1#!/usr/bin/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
285my $bondary = qr { (?<![\w\/\`\{])(?=[\w\/\`\{])|(?<=[\w\/\`\{])(?![\w\/\`\{]) }x;
286
287sub output_rest {
288	create_labels();
289
290	my $part = "";
291
292	foreach my $what (sort {
293				($data{$a}->{type} eq "File") cmp ($data{$b}->{type} eq "File") ||
294				$a cmp $b
295			       } keys %data) {
296		my $type = $data{$what}->{type};
297
298		my @file = split / /, $data{$what}->{file};
299		my @filepath = split / /, $data{$what}->{filepath};
300
301		if ($enable_lineno) {
302			printf "#define LINENO %s%s#%s\n\n",
303			       $prefix, $file[0],
304			       $data{$what}->{line_no};
305		}
306
307		my $w = $what;
308		$w =~ s/([\(\)\_\-\*\=\^\~\\])/\\$1/g;
309
310		if ($type ne "File") {
311			my $cur_part = $what;
312			if ($what =~ '/') {
313				if ($what =~ m#^(\/?(?:[\w\-]+\/?){1,2})#) {
314					$cur_part = "Symbols under $1";
315					$cur_part =~ s,/$,,;
316				}
317			}
318
319			if ($cur_part ne "" && $part ne $cur_part) {
320			    $part = $cur_part;
321			    my $bar = $part;
322			    $bar =~ s/./-/g;
323			    print "$part\n$bar\n\n";
324			}
325
326			printf ".. _%s:\n\n", $data{$what}->{label};
327
328			my @names = split /, /,$w;
329			my $len = 0;
330
331			foreach my $name (@names) {
332				$name = "**$name**";
333				$len = length($name) if (length($name) > $len);
334			}
335
336			print "+-" . "-" x $len . "-+\n";
337			foreach my $name (@names) {
338				printf "| %s", $name . " " x ($len - length($name)) . " |\n";
339				print "+-" . "-" x $len . "-+\n";
340			}
341
342			print "\n";
343		}
344
345		for (my $i = 0; $i < scalar(@filepath); $i++) {
346			my $path = $filepath[$i];
347			my $f = $file[$i];
348
349			$path =~ s,.*/(.*/.*),$1,;;
350			$path =~ s,[/\-],_,g;;
351			my $fileref = "abi_file_".$path;
352
353			if ($type eq "File") {
354				print ".. _$fileref:\n\n";
355			} else {
356				print "Defined on file :ref:`$f <$fileref>`\n\n";
357			}
358		}
359
360		if ($type eq "File") {
361			my $bar = $w;
362			$bar =~ s/./-/g;
363			print "$w\n$bar\n\n";
364		}
365
366		my $desc = "";
367		$desc = $data{$what}->{description} if (defined($data{$what}->{description}));
368		$desc =~ s/\s+$/\n/;
369
370		if (!($desc =~ /^\s*$/)) {
371			if ($description_is_rst) {
372				# Remove title markups from the description
373				# Having titles inside ABI files will only work if extra
374				# care would be taken in order to strictly follow the same
375				# level order for each markup.
376				$desc =~ s/\n[\-\*\=\^\~]+\n/\n\n/g;
377
378				# Enrich text by creating cross-references
379
380				$desc =~ s,Documentation/(?!devicetree)(\S+)\.rst,:doc:`/$1`,g;
381
382				my @matches = $desc =~ m,Documentation/ABI/([\w\/\-]+),;
383				foreach my $f (@matches) {
384					my $xref = $f;
385					my $path = $f;
386					$path =~ s,.*/(.*/.*),$1,;;
387					$path =~ s,[/\-],_,g;;
388					$xref .= " <abi_file_" . $path . ">";
389					$desc =~ s,\bDocumentation/ABI/$f\b,:ref:`$xref`,g;
390				}
391
392				@matches = $desc =~ m,$bondary(/sys/[^\s\.\,\;\:\*\s\`\'\(\)]+)$bondary,;
393
394				foreach my $s (@matches) {
395					if (defined($data{$s}) && defined($data{$s}->{label})) {
396						my $xref = $s;
397
398						$xref =~ s/([\x00-\x1f\x21-\x2f\x3a-\x40\x7b-\xff])/\\$1/g;
399						$xref = ":ref:`$xref <" . $data{$s}->{label} . ">`";
400
401						$desc =~ s,$bondary$s$bondary,$xref,g;
402					}
403				}
404
405				print "$desc\n\n";
406			} else {
407				$desc =~ s/^\s+//;
408
409				# Remove title markups from the description, as they won't work
410				$desc =~ s/\n[\-\*\=\^\~]+\n/\n\n/g;
411
412				if ($desc =~ m/\:\n/ || $desc =~ m/\n[\t ]+/  || $desc =~ m/[\x00-\x08\x0b-\x1f\x7b-\xff]/) {
413					# put everything inside a code block
414					$desc =~ s/\n/\n /g;
415
416					print "::\n\n";
417					print " $desc\n\n";
418				} else {
419					# Escape any special chars from description
420					$desc =~s/([\x00-\x08\x0b-\x1f\x21-\x2a\x2d\x2f\x3c-\x40\x5c\x5e-\x60\x7b-\xff])/\\$1/g;
421					print "$desc\n\n";
422				}
423			}
424		} else {
425			print "DESCRIPTION MISSING for $what\n\n" if (!$data{$what}->{is_file});
426		}
427
428		if ($data{$what}->{symbols}) {
429			printf "Has the following ABI:\n\n";
430
431			foreach my $content(@{$data{$what}->{symbols}}) {
432				my $label = $data{$symbols{$content}->{xref}}->{label};
433
434				# Escape special chars from content
435				$content =~s/([\x00-\x1f\x21-\x2f\x3a-\x40\x7b-\xff])/\\$1/g;
436
437				print "- :ref:`$content <$label>`\n\n";
438			}
439		}
440
441		if (defined($data{$what}->{users})) {
442			my $users = $data{$what}->{users};
443
444			$users =~ s/\n/\n\t/g;
445			printf "Users:\n\t%s\n\n", $users if ($users ne "");
446		}
447
448	}
449}
450
451#
452# Searches for ABI symbols
453#
454sub search_symbols {
455	foreach my $what (sort keys %data) {
456		next if (!($what =~ m/($arg)/));
457
458		my $type = $data{$what}->{type};
459		next if ($type eq "File");
460
461		my $file = $data{$what}->{filepath};
462
463		my $bar = $what;
464		$bar =~ s/./-/g;
465
466		print "\n$what\n$bar\n\n";
467
468		my $kernelversion = $data{$what}->{kernelversion} if (defined($data{$what}->{kernelversion}));
469		my $contact = $data{$what}->{contact} if (defined($data{$what}->{contact}));
470		my $users = $data{$what}->{users} if (defined($data{$what}->{users}));
471		my $date = $data{$what}->{date} if (defined($data{$what}->{date}));
472		my $desc = $data{$what}->{description} if (defined($data{$what}->{description}));
473
474		$kernelversion =~ s/^\s+// if ($kernelversion);
475		$contact =~ s/^\s+// if ($contact);
476		if ($users) {
477			$users =~ s/^\s+//;
478			$users =~ s/\n//g;
479		}
480		$date =~ s/^\s+// if ($date);
481		$desc =~ s/^\s+// if ($desc);
482
483		printf "Kernel version:\t\t%s\n", $kernelversion if ($kernelversion);
484		printf "Date:\t\t\t%s\n", $date if ($date);
485		printf "Contact:\t\t%s\n", $contact if ($contact);
486		printf "Users:\t\t\t%s\n", $users if ($users);
487		print "Defined on file(s):\t$file\n\n";
488		print "Description:\n\n$desc";
489	}
490}
491
492# Ensure that the prefix will always end with a slash
493# While this is not needed for find, it makes the patch nicer
494# with --enable-lineno
495$prefix =~ s,/?$,/,;
496
497#
498# Parses all ABI files located at $prefix dir
499#
500find({wanted =>\&parse_abi, no_chdir => 1}, $prefix);
501
502print STDERR Data::Dumper->Dump([\%data], [qw(*data)]) if ($debug);
503
504#
505# Handles the command
506#
507if ($cmd eq "search") {
508	search_symbols;
509} else {
510	if ($cmd eq "rest") {
511		output_rest;
512	}
513
514	# Warn about duplicated ABI entries
515	foreach my $what(sort keys %symbols) {
516		my @files = @{$symbols{$what}->{file}};
517
518		next if (scalar(@files) == 1);
519
520		printf STDERR "Warning: $what is defined %d times: @files\n",
521		    scalar(@files);
522	}
523}
524
525__END__
526
527=head1 NAME
528
529abi_book.pl - parse the Linux ABI files and produce a ReST book.
530
531=head1 SYNOPSIS
532
533B<abi_book.pl> [--debug] [--enable-lineno] [--man] [--help]
534	       [--(no-)rst-source] [--dir=<dir>] <COMAND> [<ARGUMENT>]
535
536Where <COMMAND> can be:
537
538=over 8
539
540B<search> [SEARCH_REGEX] - search for [SEARCH_REGEX] inside ABI
541
542B<rest>                  - output the ABI in ReST markup language
543
544B<validate>              - validate the ABI contents
545
546=back
547
548=head1 OPTIONS
549
550=over 8
551
552=item B<--dir>
553
554Changes the location of the ABI search. By default, it uses
555the Documentation/ABI directory.
556
557=item B<--rst-source> and B<--no-rst-source>
558
559The input file may be using ReST syntax or not. Those two options allow
560selecting between a rst-compliant source ABI (--rst-source), or a
561plain text that may be violating ReST spec, so it requres some escaping
562logic (--no-rst-source).
563
564=item B<--enable-lineno>
565
566Enable output of #define LINENO lines.
567
568=item B<--debug>
569
570Put the script in verbose mode, useful for debugging. Can be called multiple
571times, to increase verbosity.
572
573=item B<--help>
574
575Prints a brief help message and exits.
576
577=item B<--man>
578
579Prints the manual page and exits.
580
581=back
582
583=head1 DESCRIPTION
584
585Parse the Linux ABI files from ABI DIR (usually located at Documentation/ABI),
586allowing to search for ABI symbols or to produce a ReST book containing
587the Linux ABI documentation.
588
589=head1 EXAMPLES
590
591Search for all stable symbols with the word "usb":
592
593=over 8
594
595$ scripts/get_abi.pl search usb --dir Documentation/ABI/stable
596
597=back
598
599Search for all symbols that match the regex expression "usb.*cap":
600
601=over 8
602
603$ scripts/get_abi.pl search usb.*cap
604
605=back
606
607Output all obsoleted symbols in ReST format
608
609=over 8
610
611$ scripts/get_abi.pl rest --dir Documentation/ABI/obsolete
612
613=back
614
615=head1 BUGS
616
617Report bugs to Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
618
619=head1 COPYRIGHT
620
621Copyright (c) 2016-2019 by Mauro Carvalho Chehab <mchehab+samsung@kernel.org>.
622
623License GPLv2: GNU GPL version 2 <http://gnu.org/licenses/gpl.html>.
624
625This is free software: you are free to change and redistribute it.
626There is NO WARRANTY, to the extent permitted by law.
627
628=cut
629