 }
 557
 558 # Here we determine the range of possible starting bytes for a given
 559 # instruction. We need only consider the codes:
 560 # \[1234]      mean literal bytes, of course
 561 # \1[0123]     mean byte plus register value
 562 # \330         means byte plus condition code
 563 # \0 or \340   mean give up and return empty set
 564 # \34[4567]    mean PUSH/POP of segment registers: special case
 565 # \17[234]     skip is4 control byte
 566 # \26x \270    skip VEX control bytes
 567 sub startseq($$) {
 568     my ($codestr, $relax) = @_;
 569     my $word, @range;
 570     my @codes = ();
 571     my $c = $codestr;
 572     my $c0, $c1, $i;
 573     my $prefix = '';
 574
 575     @codes = decodify($codestr, $relax);
 576
 577     while ($c0 = shift(@codes)) {
 578         $c1 = $codes[0];
 579         if ($c0 >= 01 && $c0 <= 04) {
 580             # Fixed byte string
 581             my $fbs = $prefix;
 582             while (1) {
 583                 if ($c0 >= 01 && $c0 <= 04) {
 584                     while ($c0--) {
 585                         $fbs .= sprintf("%02X", shift(@codes));
 586                     }
 587                 } else {
 588                     last;
 589                 }
 590                 $c0 = shift(@codes);
 591             }
 592
 593             foreach $pfx (@disasm_prefixes) {
 594                 if (substr($fbs, 0, length($pfx)) eq $pfx) {
 595                     $prefix = $pfx;
 596                     $fbs = substr($fbs, length($pfx));
 597                     last;
 598                 }
 599             }
 600
 601             if ($fbs ne '') {
 602                 return ($prefix.substr($fbs,0,2));
 603             }
 604
 605             unshift(@codes, $c0);
 606         } elsif ($c0 >= 010 && $c0 <= 013) {
 607             return addprefix($prefix, $c1..($c1+7));
 608         } elsif (($c0 & ~013) == 0144) {
 609             return addprefix($prefix, $c1, $c1|2);
 610         } elsif ($c0 == 0330) {
 611             return addprefix($prefix, $c1..($c1+15));
 612         } elsif ($c0 == 0 || $c0 == 0340) {
 613             return $prefix;
 614         } elsif ($c0 == 0344) {
 615             return addprefix($prefix, 0x06, 0x0E, 0x16, 0x1E);
 616         } elsif ($c0 == 0345) {
 617             return addprefix($prefix, 0x07, 0x17, 0x1F);
 618         } elsif ($c0 == 0346) {
 619             return addprefix($prefix, 0xA0, 0xA8);
 620         } elsif ($c0 == 0347) {
 621             return addprefix($prefix, 0xA1, 0xA9);
 622         } elsif (($c0 & ~3) == 0260 || $c0 == 0270) {
 623             my $c,$m,$wlp;
 624             $m   = shift(@codes);
 625             $wlp = shift(@codes);
 626             $c = ($m >> 6);
 627             $m = $m & 31;
 628             $prefix .= sprintf('%s%02X%01X', $vex_class[$c], $m, $wlp & 3);
 629         } elsif ($c0 >= 0172 && $c0 <= 174) {
 630             shift(@codes);      # Skip is4 control byte
 631         } else {
 632             # We really need to be able to distinguish "forbidden"
 633             # and "ignorable" codes here
 634         }
 635     }
 636     return $prefix;
 637 }
 638
 639 #
 640 # This function takes a series of byte codes in a format which is more
 641 # typical of the Intel documentation, and encode it.
 642 #
 643 # The format looks like:
 644 #
 645 # [operands: opcodes]
 646 #
 647 # The operands word lists the order of the operands:
 648 #
 649 # r = register field in the modr/m
 650 # m = modr/m
 651 # v = VEX "v" field
 652 # d = DREX "dst" field
 653 # i = immediate
 654 # s = register field of is4/imz2 field
 655 # - = implicit (unencoded) operand
 656 #
 657 # For an operand that should be filled into more than one field,
 658 # enter it as e.g. "r+v".
 659 #
 660 sub byte_code_compile($$) {
 661     my($str, $relax) = @_;
 662     my $opr;
 663     my $opc;
 664     my @codes = ();
 665     my $litix = undef;
 666     my %oppos = ();
 667     my $i;
 668     my $op, $oq;
 669     my $opex;
 670
 671     unless ($str =~ /^(([^\s:]*)\:|)\s*(.*\S)\s*$/) {
 672         die "$fname: $line: cannot parse: [$str]\n";
 673     }
 674     $opr = "\L$2";
 675     $opc = "\L$3";
 676
 677     my $op = 0;
 678     for ($i = 0; $i < length($opr); $i++) {
 679         my $c = substr($opr,$i,1);
 680         if ($c eq '+') {
 681             $op--;
 682         } else {
 683             if ($relax & 1) {
 684                 $op--;
 685             }
 686             $relax >>= 1;
 687             $oppos{$c} = $op++;
 688         }
 689     }
 690
 691     $prefix_ok = 1;
 692     foreach $op (split(/\s*(?:\s|(?=[\/\\]))/, $opc)) {
 693         if ($op eq 'o16') {
 694             push(@codes, 0320);
 695         } elsif ($op eq 'o32') {
 696             push(@codes, 0321);
 697         } elsif ($op eq 'o64') {  # 64-bit operand size requiring REX.W
 698             push(@codes, 0324);
 699         } elsif ($op eq 'o64nw') { # Implied 64-bit operand size (no REX.W)
 700             push(@codes, 0323);
 701         } elsif ($op eq 'a16') {
 702             push(@codes, 0310);
 703         } elsif ($op eq 'a32') {
 704             push(@codes, 0311);
 705         } elsif ($op eq 'a64') {
 706             push(@codes, 0313);
 707         } elsif ($op eq '!osp') {
 708             push(@codes, 0364);
 709         } elsif ($op eq '!asp') {
 710             push(@codes, 0365);
 711         } elsif ($op eq 'rex.l') {
 712             push(@codes, 0334);
 713         } elsif ($op eq 'repe') {
 714             push(@codes, 0335);
 715         } elsif ($op eq 'nohi') { # Use spl/bpl/sil/dil even without REX
 716             push(@codes, 0325);
 717         } elsif ($op eq 'vsibx' || $op eq 'vm32x' || $op eq 'vm64x') {
 718             # This instruction takes XMM VSIB
 719             push(@codes, 0374);
 720         } elsif ($op eq 'vsiby' || $op eq 'vm32y' || $op eq 'vm64y') {
 721             push(@codes, 0375);
 722         } elsif ($prefix_ok && $op =~ /^(66|f2|f3|np)$/) {
 723             # 66/F2/F3 prefix used as an opcode extension, or np = no prefix
 724             if ($op eq '66') {
 725                 push(@codes, 0361);
 726             } elsif ($op eq 'f2') {
 727                 push(@codes, 0362);
 728             } elsif ($op eq 'f3') {
 729                 push(@codes, 0363);
 730             } else {
 731                 push(@codes, 0360);
 732             }
 733         } elsif ($op =~ /^[0-9a-f]{2}$/) {
 734             if (defined($litix) && $litix+$codes[$litix]+1 == scalar @codes &&
 735                 $codes[$litix] < 4) {
 736                 $codes[$litix]++;
 737                 push(@codes, hex $op);
 738             } else {
 739                 $litix = scalar(@codes);
 740                 push(@codes, 01, hex $op);
 741             }
 742             $prefix_ok = 0;
 743         } elsif ($op eq '/r') {
 744             if (!defined($oppos{'r'}) || !defined($oppos{'m'})) {
 745                 die "$fname: $line: $op requires r and m operands\n";
 746             }
 747             $opex = (($oppos{'m'} & 4) ? 06 : 0) |
 748                 (($oppos{'r'} & 4) ? 05 : 0);
 749             push(@codes, $opex) if ($opex);
 750             push(@codes, 0100 + (($oppos{'m'} & 3) << 3) + ($oppos{'r'} & 3));
 751             $prefix_ok = 0;
 752         } elsif ($op =~ m:^/([0-7])$:) {
 753             if (!defined($oppos{'m'})) {
 754                 die "$fname: $line: $op requires m operand\n";
 755             }
 756             push(@codes, 06) if ($oppos{'m'} & 4);
 757             push(@codes, 0200 + (($oppos{'m'} & 3) << 3) + $1);
 758             $prefix_ok = 0;
 759         } elsif ($op =~ /^(vex|xop)(|\..*)$/) {
 760             my $c = $vexmap{$1};
 761             my ($m,$w,$l,$p) = (undef,2,undef,0);
 762             my $has_nds = 0;
 763             my @subops = split(/\./, $op);
 764             shift @subops;      # Drop prefix
 765                 foreach $oq (@subops) {
 766                     if ($oq eq '128' || $oq eq 'l0' || $oq eq 'lz') {
 767                         $l = 0;
 768                     } elsif ($oq eq '256' || $oq eq 'l1') {
 769                         $l = 1;
 770                     } elsif ($oq eq 'lig') {
 771                         $l = 2;
 772                     } elsif ($oq eq 'w0') {
 773                         $w = 0;
 774                     } elsif ($oq eq 'w1') {
 775                         $w = 1;
 776                     } elsif ($oq eq 'wig') {
 777                         $w = 2;
 778                     } elsif ($oq eq 'ww') {
 779                         $w = 3;
 780                     } elsif ($oq eq 'p0') {
 781                         $p = 0;
 782                     } elsif ($oq eq '66' || $oq eq 'p1') {
 783                         $p = 1;
 784                     } elsif ($oq eq 'f3' || $oq eq 'p2') {
 785                         $p = 2;
 786                     } elsif ($oq eq 'f2' || $oq eq 'p3') {
 787                         $p = 3;
 788                     } elsif ($oq eq '0f') {
 789                         $m = 1;
 790                     } elsif ($oq eq '0f38') {
 791                         $m = 2;
 792                     } elsif ($oq eq '0f3a') {
 793                         $m = 3;
 794                     } elsif ($oq =~ /^m([0-9]+)$/) {
 795                         $m = $1+0;
 796                     } elsif ($oq eq 'nds' || $oq eq 'ndd' || $oq eq 'dds') {
 797                         if (!defined($oppos{'v'})) {
 798                             die "$fname: $line: vex.$oq without 'v' operand\n";
 799                         }
 800                         $has_nds = 1;
 801                     } else {
 802                         die "$fname: $line: undefined VEX subcode: $oq\n";
 803                     }
 804                 }
 805             if (!defined($m) || !defined($w) || !defined($l) || !defined($p)) {
 806                 die "$fname: $line: missing fields in VEX specification\n";
 807             }
 808             if (defined($oppos{'v'}) && !$has_nds) {
 809                 die "$fname: $line: 'v' operand without vex.nds or vex.ndd\n";
 810             }
 811             push(@codes, defined($oppos{'v'}) ? 0260+($oppos{'v'} & 3) : 0270,
 812                  ($c << 6)+$m, ($w << 4)+($l << 2)+$p);
 813             $prefix_ok = 0;
 814         } elsif ($op =~ /^\/drex([01])$/) {
 815             my $oc0 = $1;
 816             if (!defined($oppos{'d'})) {
 817                 die "$fname: $line: DREX without a 'd' operand\n";
 818             }
 819             # Note the use of *unshift* here, as opposed to *push*.
 820             # This is because NASM want this byte code at the start of
 821             # the instruction sequence, but the AMD documentation puts
 822             # this at (roughly) the position of the drex byte itself.
 823             # This allows us to match the AMD documentation and still
 824             # do the right thing.
 825             unshift(@codes, 0160+($oppos{'d'} & 3)+($oc0 ? 4 : 0));
 826             unshift(@codes, 05) if ($oppos{'d'} & 4);
 827         } elsif ($op =~ /^(ib\,s|ib|ibx|ib\,w|iw|iwd|id|idx|iwdq|rel|rel8|rel16|rel32|iq|seg|ibw|ibd|ibd,s)$/) {
 828             if (!defined($oppos{'i'})) {
 829                 die "$fname: $line: $op without 'i' operand\n";
 830             }
 831             if ($op eq 'ib,s') { # Signed imm8
 832                 push(@codes, 05) if ($oppos{'i'} & 4);
 833                 push(@codes, 014+($oppos{'i'} & 3));
 834             } elsif ($op eq 'ib') { # imm8
 835                 push(@codes, 05) if ($oppos{'i'} & 4);
 836                 push(@codes, 020+($oppos{'i'} & 3));
 837             } elsif ($op eq 'ib,u') { # Unsigned imm8
 838                 push(@codes, 05) if ($oppos{'i'} & 4);
 839                 push(@codes, 024+($oppos{'i'} & 3));
 840             } elsif ($op eq 'iw') { # imm16
 841                 push(@codes, 05) if ($oppos{'i'} & 4);
 842                 push(@codes, 030+($oppos{'i'} & 3));
 843             } elsif ($op eq 'ibx') { # imm8 sign-extended to opsize
 844                 push(@codes, 05) if ($oppos{'i'} & 4);
 845                 push(@codes, 0274+($oppos{'i'} & 3));
 846             } elsif ($op eq 'iwd') { # imm16 or imm32, depending on opsize
 847                 push(@codes, 05) if ($oppos{'i'} & 4);
 848                 push(@codes, 034+($oppos{'i'} & 3));
 849             } elsif ($op eq 'id') { # imm32
 850                 push(@codes, 05) if ($oppos{'i'} & 4);
 851                 push(@codes, 040+($oppos{'i'} & 3));
 852             } elsif ($op eq 'idx') { # imm32 extended to 64 bits
 853                 push(@codes, 05) if ($oppos{'i'} & 4);
 854                 push(@codes, 0254+($oppos{'i'} & 3));
 855             } elsif ($op eq 'iwdq') { # imm16/32/64, depending on opsize
 856                 push(@codes, 05) if ($oppos{'i'} & 4);
 857                 push(@codes, 044+($oppos{'i'} & 3));
 858             } elsif ($op eq 'rel8') {
 859                 push(@codes, 05) if ($oppos{'i'} & 4);
 860                 push(@codes, 050+($oppos{'i'} & 3));
 861             } elsif ($op eq 'iq') {
 862                 push(@codes, 05) if ($oppos{'i'} & 4);
 863                 push(@codes, 054+($oppos{'i'} & 3));
 864             } elsif ($op eq 'rel16') {
 865                 push(@codes, 05) if ($oppos{'i'} & 4);
 866                 push(@codes, 060+($oppos{'i'} & 3));
 867             } elsif ($op eq 'rel') { # 16 or 32 bit relative operand
 868                 push(@codes, 05) if ($oppos{'i'} & 4);
 869                 push(@codes, 064+($oppos{'i'} & 3));
 870             } elsif ($op eq 'rel32') {
 871                 push(@codes, 05) if ($oppos{'i'} & 4);
 872                 push(@codes, 070+($oppos{'i'} & 3));
 873             } elsif ($op eq 'seg') {
 874                 push(@codes, 05) if ($oppos{'i'} & 4);
 875                 push(@codes, 074+($oppos{'i'} & 3));
 876             } elsif ($op eq 'ibw') { # imm16 that can be bytified
 877                 if (!defined($s_pos)) {
 878                     die "$fname: $line: $op without a +s byte\n";
 879                 }
 880                 $codes[$s_pos] += 0144;
 881                 push(@codes, 05) if ($oppos{'i'} & 4);
 882                 push(@codes, 0140+($oppos{'i'} & 3));
 883             } elsif ($op eq 'ibd') { # imm32 that can be bytified
 884                 if (!defined($s_pos)) {
 885                     die "$fname: $line: $op without a +s byte\n";
 886                 }
 887                 $codes[$s_pos] += 0154;
 888                 push(@codes, 05) if ($oppos{'i'} & 4);
 889                 push(@codes, 0150+($oppos{'i'} & 3));
 890             } elsif ($op eq 'ibd,s') {
 891                 # imm32 that can be bytified, sign extended to 64 bits
 892                 if (!defined($s_pos)) {
 893                     die "$fname: $line: $op without a +s byte\n";
 894                 }
 895                 $codes[$s_pos] += 0154;
 896                 push(@codes, 05) if ($oppos{'i'} & 4);
 897                 push(@codes, 0250+($oppos{'i'} & 3));
 898             }
 899             $prefix_ok = 0;
 900         } elsif ($op eq '/is4') {
 901             if (!defined($oppos{'s'})) {
 902                 die "$fname: $line: $op without 's' operand\n";
 903             }
 904             if (defined($oppos{'i'})) {
 905                 push(@codes, 0172, ($oppos{'s'} << 3)+$oppos{'i'});
 906             } else {
 907                 push(@codes, 0174, $oppos{'s'});
 908             }
 909             $prefix_ok = 0;
 910         } elsif ($op =~ /^\/is4\=([0-9]+)$/) {
 911             my $imm = $1;
 912             if (!defined($oppos{'s'})) {
 913                 die "$fname: $line: $op without 's' operand\n";
 914             }
 915             if ($imm < 0 || $imm > 15) {
 916                 die "$fname: $line: invalid imm4 value for $op: $imm\n";
 917             }
 918             push(@codes, 0173, ($oppos{'s'} << 4) + $imm);
 919             $prefix_ok = 0;
 920         } elsif ($op =~ /^([0-9a-f]{2})\+s$/) {
 921             if (!defined($oppos{'i'})) {
 922                 die "$fname: $line: $op without 'i' operand\n";
 923             }
 924             $s_pos = scalar @codes;
 925             push(@codes, 05) if ($oppos{'i'} & 4);
 926             push(@codes, $oppos{'i'} & 3, hex $1);
 927             $prefix_ok = 0;
 928         } elsif ($op =~ /^([0-9a-f]{2})\+c$/) {
 929             push(@codes, 0330, hex $1);
 930             $prefix_ok = 0;
 931         } elsif ($op =~ /^\\([0-7]+|x[0-9a-f]{2})$/) {
 932             # Escape to enter literal bytecodes
 933             push(@codes, oct $1);
 934         } else {
 935             die "$fname: $line: unknown operation: $op\n";
 936         }
 937     }
 938
 939     return @codes;
 940 }