scripts/decodetree.py

   1 #!/usr/bin/env python
   2 # Copyright (c) 2018 Linaro Limited
   3 #
   4 # This library is free software; you can redistribute it and/or
   5 # modify it under the terms of the GNU Lesser General Public
   6 # License as published by the Free Software Foundation; either
   7 # version 2 of the License, or (at your option) any later version.
   8 #
   9 # This library is distributed in the hope that it will be useful,
  10 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12 # Lesser General Public License for more details.
  13 #
  14 # You should have received a copy of the GNU Lesser General Public
  15 # License along with this library; if not, see <http://www.gnu.org/licenses/>.
  16 #
  17
  18 #
  19 # Generate a decoding tree from a specification file.
  20 # See the syntax and semantics in docs/devel/decodetree.rst.
  21 #
  22
  23 import os
  24 import re
  25 import sys
  26 import getopt
  27
  28 insnwidth = 32
  29 insnmask = 0xffffffff
  30 variablewidth = False
  31 fields = {}
  32 arguments = {}
  33 formats = {}
  34 patterns = []
  35 allpatterns = []
  36
  37 translate_prefix = 'trans'
  38 translate_scope = 'static '
  39 input_file = ''
  40 output_file = None
  41 output_fd = None
  42 insntype = 'uint32_t'
  43 decode_function = 'decode'
  44
  45 re_ident = '[a-zA-Z][a-zA-Z0-9_]*'
  46
  47
  48 def error_with_file(file, lineno, *args):
  49     """Print an error message from file:line and args and exit."""
  50     global output_file
  51     global output_fd
  52
  53     if lineno:
  54         r = '{0}:{1}: error:'.format(file, lineno)
  55     elif input_file:
  56         r = '{0}: error:'.format(file)
  57     else:
  58         r = 'error:'
  59     for a in args:
  60         r += ' ' + str(a)
  61     r += '\n'
  62     sys.stderr.write(r)
  63     if output_file and output_fd:
  64         output_fd.close()
  65         os.remove(output_file)
  66     exit(1)
  67
  68 def error(lineno, *args):
  69     error_with_file(input_file, lineno, args)
  70
  71 def output(*args):
  72     global output_fd
  73     for a in args:
  74         output_fd.write(a)
  75
  76
  77 if sys.version_info >= (3, 4):
  78     re_fullmatch = re.fullmatch
  79 else:
  80     def re_fullmatch(pat, str):
  81         return re.match('^' + pat + '$', str)
  82
  83
  84 def output_autogen():
  85     output('/* This file is autogenerated by scripts/decodetree.py.  */\n\n')
  86
  87
  88 def str_indent(c):
  89     """Return a string with C spaces"""
  90     return ' ' * c
  91
  92
  93 def str_fields(fields):
  94     """Return a string uniquely identifing FIELDS"""
  95     r = ''
  96     for n in sorted(fields.keys()):
  97         r += '_' + n
  98     return r[1:]
  99
 100
 101 def str_match_bits(bits, mask):
 102     """Return a string pretty-printing BITS/MASK"""
 103     global insnwidth
 104
 105     i = 1 << (insnwidth - 1)
 106     space = 0x01010100
 107     r = ''
 108     while i != 0:
 109         if i & mask:
 110             if i & bits:
 111                 r += '1'
 112             else:
 113                 r += '0'
 114         else:
 115             r += '.'
 116         if i & space:
 117             r += ' '
 118         i >>= 1
 119     return r
 120
 121
 122 def is_pow2(x):
 123     """Return true iff X is equal to a power of 2."""
 124     return (x & (x - 1)) == 0
 125
 126
 127 def ctz(x):
 128     """Return the number of times 2 factors into X."""
 129     r = 0
 130     while ((x >> r) & 1) == 0:
 131         r += 1
 132     return r
 133
 134
 135 def is_contiguous(bits):
 136     shift = ctz(bits)
 137     if is_pow2((bits >> shift) + 1):
 138         return shift
 139     else:
 140         return -1
 141
 142
 143 def eq_fields_for_args(flds_a, flds_b):
 144     if len(flds_a) != len(flds_b):
 145         return False
 146     for k, a in flds_a.items():
 147         if k not in flds_b:
 148             return False
 149     return True
 150
 151
 152 def eq_fields_for_fmts(flds_a, flds_b):
 153     if len(flds_a) != len(flds_b):
 154         return False
 155     for k, a in flds_a.items():
 156         if k not in flds_b:
 157             return False
 158         b = flds_b[k]
 159         if a.__class__ != b.__class__ or a != b:
 160             return False
 161     return True
 162
 163
 164 class Field:
 165     """Class representing a simple instruction field"""
 166     def __init__(self, sign, pos, len):
 167         self.sign = sign
 168         self.pos = pos
 169         self.len = len
 170         self.mask = ((1 << len) - 1) << pos
 171
 172     def __str__(self):
 173         if self.sign:
 174             s = 's'
 175         else:
 176             s = ''
 177         return str(self.pos) + ':' + s + str(self.len)
 178
 179     def str_extract(self):
 180         if self.sign:
 181             extr = 'sextract32'
 182         else:
 183             extr = 'extract32'
 184         return '{0}(insn, {1}, {2})'.format(extr, self.pos, self.len)
 185
 186     def __eq__(self, other):
 187         return self.sign == other.sign and self.sign == other.sign
 188
 189     def __ne__(self, other):
 190         return not self.__eq__(other)
 191 # end Field
 192
 193
 194 class MultiField:
 195     """Class representing a compound instruction field"""
 196     def __init__(self, subs, mask):
 197         self.subs = subs
 198         self.sign = subs[0].sign
 199         self.mask = mask
 200
 201     def __str__(self):
 202         return str(self.subs)
 203
 204     def str_extract(self):
 205         ret = '0'
 206         pos = 0
 207         for f in reversed(self.subs):
 208             if pos == 0:
 209                 ret = f.str_extract()
 210             else:
 211                 ret = 'deposit32({0}, {1}, {2}, {3})' \
 212                       .format(ret, pos, 32 - pos, f.str_extract())
 213             pos += f.len
 214         return ret
 215
 216     def __ne__(self, other):
 217         if len(self.subs) != len(other.subs):
 218             return True
 219         for a, b in zip(self.subs, other.subs):
 220             if a.__class__ != b.__class__ or a != b:
 221                 return True
 222         return False
 223
 224     def __eq__(self, other):
 225         return not self.__ne__(other)
 226 # end MultiField
 227
 228
 229 class ConstField:
 230     """Class representing an argument field with constant value"""
 231     def __init__(self, value):
 232         self.value = value
 233         self.mask = 0
 234         self.sign = value < 0
 235
 236     def __str__(self):
 237         return str(self.value)
 238
 239     def str_extract(self):
 240         return str(self.value)
 241
 242     def __cmp__(self, other):
 243         return self.value - other.value
 244 # end ConstField
 245
 246
 247 class FunctionField:
 248     """Class representing a field passed through an expander"""
 249     def __init__(self, func, base):
 250         self.mask = base.mask
 251         self.sign = base.sign
 252         self.base = base
 253         self.func = func
 254
 255     def __str__(self):
 256         return self.func + '(' + str(self.base) + ')'
 257
 258     def str_extract(self):
 259         return self.func + '(' + self.base.str_extract() + ')'
 260
 261     def __eq__(self, other):
 262         return self.func == other.func and self.base == other.base
 263
 264     def __ne__(self, other):
 265         return not self.__eq__(other)
 266 # end FunctionField
 267
 268
 269 class Arguments:
 270     """Class representing the extracted fields of a format"""
 271     def __init__(self, nm, flds, extern):
 272         self.name = nm
 273         self.extern = extern
 274         self.fields = sorted(flds)
 275
 276     def __str__(self):
 277         return self.name + ' ' + str(self.fields)
 278
 279     def struct_name(self):
 280         return 'arg_' + self.name
 281
 282     def output_def(self):
 283         if not self.extern:
 284             output('typedef struct {\n')
 285             for n in self.fields:
 286                 output('    int ', n, ';\n')
 287             output('} ', self.struct_name(), ';\n\n')
 288 # end Arguments
 289
 290
 291 class General:
 292     """Common code between instruction formats and instruction patterns"""
 293     def __init__(self, name, lineno, base, fixb, fixm, udfm, fldm, flds, w):
 294         self.name = name
 295         self.file = input_file
 296         self.lineno = lineno
 297         self.base = base
 298         self.fixedbits = fixb
 299         self.fixedmask = fixm
 300         self.undefmask = udfm
 301         self.fieldmask = fldm
 302         self.fields = flds
 303         self.width = w
 304
 305     def __str__(self):
 306         return self.name + ' ' + str_match_bits(self.fixedbits, self.fixedmask)
 307
 308     def str1(self, i):
 309         return str_indent(i) + self.__str__()
 310 # end General
 311
 312
 313 class Format(General):
 314     """Class representing an instruction format"""
 315
 316     def extract_name(self):
 317         global decode_function
 318         return decode_function + '_extract_' + self.name
 319
 320     def output_extract(self):
 321         output('static void ', self.extract_name(), '(',
 322                self.base.struct_name(), ' *a, ', insntype, ' insn)\n{\n')
 323         for n, f in self.fields.items():
 324             output('    a->', n, ' = ', f.str_extract(), ';\n')
 325         output('}\n\n')
 326 # end Format
 327
 328
 329 class Pattern(General):
 330     """Class representing an instruction pattern"""
 331
 332     def output_decl(self):
 333         global translate_scope
 334         global translate_prefix
 335         output('typedef ', self.base.base.struct_name(),
 336                ' arg_', self.name, ';\n')
 337         output(translate_scope, 'bool ', translate_prefix, '_', self.name,
 338                '(DisasContext *ctx, arg_', self.name, ' *a);\n')
 339
 340     def output_code(self, i, extracted, outerbits, outermask):
 341         global translate_prefix
 342         ind = str_indent(i)
 343         arg = self.base.base.name
 344         output(ind, '/* ', self.file, ':', str(self.lineno), ' */\n')
 345         if not extracted:
 346             output(ind, self.base.extract_name(), '(&u.f_', arg, ', insn);\n')
 347         for n, f in self.fields.items():
 348             output(ind, 'u.f_', arg, '.', n, ' = ', f.str_extract(), ';\n')
 349         output(ind, 'if (', translate_prefix, '_', self.name,
 350                '(ctx, &u.f_', arg, ')) return true;\n')
 351 # end Pattern
 352
 353
 354 class MultiPattern(General):
 355     """Class representing an overlapping set of instruction patterns"""
 356
 357     def __init__(self, lineno, pats, fixb, fixm, udfm, w):
 358         self.file = input_file
 359         self.lineno = lineno
 360         self.pats = pats
 361         self.base = None
 362         self.fixedbits = fixb
 363         self.fixedmask = fixm
 364         self.undefmask = udfm
 365         self.width = w
 366
 367     def __str__(self):
 368         r = "{"
 369         for p in self.pats:
 370            r = r + ' ' + str(p)
 371         return r + "}"
 372
 373     def output_decl(self):
 374         for p in self.pats:
 375             p.output_decl()
 376
 377     def output_code(self, i, extracted, outerbits, outermask):
 378         global translate_prefix
 379         ind = str_indent(i)
 380         for p in self.pats:
 381             if outermask != p.fixedmask:
 382                 innermask = p.fixedmask & ~outermask
 383                 innerbits = p.fixedbits & ~outermask
 384                 output(ind, 'if ((insn & ',
 385                        '0x{0:08x}) == 0x{1:08x}'.format(innermask, innerbits),
 386                        ') {\n')
 387                 output(ind, '    /* ',
 388                        str_match_bits(p.fixedbits, p.fixedmask), ' */\n')
 389                 p.output_code(i + 4, extracted, p.fixedbits, p.fixedmask)
 390                 output(ind, '}\n')
 391             else:
 392                 p.output_code(i, extracted, p.fixedbits, p.fixedmask)
 393 #end MultiPattern
 394
 395
 396 def parse_field(lineno, name, toks):
 397     """Parse one instruction field from TOKS at LINENO"""
 398     global fields
 399     global re_ident
 400     global insnwidth
 401
 402     # A "simple" field will have only one entry;
 403     # a "multifield" will have several.
 404     subs = []
 405     width = 0
 406     func = None
 407     for t in toks:
 408         if re_fullmatch('!function=' + re_ident, t):
 409             if func:
 410                 error(lineno, 'duplicate function')
 411             func = t.split('=')
 412             func = func[1]
 413             continue
 414
 415         if re_fullmatch('[0-9]+:s[0-9]+', t):
 416             # Signed field extract
 417             subtoks = t.split(':s')
 418             sign = True
 419         elif re_fullmatch('[0-9]+:[0-9]+', t):
 420             # Unsigned field extract
 421             subtoks = t.split(':')
 422             sign = False
 423         else:
 424             error(lineno, 'invalid field token "{0}"'.format(t))
 425         po = int(subtoks[0])
 426         le = int(subtoks[1])
 427         if po + le > insnwidth:
 428             error(lineno, 'field {0} too large'.format(t))
 429         f = Field(sign, po, le)
 430         subs.append(f)
 431         width += le
 432
 433     if width > insnwidth:
 434         error(lineno, 'field too large')
 435     if len(subs) == 1:
 436         f = subs[0]
 437     else:
 438         mask = 0
 439         for s in subs:
 440             if mask & s.mask:
 441                 error(lineno, 'field components overlap')
 442             mask |= s.mask
 443         f = MultiField(subs, mask)
 444     if func:
 445         f = FunctionField(func, f)
 446
 447     if name in fields:
 448         error(lineno, 'duplicate field', name)
 449     fields[name] = f
 450 # end parse_field
 451
 452
 453 def parse_arguments(lineno, name, toks):
 454     """Parse one argument set from TOKS at LINENO"""
 455     global arguments
 456     global re_ident
 457
 458     flds = []
 459     extern = False
 460     for t in toks:
 461         if re_fullmatch('!extern', t):
 462             extern = True
 463             continue
 464         if not re_fullmatch(re_ident, t):
 465             error(lineno, 'invalid argument set token "{0}"'.format(t))
 466         if t in flds:
 467             error(lineno, 'duplicate argument "{0}"'.format(t))
 468         flds.append(t)
 469
 470     if name in arguments:
 471         error(lineno, 'duplicate argument set', name)
 472     arguments[name] = Arguments(name, flds, extern)
 473 # end parse_arguments
 474
 475
 476 def lookup_field(lineno, name):
 477     global fields
 478     if name in fields:
 479         return fields[name]
 480     error(lineno, 'undefined field', name)
 481
 482
 483 def add_field(lineno, flds, new_name, f):
 484     if new_name in flds:
 485         error(lineno, 'duplicate field', new_name)
 486     flds[new_name] = f
 487     return flds
 488
 489
 490 def add_field_byname(lineno, flds, new_name, old_name):
 491     return add_field(lineno, flds, new_name, lookup_field(lineno, old_name))
 492
 493
 494 def infer_argument_set(flds):
 495     global arguments
 496     global decode_function
 497
 498     for arg in arguments.values():
 499         if eq_fields_for_args(flds, arg.fields):
 500             return arg
 501
 502     name = decode_function + str(len(arguments))
 503     arg = Arguments(name, flds.keys(), False)
 504     arguments[name] = arg
 505     return arg
 506
 507
 508 def infer_format(arg, fieldmask, flds, width):
 509     global arguments
 510     global formats
 511     global decode_function
 512
 513     const_flds = {}
 514     var_flds = {}
 515     for n, c in flds.items():
 516         if c is ConstField:
 517             const_flds[n] = c
 518         else:
 519             var_flds[n] = c
 520
 521     # Look for an existing format with the same argument set and fields
 522     for fmt in formats.values():
 523         if arg and fmt.base != arg:
 524             continue
 525         if fieldmask != fmt.fieldmask:
 526             continue
 527         if width != fmt.width:
 528             continue
 529         if not eq_fields_for_fmts(flds, fmt.fields):
 530             continue
 531         return (fmt, const_flds)
 532
 533     name = decode_function + '_Fmt_' + str(len(formats))
 534     if not arg:
 535         arg = infer_argument_set(flds)
 536
 537     fmt = Format(name, 0, arg, 0, 0, 0, fieldmask, var_flds, width)
 538     formats[name] = fmt
 539
 540     return (fmt, const_flds)
 541 # end infer_format
 542
 543
 544 def parse_generic(lineno, is_format, name, toks):
 545     """Parse one instruction format from TOKS at LINENO"""
 546     global fields
 547     global arguments
 548     global formats
 549     global patterns
 550     global allpatterns
 551     global re_ident
 552     global insnwidth
 553     global insnmask
 554     global variablewidth
 555
 556     fixedmask = 0
 557     fixedbits = 0
 558     undefmask = 0
 559     width = 0
 560     flds = {}
 561     arg = None
 562     fmt = None
 563     for t in toks:
 564         # '&Foo' gives a format an explcit argument set.
 565         if t[0] == '&':
 566             tt = t[1:]
 567             if arg:
 568                 error(lineno, 'multiple argument sets')
 569             if tt in arguments:
 570                 arg = arguments[tt]
 571             else:
 572                 error(lineno, 'undefined argument set', t)
 573             continue
 574
 575         # '@Foo' gives a pattern an explicit format.
 576         if t[0] == '@':
 577             tt = t[1:]
 578             if fmt:
 579                 error(lineno, 'multiple formats')
 580             if tt in formats:
 581                 fmt = formats[tt]
 582             else:
 583                 error(lineno, 'undefined format', t)
 584             continue
 585
 586         # '%Foo' imports a field.
 587         if t[0] == '%':
 588             tt = t[1:]
 589             flds = add_field_byname(lineno, flds, tt, tt)
 590             continue
 591
 592         # 'Foo=%Bar' imports a field with a different name.
 593         if re_fullmatch(re_ident + '=%' + re_ident, t):
 594             (fname, iname) = t.split('=%')
 595             flds = add_field_byname(lineno, flds, fname, iname)
 596             continue
 597
 598         # 'Foo=number' sets an argument field to a constant value
 599         if re_fullmatch(re_ident + '=[+-]?[0-9]+', t):
 600             (fname, value) = t.split('=')
 601             value = int(value)
 602             flds = add_field(lineno, flds, fname, ConstField(value))
 603             continue
 604
 605         # Pattern of 0s, 1s, dots and dashes indicate required zeros,
 606         # required ones, or dont-cares.
 607         if re_fullmatch('[01.-]+', t):
 608             shift = len(t)
 609             fms = t.replace('0', '1')
 610             fms = fms.replace('.', '0')
 611             fms = fms.replace('-', '0')
 612             fbs = t.replace('.', '0')
 613             fbs = fbs.replace('-', '0')
 614             ubm = t.replace('1', '0')
 615             ubm = ubm.replace('.', '0')
 616             ubm = ubm.replace('-', '1')
 617             fms = int(fms, 2)
 618             fbs = int(fbs, 2)
 619             ubm = int(ubm, 2)
 620             fixedbits = (fixedbits << shift) | fbs
 621             fixedmask = (fixedmask << shift) | fms
 622             undefmask = (undefmask << shift) | ubm
 623         # Otherwise, fieldname:fieldwidth
 624         elif re_fullmatch(re_ident + ':s?[0-9]+', t):
 625             (fname, flen) = t.split(':')
 626             sign = False
 627             if flen[0] == 's':
 628                 sign = True
 629                 flen = flen[1:]
 630             shift = int(flen, 10)
 631             if shift + width > insnwidth:
 632                 error(lineno, 'field {0} exceeds insnwidth'.format(fname))
 633             f = Field(sign, insnwidth - width - shift, shift)
 634             flds = add_field(lineno, flds, fname, f)
 635             fixedbits <<= shift
 636             fixedmask <<= shift
 637             undefmask <<= shift
 638         else:
 639             error(lineno, 'invalid token "{0}"'.format(t))
 640         width += shift
 641
 642     if variablewidth and width < insnwidth and width % 8 == 0:
 643         shift = insnwidth - width
 644         fixedbits <<= shift
 645         fixedmask <<= shift
 646         undefmask <<= shift
 647         undefmask |= (1 << shift) - 1
 648
 649     # We should have filled in all of the bits of the instruction.
 650     elif not (is_format and width == 0) and width != insnwidth:
 651         error(lineno, 'definition has {0} bits'.format(width))
 652
 653     # Do not check for fields overlaping fields; one valid usage
 654     # is to be able to duplicate fields via import.
 655     fieldmask = 0
 656     for f in flds.values():
 657         fieldmask |= f.mask
 658
 659     # Fix up what we've parsed to match either a format or a pattern.
 660     if is_format:
 661         # Formats cannot reference formats.
 662         if fmt:
 663             error(lineno, 'format referencing format')
 664         # If an argument set is given, then there should be no fields
 665         # without a place to store it.
 666         if arg:
 667             for f in flds.keys():
 668                 if f not in arg.fields:
 669                     error(lineno, 'field {0} not in argument set {1}'
 670                                   .format(f, arg.name))
 671         else:
 672             arg = infer_argument_set(flds)
 673         if name in formats:
 674             error(lineno, 'duplicate format name', name)
 675         fmt = Format(name, lineno, arg, fixedbits, fixedmask,
 676                      undefmask, fieldmask, flds, width)
 677         formats[name] = fmt
 678     else:
 679         # Patterns can reference a format ...
 680         if fmt:
 681             # ... but not an argument simultaneously
 682             if arg:
 683                 error(lineno, 'pattern specifies both format and argument set')
 684             if fixedmask & fmt.fixedmask:
 685                 error(lineno, 'pattern fixed bits overlap format fixed bits')
 686             if width != fmt.width:
 687                 error(lineno, 'pattern uses format of different width')
 688             fieldmask |= fmt.fieldmask
 689             fixedbits |= fmt.fixedbits
 690             fixedmask |= fmt.fixedmask
 691             undefmask |= fmt.undefmask
 692         else:
 693             (fmt, flds) = infer_format(arg, fieldmask, flds, width)
 694         arg = fmt.base
 695         for f in flds.keys():
 696             if f not in arg.fields:
 697                 error(lineno, 'field {0} not in argument set {1}'
 698                               .format(f, arg.name))
 699             if f in fmt.fields.keys():
 700                 error(lineno, 'field {0} set by format and pattern'.format(f))
 701         for f in arg.fields:
 702             if f not in flds.keys() and f not in fmt.fields.keys():
 703                 error(lineno, 'field {0} not initialized'.format(f))
 704         pat = Pattern(name, lineno, fmt, fixedbits, fixedmask,
 705                       undefmask, fieldmask, flds, width)
 706         patterns.append(pat)
 707         allpatterns.append(pat)
 708
 709     # Validate the masks that we have assembled.
 710     if fieldmask & fixedmask:
 711         error(lineno, 'fieldmask overlaps fixedmask (0x{0:08x} & 0x{1:08x})'
 712                       .format(fieldmask, fixedmask))
 713     if fieldmask & undefmask:
 714         error(lineno, 'fieldmask overlaps undefmask (0x{0:08x} & 0x{1:08x})'
 715                       .format(fieldmask, undefmask))
 716     if fixedmask & undefmask:
 717         error(lineno, 'fixedmask overlaps undefmask (0x{0:08x} & 0x{1:08x})'
 718                       .format(fixedmask, undefmask))
 719     if not is_format:
 720         allbits = fieldmask | fixedmask | undefmask
 721         if allbits != insnmask:
 722             error(lineno, 'bits left unspecified (0x{0:08x})'
 723                           .format(allbits ^ insnmask))
 724 # end parse_general
 725
 726 def build_multi_pattern(lineno, pats):
 727     """Validate the Patterns going into a MultiPattern."""
 728     global patterns
 729     global insnmask
 730
 731     if len(pats) < 2:
 732         error(lineno, 'less than two patterns within braces')
 733
 734     fixedmask = insnmask
 735     undefmask = insnmask
 736
 737     # Collect fixed/undefmask for all of the children.
 738     # Move the defining lineno back to that of the first child.
 739     for p in pats:
 740         fixedmask &= p.fixedmask
 741         undefmask &= p.undefmask
 742         if p.lineno < lineno:
 743             lineno = p.lineno
 744
 745     width = None
 746     for p in pats:
 747         if width is None:
 748             width = p.width
 749         elif width != p.width:
 750             error(lineno, 'width mismatch in patterns within braces')
 751
 752     repeat = True
 753     while repeat:
 754         if fixedmask == 0:
 755             error(lineno, 'no overlap in patterns within braces')
 756         fixedbits = None
 757         for p in pats:
 758             thisbits = p.fixedbits & fixedmask
 759             if fixedbits is None:
 760                 fixedbits = thisbits
 761             elif fixedbits != thisbits:
 762                 fixedmask &= ~(fixedbits ^ thisbits)
 763                 break
 764         else:
 765             repeat = False
 766
 767     mp = MultiPattern(lineno, pats, fixedbits, fixedmask, undefmask, width)
 768     patterns.append(mp)
 769 # end build_multi_pattern
 770
 771 def parse_file(f):
 772     """Parse all of the patterns within a file"""
 773
 774     global patterns
 775
 776     # Read all of the lines of the file.  Concatenate lines
 777     # ending in backslash; discard empty lines and comments.
 778     toks = []
 779     lineno = 0
 780     nesting = 0
 781     saved_pats = []
 782
 783     for line in f:
 784         lineno += 1
 785
 786         # Expand and strip spaces, to find indent.
 787         line = line.rstrip()
 788         line = line.expandtabs()
 789         len1 = len(line)
 790         line = line.lstrip()
 791         len2 = len(line)
 792
 793         # Discard comments
 794         end = line.find('#')
 795         if end >= 0:
 796             line = line[:end]
 797
 798         t = line.split()
 799         if len(toks) != 0:
 800             # Next line after continuation
 801             toks.extend(t)
 802         else:
 803             # Allow completely blank lines.
 804             if len1 == 0:
 805                 continue
 806             indent = len1 - len2
 807             # Empty line due to comment.
 808             if len(t) == 0:
 809                 # Indentation must be correct, even for comment lines.
 810                 if indent != nesting:
 811                     error(lineno, 'indentation ', indent, ' != ', nesting)
 812                 continue
 813             start_lineno = lineno
 814             toks = t
 815
 816         # Continuation?
 817         if toks[-1] == '\\':
 818             toks.pop()
 819             continue
 820
 821         name = toks[0]
 822         del toks[0]
 823
 824         # End nesting?
 825         if name == '}':
 826             if nesting == 0:
 827                 error(start_lineno, 'mismatched close brace')
 828             if len(toks) != 0:
 829                 error(start_lineno, 'extra tokens after close brace')
 830             nesting -= 2
 831             if indent != nesting:
 832                 error(start_lineno, 'indentation ', indent, ' != ', nesting)
 833             pats = patterns
 834             patterns = saved_pats.pop()
 835             build_multi_pattern(lineno, pats)
 836             toks = []
 837             continue
 838
 839         # Everything else should have current indentation.
 840         if indent != nesting:
 841             error(start_lineno, 'indentation ', indent, ' != ', nesting)
 842
 843         # Start nesting?
 844         if name == '{':
 845             if len(toks) != 0:
 846                 error(start_lineno, 'extra tokens after open brace')
 847             saved_pats.append(patterns)
 848             patterns = []
 849             nesting += 2
 850             toks = []
 851             continue
 852
 853         # Determine the type of object needing to be parsed.
 854         if name[0] == '%':
 855             parse_field(start_lineno, name[1:], toks)
 856         elif name[0] == '&':
 857             parse_arguments(start_lineno, name[1:], toks)
 858         elif name[0] == '@':
 859             parse_generic(start_lineno, True, name[1:], toks)
 860         else:
 861             parse_generic(start_lineno, False, name, toks)
 862         toks = []
 863 # end parse_file
 864
 865
 866 class Tree:
 867     """Class representing a node in a decode tree"""
 868
 869     def __init__(self, fm, tm):
 870         self.fixedmask = fm
 871         self.thismask = tm
 872         self.subs = []
 873         self.base = None
 874
 875     def str1(self, i):
 876         ind = str_indent(i)
 877         r = '{0}{1:08x}'.format(ind, self.fixedmask)
 878         if self.format:
 879             r += ' ' + self.format.name
 880         r += ' [\n'
 881         for (b, s) in self.subs:
 882             r += '{0}  {1:08x}:\n'.format(ind, b)
 883             r += s.str1(i + 4) + '\n'
 884         r += ind + ']'
 885         return r
 886
 887     def __str__(self):
 888         return self.str1(0)
 889
 890     def output_code(self, i, extracted, outerbits, outermask):
 891         ind = str_indent(i)
 892
 893         # If we identified all nodes below have the same format,
 894         # extract the fields now.
 895         if not extracted and self.base:
 896             output(ind, self.base.extract_name(),
 897                    '(&u.f_', self.base.base.name, ', insn);\n')
 898             extracted = True
 899
 900         # Attempt to aid the compiler in producing compact switch statements.
 901         # If the bits in the mask are contiguous, extract them.
 902         sh = is_contiguous(self.thismask)
 903         if sh > 0:
 904             # Propagate SH down into the local functions.
 905             def str_switch(b, sh=sh):
 906                 return '(insn >> {0}) & 0x{1:x}'.format(sh, b >> sh)
 907
 908             def str_case(b, sh=sh):
 909                 return '0x{0:x}'.format(b >> sh)
 910         else:
 911             def str_switch(b):
 912                 return 'insn & 0x{0:08x}'.format(b)
 913
 914             def str_case(b):
 915                 return '0x{0:08x}'.format(b)
 916
 917         output(ind, 'switch (', str_switch(self.thismask), ') {\n')
 918         for b, s in sorted(self.subs):
 919             assert (self.thismask & ~s.fixedmask) == 0
 920             innermask = outermask | self.thismask
 921             innerbits = outerbits | b
 922             output(ind, 'case ', str_case(b), ':\n')
 923             output(ind, '    /* ',
 924                    str_match_bits(innerbits, innermask), ' */\n')
 925             s.output_code(i + 4, extracted, innerbits, innermask)
 926             output(ind, '    return false;\n')
 927         output(ind, '}\n')
 928 # end Tree
 929
 930
 931 def build_tree(pats, outerbits, outermask):
 932     # Find the intersection of all remaining fixedmask.
 933     innermask = ~outermask & insnmask
 934     for i in pats:
 935         innermask &= i.fixedmask
 936
 937     if innermask == 0:
 938         text = 'overlapping patterns:'
 939         for p in pats:
 940             text += '\n' + p.file + ':' + str(p.lineno) + ': ' + str(p)
 941         error_with_file(pats[0].file, pats[0].lineno, text)
 942
 943     fullmask = outermask | innermask
 944
 945     # Sort each element of pats into the bin selected by the mask.
 946     bins = {}
 947     for i in pats:
 948         fb = i.fixedbits & innermask
 949         if fb in bins:
 950             bins[fb].append(i)
 951         else:
 952             bins[fb] = [i]
 953
 954     # We must recurse if any bin has more than one element or if
 955     # the single element in the bin has not been fully matched.
 956     t = Tree(fullmask, innermask)
 957
 958     for b, l in bins.items():
 959         s = l[0]
 960         if len(l) > 1 or s.fixedmask & ~fullmask != 0:
 961             s = build_tree(l, b | outerbits, fullmask)
 962         t.subs.append((b, s))
 963
 964     return t
 965 # end build_tree
 966
 967
 968 def prop_format(tree):
 969     """Propagate Format objects into the decode tree"""
 970
 971     # Depth first search.
 972     for (b, s) in tree.subs:
 973         if isinstance(s, Tree):
 974             prop_format(s)
 975
 976     # If all entries in SUBS have the same format, then
 977     # propagate that into the tree.
 978     f = None
 979     for (b, s) in tree.subs:
 980         if f is None:
 981             f = s.base
 982             if f is None:
 983                 return
 984         if f is not s.base:
 985             return
 986     tree.base = f
 987 # end prop_format
 988
 989
 990 def main():
 991     global arguments
 992     global formats
 993     global patterns
 994     global allpatterns
 995     global translate_scope
 996     global translate_prefix
 997     global output_fd
 998     global output_file
 999     global input_file
1000     global insnwidth
1001     global insntype
1002     global insnmask
1003     global decode_function
1004     global variablewidth
1005
1006     decode_scope = 'static '
1007
1008     long_opts = ['decode=', 'translate=', 'output=', 'insnwidth=',
1009                  'static-decode=', 'varinsnwidth=']
1010     try:
1011         (opts, args) = getopt.getopt(sys.argv[1:], 'o:vw:', long_opts)
1012     except getopt.GetoptError as err:
1013         error(0, err)
1014     for o, a in opts:
1015         if o in ('-o', '--output'):
1016             output_file = a
1017         elif o == '--decode':
1018             decode_function = a
1019             decode_scope = ''
1020         elif o == '--static-decode':
1021             decode_function = a
1022         elif o == '--translate':
1023             translate_prefix = a
1024             translate_scope = ''
1025         elif o in ('-w', '--insnwidth', '--varinsnwidth'):
1026             if o == '--varinsnwidth':
1027                 variablewidth = True
1028             insnwidth = int(a)
1029             if insnwidth == 16:
1030                 insntype = 'uint16_t'
1031                 insnmask = 0xffff
1032             elif insnwidth != 32:
1033                 error(0, 'cannot handle insns of width', insnwidth)
1034         else:
1035             assert False, 'unhandled option'
1036
1037     if len(args) < 1:
1038         error(0, 'missing input file')
1039     for filename in args:
1040         input_file = filename
1041         f = open(filename, 'r')
1042         parse_file(f)
1043         f.close()
1044
1045     t = build_tree(patterns, 0, 0)
1046     prop_format(t)
1047
1048     if output_file:
1049         output_fd = open(output_file, 'w')
1050     else:
1051         output_fd = sys.stdout
1052
1053     output_autogen()
1054     for n in sorted(arguments.keys()):
1055         f = arguments[n]
1056         f.output_def()
1057
1058     # A single translate function can be invoked for different patterns.
1059     # Make sure that the argument sets are the same, and declare the
1060     # function only once.
1061     out_pats = {}
1062     for i in allpatterns:
1063         if i.name in out_pats:
1064             p = out_pats[i.name]
1065             if i.base.base != p.base.base:
1066                 error(0, i.name, ' has conflicting argument sets')
1067         else:
1068             i.output_decl()
1069             out_pats[i.name] = i
1070     output('\n')
1071
1072     for n in sorted(formats.keys()):
1073         f = formats[n]
1074         f.output_extract()
1075
1076     output(decode_scope, 'bool ', decode_function,
1077            '(DisasContext *ctx, ', insntype, ' insn)\n{\n')
1078
1079     i4 = str_indent(4)
1080
1081     if len(allpatterns) != 0:
1082         output(i4, 'union {\n')
1083         for n in sorted(arguments.keys()):
1084             f = arguments[n]
1085             output(i4, i4, f.struct_name(), ' f_', f.name, ';\n')
1086         output(i4, '} u;\n\n')
1087         t.output_code(4, False, 0, 0)
1088
1089     output(i4, 'return false;\n')
1090     output('}\n')
1091
1092     if output_file:
1093         output_fd.close()
1094 # end main
1095
1096
1097 if __name__ == '__main__':
1098     main()