This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
[win32] merge changes#872,873 from maintbranch
[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
35L<perllocale>), results of certain system calls (readdir, readlink,
36the gecos field of getpw* calls), and all file input are marked as
37"tainted". Tainted data may not be used directly or indirectly in any
38command that invokes a sub-shell, nor in any command that modifies
a3cb178b
GS
39files, directories, or processes. (B<Important exception>: If you pass
40a list of arguments to either C<system> or C<exec>, the elements of
41that list are B<NOT> checked for taintedness.) Any variable set
fb73857a
PP
42to a value derived from tainted data will itself be tainted,
43even if it is logically impossible for the tainted data
44to alter the variable. Because taintedness is associated with each
a034a98d 45scalar value, some elements of an array can be tainted and others not.
a0d0e21e 46
a0d0e21e
LW
47For example:
48
425e5e39
PP
49 $arg = shift; # $arg is tainted
50 $hid = $arg, 'bar'; # $hid is also tainted
51 $line = <>; # Tainted
8ebc5c01
PP
52 $line = <STDIN>; # Also tainted
53 open FOO, "/home/me/bar" or die $!;
54 $line = <FOO>; # Still tainted
a0d0e21e 55 $path = $ENV{'PATH'}; # Tainted, but see below
425e5e39 56 $data = 'abc'; # Not tainted
a0d0e21e 57
425e5e39
PP
58 system "echo $arg"; # Insecure
59 system "/bin/echo", $arg; # Secure (doesn't use sh)
60 system "echo $hid"; # Insecure
61 system "echo $data"; # Insecure until PATH set
a0d0e21e 62
425e5e39 63 $path = $ENV{'PATH'}; # $path now tainted
a0d0e21e 64
54310121 65 $ENV{'PATH'} = '/bin:/usr/bin';
c90c0ff4 66 delete @ENV{'IFS', 'CDPATH', 'ENV', 'BASH_ENV'};
a0d0e21e 67
425e5e39
PP
68 $path = $ENV{'PATH'}; # $path now NOT tainted
69 system "echo $data"; # Is secure now!
a0d0e21e 70
425e5e39
PP
71 open(FOO, "< $arg"); # OK - read-only file
72 open(FOO, "> $arg"); # Not OK - trying to write
a0d0e21e 73
425e5e39
PP
74 open(FOO,"echo $arg|"); # Not OK, but...
75 open(FOO,"-|")
76 or exec 'echo', $arg; # OK
a0d0e21e 77
425e5e39 78 $shout = `echo $arg`; # Insecure, $shout now tainted
a0d0e21e 79
425e5e39
PP
80 unlink $data, $arg; # Insecure
81 umask $arg; # Insecure
a0d0e21e 82
425e5e39
PP
83 exec "echo $arg"; # Insecure
84 exec "echo", $arg; # Secure (doesn't use the shell)
85 exec "sh", '-c', $arg; # Considered secure, alas!
a0d0e21e 86
7bac28a0
PP
87 @files = <*.c>; # Always insecure (uses csh)
88 @files = glob('*.c'); # Always insecure (uses csh)
89
a0d0e21e
LW
90If you try to do something insecure, you will get a fatal error saying
91something like "Insecure dependency" or "Insecure PATH". Note that you
425e5e39 92can still write an insecure B<system> or B<exec>, but only by explicitly
a3cb178b 93doing something like the "considered secure" example above.
425e5e39
PP
94
95=head2 Laundering and Detecting Tainted Data
96
97To test whether a variable contains tainted data, and whose use would thus
fb73857a
PP
98trigger an "Insecure dependency" message, check your nearby CPAN mirror
99for the F<Taint.pm> module, which should become available around November
1001997. Or you may be able to use the following I<is_tainted()> function.
425e5e39
PP
101
102 sub is_tainted {
54310121
PP
103 return ! eval {
104 join('',@_), kill 0;
105 1;
425e5e39
PP
106 };
107 }
108
109This function makes use of the fact that the presence of tainted data
110anywhere within an expression renders the entire expression tainted. It
111would be inefficient for every operator to test every argument for
112taintedness. Instead, the slightly more efficient and conservative
113approach is used that if any tainted value has been accessed within the
114same expression, the whole expression is considered tainted.
115
5f05dabc 116But testing for taintedness gets you only so far. Sometimes you have just
425e5e39 117to clear your data's taintedness. The only way to bypass the tainting
54310121 118mechanism is by referencing subpatterns from a regular expression match.
425e5e39
PP
119Perl presumes that if you reference a substring using $1, $2, etc., that
120you knew what you were doing when you wrote the pattern. That means using
121a bit of thought--don't just blindly untaint anything, or you defeat the
a034a98d
DD
122entire mechanism. It's better to verify that the variable has only good
123characters (for certain values of "good") rather than checking whether it
124has any bad characters. That's because it's far too easy to miss bad
125characters that you never thought of.
425e5e39
PP
126
127Here's a test to make sure that the data contains nothing but "word"
128characters (alphabetics, numerics, and underscores), a hyphen, an at sign,
129or a dot.
130
54310121 131 if ($data =~ /^([-\@\w.]+)$/) {
425e5e39
PP
132 $data = $1; # $data now untainted
133 } else {
134 die "Bad data in $data"; # log this somewhere
135 }
136
5f05dabc 137This is fairly secure because C</\w+/> doesn't normally match shell
425e5e39
PP
138metacharacters, nor are dot, dash, or at going to mean something special
139to the shell. Use of C</.+/> would have been insecure in theory because
140it lets everything through, but Perl doesn't check for that. The lesson
141is that when untainting, you must be exceedingly careful with your patterns.
142Laundering data using regular expression is the I<ONLY> mechanism for
143untainting dirty data, unless you use the strategy detailed below to fork
144a child of lesser privilege.
145
a034a98d
DD
146The example does not untaint $data if C<use locale> is in effect,
147because the characters matched by C<\w> are determined by the locale.
148Perl considers that locale definitions are untrustworthy because they
149contain data from outside the program. If you are writing a
150locale-aware program, and want to launder data with a regular expression
151containing C<\w>, put C<no locale> ahead of the expression in the same
152block. See L<perllocale/SECURITY> for further discussion and examples.
153
3a52c276
CS
154=head2 Switches On the "#!" Line
155
156When you make a script executable, in order to make it usable as a
157command, the system will pass switches to perl from the script's #!
54310121 158line. Perl checks that any command line switches given to a setuid
3a52c276 159(or setgid) script actually match the ones set on the #! line. Some
54310121 160Unix and Unix-like environments impose a one-switch limit on the #!
3a52c276 161line, so you may need to use something like C<-wU> instead of C<-w -U>
54310121
PP
162under such systems. (This issue should arise only in Unix or
163Unix-like environments that support #! and setuid or setgid scripts.)
3a52c276 164
425e5e39
PP
165=head2 Cleaning Up Your Path
166
1fef88e7 167For "Insecure C<$ENV{PATH}>" messages, you need to set C<$ENV{'PATH'}> to a
1e422769
PP
168known value, and each directory in the path must be non-writable by others
169than its owner and group. You may be surprised to get this message even
170if the pathname to your executable is fully qualified. This is I<not>
171generated because you didn't supply a full path to the program; instead,
172it's generated because you never set your PATH environment variable, or
173you didn't set it to something that was safe. Because Perl can't
174guarantee that the executable in question isn't itself going to turn
175around and execute some other program that is dependent on your PATH, it
54310121 176makes sure you set the PATH.
a0d0e21e 177
a3cb178b
GS
178The PATH isn't the only environment variable which can cause problems.
179Because some shells may use the variables IFS, CDPATH, ENV, and
180BASH_ENV, Perl checks that those are either empty or untainted when
181starting subprocesses. You may wish to add something like this to your
182setid and taint-checking scripts.
183
184 delete @ENV{qw(IFS CDPATH ENV BASH_ENV)}; # Make %ENV safer
185
a0d0e21e
LW
186It's also possible to get into trouble with other operations that don't
187care whether they use tainted values. Make judicious use of the file
188tests in dealing with any user-supplied filenames. When possible, do
fb73857a
PP
189opens and such B<after> properly dropping any special user (or group!)
190privileges. Perl doesn't prevent you from opening tainted filenames for reading,
a0d0e21e
LW
191so be careful what you print out. The tainting mechanism is intended to
192prevent stupid mistakes, not to remove the need for thought.
193
425e5e39
PP
194Perl does not call the shell to expand wild cards when you pass B<system>
195and B<exec> explicit parameter lists instead of strings with possible shell
196wildcards in them. Unfortunately, the B<open>, B<glob>, and
54310121
PP
197backtick functions provide no such alternate calling convention, so more
198subterfuge will be required.
425e5e39
PP
199
200Perl provides a reasonably safe way to open a file or pipe from a setuid
201or setgid program: just create a child process with reduced privilege who
202does the dirty work for you. First, fork a child using the special
203B<open> syntax that connects the parent and child by a pipe. Now the
204child resets its ID set and any other per-process attributes, like
205environment variables, umasks, current working directories, back to the
206originals or known safe values. Then the child process, which no longer
207has any special permissions, does the B<open> or other system call.
208Finally, the child passes the data it managed to access back to the
5f05dabc 209parent. Because the file or pipe was opened in the child while running
425e5e39
PP
210under less privilege than the parent, it's not apt to be tricked into
211doing something it shouldn't.
212
54310121 213Here's a way to do backticks reasonably safely. Notice how the B<exec> is
425e5e39
PP
214not called with a string that the shell could expand. This is by far the
215best way to call something that might be subjected to shell escapes: just
fb73857a 216never call the shell at all.
cb1a09d0 217
54310121 218 use English;
fb73857a 219 die "Can't fork: $!" unless defined $pid = open(KID, "-|");
cb1a09d0
AD
220 if ($pid) { # parent
221 while (<KID>) {
222 # do something
425e5e39 223 }
cb1a09d0
AD
224 close KID;
225 } else {
fb73857a 226 my @temp = ($EUID, $EGID);
425e5e39
PP
227 $EUID = $UID;
228 $EGID = $GID; # XXX: initgroups() not called
fb73857a
PP
229 # Make sure privs are really gone
230 ($EUID, $EGID) = @temp;
231 die "Can't drop privileges" unless
232 $UID == $EUID and
233 $GID eq $EGID; # String test
425e5e39 234 $ENV{PATH} = "/bin:/usr/bin";
fb73857a 235 exec 'myprog', 'arg1', 'arg2' or
425e5e39
PP
236 die "can't exec myprog: $!";
237 }
238
fb73857a
PP
239A similar strategy would work for wildcard expansion via C<glob>, although
240you can use C<readdir> instead.
425e5e39
PP
241
242Taint checking is most useful when although you trust yourself not to have
243written a program to give away the farm, you don't necessarily trust those
244who end up using it not to try to trick it into doing something bad. This
fb73857a 245is the kind of security checking that's useful for set-id programs and
425e5e39
PP
246programs launched on someone else's behalf, like CGI programs.
247
248This is quite different, however, from not even trusting the writer of the
249code not to try to do something evil. That's the kind of trust needed
250when someone hands you a program you've never seen before and says, "Here,
251run this." For that kind of safety, check out the Safe module,
252included standard in the Perl distribution. This module allows the
253programmer to set up special compartments in which all system operations
254are trapped and namespace access is carefully controlled.
255
256=head2 Security Bugs
257
258Beyond the obvious problems that stem from giving special privileges to
fb73857a 259systems as flexible as scripts, on many versions of Unix, set-id scripts
425e5e39
PP
260are inherently insecure right from the start. The problem is a race
261condition in the kernel. Between the time the kernel opens the file to
fb73857a 262see which interpreter to run and when the (now-set-id) interpreter turns
425e5e39
PP
263around and reopens the file to interpret it, the file in question may have
264changed, especially if you have symbolic links on your system.
265
266Fortunately, sometimes this kernel "feature" can be disabled.
267Unfortunately, there are two ways to disable it. The system can simply
fb73857a
PP
268outlaw scripts with any set-id bit set, which doesn't help much.
269Alternately, it can simply ignore the set-id bits on scripts. If the
425e5e39
PP
270latter is true, Perl can emulate the setuid and setgid mechanism when it
271notices the otherwise useless setuid/gid bits on Perl scripts. It does
272this via a special executable called B<suidperl> that is automatically
54310121 273invoked for you if it's needed.
425e5e39 274
fb73857a
PP
275However, if the kernel set-id script feature isn't disabled, Perl will
276complain loudly that your set-id script is insecure. You'll need to
277either disable the kernel set-id script feature, or put a C wrapper around
425e5e39
PP
278the script. A C wrapper is just a compiled program that does nothing
279except call your Perl program. Compiled programs are not subject to the
fb73857a 280kernel bug that plagues set-id scripts. Here's a simple wrapper, written
425e5e39
PP
281in C:
282
283 #define REAL_PATH "/path/to/script"
54310121 284 main(ac, av)
425e5e39
PP
285 char **av;
286 {
287 execv(REAL_PATH, av);
54310121 288 }
cb1a09d0 289
54310121
PP
290Compile this wrapper into a binary executable and then make I<it> rather
291than your script setuid or setgid.
425e5e39
PP
292
293See the program B<wrapsuid> in the F<eg> directory of your Perl
294distribution for a convenient way to do this automatically for all your
295setuid Perl programs. It moves setuid scripts into files with the same
296name plus a leading dot, and then compiles a wrapper like the one above
297for each of them.
298
299In recent years, vendors have begun to supply systems free of this
300inherent security bug. On such systems, when the kernel passes the name
fb73857a 301of the set-id script to open to the interpreter, rather than using a
425e5e39
PP
302pathname subject to meddling, it instead passes I</dev/fd/3>. This is a
303special file already opened on the script, so that there can be no race
304condition for evil scripts to exploit. On these systems, Perl should be
305compiled with C<-DSETUID_SCRIPTS_ARE_SECURE_NOW>. The B<Configure>
306program that builds Perl tries to figure this out for itself, so you
307should never have to specify this yourself. Most modern releases of
308SysVr4 and BSD 4.4 use this approach to avoid the kernel race condition.
309
310Prior to release 5.003 of Perl, a bug in the code of B<suidperl> could
311introduce a security hole in systems compiled with strict POSIX
312compliance.
68dc0745
PP
313
314=head2 Protecting Your Programs
315
316There are a number of ways to hide the source to your Perl programs,
317with varying levels of "security".
318
319First of all, however, you I<can't> take away read permission, because
320the source code has to be readable in order to be compiled and
321interpreted. (That doesn't mean that a CGI script's source is
322readable by people on the web, though.) So you have to leave the
323permissions at the socially friendly 0755 level.
324
325Some people regard this as a security problem. If your program does
326insecure things, and relies on people not knowing how to exploit those
327insecurities, it is not secure. It is often possible for someone to
328determine the insecure things and exploit them without viewing the
329source. Security through obscurity, the name for hiding your bugs
330instead of fixing them, is little security indeed.
331
332You can try using encryption via source filters (Filter::* from CPAN).
333But crackers might be able to decrypt it. You can try using the
54310121 334byte code compiler and interpreter described below, but crackers might
68dc0745
PP
335be able to de-compile it. You can try using the native-code compiler
336described below, but crackers might be able to disassemble it. These
337pose varying degrees of difficulty to people wanting to get at your
338code, but none can definitively conceal it (this is true of every
339language, not just Perl).
340
341If you're concerned about people profiting from your code, then the
342bottom line is that nothing but a restrictive licence will give you
343legal security. License your software and pepper it with threatening
344statements like "This is unpublished proprietary software of XYZ Corp.
345Your access to it does not give you permission to use it blah blah
346blah." You should see a lawyer to be sure your licence's wording will
347stand up in court.