1 | #!/usr/bin/perl
|
---|
2 | # Import a BIND zone file
|
---|
3 | ##
|
---|
4 | # Copyright 2020 Kris Deugau <kdeugau@deepnet.cx>
|
---|
5 | #
|
---|
6 | # This program is free software: you can redistribute it and/or modify
|
---|
7 | # it under the terms of the GNU General Public License as published by
|
---|
8 | # the Free Software Foundation, either version 3 of the License, or
|
---|
9 | # (at your option) any later version.
|
---|
10 | #
|
---|
11 | # This program is distributed in the hope that it will be useful,
|
---|
12 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
13 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
14 | # GNU General Public License for more details.
|
---|
15 | #
|
---|
16 | # You should have received a copy of the GNU General Public License
|
---|
17 | # along with this program. If not, see <http://www.gnu.org/licenses/>.
|
---|
18 | ##
|
---|
19 |
|
---|
20 | use strict;
|
---|
21 | use warnings;
|
---|
22 | use Data::Dumper;
|
---|
23 |
|
---|
24 | use lib '.';
|
---|
25 | use DNSDB;
|
---|
26 |
|
---|
27 | my $dnsdb = new DNSDB;
|
---|
28 | my $doimport = 0;
|
---|
29 |
|
---|
30 | #print Dumper(\%reverse_typemap);
|
---|
31 |
|
---|
32 | my $zname = shift @ARGV;
|
---|
33 | die "usage: bind-import zonename\n" if !$zname;
|
---|
34 | my $rev = 'n';
|
---|
35 | my $zid;
|
---|
36 |
|
---|
37 | my %amap;
|
---|
38 | my %namemap;
|
---|
39 | my %cmap;
|
---|
40 |
|
---|
41 | ##fixme: this is wrong, BIND zone files are generally complete and we're adding. merging records is an entire fridge full of worms.
|
---|
42 | if ($zname =~ /\.arpa\.?$/ || $zname =~ m,^[\d./]+$,) {
|
---|
43 | $rev = 'y';
|
---|
44 | $zname = _zone2cidr($zname) if $zname =~ /\.arpa\.?$/;
|
---|
45 | $zid = $dnsdb->revID($zname,':ANY:');
|
---|
46 | if ($zid) {
|
---|
47 | $zname = new NetAddr::IP $zname;
|
---|
48 | $zname = DNSDB::_ZONE($zname, 'ZONE', 'r', '.').($zname->{isv6} ? '.ip6.arpa' : '.in-addr.arpa');
|
---|
49 | }
|
---|
50 | } else {
|
---|
51 | $zid = $dnsdb->domainID($zname,':ANY:');
|
---|
52 | }
|
---|
53 |
|
---|
54 | die "zone $zname not on file\n" if !$zid;
|
---|
55 |
|
---|
56 | # still no sane way to expose a human-friendly view tag on the command line.
|
---|
57 | my $view = shift @ARGV;
|
---|
58 | $view = '' if !$view;
|
---|
59 |
|
---|
60 | ##fixme: retrieve defttl from SOA record
|
---|
61 | my $zonettl = 900;
|
---|
62 | my $defttl = $zonettl;
|
---|
63 | my $origin = $zname; # to append to unqualified names
|
---|
64 |
|
---|
65 | # need to spin up a full state machine-ish thing, because BIND zone files are all about context
|
---|
66 | # see ch4, p56-72 in the grasshopper book
|
---|
67 | my $prevlabel = '';
|
---|
68 | my $curlabel = '';
|
---|
69 |
|
---|
70 | my $i = 0;
|
---|
71 |
|
---|
72 | while (<>) {
|
---|
73 | chomp;
|
---|
74 | next if /^\s*$/;
|
---|
75 | next if /^\s*;/; # comments
|
---|
76 | next if /^\s*\)/; # SOA closing. arguably should do some more targeted voodoo when parsing the SOA details
|
---|
77 |
|
---|
78 | print "($_)\n";
|
---|
79 | if (my ($macro,$mdetail) = (/^\s*\$(TTL|ORIGIN|INCLUDE|GENERATE)\s+(.+)/) ) {
|
---|
80 | # macro sort of thing; $TTL and $ORIGIN most common. $INCLUDE is a thing, expect it to be rare in live use tho
|
---|
81 | if ($macro eq 'TTL') {
|
---|
82 | $mdetail =~ s/\s*;.+$//;
|
---|
83 | if ($mdetail =~ /^\d+$/) {
|
---|
84 | $defttl = $mdetail;
|
---|
85 | } else {
|
---|
86 | warn "invalid \$TTL: $_\n";
|
---|
87 | }
|
---|
88 | } elsif ($macro eq 'ORIGIN') {
|
---|
89 | ##fixme: going to skip the stupid case of "$ORIGIN com." and the like that lie
|
---|
90 | # between . and the root domain we were told we're importing; anyone using such
|
---|
91 | # a mess outside the root servers is clearly insane
|
---|
92 |
|
---|
93 | # $ORIGIN supports cascading/nesting, by watching for fully-qualified names vs partial names.
|
---|
94 |
|
---|
95 | print "origin ($mdetail)\n";
|
---|
96 | if ($mdetail =~ /\.$/) {
|
---|
97 | $origin = $mdetail;
|
---|
98 | } else {
|
---|
99 | # append current origin to unqualified origin
|
---|
100 | $origin = "$mdetail.$origin";
|
---|
101 | }
|
---|
102 |
|
---|
103 | # if ($mdetail eq '.' || $mdetail =~ /$zname\.$/ || $zname =~ /$mdetail\.$/) {
|
---|
104 | # $origin = $mdetail;
|
---|
105 | # } else {
|
---|
106 | # # if we continue, we either use an $ORIGIN that's out of zone, or ignore it and potentially publish incorrect records.
|
---|
107 | # die "bad \$ORIGIN: $_\n";
|
---|
108 | # }
|
---|
109 |
|
---|
110 | }
|
---|
111 | # not handling $INCLUDE or $GENERATE (altho the latter seems to be mostly a less-flexible version of the template types)
|
---|
112 | next;
|
---|
113 | }
|
---|
114 |
|
---|
115 | my $origrec = $_;
|
---|
116 |
|
---|
117 | # skip stale records that have no value
|
---|
118 | next if /^ip-192-168-1(12|20)-\d+/;
|
---|
119 | next if /ip.add.re.\d+\s*$/;
|
---|
120 |
|
---|
121 | # leading whitespace indicates "same label as last record"
|
---|
122 | if (/^\s/) {
|
---|
123 | $curlabel = $prevlabel;
|
---|
124 | } else {
|
---|
125 | ($curlabel) = /^([\w\@_.-]+)\s/;
|
---|
126 | }
|
---|
127 |
|
---|
128 | # magic name!
|
---|
129 | $curlabel = "$zname." if $curlabel eq '@';
|
---|
130 |
|
---|
131 | # append $ORIGIN if name is not fully qualified.
|
---|
132 | if ($curlabel !~ /\.$/) {
|
---|
133 | $curlabel .= $origin;
|
---|
134 | }
|
---|
135 |
|
---|
136 | # check for zone scope. skip bad records.
|
---|
137 | if ($curlabel !~ /$zname.$/) {
|
---|
138 | warn "bad record $origrec, maybe bad \$ORIGIN?\n";
|
---|
139 | next;
|
---|
140 | }
|
---|
141 |
|
---|
142 | # # records must begin in the first column, no leading whitespace
|
---|
143 | # my ($name) = /^([\w\@_.-]+)\s/;
|
---|
144 |
|
---|
145 | # foo IN A 1.2.3.4
|
---|
146 | # IN A 2.3.4.5
|
---|
147 | # =
|
---|
148 | # foo.zone. IN A 1.2.3.4
|
---|
149 | # foo.zone. IN A 2.3.4.5
|
---|
150 |
|
---|
151 | # # "empty" label records inherit the previous label
|
---|
152 | # # RRs start in the first column by definition, so leading whitespace indicates an inherited label
|
---|
153 | # if (/^\s+/) {
|
---|
154 | # # fatal error. if there is no previous label, we can by definition not set
|
---|
155 | # # the current label based on it. this can only happen on the very first
|
---|
156 | # # record, following records will *ALWAYS* have a previous label
|
---|
157 | # die "bad first record ($_): no previous label\n" if !$prevlabel;
|
---|
158 | # $name = $prevlabel;
|
---|
159 | # }
|
---|
160 |
|
---|
161 | print "$i ($_)\n\t$curlabel";
|
---|
162 |
|
---|
163 |
|
---|
164 |
|
---|
165 |
|
---|
166 | # # append zone name to record name if missing AND not dot-terminated;
|
---|
167 | # # this happens automagically for forward zones, but not reverse because Reasons. (fixme?)
|
---|
168 | # # suck up and deal with the error if the dot-termiated name is out of zone; should be
|
---|
169 | # # impossible with valid BIND zone file but...
|
---|
170 | # if ($name !~ /\.$/) {
|
---|
171 | # $name .= ".$zname" if $name !~ /$zname$/;
|
---|
172 | # } else {
|
---|
173 | # warn "skipping out-of-zone record:\n\t($_)\n" if $name !~ /$zname\.$/;
|
---|
174 | # next;
|
---|
175 | # }
|
---|
176 |
|
---|
177 | last if ++$i > 5;
|
---|
178 |
|
---|
179 |
|
---|
180 | s/^([\w\@_.-]+)\s+//;
|
---|
181 |
|
---|
182 | my $nc = 0;
|
---|
183 | my %seenatoms;
|
---|
184 | my $badrec;
|
---|
185 | my $curatom = 'class';
|
---|
186 | ##fixme: maybe wrap this in an eval() instead of the warn/badrec/last bits?
|
---|
187 | for (my ($atom) = /^([\w\d]+)\s/; $nc < 3; $nc++) {
|
---|
188 | if ($atom =~ /^\d+$/) {
|
---|
189 | if (defined($seenatoms{ttl})) {
|
---|
190 | warn "bad record ($origrec)\n";
|
---|
191 | $badrec = 1;
|
---|
192 | last;
|
---|
193 | } else {
|
---|
194 | if ($curatom ne 'class' && $curatom ne 'ttl') {
|
---|
195 | warn "bad record ($origrec)\n";
|
---|
196 | $badrec = 1;
|
---|
197 | last;
|
---|
198 | }
|
---|
199 | $curatom = 'ttl';
|
---|
200 | $seenatoms{ttl} = $atom;
|
---|
201 | }
|
---|
202 | }
|
---|
203 | if ($atom =~ /^IN|CS|CH|HS$/) {
|
---|
204 | if ($atom =~ /CS|CH|HS/) {
|
---|
205 | warn "unsupported class $atom in record ($origrec)\n";
|
---|
206 | $badrec = 1;
|
---|
207 | last;
|
---|
208 | }
|
---|
209 | $curatom = 'class';
|
---|
210 | }
|
---|
211 | if ($reverse_typemap{$atom}) {
|
---|
212 | print "dbg: type $atom\n";
|
---|
213 | }
|
---|
214 | # my $itype = $reverse_typemap{$type};
|
---|
215 | }
|
---|
216 |
|
---|
217 | next if $badrec;
|
---|
218 |
|
---|
219 |
|
---|
220 | ##fixme: drop curlabel? not sure it's needed
|
---|
221 | #$curlabel = $name;
|
---|
222 | $prevlabel = $curlabel;
|
---|
223 | ##todo: BIND conflates a repeated label with repeating the TTL too. Matter of opinion whether that's really correct or not.
|
---|
224 |
|
---|
225 |
|
---|
226 |
|
---|
227 | # by convention the optional TTL leads the optional class, but they're apparently swappable.
|
---|
228 | my ($ttl) = /^(\d+)?\s/;
|
---|
229 | if (defined $ttl) {
|
---|
230 | # TTL may be zero
|
---|
231 | s/(\d+)?\s+//;
|
---|
232 | } else {
|
---|
233 | # Fall back to zone default TTL
|
---|
234 | $ttl = $zonettl;
|
---|
235 | }
|
---|
236 | my ($class) = /^(IN|CS|CH|HS|\d+)\s/;
|
---|
237 | if (defined $class) {
|
---|
238 | if ($class =~ /\d+/) {
|
---|
239 |
|
---|
240 | }
|
---|
241 | if ($class ne 'IN') {
|
---|
242 | warn "Non-Internet class ($class) records not supported:\n\t$origrec\n";
|
---|
243 | next;
|
---|
244 | }
|
---|
245 | s/(IN|CS|CH|HS)\s+//;
|
---|
246 | } else {
|
---|
247 | $class = 'IN';
|
---|
248 | }
|
---|
249 | my ($type) = /([A-Z-]+)\s/;
|
---|
250 | if (!$reverse_typemap{$type}) {
|
---|
251 | warn "Unknown type $type, skipping\n\t($_)\n";
|
---|
252 | next;
|
---|
253 | }
|
---|
254 | my $itype = $reverse_typemap{$type};
|
---|
255 | s/([A-Z-]+)\s+//;
|
---|
256 | chomp;
|
---|
257 | my $rdata = $_;
|
---|
258 |
|
---|
259 | # SOA is the only type that may span multiple lines. Probably. Note even AXFRed zones write multiline SOA records:
|
---|
260 | #@ IN SOA test.example.invalid. test.example.invalid. (2020082500 7200 900 604800 3600)
|
---|
261 | # IN NS olddns.example.com.
|
---|
262 | # IN MX 1 fred.foo.bar.invalid.
|
---|
263 | #foo IN A 192.168.16.45
|
---|
264 | # AXFR'ed zone file gets written as
|
---|
265 | #$ORIGIN .
|
---|
266 | #$TTL 3600 ; 1 hour
|
---|
267 | #example.invalid IN SOA test.example.invalid. test.example.invalid. (
|
---|
268 | # 2020082500 ; serial
|
---|
269 | # 7200 ; refresh (2 hours)
|
---|
270 | # 900 ; retry (15 minutes)
|
---|
271 | # 604800 ; expire (1 week)
|
---|
272 | # 3600 ; minimum (1 hour)
|
---|
273 | # )
|
---|
274 | # NS olddns.example.com.
|
---|
275 | # MX 1 fred.foo.bar.invalid.
|
---|
276 | #$ORIGIN example.invalid.
|
---|
277 | #foo A 192.168.16.45
|
---|
278 |
|
---|
279 | if ($type eq 'SOA') {
|
---|
280 | my ($ns, $adminmail) = ($rdata =~ /([\w.]+)\s+([\w.]+)\s+\(/);
|
---|
281 | die "Can't parse gibberish SOAish record: $_\n" if !$ns;
|
---|
282 | $rdata =~ s/([\w.]+)\s+([\w.]+)\s+\(\s*//;
|
---|
283 |
|
---|
284 | # There are probably more efficient ways to do this but the SOA record
|
---|
285 | # format is essentially character based, not line-based.
|
---|
286 | # In theory the SOA serial etc may be spread over up to 5 lines, in any combination.
|
---|
287 |
|
---|
288 | # Parse fields from $rdata if present
|
---|
289 | my @soabits;
|
---|
290 | my @soafirst = split /\s+/, $rdata;
|
---|
291 | while (my $f = shift @soafirst) {
|
---|
292 | last if $f !~ /^\d/;
|
---|
293 | push @soabits, $f;
|
---|
294 | }
|
---|
295 |
|
---|
296 | # Read more lines if we don't have enough SOA fields filled
|
---|
297 | while (scalar(@soabits) < 5) {
|
---|
298 | my $tmp = <>;
|
---|
299 | $tmp =~ s/^\s*//;
|
---|
300 | my @tmpsoa = split /\s+/, $tmp;
|
---|
301 | while (my $f = shift @tmpsoa) {
|
---|
302 | last if $f !~ /^\d/;
|
---|
303 | push @soabits, $f;
|
---|
304 | }
|
---|
305 | if (scalar(@soabits) == 5) {
|
---|
306 | last;
|
---|
307 | }
|
---|
308 | }
|
---|
309 | } # SOA
|
---|
310 |
|
---|
311 | # Quotes may arguably be syntactically required, but they're not actually part of the record data
|
---|
312 | elsif ($type eq 'TXT') {
|
---|
313 | $rdata =~ s/^"//;
|
---|
314 | $rdata =~ s/"$//;
|
---|
315 | }
|
---|
316 |
|
---|
317 | # temp hack for hosts file
|
---|
318 | elsif ($type eq 'A') {
|
---|
319 | # if ($amap{$name}) {
|
---|
320 | # print "urp: dupe name $name $rdata\n";
|
---|
321 | # } else {
|
---|
322 | push @{$amap{$curlabel}}, $rdata;
|
---|
323 | # }
|
---|
324 | push @{$namemap{$rdata}}, $curlabel;
|
---|
325 | }
|
---|
326 | elsif ($type eq 'CNAME') {
|
---|
327 | push @{$cmap{$rdata}}, $curlabel;
|
---|
328 | }
|
---|
329 |
|
---|
330 | no warnings qw(uninitialized);
|
---|
331 | #print "parsed: '$name' '$class' '$ttl' '$type'->'$itype' '$rdata'\n";
|
---|
332 | #print;
|
---|
333 | #;imap IN 900 CNAME deepnet.cx.
|
---|
334 | ##fixme: not sure how to handle the case where someone leaves off the class.
|
---|
335 | if ($doimport) {
|
---|
336 | my ($code, $msg);
|
---|
337 | if ($rev eq 'n') {
|
---|
338 | ($code,$msg) = $dnsdb->addRec('n', $rev, $zid, \$curlabel, \$itype, \$rdata, $ttl);
|
---|
339 | } else {
|
---|
340 | ($code,$msg) = $dnsdb->addRec('n', $rev, $zid, \$rdata, \$itype, \$curlabel, $ttl);
|
---|
341 | }
|
---|
342 | print "$code: $msg\n";
|
---|
343 | }
|
---|
344 | $i++;
|
---|
345 | }
|
---|
346 |
|
---|
347 |
|
---|
348 | #print Dumper \%amap;
|
---|
349 | #print Dumper \%namemap;
|
---|
350 | #print Dumper \%cmap;
|
---|
351 |
|
---|
352 | foreach my $n (keys %amap) {
|
---|
353 | foreach my $ip (@{$amap{$n}}) {
|
---|
354 | #print "$ip $n\n";
|
---|
355 | push @{$namemap{$ip}}, $n unless grep $n, @{$namemap{$ip}};
|
---|
356 | }
|
---|
357 | }
|
---|
358 |
|
---|
359 | foreach my $c (keys %cmap) {
|
---|
360 | if ($amap{$c}) {
|
---|
361 | print Dumper(\@{$amap{$c}});
|
---|
362 | }
|
---|
363 | # print $amap{$c};
|
---|
364 | }
|
---|
365 |
|
---|
366 | # cname targ -> IP
|
---|
367 |
|
---|
368 | #foreach my $ip (sort keys %namemap) {
|
---|
369 | # print "$ip ".join(' ', @{$namemap{$ip}})."\n";
|
---|
370 | #}
|
---|
371 |
|
---|