1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
|
# -*- encoding: utf8 -*-
'''
This tool extracts information about structs and defines from the C headers.
You can pass either the raw header files or JSON files to this script.
The JSON input format is as follows:
[
{
'file': 'some/header.h',
'structs': {
'struct_name': [
'field1',
'field2',
'field3',
{
'field4': [
'nested1',
'nested2',
{
'nested3': [
'deep_nested1',
...
]
}
...
]
},
'field5'
],
'other_struct': [
'field1',
'field2',
...
]
},
'defines': [
'DEFINE_1',
'DEFINE_2',
['f', 'FLOAT_DEFINE'],
'DEFINE_3',
...
]
},
{
'file': 'some/other/header.h',
...
}
]
Please note that the 'f' for 'FLOAT_DEFINE' is just the format passed to printf(), you can put anything printf() understands.
If you call this script with the flag "-f" and pass a header file, it will create an automated boilerplate for you.
The JSON output format is based on the return value of Runtime.generateStructInfo().
{
'structs': {
'struct_name': {
'__size__': <the struct's size>,
'field1': <field1's offset>,
'field2': <field2's offset>,
'field3': <field3's offset>,
'field4': {
'__size__': <field4's size>,
'nested1': <nested1's offset>,
...
},
...
}
},
'defines': {
'DEFINE_1': <DEFINE_1's value>,
...
}
}
'''
import sys
import os
import json
import argparse
import tempfile
import subprocess
import shared
# Try to load pycparser.
try:
import pycparser
except ImportError:
# The import failed, warn the user.
sys.stderr.write('WARN: pycparser isn\'t available. I won\'t be able to parse C files, only .json files.\n')
def parse_header(path, cpp_opts):
# Tell the user how to get pycparser, if he tries to parse a C file.
sys.stderr.write('ERR: I need pycparser to process C files. \n')
sys.stderr.write(' Use "pip install pycparser" or go to "https://github.com/eliben/pycparser" to install it.\n')
sys.exit(1)
else:
# We successfully imported pycparser, the script will be completely functional.
class FieldVisitor(pycparser.c_ast.NodeVisitor):
def __init__(self):
self._name = None
self.structs = {}
self.named_structs = {}
def visit_Struct(self, node):
if node.decls == None:
# Skip empty struct declarations.
return
fields = []
for decl in node.decls:
# Look for nested structs.
subwalk = FieldVisitor()
subwalk.visit(decl)
if subwalk.named_structs:
# Store the nested fields.
fields.append(subwalk.named_structs)
else:
# Just store the field name.
fields.append(decl.name)
if node.name != None:
self.structs[node.name] = fields
self.named_structs[self._name] = fields
def visit_Union(self, node):
self.visit_Struct(node)
def visit_TypeDecl(self, node):
# Remember the name of this typedef, so we can access it later in visit_Struct().
old_name = self._name
self._name = node.declname
self.generic_visit(node)
self._name = old_name
def parse_header(path, cpp_opts):
sys.stderr.write('Parsing header "' + path + '"...\n')
# Use clang -E as the preprocessor for pycparser.
ast = pycparser.parse_file(path, True, cpp_path=shared.CLANG_CC, cpp_args=['-E'] + cpp_opts)
# Walk the parsed AST and filter out all the declared structs and their fields.
walker = FieldVisitor()
walker.visit(ast)
return walker.structs
# The following three functions generate C code. The output of the compiled code will be
# parsed later on and then put back together into a dict structure by parse_c_output().
def c_set(name, type_, value, code):
code.append('printf("K' + name + '\\n");')
code.append('printf("V' + type_ + '\\n", ' + value + ');')
def c_descent(name, code):
code.append('printf("D' + name + '\\n");')
def c_ascent(code):
code.append('printf("A\\n");')
def parse_c_output(lines):
result = {}
cur_level = result
parent = []
key = None
for line in lines:
arg = line[1:].strip()
if line[0] == 'K':
# This is a key
key = arg
elif line[0] == 'V':
# A value
cur_level[key] = int(arg)
elif line[0] == 'D':
# Remember the current level as the last parent.
parent.append(cur_level)
# We descend one level.
cur_level[arg] = {}
cur_level = cur_level[arg]
elif line[0] == 'A':
# We return to the parent dict. (One level up.)
cur_level = parent.pop()
return result
def gen_inspect_code(path, struct, code):
c_descent(path[-1], code)
if len(path) == 1:
c_set('__size__', '%lu', 'sizeof (struct ' + path[0] + ')', code)
else:
c_set('__size__', '%lu', 'sizeof ((struct ' + path[0] + ' *)0)->' + '.'.join(path[1:]), code)
for field in struct:
if isinstance(field, dict):
# We have to recurse to inspect the nested dict.
fname = field.keys()[0]
gen_inspect_code(path + [fname], field[fname], code)
else:
c_set(field, '%u', 'offsetof(struct ' + path[0] + ', ' + '.'.join(path[1:] + [field]) + ')', code)
c_ascent(code)
def inspect_code(headers, cpp_opts, structs, defines):
sys.stderr.write('Generating C code...\n')
code = ['#include <stdio.h>', '#include <stddef.h>']
# Include all the needed headers.
for path in headers:
code.append('#include "' + path + '"')
code.append('int main() {')
c_descent('structs', code)
for name, struct in structs.items():
gen_inspect_code([name], struct, code)
c_ascent(code)
c_descent('defines', code)
for name in defines:
if isinstance(name, list):
type_, name = name
else:
type_ = 'i'
c_set(name, '%' + type_, name, code)
code.append('return 0;')
code.append('}')
# Write the source code to a temporary file.
src_file = tempfile.mkstemp('.c')
bin_file = tempfile.mkstemp()
os.write(src_file[0], '\n'.join(code))
# Close all unneeded FDs.
os.close(src_file[0])
os.close(bin_file[0])
info = []
try:
# Compile the program.
sys.stderr.write('Compiling generated code...\n')
subprocess.check_call([shared.CLANG_CC] + cpp_opts + ['-o', bin_file[1], src_file[1]])
# Run the compiled program.
sys.stderr.write('Calling generated program...\n')
info = subprocess.check_output([bin_file[1]]).splitlines()
except subprocess.CalledProcessError:
if os.path.isfile(bin_file[1]):
sys.stderr.write('FAIL: Running the generated program failed!\n')
else:
sys.stderr.write('FAIL: Compilation failed!\n')
sys.exit(1)
finally:
# Remove all temporary files.
os.unlink(src_file[1])
if os.path.exists(bin_file[1]):
os.unlink(bin_file[1])
# Parse the output of the program into a dict.
data = parse_c_output(info)
# Convert all the define's values into the appropriate python types (based on the type passed to printf).
for name in defines:
if isinstance(name, list):
type_, name = name
else:
type_ = 'i'
if type_[-1] in ('d', 'i', 'u'):
# Integer
data['defines'][name] = int(data['defines'][name])
elif type_[-1] in ('x', 'X', 'a', 'A'):
# Hexadecimal
data['defines'][name] = float.fromhex(data['defines'][name])
elif type_[-1] in ('f', 'F', 'e', 'E', 'g', 'G'):
# Float
data['defines'][name] = float(data['defines'][name])
# Leave everything else untouched.
return data
def main():
parser = argparse.ArgumentParser(description='Generate JSON infos for structs.')
parser.add_argument('headers', nargs='+', help='A header (.h) file or a JSON file with a list of structs and their fields')
parser.add_argument('-f', dest='list_fields', action='store_true', default=False, help='Output a list of structs and fields for the first header.')
parser.add_argument('-p', dest='pretty_print', action='store_true', default=False, help='Pretty print the outputted JSON.')
parser.add_argument('-o', dest='output', metavar='path', default=None, help='Path to the JSON file that will be written. If omitted, the generated data will be printed to stdout.')
parser.add_argument('-I', dest='includes', metavar='dir', action='append', default=[], help='Add directory to include search path')
parser.add_argument('-D', dest='defines', metavar='define', action='append', default=[], help='Pass a define to the preprocessor')
parser.add_argument('-U', dest='undefines', metavar='undefine', action='append', default=[], help='Pass an undefine to the preprocessor')
args = parser.parse_args()
# Avoid parsing problems due to gcc specifc syntax.
cpp_opts = ['-U__GNUC__']
# Only apply compiler options regarding syntax, includes and defines.
# We have to compile for the current system, we aren't compiling to bitcode after all.
for flag in shared.COMPILER_OPTS:
if flag[:2] in ('-f', '-I', '-i', '-D', '-U'):
cpp_opts.append(flag)
# Add the user options to the list as well.
for path in args.includes:
cpp_opts.append('-I' + path)
for arg in args.defines:
cpp_opts.append('-D' + arg)
for arg in args.undefines:
cpp_opts.append('-U' + arg)
if args.list_fields:
# Just parse the first header and output the result.
structs = parse_header(args.headers[0], cpp_opts)
data = {
'file': args.headers[0],
'structs': structs,
'defines': []
}
if args
|