This is a live mirror of the Perl 5 development currently hosted at
Re: P and V
[perl5.git] / pod / perlsec.pod
1=head1 NAME
3perlsec - Perl security
425e5e39 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 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 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 21someone else, such as a CGI script. Once taint mode is on, it's on for
22the remainder of your script.
a0d0e21e 23
1e422769 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 30program more secure than the corresponding C program.
fb73857a 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(),
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,
41directories, or processes, B<with the following exceptions>:
43=over 4
45=item *
47Arguments to C<print> and C<syswrite> are B<not> checked for taintedness.
49=item *
51Symbolic methods
53 $obj->$method(@args);
55and symbolic sub references
57 &{$foo}(@args);
58 $foo->(@args);
60are not checked for taintedness. This requires extra carefulness
61unless you want external data to affect your control flow. Unless
62you carefully limit what these symbolic values are, people are able
63to call functions B<outside> your Perl code, such as POSIX::system,
64in which case they are able to run arbitrary external code.
68The value of an expression containing tainted data will itself be
69tainted, even if it is logically impossible for the tainted data to
70affect the value.
72Because taintedness is associated with each scalar value, some
73elements of an array can be tainted and others not.
a0d0e21e 74
75For example:
425e5e39 77 $arg = shift; # $arg is tainted
78 $hid = $arg, 'bar'; # $hid is also tainted
79 $line = <>; # Tainted
8ebc5c01 80 $line = <STDIN>; # Also tainted
81 open FOO, "/home/me/bar" or die $!;
82 $line = <FOO>; # Still tainted
a0d0e21e 83 $path = $ENV{'PATH'}; # Tainted, but see below
425e5e39 84 $data = 'abc'; # Not tainted
a0d0e21e 85
425e5e39 86 system "echo $arg"; # Insecure
7de90c4d 87 system "/bin/echo", $arg; # Considered insecure
bbd7eb8a 88 # (Perl doesn't know about /bin/echo)
425e5e39 89 system "echo $hid"; # Insecure
90 system "echo $data"; # Insecure until PATH set
a0d0e21e 91
425e5e39 92 $path = $ENV{'PATH'}; # $path now tainted
a0d0e21e 93
54310121 94 $ENV{'PATH'} = '/bin:/usr/bin';
c90c0ff4 95 delete @ENV{'IFS', 'CDPATH', 'ENV', 'BASH_ENV'};
a0d0e21e 96
425e5e39 97 $path = $ENV{'PATH'}; # $path now NOT tainted
98 system "echo $data"; # Is secure now!
a0d0e21e 99
425e5e39 100 open(FOO, "< $arg"); # OK - read-only file
101 open(FOO, "> $arg"); # Not OK - trying to write
a0d0e21e 102
bbd7eb8a 103 open(FOO,"echo $arg|"); # Not OK
425e5e39 104 open(FOO,"-|")
7de90c4d 105 or exec 'echo', $arg; # Also not OK
a0d0e21e 106
425e5e39 107 $shout = `echo $arg`; # Insecure, $shout now tainted
a0d0e21e 108
425e5e39 109 unlink $data, $arg; # Insecure
110 umask $arg; # Insecure
a0d0e21e 111
bbd7eb8a 112 exec "echo $arg"; # Insecure
113 exec "echo", $arg; # Insecure
114 exec "sh", '-c', $arg; # Very insecure!
a0d0e21e 115
116 @files = <*.c>; # insecure (uses readdir() or similar)
117 @files = glob('*.c'); # insecure (uses readdir() or similar)
7bac28a0 118
119 # In Perl releases older than 5.6.0 the <*.c> and glob('*.c') would
120 # have used an external program to do the filename expansion; but in
121 # either case the result is tainted since the list of filenames comes
122 # from outside of the program.
124 $bad = ($arg, 23); # $bad will be tainted
125 $arg, `true`; # Insecure (although it isn't really)
a0d0e21e 127If you try to do something insecure, you will get a fatal error saying
7de90c4d 128something like "Insecure dependency" or "Insecure $ENV{PATH}".
425e5e39 129
130=head2 Laundering and Detecting Tainted Data
132To test whether a variable contains tainted data, and whose use would
133thus trigger an "Insecure dependency" message, you can use the
134tainted() function of the Scalar::Util module, available in your
135nearby CPAN mirror, and included in Perl starting from the release 5.8.0.
136Or you may be able to use the following I<is_tainted()> function.
425e5e39 137
138 sub is_tainted {
61890e45 139 return ! eval { eval("#" . substr(join("", @_), 0, 0)); 1 };
425e5e39 140 }
142This function makes use of the fact that the presence of tainted data
143anywhere within an expression renders the entire expression tainted. It
144would be inefficient for every operator to test every argument for
145taintedness. Instead, the slightly more efficient and conservative
146approach is used that if any tainted value has been accessed within the
147same expression, the whole expression is considered tainted.
5f05dabc 149But testing for taintedness gets you only so far. Sometimes you have just
425e5e39 150to clear your data's taintedness. The only way to bypass the tainting
54310121 151mechanism is by referencing subpatterns from a regular expression match.
425e5e39 152Perl presumes that if you reference a substring using $1, $2, etc., that
153you knew what you were doing when you wrote the pattern. That means using
154a bit of thought--don't just blindly untaint anything, or you defeat the
155entire mechanism. It's better to verify that the variable has only good
156characters (for certain values of "good") rather than checking whether it
157has any bad characters. That's because it's far too easy to miss bad
158characters that you never thought of.
425e5e39 159
160Here's a test to make sure that the data contains nothing but "word"
161characters (alphabetics, numerics, and underscores), a hyphen, an at sign,
162or a dot.
54310121 164 if ($data =~ /^([-\@\w.]+)$/) {
425e5e39 165 $data = $1; # $data now untainted
166 } else {
3a2263fe 167 die "Bad data in '$data'"; # log this somewhere
425e5e39 168 }
5f05dabc 170This is fairly secure because C</\w+/> doesn't normally match shell
425e5e39 171metacharacters, nor are dot, dash, or at going to mean something special
172to the shell. Use of C</.+/> would have been insecure in theory because
173it lets everything through, but Perl doesn't check for that. The lesson
174is that when untainting, you must be exceedingly careful with your patterns.
19799a22 175Laundering data using regular expression is the I<only> mechanism for
425e5e39 176untainting dirty data, unless you use the strategy detailed below to fork
177a child of lesser privilege.
179The example does not untaint $data if C<use locale> is in effect,
180because the characters matched by C<\w> are determined by the locale.
181Perl considers that locale definitions are untrustworthy because they
182contain data from outside the program. If you are writing a
183locale-aware program, and want to launder data with a regular expression
184containing C<\w>, put C<no locale> ahead of the expression in the same
185block. See L<perllocale/SECURITY> for further discussion and examples.
187=head2 Switches On the "#!" Line
189When you make a script executable, in order to make it usable as a
190command, the system will pass switches to perl from the script's #!
54310121 191line. Perl checks that any command line switches given to a setuid
3a52c276 192(or setgid) script actually match the ones set on the #! line. Some
54310121 193Unix and Unix-like environments impose a one-switch limit on the #!
3a52c276 194line, so you may need to use something like C<-wU> instead of C<-w -U>
54310121 195under such systems. (This issue should arise only in Unix or
196Unix-like environments that support #! and setuid or setgid scripts.)
3a52c276 197
425e5e39 198=head2 Cleaning Up Your Path
1fef88e7 200For "Insecure C<$ENV{PATH}>" messages, you need to set C<$ENV{'PATH'}> to a
1e422769 201known value, and each directory in the path must be non-writable by others
202than its owner and group. You may be surprised to get this message even
203if the pathname to your executable is fully qualified. This is I<not>
204generated because you didn't supply a full path to the program; instead,
205it's generated because you never set your PATH environment variable, or
206you didn't set it to something that was safe. Because Perl can't
207guarantee that the executable in question isn't itself going to turn
208around and execute some other program that is dependent on your PATH, it
54310121 209makes sure you set the PATH.
a0d0e21e 210
211The PATH isn't the only environment variable which can cause problems.
212Because some shells may use the variables IFS, CDPATH, ENV, and
213BASH_ENV, Perl checks that those are either empty or untainted when
214starting subprocesses. You may wish to add something like this to your
215setid and taint-checking scripts.
217 delete @ENV{qw(IFS CDPATH ENV BASH_ENV)}; # Make %ENV safer
219It's also possible to get into trouble with other operations that don't
220care whether they use tainted values. Make judicious use of the file
221tests in dealing with any user-supplied filenames. When possible, do
fb73857a 222opens and such B<after> properly dropping any special user (or group!)
223privileges. Perl doesn't prevent you from opening tainted filenames for reading,
224so be careful what you print out. The tainting mechanism is intended to
225prevent stupid mistakes, not to remove the need for thought.
425e5e39 227Perl does not call the shell to expand wild cards when you pass B<system>
228and B<exec> explicit parameter lists instead of strings with possible shell
229wildcards in them. Unfortunately, the B<open>, B<glob>, and
54310121 230backtick functions provide no such alternate calling convention, so more
231subterfuge will be required.
425e5e39 232
233Perl provides a reasonably safe way to open a file or pipe from a setuid
234or setgid program: just create a child process with reduced privilege who
235does the dirty work for you. First, fork a child using the special
236B<open> syntax that connects the parent and child by a pipe. Now the
237child resets its ID set and any other per-process attributes, like
238environment variables, umasks, current working directories, back to the
239originals or known safe values. Then the child process, which no longer
240has any special permissions, does the B<open> or other system call.
241Finally, the child passes the data it managed to access back to the
5f05dabc 242parent. Because the file or pipe was opened in the child while running
425e5e39 243under less privilege than the parent, it's not apt to be tricked into
244doing something it shouldn't.
54310121 246Here's a way to do backticks reasonably safely. Notice how the B<exec> is
425e5e39 247not called with a string that the shell could expand. This is by far the
248best way to call something that might be subjected to shell escapes: just
fb73857a 249never call the shell at all.
cb1a09d0 250
a1ce9542 251 use English '-no_match_vars';
252 die "Can't fork: $!" unless defined($pid = open(KID, "-|"));
253 if ($pid) { # parent
254 while (<KID>) {
255 # do something
256 }
257 close KID;
258 } else {
259 my @temp = ($EUID, $EGID);
260 my $orig_uid = $UID;
261 my $orig_gid = $GID;
262 $EUID = $UID;
263 $EGID = $GID;
264 # Drop privileges
265 $UID = $orig_uid;
266 $GID = $orig_gid;
267 # Make sure privs are really gone
268 ($EUID, $EGID) = @temp;
269 die "Can't drop privileges"
270 unless $UID == $EUID && $GID eq $EGID;
271 $ENV{PATH} = "/bin:/usr/bin"; # Minimal PATH.
272 # Consider sanitizing the environment even more.
273 exec 'myprog', 'arg1', 'arg2'
274 or die "can't exec myprog: $!";
275 }
425e5e39 276
fb73857a 277A similar strategy would work for wildcard expansion via C<glob>, although
278you can use C<readdir> instead.
425e5e39 279
280Taint checking is most useful when although you trust yourself not to have
281written a program to give away the farm, you don't necessarily trust those
282who end up using it not to try to trick it into doing something bad. This
fb73857a 283is the kind of security checking that's useful for set-id programs and
425e5e39 284programs launched on someone else's behalf, like CGI programs.
286This is quite different, however, from not even trusting the writer of the
287code not to try to do something evil. That's the kind of trust needed
288when someone hands you a program you've never seen before and says, "Here,
289run this." For that kind of safety, check out the Safe module,
290included standard in the Perl distribution. This module allows the
291programmer to set up special compartments in which all system operations
292are trapped and namespace access is carefully controlled.
294=head2 Security Bugs
296Beyond the obvious problems that stem from giving special privileges to
fb73857a 297systems as flexible as scripts, on many versions of Unix, set-id scripts
425e5e39 298are inherently insecure right from the start. The problem is a race
299condition in the kernel. Between the time the kernel opens the file to
fb73857a 300see which interpreter to run and when the (now-set-id) interpreter turns
425e5e39 301around and reopens the file to interpret it, the file in question may have
302changed, especially if you have symbolic links on your system.
304Fortunately, sometimes this kernel "feature" can be disabled.
305Unfortunately, there are two ways to disable it. The system can simply
fb73857a 306outlaw scripts with any set-id bit set, which doesn't help much.
307Alternately, it can simply ignore the set-id bits on scripts. If the
425e5e39 308latter is true, Perl can emulate the setuid and setgid mechanism when it
309notices the otherwise useless setuid/gid bits on Perl scripts. It does
310this via a special executable called B<suidperl> that is automatically
54310121 311invoked for you if it's needed.
425e5e39 312
fb73857a 313However, if the kernel set-id script feature isn't disabled, Perl will
314complain loudly that your set-id script is insecure. You'll need to
315either disable the kernel set-id script feature, or put a C wrapper around
425e5e39 316the script. A C wrapper is just a compiled program that does nothing
317except call your Perl program. Compiled programs are not subject to the
fb73857a 318kernel bug that plagues set-id scripts. Here's a simple wrapper, written
425e5e39 319in C:
321 #define REAL_PATH "/path/to/script"
54310121 322 main(ac, av)
425e5e39 323 char **av;
324 {
325 execv(REAL_PATH, av);
54310121 326 }
cb1a09d0 327
54310121 328Compile this wrapper into a binary executable and then make I<it> rather
329than your script setuid or setgid.
425e5e39 330
425e5e39 331In recent years, vendors have begun to supply systems free of this
332inherent security bug. On such systems, when the kernel passes the name
fb73857a 333of the set-id script to open to the interpreter, rather than using a
425e5e39 334pathname subject to meddling, it instead passes I</dev/fd/3>. This is a
335special file already opened on the script, so that there can be no race
336condition for evil scripts to exploit. On these systems, Perl should be
337compiled with C<-DSETUID_SCRIPTS_ARE_SECURE_NOW>. The B<Configure>
338program that builds Perl tries to figure this out for itself, so you
339should never have to specify this yourself. Most modern releases of
340SysVr4 and BSD 4.4 use this approach to avoid the kernel race condition.
342Prior to release 5.6.1 of Perl, bugs in the code of B<suidperl> could
343introduce a security hole.
68dc0745 344
345=head2 Protecting Your Programs
347There are a number of ways to hide the source to your Perl programs,
348with varying levels of "security".
350First of all, however, you I<can't> take away read permission, because
351the source code has to be readable in order to be compiled and
352interpreted. (That doesn't mean that a CGI script's source is
353readable by people on the web, though.) So you have to leave the
354permissions at the socially friendly 0755 level. This lets
355people on your local system only see your source.
68dc0745 356
5a964f20 357Some people mistakenly regard this as a security problem. If your program does
68dc0745 358insecure things, and relies on people not knowing how to exploit those
359insecurities, it is not secure. It is often possible for someone to
360determine the insecure things and exploit them without viewing the
361source. Security through obscurity, the name for hiding your bugs
362instead of fixing them, is little security indeed.
364You can try using encryption via source filters (Filter::* from CPAN,
365or Filter::Util::Call and Filter::Simple since Perl 5.8).
366But crackers might be able to decrypt it. You can try using the byte
367code compiler and interpreter described below, but crackers might be
368able to de-compile it. You can try using the native-code compiler
68dc0745 369described below, but crackers might be able to disassemble it. These
370pose varying degrees of difficulty to people wanting to get at your
371code, but none can definitively conceal it (this is true of every
372language, not just Perl).
374If you're concerned about people profiting from your code, then the
375bottom line is that nothing but a restrictive licence will give you
376legal security. License your software and pepper it with threatening
377statements like "This is unpublished proprietary software of XYZ Corp.
378Your access to it does not give you permission to use it blah blah
379blah." You should see a lawyer to be sure your licence's wording will
380stand up in court.
5a964f20 381
382=head2 Unicode
384Unicode is a new and complex technology and one may easily overlook
385certain security pitfalls. See L<perluniintro> for an overview and
386L<perlunicode> for details, and L<perlunicode/"Security Implications
387of Unicode"> for security implications in particular.
389=head1 SEE ALSO
391L<perlrun> for its description of cleaning up environment variables.