Commit | Line | Data |
7766f137 |
1 | =head1 NAME |
2 | |
c3c83ace |
3 | perlfork - Perl's fork() emulation |
7766f137 |
4 | |
5 | =head1 SYNOPSIS |
6 | |
c3c83ace |
7 | NOTE: As of the 5.8.0 release, fork() emulation has considerably |
8 | matured. However, there are still a few known bugs and differences |
9 | from real fork() that might affect you. See the "BUGS" and |
10 | "CAVEATS AND LIMITATIONS" sections below. |
c7fa416b |
11 | |
7766f137 |
12 | Perl provides a fork() keyword that corresponds to the Unix system call |
13 | of the same name. On most Unix-like platforms where the fork() system |
14 | call is available, Perl's fork() simply calls it. |
15 | |
16 | On some platforms such as Windows where the fork() system call is not |
17 | available, Perl can be built to emulate fork() at the interpreter level. |
18 | While the emulation is designed to be as compatible as possible with the |
106325ad |
19 | real fork() at the level of the Perl program, there are certain |
7766f137 |
20 | important differences that stem from the fact that all the pseudo child |
21 | "processes" created this way live in the same real process as far as the |
22 | operating system is concerned. |
23 | |
24 | This document provides a general overview of the capabilities and |
25 | limitations of the fork() emulation. Note that the issues discussed here |
26 | are not applicable to platforms where a real fork() is available and Perl |
27 | has been configured to use it. |
28 | |
29 | =head1 DESCRIPTION |
30 | |
31 | The fork() emulation is implemented at the level of the Perl interpreter. |
32 | What this means in general is that running fork() will actually clone the |
33 | running interpreter and all its state, and run the cloned interpreter in |
34 | a separate thread, beginning execution in the new thread just after the |
35 | point where the fork() was called in the parent. We will refer to the |
36 | thread that implements this child "process" as the pseudo-process. |
37 | |
38 | To the Perl program that called fork(), all this is designed to be |
39 | transparent. The parent returns from the fork() with a pseudo-process |
40 | ID that can be subsequently used in any process manipulation functions; |
41 | the child returns from the fork() with a value of C<0> to signify that |
42 | it is the child pseudo-process. |
43 | |
44 | =head2 Behavior of other Perl features in forked pseudo-processes |
45 | |
46 | Most Perl features behave in a natural way within pseudo-processes. |
47 | |
48 | =over 8 |
49 | |
50 | =item $$ or $PROCESS_ID |
51 | |
52 | This special variable is correctly set to the pseudo-process ID. |
53 | It can be used to identify pseudo-processes within a particular |
54 | session. Note that this value is subject to recycling if any |
55 | pseudo-processes are launched after others have been wait()-ed on. |
56 | |
57 | =item %ENV |
58 | |
4375e838 |
59 | Each pseudo-process maintains its own virtual environment. Modifications |
7766f137 |
60 | to %ENV affect the virtual environment, and are only visible within that |
61 | pseudo-process, and in any processes (or pseudo-processes) launched from |
62 | it. |
63 | |
64 | =item chdir() and all other builtins that accept filenames |
65 | |
66 | Each pseudo-process maintains its own virtual idea of the current directory. |
67 | Modifications to the current directory using chdir() are only visible within |
68 | that pseudo-process, and in any processes (or pseudo-processes) launched from |
69 | it. All file and directory accesses from the pseudo-process will correctly |
70 | map the virtual working directory to the real working directory appropriately. |
71 | |
72 | =item wait() and waitpid() |
73 | |
74 | wait() and waitpid() can be passed a pseudo-process ID returned by fork(). |
75 | These calls will properly wait for the termination of the pseudo-process |
76 | and return its status. |
77 | |
78 | =item kill() |
79 | |
80 | kill() can be used to terminate a pseudo-process by passing it the ID returned |
81 | by fork(). This should not be used except under dire circumstances, because |
82 | the operating system may not guarantee integrity of the process resources |
83 | when a running thread is terminated. Note that using kill() on a |
84 | pseudo-process() may typically cause memory leaks, because the thread that |
85 | implements the pseudo-process does not get a chance to clean up its resources. |
86 | |
87 | =item exec() |
88 | |
89 | Calling exec() within a pseudo-process actually spawns the requested |
90 | executable in a separate process and waits for it to complete before |
91 | exiting with the same exit status as that process. This means that the |
92 | process ID reported within the running executable will be different from |
93 | what the earlier Perl fork() might have returned. Similarly, any process |
94 | manipulation functions applied to the ID returned by fork() will affect the |
95 | waiting pseudo-process that called exec(), not the real process it is |
96 | waiting for after the exec(). |
97 | |
98 | =item exit() |
99 | |
100 | exit() always exits just the executing pseudo-process, after automatically |
101 | wait()-ing for any outstanding child pseudo-processes. Note that this means |
102 | that the process as a whole will not exit unless all running pseudo-processes |
1d335e36 |
103 | have exited. See below for some limitations with open filehandles. |
7766f137 |
104 | |
105 | =item Open handles to files, directories and network sockets |
106 | |
107 | All open handles are dup()-ed in pseudo-processes, so that closing |
108 | any handles in one process does not affect the others. See below for |
109 | some limitations. |
110 | |
111 | =back |
112 | |
113 | =head2 Resource limits |
114 | |
115 | In the eyes of the operating system, pseudo-processes created via the fork() |
116 | emulation are simply threads in the same process. This means that any |
117 | process-level limits imposed by the operating system apply to all |
118 | pseudo-processes taken together. This includes any limits imposed by the |
119 | operating system on the number of open file, directory and socket handles, |
120 | limits on disk space usage, limits on memory size, limits on CPU utilization |
121 | etc. |
122 | |
123 | =head2 Killing the parent process |
124 | |
125 | If the parent process is killed (either using Perl's kill() builtin, or |
126 | using some external means) all the pseudo-processes are killed as well, |
127 | and the whole process exits. |
128 | |
129 | =head2 Lifetime of the parent process and pseudo-processes |
130 | |
131 | During the normal course of events, the parent process and every |
132 | pseudo-process started by it will wait for their respective pseudo-children |
133 | to complete before they exit. This means that the parent and every |
134 | pseudo-child created by it that is also a pseudo-parent will only exit |
135 | after their pseudo-children have exited. |
136 | |
137 | A way to mark a pseudo-processes as running detached from their parent (so |
138 | that the parent would not have to wait() for them if it doesn't want to) |
139 | will be provided in future. |
140 | |
141 | =head2 CAVEATS AND LIMITATIONS |
142 | |
143 | =over 8 |
144 | |
145 | =item BEGIN blocks |
146 | |
147 | The fork() emulation will not work entirely correctly when called from |
148 | within a BEGIN block. The forked copy will run the contents of the |
149 | BEGIN block, but will not continue parsing the source stream after the |
150 | BEGIN block. For example, consider the following code: |
151 | |
152 | BEGIN { |
153 | fork and exit; # fork child and exit the parent |
154 | print "inner\n"; |
155 | } |
156 | print "outer\n"; |
157 | |
158 | This will print: |
159 | |
160 | inner |
161 | |
162 | rather than the expected: |
163 | |
164 | inner |
165 | outer |
166 | |
167 | This limitation arises from fundamental technical difficulties in |
168 | cloning and restarting the stacks used by the Perl parser in the |
169 | middle of a parse. |
170 | |
171 | =item Open filehandles |
172 | |
173 | Any filehandles open at the time of the fork() will be dup()-ed. Thus, |
174 | the files can be closed independently in the parent and child, but beware |
175 | that the dup()-ed handles will still share the same seek pointer. Changing |
176 | the seek position in the parent will change it in the child and vice-versa. |
177 | One can avoid this by opening files that need distinct seek pointers |
178 | separately in the child. |
179 | |
1d335e36 |
180 | On some operating systems, notably Solaris and Unixware, calling C<exit()> |
181 | from a child process will flush and close open filehandles in the parent, |
182 | thereby corrupting the filehandles. On these systems, calling C<_exit()> |
183 | is suggested instead. C<_exit()> is available in Perl through the |
184 | C<POSIX> module. Please consult your systems manpages for more information |
185 | on this. |
186 | |
030866aa |
187 | =item Forking pipe open() not yet implemented |
188 | |
189 | The C<open(FOO, "|-")> and C<open(BAR, "-|")> constructs are not yet |
190 | implemented. This limitation can be easily worked around in new code |
191 | by creating a pipe explicitly. The following example shows how to |
192 | write to a forked child: |
193 | |
194 | # simulate open(FOO, "|-") |
195 | sub pipe_to_fork ($) { |
196 | my $parent = shift; |
197 | pipe my $child, $parent or die; |
198 | my $pid = fork(); |
199 | die "fork() failed: $!" unless defined $pid; |
200 | if ($pid) { |
201 | close $child; |
202 | } |
203 | else { |
204 | close $parent; |
205 | open(STDIN, "<&=" . fileno($child)) or die; |
206 | } |
207 | $pid; |
208 | } |
209 | |
210 | if (pipe_to_fork('FOO')) { |
211 | # parent |
212 | print FOO "pipe_to_fork\n"; |
213 | close FOO; |
214 | } |
215 | else { |
216 | # child |
217 | while (<STDIN>) { print; } |
030866aa |
218 | exit(0); |
219 | } |
220 | |
221 | And this one reads from the child: |
222 | |
223 | # simulate open(FOO, "-|") |
224 | sub pipe_from_fork ($) { |
225 | my $parent = shift; |
226 | pipe $parent, my $child or die; |
227 | my $pid = fork(); |
228 | die "fork() failed: $!" unless defined $pid; |
229 | if ($pid) { |
230 | close $child; |
231 | } |
232 | else { |
233 | close $parent; |
234 | open(STDOUT, ">&=" . fileno($child)) or die; |
235 | } |
236 | $pid; |
237 | } |
238 | |
239 | if (pipe_from_fork('BAR')) { |
240 | # parent |
241 | while (<BAR>) { print; } |
242 | close BAR; |
243 | } |
244 | else { |
245 | # child |
246 | print "pipe_from_fork\n"; |
030866aa |
247 | exit(0); |
248 | } |
249 | |
250 | Forking pipe open() constructs will be supported in future. |
251 | |
7766f137 |
252 | =item Global state maintained by XSUBs |
253 | |
254 | External subroutines (XSUBs) that maintain their own global state may |
255 | not work correctly. Such XSUBs will either need to maintain locks to |
256 | protect simultaneous access to global data from different pseudo-processes, |
257 | or maintain all their state on the Perl symbol table, which is copied |
258 | naturally when fork() is called. A callback mechanism that provides |
259 | extensions an opportunity to clone their state will be provided in the |
260 | near future. |
261 | |
262 | =item Interpreter embedded in larger application |
263 | |
264 | The fork() emulation may not behave as expected when it is executed in an |
265 | application which embeds a Perl interpreter and calls Perl APIs that can |
266 | evaluate bits of Perl code. This stems from the fact that the emulation |
267 | only has knowledge about the Perl interpreter's own data structures and |
268 | knows nothing about the containing application's state. For example, any |
269 | state carried on the application's own call stack is out of reach. |
270 | |
7e396c59 |
271 | =item Thread-safety of extensions |
272 | |
273 | Since the fork() emulation runs code in multiple threads, extensions |
274 | calling into non-thread-safe libraries may not work reliably when |
275 | calling fork(). As Perl's threading support gradually becomes more |
276 | widely adopted even on platforms with a native fork(), such extensions |
277 | are expected to be fixed for thread-safety. |
278 | |
7766f137 |
279 | =back |
280 | |
281 | =head1 BUGS |
282 | |
283 | =over 8 |
284 | |
285 | =item * |
286 | |
287 | Having pseudo-process IDs be negative integers breaks down for the integer |
288 | C<-1> because the wait() and waitpid() functions treat this number as |
289 | being special. The tacit assumption in the current implementation is that |
290 | the system never allocates a thread ID of C<1> for user threads. A better |
291 | representation for pseudo-process IDs will be implemented in future. |
292 | |
293 | =item * |
294 | |
c3c83ace |
295 | In certain cases, the OS-level handles created by the pipe(), socket(), |
296 | and accept() operators are apparently not duplicated accurately in |
297 | pseudo-processes. This only happens in some situations, but where it |
298 | does happen, it may result in deadlocks between the read and write ends |
299 | of pipe handles, or inability to send or receive data across socket |
300 | handles. |
301 | |
302 | =item * |
303 | |
7766f137 |
304 | This document may be incomplete in some respects. |
305 | |
a45bd81d |
306 | =back |
307 | |
7766f137 |
308 | =head1 AUTHOR |
309 | |
7e396c59 |
310 | Support for concurrent interpreters and the fork() emulation was implemented |
311 | by ActiveState, with funding from Microsoft Corporation. |
7766f137 |
312 | |
313 | This document is authored and maintained by Gurusamy Sarathy |
314 | E<lt>gsar@activestate.comE<gt>. |
315 | |
316 | =head1 SEE ALSO |
317 | |
318 | L<perlfunc/"fork">, L<perlipc> |
319 | |
320 | =cut |