| 1 |
package Opcode; |
| 2 |
|
| 3 |
use 5.006_001; |
| 4 |
|
| 5 |
use strict; |
| 6 |
|
| 7 |
our($VERSION, $XS_VERSION, @ISA, @EXPORT_OK); |
| 8 |
|
| 9 |
$VERSION = "1.05"; |
| 10 |
$XS_VERSION = "1.03"; |
| 11 |
|
| 12 |
use Carp; |
| 13 |
use Exporter (); |
| 14 |
use XSLoader (); |
| 15 |
|
| 16 |
BEGIN { |
| 17 |
@ISA = qw(Exporter); |
| 18 |
@EXPORT_OK = qw( |
| 19 |
opset ops_to_opset |
| 20 |
opset_to_ops opset_to_hex invert_opset |
| 21 |
empty_opset full_opset |
| 22 |
opdesc opcodes opmask define_optag |
| 23 |
opmask_add verify_opset opdump |
| 24 |
); |
| 25 |
} |
| 26 |
|
| 27 |
sub opset (;@); |
| 28 |
sub opset_to_hex ($); |
| 29 |
sub opdump (;$); |
| 30 |
use subs @EXPORT_OK; |
| 31 |
|
| 32 |
XSLoader::load 'Opcode', $XS_VERSION; |
| 33 |
|
| 34 |
_init_optags(); |
| 35 |
|
| 36 |
sub ops_to_opset { opset @_ } # alias for old name |
| 37 |
|
| 38 |
sub opset_to_hex ($) { |
| 39 |
return "(invalid opset)" unless verify_opset($_[0]); |
| 40 |
unpack("h*",$_[0]); |
| 41 |
} |
| 42 |
|
| 43 |
sub opdump (;$) { |
| 44 |
my $pat = shift; |
| 45 |
# handy utility: perl -MOpcode=opdump -e 'opdump File' |
| 46 |
foreach(opset_to_ops(full_opset)) { |
| 47 |
my $op = sprintf " %12s %s\n", $_, opdesc($_); |
| 48 |
next if defined $pat and $op !~ m/$pat/i; |
| 49 |
print $op; |
| 50 |
} |
| 51 |
} |
| 52 |
|
| 53 |
|
| 54 |
|
| 55 |
sub _init_optags { |
| 56 |
my(%all, %seen); |
| 57 |
@all{opset_to_ops(full_opset)} = (); # keys only |
| 58 |
|
| 59 |
local($_); |
| 60 |
local($/) = "\n=cut"; # skip to optags definition section |
| 61 |
<DATA>; |
| 62 |
$/ = "\n="; # now read in 'pod section' chunks |
| 63 |
while(<DATA>) { |
| 64 |
next unless m/^item\s+(:\w+)/; |
| 65 |
my $tag = $1; |
| 66 |
|
| 67 |
# Split into lines, keep only indented lines |
| 68 |
my @lines = grep { m/^\s/ } split(/\n/); |
| 69 |
foreach (@lines) { s/--.*// } # delete comments |
| 70 |
my @ops = map { split ' ' } @lines; # get op words |
| 71 |
|
| 72 |
foreach(@ops) { |
| 73 |
warn "$tag - $_ already tagged in $seen{$_}\n" if $seen{$_}; |
| 74 |
$seen{$_} = $tag; |
| 75 |
delete $all{$_}; |
| 76 |
} |
| 77 |
# opset will croak on invalid names |
| 78 |
define_optag($tag, opset(@ops)); |
| 79 |
} |
| 80 |
close(DATA); |
| 81 |
warn "Untagged opnames: ".join(' ',keys %all)."\n" if %all; |
| 82 |
} |
| 83 |
|
| 84 |
|
| 85 |
1; |
| 86 |
|
| 87 |
__DATA__ |
| 88 |
|
| 89 |
=head1 NAME |
| 90 |
|
| 91 |
Opcode - Disable named opcodes when compiling perl code |
| 92 |
|
| 93 |
=head1 SYNOPSIS |
| 94 |
|
| 95 |
use Opcode; |
| 96 |
|
| 97 |
|
| 98 |
=head1 DESCRIPTION |
| 99 |
|
| 100 |
Perl code is always compiled into an internal format before execution. |
| 101 |
|
| 102 |
Evaluating perl code (e.g. via "eval" or "do 'file'") causes |
| 103 |
the code to be compiled into an internal format and then, |
| 104 |
provided there was no error in the compilation, executed. |
| 105 |
The internal format is based on many distinct I<opcodes>. |
| 106 |
|
| 107 |
By default no opmask is in effect and any code can be compiled. |
| 108 |
|
| 109 |
The Opcode module allow you to define an I<operator mask> to be in |
| 110 |
effect when perl I<next> compiles any code. Attempting to compile code |
| 111 |
which contains a masked opcode will cause the compilation to fail |
| 112 |
with an error. The code will not be executed. |
| 113 |
|
| 114 |
=head1 NOTE |
| 115 |
|
| 116 |
The Opcode module is not usually used directly. See the ops pragma and |
| 117 |
Safe modules for more typical uses. |
| 118 |
|
| 119 |
=head1 WARNING |
| 120 |
|
| 121 |
The authors make B<no warranty>, implied or otherwise, about the |
| 122 |
suitability of this software for safety or security purposes. |
| 123 |
|
| 124 |
The authors shall not in any case be liable for special, incidental, |
| 125 |
consequential, indirect or other similar damages arising from the use |
| 126 |
of this software. |
| 127 |
|
| 128 |
Your mileage will vary. If in any doubt B<do not use it>. |
| 129 |
|
| 130 |
|
| 131 |
=head1 Operator Names and Operator Lists |
| 132 |
|
| 133 |
The canonical list of operator names is the contents of the array |
| 134 |
PL_op_name defined and initialised in file F<opcode.h> of the Perl |
| 135 |
source distribution (and installed into the perl library). |
| 136 |
|
| 137 |
Each operator has both a terse name (its opname) and a more verbose or |
| 138 |
recognisable descriptive name. The opdesc function can be used to |
| 139 |
return a list of descriptions for a list of operators. |
| 140 |
|
| 141 |
Many of the functions and methods listed below take a list of |
| 142 |
operators as parameters. Most operator lists can be made up of several |
| 143 |
types of element. Each element can be one of |
| 144 |
|
| 145 |
=over 8 |
| 146 |
|
| 147 |
=item an operator name (opname) |
| 148 |
|
| 149 |
Operator names are typically small lowercase words like enterloop, |
| 150 |
leaveloop, last, next, redo etc. Sometimes they are rather cryptic |
| 151 |
like gv2cv, i_ncmp and ftsvtx. |
| 152 |
|
| 153 |
=item an operator tag name (optag) |
| 154 |
|
| 155 |
Operator tags can be used to refer to groups (or sets) of operators. |
| 156 |
Tag names always begin with a colon. The Opcode module defines several |
| 157 |
optags and the user can define others using the define_optag function. |
| 158 |
|
| 159 |
=item a negated opname or optag |
| 160 |
|
| 161 |
An opname or optag can be prefixed with an exclamation mark, e.g., !mkdir. |
| 162 |
Negating an opname or optag means remove the corresponding ops from the |
| 163 |
accumulated set of ops at that point. |
| 164 |
|
| 165 |
=item an operator set (opset) |
| 166 |
|
| 167 |
An I<opset> as a binary string of approximately 44 bytes which holds a |
| 168 |
set or zero or more operators. |
| 169 |
|
| 170 |
The opset and opset_to_ops functions can be used to convert from |
| 171 |
a list of operators to an opset and I<vice versa>. |
| 172 |
|
| 173 |
Wherever a list of operators can be given you can use one or more opsets. |
| 174 |
See also Manipulating Opsets below. |
| 175 |
|
| 176 |
=back |
| 177 |
|
| 178 |
|
| 179 |
=head1 Opcode Functions |
| 180 |
|
| 181 |
The Opcode package contains functions for manipulating operator names |
| 182 |
tags and sets. All are available for export by the package. |
| 183 |
|
| 184 |
=over 8 |
| 185 |
|
| 186 |
=item opcodes |
| 187 |
|
| 188 |
In a scalar context opcodes returns the number of opcodes in this |
| 189 |
version of perl (around 350 for perl-5.7.0). |
| 190 |
|
| 191 |
In a list context it returns a list of all the operator names. |
| 192 |
(Not yet implemented, use @names = opset_to_ops(full_opset).) |
| 193 |
|
| 194 |
=item opset (OP, ...) |
| 195 |
|
| 196 |
Returns an opset containing the listed operators. |
| 197 |
|
| 198 |
=item opset_to_ops (OPSET) |
| 199 |
|
| 200 |
Returns a list of operator names corresponding to those operators in |
| 201 |
the set. |
| 202 |
|
| 203 |
=item opset_to_hex (OPSET) |
| 204 |
|
| 205 |
Returns a string representation of an opset. Can be handy for debugging. |
| 206 |
|
| 207 |
=item full_opset |
| 208 |
|
| 209 |
Returns an opset which includes all operators. |
| 210 |
|
| 211 |
=item empty_opset |
| 212 |
|
| 213 |
Returns an opset which contains no operators. |
| 214 |
|
| 215 |
=item invert_opset (OPSET) |
| 216 |
|
| 217 |
Returns an opset which is the inverse set of the one supplied. |
| 218 |
|
| 219 |
=item verify_opset (OPSET, ...) |
| 220 |
|
| 221 |
Returns true if the supplied opset looks like a valid opset (is the |
| 222 |
right length etc) otherwise it returns false. If an optional second |
| 223 |
parameter is true then verify_opset will croak on an invalid opset |
| 224 |
instead of returning false. |
| 225 |
|
| 226 |
Most of the other Opcode functions call verify_opset automatically |
| 227 |
and will croak if given an invalid opset. |
| 228 |
|
| 229 |
=item define_optag (OPTAG, OPSET) |
| 230 |
|
| 231 |
Define OPTAG as a symbolic name for OPSET. Optag names always start |
| 232 |
with a colon C<:>. |
| 233 |
|
| 234 |
The optag name used must not be defined already (define_optag will |
| 235 |
croak if it is already defined). Optag names are global to the perl |
| 236 |
process and optag definitions cannot be altered or deleted once |
| 237 |
defined. |
| 238 |
|
| 239 |
It is strongly recommended that applications using Opcode should use a |
| 240 |
leading capital letter on their tag names since lowercase names are |
| 241 |
reserved for use by the Opcode module. If using Opcode within a module |
| 242 |
you should prefix your tags names with the name of your module to |
| 243 |
ensure uniqueness and thus avoid clashes with other modules. |
| 244 |
|
| 245 |
=item opmask_add (OPSET) |
| 246 |
|
| 247 |
Adds the supplied opset to the current opmask. Note that there is |
| 248 |
currently I<no> mechanism for unmasking ops once they have been masked. |
| 249 |
This is intentional. |
| 250 |
|
| 251 |
=item opmask |
| 252 |
|
| 253 |
Returns an opset corresponding to the current opmask. |
| 254 |
|
| 255 |
=item opdesc (OP, ...) |
| 256 |
|
| 257 |
This takes a list of operator names and returns the corresponding list |
| 258 |
of operator descriptions. |
| 259 |
|
| 260 |
=item opdump (PAT) |
| 261 |
|
| 262 |
Dumps to STDOUT a two column list of op names and op descriptions. |
| 263 |
If an optional pattern is given then only lines which match the |
| 264 |
(case insensitive) pattern will be output. |
| 265 |
|
| 266 |
It's designed to be used as a handy command line utility: |
| 267 |
|
| 268 |
perl -MOpcode=opdump -e opdump |
| 269 |
perl -MOpcode=opdump -e 'opdump Eval' |
| 270 |
|
| 271 |
=back |
| 272 |
|
| 273 |
=head1 Manipulating Opsets |
| 274 |
|
| 275 |
Opsets may be manipulated using the perl bit vector operators & (and), | (or), |
| 276 |
^ (xor) and ~ (negate/invert). |
| 277 |
|
| 278 |
However you should never rely on the numerical position of any opcode |
| 279 |
within the opset. In other words both sides of a bit vector operator |
| 280 |
should be opsets returned from Opcode functions. |
| 281 |
|
| 282 |
Also, since the number of opcodes in your current version of perl might |
| 283 |
not be an exact multiple of eight, there may be unused bits in the last |
| 284 |
byte of an upset. This should not cause any problems (Opcode functions |
| 285 |
ignore those extra bits) but it does mean that using the ~ operator |
| 286 |
will typically not produce the same 'physical' opset 'string' as the |
| 287 |
invert_opset function. |
| 288 |
|
| 289 |
|
| 290 |
=head1 TO DO (maybe) |
| 291 |
|
| 292 |
$bool = opset_eq($opset1, $opset2) true if opsets are logically eqiv |
| 293 |
|
| 294 |
$yes = opset_can($opset, @ops) true if $opset has all @ops set |
| 295 |
|
| 296 |
@diff = opset_diff($opset1, $opset2) => ('foo', '!bar', ...) |
| 297 |
|
| 298 |
=cut |
| 299 |
|
| 300 |
# the =cut above is used by _init_optags() to get here quickly |
| 301 |
|
| 302 |
=head1 Predefined Opcode Tags |
| 303 |
|
| 304 |
=over 5 |
| 305 |
|
| 306 |
=item :base_core |
| 307 |
|
| 308 |
null stub scalar pushmark wantarray const defined undef |
| 309 |
|
| 310 |
rv2sv sassign |
| 311 |
|
| 312 |
rv2av aassign aelem aelemfast aslice av2arylen |
| 313 |
|
| 314 |
rv2hv helem hslice each values keys exists delete |
| 315 |
|
| 316 |
preinc i_preinc predec i_predec postinc i_postinc postdec i_postdec |
| 317 |
int hex oct abs pow multiply i_multiply divide i_divide |
| 318 |
modulo i_modulo add i_add subtract i_subtract |
| 319 |
|
| 320 |
left_shift right_shift bit_and bit_xor bit_or negate i_negate |
| 321 |
not complement |
| 322 |
|
| 323 |
lt i_lt gt i_gt le i_le ge i_ge eq i_eq ne i_ne ncmp i_ncmp |
| 324 |
slt sgt sle sge seq sne scmp |
| 325 |
|
| 326 |
substr vec stringify study pos length index rindex ord chr |
| 327 |
|
| 328 |
ucfirst lcfirst uc lc quotemeta trans chop schop chomp schomp |
| 329 |
|
| 330 |
match split qr |
| 331 |
|
| 332 |
list lslice splice push pop shift unshift reverse |
| 333 |
|
| 334 |
cond_expr flip flop andassign orassign and or xor |
| 335 |
|
| 336 |
warn die lineseq nextstate scope enter leave setstate |
| 337 |
|
| 338 |
rv2cv anoncode prototype |
| 339 |
|
| 340 |
entersub leavesub leavesublv return method method_named -- XXX loops via recursion? |
| 341 |
|
| 342 |
leaveeval -- needed for Safe to operate, is safe without entereval |
| 343 |
|
| 344 |
=item :base_mem |
| 345 |
|
| 346 |
These memory related ops are not included in :base_core because they |
| 347 |
can easily be used to implement a resource attack (e.g., consume all |
| 348 |
available memory). |
| 349 |
|
| 350 |
concat repeat join range |
| 351 |
|
| 352 |
anonlist anonhash |
| 353 |
|
| 354 |
Note that despite the existance of this optag a memory resource attack |
| 355 |
may still be possible using only :base_core ops. |
| 356 |
|
| 357 |
Disabling these ops is a I<very> heavy handed way to attempt to prevent |
| 358 |
a memory resource attack. It's probable that a specific memory limit |
| 359 |
mechanism will be added to perl in the near future. |
| 360 |
|
| 361 |
=item :base_loop |
| 362 |
|
| 363 |
These loop ops are not included in :base_core because they can easily be |
| 364 |
used to implement a resource attack (e.g., consume all available CPU time). |
| 365 |
|
| 366 |
grepstart grepwhile |
| 367 |
mapstart mapwhile |
| 368 |
enteriter iter |
| 369 |
enterloop leaveloop unstack |
| 370 |
last next redo |
| 371 |
goto |
| 372 |
|
| 373 |
=item :base_io |
| 374 |
|
| 375 |
These ops enable I<filehandle> (rather than filename) based input and |
| 376 |
output. These are safe on the assumption that only pre-existing |
| 377 |
filehandles are available for use. To create new filehandles other ops |
| 378 |
such as open would need to be enabled. |
| 379 |
|
| 380 |
readline rcatline getc read |
| 381 |
|
| 382 |
formline enterwrite leavewrite |
| 383 |
|
| 384 |
print sysread syswrite send recv |
| 385 |
|
| 386 |
eof tell seek sysseek |
| 387 |
|
| 388 |
readdir telldir seekdir rewinddir |
| 389 |
|
| 390 |
=item :base_orig |
| 391 |
|
| 392 |
These are a hotchpotch of opcodes still waiting to be considered |
| 393 |
|
| 394 |
gvsv gv gelem |
| 395 |
|
| 396 |
padsv padav padhv padany |
| 397 |
|
| 398 |
rv2gv refgen srefgen ref |
| 399 |
|
| 400 |
bless -- could be used to change ownership of objects (reblessing) |
| 401 |
|
| 402 |
pushre regcmaybe regcreset regcomp subst substcont |
| 403 |
|
| 404 |
sprintf prtf -- can core dump |
| 405 |
|
| 406 |
crypt |
| 407 |
|
| 408 |
tie untie |
| 409 |
|
| 410 |
dbmopen dbmclose |
| 411 |
sselect select |
| 412 |
pipe_op sockpair |
| 413 |
|
| 414 |
getppid getpgrp setpgrp getpriority setpriority localtime gmtime |
| 415 |
|
| 416 |
entertry leavetry -- can be used to 'hide' fatal errors |
| 417 |
|
| 418 |
custom -- where should this go |
| 419 |
|
| 420 |
=item :base_math |
| 421 |
|
| 422 |
These ops are not included in :base_core because of the risk of them being |
| 423 |
used to generate floating point exceptions (which would have to be caught |
| 424 |
using a $SIG{FPE} handler). |
| 425 |
|
| 426 |
atan2 sin cos exp log sqrt |
| 427 |
|
| 428 |
These ops are not included in :base_core because they have an effect |
| 429 |
beyond the scope of the compartment. |
| 430 |
|
| 431 |
rand srand |
| 432 |
|
| 433 |
=item :base_thread |
| 434 |
|
| 435 |
These ops are related to multi-threading. |
| 436 |
|
| 437 |
lock threadsv |
| 438 |
|
| 439 |
=item :default |
| 440 |
|
| 441 |
A handy tag name for a I<reasonable> default set of ops. (The current ops |
| 442 |
allowed are unstable while development continues. It will change.) |
| 443 |
|
| 444 |
:base_core :base_mem :base_loop :base_io :base_orig :base_thread |
| 445 |
|
| 446 |
If safety matters to you (and why else would you be using the Opcode module?) |
| 447 |
then you should not rely on the definition of this, or indeed any other, optag! |
| 448 |
|
| 449 |
|
| 450 |
=item :filesys_read |
| 451 |
|
| 452 |
stat lstat readlink |
| 453 |
|
| 454 |
ftatime ftblk ftchr ftctime ftdir fteexec fteowned fteread |
| 455 |
ftewrite ftfile ftis ftlink ftmtime ftpipe ftrexec ftrowned |
| 456 |
ftrread ftsgid ftsize ftsock ftsuid fttty ftzero ftrwrite ftsvtx |
| 457 |
|
| 458 |
fttext ftbinary |
| 459 |
|
| 460 |
fileno |
| 461 |
|
| 462 |
=item :sys_db |
| 463 |
|
| 464 |
ghbyname ghbyaddr ghostent shostent ehostent -- hosts |
| 465 |
gnbyname gnbyaddr gnetent snetent enetent -- networks |
| 466 |
gpbyname gpbynumber gprotoent sprotoent eprotoent -- protocols |
| 467 |
gsbyname gsbyport gservent sservent eservent -- services |
| 468 |
|
| 469 |
gpwnam gpwuid gpwent spwent epwent getlogin -- users |
| 470 |
ggrnam ggrgid ggrent sgrent egrent -- groups |
| 471 |
|
| 472 |
=item :browse |
| 473 |
|
| 474 |
A handy tag name for a I<reasonable> default set of ops beyond the |
| 475 |
:default optag. Like :default (and indeed all the other optags) its |
| 476 |
current definition is unstable while development continues. It will change. |
| 477 |
|
| 478 |
The :browse tag represents the next step beyond :default. It it a |
| 479 |
superset of the :default ops and adds :filesys_read the :sys_db. |
| 480 |
The intent being that scripts can access more (possibly sensitive) |
| 481 |
information about your system but not be able to change it. |
| 482 |
|
| 483 |
:default :filesys_read :sys_db |
| 484 |
|
| 485 |
=item :filesys_open |
| 486 |
|
| 487 |
sysopen open close |
| 488 |
umask binmode |
| 489 |
|
| 490 |
open_dir closedir -- other dir ops are in :base_io |
| 491 |
|
| 492 |
=item :filesys_write |
| 493 |
|
| 494 |
link unlink rename symlink truncate |
| 495 |
|
| 496 |
mkdir rmdir |
| 497 |
|
| 498 |
utime chmod chown |
| 499 |
|
| 500 |
fcntl -- not strictly filesys related, but possibly as dangerous? |
| 501 |
|
| 502 |
=item :subprocess |
| 503 |
|
| 504 |
backtick system |
| 505 |
|
| 506 |
fork |
| 507 |
|
| 508 |
wait waitpid |
| 509 |
|
| 510 |
glob -- access to Cshell via <`rm *`> |
| 511 |
|
| 512 |
=item :ownprocess |
| 513 |
|
| 514 |
exec exit kill |
| 515 |
|
| 516 |
time tms -- could be used for timing attacks (paranoid?) |
| 517 |
|
| 518 |
=item :others |
| 519 |
|
| 520 |
This tag holds groups of assorted specialist opcodes that don't warrant |
| 521 |
having optags defined for them. |
| 522 |
|
| 523 |
SystemV Interprocess Communications: |
| 524 |
|
| 525 |
msgctl msgget msgrcv msgsnd |
| 526 |
|
| 527 |
semctl semget semop |
| 528 |
|
| 529 |
shmctl shmget shmread shmwrite |
| 530 |
|
| 531 |
=item :still_to_be_decided |
| 532 |
|
| 533 |
chdir |
| 534 |
flock ioctl |
| 535 |
|
| 536 |
socket getpeername ssockopt |
| 537 |
bind connect listen accept shutdown gsockopt getsockname |
| 538 |
|
| 539 |
sleep alarm -- changes global timer state and signal handling |
| 540 |
sort -- assorted problems including core dumps |
| 541 |
tied -- can be used to access object implementing a tie |
| 542 |
pack unpack -- can be used to create/use memory pointers |
| 543 |
|
| 544 |
entereval -- can be used to hide code from initial compile |
| 545 |
require dofile |
| 546 |
|
| 547 |
caller -- get info about calling environment and args |
| 548 |
|
| 549 |
reset |
| 550 |
|
| 551 |
dbstate -- perl -d version of nextstate(ment) opcode |
| 552 |
|
| 553 |
=item :dangerous |
| 554 |
|
| 555 |
This tag is simply a bucket for opcodes that are unlikely to be used via |
| 556 |
a tag name but need to be tagged for completness and documentation. |
| 557 |
|
| 558 |
syscall dump chroot |
| 559 |
|
| 560 |
|
| 561 |
=back |
| 562 |
|
| 563 |
=head1 SEE ALSO |
| 564 |
|
| 565 |
ops(3) -- perl pragma interface to Opcode module. |
| 566 |
|
| 567 |
Safe(3) -- Opcode and namespace limited execution compartments |
| 568 |
|
| 569 |
=head1 AUTHORS |
| 570 |
|
| 571 |
Originally designed and implemented by Malcolm Beattie, |
| 572 |
mbeattie@sable.ox.ac.uk as part of Safe version 1. |
| 573 |
|
| 574 |
Split out from Safe module version 1, named opcode tags and other |
| 575 |
changes added by Tim Bunce. |
| 576 |
|
| 577 |
=cut |
| 578 |
|
