3 perlipc - Perl interprocess communication (signals, fifos, pipes, safe subprocesses, sockets, and semaphores)
7 The basic IPC facilities of Perl are built out of the good old Unix
8 signals, named pipes, pipe opens, the Berkeley socket routines, and SysV
9 IPC calls. Each is used in slightly different situations.
13 Perl uses a simple signal handling model: the %SIG hash contains names
14 or references of user-installed signal handlers. These handlers will
15 be called with an argument which is the name of the signal that
16 triggered it. A signal may be generated intentionally from a
17 particular keyboard sequence like control-C or control-Z, sent to you
18 from another process, or triggered automatically by the kernel when
19 special events transpire, like a child process exiting, your own process
20 running out of stack space, or hitting a process file-size limit.
22 For example, to trap an interrupt signal, set up a handler like this:
29 die "Somebody sent me a SIG$signame";
31 $SIG{INT} = __PACKAGE__ . "::catch_zap";
32 $SIG{INT} = \&catch_zap; # best strategy
34 Prior to Perl 5.7.3 it was necessary to do as little as you possibly
35 could in your handler; notice how all we do is set a global variable
36 and then raise an exception. That's because on most systems,
37 libraries are not re-entrant; particularly, memory allocation and I/O
38 routines are not. That meant that doing nearly I<anything> in your
39 handler could in theory trigger a memory fault and subsequent core
40 dump - see L</Deferred Signals (Safe Signals)> below.
42 The names of the signals are the ones listed out by C<kill -l> on your
43 system, or you can retrieve them using the CPAN module L<IPC::Signal>.
45 You may also choose to assign the strings C<"IGNORE"> or C<"DEFAULT"> as
46 the handler, in which case Perl will try to discard the signal or do the
49 On most Unix platforms, the C<CHLD> (sometimes also known as C<CLD>) signal
50 has special behavior with respect to a value of C<"IGNORE">.
51 Setting C<$SIG{CHLD}> to C<"IGNORE"> on such a platform has the effect of
52 not creating zombie processes when the parent process fails to C<wait()>
53 on its child processes (i.e., child processes are automatically reaped).
54 Calling C<wait()> with C<$SIG{CHLD}> set to C<"IGNORE"> usually returns
55 C<-1> on such platforms.
57 Some signals can be neither trapped nor ignored, such as the KILL and STOP
58 (but not the TSTP) signals. Note that ignoring signals makes them disappear.
59 If you only want them blocked temporarily without them getting lost you'll
60 have to use POSIX' sigprocmask.
62 Sending a signal to a negative process ID means that you send the signal
63 to the entire Unix process group. This code sends a hang-up signal to all
64 processes in the current process group, and also sets $SIG{HUP} to C<"IGNORE">
65 so it doesn't kill itself:
67 # block scope for local
69 local $SIG{HUP} = "IGNORE";
71 # snazzy writing of: kill("HUP", -$$)
74 Another interesting signal to send is signal number zero. This doesn't
75 actually affect a child process, but instead checks whether it's alive
76 or has changed its UIDs.
78 unless (kill 0 => $kid_pid) {
79 warn "something wicked happened to $kid_pid";
82 Signal number zero may fail because you lack permission to send the
83 signal when directed at a process whose real or saved UID is not
84 identical to the real or effective UID of the sending process, even
85 though the process is alive. You may be able to determine the cause of
86 failure using C<$!> or C<%!>.
88 unless (kill(0 => $pid) || $!{EPERM}) {
89 warn "$pid looks dead";
92 You might also want to employ anonymous functions for simple signal
95 $SIG{INT} = sub { die "\nOutta here!\n" };
97 SIGCHLD handlers require some special care. If a second child dies
98 while in the signal handler caused by the first death, we won't get
99 another signal. So must loop here else we will leave the unreaped child
100 as a zombie. And the next time two children die we get another zombie.
103 use POSIX ":sys_wait_h";
105 while ((my $child = waitpid(-1, WNOHANG)) > 0) {
106 $Kid_Status{$child} = $?;
109 # do something that forks...
111 Be careful: qx(), system(), and some modules for calling external commands
112 do a fork(), then wait() for the result. Thus, your signal handler
113 will be called. Because wait() was already called by system() or qx(),
114 the wait() in the signal handler will see no more zombies and will
117 The best way to prevent this issue is to use waitpid(), as in the following
120 use POSIX ":sys_wait_h"; # for nonblocking read
125 # don't change $! and $? outside handler
127 my $pid = waitpid(-1, WNOHANG);
128 return if $pid == -1;
129 return unless defined $children{$pid};
130 delete $children{$pid};
131 cleanup_child($pid, $?);
136 die "cannot fork" unless defined $pid;
148 Signal handling is also used for timeouts in Unix. While safely
149 protected within an C<eval{}> block, you set a signal handler to trap
150 alarm signals and then schedule to have one delivered to you in some
151 number of seconds. Then try your blocking operation, clearing the alarm
152 when it's done but not before you've exited your C<eval{}> block. If it
153 goes off, you'll use die() to jump out of the block.
157 my $ALARM_EXCEPTION = "alarm clock restart";
159 local $SIG{ALRM} = sub { die $ALARM_EXCEPTION };
161 flock(FH, 2) # blocking write lock
162 || die "cannot flock: $!";
165 if ($@ && $@ !~ quotemeta($ALARM_EXCEPTION)) { die }
167 If the operation being timed out is system() or qx(), this technique
168 is liable to generate zombies. If this matters to you, you'll
169 need to do your own fork() and exec(), and kill the errant child process.
171 For more complex signal handling, you might see the standard POSIX
172 module. Lamentably, this is almost entirely undocumented, but
173 the F<t/lib/posix.t> file from the Perl source distribution has some
176 =head2 Handling the SIGHUP Signal in Daemons
178 A process that usually starts when the system boots and shuts down
179 when the system is shut down is called a daemon (Disk And Execution
180 MONitor). If a daemon process has a configuration file which is
181 modified after the process has been started, there should be a way to
182 tell that process to reread its configuration file without stopping
183 the process. Many daemons provide this mechanism using a C<SIGHUP>
184 signal handler. When you want to tell the daemon to reread the file,
185 simply send it the C<SIGHUP> signal.
187 The following example implements a simple daemon, which restarts
188 itself every time the C<SIGHUP> signal is received. The actual code is
189 located in the subroutine C<code()>, which just prints some debugging
190 info to show that it works; it should be replaced with the real code.
196 use File::Basename ();
197 use File::Spec::Functions;
201 # make the daemon cross-platform, so exec always calls the script
202 # itself with the right path, no matter how the script was invoked.
203 my $script = File::Basename::basename($0);
204 my $SELF = catfile($FindBin::Bin, $script);
206 # POSIX unmasks the sigprocmask properly
208 print "got SIGHUP\n";
209 exec($SELF, @ARGV) || die "$0: couldn't restart: $!";
216 print "ARGV: @ARGV\n";
225 =head2 Deferred Signals (Safe Signals)
227 Before Perl 5.7.3, installing Perl code to deal with signals exposed you to
228 danger from two things. First, few system library functions are
229 re-entrant. If the signal interrupts while Perl is executing one function
230 (like malloc(3) or printf(3)), and your signal handler then calls the same
231 function again, you could get unpredictable behavior--often, a core dump.
232 Second, Perl isn't itself re-entrant at the lowest levels. If the signal
233 interrupts Perl while Perl is changing its own internal data structures,
234 similarly unpredictable behavior may result.
236 There were two things you could do, knowing this: be paranoid or be
237 pragmatic. The paranoid approach was to do as little as possible in your
238 signal handler. Set an existing integer variable that already has a
239 value, and return. This doesn't help you if you're in a slow system call,
240 which will just restart. That means you have to C<die> to longjmp(3) out
241 of the handler. Even this is a little cavalier for the true paranoiac,
242 who avoids C<die> in a handler because the system I<is> out to get you.
243 The pragmatic approach was to say "I know the risks, but prefer the
244 convenience", and to do anything you wanted in your signal handler,
245 and be prepared to clean up core dumps now and again.
247 Perl 5.7.3 and later avoid these problems by "deferring" signals. That is,
248 when the signal is delivered to the process by the system (to the C code
249 that implements Perl) a flag is set, and the handler returns immediately.
250 Then at strategic "safe" points in the Perl interpreter (e.g. when it is
251 about to execute a new opcode) the flags are checked and the Perl level
252 handler from %SIG is executed. The "deferred" scheme allows much more
253 flexibility in the coding of signal handlers as we know the Perl
254 interpreter is in a safe state, and that we are not in a system library
255 function when the handler is called. However the implementation does
256 differ from previous Perls in the following ways:
260 =item Long-running opcodes
262 As the Perl interpreter looks at signal flags only when it is about
263 to execute a new opcode, a signal that arrives during a long-running
264 opcode (e.g. a regular expression operation on a very large string) will
265 not be seen until the current opcode completes.
267 If a signal of any given type fires multiple times during an opcode
268 (such as from a fine-grained timer), the handler for that signal will
269 be called only once, after the opcode completes; all other
270 instances will be discarded. Furthermore, if your system's signal queue
271 gets flooded to the point that there are signals that have been raised
272 but not yet caught (and thus not deferred) at the time an opcode
273 completes, those signals may well be caught and deferred during
274 subsequent opcodes, with sometimes surprising results. For example, you
275 may see alarms delivered even after calling C<alarm(0)> as the latter
276 stops the raising of alarms but does not cancel the delivery of alarms
277 raised but not yet caught. Do not depend on the behaviors described in
278 this paragraph as they are side effects of the current implementation and
279 may change in future versions of Perl.
281 =item Interrupting IO
283 When a signal is delivered (e.g., SIGINT from a control-C) the operating
284 system breaks into IO operations like I<read>(2), which is used to
285 implement Perl's readline() function, the C<< <> >> operator. On older
286 Perls the handler was called immediately (and as C<read> is not "unsafe",
287 this worked well). With the "deferred" scheme the handler is I<not> called
288 immediately, and if Perl is using the system's C<stdio> library that
289 library may restart the C<read> without returning to Perl to give it a
290 chance to call the %SIG handler. If this happens on your system the
291 solution is to use the C<:perlio> layer to do IO--at least on those handles
292 that you want to be able to break into with signals. (The C<:perlio> layer
293 checks the signal flags and calls %SIG handlers before resuming IO
296 The default in Perl 5.7.3 and later is to automatically use
297 the C<:perlio> layer.
299 Note that it is not advisable to access a file handle within a signal
300 handler where that signal has interrupted an I/O operation on that same
301 handle. While perl will at least try hard not to crash, there are no
302 guarantees of data integrity; for example, some data might get dropped or
305 Some networking library functions like gethostbyname() are known to have
306 their own implementations of timeouts which may conflict with your
307 timeouts. If you have problems with such functions, try using the POSIX
308 sigaction() function, which bypasses Perl safe signals. Be warned that
309 this does subject you to possible memory corruption, as described above.
311 Instead of setting C<$SIG{ALRM}>:
313 local $SIG{ALRM} = sub { die "alarm" };
315 try something like the following:
317 use POSIX qw(SIGALRM);
318 POSIX::sigaction(SIGALRM, POSIX::SigAction->new(sub { die "alarm" }))
319 || die "Error setting SIGALRM handler: $!\n";
321 Another way to disable the safe signal behavior locally is to use
322 the C<Perl::Unsafe::Signals> module from CPAN, which affects
325 =item Restartable system calls
327 On systems that supported it, older versions of Perl used the
328 SA_RESTART flag when installing %SIG handlers. This meant that
329 restartable system calls would continue rather than returning when
330 a signal arrived. In order to deliver deferred signals promptly,
331 Perl 5.7.3 and later do I<not> use SA_RESTART. Consequently,
332 restartable system calls can fail (with $! set to C<EINTR>) in places
333 where they previously would have succeeded.
335 The default C<:perlio> layer retries C<read>, C<write>
336 and C<close> as described above; interrupted C<wait> and
337 C<waitpid> calls will always be retried.
339 =item Signals as "faults"
341 Certain signals like SEGV, ILL, and BUS are generated by virtual memory
342 addressing errors and similar "faults". These are normally fatal: there is
343 little a Perl-level handler can do with them. So Perl delivers them
344 immediately rather than attempting to defer them.
346 =item Signals triggered by operating system state
348 On some operating systems certain signal handlers are supposed to "do
349 something" before returning. One example can be CHLD or CLD, which
350 indicates a child process has completed. On some operating systems the
351 signal handler is expected to C<wait> for the completed child
352 process. On such systems the deferred signal scheme will not work for
353 those signals: it does not do the C<wait>. Again the failure will
354 look like a loop as the operating system will reissue the signal because
355 there are completed child processes that have not yet been C<wait>ed for.
359 If you want the old signal behavior back despite possible
360 memory corruption, set the environment variable C<PERL_SIGNALS> to
361 C<"unsafe">. This feature first appeared in Perl 5.8.1.
365 A named pipe (often referred to as a FIFO) is an old Unix IPC
366 mechanism for processes communicating on the same machine. It works
367 just like regular anonymous pipes, except that the
368 processes rendezvous using a filename and need not be related.
370 To create a named pipe, use the C<POSIX::mkfifo()> function.
372 use POSIX qw(mkfifo);
373 mkfifo($path, 0700) || die "mkfifo $path failed: $!";
375 You can also use the Unix command mknod(1), or on some
376 systems, mkfifo(1). These may not be in your normal path, though.
378 # system return val is backwards, so && not ||
380 $ENV{PATH} .= ":/etc:/usr/etc";
381 if ( system("mknod", $path, "p")
382 && system("mkfifo", $path) )
384 die "mk{nod,fifo} $path failed";
388 A fifo is convenient when you want to connect a process to an unrelated
389 one. When you open a fifo, the program will block until there's something
392 For example, let's say you'd like to have your F<.signature> file be a
393 named pipe that has a Perl program on the other end. Now every time any
394 program (like a mailer, news reader, finger program, etc.) tries to read
395 from that file, the reading program will read the new signature from your
396 program. We'll use the pipe-checking file-test operator, B<-p>, to find
397 out whether anyone (or anything) has accidentally removed our fifo.
400 my $FIFO = ".signature";
404 unlink $FIFO; # discard any failure, will catch later
405 require POSIX; # delayed loading of heavy module
406 POSIX::mkfifo($FIFO, 0700)
407 || die "can't mkfifo $FIFO: $!";
410 # next line blocks till there's a reader
411 open (FIFO, "> $FIFO") || die "can't open $FIFO: $!";
412 print FIFO "John Smith (smith\@host.org)\n", `fortune -s`;
413 close(FIFO) || die "can't close $FIFO: $!";
414 sleep 2; # to avoid dup signals
417 =head1 Using open() for IPC
419 Perl's basic open() statement can also be used for unidirectional
420 interprocess communication by either appending or prepending a pipe
421 symbol to the second argument to open(). Here's how to start
422 something up in a child process you intend to write to:
424 open(SPOOLER, "| cat -v | lpr -h 2>/dev/null")
425 || die "can't fork: $!";
426 local $SIG{PIPE} = sub { die "spooler pipe broke" };
427 print SPOOLER "stuff\n";
428 close SPOOLER || die "bad spool: $! $?";
430 And here's how to start up a child process you intend to read from:
432 open(STATUS, "netstat -an 2>&1 |")
433 || die "can't fork: $!";
435 next if /^(tcp|udp)/;
438 close STATUS || die "bad netstat: $! $?";
440 If one can be sure that a particular program is a Perl script expecting
441 filenames in @ARGV, the clever programmer can write something like this:
443 % program f1 "cmd1|" - f2 "cmd2|" f3 < tmpfile
445 and no matter which sort of shell it's called from, the Perl program will
446 read from the file F<f1>, the process F<cmd1>, standard input (F<tmpfile>
447 in this case), the F<f2> file, the F<cmd2> command, and finally the F<f3>
448 file. Pretty nifty, eh?
450 You might notice that you could use backticks for much the
451 same effect as opening a pipe for reading:
453 print grep { !/^(tcp|udp)/ } `netstat -an 2>&1`;
454 die "bad netstatus ($?)" if $?;
456 While this is true on the surface, it's much more efficient to process the
457 file one line or record at a time because then you don't have to read the
458 whole thing into memory at once. It also gives you finer control of the
459 whole process, letting you kill off the child process early if you'd like.
461 Be careful to check the return values from both open() and close(). If
462 you're I<writing> to a pipe, you should also trap SIGPIPE. Otherwise,
463 think of what happens when you start up a pipe to a command that doesn't
464 exist: the open() will in all likelihood succeed (it only reflects the
465 fork()'s success), but then your output will fail--spectacularly. Perl
466 can't know whether the command worked, because your command is actually
467 running in a separate process whose exec() might have failed. Therefore,
468 while readers of bogus commands return just a quick EOF, writers
469 to bogus commands will get hit with a signal, which they'd best be prepared
472 open(FH, "|bogus") || die "can't fork: $!";
473 print FH "bang\n"; # neither necessary nor sufficient
474 # to check print retval!
475 close(FH) || die "can't close: $!";
477 The reason for not checking the return value from print() is because of
478 pipe buffering; physical writes are delayed. That won't blow up until the
479 close, and it will blow up with a SIGPIPE. To catch it, you could use
482 $SIG{PIPE} = "IGNORE";
483 open(FH, "|bogus") || die "can't fork: $!";
485 close(FH) || die "can't close: status=$?";
489 Both the main process and any child processes it forks share the same
490 STDIN, STDOUT, and STDERR filehandles. If both processes try to access
491 them at once, strange things can happen. You may also want to close
492 or reopen the filehandles for the child. You can get around this by
493 opening your pipe with open(), but on some systems this means that the
494 child process cannot outlive the parent.
496 =head2 Background Processes
498 You can run a command in the background with:
502 The command's STDOUT and STDERR (and possibly STDIN, depending on your
503 shell) will be the same as the parent's. You won't need to catch
504 SIGCHLD because of the double-fork taking place; see below for details.
506 =head2 Complete Dissociation of Child from Parent
508 In some cases (starting server processes, for instance) you'll want to
509 completely dissociate the child process from the parent. This is
510 often called daemonization. A well-behaved daemon will also chdir()
511 to the root directory so it doesn't prevent unmounting the filesystem
512 containing the directory from which it was launched, and redirect its
513 standard file descriptors from and to F</dev/null> so that random
514 output doesn't wind up on the user's terminal.
519 chdir("/") || die "can't chdir to /: $!";
520 open(STDIN, "< /dev/null") || die "can't read /dev/null: $!";
521 open(STDOUT, "> /dev/null") || die "can't write to /dev/null: $!";
522 defined(my $pid = fork()) || die "can't fork: $!";
523 exit if $pid; # non-zero now means I am the parent
524 (setsid() != -1) || die "Can't start a new session: $!"
525 open(STDERR, ">&STDOUT") || die "can't dup stdout: $!";
528 The fork() has to come before the setsid() to ensure you aren't a
529 process group leader; the setsid() will fail if you are. If your
530 system doesn't have the setsid() function, open F</dev/tty> and use the
531 C<TIOCNOTTY> ioctl() on it instead. See tty(4) for details.
533 Non-Unix users should check their C<< I<Your_OS>::Process >> module for
534 other possible solutions.
536 =head2 Safe Pipe Opens
538 Another interesting approach to IPC is making your single program go
539 multiprocess and communicate between--or even amongst--yourselves. The
540 open() function will accept a file argument of either C<"-|"> or C<"|-">
541 to do a very interesting thing: it forks a child connected to the
542 filehandle you've opened. The child is running the same program as the
543 parent. This is useful for safely opening a file when running under an
544 assumed UID or GID, for example. If you open a pipe I<to> minus, you can
545 write to the filehandle you opened and your kid will find it in I<his>
546 STDIN. If you open a pipe I<from> minus, you can read from the filehandle
547 you opened whatever your kid writes to I<his> STDOUT.
549 use English qw[ -no_match_vars ];
550 my $PRECIOUS = "/path/to/some/safe/file";
555 $pid = open(KID_TO_WRITE, "|-");
556 unless (defined $pid) {
557 warn "cannot fork: $!";
558 die "bailing out" if $sleep_count++ > 6;
561 } until defined $pid;
563 if ($pid) { # I am the parent
564 print KID_TO_WRITE @some_data;
565 close(KID_TO_WRITE) || warn "kid exited $?";
566 } else { # I am the child
567 # drop permissions in setuid and/or setgid programs:
568 ($EUID, $EGID) = ($UID, $GID);
569 open (OUTFILE, "> $PRECIOUS")
570 || die "can't open $PRECIOUS: $!";
572 print OUTFILE; # child's STDIN is parent's KID_TO_WRITE
574 close(OUTFILE) || die "can't close $PRECIOUS: $!";
575 exit(0); # don't forget this!!
578 Another common use for this construct is when you need to execute
579 something without the shell's interference. With system(), it's
580 straightforward, but you can't use a pipe open or backticks safely.
581 That's because there's no way to stop the shell from getting its hands on
582 your arguments. Instead, use lower-level control to call exec() directly.
584 Here's a safe backtick or pipe open for read:
586 my $pid = open(KID_TO_READ, "-|");
587 defined($pid) || die "can't fork: $!";
590 while (<KID_TO_READ>) {
591 # do something interesting
593 close(KID_TO_READ) || warn "kid exited $?";
596 ($EUID, $EGID) = ($UID, $GID); # suid only
597 exec($program, @options, @args)
598 || die "can't exec program: $!";
602 And here's a safe pipe open for writing:
604 my $pid = open(KID_TO_WRITE, "|-");
605 defined($pid) || die "can't fork: $!";
607 $SIG{PIPE} = sub { die "whoops, $program pipe broke" };
610 print KID_TO_WRITE @data;
611 close(KID_TO_WRITE) || warn "kid exited $?";
614 ($EUID, $EGID) = ($UID, $GID);
615 exec($program, @options, @args)
616 || die "can't exec program: $!";
620 It is very easy to dead-lock a process using this form of open(), or
621 indeed with any use of pipe() with multiple subprocesses. The
622 example above is "safe" because it is simple and calls exec(). See
623 L</"Avoiding Pipe Deadlocks"> for general safety principles, but there
624 are extra gotchas with Safe Pipe Opens.
626 In particular, if you opened the pipe using C<open FH, "|-">, then you
627 cannot simply use close() in the parent process to close an unwanted
628 writer. Consider this code:
630 my $pid = open(WRITER, "|-"); # fork open a kid
631 defined($pid) || die "first fork failed: $!";
633 if (my $sub_pid = fork()) {
634 defined($sub_pid) || die "second fork failed: $!";
635 close(WRITER) || die "couldn't close WRITER: $!";
636 # now do something else...
639 # first write to WRITER
642 close(WRITER) || die "couldn't close WRITER: $!";
647 # first do something with STDIN, then
651 In the example above, the true parent does not want to write to the WRITER
652 filehandle, so it closes it. However, because WRITER was opened using
653 C<open FH, "|-">, it has a special behavior: closing it calls
654 waitpid() (see L<perlfunc/waitpid>), which waits for the subprocess
655 to exit. If the child process ends up waiting for something happening
656 in the section marked "do something else", you have deadlock.
658 This can also be a problem with intermediate subprocesses in more
659 complicated code, which will call waitpid() on all open filehandles
660 during global destruction--in no predictable order.
662 To solve this, you must manually use pipe(), fork(), and the form of
663 open() which sets one file descriptor to another, as shown below:
665 pipe(READER, WRITER) || die "pipe failed: $!";
667 defined($pid) || die "first fork failed: $!";
670 if (my $sub_pid = fork()) {
671 defined($sub_pid) || die "first fork failed: $!";
672 close(WRITER) || die "can't close WRITER: $!";
678 close(WRITER) || die "can't close WRITER: $!";
684 open(STDIN, "<&READER") || die "can't reopen STDIN: $!";
685 close(WRITER) || die "can't close WRITER: $!";
690 Since Perl 5.8.0, you can also use the list form of C<open> for pipes.
691 This is preferred when you wish to avoid having the shell interpret
692 metacharacters that may be in your command string.
694 So for example, instead of using:
696 open(PS_PIPE, "ps aux|") || die "can't open ps pipe: $!";
698 One would use either of these:
700 open(PS_PIPE, "-|", "ps", "aux")
701 || die "can't open ps pipe: $!";
703 @ps_args = qw[ ps aux ];
704 open(PS_PIPE, "-|", @ps_args)
705 || die "can't open @ps_args|: $!";
707 Because there are more than three arguments to open(), forks the ps(1)
708 command I<without> spawning a shell, and reads its standard output via the
709 C<PS_PIPE> filehandle. The corresponding syntax to I<write> to command
710 pipes is to use C<"|-"> in place of C<"-|">.
712 This was admittedly a rather silly example, because you're using string
713 literals whose content is perfectly safe. There is therefore no cause to
714 resort to the harder-to-read, multi-argument form of pipe open(). However,
715 whenever you cannot be assured that the program arguments are free of shell
716 metacharacters, the fancier form of open() should be used. For example:
718 @grep_args = ("egrep", "-i", $some_pattern, @many_files);
719 open(GREP_PIPE, "-|", @grep_args)
720 || die "can't open @grep_args|: $!";
722 Here the multi-argument form of pipe open() is preferred because the
723 pattern and indeed even the filenames themselves might hold metacharacters.
725 Be aware that these operations are full Unix forks, which means they may
726 not be correctly implemented on all alien systems. Additionally, these are
727 not true multithreading. To learn more about threading, see the F<modules>
728 file mentioned below in the SEE ALSO section.
730 =head2 Avoiding Pipe Deadlocks
732 Whenever you have more than one subprocess, you must be careful that each
733 closes whichever half of any pipes created for interprocess communication
734 it is not using. This is because any child process reading from the pipe
735 and expecting an EOF will never receive it, and therefore never exit. A
736 single process closing a pipe is not enough to close it; the last process
737 with the pipe open must close it for it to read EOF.
739 Certain built-in Unix features help prevent this most of the time. For
740 instance, filehandles have a "close on exec" flag, which is set I<en masse>
741 under control of the C<$^F> variable. This is so any filehandles you
742 didn't explicitly route to the STDIN, STDOUT or STDERR of a child
743 I<program> will be automatically closed.
745 Always explicitly and immediately call close() on the writable end of any
746 pipe, unless that process is actually writing to it. Even if you don't
747 explicitly call close(), Perl will still close() all filehandles during
748 global destruction. As previously discussed, if those filehandles have
749 been opened with Safe Pipe Open, this will result in calling waitpid(),
750 which may again deadlock.
752 =head2 Bidirectional Communication with Another Process
754 While this works reasonably well for unidirectional communication, what
755 about bidirectional communication? The most obvious approach doesn't work:
757 # THIS DOES NOT WORK!!
758 open(PROG_FOR_READING_AND_WRITING, "| some program |")
760 If you forget to C<use warnings>, you'll miss out entirely on the
761 helpful diagnostic message:
763 Can't do bidirectional pipe at -e line 1.
765 If you really want to, you can use the standard open2() from the
766 C<IPC::Open2> module to catch both ends. There's also an open3() in
767 C<IPC::Open3> for tridirectional I/O so you can also catch your child's
768 STDERR, but doing so would then require an awkward select() loop and
769 wouldn't allow you to use normal Perl input operations.
771 If you look at its source, you'll see that open2() uses low-level
772 primitives like the pipe() and exec() syscalls to create all the
773 connections. Although it might have been more efficient by using
774 socketpair(), this would have been even less portable than it already
775 is. The open2() and open3() functions are unlikely to work anywhere
776 except on a Unix system, or at least one purporting POSIX compliance.
779 Hold on, is this even true? First it says that socketpair() is avoided
780 for portability, but then it says it probably won't work except on
781 Unixy systems anyway. Which one of those is true?
783 Here's an example of using open2():
787 $pid = open2(*Reader, *Writer, "cat -un");
788 print Writer "stuff\n";
791 The problem with this is that buffering is really going to ruin your
792 day. Even though your C<Writer> filehandle is auto-flushed so the process
793 on the other end gets your data in a timely manner, you can't usually do
794 anything to force that process to give its data to you in a similarly quick
795 fashion. In this special case, we could actually so, because we gave
796 I<cat> a B<-u> flag to make it unbuffered. But very few commands are
797 designed to operate over pipes, so this seldom works unless you yourself
798 wrote the program on the other end of the double-ended pipe.
800 A solution to this is to use a library which uses pseudottys to make your
801 program behave more reasonably. This way you don't have to have control
802 over the source code of the program you're using. The C<Expect> module
803 from CPAN also addresses this kind of thing. This module requires two
804 other modules from CPAN, C<IO::Pty> and C<IO::Stty>. It sets up a pseudo
805 terminal to interact with programs that insist on talking to the terminal
806 device driver. If your system is supported, this may be your best bet.
808 =head2 Bidirectional Communication with Yourself
810 If you want, you may make low-level pipe() and fork() syscalls to stitch
811 this together by hand. This example only talks to itself, but you could
812 reopen the appropriate handles to STDIN and STDOUT and call other processes.
813 (The following example lacks proper error checking.)
816 # pipe1 - bidirectional communication using two pipe pairs
817 # designed for the socketpair-challenged
818 use IO::Handle; # thousands of lines just for autoflush :-(
819 pipe(PARENT_RDR, CHILD_WTR); # XXX: check failure?
820 pipe(CHILD_RDR, PARENT_WTR); # XXX: check failure?
821 CHILD_WTR->autoflush(1);
822 PARENT_WTR->autoflush(1);
827 print CHILD_WTR "Parent Pid $$ is sending this\n";
828 chomp($line = <CHILD_RDR>);
829 print "Parent Pid $$ just read this: '$line'\n";
830 close CHILD_RDR; close CHILD_WTR;
833 die "cannot fork: $!" unless defined $pid;
836 chomp($line = <PARENT_RDR>);
837 print "Child Pid $$ just read this: '$line'\n";
838 print PARENT_WTR "Child Pid $$ is sending this\n";
844 But you don't actually have to make two pipe calls. If you
845 have the socketpair() system call, it will do this all for you.
848 # pipe2 - bidirectional communication using socketpair
849 # "the best ones always go both ways"
852 use IO::Handle; # thousands of lines just for autoflush :-(
854 # We say AF_UNIX because although *_LOCAL is the
855 # POSIX 1003.1g form of the constant, many machines
856 # still don't have it.
857 socketpair(CHILD, PARENT, AF_UNIX, SOCK_STREAM, PF_UNSPEC)
858 || die "socketpair: $!";
861 PARENT->autoflush(1);
865 print CHILD "Parent Pid $$ is sending this\n";
866 chomp($line = <CHILD>);
867 print "Parent Pid $$ just read this: '$line'\n";
871 die "cannot fork: $!" unless defined $pid;
873 chomp($line = <PARENT>);
874 print "Child Pid $$ just read this: '$line'\n";
875 print PARENT "Child Pid $$ is sending this\n";
880 =head1 Sockets: Client/Server Communication
882 While not entirely limited to Unix-derived operating systems (e.g., WinSock
883 on PCs provides socket support, as do some VMS libraries), you might not have
884 sockets on your system, in which case this section probably isn't going to
885 do you much good. With sockets, you can do both virtual circuits like TCP
886 streams and datagrams like UDP packets. You may be able to do even more
887 depending on your system.
889 The Perl functions for dealing with sockets have the same names as
890 the corresponding system calls in C, but their arguments tend to differ
891 for two reasons. First, Perl filehandles work differently than C file
892 descriptors. Second, Perl already knows the length of its strings, so you
893 don't need to pass that information.
895 One of the major problems with ancient, antemillennial socket code in Perl
896 was that it used hard-coded values for some of the constants, which
897 severely hurt portability. If you ever see code that does anything like
898 explicitly setting C<$AF_INET = 2>, you know you're in for big trouble.
899 An immeasurably superior approach is to use the C<Socket> module, which more
900 reliably grants access to the various constants and functions you'll need.
902 If you're not writing a server/client for an existing protocol like
903 NNTP or SMTP, you should give some thought to how your server will
904 know when the client has finished talking, and vice-versa. Most
905 protocols are based on one-line messages and responses (so one party
906 knows the other has finished when a "\n" is received) or multi-line
907 messages and responses that end with a period on an empty line
908 ("\n.\n" terminates a message/response).
910 =head2 Internet Line Terminators
912 The Internet line terminator is "\015\012". Under ASCII variants of
913 Unix, that could usually be written as "\r\n", but under other systems,
914 "\r\n" might at times be "\015\015\012", "\012\012\015", or something
915 completely different. The standards specify writing "\015\012" to be
916 conformant (be strict in what you provide), but they also recommend
917 accepting a lone "\012" on input (be lenient in what you require).
918 We haven't always been very good about that in the code in this manpage,
919 but unless you're on a Mac from way back in its pre-Unix dark ages, you'll
922 =head2 Internet TCP Clients and Servers
924 Use Internet-domain sockets when you want to do client-server
925 communication that might extend to machines outside of your own system.
927 Here's a sample TCP client using Internet-domain sockets:
932 my ($remote, $port, $iaddr, $paddr, $proto, $line);
934 $remote = shift || "localhost";
935 $port = shift || 2345; # random port
936 if ($port =~ /\D/) { $port = getservbyname($port, "tcp") }
937 die "No port" unless $port;
938 $iaddr = inet_aton($remote) || die "no host: $remote";
939 $paddr = sockaddr_in($port, $iaddr);
941 $proto = getprotobyname("tcp");
942 socket(SOCK, PF_INET, SOCK_STREAM, $proto) || die "socket: $!";
943 connect(SOCK, $paddr) || die "connect: $!";
944 while ($line = <SOCK>) {
948 close (SOCK) || die "close: $!";
951 And here's a corresponding server to go along with it. We'll
952 leave the address as C<INADDR_ANY> so that the kernel can choose
953 the appropriate interface on multihomed hosts. If you want sit
954 on a particular interface (like the external side of a gateway
955 or firewall machine), fill this in with your real address instead.
959 BEGIN { $ENV{PATH} = "/usr/bin:/bin" }
962 my $EOL = "\015\012";
964 sub logmsg { print "$0 $$: @_ at ", scalar localtime(), "\n" }
966 my $port = shift || 2345;
967 die "invalid port" unless if $port =~ /^ \d+ $/x;
969 my $proto = getprotobyname("tcp");
971 socket(Server, PF_INET, SOCK_STREAM, $proto) || die "socket: $!";
972 setsockopt(Server, SOL_SOCKET, SO_REUSEADDR, pack("l", 1))
973 || die "setsockopt: $!";
974 bind(Server, sockaddr_in($port, INADDR_ANY)) || die "bind: $!";
975 listen(Server, SOMAXCONN) || die "listen: $!";
977 logmsg "server started on port $port";
981 $SIG{CHLD} = \&REAPER;
983 for ( ; $paddr = accept(Client, Server); close Client) {
984 my($port, $iaddr) = sockaddr_in($paddr);
985 my $name = gethostbyaddr($iaddr, AF_INET);
987 logmsg "connection from $name [",
988 inet_ntoa($iaddr), "]
991 print Client "Hello there, $name, it's now ",
992 scalar localtime(), $EOL;
995 And here's a multithreaded version. It's multithreaded in that
996 like most typical servers, it spawns (fork()s) a slave server to
997 handle the client request so that the master server can quickly
998 go back to service a new client.
1002 BEGIN { $ENV{PATH} = "/usr/bin:/bin" }
1005 my $EOL = "\015\012";
1007 sub spawn; # forward declaration
1008 sub logmsg { print "$0 $$: @_ at ", scalar localtime(), "\n" }
1010 my $port = shift || 2345;
1011 die "invalid port" unless if $port =~ /^ \d+ $/x;
1013 my $proto = getprotobyname("tcp");
1015 socket(Server, PF_INET, SOCK_STREAM, $proto) || die "socket: $!";
1016 setsockopt(Server, SOL_SOCKET, SO_REUSEADDR, pack("l", 1))
1017 || die "setsockopt: $!";
1018 bind(Server, sockaddr_in($port, INADDR_ANY)) || die "bind: $!";
1019 listen(Server, SOMAXCONN) || die "listen: $!";
1021 logmsg "server started on port $port";
1026 use POSIX ":sys_wait_h";
1030 local $!; # don't let waitpid() overwrite current error
1031 while ((my $pid = waitpid(-1, WNOHANG)) > 0 && WIFEXITED($?)) {
1032 logmsg "reaped $waitedpid" . ($? ? " with exit $?" : "");
1034 $SIG{CHLD} = \&REAPER; # loathe SysV
1037 $SIG{CHLD} = \&REAPER;
1040 $paddr = accept(Client, Server) || do {
1041 # try again if accept() returned because got a signal
1045 my ($port, $iaddr) = sockaddr_in($paddr);
1046 my $name = gethostbyaddr($iaddr, AF_INET);
1048 logmsg "connection from $name [",
1054 print "Hello there, $name, it's now ", scalar localtime(), $EOL;
1055 exec "/usr/games/fortune" # XXX: "wrong" line terminators
1056 or confess "can't exec fortune: $!";
1062 my $coderef = shift;
1064 unless (@_ == 0 && $coderef && ref($coderef) eq "CODE") {
1065 confess "usage: spawn CODEREF";
1069 unless (defined($pid = fork())) {
1070 logmsg "cannot fork: $!";
1074 logmsg "begat $pid";
1075 return; # I'm the parent
1077 # else I'm the child -- go spawn
1079 open(STDIN, "<&Client") || die "can't dup client to stdin";
1080 open(STDOUT, ">&Client") || die "can't dup client to stdout";
1081 ## open(STDERR, ">&STDOUT") || die "can't dup stdout to stderr";
1085 This server takes the trouble to clone off a child version via fork()
1086 for each incoming request. That way it can handle many requests at
1087 once, which you might not always want. Even if you don't fork(), the
1088 listen() will allow that many pending connections. Forking servers
1089 have to be particularly careful about cleaning up their dead children
1090 (called "zombies" in Unix parlance), because otherwise you'll quickly
1091 fill up your process table. The REAPER subroutine is used here to
1092 call waitpid() for any child processes that have finished, thereby
1093 ensuring that they terminate cleanly and don't join the ranks of the
1096 Within the while loop we call accept() and check to see if it returns
1097 a false value. This would normally indicate a system error needs
1098 to be reported. However, the introduction of safe signals (see
1099 L</Deferred Signals (Safe Signals)> above) in Perl 5.7.3 means that
1100 accept() might also be interrupted when the process receives a signal.
1101 This typically happens when one of the forked subprocesses exits and
1102 notifies the parent process with a CHLD signal.
1104 If accept() is interrupted by a signal, $! will be set to EINTR.
1105 If this happens, we can safely continue to the next iteration of
1106 the loop and another call to accept(). It is important that your
1107 signal handling code not modify the value of $!, or else this test
1108 will likely fail. In the REAPER subroutine we create a local version
1109 of $! before calling waitpid(). When waitpid() sets $! to ECHILD as
1110 it inevitably does when it has no more children waiting, it
1111 updates the local copy and leaves the original unchanged.
1113 You should use the B<-T> flag to enable taint checking (see L<perlsec>)
1114 even if we aren't running setuid or setgid. This is always a good idea
1115 for servers or any program run on behalf of someone else (like CGI
1116 scripts), because it lessens the chances that people from the outside will
1117 be able to compromise your system.
1119 Let's look at another TCP client. This one connects to the TCP "time"
1120 service on a number of different machines and shows how far their clocks
1121 differ from the system on which it's being run:
1127 my $SECS_OF_70_YEARS = 2208988800;
1128 sub ctime { scalar localtime(shift() || time()) }
1130 my $iaddr = gethostbyname("localhost");
1131 my $proto = getprotobyname("tcp");
1132 my $port = getservbyname("time", "tcp");
1133 my $paddr = sockaddr_in(0, $iaddr);
1137 printf "%-24s %8s %s\n", "localhost", 0, ctime();
1139 foreach $host (@ARGV) {
1140 printf "%-24s ", $host;
1141 my $hisiaddr = inet_aton($host) || die "unknown host";
1142 my $hispaddr = sockaddr_in($port, $hisiaddr);
1143 socket(SOCKET, PF_INET, SOCK_STREAM, $proto)
1144 || die "socket: $!";
1145 connect(SOCKET, $hispaddr) || die "connect: $!";
1146 my $rtime = pack("C4", ());
1147 read(SOCKET, $rtime, 4);
1149 my $histime = unpack("N", $rtime) - $SECS_OF_70_YEARS;
1150 printf "%8d %s\n", $histime - time(), ctime($histime);
1153 =head2 Unix-Domain TCP Clients and Servers
1155 That's fine for Internet-domain clients and servers, but what about local
1156 communications? While you can use the same setup, sometimes you don't
1157 want to. Unix-domain sockets are local to the current host, and are often
1158 used internally to implement pipes. Unlike Internet domain sockets, Unix
1159 domain sockets can show up in the file system with an ls(1) listing.
1162 srw-rw-rw- 1 root 0 Oct 31 07:23 /dev/log
1164 You can test for these with Perl's B<-S> file test:
1166 unless (-S "/dev/log") {
1167 die "something's wicked with the log system";
1170 Here's a sample Unix-domain client:
1175 my ($rendezvous, $line);
1177 $rendezvous = shift || "catsock";
1178 socket(SOCK, PF_UNIX, SOCK_STREAM, 0) || die "socket: $!";
1179 connect(SOCK, sockaddr_un($rendezvous)) || die "connect: $!";
1180 while (defined($line = <SOCK>)) {
1185 And here's a corresponding server. You don't have to worry about silly
1186 network terminators here because Unix domain sockets are guaranteed
1187 to be on the localhost, and thus everything works right.
1194 BEGIN { $ENV{PATH} = "/usr/bin:/bin" }
1195 sub spawn; # forward declaration
1196 sub logmsg { print "$0 $$: @_ at ", scalar localtime(), "\n" }
1198 my $NAME = "catsock";
1199 my $uaddr = sockaddr_un($NAME);
1200 my $proto = getprotobyname("tcp");
1202 socket(Server, PF_UNIX, SOCK_STREAM, 0) || die "socket: $!";
1204 bind (Server, $uaddr) || die "bind: $!";
1205 listen(Server, SOMAXCONN) || die "listen: $!";
1207 logmsg "server started on $NAME";
1211 use POSIX ":sys_wait_h";
1214 while (($waitedpid = waitpid(-1, WNOHANG)) > 0) {
1215 logmsg "reaped $waitedpid" . ($? ? " with exit $?" : "");
1217 $SIG{CHLD} = \&REAPER; # loathe SysV
1220 $SIG{CHLD} = \&REAPER;
1223 for ( $waitedpid = 0;
1224 accept(Client, Server) || $waitedpid;
1225 $waitedpid = 0, close Client)
1228 logmsg "connection on $NAME";
1230 print "Hello there, it's now ", scalar localtime(), "\n";
1231 exec("/usr/games/fortune") || die "can't exec fortune: $!";
1236 my $coderef = shift();
1238 unless (@_ == 0 && $coderef && ref($coderef) eq "CODE") {
1239 confess "usage: spawn CODEREF";
1243 unless (defined($pid = fork())) {
1244 logmsg "cannot fork: $!";
1248 logmsg "begat $pid";
1249 return; # I'm the parent
1252 # I'm the child -- go spawn
1255 open(STDIN, "<&Client") || die "can't dup client to stdin";
1256 open(STDOUT, ">&Client") || die "can't dup client to stdout";
1257 ## open(STDERR, ">&STDOUT") || die "can't dup stdout to stderr";
1261 As you see, it's remarkably similar to the Internet domain TCP server, so
1262 much so, in fact, that we've omitted several duplicate functions--spawn(),
1263 logmsg(), ctime(), and REAPER()--which are the same as in the other server.
1265 So why would you ever want to use a Unix domain socket instead of a
1266 simpler named pipe? Because a named pipe doesn't give you sessions. You
1267 can't tell one process's data from another's. With socket programming,
1268 you get a separate session for each client; that's why accept() takes two
1271 For example, let's say that you have a long-running database server daemon
1272 that you want folks to be able to access from the Web, but only
1273 if they go through a CGI interface. You'd have a small, simple CGI
1274 program that does whatever checks and logging you feel like, and then acts
1275 as a Unix-domain client and connects to your private server.
1277 =head1 TCP Clients with IO::Socket
1279 For those preferring a higher-level interface to socket programming, the
1280 IO::Socket module provides an object-oriented approach. IO::Socket has
1281 been included in the standard Perl distribution ever since Perl 5.004. If
1282 you're running an earlier version of Perl (in which case, how are you
1283 reading this manpage?), just fetch IO::Socket from CPAN, where you'll also
1284 find modules providing easy interfaces to the following systems: DNS, FTP,
1285 Ident (RFC 931), NIS and NISPlus, NNTP, Ping, POP3, SMTP, SNMP, SSLeay,
1286 Telnet, and Time--to name just a few.
1288 =head2 A Simple Client
1290 Here's a client that creates a TCP connection to the "daytime"
1291 service at port 13 of the host name "localhost" and prints out everything
1292 that the server there cares to provide.
1296 $remote = IO::Socket::INET->new(
1298 PeerAddr => "localhost",
1299 PeerPort => "daytime(13)",
1301 || die "can't connect to daytime service on localhost";
1302 while (<$remote>) { print }
1304 When you run this program, you should get something back that
1307 Wed May 14 08:40:46 MDT 1997
1309 Here are what those parameters to the new() constructor mean:
1315 This is which protocol to use. In this case, the socket handle returned
1316 will be connected to a TCP socket, because we want a stream-oriented
1317 connection, that is, one that acts pretty much like a plain old file.
1318 Not all sockets are this of this type. For example, the UDP protocol
1319 can be used to make a datagram socket, used for message-passing.
1323 This is the name or Internet address of the remote host the server is
1324 running on. We could have specified a longer name like C<"www.perl.com">,
1325 or an address like C<"207.171.7.72">. For demonstration purposes, we've
1326 used the special hostname C<"localhost">, which should always mean the
1327 current machine you're running on. The corresponding Internet address
1328 for localhost is C<"127.0.0.1">, if you'd rather use that.
1332 This is the service name or port number we'd like to connect to.
1333 We could have gotten away with using just C<"daytime"> on systems with a
1334 well-configured system services file,[FOOTNOTE: The system services file
1335 is found in I</etc/services> under Unixy systems.] but here we've specified the
1336 port number (13) in parentheses. Using just the number would have also
1337 worked, but numeric literals make careful programmers nervous.
1341 Notice how the return value from the C<new> constructor is used as
1342 a filehandle in the C<while> loop? That's what's called an I<indirect
1343 filehandle>, a scalar variable containing a filehandle. You can use
1344 it the same way you would a normal filehandle. For example, you
1345 can read one line from it this way:
1349 all remaining lines from is this way:
1353 and send a line of data to it this way:
1355 print $handle "some data\n";
1357 =head2 A Webget Client
1359 Here's a simple client that takes a remote host to fetch a document
1360 from, and then a list of files to get from that host. This is a
1361 more interesting client than the previous one because it first sends
1362 something to the server before fetching the server's response.
1366 unless (@ARGV > 1) { die "usage: $0 host url ..." }
1367 $host = shift(@ARGV);
1370 for my $document (@ARGV) {
1371 $remote = IO::Socket::INET->new( Proto => "tcp",
1373 PeerPort => "http(80)",
1374 ) || die "cannot connect to httpd on $host";
1375 $remote->autoflush(1);
1376 print $remote "GET $document HTTP/1.0" . $BLANK;
1377 while ( <$remote> ) { print }
1381 The web server handling the HTTP service is assumed to be at
1382 its standard port, number 80. If the server you're trying to
1383 connect to is at a different port, like 1080 or 8080, you should specify it
1384 as the named-parameter pair, C<< PeerPort => 8080 >>. The C<autoflush>
1385 method is used on the socket because otherwise the system would buffer
1386 up the output we sent it. (If you're on a prehistoric Mac, you'll also
1387 need to change every C<"\n"> in your code that sends data over the network
1388 to be a C<"\015\012"> instead.)
1390 Connecting to the server is only the first part of the process: once you
1391 have the connection, you have to use the server's language. Each server
1392 on the network has its own little command language that it expects as
1393 input. The string that we send to the server starting with "GET" is in
1394 HTTP syntax. In this case, we simply request each specified document.
1395 Yes, we really are making a new connection for each document, even though
1396 it's the same host. That's the way you always used to have to speak HTTP.
1397 Recent versions of web browsers may request that the remote server leave
1398 the connection open a little while, but the server doesn't have to honor
1401 Here's an example of running that program, which we'll call I<webget>:
1403 % webget www.perl.com /guanaco.html
1404 HTTP/1.1 404 File Not Found
1405 Date: Thu, 08 May 1997 18:02:32 GMT
1406 Server: Apache/1.2b6
1408 Content-type: text/html
1410 <HEAD><TITLE>404 File Not Found</TITLE></HEAD>
1411 <BODY><H1>File Not Found</H1>
1412 The requested URL /guanaco.html was not found on this server.<P>
1415 Ok, so that's not very interesting, because it didn't find that
1416 particular document. But a long response wouldn't have fit on this page.
1418 For a more featureful version of this program, you should look to
1419 the I<lwp-request> program included with the LWP modules from CPAN.
1421 =head2 Interactive Client with IO::Socket
1423 Well, that's all fine if you want to send one command and get one answer,
1424 but what about setting up something fully interactive, somewhat like
1425 the way I<telnet> works? That way you can type a line, get the answer,
1426 type a line, get the answer, etc.
1428 This client is more complicated than the two we've done so far, but if
1429 you're on a system that supports the powerful C<fork> call, the solution
1430 isn't that rough. Once you've made the connection to whatever service
1431 you'd like to chat with, call C<fork> to clone your process. Each of
1432 these two identical process has a very simple job to do: the parent
1433 copies everything from the socket to standard output, while the child
1434 simultaneously copies everything from standard input to the socket.
1435 To accomplish the same thing using just one process would be I<much>
1436 harder, because it's easier to code two processes to do one thing than it
1437 is to code one process to do two things. (This keep-it-simple principle
1438 a cornerstones of the Unix philosophy, and good software engineering as
1439 well, which is probably why it's spread to other systems.)
1446 my ($host, $port, $kidpid, $handle, $line);
1448 unless (@ARGV == 2) { die "usage: $0 host port" }
1449 ($host, $port) = @ARGV;
1451 # create a tcp connection to the specified host and port
1452 $handle = IO::Socket::INET->new(Proto => "tcp",
1455 || die "can't connect to port $port on $host: $!";
1457 $handle->autoflush(1); # so output gets there right away
1458 print STDERR "[Connected to $host:$port]\n";
1460 # split the program into two processes, identical twins
1461 die "can't fork: $!" unless defined($kidpid = fork());
1463 # the if{} block runs only in the parent process
1465 # copy the socket to standard output
1466 while (defined ($line = <$handle>)) {
1469 kill("TERM", $kidpid); # send SIGTERM to child
1471 # the else{} block runs only in the child process
1473 # copy standard input to the socket
1474 while (defined ($line = <STDIN>)) {
1475 print $handle $line;
1477 exit(0); # just in case
1480 The C<kill> function in the parent's C<if> block is there to send a
1481 signal to our child process, currently running in the C<else> block,
1482 as soon as the remote server has closed its end of the connection.
1484 If the remote server sends data a byte at time, and you need that
1485 data immediately without waiting for a newline (which might not happen),
1486 you may wish to replace the C<while> loop in the parent with the
1490 while (sysread($handle, $byte, 1) == 1) {
1494 Making a system call for each byte you want to read is not very efficient
1495 (to put it mildly) but is the simplest to explain and works reasonably
1498 =head1 TCP Servers with IO::Socket
1500 As always, setting up a server is little bit more involved than running a client.
1501 The model is that the server creates a special kind of socket that
1502 does nothing but listen on a particular port for incoming connections.
1503 It does this by calling the C<< IO::Socket::INET->new() >> method with
1504 slightly different arguments than the client did.
1510 This is which protocol to use. Like our clients, we'll
1511 still specify C<"tcp"> here.
1516 port in the C<LocalPort> argument, which we didn't do for the client.
1517 This is service name or port number for which you want to be the
1518 server. (Under Unix, ports under 1024 are restricted to the
1519 superuser.) In our sample, we'll use port 9000, but you can use
1520 any port that's not currently in use on your system. If you try
1521 to use one already in used, you'll get an "Address already in use"
1522 message. Under Unix, the C<netstat -a> command will show
1523 which services current have servers.
1527 The C<Listen> parameter is set to the maximum number of
1528 pending connections we can accept until we turn away incoming clients.
1529 Think of it as a call-waiting queue for your telephone.
1530 The low-level Socket module has a special symbol for the system maximum, which
1535 The C<Reuse> parameter is needed so that we restart our server
1536 manually without waiting a few minutes to allow system buffers to
1541 Once the generic server socket has been created using the parameters
1542 listed above, the server then waits for a new client to connect
1543 to it. The server blocks in the C<accept> method, which eventually accepts a
1544 bidirectional connection from the remote client. (Make sure to autoflush
1545 this handle to circumvent buffering.)
1547 To add to user-friendliness, our server prompts the user for commands.
1548 Most servers don't do this. Because of the prompt without a newline,
1549 you'll have to use the C<sysread> variant of the interactive client above.
1551 This server accepts one of five different commands, sending output back to
1552 the client. Unlike most network servers, this one handles only one
1553 incoming client at a time. Multithreaded servers are covered in
1554 Chapter 16 of the Camel.
1556 Here's the code. We'll
1560 use Net::hostent; # for OOish version of gethostbyaddr
1562 $PORT = 9000; # pick something not in use
1564 $server = IO::Socket::INET->new( Proto => "tcp",
1566 Listen => SOMAXCONN,
1569 die "can't setup server" unless $server;
1570 print "[Server $0 accepting clients]\n";
1572 while ($client = $server->accept()) {
1573 $client->autoflush(1);
1574 print $client "Welcome to $0; type help for command list.\n";
1575 $hostinfo = gethostbyaddr($client->peeraddr);
1576 printf "[Connect from %s]\n", $hostinfo ? $hostinfo->name : $client->peerhost;
1577 print $client "Command? ";
1578 while ( <$client>) {
1579 next unless /\S/; # blank line
1580 if (/quit|exit/i) { last }
1581 elsif (/date|time/i) { printf $client "%s\n", scalar localtime() }
1582 elsif (/who/i ) { print $client `who 2>&1` }
1583 elsif (/cookie/i ) { print $client `/usr/games/fortune 2>&1` }
1584 elsif (/motd/i ) { print $client `cat /etc/motd 2>&1` }
1586 print $client "Commands: quit date who cookie motd\n";
1589 print $client "Command? ";
1594 =head1 UDP: Message Passing
1596 Another kind of client-server setup is one that uses not connections, but
1597 messages. UDP communications involve much lower overhead but also provide
1598 less reliability, as there are no promises that messages will arrive at
1599 all, let alone in order and unmangled. Still, UDP offers some advantages
1600 over TCP, including being able to "broadcast" or "multicast" to a whole
1601 bunch of destination hosts at once (usually on your local subnet). If you
1602 find yourself overly concerned about reliability and start building checks
1603 into your message system, then you probably should use just TCP to start
1606 UDP datagrams are I<not> a bytestream and should not be treated as such.
1607 This makes using I/O mechanisms with internal buffering like stdio (i.e.
1608 print() and friends) especially cumbersome. Use syswrite(), or better
1609 send(), like in the example below.
1611 Here's a UDP program similar to the sample Internet TCP client given
1612 earlier. However, instead of checking one host at a time, the UDP version
1613 will check many of them asynchronously by simulating a multicast and then
1614 using select() to do a timed-out wait for I/O. To do something similar
1615 with TCP, you'd have to use a different socket handle for each host.
1622 my ( $count, $hisiaddr, $hispaddr, $histime,
1623 $host, $iaddr, $paddr, $port, $proto,
1624 $rin, $rout, $rtime, $SECS_OF_70_YEARS);
1626 $SECS_OF_70_YEARS = 2_208_988_800;
1628 $iaddr = gethostbyname(hostname());
1629 $proto = getprotobyname("udp");
1630 $port = getservbyname("time", "udp");
1631 $paddr = sockaddr_in(0, $iaddr); # 0 means let kernel pick
1633 socket(SOCKET, PF_INET, SOCK_DGRAM, $proto) || die "socket: $!";
1634 bind(SOCKET, $paddr) || die "bind: $!";
1637 printf "%-12s %8s %s\n", "localhost", 0, scalar localtime();
1641 $hisiaddr = inet_aton($host) || die "unknown host";
1642 $hispaddr = sockaddr_in($port, $hisiaddr);
1643 defined(send(SOCKET, 0, 0, $hispaddr)) || die "send $host: $!";
1647 vec($rin, fileno(SOCKET), 1) = 1;
1649 # timeout after 10.0 seconds
1650 while ($count && select($rout = $rin, undef, undef, 10.0)) {
1652 $hispaddr = recv(SOCKET, $rtime, 4, 0) || die "recv: $!";
1653 ($port, $hisiaddr) = sockaddr_in($hispaddr);
1654 $host = gethostbyaddr($hisiaddr, AF_INET);
1655 $histime = unpack("N", $rtime) - $SECS_OF_70_YEARS;
1656 printf "%-12s ", $host;
1657 printf "%8d %s\n", $histime - time(), scalar localtime($histime);
1661 This example does not include any retries and may consequently fail to
1662 contact a reachable host. The most prominent reason for this is congestion
1663 of the queues on the sending host if the number of hosts to contact is
1668 While System V IPC isn't so widely used as sockets, it still has some
1669 interesting uses. However, you cannot use SysV IPC or Berkeley mmap() to
1670 have a variable shared amongst several processes. That's because Perl
1671 would reallocate your string when you weren't wanting it to. You might
1672 look into the C<IPC::Shareable> or C<threads::shared> modules for that.
1674 Here's a small example showing shared memory usage.
1676 use IPC::SysV qw(IPC_PRIVATE IPC_RMID S_IRUSR S_IWUSR);
1679 $id = shmget(IPC_PRIVATE, $size, S_IRUSR | S_IWUSR);
1680 defined($id) || die "shmget: $!";
1681 print "shm key $id\n";
1683 $message = "Message #1";
1684 shmwrite($id, $message, 0, 60) || die "shmwrite: $!";
1685 print "wrote: '$message'\n";
1686 shmread($id, $buff, 0, 60) || die "shmread: $!";
1687 print "read : '$buff'\n";
1689 # the buffer of shmread is zero-character end-padded.
1690 substr($buff, index($buff, "\0")) = "";
1691 print "un" unless $buff eq $message;
1694 print "deleting shm $id\n";
1695 shmctl($id, IPC_RMID, 0) || die "shmctl: $!";
1697 Here's an example of a semaphore:
1699 use IPC::SysV qw(IPC_CREAT);
1702 $id = semget($IPC_KEY, 10, 0666 | IPC_CREAT);
1703 defined($id) || die "shmget: $!";
1704 print "shm key $id\n";
1706 Put this code in a separate file to be run in more than one process.
1707 Call the file F<take>:
1709 # create a semaphore
1712 $id = semget($IPC_KEY, 0, 0);
1713 defined($id) || die "shmget: $!";
1719 # wait for semaphore to be zero
1721 $opstring1 = pack("s!s!s!", $semnum, $semop, $semflag);
1723 # Increment the semaphore count
1725 $opstring2 = pack("s!s!s!", $semnum, $semop, $semflag);
1726 $opstring = $opstring1 . $opstring2;
1728 semop($id, $opstring) || die "semop: $!";
1730 Put this code in a separate file to be run in more than one process.
1731 Call this file F<give>:
1733 # "give" the semaphore
1734 # run this in the original process and you will see
1735 # that the second process continues
1738 $id = semget($IPC_KEY, 0, 0);
1739 die unless defined($id);
1744 # Decrement the semaphore count
1746 $opstring = pack("s!s!s!", $semnum, $semop, $semflag);
1748 semop($id, $opstring) || die "semop: $!";
1750 The SysV IPC code above was written long ago, and it's definitely
1751 clunky looking. For a more modern look, see the IPC::SysV module
1752 which is included with Perl starting from Perl 5.005.
1754 A small example demonstrating SysV message queues:
1756 use IPC::SysV qw(IPC_PRIVATE IPC_RMID IPC_CREAT S_IRUSR S_IWUSR);
1758 my $id = msgget(IPC_PRIVATE, IPC_CREAT | S_IRUSR | S_IWUSR);
1759 defined($id) || die "msgget failed: $!";
1761 my $sent = "message";
1762 my $type_sent = 1234;
1764 msgsnd($id, pack("l! a*", $type_sent, $sent), 0)
1765 || die "msgsnd failed: $!";
1767 msgrcv($id, my $rcvd_buf, 60, 0, 0)
1768 || die "msgrcv failed: $!";
1770 my($type_rcvd, $rcvd) = unpack("l! a*", $rcvd_buf);
1772 if ($rcvd eq $sent) {
1778 msgctl($id, IPC_RMID, 0) || die "msgctl failed: $!\n";
1782 Most of these routines quietly but politely return C<undef> when they
1783 fail instead of causing your program to die right then and there due to
1784 an uncaught exception. (Actually, some of the new I<Socket> conversion
1785 functions do croak() on bad arguments.) It is therefore essential to
1786 check return values from these functions. Always begin your socket
1787 programs this way for optimal success, and don't forget to add the B<-T>
1788 taint-checking flag to the C<#!> line for servers:
1797 These routines all create system-specific portability problems. As noted
1798 elsewhere, Perl is at the mercy of your C libraries for much of its system
1799 behavior. It's probably safest to assume broken SysV semantics for
1800 signals and to stick with simple TCP and UDP socket operations; e.g., don't
1801 try to pass open file descriptors over a local UDP datagram socket if you
1802 want your code to stand a chance of being portable.
1806 Tom Christiansen, with occasional vestiges of Larry Wall's original
1807 version and suggestions from the Perl Porters.
1811 There's a lot more to networking than this, but this should get you
1814 For intrepid programmers, the indispensable textbook is I<Unix Network
1815 Programming, 2nd Edition, Volume 1> by W. Richard Stevens (published by
1816 Prentice-Hall). Most books on networking address the subject from the
1817 perspective of a C programmer; translation to Perl is left as an exercise
1820 The IO::Socket(3) manpage describes the object library, and the Socket(3)
1821 manpage describes the low-level interface to sockets. Besides the obvious
1822 functions in L<perlfunc>, you should also check out the F<modules> file at
1823 your nearest CPAN site, especially
1824 L<http://www.cpan.org/modules/00modlist.long.html#ID5_Networking_>.
1825 See L<perlmodlib> or best yet, the F<Perl FAQ> for a description
1826 of what CPAN is and where to get it if the previous link doesn't work
1829 Section 5 of CPAN's F<modules> file is devoted to "Networking, Device
1830 Control (modems), and Interprocess Communication", and contains numerous
1831 unbundled modules numerous networking modules, Chat and Expect operations,
1832 CGI programming, DCE, FTP, IPC, NNTP, Proxy, Ptty, RPC, SNMP, SMTP, Telnet,
1833 Threads, and ToolTalk--to name just a few.