xref: /linux/Documentation/sphinx/parse-headers.pl (revision b1cce98493a095925fb51be045ccf6e08edb4aa0)
1#!/usr/bin/env perl
2# SPDX-License-Identifier: GPL-2.0
3# Copyright (c) 2016 by Mauro Carvalho Chehab <mchehab@kernel.org>.
4
5use strict;
6use Text::Tabs;
7use Getopt::Long;
8use Pod::Usage;
9
10my $debug;
11my $help;
12my $man;
13
14GetOptions(
15	"debug" => \$debug,
16	'usage|?' => \$help,
17	'help' => \$man
18) or pod2usage(2);
19
20pod2usage(1) if $help;
21pod2usage(-exitstatus => 0, -verbose => 2) if $man;
22pod2usage(2) if (scalar @ARGV < 2 || scalar @ARGV > 3);
23
24my ($file_in, $file_out, $file_exceptions) = @ARGV;
25
26my $data;
27my %ioctls;
28my %defines;
29my %typedefs;
30my %enums;
31my %enum_symbols;
32my %structs;
33
34require Data::Dumper if ($debug);
35
36#
37# read the file and get identifiers
38#
39
40my $is_enum = 0;
41my $is_comment = 0;
42open IN, $file_in or die "Can't open $file_in";
43while (<IN>) {
44	$data .= $_;
45
46	my $ln = $_;
47	if (!$is_comment) {
48		$ln =~ s,/\*.*(\*/),,g;
49
50		$is_comment = 1 if ($ln =~ s,/\*.*,,);
51	} else {
52		if ($ln =~ s,^(.*\*/),,) {
53			$is_comment = 0;
54		} else {
55			next;
56		}
57	}
58
59	if ($is_enum && $ln =~ m/^\s*([_\w][\w\d_]+)\s*[\,=]?/) {
60		my $s = $1;
61		my $n = $1;
62		$n =~ tr/A-Z/a-z/;
63		$n =~ tr/_/-/;
64
65		$enum_symbols{$s} =  "\\ :ref:`$s <$n>`\\ ";
66
67		$is_enum = 0 if ($is_enum && m/\}/);
68		next;
69	}
70	$is_enum = 0 if ($is_enum && m/\}/);
71
72	if ($ln =~ m/^\s*#\s*define\s+([_\w][\w\d_]+)\s+_IO/) {
73		my $s = $1;
74		my $n = $1;
75		$n =~ tr/A-Z/a-z/;
76
77		$ioctls{$s} = "\\ :ref:`$s <$n>`\\ ";
78		next;
79	}
80
81	if ($ln =~ m/^\s*#\s*define\s+([_\w][\w\d_]+)\s+/) {
82		my $s = $1;
83		my $n = $1;
84		$n =~ tr/A-Z/a-z/;
85		$n =~ tr/_/-/;
86
87		$defines{$s} = "\\ :ref:`$s <$n>`\\ ";
88		next;
89	}
90
91	if ($ln =~ m/^\s*typedef\s+([_\w][\w\d_]+)\s+(.*)\s+([_\w][\w\d_]+);/) {
92		my $s = $2;
93		my $n = $3;
94
95		$typedefs{$n} = "\\ :c:type:`$n <$s>`\\ ";
96		next;
97	}
98	if ($ln =~ m/^\s*enum\s+([_\w][\w\d_]+)\s+\{/
99	    || $ln =~ m/^\s*enum\s+([_\w][\w\d_]+)$/
100	    || $ln =~ m/^\s*typedef\s*enum\s+([_\w][\w\d_]+)\s+\{/
101	    || $ln =~ m/^\s*typedef\s*enum\s+([_\w][\w\d_]+)$/) {
102		my $s = $1;
103
104		$enums{$s} =  "enum :c:type:`$s`\\ ";
105
106		$is_enum = $1;
107		next;
108	}
109	if ($ln =~ m/^\s*struct\s+([_\w][\w\d_]+)\s+\{/
110	    || $ln =~ m/^\s*struct\s+([[_\w][\w\d_]+)$/
111	    || $ln =~ m/^\s*typedef\s*struct\s+([_\w][\w\d_]+)\s+\{/
112	    || $ln =~ m/^\s*typedef\s*struct\s+([[_\w][\w\d_]+)$/
113	    ) {
114		my $s = $1;
115
116		$structs{$s} = "struct $s\\ ";
117		next;
118	}
119}
120close IN;
121
122#
123# Handle multi-line typedefs
124#
125
126my @matches = ($data =~ m/typedef\s+struct\s+\S+?\s*\{[^\}]+\}\s*(\S+)\s*\;/g,
127	       $data =~ m/typedef\s+enum\s+\S+?\s*\{[^\}]+\}\s*(\S+)\s*\;/g,);
128foreach my $m (@matches) {
129	my $s = $m;
130
131	$typedefs{$s} = "\\ :c:type:`$s`\\ ";
132	next;
133}
134
135#
136# Handle exceptions, if any
137#
138
139my %def_reftype = (
140	"ioctl"   => ":ref",
141	"define"  => ":ref",
142	"symbol"  => ":ref",
143	"typedef" => ":c:type",
144	"enum"    => ":c:type",
145	"struct"  => ":c:type",
146);
147
148if ($file_exceptions) {
149	open IN, $file_exceptions or die "Can't read $file_exceptions";
150	while (<IN>) {
151		next if (m/^\s*$/ || m/^\s*#/);
152
153		# Parsers to ignore a symbol
154
155		if (m/^ignore\s+ioctl\s+(\S+)/) {
156			delete $ioctls{$1} if (exists($ioctls{$1}));
157			next;
158		}
159		if (m/^ignore\s+define\s+(\S+)/) {
160			delete $defines{$1} if (exists($defines{$1}));
161			next;
162		}
163		if (m/^ignore\s+typedef\s+(\S+)/) {
164			delete $typedefs{$1} if (exists($typedefs{$1}));
165			next;
166		}
167		if (m/^ignore\s+enum\s+(\S+)/) {
168			delete $enums{$1} if (exists($enums{$1}));
169			next;
170		}
171		if (m/^ignore\s+struct\s+(\S+)/) {
172			delete $structs{$1} if (exists($structs{$1}));
173			next;
174		}
175		if (m/^ignore\s+symbol\s+(\S+)/) {
176			delete $enum_symbols{$1} if (exists($enum_symbols{$1}));
177			next;
178		}
179
180		# Parsers to replace a symbol
181		my ($type, $old, $new, $reftype);
182
183		if (m/^replace\s+(\S+)\s+(\S+)\s+(\S+)/) {
184			$type = $1;
185			$old = $2;
186			$new = $3;
187		} else {
188			die "Can't parse $file_exceptions: $_";
189		}
190
191		if ($new =~ m/^\:c\:(data|func|macro|type)\:\`(.+)\`/) {
192			$reftype = ":c:$1";
193			$new = $2;
194		} elsif ($new =~ m/\:ref\:\`(.+)\`/) {
195			$reftype = ":ref";
196			$new = $1;
197		} else {
198			$reftype = $def_reftype{$type};
199		}
200		$new = "$reftype:`$old <$new>`";
201
202		if ($type eq "ioctl") {
203			$ioctls{$old} = $new if (exists($ioctls{$old}));
204			next;
205		}
206		if ($type eq "define") {
207			$defines{$old} = $new if (exists($defines{$old}));
208			next;
209		}
210		if ($type eq "symbol") {
211			$enum_symbols{$old} = $new if (exists($enum_symbols{$old}));
212			next;
213		}
214		if ($type eq "typedef") {
215			$typedefs{$old} = $new if (exists($typedefs{$old}));
216			next;
217		}
218		if ($type eq "enum") {
219			$enums{$old} = $new if (exists($enums{$old}));
220			next;
221		}
222		if ($type eq "struct") {
223			$structs{$old} = $new if (exists($structs{$old}));
224			next;
225		}
226
227		die "Can't parse $file_exceptions: $_";
228	}
229}
230
231if ($debug) {
232	print Data::Dumper->Dump([\%ioctls], [qw(*ioctls)]) if (%ioctls);
233	print Data::Dumper->Dump([\%typedefs], [qw(*typedefs)]) if (%typedefs);
234	print Data::Dumper->Dump([\%enums], [qw(*enums)]) if (%enums);
235	print Data::Dumper->Dump([\%structs], [qw(*structs)]) if (%structs);
236	print Data::Dumper->Dump([\%defines], [qw(*defines)]) if (%defines);
237	print Data::Dumper->Dump([\%enum_symbols], [qw(*enum_symbols)]) if (%enum_symbols);
238}
239
240#
241# Align block
242#
243$data = expand($data);
244$data = "    " . $data;
245$data =~ s/\n/\n    /g;
246$data =~ s/\n\s+$/\n/g;
247$data =~ s/\n\s+\n/\n\n/g;
248
249#
250# Add escape codes for special characters
251#
252$data =~ s,([\_\`\*\<\>\&\\\\:\/\|\%\$\#\{\}\~\^]),\\$1,g;
253
254$data =~ s,DEPRECATED,**DEPRECATED**,g;
255
256#
257# Add references
258#
259
260my $start_delim = "[ \n\t\(\=\*\@]";
261my $end_delim = "(\\s|,|\\\\=|\\\\:|\\;|\\\)|\\}|\\{)";
262
263foreach my $r (keys %ioctls) {
264	my $s = $ioctls{$r};
265
266	$r =~ s,([\_\`\*\<\>\&\\\\:\/]),\\\\$1,g;
267
268	print "$r -> $s\n" if ($debug);
269
270	$data =~ s/($start_delim)($r)$end_delim/$1$s$3/g;
271}
272
273foreach my $r (keys %defines) {
274	my $s = $defines{$r};
275
276	$r =~ s,([\_\`\*\<\>\&\\\\:\/]),\\\\$1,g;
277
278	print "$r -> $s\n" if ($debug);
279
280	$data =~ s/($start_delim)($r)$end_delim/$1$s$3/g;
281}
282
283foreach my $r (keys %enum_symbols) {
284	my $s = $enum_symbols{$r};
285
286	$r =~ s,([\_\`\*\<\>\&\\\\:\/]),\\\\$1,g;
287
288	print "$r -> $s\n" if ($debug);
289
290	$data =~ s/($start_delim)($r)$end_delim/$1$s$3/g;
291}
292
293foreach my $r (keys %enums) {
294	my $s = $enums{$r};
295
296	$r =~ s,([\_\`\*\<\>\&\\\\:\/]),\\\\$1,g;
297
298	print "$r -> $s\n" if ($debug);
299
300	$data =~ s/enum\s+($r)$end_delim/$s$2/g;
301}
302
303foreach my $r (keys %structs) {
304	my $s = $structs{$r};
305
306	$r =~ s,([\_\`\*\<\>\&\\\\:\/]),\\\\$1,g;
307
308	print "$r -> $s\n" if ($debug);
309
310	$data =~ s/struct\s+($r)$end_delim/$s$2/g;
311}
312
313foreach my $r (keys %typedefs) {
314	my $s = $typedefs{$r};
315
316	$r =~ s,([\_\`\*\<\>\&\\\\:\/]),\\\\$1,g;
317
318	print "$r -> $s\n" if ($debug);
319	$data =~ s/($start_delim)($r)$end_delim/$1$s$3/g;
320}
321
322$data =~ s/\\ ([\n\s])/\1/g;
323
324#
325# Generate output file
326#
327
328my $title = $file_in;
329$title =~ s,.*/,,;
330
331open OUT, "> $file_out" or die "Can't open $file_out";
332print OUT ".. -*- coding: utf-8; mode: rst -*-\n\n";
333print OUT "$title\n";
334print OUT "=" x length($title);
335print OUT "\n\n.. parsed-literal::\n\n";
336print OUT $data;
337close OUT;
338
339__END__
340
341=head1 NAME
342
343parse_headers.pl - parse a C file, in order to identify functions, structs,
344enums and defines and create cross-references to a Sphinx book.
345
346=head1 SYNOPSIS
347
348B<parse_headers.pl> [<options>] <C_FILE> <OUT_FILE> [<EXCEPTIONS_FILE>]
349
350Where <options> can be: --debug, --help or --usage.
351
352=head1 OPTIONS
353
354=over 8
355
356=item B<--debug>
357
358Put the script in verbose mode, useful for debugging.
359
360=item B<--usage>
361
362Prints a brief help message and exits.
363
364=item B<--help>
365
366Prints a more detailed help message and exits.
367
368=back
369
370=head1 DESCRIPTION
371
372Convert a C header or source file (C_FILE), into a ReStructured Text
373included via ..parsed-literal block with cross-references for the
374documentation files that describe the API. It accepts an optional
375EXCEPTIONS_FILE with describes what elements will be either ignored or
376be pointed to a non-default reference.
377
378The output is written at the (OUT_FILE).
379
380It is capable of identifying defines, functions, structs, typedefs,
381enums and enum symbols and create cross-references for all of them.
382It is also capable of distinguish #define used for specifying a Linux
383ioctl.
384
385The EXCEPTIONS_FILE contain two rules to allow ignoring a symbol or
386to replace the default references by a custom one.
387
388Please read Documentation/doc-guide/parse-headers.rst at the Kernel's
389tree for more details.
390
391=head1 BUGS
392
393Report bugs to Mauro Carvalho Chehab <mchehab@kernel.org>
394
395=head1 COPYRIGHT
396
397Copyright (c) 2016 by Mauro Carvalho Chehab <mchehab@kernel.org>.
398
399License GPLv2: GNU GPL version 2 <https://gnu.org/licenses/gpl.html>.
400
401This is free software: you are free to change and redistribute it.
402There is NO WARRANTY, to the extent permitted by law.
403
404=cut
405