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."""
18 _DEFAULT_OUTPUT = 'compile_commands.json'
19 _DEFAULT_LOG_LEVEL = 'WARNING'
21 _FILENAME_PATTERN = r'^\..*\.cmd$'
22 _LINE_PATTERN = r'^cmd_[^ ]*\.o := (.* )([^ ]*\.c)$'
23 _VALID_LOG_LEVELS = ['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']
26 def parse_arguments():
27 """Sets up and parses command-line arguments.
30 log_level: A logging level to filter log output.
31 directory: The work directory where the objects were built.
32 ar: Command used for parsing .a archives.
33 output: Where to write the compile-commands JSON file.
34 paths: The list of files/directories to handle to find .cmd files.
36 usage = 'Creates a compile_commands.json database from kernel .cmd files'
37 parser = argparse.ArgumentParser(description=usage)
39 directory_help = ('specify the output directory used for the kernel build '
40 '(defaults to the working directory)')
41 parser.add_argument('-d', '--directory', type=str, default='.',
44 output_help = ('path to the output command database (defaults to ' +
45 _DEFAULT_OUTPUT + ')')
46 parser.add_argument('-o', '--output', type=str, default=_DEFAULT_OUTPUT,
49 log_level_help = ('the level of log messages to produce (defaults to ' +
50 _DEFAULT_LOG_LEVEL + ')')
51 parser.add_argument('--log_level', choices=_VALID_LOG_LEVELS,
52 default=_DEFAULT_LOG_LEVEL, help=log_level_help)
54 ar_help = 'command used for parsing .a archives'
55 parser.add_argument('-a', '--ar', type=str, default='llvm-ar', help=ar_help)
57 paths_help = ('directories to search or files to parse '
58 '(files should be *.o, *.a, or modules.order). '
59 'If nothing is specified, the current directory is searched')
60 parser.add_argument('paths', type=str, nargs='*', help=paths_help)
62 args = parser.parse_args()
64 return (args.log_level,
65 os.path.abspath(args.directory),
68 args.paths if len(args.paths) > 0 else [args.directory])
71 def cmdfiles_in_dir(directory):
72 """Generate the iterator of .cmd files found under the directory.
74 Walk under the given directory, and yield every .cmd file found.
77 directory: The directory to search for .cmd files.
80 The path to a .cmd file.
83 filename_matcher = re.compile(_FILENAME_PATTERN)
85 for dirpath, _, filenames in os.walk(directory):
86 for filename in filenames:
87 if filename_matcher.match(filename):
88 yield os.path.join(dirpath, filename)
92 """Return the path of .cmd file used for the given build artifact
100 dir, base = os.path.split(path)
101 return os.path.join(dir, '.' + base + '.cmd')
104 def cmdfiles_for_o(obj):
105 """Generate the iterator of .cmd files associated with the object
107 Yield the .cmd file used to build the given object
113 The path to .cmd file
115 yield to_cmdfile(obj)
118 def cmdfiles_for_a(archive, ar):
119 """Generate the iterator of .cmd files associated with the archive.
121 Parse the given archive, and yield every .cmd file used to build it.
124 archive: The archive to parse
127 The path to every .cmd file found
129 for obj in subprocess.check_output([ar, '-t', archive]).decode().split():
130 yield to_cmdfile(obj)
133 def cmdfiles_for_modorder(modorder):
134 """Generate the iterator of .cmd files associated with the modules.order.
136 Parse the given modules.order, and yield every .cmd file used to build the
140 modorder: The modules.order file to parse
143 The path to every .cmd file found
145 with open(modorder) as f:
148 base, ext = os.path.splitext(ko)
150 sys.exit('{}: module path must end with .ko'.format(ko))
152 # The first line of *.mod lists the objects that compose the module.
154 for obj in m.readline().split():
155 yield to_cmdfile(obj)
158 def process_line(root_directory, command_prefix, file_path):
159 """Extracts information from a .cmd line and creates an entry from it.
162 root_directory: The directory that was searched for .cmd files. Usually
163 used directly in the "directory" entry in compile_commands.json.
164 command_prefix: The extracted command line, up to the last element.
165 file_path: The .c file from the end of the extracted command.
166 Usually relative to root_directory, but sometimes absolute.
169 An entry to append to compile_commands.
172 ValueError: Could not find the extracted file based on file_path and
173 root_directory or file_directory.
175 # The .cmd files are intended to be included directly by Make, so they
176 # escape the pound sign '#', either as '\#' or '$(pound)' (depending on the
177 # kernel version). The compile_commands.json file is not interepreted
178 # by Make, so this code replaces the escaped version with '#'.
179 prefix = command_prefix.replace(r'\#', '#').replace('$(pound)', '#')
181 # Use os.path.abspath() to normalize the path resolving '.' and '..' .
182 abs_path = os.path.abspath(os.path.join(root_directory, file_path))
183 if not os.path.exists(abs_path):
184 raise ValueError('File %s not found' % abs_path)
186 'directory': root_directory,
188 'command': prefix + file_path,
193 """Walks through the directory and finds and parses .cmd files."""
194 log_level, directory, output, ar, paths = parse_arguments()
196 level = getattr(logging, log_level)
197 logging.basicConfig(format='%(levelname)s: %(message)s', level=level)
199 line_matcher = re.compile(_LINE_PATTERN)
201 compile_commands = []
204 # If 'path' is a directory, handle all .cmd files under it.
205 # Otherwise, handle .cmd files associated with the file.
206 # Most of built-in objects are linked via archives (built-in.a or lib.a)
207 # but some objects are linked to vmlinux directly.
208 # Modules are listed in modules.order.
209 if os.path.isdir(path):
210 cmdfiles = cmdfiles_in_dir(path)
211 elif path.endswith('.o'):
212 cmdfiles = cmdfiles_for_o(path)
213 elif path.endswith('.a'):
214 cmdfiles = cmdfiles_for_a(path, ar)
215 elif path.endswith('modules.order'):
216 cmdfiles = cmdfiles_for_modorder(path)
218 sys.exit('{}: unknown file type'.format(path))
220 for cmdfile in cmdfiles:
221 with open(cmdfile, 'rt') as f:
222 result = line_matcher.match(f.readline())
225 entry = process_line(directory, result.group(1),
227 compile_commands.append(entry)
228 except ValueError as err:
229 logging.info('Could not add line from %s: %s',
232 with open(output, 'wt') as f:
233 json.dump(compile_commands, f, indent=2, sort_keys=True)
236 if __name__ == '__main__':