micro_asm.py revision 4566
1# Copyright (c) 2003-2005 The Regents of The University of Michigan 2# All rights reserved. 3# 4# Redistribution and use in source and binary forms, with or without 5# modification, are permitted provided that the following conditions are 6# met: redistributions of source code must retain the above copyright 7# notice, this list of conditions and the following disclaimer; 8# redistributions in binary form must reproduce the above copyright 9# notice, this list of conditions and the following disclaimer in the 10# documentation and/or other materials provided with the distribution; 11# neither the name of the copyright holders nor the names of its 12# contributors may be used to endorse or promote products derived from 13# this software without specific prior written permission. 14# 15# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 16# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 17# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 18# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT 19# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 20# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 21# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 22# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 23# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 24# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 25# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 26# 27# Authors: Gabe Black 28 29import os 30import sys 31import re 32import string 33import traceback 34# get type names 35from types import * 36 37# Prepend the directory where the PLY lex & yacc modules are found 38# to the search path. 39sys.path[0:0] = [os.environ['M5_PLY']] 40 41from ply import lex 42from ply import yacc 43 44########################################################################## 45# 46# Base classes for use outside of the assembler 47# 48########################################################################## 49 50class Micro_Container(object): 51 def __init__(self, name): 52 self.microops = [] 53 self.name = name 54 self.directives = {} 55 self.micro_classes = {} 56 self.labels = {} 57 58 def add_microop(self, microop): 59 self.microops.append(microop) 60 61 def __str__(self): 62 string = "%s:\n" % self.name 63 for microop in self.microops: 64 string += " %s\n" % microop 65 return string 66 67class Combinational_Macroop(Micro_Container): 68 pass 69 70class Rom_Macroop(object): 71 def __init__(self, name, target): 72 self.name = name 73 self.target = target 74 75 def __str__(self): 76 return "%s: %s\n" % (self.name, self.target) 77 78class Rom(Micro_Container): 79 def __init__(self, name): 80 super(Rom, self).__init__(name) 81 self.externs = {} 82 83########################################################################## 84# 85# Support classes 86# 87########################################################################## 88 89class Label(object): 90 def __init__(self): 91 self.extern = False 92 self.name = "" 93 94class Block(object): 95 def __init__(self): 96 self.statements = [] 97 98class Statement(object): 99 def __init__(self): 100 self.is_microop = False 101 self.is_directive = False 102 self.params = "" 103 104class Microop(Statement): 105 def __init__(self): 106 super(Microop, self).__init__() 107 self.mnemonic = "" 108 self.labels = [] 109 self.is_microop = True 110 111class Directive(Statement): 112 def __init__(self): 113 super(Directive, self).__init__() 114 self.name = "" 115 self.is_directive = True 116 117########################################################################## 118# 119# Functions that handle common tasks 120# 121########################################################################## 122 123def print_error(message): 124 print 125 print "*** %s" % message 126 print 127 128def handle_statement(parser, container, statement): 129 if statement.is_microop: 130 try: 131 microop = eval('parser.microops[statement.mnemonic](%s)' % 132 statement.params) 133 except: 134 print_error("Error creating microop object.") 135 raise 136 try: 137 for label in statement.labels: 138 container.labels[label.name] = microop 139 if label.extern: 140 container.externs[label.name] = microop 141 container.add_microop(microop) 142 except: 143 print_error("Error adding microop.") 144 raise 145 elif statement.is_directive: 146 try: 147 eval('container.directives[statement.name](%s)' % statement.params) 148 except: 149 print_error("Error executing directive.") 150 print container.directives 151 raise 152 else: 153 raise Exception, "Didn't recognize the type of statement", statement 154 155########################################################################## 156# 157# Lexer specification 158# 159########################################################################## 160 161# Error handler. Just call exit. Output formatted to work under 162# Emacs compile-mode. Optional 'print_traceback' arg, if set to True, 163# prints a Python stack backtrace too (can be handy when trying to 164# debug the parser itself). 165def error(lineno, string, print_traceback = False): 166 # Print a Python stack backtrace if requested. 167 if (print_traceback): 168 traceback.print_exc() 169 if lineno != 0: 170 line_str = "%d:" % lineno 171 else: 172 line_str = "" 173 sys.exit("%s %s" % (line_str, string)) 174 175reserved = ('DEF', 'MACROOP', 'ROM', 'EXTERN') 176 177tokens = reserved + ( 178 # identifier 179 'ID', 180 # arguments for microops and directives 181 'PARAMS', 182 183 'LPAREN', 'RPAREN', 184 'LBRACE', 'RBRACE', 185 'COLON', 'SEMI', 'DOT', 186 'NEWLINE' 187 ) 188 189# New lines are ignored at the top level, but they end statements in the 190# assembler 191states = ( 192 ('asm', 'exclusive'), 193 ('params', 'exclusive'), 194) 195 196reserved_map = { } 197for r in reserved: 198 reserved_map[r.lower()] = r 199 200# Ignore comments 201def t_ANY_COMMENT(t): 202 r'\#[^\n]*(?=\n)' 203 204def t_ANY_MULTILINECOMMENT(t): 205 r'/\*([^/]|((?<!\*)/))*\*/' 206 207# A colon marks the end of a label. It should follow an ID which will 208# put the lexer in the "params" state. Seeing the colon will put it back 209# in the "asm" state since it knows it saw a label and not a mnemonic. 210def t_params_COLON(t): 211 r':' 212 t.lexer.begin('asm') 213 return t 214 215# An "ID" in the micro assembler is either a label, directive, or mnemonic 216# If it's either a directive or a mnemonic, it will be optionally followed by 217# parameters. If it's a label, the following colon will make the lexer stop 218# looking for parameters. 219def t_asm_ID(t): 220 r'[A-Za-z_]\w*' 221 t.type = reserved_map.get(t.value, 'ID') 222 t.lexer.begin('params') 223 return t 224 225# If there is a label and you're -not- in the assember (which would be caught 226# above), don't start looking for parameters. 227def t_ANY_ID(t): 228 r'[A-Za-z_]\w*' 229 t.type = reserved_map.get(t.value, 'ID') 230 return t 231 232# Parameters are a string of text which don't contain an unescaped statement 233# statement terminator, ie a newline or semi colon. 234def t_params_PARAMS(t): 235 r'([^\n;\\]|(\\[\n;\\]))+' 236 t.lineno += t.value.count('\n') 237 unescapeParamsRE = re.compile(r'(\\[\n;\\])') 238 def unescapeParams(mo): 239 val = mo.group(0) 240 print "About to sub %s for %s" % (val[1], val) 241 return val[1] 242 print "Looking for matches in %s" % t.value 243 t.value = unescapeParamsRE.sub(unescapeParams, t.value) 244 t.lexer.begin('asm') 245 return t 246 247# Braces enter and exit micro assembly 248def t_INITIAL_LBRACE(t): 249 r'\{' 250 t.lexer.begin('asm') 251 return t 252 253def t_asm_RBRACE(t): 254 r'\}' 255 t.lexer.begin('INITIAL') 256 return t 257 258# At the top level, keep track of newlines only for line counting. 259def t_INITIAL_NEWLINE(t): 260 r'\n+' 261 t.lineno += t.value.count('\n') 262 263# In the micro assembler, do line counting but also return a token. The 264# token is needed by the parser to detect the end of a statement. 265def t_asm_NEWLINE(t): 266 r'\n+' 267 t.lineno += t.value.count('\n') 268 return t 269 270# A newline or semi colon when looking for params signals that the statement 271# is over and the lexer should go back to looking for regular assembly. 272def t_params_NEWLINE(t): 273 r'\n+' 274 t.lineno += t.value.count('\n') 275 t.lexer.begin('asm') 276 return t 277 278def t_params_SEMI(t): 279 r';' 280 t.lexer.begin('asm') 281 return t 282 283# Basic regular expressions to pick out simple tokens 284t_ANY_LPAREN = r'\(' 285t_ANY_RPAREN = r'\)' 286t_ANY_SEMI = r';' 287t_ANY_DOT = r'\.' 288 289t_ANY_ignore = ' \t\x0c' 290 291def t_ANY_error(t): 292 error(t.lineno, "illegal character '%s'" % t.value[0]) 293 t.skip(1) 294 295########################################################################## 296# 297# Parser specification 298# 299########################################################################## 300 301# Start symbol for a file which may have more than one macroop or rom 302# specification. 303def p_file(t): 304 'file : opt_rom_or_macros' 305 306def p_opt_rom_or_macros_0(t): 307 'opt_rom_or_macros : ' 308 309def p_opt_rom_or_macros_1(t): 310 'opt_rom_or_macros : rom_or_macros' 311 312def p_rom_or_macros_0(t): 313 'rom_or_macros : rom_or_macro' 314 315def p_rom_or_macros_1(t): 316 'rom_or_macros : rom_or_macros rom_or_macro' 317 318def p_rom_or_macro_0(t): 319 '''rom_or_macro : rom_block 320 | macroop_def''' 321 322# Defines a section of microcode that should go in the current ROM 323def p_rom_block(t): 324 'rom_block : DEF ROM block SEMI' 325 if not t.parser.rom: 326 print_error("Rom block found, but no Rom object specified.") 327 raise TypeError, "Rom block found, but no Rom object was specified." 328 for statement in t[3].statements: 329 handle_statement(t.parser, t.parser.rom, statement) 330 t[0] = t.parser.rom 331 332# Defines a macroop that jumps to an external label in the ROM 333def p_macroop_def_0(t): 334 'macroop_def : DEF MACROOP ID LPAREN ID RPAREN SEMI' 335 if not t.parser.rom_macroop_type: 336 print_error("ROM based macroop found, but no ROM macroop class was specified.") 337 raise TypeError, "ROM based macroop found, but no ROM macroop class was specified." 338 macroop = t.parser.rom_macroop_type(t[3], t[5]) 339 t.parser.macroops[t[3]] = macroop 340 341 342# Defines a macroop that is combinationally generated 343def p_macroop_def_1(t): 344 'macroop_def : DEF MACROOP ID block SEMI' 345 try: 346 curop = t.parser.macro_type(t[3]) 347 except TypeError: 348 print_error("Error creating macroop object.") 349 raise 350 for statement in t[4].statements: 351 handle_statement(t.parser, curop, statement) 352 t.parser.macroops[t[3]] = curop 353 354# A block of statements 355def p_block(t): 356 'block : LBRACE statements RBRACE' 357 block = Block() 358 block.statements = t[2] 359 t[0] = block 360 361def p_statements_0(t): 362 'statements : statement' 363 if t[1]: 364 t[0] = [t[1]] 365 else: 366 t[0] = [] 367 368def p_statements_1(t): 369 'statements : statements statement' 370 if t[2]: 371 t[1].append(t[2]) 372 t[0] = t[1] 373 374def p_statement(t): 375 'statement : content_of_statement end_of_statement' 376 t[0] = t[1] 377 378# A statement can be a microop or an assembler directive 379def p_content_of_statement_0(t): 380 '''content_of_statement : microop 381 | directive''' 382 t[0] = t[1] 383 384# Ignore empty statements 385def p_content_of_statement_1(t): 386 'content_of_statement : ' 387 pass 388 389# Statements are ended by newlines or a semi colon 390def p_end_of_statement(t): 391 '''end_of_statement : NEWLINE 392 | SEMI''' 393 pass 394 395# Different flavors of microop to avoid shift/reduce errors 396def p_microop_0(t): 397 'microop : labels ID' 398 microop = Microop() 399 microop.labels = t[1] 400 microop.mnemonic = t[2] 401 t[0] = microop 402 403def p_microop_1(t): 404 'microop : ID' 405 microop = Microop() 406 microop.mnemonic = t[1] 407 t[0] = microop 408 409def p_microop_2(t): 410 'microop : labels ID PARAMS' 411 microop = Microop() 412 microop.labels = t[1] 413 microop.mnemonic = t[2] 414 microop.params = t[3] 415 t[0] = microop 416 417def p_microop_3(t): 418 'microop : ID PARAMS' 419 microop = Microop() 420 microop.mnemonic = t[1] 421 microop.params = t[2] 422 t[0] = microop 423 424# Labels in the microcode 425def p_labels_0(t): 426 'labels : label' 427 t[0] = [t[1]] 428 429def p_labels_1(t): 430 'labels : labels label' 431 t[1].append(t[2]) 432 t[0] = t[1] 433 434def p_label_0(t): 435 'label : ID COLON' 436 label = Label() 437 label.is_extern = False 438 label.text = t[1] 439 t[0] = label 440 441def p_label_1(t): 442 'label : EXTERN ID COLON' 443 label = Label() 444 label.is_extern = True 445 label.text = t[2] 446 t[0] = label 447 448# Directives for the macroop 449def p_directive_0(t): 450 'directive : DOT ID' 451 directive = Directive() 452 directive.name = t[2] 453 t[0] = directive 454 455def p_directive_1(t): 456 'directive : DOT ID PARAMS' 457 directive = Directive() 458 directive.name = t[2] 459 directive.params = t[3] 460 t[0] = directive 461 462# Parse error handler. Note that the argument here is the offending 463# *token*, not a grammar symbol (hence the need to use t.value) 464def p_error(t): 465 if t: 466 error(t.lineno, "syntax error at '%s'" % t.value) 467 else: 468 error(0, "unknown syntax error", True) 469 470class MicroAssembler(object): 471 472 def __init__(self, macro_type, microops, 473 rom = None, rom_macroop_type = None): 474 self.lexer = lex.lex() 475 self.parser = yacc.yacc() 476 self.parser.macro_type = macro_type 477 self.parser.macroops = {} 478 self.parser.microops = microops 479 self.parser.rom = rom 480 self.parser.rom_macroop_type = rom_macroop_type 481 482 def assemble(self, asm): 483 self.parser.parse(asm, lexer=self.lexer) 484 # Begin debug printing 485 for macroop in self.parser.macroops.values(): 486 print macroop 487 print self.parser.rom 488 # End debug printing 489 macroops = self.parser.macroops 490 self.parser.macroops = {} 491 return macroops 492