scripts/checkkconfigsymbols.py

   1 #!/usr/bin/env python3
   2 # SPDX-License-Identifier: GPL-2.0-only
   3
   4 """Find Kconfig symbols that are referenced but not defined."""
   5
   6 # (c) 2014-2017 Valentin Rothberg <valentinrothberg@gmail.com>
   7 # (c) 2014 Stefan Hengelein <stefan.hengelein@fau.de>
   8 #
   9
  10
  11 import argparse
  12 import difflib
  13 import os
  14 import re
  15 import signal
  16 import subprocess
  17 import sys
  18 from multiprocessing import Pool, cpu_count
  19
  20
  21 # regex expressions
  22 OPERATORS = r"&|\(|\)|\||\!"
  23 SYMBOL = r"(?:\w*[A-Z0-9]\w*){2,}"
  24 DEF = r"^\s*(?:menu){,1}config\s+(" + SYMBOL + r")\s*"
  25 EXPR = r"(?:" + OPERATORS + r"|\s|" + SYMBOL + r")+"
  26 DEFAULT = r"default\s+.*?(?:if\s.+){,1}"
  27 STMT = r"^\s*(?:if|select|imply|depends\s+on|(?:" + DEFAULT + r"))\s+" + EXPR
  28 SOURCE_SYMBOL = r"(?:\W|\b)+[D]{,1}CONFIG_(" + SYMBOL + r")"
  29
  30 # regex objects
  31 REGEX_FILE_KCONFIG = re.compile(r".*Kconfig[\.\w+\-]*$")
  32 REGEX_SYMBOL = re.compile(r'(?!\B)' + SYMBOL + r'(?!\B)')
  33 REGEX_SOURCE_SYMBOL = re.compile(SOURCE_SYMBOL)
  34 REGEX_KCONFIG_DEF = re.compile(DEF)
  35 REGEX_KCONFIG_EXPR = re.compile(EXPR)
  36 REGEX_KCONFIG_STMT = re.compile(STMT)
  37 REGEX_KCONFIG_HELP = re.compile(r"^\s+(help|---help---)\s*$")
  38 REGEX_FILTER_SYMBOLS = re.compile(r"[A-Za-z0-9]$")
  39 REGEX_NUMERIC = re.compile(r"0[xX][0-9a-fA-F]+|[0-9]+")
  40 REGEX_QUOTES = re.compile("(\"(.*?)\")")
  41
  42
  43 def parse_options():
  44     """The user interface of this module."""
  45     usage = "Run this tool to detect Kconfig symbols that are referenced but " \
  46             "not defined in Kconfig.  If no option is specified, "             \
  47             "checkkconfigsymbols defaults to check your current tree.  "       \
  48             "Please note that specifying commits will 'git reset --hard\' "    \
  49             "your current tree!  You may save uncommitted changes to avoid "   \
  50             "losing data."
  51
  52     parser = argparse.ArgumentParser(description=usage)
  53
  54     parser.add_argument('-c', '--commit', dest='commit', action='store',
  55                         default="",
  56                         help="check if the specified commit (hash) introduces "
  57                              "undefined Kconfig symbols")
  58
  59     parser.add_argument('-d', '--diff', dest='diff', action='store',
  60                         default="",
  61                         help="diff undefined symbols between two commits "
  62                              "(e.g., -d commmit1..commit2)")
  63
  64     parser.add_argument('-f', '--find', dest='find', action='store_true',
  65                         default=False,
  66                         help="find and show commits that may cause symbols to be "
  67                              "missing (required to run with --diff)")
  68
  69     parser.add_argument('-i', '--ignore', dest='ignore', action='store',
  70                         default="",
  71                         help="ignore files matching this Python regex "
  72                              "(e.g., -i '.*defconfig')")
  73
  74     parser.add_argument('-s', '--sim', dest='sim', action='store', default="",
  75                         help="print a list of max. 10 string-similar symbols")
  76
  77     parser.add_argument('--force', dest='force', action='store_true',
  78                         default=False,
  79                         help="reset current Git tree even when it's dirty")
  80
  81     parser.add_argument('--no-color', dest='color', action='store_false',
  82                         default=True,
  83                         help="don't print colored output (default when not "
  84                              "outputting to a terminal)")
  85
  86     args = parser.parse_args()
  87
  88     if args.commit and args.diff:
  89         sys.exit("Please specify only one option at once.")
  90
  91     if args.diff and not re.match(r"^[\w\-\.\^]+\.\.[\w\-\.\^]+$", args.diff):
  92         sys.exit("Please specify valid input in the following format: "
  93                  "\'commit1..commit2\'")
  94
  95     if args.commit or args.diff:
  96         if not args.force and tree_is_dirty():
  97             sys.exit("The current Git tree is dirty (see 'git status').  "
  98                      "Running this script may\ndelete important data since it "
  99                      "calls 'git reset --hard' for some performance\nreasons. "
 100                      " Please run this script in a clean Git tree or pass "
 101                      "'--force' if you\nwant to ignore this warning and "
 102                      "continue.")
 103
 104     if args.commit:
 105         args.find = False
 106
 107     if args.ignore:
 108         try:
 109             re.match(args.ignore, "this/is/just/a/test.c")
 110         except:
 111             sys.exit("Please specify a valid Python regex.")
 112
 113     return args
 114
 115
 116 def main():
 117     """Main function of this module."""
 118     args = parse_options()
 119
 120     global COLOR
 121     COLOR = args.color and sys.stdout.isatty()
 122
 123     if args.sim and not args.commit and not args.diff:
 124         sims = find_sims(args.sim, args.ignore)
 125         if sims:
 126             print("%s: %s" % (yel("Similar symbols"), ', '.join(sims)))
 127         else:
 128             print("%s: no similar symbols found" % yel("Similar symbols"))
 129         sys.exit(0)
 130
 131     # dictionary of (un)defined symbols
 132     defined = {}
 133     undefined = {}
 134
 135     if args.commit or args.diff:
 136         head = get_head()
 137
 138         # get commit range
 139         commit_a = None
 140         commit_b = None
 141         if args.commit:
 142             commit_a = args.commit + "~"
 143             commit_b = args.commit
 144         elif args.diff:
 145             split = args.diff.split("..")
 146             commit_a = split[0]
 147             commit_b = split[1]
 148             undefined_a = {}
 149             undefined_b = {}
 150
 151         # get undefined items before the commit
 152         reset(commit_a)
 153         undefined_a, _ = check_symbols(args.ignore)
 154
 155         # get undefined items for the commit
 156         reset(commit_b)
 157         undefined_b, defined = check_symbols(args.ignore)
 158
 159         # report cases that are present for the commit but not before
 160         for symbol in sorted(undefined_b):
 161             # symbol has not been undefined before
 162             if symbol not in undefined_a:
 163                 files = sorted(undefined_b.get(symbol))
 164                 undefined[symbol] = files
 165             # check if there are new files that reference the undefined symbol
 166             else:
 167                 files = sorted(undefined_b.get(symbol) -
 168                                undefined_a.get(symbol))
 169                 if files:
 170                     undefined[symbol] = files
 171
 172         # reset to head
 173         reset(head)
 174
 175     # default to check the entire tree
 176     else:
 177         undefined, defined = check_symbols(args.ignore)
 178
 179     # now print the output
 180     for symbol in sorted(undefined):
 181         print(red(symbol))
 182
 183         files = sorted(undefined.get(symbol))
 184         print("%s: %s" % (yel("Referencing files"), ", ".join(files)))
 185
 186         sims = find_sims(symbol, args.ignore, defined)
 187         sims_out = yel("Similar symbols")
 188         if sims:
 189             print("%s: %s" % (sims_out, ', '.join(sims)))
 190         else:
 191             print("%s: %s" % (sims_out, "no similar symbols found"))
 192
 193         if args.find:
 194             print("%s:" % yel("Commits changing symbol"))
 195             commits = find_commits(symbol, args.diff)
 196             if commits:
 197                 for commit in commits:
 198                     commit = commit.split(" ", 1)
 199                     print("\t- %s (\"%s\")" % (yel(commit[0]), commit[1]))
 200             else:
 201                 print("\t- no commit found")
 202         print()  # new line
 203
 204
 205 def reset(commit):
 206     """Reset current git tree to %commit."""
 207     execute(["git", "reset", "--hard", commit])
 208
 209
 210 def yel(string):
 211     """
 212     Color %string yellow.
 213     """
 214     return "\033[33m%s\033[0m" % string if COLOR else string
 215
 216
 217 def red(string):
 218     """
 219     Color %string red.
 220     """
 221     return "\033[31m%s\033[0m" % string if COLOR else string
 222
 223
 224 def execute(cmd):
 225     """Execute %cmd and return stdout.  Exit in case of error."""
 226     try:
 227         stdout = subprocess.check_output(cmd, stderr=subprocess.STDOUT, shell=False)
 228         stdout = stdout.decode(errors='replace')
 229     except subprocess.CalledProcessError as fail:
 230         exit(fail)
 231     return stdout
 232
 233
 234 def find_commits(symbol, diff):
 235     """Find commits changing %symbol in the given range of %diff."""
 236     commits = execute(["git", "log", "--pretty=oneline",
 237                        "--abbrev-commit", "-G",
 238                        symbol, diff])
 239     return [x for x in commits.split("\n") if x]
 240
 241
 242 def tree_is_dirty():
 243     """Return true if the current working tree is dirty (i.e., if any file has
 244     been added, deleted, modified, renamed or copied but not committed)."""
 245     stdout = execute(["git", "status", "--porcelain"])
 246     for line in stdout:
 247         if re.findall(r"[URMADC]{1}", line[:2]):
 248             return True
 249     return False
 250
 251
 252 def get_head():
 253     """Return commit hash of current HEAD."""
 254     stdout = execute(["git", "rev-parse", "HEAD"])
 255     return stdout.strip('\n')
 256
 257
 258 def partition(lst, size):
 259     """Partition list @lst into eveni-sized lists of size @size."""
 260     return [lst[i::size] for i in range(size)]
 261
 262
 263 def init_worker():
 264     """Set signal handler to ignore SIGINT."""
 265     signal.signal(signal.SIGINT, signal.SIG_IGN)
 266
 267
 268 def find_sims(symbol, ignore, defined=[]):
 269     """Return a list of max. ten Kconfig symbols that are string-similar to
 270     @symbol."""
 271     if defined:
 272         return difflib.get_close_matches(symbol, set(defined), 10)
 273
 274     pool = Pool(cpu_count(), init_worker)
 275     kfiles = []
 276     for gitfile in get_files():
 277         if REGEX_FILE_KCONFIG.match(gitfile):
 278             kfiles.append(gitfile)
 279
 280     arglist = []
 281     for part in partition(kfiles, cpu_count()):
 282         arglist.append((part, ignore))
 283
 284     for res in pool.map(parse_kconfig_files, arglist):
 285         defined.extend(res[0])
 286
 287     return difflib.get_close_matches(symbol, set(defined), 10)
 288
 289
 290 def get_files():
 291     """Return a list of all files in the current git directory."""
 292     # use 'git ls-files' to get the worklist
 293     stdout = execute(["git", "ls-files"])
 294     if len(stdout) > 0 and stdout[-1] == "\n":
 295         stdout = stdout[:-1]
 296
 297     files = []
 298     for gitfile in stdout.rsplit("\n"):
 299         if ".git" in gitfile or "ChangeLog" in gitfile or      \
 300                 ".log" in gitfile or os.path.isdir(gitfile) or \
 301                 gitfile.startswith("tools/"):
 302             continue
 303         files.append(gitfile)
 304     return files
 305
 306
 307 def check_symbols(ignore):
 308     """Find undefined Kconfig symbols and return a dict with the symbol as key
 309     and a list of referencing files as value.  Files matching %ignore are not
 310     checked for undefined symbols."""
 311     pool = Pool(cpu_count(), init_worker)
 312     try:
 313         return check_symbols_helper(pool, ignore)
 314     except KeyboardInterrupt:
 315         pool.terminate()
 316         pool.join()
 317         sys.exit(1)
 318
 319
 320 def check_symbols_helper(pool, ignore):
 321     """Helper method for check_symbols().  Used to catch keyboard interrupts in
 322     check_symbols() in order to properly terminate running worker processes."""
 323     source_files = []
 324     kconfig_files = []
 325     defined_symbols = []
 326     referenced_symbols = dict()  # {file: [symbols]}
 327
 328     for gitfile in get_files():
 329         if REGEX_FILE_KCONFIG.match(gitfile):
 330             kconfig_files.append(gitfile)
 331         else:
 332             if ignore and not re.match(ignore, gitfile):
 333                 continue
 334             # add source files that do not match the ignore pattern
 335             source_files.append(gitfile)
 336
 337     # parse source files
 338     arglist = partition(source_files, cpu_count())
 339     for res in pool.map(parse_source_files, arglist):
 340         referenced_symbols.update(res)
 341
 342     # parse kconfig files
 343     arglist = []
 344     for part in partition(kconfig_files, cpu_count()):
 345         arglist.append((part, ignore))
 346     for res in pool.map(parse_kconfig_files, arglist):
 347         defined_symbols.extend(res[0])
 348         referenced_symbols.update(res[1])
 349     defined_symbols = set(defined_symbols)
 350
 351     # inverse mapping of referenced_symbols to dict(symbol: [files])
 352     inv_map = dict()
 353     for _file, symbols in referenced_symbols.items():
 354         for symbol in symbols:
 355             inv_map[symbol] = inv_map.get(symbol, set())
 356             inv_map[symbol].add(_file)
 357     referenced_symbols = inv_map
 358
 359     undefined = {}  # {symbol: [files]}
 360     for symbol in sorted(referenced_symbols):
 361         # filter some false positives
 362         if symbol == "FOO" or symbol == "BAR" or \
 363                 symbol == "FOO_BAR" or symbol == "XXX":
 364             continue
 365         if symbol not in defined_symbols:
 366             if symbol.endswith("_MODULE"):
 367                 # avoid false positives for kernel modules
 368                 if symbol[:-len("_MODULE")] in defined_symbols:
 369                     continue
 370             undefined[symbol] = referenced_symbols.get(symbol)
 371     return undefined, defined_symbols
 372
 373
 374 def parse_source_files(source_files):
 375     """Parse each source file in @source_files and return dictionary with source
 376     files as keys and lists of references Kconfig symbols as values."""
 377     referenced_symbols = dict()
 378     for sfile in source_files:
 379         referenced_symbols[sfile] = parse_source_file(sfile)
 380     return referenced_symbols
 381
 382
 383 def parse_source_file(sfile):
 384     """Parse @sfile and return a list of referenced Kconfig symbols."""
 385     lines = []
 386     references = []
 387
 388     if not os.path.exists(sfile):
 389         return references
 390
 391     with open(sfile, "r", encoding='utf-8', errors='replace') as stream:
 392         lines = stream.readlines()
 393
 394     for line in lines:
 395         if "CONFIG_" not in line:
 396             continue
 397         symbols = REGEX_SOURCE_SYMBOL.findall(line)
 398         for symbol in symbols:
 399             if not REGEX_FILTER_SYMBOLS.search(symbol):
 400                 continue
 401             references.append(symbol)
 402
 403     return references
 404
 405
 406 def get_symbols_in_line(line):
 407     """Return mentioned Kconfig symbols in @line."""
 408     return REGEX_SYMBOL.findall(line)
 409
 410
 411 def parse_kconfig_files(args):
 412     """Parse kconfig files and return tuple of defined and references Kconfig
 413     symbols.  Note, @args is a tuple of a list of files and the @ignore
 414     pattern."""
 415     kconfig_files = args[0]
 416     ignore = args[1]
 417     defined_symbols = []
 418     referenced_symbols = dict()
 419
 420     for kfile in kconfig_files:
 421         defined, references = parse_kconfig_file(kfile)
 422         defined_symbols.extend(defined)
 423         if ignore and re.match(ignore, kfile):
 424             # do not collect references for files that match the ignore pattern
 425             continue
 426         referenced_symbols[kfile] = references
 427     return (defined_symbols, referenced_symbols)
 428
 429
 430 def parse_kconfig_file(kfile):
 431     """Parse @kfile and update symbol definitions and references."""
 432     lines = []
 433     defined = []
 434     references = []
 435     skip = False
 436
 437     if not os.path.exists(kfile):
 438         return defined, references
 439
 440     with open(kfile, "r", encoding='utf-8', errors='replace') as stream:
 441         lines = stream.readlines()
 442
 443     for i in range(len(lines)):
 444         line = lines[i]
 445         line = line.strip('\n')
 446         line = line.split("#")[0]  # ignore comments
 447
 448         if REGEX_KCONFIG_DEF.match(line):
 449             symbol_def = REGEX_KCONFIG_DEF.findall(line)
 450             defined.append(symbol_def[0])
 451             skip = False
 452         elif REGEX_KCONFIG_HELP.match(line):
 453             skip = True
 454         elif skip:
 455             # ignore content of help messages
 456             pass
 457         elif REGEX_KCONFIG_STMT.match(line):
 458             line = REGEX_QUOTES.sub("", line)
 459             symbols = get_symbols_in_line(line)
 460             # multi-line statements
 461             while line.endswith("\\"):
 462                 i += 1
 463                 line = lines[i]
 464                 line = line.strip('\n')
 465                 symbols.extend(get_symbols_in_line(line))
 466             for symbol in set(symbols):
 467                 if REGEX_NUMERIC.match(symbol):
 468                     # ignore numeric values
 469                     continue
 470                 references.append(symbol)
 471
 472     return defined, references
 473
 474
 475 if __name__ == "__main__":
 476     main()