1#!/usr/bin/env python3
2# Wireshark - Network traffic analyzer
3# By Gerald Combs <gerald@wireshark.org>
4# Copyright 1998 Gerald Combs
5#
6# SPDX-License-Identifier: GPL-2.0-or-later
7
8import os
9import re
10import subprocess
11import argparse
12import signal
13
14# This utility scans for tfs items, and works out if standard ones
15# could have been used intead (from epan/tfs.c)
16
17# TODO:
18# - check how many of the definitions in epan/tfs.c are used in other dissectors
19# - see if there are other values that should be in epan/tfs.c and shared
20
21
22# Try to exit soon after Ctrl-C is pressed.
23should_exit = False
24
25def signal_handler(sig, frame):
26    global should_exit
27    should_exit = True
28    print('You pressed Ctrl+C - exiting')
29
30signal.signal(signal.SIGINT, signal_handler)
31
32
33# Keep track of custom entries that might appear in multiple dissectors,
34# so we can consider adding them to tfs.c
35custom_tfs_entries = {}
36def AddCustomEntry(val1, val2, file):
37    global custom_tfs_entries
38    if (val1, val2) in custom_tfs_entries:
39        custom_tfs_entries[(val1, val2)].append(file)
40    else:
41        custom_tfs_entries[(val1, val2)] = [file]
42
43
44
45class TFS:
46    def __init__(self, file, name, val1, val2):
47        self.file = file
48        self.name = name
49        self.val1 = val1
50        self.val2 = val2
51
52        # Do some extra checks on values.
53        if val1.startswith(' ') or val1.endswith(' '):
54            print('N.B.: file=' + self.file + ' ' + self.name + ' - false val begins or ends with space \"' + self.val1 + '\"')
55        if val2.startswith(' ') or val2.endswith(' '):
56            print('N.B.: file=' + self.file + ' ' + self.name + ' - true val begins or ends with space \"' + self.val2 + '\"')
57
58    def __str__(self):
59        return '{' + '"' + self.val1 + '", "' + self.val2 + '"}'
60
61
62def removeComments(code_string):
63    code_string = re.sub(re.compile(r"/\*.*?\*/",re.DOTALL ) ,"" ,code_string) # C-style comment
64    code_string = re.sub(re.compile(r"//.*?\n" ) ,"" ,code_string)             # C++-style comment
65    return code_string
66
67
68# Look for hf items in a dissector file.
69def findItems(filename):
70    items = {}
71
72    with open(filename, 'r') as f:
73        contents = f.read()
74        # Example: const true_false_string tfs_true_false = { "True", "False" };
75
76        # Remove comments so as not to trip up RE.
77        contents = removeComments(contents)
78
79        matches =   re.finditer(r'.*const\s*true_false_string\s*([a-z_]*)\s*=\s*{\s*\"([a-zA-Z_ ]*)\"\s*,\s*\"([a-zA-Z_ ]*)\"', contents)
80        for m in matches:
81            name = m.group(1)
82            val1 = m.group(2)
83            val2 = m.group(3)
84            # Store this entry.
85            items[name] = TFS(filename, name, val1, val2)
86
87    return items
88
89
90
91def is_dissector_file(filename):
92    p = re.compile(r'.*packet-.*\.c')
93    return p.match(filename)
94
95def findDissectorFilesInFolder(folder):
96    # Look at files in sorted order, to give some idea of how far through is.
97    files = []
98
99    for f in sorted(os.listdir(folder)):
100        if should_exit:
101            return
102        if is_dissector_file(f):
103            filename = os.path.join(folder, f)
104            files.append(filename)
105    return files
106
107warnings_found = 0
108errors_found = 0
109
110# Check the given dissector file.
111def checkFile(filename, tfs_items, look_for_common=False):
112    global warnings_found
113    global errors_found
114
115    # Check file exists - e.g. may have been deleted in a recent commit.
116    if not os.path.exists(filename):
117        print(filename, 'does not exist!')
118        return
119
120    # Find items.
121    items = findItems(filename)
122
123    # See if any of these items already existed in tfs.c
124    for i in items:
125        for t in tfs_items:
126            found = False
127
128            #
129            # Do not do this check for plugins; plugins cannot import
130            # data values from libwireshark (functions, yes; data
131            # values, no).
132            #
133            # Test whether there's a common prefix for the file name
134            # and "plugin/epan/"; if so, this is a plugin, and there
135            # is no common path and os.path.commonprefix returns an
136            # empty string, otherwise it returns the common path, so
137            # we check whether the common path is an empty string.
138            #
139            if os.path.commonprefix([filename, 'plugin/epan/']) == '':
140                exact_case = False
141                if tfs_items[t].val1 == items[i].val1 and tfs_items[t].val2 == items[i].val2:
142                    found = True
143                    exact_case = True
144                elif tfs_items[t].val1.upper() == items[i].val1.upper() and tfs_items[t].val2.upper() == items[i].val2.upper():
145                    found = True
146
147                if found:
148                    print(filename, i, "- could have used", t, 'from tfs.c instead: ', tfs_items[t],
149                          '' if exact_case else '  (capitalisation differs)')
150                    if exact_case:
151                        errors_found += 1
152                    else:
153                        warnings_found += 1
154                    break
155        if not found:
156            if look_for_common:
157                AddCustomEntry(items[i].val1, items[i].val2, filename)
158
159
160#################################################################
161# Main logic.
162
163# command-line args.  Controls which dissector files should be checked.
164# If no args given, will just scan epan/dissectors folder.
165parser = argparse.ArgumentParser(description='Check calls in dissectors')
166parser.add_argument('--file', action='store', default='',
167                    help='specify individual dissector file to test')
168parser.add_argument('--commits', action='store',
169                    help='last N commits to check')
170parser.add_argument('--open', action='store_true',
171                    help='check open files')
172parser.add_argument('--common', action='store_true',
173                    help='check for potential new entries for tfs.c')
174
175
176args = parser.parse_args()
177
178
179# Get files from wherever command-line args indicate.
180files = []
181if args.file:
182    # Add single specified file..
183    if not args.file.startswith('epan'):
184        files.append(os.path.join('epan', 'dissectors', args.file))
185    else:
186        files.append(args.file)
187elif args.commits:
188    # Get files affected by specified number of commits.
189    command = ['git', 'diff', '--name-only', 'HEAD~' + args.commits]
190    files = [f.decode('utf-8')
191             for f in subprocess.check_output(command).splitlines()]
192    # Will examine dissector files only
193    files = list(filter(lambda f : is_dissector_file(f), files))
194elif args.open:
195    # Unstaged changes.
196    command = ['git', 'diff', '--name-only']
197    files = [f.decode('utf-8')
198             for f in subprocess.check_output(command).splitlines()]
199    # Only interested in dissector files.
200    files = list(filter(lambda f : is_dissector_file(f), files))
201    # Staged changes.
202    command = ['git', 'diff', '--staged', '--name-only']
203    files_staged = [f.decode('utf-8')
204                    for f in subprocess.check_output(command).splitlines()]
205    # Only interested in dissector files.
206    files_staged = list(filter(lambda f : is_dissector_file(f), files_staged))
207    for f in files_staged:
208        if not f in files:
209            files.append(f)
210else:
211    # Find all dissector files from folder.
212    files = findDissectorFilesInFolder(os.path.join('epan', 'dissectors'))
213
214
215# If scanning a subset of files, list them here.
216print('Examining:')
217if args.file or args.commits or args.open:
218    if files:
219        print(' '.join(files), '\n')
220    else:
221        print('No files to check.\n')
222else:
223    print('All dissector modules\n')
224
225
226# Get standard/ shared ones.
227tfs_entries = findItems(os.path.join('epan', 'tfs.c'))
228
229# Now check the files to see if they could have used shared ones instead.
230for f in files:
231    if should_exit:
232        exit(1)
233    checkFile(f, tfs_entries, look_for_common=args.common)
234
235
236# Show summary.
237print(warnings_found, 'warnings found')
238if errors_found:
239    print(errors_found, 'errors found')
240    exit(1)
241
242if args.common:
243    # Looking for items that could potentially be moved to tfs.c
244    for c in custom_tfs_entries:
245        # Only want to see items that have 3 or more occurrences.
246        # Even then, probably only want to consider ones that sound generic.
247        if len(custom_tfs_entries[c]) > 2:
248            print(c, 'appears', len(custom_tfs_entries[c]), 'times, in: ', custom_tfs_entries[c])
249