2 # SPDX-License-Identifier: GPL-2.0
4 # Copyright (C) Google LLC, 2018
6 # Author: Tom Roeder <tmroeder@google.com>
8 """A tool for generating compile_commands.json in the Linux kernel."""
17 _DEFAULT_OUTPUT = 'compile_commands.json'
18 _DEFAULT_LOG_LEVEL = 'WARNING'
20 _FILENAME_PATTERN = r'^\..*\.cmd$'
21 _LINE_PATTERN = r'^cmd_[^ ]*\.o := (.* )([^ ]*\.c)$'
22 _VALID_LOG_LEVELS = ['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']
24 # A kernel build generally has over 2000 entries in its compile_commands.json
25 # database. If this code finds 300 or fewer, then warn the user that they might
26 # not have all the .cmd files, and they might need to compile the kernel.
27 _LOW_COUNT_THRESHOLD = 300
30 def parse_arguments():
31 """Sets up and parses command-line arguments.
34 log_level: A logging level to filter log output.
35 directory: The work directory where the objects were built.
36 ar: Command used for parsing .a archives.
37 output: Where to write the compile-commands JSON file.
38 paths: The list of files/directories to handle to find .cmd files.
40 usage = 'Creates a compile_commands.json database from kernel .cmd files'
41 parser = argparse.ArgumentParser(description=usage)
43 directory_help = ('specify the output directory used for the kernel build '
44 '(defaults to the working directory)')
45 parser.add_argument('-d', '--directory', type=str, default='.',
48 output_help = ('path to the output command database (defaults to ' +
49 _DEFAULT_OUTPUT + ')')
50 parser.add_argument('-o', '--output', type=str, default=_DEFAULT_OUTPUT,
53 log_level_help = ('the level of log messages to produce (defaults to ' +
54 _DEFAULT_LOG_LEVEL + ')')
55 parser.add_argument('--log_level', choices=_VALID_LOG_LEVELS,
56 default=_DEFAULT_LOG_LEVEL, help=log_level_help)
58 ar_help = 'command used for parsing .a archives'
59 parser.add_argument('-a', '--ar', type=str, default='llvm-ar', help=ar_help)
61 paths_help = ('directories to search or files to parse '
62 '(files should be *.o, *.a, or modules.order). '
63 'If nothing is specified, the current directory is searched')
64 parser.add_argument('paths', type=str, nargs='*', help=paths_help)
66 args = parser.parse_args()
68 return (args.log_level,
69 os.path.abspath(args.directory),
72 args.paths if len(args.paths) > 0 else [args.directory])
75 def cmdfiles_in_dir(directory):
76 """Generate the iterator of .cmd files found under the directory.
78 Walk under the given directory, and yield every .cmd file found.
81 directory: The directory to search for .cmd files.
84 The path to a .cmd file.
87 filename_matcher = re.compile(_FILENAME_PATTERN)
89 for dirpath, _, filenames in os.walk(directory):
90 for filename in filenames:
91 if filename_matcher.match(filename):
92 yield os.path.join(dirpath, filename)
96 """Return the path of .cmd file used for the given build artifact
102 The path to .cmd file
104 dir, base = os.path.split(path)
105 return os.path.join(dir, '.' + base + '.cmd')
108 def cmdfiles_for_o(obj):
109 """Generate the iterator of .cmd files associated with the object
111 Yield the .cmd file used to build the given object
117 The path to .cmd file
119 yield to_cmdfile(obj)
122 def cmdfiles_for_a(archive, ar):
123 """Generate the iterator of .cmd files associated with the archive.
125 Parse the given archive, and yield every .cmd file used to build it.
128 archive: The archive to parse
131 The path to every .cmd file found
133 for obj in subprocess.check_output([ar, '-t', archive]).decode().split():
134 yield to_cmdfile(obj)
137 def cmdfiles_for_modorder(modorder):
138 """Generate the iterator of .cmd files associated with the modules.order.
140 Parse the given modules.order, and yield every .cmd file used to build the
144 modorder: The modules.order file to parse
147 The path to every .cmd file found
149 with open(modorder) as f:
152 base, ext = os.path.splitext(ko)
154 sys.exit('{}: module path must end with .ko'.format(ko))
156 # The first line of *.mod lists the objects that compose the module.
158 for obj in m.readline().split():
159 yield to_cmdfile(obj)
162 def process_line(root_directory, command_prefix, file_path):
163 """Extracts information from a .cmd line and creates an entry from it.
166 root_directory: The directory that was searched for .cmd files. Usually
167 used directly in the "directory" entry in compile_commands.json.
168 command_prefix: The extracted command line, up to the last element.
169 file_path: The .c file from the end of the extracted command.
170 Usually relative to root_directory, but sometimes absolute.
173 An entry to append to compile_commands.
176 ValueError: Could not find the extracted file based on file_path and
177 root_directory or file_directory.
179 # The .cmd files are intended to be included directly by Make, so they
180 # escape the pound sign '#', either as '\#' or '$(pound)' (depending on the
181 # kernel version). The compile_commands.json file is not interepreted
182 # by Make, so this code replaces the escaped version with '#'.
183 prefix = command_prefix.replace('\#', '#').replace('$(pound)', '#')
185 # Use os.path.abspath() to normalize the path resolving '.' and '..' .
186 abs_path = os.path.abspath(os.path.join(root_directory, file_path))
187 if not os.path.exists(abs_path):
188 raise ValueError('File %s not found' % abs_path)
190 'directory': root_directory,
192 'command': prefix + file_path,
197 """Walks through the directory and finds and parses .cmd files."""
198 log_level, directory, output, ar, paths = parse_arguments()
200 level = getattr(logging, log_level)
201 logging.basicConfig(format='%(levelname)s: %(message)s', level=level)
203 line_matcher = re.compile(_LINE_PATTERN)
205 compile_commands = []
208 # If 'path' is a directory, handle all .cmd files under it.
209 # Otherwise, handle .cmd files associated with the file.
210 # Most of built-in objects are linked via archives (built-in.a or lib.a)
211 # but some objects are linked to vmlinux directly.
212 # Modules are listed in modules.order.
213 if os.path.isdir(path):
214 cmdfiles = cmdfiles_in_dir(path)
215 elif path.endswith('.o'):
216 cmdfiles = cmdfiles_for_o(path)
217 elif path.endswith('.a'):
218 cmdfiles = cmdfiles_for_a(path, ar)
219 elif path.endswith('modules.order'):
220 cmdfiles = cmdfiles_for_modorder(path)
222 sys.exit('{}: unknown file type'.format(path))
224 for cmdfile in cmdfiles:
225 with open(cmdfile, 'rt') as f:
226 result = line_matcher.match(f.readline())
229 entry = process_line(directory, result.group(1),
231 compile_commands.append(entry)
232 except ValueError as err:
233 logging.info('Could not add line from %s: %s',
236 with open(output, 'wt') as f:
237 json.dump(compile_commands, f, indent=2, sort_keys=True)
239 count = len(compile_commands)
240 if count < _LOW_COUNT_THRESHOLD:
242 'Found %s entries. Have you compiled the kernel?', count)
245 if __name__ == '__main__':