This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
[perl #15987] Tests fail when sources are under CVS control
[perl5.git] / pod / perlsec.pod
CommitLineData
a0d0e21e
LW
1=head1 NAME
2
3perlsec - Perl security
4
5=head1 DESCRIPTION
6
425e5e39
PP
7Perl is designed to make it easy to program securely even when running
8with extra privileges, like setuid or setgid programs. Unlike most
54310121 9command line shells, which are based on multiple substitution passes on
425e5e39
PP
10each line of the script, Perl uses a more conventional evaluation scheme
11with fewer hidden snags. Additionally, because the language has more
54310121 12builtin functionality, it can rely less upon external (and possibly
425e5e39 13untrustworthy) programs to accomplish its purposes.
a0d0e21e 14
425e5e39
PP
15Perl automatically enables a set of special security checks, called I<taint
16mode>, when it detects its program running with differing real and effective
17user or group IDs. The setuid bit in Unix permissions is mode 04000, the
18setgid bit mode 02000; either or both may be set. You can also enable taint
5f05dabc 19mode explicitly by using the B<-T> command line flag. This flag is
425e5e39 20I<strongly> suggested for server programs and any program run on behalf of
fb73857a
PP
21someone else, such as a CGI script. Once taint mode is on, it's on for
22the remainder of your script.
a0d0e21e 23
1e422769
PP
24While in this mode, Perl takes special precautions called I<taint
25checks> to prevent both obvious and subtle traps. Some of these checks
26are reasonably simple, such as verifying that path directories aren't
27writable by others; careful programmers have always used checks like
28these. Other checks, however, are best supported by the language itself,
fb73857a 29and it is these checks especially that contribute to making a set-id Perl
425e5e39
PP
30program more secure than the corresponding C program.
31
fb73857a
PP
32You may not use data derived from outside your program to affect
33something else outside your program--at least, not by accident. All
34command line arguments, environment variables, locale information (see
d929ce6f 35L<perllocale>), results of certain system calls (readdir(),
41d6edb2
JH
36readlink(), the variable of shmread(), the messages returned by
37msgrcv(), the password, gcos and shell fields returned by the
38getpwxxx() calls), and all file input are marked as "tainted".
39Tainted data may not be used directly or indirectly in any command
40that invokes a sub-shell, nor in any command that modifies files,
b7ee89ce
AP
41directories, or processes, B<with the following exceptions>:
42
43=over 4
44
45=item *
46
ee556d55 47If you pass more than one argument to either C<system> or C<exec>,
bbd7eb8a
RD
48the arguments are checked for taintedness B<but> the operation will still
49be attempted, emitting an optional warning. This will be fatal in a
50future version of perl so do not rely on it to bypass the tainting
51mechanism.
b7ee89ce
AP
52
53=item *
54
55Arguments to C<print> and C<syswrite> are B<not> checked for taintedness.
56
7f6513c1
JH
57=item *
58
59Symbolic methods
60
61 $obj->$method(@args);
62
63and symbolic sub references
64
65 &{$foo}(@args);
66 $foo->(@args);
67
68are not checked for taintedness. This requires extra carefulness
69unless you want external data to affect your control flow. Unless
70you carefully limit what these symbolic values are, people are able
71to call functions B<outside> your Perl code, such as POSIX::system,
72in which case they are able to run arbitrary external code.
73
b7ee89ce
AP
74=back
75
ee556d55
MG
76The value of an expression containing tainted data will itself be
77tainted, even if it is logically impossible for the tainted data to
78affect the value.
79
d929ce6f
JH
80Because taintedness is associated with each scalar value, some
81elements of an array can be tainted and others not.
a0d0e21e 82
a0d0e21e
LW
83For example:
84
425e5e39
PP
85 $arg = shift; # $arg is tainted
86 $hid = $arg, 'bar'; # $hid is also tainted
87 $line = <>; # Tainted
8ebc5c01
PP
88 $line = <STDIN>; # Also tainted
89 open FOO, "/home/me/bar" or die $!;
90 $line = <FOO>; # Still tainted
a0d0e21e 91 $path = $ENV{'PATH'}; # Tainted, but see below
425e5e39 92 $data = 'abc'; # Not tainted
a0d0e21e 93
425e5e39 94 system "echo $arg"; # Insecure
bbd7eb8a
RD
95 system "/bin/echo", $arg; # Allowed but considered insecure
96 # (Perl doesn't know about /bin/echo)
425e5e39
PP
97 system "echo $hid"; # Insecure
98 system "echo $data"; # Insecure until PATH set
a0d0e21e 99
425e5e39 100 $path = $ENV{'PATH'}; # $path now tainted
a0d0e21e 101
54310121 102 $ENV{'PATH'} = '/bin:/usr/bin';
c90c0ff4 103 delete @ENV{'IFS', 'CDPATH', 'ENV', 'BASH_ENV'};
a0d0e21e 104
425e5e39
PP
105 $path = $ENV{'PATH'}; # $path now NOT tainted
106 system "echo $data"; # Is secure now!
a0d0e21e 107
425e5e39
PP
108 open(FOO, "< $arg"); # OK - read-only file
109 open(FOO, "> $arg"); # Not OK - trying to write
a0d0e21e 110
bbd7eb8a 111 open(FOO,"echo $arg|"); # Not OK
425e5e39 112 open(FOO,"-|")
bbd7eb8a 113 or exec 'echo', $arg; # Allowed but not really OK
a0d0e21e 114
425e5e39 115 $shout = `echo $arg`; # Insecure, $shout now tainted
a0d0e21e 116
425e5e39
PP
117 unlink $data, $arg; # Insecure
118 umask $arg; # Insecure
a0d0e21e 119
bbd7eb8a
RD
120 exec "echo $arg"; # Insecure
121 exec "echo", $arg; # Allowed but considered insecure
425e5e39 122 exec "sh", '-c', $arg; # Considered secure, alas!
a0d0e21e 123
3a4b19e4
GS
124 @files = <*.c>; # insecure (uses readdir() or similar)
125 @files = glob('*.c'); # insecure (uses readdir() or similar)
7bac28a0 126
3f7d42d8
JH
127 # In Perl releases older than 5.6.0 the <*.c> and glob('*.c') would
128 # have used an external program to do the filename expansion; but in
129 # either case the result is tainted since the list of filenames comes
130 # from outside of the program.
131
ee556d55
MG
132 $bad = ($arg, 23); # $bad will be tainted
133 $arg, `true`; # Insecure (although it isn't really)
134
a0d0e21e 135If you try to do something insecure, you will get a fatal error saying
62f468fc 136something like "Insecure dependency" or "Insecure $ENV{PATH}". Note that you
425e5e39 137can still write an insecure B<system> or B<exec>, but only by explicitly
bbd7eb8a
RD
138doing something like the "considered secure" example above. This will not
139be possible in a future version of Perl.
425e5e39
PP
140
141=head2 Laundering and Detecting Tainted Data
142
3f7d42d8
JH
143To test whether a variable contains tainted data, and whose use would
144thus trigger an "Insecure dependency" message, you can use the
145tainted() function of the Scalar::Util module, available in your
146nearby CPAN mirror, and included in Perl starting from the release 5.8.0.
147Or you may be able to use the following I<is_tainted()> function.
425e5e39
PP
148
149 sub is_tainted {
61890e45 150 return ! eval { eval("#" . substr(join("", @_), 0, 0)); 1 };
425e5e39
PP
151 }
152
153This function makes use of the fact that the presence of tainted data
154anywhere within an expression renders the entire expression tainted. It
155would be inefficient for every operator to test every argument for
156taintedness. Instead, the slightly more efficient and conservative
157approach is used that if any tainted value has been accessed within the
158same expression, the whole expression is considered tainted.
159
5f05dabc 160But testing for taintedness gets you only so far. Sometimes you have just
425e5e39 161to clear your data's taintedness. The only way to bypass the tainting
54310121 162mechanism is by referencing subpatterns from a regular expression match.
425e5e39
PP
163Perl presumes that if you reference a substring using $1, $2, etc., that
164you knew what you were doing when you wrote the pattern. That means using
165a bit of thought--don't just blindly untaint anything, or you defeat the
a034a98d
DD
166entire mechanism. It's better to verify that the variable has only good
167characters (for certain values of "good") rather than checking whether it
168has any bad characters. That's because it's far too easy to miss bad
169characters that you never thought of.
425e5e39
PP
170
171Here's a test to make sure that the data contains nothing but "word"
172characters (alphabetics, numerics, and underscores), a hyphen, an at sign,
173or a dot.
174
54310121 175 if ($data =~ /^([-\@\w.]+)$/) {
425e5e39
PP
176 $data = $1; # $data now untainted
177 } else {
178 die "Bad data in $data"; # log this somewhere
179 }
180
5f05dabc 181This is fairly secure because C</\w+/> doesn't normally match shell
425e5e39
PP
182metacharacters, nor are dot, dash, or at going to mean something special
183to the shell. Use of C</.+/> would have been insecure in theory because
184it lets everything through, but Perl doesn't check for that. The lesson
185is that when untainting, you must be exceedingly careful with your patterns.
19799a22 186Laundering data using regular expression is the I<only> mechanism for
425e5e39
PP
187untainting dirty data, unless you use the strategy detailed below to fork
188a child of lesser privilege.
189
a034a98d
DD
190The example does not untaint $data if C<use locale> is in effect,
191because the characters matched by C<\w> are determined by the locale.
192Perl considers that locale definitions are untrustworthy because they
193contain data from outside the program. If you are writing a
194locale-aware program, and want to launder data with a regular expression
195containing C<\w>, put C<no locale> ahead of the expression in the same
196block. See L<perllocale/SECURITY> for further discussion and examples.
197
3a52c276
CS
198=head2 Switches On the "#!" Line
199
200When you make a script executable, in order to make it usable as a
201command, the system will pass switches to perl from the script's #!
54310121 202line. Perl checks that any command line switches given to a setuid
3a52c276 203(or setgid) script actually match the ones set on the #! line. Some
54310121 204Unix and Unix-like environments impose a one-switch limit on the #!
3a52c276 205line, so you may need to use something like C<-wU> instead of C<-w -U>
54310121
PP
206under such systems. (This issue should arise only in Unix or
207Unix-like environments that support #! and setuid or setgid scripts.)
3a52c276 208
425e5e39
PP
209=head2 Cleaning Up Your Path
210
1fef88e7 211For "Insecure C<$ENV{PATH}>" messages, you need to set C<$ENV{'PATH'}> to a
1e422769
PP
212known value, and each directory in the path must be non-writable by others
213than its owner and group. You may be surprised to get this message even
214if the pathname to your executable is fully qualified. This is I<not>
215generated because you didn't supply a full path to the program; instead,
216it's generated because you never set your PATH environment variable, or
217you didn't set it to something that was safe. Because Perl can't
218guarantee that the executable in question isn't itself going to turn
219around and execute some other program that is dependent on your PATH, it
54310121 220makes sure you set the PATH.
a0d0e21e 221
a3cb178b
GS
222The PATH isn't the only environment variable which can cause problems.
223Because some shells may use the variables IFS, CDPATH, ENV, and
224BASH_ENV, Perl checks that those are either empty or untainted when
225starting subprocesses. You may wish to add something like this to your
226setid and taint-checking scripts.
227
228 delete @ENV{qw(IFS CDPATH ENV BASH_ENV)}; # Make %ENV safer
229
a0d0e21e
LW
230It's also possible to get into trouble with other operations that don't
231care whether they use tainted values. Make judicious use of the file
232tests in dealing with any user-supplied filenames. When possible, do
fb73857a
PP
233opens and such B<after> properly dropping any special user (or group!)
234privileges. Perl doesn't prevent you from opening tainted filenames for reading,
a0d0e21e
LW
235so be careful what you print out. The tainting mechanism is intended to
236prevent stupid mistakes, not to remove the need for thought.
237
425e5e39
PP
238Perl does not call the shell to expand wild cards when you pass B<system>
239and B<exec> explicit parameter lists instead of strings with possible shell
240wildcards in them. Unfortunately, the B<open>, B<glob>, and
54310121
PP
241backtick functions provide no such alternate calling convention, so more
242subterfuge will be required.
425e5e39
PP
243
244Perl provides a reasonably safe way to open a file or pipe from a setuid
245or setgid program: just create a child process with reduced privilege who
246does the dirty work for you. First, fork a child using the special
247B<open> syntax that connects the parent and child by a pipe. Now the
248child resets its ID set and any other per-process attributes, like
249environment variables, umasks, current working directories, back to the
250originals or known safe values. Then the child process, which no longer
251has any special permissions, does the B<open> or other system call.
252Finally, the child passes the data it managed to access back to the
5f05dabc 253parent. Because the file or pipe was opened in the child while running
425e5e39
PP
254under less privilege than the parent, it's not apt to be tricked into
255doing something it shouldn't.
256
54310121 257Here's a way to do backticks reasonably safely. Notice how the B<exec> is
425e5e39
PP
258not called with a string that the shell could expand. This is by far the
259best way to call something that might be subjected to shell escapes: just
fb73857a 260never call the shell at all.
cb1a09d0 261
a1ce9542 262 use English '-no_match_vars';
e093bcf0
GW
263 die "Can't fork: $!" unless defined($pid = open(KID, "-|"));
264 if ($pid) { # parent
265 while (<KID>) {
266 # do something
267 }
268 close KID;
269 } else {
270 my @temp = ($EUID, $EGID);
271 my $orig_uid = $UID;
272 my $orig_gid = $GID;
273 $EUID = $UID;
274 $EGID = $GID;
275 # Drop privileges
276 $UID = $orig_uid;
277 $GID = $orig_gid;
278 # Make sure privs are really gone
279 ($EUID, $EGID) = @temp;
280 die "Can't drop privileges"
281 unless $UID == $EUID && $GID eq $EGID;
282 $ENV{PATH} = "/bin:/usr/bin"; # Minimal PATH.
283 # Consider sanitizing the environment even more.
284 exec 'myprog', 'arg1', 'arg2'
285 or die "can't exec myprog: $!";
286 }
425e5e39 287
fb73857a
PP
288A similar strategy would work for wildcard expansion via C<glob>, although
289you can use C<readdir> instead.
425e5e39
PP
290
291Taint checking is most useful when although you trust yourself not to have
292written a program to give away the farm, you don't necessarily trust those
293who end up using it not to try to trick it into doing something bad. This
fb73857a 294is the kind of security checking that's useful for set-id programs and
425e5e39
PP
295programs launched on someone else's behalf, like CGI programs.
296
297This is quite different, however, from not even trusting the writer of the
298code not to try to do something evil. That's the kind of trust needed
299when someone hands you a program you've never seen before and says, "Here,
300run this." For that kind of safety, check out the Safe module,
301included standard in the Perl distribution. This module allows the
302programmer to set up special compartments in which all system operations
303are trapped and namespace access is carefully controlled.
304
305=head2 Security Bugs
306
307Beyond the obvious problems that stem from giving special privileges to
fb73857a 308systems as flexible as scripts, on many versions of Unix, set-id scripts
425e5e39
PP
309are inherently insecure right from the start. The problem is a race
310condition in the kernel. Between the time the kernel opens the file to
fb73857a 311see which interpreter to run and when the (now-set-id) interpreter turns
425e5e39
PP
312around and reopens the file to interpret it, the file in question may have
313changed, especially if you have symbolic links on your system.
314
315Fortunately, sometimes this kernel "feature" can be disabled.
316Unfortunately, there are two ways to disable it. The system can simply
fb73857a
PP
317outlaw scripts with any set-id bit set, which doesn't help much.
318Alternately, it can simply ignore the set-id bits on scripts. If the
425e5e39
PP
319latter is true, Perl can emulate the setuid and setgid mechanism when it
320notices the otherwise useless setuid/gid bits on Perl scripts. It does
321this via a special executable called B<suidperl> that is automatically
54310121 322invoked for you if it's needed.
425e5e39 323
fb73857a
PP
324However, if the kernel set-id script feature isn't disabled, Perl will
325complain loudly that your set-id script is insecure. You'll need to
326either disable the kernel set-id script feature, or put a C wrapper around
425e5e39
PP
327the script. A C wrapper is just a compiled program that does nothing
328except call your Perl program. Compiled programs are not subject to the
fb73857a 329kernel bug that plagues set-id scripts. Here's a simple wrapper, written
425e5e39
PP
330in C:
331
332 #define REAL_PATH "/path/to/script"
54310121 333 main(ac, av)
425e5e39
PP
334 char **av;
335 {
336 execv(REAL_PATH, av);
54310121 337 }
cb1a09d0 338
54310121
PP
339Compile this wrapper into a binary executable and then make I<it> rather
340than your script setuid or setgid.
425e5e39 341
425e5e39
PP
342In recent years, vendors have begun to supply systems free of this
343inherent security bug. On such systems, when the kernel passes the name
fb73857a 344of the set-id script to open to the interpreter, rather than using a
425e5e39
PP
345pathname subject to meddling, it instead passes I</dev/fd/3>. This is a
346special file already opened on the script, so that there can be no race
347condition for evil scripts to exploit. On these systems, Perl should be
348compiled with C<-DSETUID_SCRIPTS_ARE_SECURE_NOW>. The B<Configure>
349program that builds Perl tries to figure this out for itself, so you
350should never have to specify this yourself. Most modern releases of
351SysVr4 and BSD 4.4 use this approach to avoid the kernel race condition.
352
0325b4c4
JH
353Prior to release 5.6.1 of Perl, bugs in the code of B<suidperl> could
354introduce a security hole.
68dc0745
PP
355
356=head2 Protecting Your Programs
357
358There are a number of ways to hide the source to your Perl programs,
359with varying levels of "security".
360
361First of all, however, you I<can't> take away read permission, because
362the source code has to be readable in order to be compiled and
363interpreted. (That doesn't mean that a CGI script's source is
364readable by people on the web, though.) So you have to leave the
5a964f20
TC
365permissions at the socially friendly 0755 level. This lets
366people on your local system only see your source.
68dc0745 367
5a964f20 368Some people mistakenly regard this as a security problem. If your program does
68dc0745
PP
369insecure things, and relies on people not knowing how to exploit those
370insecurities, it is not secure. It is often possible for someone to
371determine the insecure things and exploit them without viewing the
372source. Security through obscurity, the name for hiding your bugs
373instead of fixing them, is little security indeed.
374
83df6a1d
JH
375You can try using encryption via source filters (Filter::* from CPAN,
376or Filter::Util::Call and Filter::Simple since Perl 5.8).
377But crackers might be able to decrypt it. You can try using the byte
378code compiler and interpreter described below, but crackers might be
379able to de-compile it. You can try using the native-code compiler
68dc0745
PP
380described below, but crackers might be able to disassemble it. These
381pose varying degrees of difficulty to people wanting to get at your
382code, but none can definitively conceal it (this is true of every
383language, not just Perl).
384
385If you're concerned about people profiting from your code, then the
386bottom line is that nothing but a restrictive licence will give you
387legal security. License your software and pepper it with threatening
388statements like "This is unpublished proprietary software of XYZ Corp.
389Your access to it does not give you permission to use it blah blah
390blah." You should see a lawyer to be sure your licence's wording will
391stand up in court.
5a964f20 392
0d7c09bb
JH
393=head2 Unicode
394
395Unicode is a new and complex technology and one may easily overlook
396certain security pitfalls. See L<perluniintro> for an overview and
397L<perlunicode> for details, and L<perlunicode/"Security Implications
398of Unicode"> for security implications in particular.
399
5a964f20
TC
400=head1 SEE ALSO
401
402L<perlrun> for its description of cleaning up environment variables.