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+\-]*$")
  32 REGEX_SYMBOL = re.compile(r'(?!\B)' + SYMBOL + r'(?!\B)')
  33 REGEX_SOURCE_SYMBOL = re.compile(SOURCE_SYMBOL)
  34 REGEX_KCONFIG_DEF = re.compile(DEF)
  35 REGEX_KCONFIG_EXPR = re.compile(EXPR)
  36 REGEX_KCONFIG_STMT = re.compile(STMT)
  37 REGEX_FILTER_SYMBOLS = re.compile(r"[A-Za-z0-9]$")
  38 REGEX_NUMERIC = re.compile(r"0[xX][0-9a-fA-F]+|[0-9]+")
  39 REGEX_QUOTES = re.compile("(\"(.*?)\")")
  40
  41
  42 def parse_options():
  43     """The user interface of this module."""
  44     usage = "Run this tool to detect Kconfig symbols that are referenced but " \
  45             "not defined in Kconfig.  If no option is specified, "             \
  46             "checkkconfigsymbols defaults to check your current tree.  "       \
  47             "Please note that specifying commits will 'git reset --hard\' "    \
  48             "your current tree!  You may save uncommitted changes to avoid "   \
  49             "losing data."
  50
  51     parser = argparse.ArgumentParser(description=usage)
  52
  53     parser.add_argument('-c', '--commit', dest='commit', action='store',
  54                         default="",
  55                         help="check if the specified commit (hash) introduces "
  56                              "undefined Kconfig symbols")
  57
  58     parser.add_argument('-d', '--diff', dest='diff', action='store',
  59                         default="",
  60                         help="diff undefined symbols between two commits "
  61                              "(e.g., -d commmit1..commit2)")
  62
  63     parser.add_argument('-f', '--find', dest='find', action='store_true',
  64                         default=False,
  65                         help="find and show commits that may cause symbols to be "
  66                              "missing (required to run with --diff)")
  67
  68     parser.add_argument('-i', '--ignore', dest='ignore', action='store',
  69                         default="",
  70                         help="ignore files matching this Python regex "
  71                              "(e.g., -i '.*defconfig')")
  72
  73     parser.add_argument('-s', '--sim', dest='sim', action='store', default="",
  74                         help="print a list of max. 10 string-similar symbols")
  75
  76     parser.add_argument('--force', dest='force', action='store_true',
  77                         default=False,
  78                         help="reset current Git tree even when it's dirty")
  79
  80     parser.add_argument('--no-color', dest='color', action='store_false',
  81                         default=True,
  82                         help="don't print colored output (default when not "
  83                              "outputting to a terminal)")
  84
  85     args = parser.parse_args()
  86
  87     if args.commit and args.diff:
  88         sys.exit("Please specify only one option at once.")
  89
  90     if args.diff and not re.match(r"^[\w\-\.\^]+\.\.[\w\-\.\^]+$", args.diff):
  91         sys.exit("Please specify valid input in the following format: "
  92                  "\'commit1..commit2\'")
  93
  94     if args.commit or args.diff:
  95         if not args.force and tree_is_dirty():
  96             sys.exit("The current Git tree is dirty (see 'git status').  "
  97                      "Running this script may\ndelete important data since it "
  98                      "calls 'git reset --hard' for some performance\nreasons. "
  99                      " Please run this script in a clean Git tree or pass "
 100                      "'--force' if you\nwant to ignore this warning and "
 101                      "continue.")
 102
 103     if args.commit:
 104         if args.commit.startswith('HEAD'):
 105             sys.exit("The --commit option can't use the HEAD ref")
 106
 107         args.find = False
 108
 109     if args.ignore:
 110         try:
 111             re.match(args.ignore, "this/is/just/a/test.c")
 112         except:
 113             sys.exit("Please specify a valid Python regex.")
 114
 115     return args
 116
 117
 118 def print_undefined_symbols():
 119     """Main function of this module."""
 120     args = parse_options()
 121
 122     global COLOR
 123     COLOR = args.color and sys.stdout.isatty()
 124
 125     if args.sim and not args.commit and not args.diff:
 126         sims = find_sims(args.sim, args.ignore)
 127         if sims:
 128             print("%s: %s" % (yel("Similar symbols"), ', '.join(sims)))
 129         else:
 130             print("%s: no similar symbols found" % yel("Similar symbols"))
 131         sys.exit(0)
 132
 133     # dictionary of (un)defined symbols
 134     defined = {}
 135     undefined = {}
 136
 137     if args.commit or args.diff:
 138         head = get_head()
 139
 140         # get commit range
 141         commit_a = None
 142         commit_b = None
 143         if args.commit:
 144             commit_a = args.commit + "~"
 145             commit_b = args.commit
 146         elif args.diff:
 147             split = args.diff.split("..")
 148             commit_a = split[0]
 149             commit_b = split[1]
 150             undefined_a = {}
 151             undefined_b = {}
 152
 153         # get undefined items before the commit
 154         reset(commit_a)
 155         undefined_a, _ = check_symbols(args.ignore)
 156
 157         # get undefined items for the commit
 158         reset(commit_b)
 159         undefined_b, defined = check_symbols(args.ignore)
 160
 161         # report cases that are present for the commit but not before
 162         for symbol in sorted(undefined_b):
 163             # symbol has not been undefined before
 164             if symbol not in undefined_a:
 165                 files = sorted(undefined_b.get(symbol))
 166                 undefined[symbol] = files
 167             # check if there are new files that reference the undefined symbol
 168             else:
 169                 files = sorted(undefined_b.get(symbol) -
 170                                undefined_a.get(symbol))
 171                 if files:
 172                     undefined[symbol] = files
 173
 174         # reset to head
 175         reset(head)
 176
 177     # default to check the entire tree
 178     else:
 179         undefined, defined = check_symbols(args.ignore)
 180
 181     # now print the output
 182     for symbol in sorted(undefined):
 183         print(red(symbol))
 184
 185         files = sorted(undefined.get(symbol))
 186         print("%s: %s" % (yel("Referencing files"), ", ".join(files)))
 187
 188         sims = find_sims(symbol, args.ignore, defined)
 189         sims_out = yel("Similar symbols")
 190         if sims:
 191             print("%s: %s" % (sims_out, ', '.join(sims)))
 192         else:
 193             print("%s: %s" % (sims_out, "no similar symbols found"))
 194
 195         if args.find:
 196             print("%s:" % yel("Commits changing symbol"))
 197             commits = find_commits(symbol, args.diff)
 198             if commits:
 199                 for commit in commits:
 200                     commit = commit.split(" ", 1)
 201                     print("\t- %s (\"%s\")" % (yel(commit[0]), commit[1]))
 202             else:
 203                 print("\t- no commit found")
 204         print()  # new line
 205
 206
 207 def reset(commit):
 208     """Reset current git tree to %commit."""
 209     execute(["git", "reset", "--hard", commit])
 210
 211
 212 def yel(string):
 213     """
 214     Color %string yellow.
 215     """
 216     return "\033[33m%s\033[0m" % string if COLOR else string
 217
 218
 219 def red(string):
 220     """
 221     Color %string red.
 222     """
 223     return "\033[31m%s\033[0m" % string if COLOR else string
 224
 225
 226 def execute(cmd):
 227     """Execute %cmd and return stdout.  Exit in case of error."""
 228     try:
 229         stdout = subprocess.check_output(cmd, stderr=subprocess.STDOUT, shell=False)
 230         stdout = stdout.decode(errors='replace')
 231     except subprocess.CalledProcessError as fail:
 232         exit(fail)
 233     return stdout
 234
 235
 236 def find_commits(symbol, diff):
 237     """Find commits changing %symbol in the given range of %diff."""
 238     commits = execute(["git", "log", "--pretty=oneline",
 239                        "--abbrev-commit", "-G",
 240                        symbol, diff])
 241     return [x for x in commits.split("\n") if x]
 242
 243
 244 def tree_is_dirty():
 245     """Return true if the current working tree is dirty (i.e., if any file has
 246     been added, deleted, modified, renamed or copied but not committed)."""
 247     stdout = execute(["git", "status", "--porcelain"])
 248     for line in stdout:
 249         if re.findall(r"[URMADC]{1}", line[:2]):
 250             return True
 251     return False
 252
 253
 254 def get_head():
 255     """Return commit hash of current HEAD."""
 256     stdout = execute(["git", "rev-parse", "HEAD"])
 257     return stdout.strip('\n')
 258
 259
 260 def partition(lst, size):
 261     """Partition list @lst into eveni-sized lists of size @size."""
 262     return [lst[i::size] for i in range(size)]
 263
 264
 265 def init_worker():
 266     """Set signal handler to ignore SIGINT."""
 267     signal.signal(signal.SIGINT, signal.SIG_IGN)
 268
 269
 270 def find_sims(symbol, ignore, defined=[]):
 271     """Return a list of max. ten Kconfig symbols that are string-similar to
 272     @symbol."""
 273     if defined:
 274         return difflib.get_close_matches(symbol, set(defined), 10)
 275
 276     pool = Pool(cpu_count(), init_worker)
 277     kfiles = []
 278     for gitfile in get_files():
 279         if REGEX_FILE_KCONFIG.match(gitfile):
 280             kfiles.append(gitfile)
 281
 282     arglist = []
 283     for part in partition(kfiles, cpu_count()):
 284         arglist.append((part, ignore))
 285
 286     for res in pool.map(parse_kconfig_files, arglist):
 287         defined.extend(res[0])
 288
 289     return difflib.get_close_matches(symbol, set(defined), 10)
 290
 291
 292 def get_files():
 293     """Return a list of all files in the current git directory."""
 294     # use 'git ls-files' to get the worklist
 295     stdout = execute(["git", "ls-files"])
 296     if len(stdout) > 0 and stdout[-1] == "\n":
 297         stdout = stdout[:-1]
 298
 299     files = []
 300     for gitfile in stdout.rsplit("\n"):
 301         if ".git" in gitfile or "ChangeLog" in gitfile or      \
 302                 ".log" in gitfile or os.path.isdir(gitfile) or \
 303                 gitfile.startswith("tools/"):
 304             continue
 305         files.append(gitfile)
 306     return files
 307
 308
 309 def check_symbols(ignore):
 310     """Find undefined Kconfig symbols and return a dict with the symbol as key
 311     and a list of referencing files as value.  Files matching %ignore are not
 312     checked for undefined symbols."""
 313     pool = Pool(cpu_count(), init_worker)
 314     try:
 315         return check_symbols_helper(pool, ignore)
 316     except KeyboardInterrupt:
 317         pool.terminate()
 318         pool.join()
 319         sys.exit(1)
 320
 321
 322 def check_symbols_helper(pool, ignore):
 323     """Helper method for check_symbols().  Used to catch keyboard interrupts in
 324     check_symbols() in order to properly terminate running worker processes."""
 325     source_files = []
 326     kconfig_files = []
 327     defined_symbols = []
 328     referenced_symbols = dict()  # {file: [symbols]}
 329
 330     for gitfile in get_files():
 331         if REGEX_FILE_KCONFIG.match(gitfile):
 332             kconfig_files.append(gitfile)
 333         else:
 334             if ignore and re.match(ignore, gitfile):
 335                 continue
 336             # add source files that do not match the ignore pattern
 337             source_files.append(gitfile)
 338
 339     # parse source files
 340     arglist = partition(source_files, cpu_count())
 341     for res in pool.map(parse_source_files, arglist):
 342         referenced_symbols.update(res)
 343
 344     # parse kconfig files
 345     arglist = []
 346     for part in partition(kconfig_files, cpu_count()):
 347         arglist.append((part, ignore))
 348     for res in pool.map(parse_kconfig_files, arglist):
 349         defined_symbols.extend(res[0])
 350         referenced_symbols.update(res[1])
 351     defined_symbols = set(defined_symbols)
 352
 353     # inverse mapping of referenced_symbols to dict(symbol: [files])
 354     inv_map = dict()
 355     for _file, symbols in referenced_symbols.items():
 356         for symbol in symbols:
 357             inv_map[symbol] = inv_map.get(symbol, set())
 358             inv_map[symbol].add(_file)
 359     referenced_symbols = inv_map
 360
 361     undefined = {}  # {symbol: [files]}
 362     for symbol in sorted(referenced_symbols):
 363         # filter some false positives
 364         if symbol == "FOO" or symbol == "BAR" or \
 365                 symbol == "FOO_BAR" or symbol == "XXX":
 366             continue
 367         if symbol not in defined_symbols:
 368             if symbol.endswith("_MODULE"):
 369                 # avoid false positives for kernel modules
 370                 if symbol[:-len("_MODULE")] in defined_symbols:
 371                     continue
 372             undefined[symbol] = referenced_symbols.get(symbol)
 373     return undefined, defined_symbols
 374
 375
 376 def parse_source_files(source_files):
 377     """Parse each source file in @source_files and return dictionary with source
 378     files as keys and lists of references Kconfig symbols as values."""
 379     referenced_symbols = dict()
 380     for sfile in source_files:
 381         referenced_symbols[sfile] = parse_source_file(sfile)
 382     return referenced_symbols
 383
 384
 385 def parse_source_file(sfile):
 386     """Parse @sfile and return a list of referenced Kconfig symbols."""
 387     lines = []
 388     references = []
 389
 390     if not os.path.exists(sfile):
 391         return references
 392
 393     with open(sfile, "r", encoding='utf-8', errors='replace') as stream:
 394         lines = stream.readlines()
 395
 396     for line in lines:
 397         if "CONFIG_" not in line:
 398             continue
 399         symbols = REGEX_SOURCE_SYMBOL.findall(line)
 400         for symbol in symbols:
 401             if not REGEX_FILTER_SYMBOLS.search(symbol):
 402                 continue
 403             references.append(symbol)
 404
 405     return references
 406
 407
 408 def get_symbols_in_line(line):
 409     """Return mentioned Kconfig symbols in @line."""
 410     return REGEX_SYMBOL.findall(line)
 411
 412
 413 def parse_kconfig_files(args):
 414     """Parse kconfig files and return tuple of defined and references Kconfig
 415     symbols.  Note, @args is a tuple of a list of files and the @ignore
 416     pattern."""
 417     kconfig_files = args[0]
 418     ignore = args[1]
 419     defined_symbols = []
 420     referenced_symbols = dict()
 421
 422     for kfile in kconfig_files:
 423         defined, references = parse_kconfig_file(kfile)
 424         defined_symbols.extend(defined)
 425         if ignore and re.match(ignore, kfile):
 426             # do not collect references for files that match the ignore pattern
 427             continue
 428         referenced_symbols[kfile] = references
 429     return (defined_symbols, referenced_symbols)
 430
 431
 432 def parse_kconfig_file(kfile):
 433     """Parse @kfile and update symbol definitions and references."""
 434     lines = []
 435     defined = []
 436     references = []
 437
 438     if not os.path.exists(kfile):
 439         return defined, references
 440
 441     with open(kfile, "r", encoding='utf-8', errors='replace') as stream:
 442         lines = stream.readlines()
 443
 444     for i in range(len(lines)):
 445         line = lines[i]
 446         line = line.strip('\n')
 447         line = line.split("#")[0]  # ignore comments
 448
 449         if REGEX_KCONFIG_DEF.match(line):
 450             symbol_def = REGEX_KCONFIG_DEF.findall(line)
 451             defined.append(symbol_def[0])
 452         elif REGEX_KCONFIG_STMT.match(line):
 453             line = REGEX_QUOTES.sub("", line)
 454             symbols = get_symbols_in_line(line)
 455             # multi-line statements
 456             while line.endswith("\\"):
 457                 i += 1
 458                 line = lines[i]
 459                 line = line.strip('\n')
 460                 symbols.extend(get_symbols_in_line(line))
 461             for symbol in set(symbols):
 462                 if REGEX_NUMERIC.match(symbol):
 463                     # ignore numeric values
 464                     continue
 465                 references.append(symbol)
 466
 467     return defined, references
 468
 469
 470 def main():
 471     try:
 472         print_undefined_symbols()
 473     except BrokenPipeError:
 474         # Python flushes standard streams on exit; redirect remaining output
 475         # to devnull to avoid another BrokenPipeError at shutdown
 476         devnull = os.open(os.devnull, os.O_WRONLY)
 477         os.dup2(devnull, sys.stdout.fileno())
 478         sys.exit(1)  # Python exits with error code 1 on EPIPE
 479
 480
 481 if __name__ == "__main__":
 482     main()