Commit | Line | Data |
6badd1a5 |
1 | package Opcode; |
2 | |
3 | require 5.002; |
4 | |
5 | use vars qw($VERSION @ISA @EXPORT_OK); |
6 | |
7 | $VERSION = "1.01"; |
8 | |
9 | use strict; |
10 | use Carp; |
11 | use Exporter (); |
12 | use DynaLoader (); |
13 | @ISA = qw(Exporter DynaLoader); |
14 | |
15 | BEGIN { |
16 | @EXPORT_OK = qw( |
17 | opset ops_to_opset |
18 | opset_to_ops opset_to_hex invert_opset |
19 | empty_opset full_opset |
20 | opdesc opcodes opmask define_optag |
21 | opmask_add verify_opset opdump |
22 | ); |
23 | } |
24 | |
25 | use subs @EXPORT_OK; |
26 | |
27 | bootstrap Opcode $VERSION; |
28 | |
29 | _init_optags(); |
30 | |
31 | |
32 | *ops_to_opset = \&opset; # alias for old name |
33 | |
34 | |
35 | sub opset_to_hex ($) { |
36 | return "(invalid opset)" unless verify_opset($_[0]); |
37 | unpack("h*",$_[0]); |
38 | } |
39 | |
40 | sub opdump (;$) { |
41 | my $pat = shift; |
42 | # handy utility: perl -MOpcode=opdump -e 'opdump File' |
43 | foreach(opset_to_ops(full_opset)) { |
44 | my $op = sprintf " %12s %s\n", $_, opdesc($_); |
45 | next if defined $pat and $op !~ m/$pat/i; |
46 | print $op; |
47 | } |
48 | } |
49 | |
50 | |
51 | |
52 | sub _init_optags { |
53 | my(%all, %seen); |
54 | @all{opset_to_ops(full_opset)} = (); # keys only |
55 | |
7a57407b |
56 | local($_); |
6badd1a5 |
57 | local($/) = "\n=cut"; # skip to optags definition section |
58 | <DATA>; |
59 | $/ = "\n="; # now read in 'pod section' chunks |
60 | while(<DATA>) { |
61 | next unless m/^item\s+(:\w+)/; |
62 | my $tag = $1; |
63 | |
64 | # Split into lines, keep only indented lines |
65 | my @lines = grep { m/^\s/ } split(/\n/); |
66 | foreach (@lines) { s/--.*// } # delete comments |
67 | my @ops = map { split ' ' } @lines; # get op words |
68 | |
69 | foreach(@ops) { |
70 | warn "$tag - $_ already tagged in $seen{$_}\n" if $seen{$_}; |
71 | $seen{$_} = $tag; |
72 | delete $all{$_}; |
73 | } |
74 | # opset will croak on invalid names |
75 | define_optag($tag, opset(@ops)); |
76 | } |
77 | close(DATA); |
78 | warn "Untagged opnames: ".join(' ',keys %all)."\n" if %all; |
79 | } |
80 | |
81 | |
82 | 1; |
83 | |
84 | __DATA__ |
85 | |
86 | =head1 NAME |
87 | |
88 | Opcode - Disable named opcodes when compiling perl code |
89 | |
90 | =head1 SYNOPSIS |
91 | |
92 | use Opcode; |
93 | |
94 | |
95 | =head1 DESCRIPTION |
96 | |
97 | Perl code is always compiled into an internal format before execution. |
98 | |
99 | Evaluating perl code (e.g. via "eval" or "do 'file'") causes |
100 | the code to be compiled into an internal format and then, |
101 | provided there was no error in the compilation, executed. |
102 | The internal format is based on many distinct I<opcodes>. |
103 | |
104 | By default no opmask is in effect and any code can be compiled. |
105 | |
106 | The Opcode module allow you to define an I<operator mask> to be in |
107 | effect when perl I<next> compiles any code. Attempting to compile code |
108 | which contains a masked opcode will cause the compilation to fail |
109 | with an error. The code will not be executed. |
110 | |
111 | =head1 NOTE |
112 | |
113 | The Opcode module is not usually used directly. See the ops pragma and |
114 | Safe modules for more typical uses. |
115 | |
116 | =head1 WARNING |
117 | |
118 | The authors make B<no warranty>, implied or otherwise, about the |
119 | suitability of this software for safety or security purposes. |
120 | |
121 | The authors shall not in any case be liable for special, incidental, |
122 | consequential, indirect or other similar damages arising from the use |
123 | of this software. |
124 | |
125 | Your mileage will vary. If in any doubt B<do not use it>. |
126 | |
127 | |
128 | =head1 Operator Names and Operator Lists |
129 | |
130 | The canonical list of operator names is the contents of the array |
131 | op_name defined and initialised in file F<opcode.h> of the Perl |
132 | source distribution (and installed into the perl library). |
133 | |
134 | Each operator has both a terse name (its opname) and a more verbose or |
135 | recognisable descriptive name. The opdesc function can be used to |
136 | return a list of descriptions for a list of operators. |
137 | |
138 | Many of the functions and methods listed below take a list of |
139 | operators as parameters. Most operator lists can be made up of several |
140 | types of element. Each element can be one of |
141 | |
142 | =over 8 |
143 | |
144 | =item an operator name (opname) |
145 | |
146 | Operator names are typically small lowercase words like enterloop, |
147 | leaveloop, last, next, redo etc. Sometimes they are rather cryptic |
148 | like gv2cv, i_ncmp and ftsvtx. |
149 | |
150 | =item an operator tag name (optag) |
151 | |
152 | Operator tags can be used to refer to groups (or sets) of operators. |
153 | Tag names always being with a colon. The Opcode module defines several |
154 | optags and the user can define others using the define_optag function. |
155 | |
156 | =item a negated opname or optag |
157 | |
158 | An opname or optag can be prefixed with an exclamation mark, e.g., !mkdir. |
159 | Negating an opname or optag means remove the corresponding ops from the |
160 | accumulated set of ops at that point. |
161 | |
162 | =item an operator set (opset) |
163 | |
164 | An I<opset> as a binary string of approximately 43 bytes which holds a |
165 | set or zero or more operators. |
166 | |
167 | The opset and opset_to_ops functions can be used to convert from |
168 | a list of operators to an opset and I<vice versa>. |
169 | |
170 | Wherever a list of operators can be given you can use one or more opsets. |
171 | See also Manipulating Opsets below. |
172 | |
173 | =back |
174 | |
175 | |
176 | =head1 Opcode Functions |
177 | |
178 | The Opcode package contains functions for manipulating operator names |
179 | tags and sets. All are available for export by the package. |
180 | |
181 | =over 8 |
182 | |
183 | =item opcodes |
184 | |
185 | In a scalar context opcodes returns the number of opcodes in this |
186 | version of perl (around 340 for perl5.002). |
187 | |
188 | In a list context it returns a list of all the operator names. |
189 | (Not yet implemented, use @names = opset_to_ops(full_opset).) |
190 | |
191 | =item opset (OP, ...) |
192 | |
193 | Returns an opset containing the listed operators. |
194 | |
195 | =item opset_to_ops (OPSET) |
196 | |
197 | Returns a list of operator names corresponding to those operators in |
198 | the set. |
199 | |
200 | =item opset_to_hex (OPSET) |
201 | |
202 | Returns a string representation of an opset. Can be handy for debugging. |
203 | |
204 | =item full_opset |
205 | |
206 | Returns an opset which includes all operators. |
207 | |
208 | =item empty_opset |
209 | |
210 | Returns an opset which contains no operators. |
211 | |
212 | =item invert_opset (OPSET) |
213 | |
214 | Returns an opset which is the inverse set of the one supplied. |
215 | |
216 | =item verify_opset (OPSET, ...) |
217 | |
218 | Returns true if the supplied opset looks like a valid opset (is the |
219 | right length etc) otherwise it returns false. If an optional second |
220 | parameter is true then verify_opset will croak on an invalid opset |
221 | instead of returning false. |
222 | |
223 | Most of the other Opcode functions call verify_opset automatically |
224 | and will croak if given an invalid opset. |
225 | |
226 | =item define_optag (OPTAG, OPSET) |
227 | |
228 | Define OPTAG as a symbolic name for OPSET. Optag names always start |
229 | with a colon C<:>. |
230 | |
231 | The optag name used must not be defined already (define_optag will |
232 | croak if it is already defined). Optag names are global to the perl |
233 | process and optag definitions cannot be altered or deleted once |
234 | defined. |
235 | |
236 | It is strongly recommended that applications using Opcode should use a |
237 | leading capital letter on their tag names since lowercase names are |
238 | reserved for use by the Opcode module. If using Opcode within a module |
239 | you should prefix your tags names with the name of your module to |
240 | ensure uniqueness and thus avoid clashes with other modules. |
241 | |
242 | =item opmask_add (OPSET) |
243 | |
244 | Adds the supplied opset to the current opmask. Note that there is |
245 | currently I<no> mechanism for unmasking ops once they have been masked. |
246 | This is intentional. |
247 | |
248 | =item opmask |
249 | |
250 | Returns an opset corresponding to the current opmask. |
251 | |
252 | =item opdesc (OP, ...) |
253 | |
254 | This takes a list of operator names and returns the corresponding list |
255 | of operator descriptions. |
256 | |
257 | =item opdump (PAT) |
258 | |
259 | Dumps to STDOUT a two column list of op names and op descriptions. |
260 | If an optional pattern is given then only lines which match the |
261 | (case insensitive) pattern will be output. |
262 | |
263 | It's designed to be used as a handy command line utility: |
264 | |
265 | perl -MOpcode=opdump -e opdump |
266 | perl -MOpcode=opdump -e 'opdump Eval' |
267 | |
268 | =back |
269 | |
270 | =head1 Manipulating Opsets |
271 | |
272 | Opsets may be manipulated using the perl bit vector operators & (and), | (or), |
273 | ^ (xor) and ~ (negate/invert). |
274 | |
275 | However you should never rely on the numerical position of any opcode |
276 | within the opset. In other words both sides of a bit vector operator |
277 | should be opsets returned from Opcode functions. |
278 | |
279 | Also, since the number of opcodes in your current version of perl might |
280 | not be an exact multiple of eight, there may be unused bits in the last |
281 | byte of an upset. This should not cause any problems (Opcode functions |
282 | ignore those extra bits) but it does mean that using the ~ operator |
283 | will typically not produce the same 'physical' opset 'string' as the |
284 | invert_opset function. |
285 | |
286 | |
287 | =head1 TO DO (maybe) |
288 | |
289 | $bool = opset_eq($opset1, $opset2) true if opsets are logically eqiv |
290 | |
291 | $yes = opset_can($opset, @ops) true if $opset has all @ops set |
292 | |
293 | @diff = opset_diff($opset1, $opset2) => ('foo', '!bar', ...) |
294 | |
295 | =cut |
296 | |
297 | # the =cut above is used by _init_optags() to get here quickly |
298 | |
299 | =head1 Predefined Opcode Tags |
300 | |
301 | =over 5 |
302 | |
303 | =item :base_core |
304 | |
305 | null stub scalar pushmark wantarray const defined undef |
306 | |
307 | rv2sv sassign |
308 | |
309 | rv2av aassign aelem aelemfast aslice av2arylen |
310 | |
311 | rv2hv helem hslice each values keys exists delete |
312 | |
313 | preinc i_preinc predec i_predec postinc i_postinc postdec i_postdec |
314 | int hex oct abs pow multiply i_multiply divide i_divide |
315 | modulo i_modulo add i_add subtract i_subtract |
316 | |
317 | left_shift right_shift bit_and bit_xor bit_or negate i_negate |
318 | not complement |
319 | |
320 | lt i_lt gt i_gt le i_le ge i_ge eq i_eq ne i_ne ncmp i_ncmp |
321 | slt sgt sle sge seq sne scmp |
322 | |
323 | substr vec stringify study pos length index rindex ord chr |
324 | |
325 | ucfirst lcfirst uc lc quotemeta trans chop schop chomp schomp |
326 | |
327 | match split |
328 | |
329 | list lslice splice push pop shift unshift reverse |
330 | |
331 | cond_expr flip flop andassign orassign and or xor |
332 | |
333 | warn die lineseq nextstate unstack scope enter leave |
334 | |
335 | rv2cv anoncode prototype |
336 | |
337 | entersub leavesub return method -- XXX loops via recursion? |
338 | |
339 | leaveeval -- needed for Safe to operate, is safe without entereval |
340 | |
341 | =item :base_mem |
342 | |
343 | These memory related ops are not included in :base_core because they |
344 | can easily be used to implement a resource attack (e.g., consume all |
345 | available memory). |
346 | |
347 | concat repeat join range |
348 | |
349 | anonlist anonhash |
350 | |
351 | Note that despite the existance of this optag a memory resource attack |
352 | may still be possible using only :base_core ops. |
353 | |
354 | Disabling these ops is a I<very> heavy handed way to attempt to prevent |
355 | a memory resource attack. It's probable that a specific memory limit |
356 | mechanism will be added to perl in the near future. |
357 | |
358 | =item :base_loop |
359 | |
360 | These loop ops are not included in :base_core because they can easily be |
361 | used to implement a resource attack (e.g., consume all available CPU time). |
362 | |
363 | grepstart grepwhile |
364 | mapstart mapwhile |
365 | enteriter iter |
366 | enterloop leaveloop |
367 | last next redo |
368 | goto |
369 | |
370 | =item :base_io |
371 | |
372 | These ops enable I<filehandle> (rather than filename) based input and |
373 | output. These are safe on the assumption that only pre-existing |
374 | filehandles are available for use. To create new filehandles other ops |
375 | such as open would need to be enabled. |
376 | |
377 | readline rcatline getc read |
378 | |
379 | formline enterwrite leavewrite |
380 | |
381 | print sysread syswrite send recv eof tell seek |
382 | |
383 | readdir telldir seekdir rewinddir |
384 | |
385 | =item :base_orig |
386 | |
387 | These are a hotchpotch of opcodes still waiting to be considered |
388 | |
389 | gvsv gv gelem |
390 | |
391 | padsv padav padhv padany |
392 | |
393 | rv2gv refgen srefgen ref |
394 | |
395 | bless -- could be used to change ownership of objects (reblessing) |
396 | |
6badd1a5 |
397 | pushre regcmaybe regcomp subst substcont |
398 | |
399 | sprintf prtf -- can core dump |
400 | |
401 | crypt |
402 | |
403 | tie untie |
404 | |
405 | dbmopen dbmclose |
406 | sselect select |
407 | pipe_op sockpair |
408 | |
409 | getppid getpgrp setpgrp getpriority setpriority localtime gmtime |
410 | |
411 | entertry leavetry -- can be used to 'hide' fatal errors |
412 | |
413 | =item :base_math |
414 | |
415 | These ops are not included in :base_core because of the risk of them being |
416 | used to generate floating point exceptions (which would have to be caught |
417 | using a $SIG{FPE} handler). |
418 | |
419 | atan2 sin cos exp log sqrt |
420 | |
421 | These ops are not included in :base_core because they have an effect |
422 | beyond the scope of the compartment. |
423 | |
424 | rand srand |
425 | |
426 | =item :default |
427 | |
428 | A handy tag name for a I<reasonable> default set of ops. (The current ops |
429 | allowed are unstable while development continues. It will change.) |
430 | |
431 | :base_core :base_mem :base_loop :base_io :base_orig |
432 | |
433 | If safety matters to you (and why else would you be using the Opcode module?) |
434 | then you should not rely on the definition of this, or indeed any other, optag! |
435 | |
436 | |
437 | =item :filesys_read |
438 | |
439 | stat lstat readlink |
440 | |
441 | ftatime ftblk ftchr ftctime ftdir fteexec fteowned fteread |
442 | ftewrite ftfile ftis ftlink ftmtime ftpipe ftrexec ftrowned |
443 | ftrread ftsgid ftsize ftsock ftsuid fttty ftzero ftrwrite ftsvtx |
444 | |
445 | fttext ftbinary |
446 | |
447 | fileno |
448 | |
449 | =item :sys_db |
450 | |
451 | ghbyname ghbyaddr ghostent shostent ehostent -- hosts |
452 | gnbyname gnbyaddr gnetent snetent enetent -- networks |
453 | gpbyname gpbynumber gprotoent sprotoent eprotoent -- protocols |
454 | gsbyname gsbyport gservent sservent eservent -- services |
455 | |
456 | gpwnam gpwuid gpwent spwent epwent getlogin -- users |
457 | ggrnam ggrgid ggrent sgrent egrent -- groups |
458 | |
459 | =item :browse |
460 | |
461 | A handy tag name for a I<reasonable> default set of ops beyond the |
462 | :default optag. Like :default (and indeed all the other optags) its |
463 | current definition is unstable while development continues. It will change. |
464 | |
465 | The :browse tag represents the next step beyond :default. It it a |
466 | superset of the :default ops and adds :filesys_read the :sys_db. |
467 | The intent being that scripts can access more (possibly sensitive) |
468 | information about your system but not be able to change it. |
469 | |
470 | :default :filesys_read :sys_db |
471 | |
472 | =item :filesys_open |
473 | |
474 | sysopen open close |
475 | umask binmode |
476 | |
477 | open_dir closedir -- other dir ops are in :base_io |
478 | |
479 | =item :filesys_write |
480 | |
481 | link unlink rename symlink truncate |
482 | |
483 | mkdir rmdir |
484 | |
485 | utime chmod chown |
486 | |
487 | fcntl -- not strictly filesys related, but possibly as dangerous? |
488 | |
489 | =item :subprocess |
490 | |
491 | backtick system |
492 | |
493 | fork |
494 | |
495 | wait waitpid |
496 | |
f812a825 |
497 | glob -- access to Cshell via <`rm *`> |
498 | |
6badd1a5 |
499 | =item :ownprocess |
500 | |
501 | exec exit kill |
502 | |
503 | time tms -- could be used for timing attacks (paranoid?) |
504 | |
505 | =item :others |
506 | |
507 | This tag holds groups of assorted specialist opcodes that don't warrant |
508 | having optags defined for them. |
509 | |
510 | SystemV Interprocess Communications: |
511 | |
512 | msgctl msgget msgrcv msgsnd |
513 | |
514 | semctl semget semop |
515 | |
516 | shmctl shmget shmread shmwrite |
517 | |
518 | =item :still_to_be_decided |
519 | |
520 | chdir |
521 | flock ioctl |
522 | |
523 | socket getpeername ssockopt |
524 | bind connect listen accept shutdown gsockopt getsockname |
525 | |
526 | sleep alarm -- changes global timer state and signal handling |
527 | sort -- assorted problems including core dumps |
528 | tied -- can be used to access object implementing a tie |
529 | pack unpack -- can be used to create/use memory pointers |
530 | |
531 | entereval -- can be used to hide code from initial compile |
532 | require dofile |
533 | |
534 | caller -- get info about calling environment and args |
535 | |
536 | reset |
537 | |
538 | dbstate -- perl -d version of nextstate(ment) opcode |
539 | |
540 | =item :dangerous |
541 | |
542 | This tag is simply a bucket for opcodes that are unlikely to be used via |
543 | a tag name but need to be tagged for completness and documentation. |
544 | |
545 | syscall dump chroot |
546 | |
547 | |
548 | =back |
549 | |
550 | =head1 SEE ALSO |
551 | |
552 | ops(3) -- perl pragma interface to Opcode module. |
553 | |
554 | Safe(3) -- Opcode and namespace limited execution compartments |
555 | |
556 | =head1 AUTHORS |
557 | |
558 | Originally designed and implemented by Malcolm Beattie, |
559 | mbeattie@sable.ox.ac.uk as part of Safe version 1. |
560 | |
561 | Split out from Safe module version 1, named opcode tags and other |
7a57407b |
562 | changes added by Tim Bunce E<lt>F<Tim.Bunce@ig.co.uk>E<gt>. |
6badd1a5 |
563 | |
564 | =cut |
565 | |