… | |
… | |
64 | |
64 | |
65 | package CBOR::XS; |
65 | package CBOR::XS; |
66 | |
66 | |
67 | use common::sense; |
67 | use common::sense; |
68 | |
68 | |
69 | our $VERSION = 1.51; |
69 | our $VERSION = 1.71; |
70 | our @ISA = qw(Exporter); |
70 | our @ISA = qw(Exporter); |
71 | |
71 | |
72 | our @EXPORT = qw(encode_cbor decode_cbor); |
72 | our @EXPORT = qw(encode_cbor decode_cbor); |
73 | |
73 | |
74 | use Exporter; |
74 | use Exporter; |
… | |
… | |
112 | |
112 | |
113 | The mutators for flags all return the CBOR object again and thus calls can |
113 | The mutators for flags all return the CBOR object again and thus calls can |
114 | be chained: |
114 | be chained: |
115 | |
115 | |
116 | my $cbor = CBOR::XS->new->encode ({a => [1,2]}); |
116 | my $cbor = CBOR::XS->new->encode ({a => [1,2]}); |
|
|
117 | |
|
|
118 | =item $cbor = new_safe CBOR::XS |
|
|
119 | |
|
|
120 | Create a new, safe/secure CBOR::XS object. This is similar to C<new>, |
|
|
121 | but configures the coder object to be safe to use with untrusted |
|
|
122 | data. Currently, this is equivalent to: |
|
|
123 | |
|
|
124 | my $cbor = CBOR::XS |
|
|
125 | ->new |
|
|
126 | ->forbid_objects |
|
|
127 | ->filter (\&CBOR::XS::safe_filter) |
|
|
128 | ->max_size (1e8); |
|
|
129 | |
|
|
130 | But is more future proof (it is better to crash because of a change than |
|
|
131 | to be exploited in other ways). |
|
|
132 | |
|
|
133 | =cut |
|
|
134 | |
|
|
135 | sub new_safe { |
|
|
136 | CBOR::XS |
|
|
137 | ->new |
|
|
138 | ->forbid_objects |
|
|
139 | ->filter (\&CBOR::XS::safe_filter) |
|
|
140 | ->max_size (1e8) |
|
|
141 | } |
117 | |
142 | |
118 | =item $cbor = $cbor->max_depth ([$maximum_nesting_depth]) |
143 | =item $cbor = $cbor->max_depth ([$maximum_nesting_depth]) |
119 | |
144 | |
120 | =item $max_depth = $cbor->get_max_depth |
145 | =item $max_depth = $cbor->get_max_depth |
121 | |
146 | |
… | |
… | |
137 | |
162 | |
138 | Note that nesting is implemented by recursion in C. The default value has |
163 | Note that nesting is implemented by recursion in C. The default value has |
139 | been chosen to be as large as typical operating systems allow without |
164 | been chosen to be as large as typical operating systems allow without |
140 | crashing. |
165 | crashing. |
141 | |
166 | |
142 | See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
167 | See L<SECURITY CONSIDERATIONS>, below, for more info on why this is useful. |
143 | |
168 | |
144 | =item $cbor = $cbor->max_size ([$maximum_string_size]) |
169 | =item $cbor = $cbor->max_size ([$maximum_string_size]) |
145 | |
170 | |
146 | =item $max_size = $cbor->get_max_size |
171 | =item $max_size = $cbor->get_max_size |
147 | |
172 | |
… | |
… | |
152 | effect on C<encode> (yet). |
177 | effect on C<encode> (yet). |
153 | |
178 | |
154 | If no argument is given, the limit check will be deactivated (same as when |
179 | If no argument is given, the limit check will be deactivated (same as when |
155 | C<0> is specified). |
180 | C<0> is specified). |
156 | |
181 | |
157 | See SECURITY CONSIDERATIONS, below, for more info on why this is useful. |
182 | See L<SECURITY CONSIDERATIONS>, below, for more info on why this is useful. |
158 | |
183 | |
159 | =item $cbor = $cbor->allow_unknown ([$enable]) |
184 | =item $cbor = $cbor->allow_unknown ([$enable]) |
160 | |
185 | |
161 | =item $enabled = $cbor->get_allow_unknown |
186 | =item $enabled = $cbor->get_allow_unknown |
162 | |
187 | |
… | |
… | |
226 | throwing an error. |
251 | throwing an error. |
227 | |
252 | |
228 | This option does not affect C<encode> in any way - shared values and |
253 | This option does not affect C<encode> in any way - shared values and |
229 | references will always be encoded properly if present. |
254 | references will always be encoded properly if present. |
230 | |
255 | |
|
|
256 | =item $cbor = $cbor->forbid_objects ([$enable]) |
|
|
257 | |
|
|
258 | =item $enabled = $cbor->get_forbid_objects |
|
|
259 | |
|
|
260 | Disables the use of the object serialiser protocol. |
|
|
261 | |
|
|
262 | If C<$enable> is true (or missing), then C<encode> will will throw an |
|
|
263 | exception when it encounters perl objects that would be encoded using the |
|
|
264 | perl-object tag (26). When C<decode> encounters such tags, it will fall |
|
|
265 | back to the general filter/tagged logic as if this were an unknown tag (by |
|
|
266 | default resulting in a C<CBOR::XC::Tagged> object). |
|
|
267 | |
|
|
268 | If C<$enable> is false (the default), then C<encode> will use the |
|
|
269 | L<Types::Serialiser> object serialisation protocol to serialise objects |
|
|
270 | into perl-object tags, and C<decode> will do the same to decode such tags. |
|
|
271 | |
|
|
272 | See L<SECURITY CONSIDERATIONS>, below, for more info on why forbidding this |
|
|
273 | protocol can be useful. |
|
|
274 | |
231 | =item $cbor = $cbor->pack_strings ([$enable]) |
275 | =item $cbor = $cbor->pack_strings ([$enable]) |
232 | |
276 | |
233 | =item $enabled = $cbor->get_pack_strings |
277 | =item $enabled = $cbor->get_pack_strings |
234 | |
278 | |
235 | If C<$enable> is true (or missing), then C<encode> will try not to encode |
279 | If C<$enable> is true (or missing), then C<encode> will try not to encode |
… | |
… | |
337 | replace the tagged value in the decoded data structure, or no values, |
381 | replace the tagged value in the decoded data structure, or no values, |
338 | which will result in default handling, which currently means the decoder |
382 | which will result in default handling, which currently means the decoder |
339 | creates a C<CBOR::XS::Tagged> object to hold the tag and the value. |
383 | creates a C<CBOR::XS::Tagged> object to hold the tag and the value. |
340 | |
384 | |
341 | When the filter is cleared (the default state), the default filter |
385 | When the filter is cleared (the default state), the default filter |
342 | function, C<CBOR::XS::default_filter>, is used. This function simply looks |
386 | function, C<CBOR::XS::default_filter>, is used. This function simply |
343 | up the tag in the C<%CBOR::XS::FILTER> hash. If an entry exists it must be |
387 | looks up the tag in the C<%CBOR::XS::FILTER> hash. If an entry exists |
344 | a code reference that is called with tag and value, and is responsible for |
388 | it must be a code reference that is called with tag and value, and is |
345 | decoding the value. If no entry exists, it returns no values. |
389 | responsible for decoding the value. If no entry exists, it returns no |
|
|
390 | values. C<CBOR::XS> provides a number of default filter functions already, |
|
|
391 | the the C<%CBOR::XS::FILTER> hash can be freely extended with more. |
|
|
392 | |
|
|
393 | C<CBOR::XS> additionally provides an alternative filter function that is |
|
|
394 | supposed to be safe to use with untrusted data (which the default filter |
|
|
395 | might not), called C<CBOR::XS::safe_filter>, which works the same as |
|
|
396 | the C<default_filter> but uses the C<%CBOR::XS::SAFE_FILTER> variable |
|
|
397 | instead. It is prepopulated with the tag decoding functions that are |
|
|
398 | deemed safe (basically the same as C<%CBOR::XS::FILTER> without all |
|
|
399 | the bignum tags), and can be extended by user code as wlel, although, |
|
|
400 | obviously, one should be very careful about adding decoding functions |
|
|
401 | here, since the expectation is that they are safe to use on untrusted |
|
|
402 | data, after all. |
346 | |
403 | |
347 | Example: decode all tags not handled internally into C<CBOR::XS::Tagged> |
404 | Example: decode all tags not handled internally into C<CBOR::XS::Tagged> |
348 | objects, with no other special handling (useful when working with |
405 | objects, with no other special handling (useful when working with |
349 | potentially "unsafe" CBOR data). |
406 | potentially "unsafe" CBOR data). |
350 | |
407 | |
… | |
… | |
356 | $CBOR::XS::FILTER{1347375694} = sub { |
413 | $CBOR::XS::FILTER{1347375694} = sub { |
357 | my ($tag, $value); |
414 | my ($tag, $value); |
358 | |
415 | |
359 | "tag 1347375694 value $value" |
416 | "tag 1347375694 value $value" |
360 | }; |
417 | }; |
|
|
418 | |
|
|
419 | Example: provide your own filter function that looks up tags in your own |
|
|
420 | hash: |
|
|
421 | |
|
|
422 | my %my_filter = ( |
|
|
423 | 998347484 => sub { |
|
|
424 | my ($tag, $value); |
|
|
425 | |
|
|
426 | "tag 998347484 value $value" |
|
|
427 | }; |
|
|
428 | ); |
|
|
429 | |
|
|
430 | my $coder = CBOR::XS->new->filter (sub { |
|
|
431 | &{ $my_filter{$_[0]} or return } |
|
|
432 | }); |
|
|
433 | |
|
|
434 | |
|
|
435 | Example: use the safe filter function (see L<SECURITY CONSIDERATIONS> for |
|
|
436 | more considerations on security). |
|
|
437 | |
|
|
438 | CBOR::XS->new->filter (\&CBOR::XS::safe_filter)->decode ($cbor_data); |
361 | |
439 | |
362 | =item $cbor_data = $cbor->encode ($perl_scalar) |
440 | =item $cbor_data = $cbor->encode ($perl_scalar) |
363 | |
441 | |
364 | Converts the given Perl data structure (a scalar value) to its CBOR |
442 | Converts the given Perl data structure (a scalar value) to its CBOR |
365 | representation. |
443 | representation. |
… | |
… | |
978 | CBOR intact. |
1056 | CBOR intact. |
979 | |
1057 | |
980 | |
1058 | |
981 | =head1 SECURITY CONSIDERATIONS |
1059 | =head1 SECURITY CONSIDERATIONS |
982 | |
1060 | |
983 | When you are using CBOR in a protocol, talking to untrusted potentially |
1061 | Tl;dr... if you want to decode or encode CBOR from untrusted sources, you |
984 | hostile creatures requires relatively few measures. |
1062 | should start with a coder object created via C<new_safe>: |
985 | |
1063 | |
|
|
1064 | my $coder = CBOR::XS->new_safe; |
|
|
1065 | |
|
|
1066 | my $data = $coder->decode ($cbor_text); |
|
|
1067 | my $cbor = $coder->encode ($data); |
|
|
1068 | |
|
|
1069 | Longer version: When you are using CBOR in a protocol, talking to |
|
|
1070 | untrusted potentially hostile creatures requires some thought: |
|
|
1071 | |
|
|
1072 | =over 4 |
|
|
1073 | |
|
|
1074 | =item Security of the CBOR decoder itself |
|
|
1075 | |
986 | First of all, your CBOR decoder should be secure, that is, should not have |
1076 | First and foremost, your CBOR decoder should be secure, that is, should |
|
|
1077 | not have any buffer overflows or similar bugs that could potentially be |
987 | any buffer overflows. Obviously, this module should ensure that and I am |
1078 | exploited. Obviously, this module should ensure that and I am trying hard |
988 | trying hard on making that true, but you never know. |
1079 | on making that true, but you never know. |
989 | |
1080 | |
|
|
1081 | =item CBOR::XS can invoke almost arbitrary callbacks during decoding |
|
|
1082 | |
|
|
1083 | CBOR::XS supports object serialisation - decoding CBOR can cause calls |
|
|
1084 | to I<any> C<THAW> method in I<any> package that exists in your process |
|
|
1085 | (that is, CBOR::XS will not try to load modules, but any existing C<THAW> |
|
|
1086 | method or function can be called, so they all have to be secure). |
|
|
1087 | |
|
|
1088 | Less obviously, it will also invoke C<TO_CBOR> and C<FREEZE> methods - |
|
|
1089 | even if all your C<THAW> methods are secure, encoding data structures from |
|
|
1090 | untrusted sources can invoke those and trigger bugs in those. |
|
|
1091 | |
|
|
1092 | So, if you are not sure about the security of all the modules you |
|
|
1093 | have loaded (you shouldn't), you should disable this part using |
|
|
1094 | C<forbid_objects>. |
|
|
1095 | |
|
|
1096 | =item CBOR can be extended with tags that call library code |
|
|
1097 | |
|
|
1098 | CBOR can be extended with tags, and C<CBOR::XS> has a registry of |
|
|
1099 | conversion functions for many existing tags that can be extended via |
|
|
1100 | third-party modules (see the C<filter> method). |
|
|
1101 | |
|
|
1102 | If you don't trust these, you should configure the "safe" filter function, |
|
|
1103 | C<CBOR::XS::safe_filter>, which by default only includes conversion |
|
|
1104 | functions that are considered "safe" by the author (but again, they can be |
|
|
1105 | extended by third party modules). |
|
|
1106 | |
|
|
1107 | Depending on your level of paranoia, you can use the "safe" filter: |
|
|
1108 | |
|
|
1109 | $cbor->filter (\&CBOR::XS::safe_filter); |
|
|
1110 | |
|
|
1111 | ... your own filter... |
|
|
1112 | |
|
|
1113 | $cbor->filter (sub { ... do your stuffs here ... }); |
|
|
1114 | |
|
|
1115 | ... or even no filter at all, disabling all tag decoding: |
|
|
1116 | |
|
|
1117 | $cbor->filter (sub { }); |
|
|
1118 | |
|
|
1119 | This is never a problem for encoding, as the tag mechanism only exists in |
|
|
1120 | CBOR texts. |
|
|
1121 | |
|
|
1122 | =item Resource-starving attacks: object memory usage |
|
|
1123 | |
990 | Second, you need to avoid resource-starving attacks. That means you should |
1124 | You need to avoid resource-starving attacks. That means you should limit |
991 | limit the size of CBOR data you accept, or make sure then when your |
1125 | the size of CBOR data you accept, or make sure then when your resources |
992 | resources run out, that's just fine (e.g. by using a separate process that |
1126 | run out, that's just fine (e.g. by using a separate process that can |
993 | can crash safely). The size of a CBOR string in octets is usually a good |
1127 | crash safely). The size of a CBOR string in octets is usually a good |
994 | indication of the size of the resources required to decode it into a Perl |
1128 | indication of the size of the resources required to decode it into a Perl |
995 | structure. While CBOR::XS can check the size of the CBOR text, it might be |
1129 | structure. While CBOR::XS can check the size of the CBOR text (using |
996 | too late when you already have it in memory, so you might want to check |
1130 | C<max_size>), it might be too late when you already have it in memory, so |
997 | the size before you accept the string. |
1131 | you might want to check the size before you accept the string. |
998 | |
1132 | |
|
|
1133 | As for encoding, it is possible to construct data structures that are |
|
|
1134 | relatively small but result in large CBOR texts (for example by having an |
|
|
1135 | array full of references to the same big data structure, which will all be |
|
|
1136 | deep-cloned during encoding by default). This is rarely an actual issue |
|
|
1137 | (and the worst case is still just running out of memory), but you can |
|
|
1138 | reduce this risk by using C<allow_sharing>. |
|
|
1139 | |
|
|
1140 | =item Resource-starving attacks: stack overflows |
|
|
1141 | |
999 | Third, CBOR::XS recurses using the C stack when decoding objects and |
1142 | CBOR::XS recurses using the C stack when decoding objects and arrays. The |
1000 | arrays. The C stack is a limited resource: for instance, on my amd64 |
1143 | C stack is a limited resource: for instance, on my amd64 machine with 8MB |
1001 | machine with 8MB of stack size I can decode around 180k nested arrays but |
1144 | of stack size I can decode around 180k nested arrays but only 14k nested |
1002 | only 14k nested CBOR objects (due to perl itself recursing deeply on croak |
1145 | CBOR objects (due to perl itself recursing deeply on croak to free the |
1003 | to free the temporary). If that is exceeded, the program crashes. To be |
1146 | temporary). If that is exceeded, the program crashes. To be conservative, |
1004 | conservative, the default nesting limit is set to 512. If your process |
1147 | the default nesting limit is set to 512. If your process has a smaller |
1005 | has a smaller stack, you should adjust this setting accordingly with the |
1148 | stack, you should adjust this setting accordingly with the C<max_depth> |
1006 | C<max_depth> method. |
1149 | method. |
|
|
1150 | |
|
|
1151 | =item Resource-starving attacks: CPU en-/decoding complexity |
|
|
1152 | |
|
|
1153 | CBOR::XS will use the L<Math::BigInt>, L<Math::BigFloat> and |
|
|
1154 | L<Math::BigRat> libraries to represent encode/decode bignums. These can |
|
|
1155 | be very slow (as in, centuries of CPU time) and can even crash your |
|
|
1156 | program (and are generally not very trustworthy). See the next section for |
|
|
1157 | details. |
|
|
1158 | |
|
|
1159 | =item Data breaches: leaking information in error messages |
|
|
1160 | |
|
|
1161 | CBOR::XS might leak contents of your Perl data structures in its error |
|
|
1162 | messages, so when you serialise sensitive information you might want to |
|
|
1163 | make sure that exceptions thrown by CBOR::XS will not end up in front of |
|
|
1164 | untrusted eyes. |
|
|
1165 | |
|
|
1166 | =item Something else... |
1007 | |
1167 | |
1008 | Something else could bomb you, too, that I forgot to think of. In that |
1168 | Something else could bomb you, too, that I forgot to think of. In that |
1009 | case, you get to keep the pieces. I am always open for hints, though... |
1169 | case, you get to keep the pieces. I am always open for hints, though... |
1010 | |
1170 | |
1011 | Also keep in mind that CBOR::XS might leak contents of your Perl data |
1171 | =back |
1012 | structures in its error messages, so when you serialise sensitive |
|
|
1013 | information you might want to make sure that exceptions thrown by CBOR::XS |
|
|
1014 | will not end up in front of untrusted eyes. |
|
|
1015 | |
1172 | |
1016 | |
1173 | |
1017 | =head1 BIGNUM SECURITY CONSIDERATIONS |
1174 | =head1 BIGNUM SECURITY CONSIDERATIONS |
1018 | |
1175 | |
1019 | CBOR::XS provides a C<TO_CBOR> method for both L<Math::BigInt> and |
1176 | CBOR::XS provides a C<TO_CBOR> method for both L<Math::BigInt> and |
… | |
… | |
1093 | |
1250 | |
1094 | Please refrain from using rt.cpan.org or any other bug reporting |
1251 | Please refrain from using rt.cpan.org or any other bug reporting |
1095 | service. I put the contact address into my modules for a reason. |
1252 | service. I put the contact address into my modules for a reason. |
1096 | |
1253 | |
1097 | =cut |
1254 | =cut |
|
|
1255 | |
|
|
1256 | # clumsy and slow hv_store-in-hash helper function |
|
|
1257 | sub _hv_store { |
|
|
1258 | $_[0]{$_[1]} = $_[2]; |
|
|
1259 | } |
1098 | |
1260 | |
1099 | our %FILTER = ( |
1261 | our %FILTER = ( |
1100 | 0 => sub { # rfc4287 datetime, utf-8 |
1262 | 0 => sub { # rfc4287 datetime, utf-8 |
1101 | require Time::Piece; |
1263 | require Time::Piece; |
1102 | # Time::Piece::Strptime uses the "incredibly flexible date parsing routine" |
1264 | # Time::Piece::Strptime uses the "incredibly flexible date parsing routine" |
… | |
… | |
1175 | # 34 # base64 rfc46484, utf-8 |
1337 | # 34 # base64 rfc46484, utf-8 |
1176 | # 35 # regex pcre/ecma262, utf-8 |
1338 | # 35 # regex pcre/ecma262, utf-8 |
1177 | # 36 # mime message rfc2045, utf-8 |
1339 | # 36 # mime message rfc2045, utf-8 |
1178 | ); |
1340 | ); |
1179 | |
1341 | |
1180 | sub CBOR::XS::default_filter { |
1342 | sub default_filter { |
1181 | &{ $FILTER{$_[0]} or return } |
1343 | &{ $FILTER{$_[0]} or return } |
|
|
1344 | } |
|
|
1345 | |
|
|
1346 | our %SAFE_FILTER = map { $_ => $FILTER{$_} } 0, 1, 21, 22, 23, 32; |
|
|
1347 | |
|
|
1348 | sub safe_filter { |
|
|
1349 | &{ $SAFE_FILTER{$_[0]} or return } |
1182 | } |
1350 | } |
1183 | |
1351 | |
1184 | sub URI::TO_CBOR { |
1352 | sub URI::TO_CBOR { |
1185 | my $uri = $_[0]->as_string; |
1353 | my $uri = $_[0]->as_string; |
1186 | utf8::upgrade $uri; |
1354 | utf8::upgrade $uri; |