scripts: get_abi.pl: parse description line per line
[linux-2.6-block.git] / scripts / get_abi.pl
... / ...
CommitLineData
1#!/usr/bin/env perl
2# SPDX-License-Identifier: GPL-2.0
3
4use strict;
5use warnings;
6use utf8;
7use Pod::Usage;
8use Getopt::Long;
9use File::Find;
10use Fcntl ':mode';
11
12my $help = 0;
13my $man = 0;
14my $debug = 0;
15my $enable_lineno = 0;
16my $prefix="Documentation/ABI";
17
18#
19# If true, assumes that the description is formatted with ReST
20#
21my $description_is_rst = 1;
22
23GetOptions(
24 "debug|d+" => \$debug,
25 "enable-lineno" => \$enable_lineno,
26 "rst-source!" => \$description_is_rst,
27 "dir=s" => \$prefix,
28 'help|?' => \$help,
29 man => \$man
30) or pod2usage(2);
31
32pod2usage(1) if $help;
33pod2usage(-exitstatus => 0, -verbose => 2) if $man;
34
35pod2usage(2) if (scalar @ARGV < 1 || @ARGV > 2);
36
37my ($cmd, $arg) = @ARGV;
38
39pod2usage(2) if ($cmd ne "search" && $cmd ne "rest" && $cmd ne "validate");
40pod2usage(2) if ($cmd eq "search" && !$arg);
41
42require Data::Dumper if ($debug);
43
44my %data;
45my %symbols;
46
47#
48# Displays an error message, printing file name and line
49#
50sub parse_error($$$$) {
51 my ($file, $ln, $msg, $data) = @_;
52
53 $data =~ s/\s+$/\n/;
54
55 print STDERR "Warning: file $file#$ln:\n\t$msg";
56
57 if ($data ne "") {
58 print STDERR ". Line\n\t\t$data";
59 } else {
60 print STDERR "\n";
61 }
62}
63
64#
65# Parse an ABI file, storing its contents at %data
66#
67sub parse_abi {
68 my $file = $File::Find::name;
69
70 my $mode = (stat($file))[2];
71 return if ($mode & S_IFDIR);
72 return if ($file =~ m,/README,);
73
74 my $name = $file;
75 $name =~ s,.*/,,;
76
77 my $fn = $file;
78 $fn =~ s,Documentation/ABI/,,;
79
80 my $nametag = "File $fn";
81 $data{$nametag}->{what} = "File $name";
82 $data{$nametag}->{type} = "File";
83 $data{$nametag}->{file} = $name;
84 $data{$nametag}->{filepath} = $file;
85 $data{$nametag}->{is_file} = 1;
86 $data{$nametag}->{line_no} = 1;
87
88 my $type = $file;
89 $type =~ s,.*/(.*)/.*,$1,;
90
91 my $what;
92 my $new_what;
93 my $tag = "";
94 my $ln;
95 my $xrefs;
96 my $space;
97 my @labels;
98 my $label = "";
99
100 print STDERR "Opening $file\n" if ($debug > 1);
101 open IN, $file;
102 while(<IN>) {
103 $ln++;
104 if (m/^(\S+)(:\s*)(.*)/i) {
105 my $new_tag = lc($1);
106 my $sep = $2;
107 my $content = $3;
108
109 if (!($new_tag =~ m/(what|where|date|kernelversion|contact|description|users)/)) {
110 if ($tag eq "description") {
111 # New "tag" is actually part of
112 # description. Don't consider it a tag
113 $new_tag = "";
114 } elsif ($tag ne "") {
115 parse_error($file, $ln, "tag '$tag' is invalid", $_);
116 }
117 }
118
119 # Invalid, but it is a common mistake
120 if ($new_tag eq "where") {
121 parse_error($file, $ln, "tag 'Where' is invalid. Should be 'What:' instead", "");
122 $new_tag = "what";
123 }
124
125 if ($new_tag =~ m/what/) {
126 $space = "";
127 $content =~ s/[,.;]$//;
128
129 push @{$symbols{$content}->{file}}, " $file:" . ($ln - 1);
130
131 if ($tag =~ m/what/) {
132 $what .= ", " . $content;
133 } else {
134 if ($what) {
135 parse_error($file, $ln, "What '$what' doesn't have a description", "") if (!$data{$what}->{description});
136
137 foreach my $w(split /, /, $what) {
138 $symbols{$w}->{xref} = $what;
139 };
140 }
141
142 $what = $content;
143 $label = $content;
144 $new_what = 1;
145 }
146 push @labels, [($content, $label)];
147 $tag = $new_tag;
148
149 push @{$data{$nametag}->{symbols}}, $content if ($data{$nametag}->{what});
150 next;
151 }
152
153 if ($tag ne "" && $new_tag) {
154 $tag = $new_tag;
155
156 if ($new_what) {
157 @{$data{$what}->{label_list}} = @labels if ($data{$nametag}->{what});
158 @labels = ();
159 $label = "";
160 $new_what = 0;
161
162 $data{$what}->{type} = $type;
163 if (!defined($data{$what}->{file})) {
164 $data{$what}->{file} = $name;
165 $data{$what}->{filepath} = $file;
166 } else {
167 if ($name ne $data{$what}->{file}) {
168 $data{$what}->{file} .= " " . $name;
169 $data{$what}->{filepath} .= " " . $file;
170 }
171 }
172 print STDERR "\twhat: $what\n" if ($debug > 1);
173 $data{$what}->{line_no} = $ln;
174 } else {
175 $data{$what}->{line_no} = $ln if (!defined($data{$what}->{line_no}));
176 }
177
178 if (!$what) {
179 parse_error($file, $ln, "'What:' should come first:", $_);
180 next;
181 }
182 if ($new_tag eq "description") {
183 $sep =~ s,:, ,;
184 $content = ' ' x length($new_tag) . $sep . $content;
185 while ($content =~ s/\t+/' ' x (length($&) * 8 - length($`) % 8)/e) {}
186 if ($content =~ m/^(\s*)(\S.*)$/) {
187 # Preserve initial spaces for the first line
188 $space = $1;
189 $content = "$2\n";
190 $data{$what}->{$tag} .= $content;
191 } else {
192 undef($space);
193 }
194
195 } else {
196 $data{$what}->{$tag} = $content;
197 }
198 next;
199 }
200 }
201
202 # Store any contents before tags at the database
203 if (!$tag && $data{$nametag}->{what}) {
204 $data{$nametag}->{description} .= $_;
205 next;
206 }
207
208 if ($tag eq "description") {
209 my $content = $_;
210 while ($content =~ s/\t+/' ' x (length($&) * 8 - length($`) % 8)/e) {}
211 if (m/^\s*\n/) {
212 $data{$what}->{$tag} .= "\n";
213 next;
214 }
215
216 if (!defined($space)) {
217 # Preserve initial spaces for the first line
218 if ($content =~ m/^(\s*)(\S.*)$/) {
219 $space = $1;
220 $content = "$2\n";
221 }
222 } else {
223 $space = "" if (!($content =~ s/^($space)//));
224 }
225 $data{$what}->{$tag} .= $content;
226
227 next;
228 }
229 if (m/^\s*(.*)/) {
230 $data{$what}->{$tag} .= "\n$1";
231 $data{$what}->{$tag} =~ s/\n+$//;
232 next;
233 }
234
235 # Everything else is error
236 parse_error($file, $ln, "Unexpected content", $_);
237 }
238 $data{$nametag}->{description} =~ s/^\n+// if ($data{$nametag}->{description});
239 if ($what) {
240 parse_error($file, $ln, "What '$what' doesn't have a description", "") if (!$data{$what}->{description});
241
242 foreach my $w(split /, /,$what) {
243 $symbols{$w}->{xref} = $what;
244 };
245 }
246 close IN;
247}
248
249sub create_labels {
250 my %labels;
251
252 foreach my $what (keys %data) {
253 next if ($data{$what}->{file} eq "File");
254
255 foreach my $p (@{$data{$what}->{label_list}}) {
256 my ($content, $label) = @{$p};
257 $label = "abi_" . $label . " ";
258 $label =~ tr/A-Z/a-z/;
259
260 # Convert special chars to "_"
261 $label =~s/([\x00-\x2f\x3a-\x40\x5b-\x60\x7b-\xff])/_/g;
262 $label =~ s,_+,_,g;
263 $label =~ s,_$,,;
264
265 # Avoid duplicated labels
266 while (defined($labels{$label})) {
267 my @chars = ("A".."Z", "a".."z");
268 $label .= $chars[rand @chars];
269 }
270 $labels{$label} = 1;
271
272 $data{$what}->{label} = $label;
273
274 # only one label is enough
275 last;
276 }
277 }
278}
279
280#
281# Outputs the book on ReST format
282#
283
284# \b doesn't work well with paths. So, we need to define something else:
285# Boundaries are punct characters, spaces and end-of-line
286my $start = qr {(^|\s|\() }x;
287my $bondary = qr { ([,.:;\)\s]|\z) }x;
288my $xref_match = qr { $start(\/(sys|config|proc|dev|kvd)\/[^,.:;\)\s]+)$bondary }x;
289my $symbols = qr { ([\x01-\x08\x0e-\x1f\x21-\x2f\x3a-\x40\x7b-\xff]) }x;
290
291sub output_rest {
292 create_labels();
293
294 my $part = "";
295
296 foreach my $what (sort {
297 ($data{$a}->{type} eq "File") cmp ($data{$b}->{type} eq "File") ||
298 $a cmp $b
299 } keys %data) {
300 my $type = $data{$what}->{type};
301
302 my @file = split / /, $data{$what}->{file};
303 my @filepath = split / /, $data{$what}->{filepath};
304
305 if ($enable_lineno) {
306 printf "#define LINENO %s%s#%s\n\n",
307 $prefix, $file[0],
308 $data{$what}->{line_no};
309 }
310
311 my $w = $what;
312
313 if ($type ne "File") {
314 my $cur_part = $what;
315 if ($what =~ '/') {
316 if ($what =~ m#^(\/?(?:[\w\-]+\/?){1,2})#) {
317 $cur_part = "Symbols under $1";
318 $cur_part =~ s,/$,,;
319 }
320 }
321
322 if ($cur_part ne "" && $part ne $cur_part) {
323 $part = $cur_part;
324 my $bar = $part;
325 $bar =~ s/./-/g;
326 print "$part\n$bar\n\n";
327 }
328
329 printf ".. _%s:\n\n", $data{$what}->{label};
330
331 my @names = split /, /,$w;
332 my $len = 0;
333
334 foreach my $name (@names) {
335 $name =~ s/$symbols/\\$1/g;
336 $name = "**$name**";
337 $len = length($name) if (length($name) > $len);
338 }
339
340 print "+-" . "-" x $len . "-+\n";
341 foreach my $name (@names) {
342 printf "| %s", $name . " " x ($len - length($name)) . " |\n";
343 print "+-" . "-" x $len . "-+\n";
344 }
345
346 print "\n";
347 }
348
349 for (my $i = 0; $i < scalar(@filepath); $i++) {
350 my $path = $filepath[$i];
351 my $f = $file[$i];
352
353 $path =~ s,.*/(.*/.*),$1,;;
354 $path =~ s,[/\-],_,g;;
355 my $fileref = "abi_file_".$path;
356
357 if ($type eq "File") {
358 print ".. _$fileref:\n\n";
359 } else {
360 print "Defined on file :ref:`$f <$fileref>`\n\n";
361 }
362 }
363
364 if ($type eq "File") {
365 my $bar = $w;
366 $bar =~ s/./-/g;
367 print "$w\n$bar\n\n";
368 }
369
370 my $desc = "";
371 $desc = $data{$what}->{description} if (defined($data{$what}->{description}));
372 $desc =~ s/\s+$/\n/;
373
374 if (!($desc =~ /^\s*$/)) {
375 if ($description_is_rst) {
376 # Remove title markups from the description
377 # Having titles inside ABI files will only work if extra
378 # care would be taken in order to strictly follow the same
379 # level order for each markup.
380 $desc =~ s/\n[\-\*\=\^\~]+\n/\n\n/g;
381
382 # Enrich text by creating cross-references
383
384 my $new_desc = "";
385 open(my $fh, "+<", \$desc);
386 while (my $d = <$fh>) {
387 $d =~ s,Documentation/(?!devicetree)(\S+)\.rst,:doc:`/$1`,g;
388
389 my @matches = $d =~ m,Documentation/ABI/([\w\/\-]+),g;
390 foreach my $f (@matches) {
391 my $xref = $f;
392 my $path = $f;
393 $path =~ s,.*/(.*/.*),$1,;;
394 $path =~ s,[/\-],_,g;;
395 $xref .= " <abi_file_" . $path . ">";
396 $d =~ s,\bDocumentation/ABI/$f\b,:ref:`$xref`,g;
397 }
398
399 # Seek for cross reference symbols like /sys/...
400 @matches = $d =~ m/$xref_match/g;
401
402 foreach my $s (@matches) {
403 next if (!($s =~ m,/,));
404 if (defined($data{$s}) && defined($data{$s}->{label})) {
405 my $xref = $s;
406
407 $xref =~ s/$symbols/\\$1/g;
408 $xref = ":ref:`$xref <" . $data{$s}->{label} . ">`";
409
410 $d =~ s,$start$s$bondary,$1$xref$2,g;
411 }
412 }
413 $new_desc .= $d;
414 }
415 close $fh;
416
417
418 print "$new_desc\n\n";
419 } else {
420 $desc =~ s/^\s+//;
421
422 # Remove title markups from the description, as they won't work
423 $desc =~ s/\n[\-\*\=\^\~]+\n/\n\n/g;
424
425 if ($desc =~ m/\:\n/ || $desc =~ m/\n[\t ]+/ || $desc =~ m/[\x00-\x08\x0b-\x1f\x7b-\xff]/) {
426 # put everything inside a code block
427 $desc =~ s/\n/\n /g;
428
429 print "::\n\n";
430 print " $desc\n\n";
431 } else {
432 # Escape any special chars from description
433 $desc =~s/([\x00-\x08\x0b-\x1f\x21-\x2a\x2d\x2f\x3c-\x40\x5c\x5e-\x60\x7b-\xff])/\\$1/g;
434 print "$desc\n\n";
435 }
436 }
437 } else {
438 print "DESCRIPTION MISSING for $what\n\n" if (!$data{$what}->{is_file});
439 }
440
441 if ($data{$what}->{symbols}) {
442 printf "Has the following ABI:\n\n";
443
444 foreach my $content(@{$data{$what}->{symbols}}) {
445 my $label = $data{$symbols{$content}->{xref}}->{label};
446
447 # Escape special chars from content
448 $content =~s/([\x00-\x1f\x21-\x2f\x3a-\x40\x7b-\xff])/\\$1/g;
449
450 print "- :ref:`$content <$label>`\n\n";
451 }
452 }
453
454 if (defined($data{$what}->{users})) {
455 my $users = $data{$what}->{users};
456
457 $users =~ s/\n/\n\t/g;
458 printf "Users:\n\t%s\n\n", $users if ($users ne "");
459 }
460
461 }
462}
463
464#
465# Searches for ABI symbols
466#
467sub search_symbols {
468 foreach my $what (sort keys %data) {
469 next if (!($what =~ m/($arg)/));
470
471 my $type = $data{$what}->{type};
472 next if ($type eq "File");
473
474 my $file = $data{$what}->{filepath};
475
476 my $bar = $what;
477 $bar =~ s/./-/g;
478
479 print "\n$what\n$bar\n\n";
480
481 my $kernelversion = $data{$what}->{kernelversion} if (defined($data{$what}->{kernelversion}));
482 my $contact = $data{$what}->{contact} if (defined($data{$what}->{contact}));
483 my $users = $data{$what}->{users} if (defined($data{$what}->{users}));
484 my $date = $data{$what}->{date} if (defined($data{$what}->{date}));
485 my $desc = $data{$what}->{description} if (defined($data{$what}->{description}));
486
487 $kernelversion =~ s/^\s+// if ($kernelversion);
488 $contact =~ s/^\s+// if ($contact);
489 if ($users) {
490 $users =~ s/^\s+//;
491 $users =~ s/\n//g;
492 }
493 $date =~ s/^\s+// if ($date);
494 $desc =~ s/^\s+// if ($desc);
495
496 printf "Kernel version:\t\t%s\n", $kernelversion if ($kernelversion);
497 printf "Date:\t\t\t%s\n", $date if ($date);
498 printf "Contact:\t\t%s\n", $contact if ($contact);
499 printf "Users:\t\t\t%s\n", $users if ($users);
500 print "Defined on file(s):\t$file\n\n";
501 print "Description:\n\n$desc";
502 }
503}
504
505# Ensure that the prefix will always end with a slash
506# While this is not needed for find, it makes the patch nicer
507# with --enable-lineno
508$prefix =~ s,/?$,/,;
509
510#
511# Parses all ABI files located at $prefix dir
512#
513find({wanted =>\&parse_abi, no_chdir => 1}, $prefix);
514
515print STDERR Data::Dumper->Dump([\%data], [qw(*data)]) if ($debug);
516
517#
518# Handles the command
519#
520if ($cmd eq "search") {
521 search_symbols;
522} else {
523 if ($cmd eq "rest") {
524 output_rest;
525 }
526
527 # Warn about duplicated ABI entries
528 foreach my $what(sort keys %symbols) {
529 my @files = @{$symbols{$what}->{file}};
530
531 next if (scalar(@files) == 1);
532
533 printf STDERR "Warning: $what is defined %d times: @files\n",
534 scalar(@files);
535 }
536}
537
538__END__
539
540=head1 NAME
541
542abi_book.pl - parse the Linux ABI files and produce a ReST book.
543
544=head1 SYNOPSIS
545
546B<abi_book.pl> [--debug] [--enable-lineno] [--man] [--help]
547 [--(no-)rst-source] [--dir=<dir>] <COMAND> [<ARGUMENT>]
548
549Where <COMMAND> can be:
550
551=over 8
552
553B<search> [SEARCH_REGEX] - search for [SEARCH_REGEX] inside ABI
554
555B<rest> - output the ABI in ReST markup language
556
557B<validate> - validate the ABI contents
558
559=back
560
561=head1 OPTIONS
562
563=over 8
564
565=item B<--dir>
566
567Changes the location of the ABI search. By default, it uses
568the Documentation/ABI directory.
569
570=item B<--rst-source> and B<--no-rst-source>
571
572The input file may be using ReST syntax or not. Those two options allow
573selecting between a rst-compliant source ABI (--rst-source), or a
574plain text that may be violating ReST spec, so it requres some escaping
575logic (--no-rst-source).
576
577=item B<--enable-lineno>
578
579Enable output of #define LINENO lines.
580
581=item B<--debug>
582
583Put the script in verbose mode, useful for debugging. Can be called multiple
584times, to increase verbosity.
585
586=item B<--help>
587
588Prints a brief help message and exits.
589
590=item B<--man>
591
592Prints the manual page and exits.
593
594=back
595
596=head1 DESCRIPTION
597
598Parse the Linux ABI files from ABI DIR (usually located at Documentation/ABI),
599allowing to search for ABI symbols or to produce a ReST book containing
600the Linux ABI documentation.
601
602=head1 EXAMPLES
603
604Search for all stable symbols with the word "usb":
605
606=over 8
607
608$ scripts/get_abi.pl search usb --dir Documentation/ABI/stable
609
610=back
611
612Search for all symbols that match the regex expression "usb.*cap":
613
614=over 8
615
616$ scripts/get_abi.pl search usb.*cap
617
618=back
619
620Output all obsoleted symbols in ReST format
621
622=over 8
623
624$ scripts/get_abi.pl rest --dir Documentation/ABI/obsolete
625
626=back
627
628=head1 BUGS
629
630Report bugs to Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
631
632=head1 COPYRIGHT
633
634Copyright (c) 2016-2019 by Mauro Carvalho Chehab <mchehab+samsung@kernel.org>.
635
636License GPLv2: GNU GPL version 2 <http://gnu.org/licenses/gpl.html>.
637
638This is free software: you are free to change and redistribute it.
639There is NO WARRANTY, to the extent permitted by law.
640
641=cut