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 <[email protected]>
7# (c) 2014 Stefan Hengelein <[email protected]>
8#
9
10
11import argparse
12import difflib
13import os
14import re
15import signal
16import subprocess
17import sys
18from multiprocessing import Pool, cpu_count
19
20
21# regex expressions
22OPERATORS = r"&|\(|\)|\||\!"
23SYMBOL = r"(?:\w*[A-Z0-9]\w*){2,}"
24DEF = r"^\s*(?:menu){,1}config\s+(" + SYMBOL + r")\s*"
25EXPR = r"(?:" + OPERATORS + r"|\s|" + SYMBOL + r")+"
26DEFAULT = r"default\s+.*?(?:if\s.+){,1}"
27STMT = r"^\s*(?:if|select|imply|depends\s+on|(?:" + DEFAULT + r"))\s+" + EXPR
28SOURCE_SYMBOL = r"(?:\W|\b)+[D]{,1}CONFIG_(" + SYMBOL + r")"
29
30# regex objects
31REGEX_FILE_KCONFIG = re.compile(r".*Kconfig[\.\w+\-]*$")
32REGEX_SYMBOL = re.compile(r'(?!\B)' + SYMBOL + r'(?!\B)')
33REGEX_SOURCE_SYMBOL = re.compile(SOURCE_SYMBOL)
34REGEX_KCONFIG_DEF = re.compile(DEF)
35REGEX_KCONFIG_EXPR = re.compile(EXPR)
36REGEX_KCONFIG_STMT = re.compile(STMT)
37REGEX_KCONFIG_HELP = re.compile(r"^\s+help\s*$")
38REGEX_FILTER_SYMBOLS = re.compile(r"[A-Za-z0-9]$")
39REGEX_NUMERIC = re.compile(r"0[xX][0-9a-fA-F]+|[0-9]+")
40REGEX_QUOTES = re.compile("(\"(.*?)\")")
41
42
43def 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        if args.commit.startswith('HEAD'):
106            sys.exit("The --commit option can't use the HEAD ref")
107
108        args.find = False
109
110    if args.ignore:
111        try:
112            re.match(args.ignore, "this/is/just/a/test.c")
113        except:
114            sys.exit("Please specify a valid Python regex.")
115
116    return args
117
118
119def main():
120    """Main function of this module."""
121    args = parse_options()
122
123    global COLOR
124    COLOR = args.color and sys.stdout.isatty()
125
126    if args.sim and not args.commit and not args.diff:
127        sims = find_sims(args.sim, args.ignore)
128        if sims:
129            print("%s: %s" % (yel("Similar symbols"), ', '.join(sims)))
130        else:
131            print("%s: no similar symbols found" % yel("Similar symbols"))
132        sys.exit(0)
133
134    # dictionary of (un)defined symbols
135    defined = {}
136    undefined = {}
137
138    if args.commit or args.diff:
139        head = get_head()
140
141        # get commit range
142        commit_a = None
143        commit_b = None
144        if args.commit:
145            commit_a = args.commit + "~"
146            commit_b = args.commit
147        elif args.diff:
148            split = args.diff.split("..")
149            commit_a = split[0]
150            commit_b = split[1]
151            undefined_a = {}
152            undefined_b = {}
153
154        # get undefined items before the commit
155        reset(commit_a)
156        undefined_a, _ = check_symbols(args.ignore)
157
158        # get undefined items for the commit
159        reset(commit_b)
160        undefined_b, defined = check_symbols(args.ignore)
161
162        # report cases that are present for the commit but not before
163        for symbol in sorted(undefined_b):
164            # symbol has not been undefined before
165            if symbol not in undefined_a:
166                files = sorted(undefined_b.get(symbol))
167                undefined[symbol] = files
168            # check if there are new files that reference the undefined symbol
169            else:
170                files = sorted(undefined_b.get(symbol) -
171                               undefined_a.get(symbol))
172                if files:
173                    undefined[symbol] = files
174
175        # reset to head
176        reset(head)
177
178    # default to check the entire tree
179    else:
180        undefined, defined = check_symbols(args.ignore)
181
182    # now print the output
183    for symbol in sorted(undefined):
184        print(red(symbol))
185
186        files = sorted(undefined.get(symbol))
187        print("%s: %s" % (yel("Referencing files"), ", ".join(files)))
188
189        sims = find_sims(symbol, args.ignore, defined)
190        sims_out = yel("Similar symbols")
191        if sims:
192            print("%s: %s" % (sims_out, ', '.join(sims)))
193        else:
194            print("%s: %s" % (sims_out, "no similar symbols found"))
195
196        if args.find:
197            print("%s:" % yel("Commits changing symbol"))
198            commits = find_commits(symbol, args.diff)
199            if commits:
200                for commit in commits:
201                    commit = commit.split(" ", 1)
202                    print("\t- %s (\"%s\")" % (yel(commit[0]), commit[1]))
203            else:
204                print("\t- no commit found")
205        print()  # new line
206
207
208def reset(commit):
209    """Reset current git tree to %commit."""
210    execute(["git", "reset", "--hard", commit])
211
212
213def yel(string):
214    """
215    Color %string yellow.
216    """
217    return "\033[33m%s\033[0m" % string if COLOR else string
218
219
220def red(string):
221    """
222    Color %string red.
223    """
224    return "\033[31m%s\033[0m" % string if COLOR else string
225
226
227def execute(cmd):
228    """Execute %cmd and return stdout.  Exit in case of error."""
229    try:
230        stdout = subprocess.check_output(cmd, stderr=subprocess.STDOUT, shell=False)
231        stdout = stdout.decode(errors='replace')
232    except subprocess.CalledProcessError as fail:
233        exit(fail)
234    return stdout
235
236
237def find_commits(symbol, diff):
238    """Find commits changing %symbol in the given range of %diff."""
239    commits = execute(["git", "log", "--pretty=oneline",
240                       "--abbrev-commit", "-G",
241                       symbol, diff])
242    return [x for x in commits.split("\n") if x]
243
244
245def tree_is_dirty():
246    """Return true if the current working tree is dirty (i.e., if any file has
247    been added, deleted, modified, renamed or copied but not committed)."""
248    stdout = execute(["git", "status", "--porcelain"])
249    for line in stdout:
250        if re.findall(r"[URMADC]{1}", line[:2]):
251            return True
252    return False
253
254
255def get_head():
256    """Return commit hash of current HEAD."""
257    stdout = execute(["git", "rev-parse", "HEAD"])
258    return stdout.strip('\n')
259
260
261def partition(lst, size):
262    """Partition list @lst into eveni-sized lists of size @size."""
263    return [lst[i::size] for i in range(size)]
264
265
266def init_worker():
267    """Set signal handler to ignore SIGINT."""
268    signal.signal(signal.SIGINT, signal.SIG_IGN)
269
270
271def find_sims(symbol, ignore, defined=[]):
272    """Return a list of max. ten Kconfig symbols that are string-similar to
273    @symbol."""
274    if defined:
275        return difflib.get_close_matches(symbol, set(defined), 10)
276
277    pool = Pool(cpu_count(), init_worker)
278    kfiles = []
279    for gitfile in get_files():
280        if REGEX_FILE_KCONFIG.match(gitfile):
281            kfiles.append(gitfile)
282
283    arglist = []
284    for part in partition(kfiles, cpu_count()):
285        arglist.append((part, ignore))
286
287    for res in pool.map(parse_kconfig_files, arglist):
288        defined.extend(res[0])
289
290    return difflib.get_close_matches(symbol, set(defined), 10)
291
292
293def get_files():
294    """Return a list of all files in the current git directory."""
295    # use 'git ls-files' to get the worklist
296    stdout = execute(["git", "ls-files"])
297    if len(stdout) > 0 and stdout[-1] == "\n":
298        stdout = stdout[:-1]
299
300    files = []
301    for gitfile in stdout.rsplit("\n"):
302        if ".git" in gitfile or "ChangeLog" in gitfile or      \
303                ".log" in gitfile or os.path.isdir(gitfile) or \
304                gitfile.startswith("tools/"):
305            continue
306        files.append(gitfile)
307    return files
308
309
310def check_symbols(ignore):
311    """Find undefined Kconfig symbols and return a dict with the symbol as key
312    and a list of referencing files as value.  Files matching %ignore are not
313    checked for undefined symbols."""
314    pool = Pool(cpu_count(), init_worker)
315    try:
316        return check_symbols_helper(pool, ignore)
317    except KeyboardInterrupt:
318        pool.terminate()
319        pool.join()
320        sys.exit(1)
321
322
323def check_symbols_helper(pool, ignore):
324    """Helper method for check_symbols().  Used to catch keyboard interrupts in
325    check_symbols() in order to properly terminate running worker processes."""
326    source_files = []
327    kconfig_files = []
328    defined_symbols = []
329    referenced_symbols = dict()  # {file: [symbols]}
330
331    for gitfile in get_files():
332        if REGEX_FILE_KCONFIG.match(gitfile):
333            kconfig_files.append(gitfile)
334        else:
335            if ignore and re.match(ignore, gitfile):
336                continue
337            # add source files that do not match the ignore pattern
338            source_files.append(gitfile)
339
340    # parse source files
341    arglist = partition(source_files, cpu_count())
342    for res in pool.map(parse_source_files, arglist):
343        referenced_symbols.update(res)
344
345    # parse kconfig files
346    arglist = []
347    for part in partition(kconfig_files, cpu_count()):
348        arglist.append((part, ignore))
349    for res in pool.map(parse_kconfig_files, arglist):
350        defined_symbols.extend(res[0])
351        referenced_symbols.update(res[1])
352    defined_symbols = set(defined_symbols)
353
354    # inverse mapping of referenced_symbols to dict(symbol: [files])
355    inv_map = dict()
356    for _file, symbols in referenced_symbols.items():
357        for symbol in symbols:
358            inv_map[symbol] = inv_map.get(symbol, set())
359            inv_map[symbol].add(_file)
360    referenced_symbols = inv_map
361
362    undefined = {}  # {symbol: [files]}
363    for symbol in sorted(referenced_symbols):
364        # filter some false positives
365        if symbol == "FOO" or symbol == "BAR" or \
366                symbol == "FOO_BAR" or symbol == "XXX":
367            continue
368        if symbol not in defined_symbols:
369            if symbol.endswith("_MODULE"):
370                # avoid false positives for kernel modules
371                if symbol[:-len("_MODULE")] in defined_symbols:
372                    continue
373            undefined[symbol] = referenced_symbols.get(symbol)
374    return undefined, defined_symbols
375
376
377def parse_source_files(source_files):
378    """Parse each source file in @source_files and return dictionary with source
379    files as keys and lists of references Kconfig symbols as values."""
380    referenced_symbols = dict()
381    for sfile in source_files:
382        referenced_symbols[sfile] = parse_source_file(sfile)
383    return referenced_symbols
384
385
386def parse_source_file(sfile):
387    """Parse @sfile and return a list of referenced Kconfig symbols."""
388    lines = []
389    references = []
390
391    if not os.path.exists(sfile):
392        return references
393
394    with open(sfile, "r", encoding='utf-8', errors='replace') as stream:
395        lines = stream.readlines()
396
397    for line in lines:
398        if "CONFIG_" not in line:
399            continue
400        symbols = REGEX_SOURCE_SYMBOL.findall(line)
401        for symbol in symbols:
402            if not REGEX_FILTER_SYMBOLS.search(symbol):
403                continue
404            references.append(symbol)
405
406    return references
407
408
409def get_symbols_in_line(line):
410    """Return mentioned Kconfig symbols in @line."""
411    return REGEX_SYMBOL.findall(line)
412
413
414def parse_kconfig_files(args):
415    """Parse kconfig files and return tuple of defined and references Kconfig
416    symbols.  Note, @args is a tuple of a list of files and the @ignore
417    pattern."""
418    kconfig_files = args[0]
419    ignore = args[1]
420    defined_symbols = []
421    referenced_symbols = dict()
422
423    for kfile in kconfig_files:
424        defined, references = parse_kconfig_file(kfile)
425        defined_symbols.extend(defined)
426        if ignore and re.match(ignore, kfile):
427            # do not collect references for files that match the ignore pattern
428            continue
429        referenced_symbols[kfile] = references
430    return (defined_symbols, referenced_symbols)
431
432
433def parse_kconfig_file(kfile):
434    """Parse @kfile and update symbol definitions and references."""
435    lines = []
436    defined = []
437    references = []
438    skip = False
439
440    if not os.path.exists(kfile):
441        return defined, references
442
443    with open(kfile, "r", encoding='utf-8', errors='replace') as stream:
444        lines = stream.readlines()
445
446    for i in range(len(lines)):
447        line = lines[i]
448        line = line.strip('\n')
449        line = line.split("#")[0]  # ignore comments
450
451        if REGEX_KCONFIG_DEF.match(line):
452            symbol_def = REGEX_KCONFIG_DEF.findall(line)
453            defined.append(symbol_def[0])
454            skip = False
455        elif REGEX_KCONFIG_HELP.match(line):
456            skip = True
457        elif skip:
458            # ignore content of help messages
459            pass
460        elif REGEX_KCONFIG_STMT.match(line):
461            line = REGEX_QUOTES.sub("", line)
462            symbols = get_symbols_in_line(line)
463            # multi-line statements
464            while line.endswith("\\"):
465                i += 1
466                line = lines[i]
467                line = line.strip('\n')
468                symbols.extend(get_symbols_in_line(line))
469            for symbol in set(symbols):
470                if REGEX_NUMERIC.match(symbol):
471                    # ignore numeric values
472                    continue
473                references.append(symbol)
474
475    return defined, references
476
477
478if __name__ == "__main__":
479    main()
480