[ViewVC] Diff of: cvs/AnyEvent-Fork/Fork.pm

Comparing AnyEvent-Fork/Fork.pm (file contents):
Revision 1.10 by root, Thu Apr 4 06:09:15 2013 UTC vs.
Revision 1.49 by root, Fri Apr 19 12:56:53 2013 UTC

…		…
1	=head1 NAME	1	=head1 NAME
2		2
3	AnyEvent::Fork - everything you wanted to use fork() for, but couldn't	3	AnyEvent::Fork - everything you wanted to use fork() for, but couldn't
4		4
5	ATTENTION, this is a very early release, and very untested. Consider it a
6	technology preview.
7
8	=head1 SYNOPSIS	5	=head1 SYNOPSIS
9		6
10	use AnyEvent::Fork;	7	use AnyEvent::Fork;
11		8
12	##################################################################	9	AnyEvent::Fork
		10	->new
		11	->require ("MyModule")
		12	->run ("MyModule::server", my $cv = AE::cv);
		13
		14	my $fh = $cv->recv;
		15
		16	=head1 DESCRIPTION
		17
		18	This module allows you to create new processes, without actually forking
		19	them from your current process (avoiding the problems of forking), but
		20	preserving most of the advantages of fork.
		21
		22	It can be used to create new worker processes or new independent
		23	subprocesses for short- and long-running jobs, process pools (e.g. for use
		24	in pre-forked servers) but also to spawn new external processes (such as
		25	CGI scripts from a web server), which can be faster (and more well behaved)
		26	than using fork+exec in big processes.
		27
		28	Special care has been taken to make this module useful from other modules,
		29	while still supporting specialised environments such as L<App::Staticperl>
		30	or L<PAR::Packer>.
		31
		32	=head2 WHAT THIS MODULE IS NOT
		33
		34	This module only creates processes and lets you pass file handles and
		35	strings to it, and run perl code. It does not implement any kind of RPC -
		36	there is no back channel from the process back to you, and there is no RPC
		37	or message passing going on.
		38
		39	If you need some form of RPC, you could use the L<AnyEvent::Fork::RPC>
		40	companion module, which adds simple RPC/job queueing to a process created
		41	by this module.
		42
		43	Or you can implement it yourself in whatever way you like, use some
		44	message-passing module such as L<AnyEvent::MP>, some pipe such as
		45	L<AnyEvent::ZeroMQ>, use L<AnyEvent::Handle> on both sides to send
		46	e.g. JSON or Storable messages, and so on.
		47
		48	=head2 COMPARISON TO OTHER MODULES
		49
		50	There is an abundance of modules on CPAN that do "something fork", such as
		51	L<Parallel::ForkManager>, L<AnyEvent::ForkManager>, L<AnyEvent::Worker>
		52	or L<AnyEvent::Subprocess>. There are modules that implement their own
		53	process management, such as L<AnyEvent::DBI>.
		54
		55	The problems that all these modules try to solve are real, however, none
		56	of them (from what I have seen) tackle the very real problems of unwanted
		57	memory sharing, efficiency, not being able to use event processing or
		58	similar modules in the processes they create.
		59
		60	This module doesn't try to replace any of them - instead it tries to solve
		61	the problem of creating processes with a minimum of fuss and overhead (and
		62	also luxury). Ideally, most of these would use AnyEvent::Fork internally,
		63	except they were written before AnyEvent:Fork was available, so obviously
		64	had to roll their own.
		65
		66	=head2 PROBLEM STATEMENT
		67
		68	There are two traditional ways to implement parallel processing on UNIX
		69	like operating systems - fork and process, and fork+exec and process. They
		70	have different advantages and disadvantages that I describe below,
		71	together with how this module tries to mitigate the disadvantages.
		72
		73	=over 4
		74
		75	=item Forking from a big process can be very slow.
		76
		77	A 5GB process needs 0.05s to fork on my 3.6GHz amd64 GNU/Linux box. This
		78	overhead is often shared with exec (because you have to fork first), but
		79	in some circumstances (e.g. when vfork is used), fork+exec can be much
		80	faster.
		81
		82	This module can help here by telling a small(er) helper process to fork,
		83	which is faster then forking the main process, and also uses vfork where
		84	possible. This gives the speed of vfork, with the flexibility of fork.
		85
		86	=item Forking usually creates a copy-on-write copy of the parent
		87	process.
		88
		89	For example, modules or data files that are loaded will not use additional
		90	memory after a fork. When exec'ing a new process, modules and data files
		91	might need to be loaded again, at extra CPU and memory cost. But when
		92	forking, literally all data structures are copied - if the program frees
		93	them and replaces them by new data, the child processes will retain the
		94	old version even if it isn't used, which can suddenly and unexpectedly
		95	increase memory usage when freeing memory.
		96
		97	The trade-off is between more sharing with fork (which can be good or
		98	bad), and no sharing with exec.
		99
		100	This module allows the main program to do a controlled fork, and allows
		101	modules to exec processes safely at any time. When creating a custom
		102	process pool you can take advantage of data sharing via fork without
		103	risking to share large dynamic data structures that will blow up child
		104	memory usage.
		105
		106	In other words, this module puts you into control over what is being
		107	shared and what isn't, at all times.
		108
		109	=item Exec'ing a new perl process might be difficult.
		110
		111	For example, it is not easy to find the correct path to the perl
		112	interpreter - C<$^X> might not be a perl interpreter at all.
		113
		114	This module tries hard to identify the correct path to the perl
		115	interpreter. With a cooperative main program, exec'ing the interpreter
		116	might not even be necessary, but even without help from the main program,
		117	it will still work when used from a module.
		118
		119	=item Exec'ing a new perl process might be slow, as all necessary modules
		120	have to be loaded from disk again, with no guarantees of success.
		121
		122	Long running processes might run into problems when perl is upgraded
		123	and modules are no longer loadable because they refer to a different
		124	perl version, or parts of a distribution are newer than the ones already
		125	loaded.
		126
		127	This module supports creating pre-initialised perl processes to be used as
		128	a template for new processes.
		129
		130	=item Forking might be impossible when a program is running.
		131
		132	For example, POSIX makes it almost impossible to fork from a
		133	multi-threaded program while doing anything useful in the child - in
		134	fact, if your perl program uses POSIX threads (even indirectly via
		135	e.g. L<IO::AIO> or L<threads>), you cannot call fork on the perl level
		136	anymore without risking corruption issues on a number of operating
		137	systems.
		138
		139	This module can safely fork helper processes at any time, by calling
		140	fork+exec in C, in a POSIX-compatible way (via L<Proc::FastSpawn>).
		141
		142	=item Parallel processing with fork might be inconvenient or difficult
		143	to implement. Modules might not work in both parent and child.
		144
		145	For example, when a program uses an event loop and creates watchers it
		146	becomes very hard to use the event loop from a child program, as the
		147	watchers already exist but are only meaningful in the parent. Worse, a
		148	module might want to use such a module, not knowing whether another module
		149	or the main program also does, leading to problems.
		150
		151	Apart from event loops, graphical toolkits also commonly fall into the
		152	"unsafe module" category, or just about anything that communicates with
		153	the external world, such as network libraries and file I/O modules, which
		154	usually don't like being copied and then allowed to continue in two
		155	processes.
		156
		157	With this module only the main program is allowed to create new processes
		158	by forking (because only the main program can know when it is still safe
		159	to do so) - all other processes are created via fork+exec, which makes it
		160	possible to use modules such as event loops or window interfaces safely.
		161
		162	=back
		163
		164	=head1 EXAMPLES
		165
13	# create a single new process, tell it to run your worker function	166	=head2 Create a single new process, tell it to run your worker function.
14		167
15	AnyEvent::Fork	168	AnyEvent::Fork
16	->new	169	->new
17	->require ("MyModule")	170	->require ("MyModule")
18	->run ("MyModule::worker, sub {	171	->run ("MyModule::worker, sub {
20		173
21	# now $master_filehandle is connected to the	174	# now $master_filehandle is connected to the
22	# $slave_filehandle in the new process.	175	# $slave_filehandle in the new process.
23	});	176	});
24		177
25	# MyModule::worker might look like this	178	C<MyModule> might look like this:
		179
		180	package MyModule;
		181
26	sub MyModule::worker {	182	sub worker {
27	my ($slave_filehandle) = @_;	183	my ($slave_filehandle) = @_;
28		184
29	# now $slave_filehandle is connected to the $master_filehandle	185	# now $slave_filehandle is connected to the $master_filehandle
30	# in the original prorcess. have fun!	186	# in the original prorcess. have fun!
31	}	187	}
32		188
33	##################################################################
34	# create a pool of server processes all accepting on the same socket	189	=head2 Create a pool of server processes all accepting on the same socket.
35		190
36	# create listener socket	191	# create listener socket
37	my $listener = ...;	192	my $listener = ...;
38		193
39	# create a pool template, initialise it and give it the socket	194	# create a pool template, initialise it and give it the socket
…		…
51	}	206	}
52		207
53	# now do other things - maybe use the filehandle provided by run	208	# now do other things - maybe use the filehandle provided by run
54	# to wait for the processes to die. or whatever.	209	# to wait for the processes to die. or whatever.
55		210
56	# My::Server::run might look like this	211	C<My::Server> might look like this:
57	sub My::Server::run {	212
		213	package My::Server;
		214
		215	sub run {
58	my ($slave, $listener, $id) = @_;	216	my ($slave, $listener, $id) = @_;
59		217
60	close $slave; # we do not use the socket, so close it to save resources	218	close $slave; # we do not use the socket, so close it to save resources
61		219
62	# we could go ballistic and use e.g. AnyEvent here, or IO::AIO,	220	# we could go ballistic and use e.g. AnyEvent here, or IO::AIO,
…		…
64	while (my $socket = $listener->accept) {	222	while (my $socket = $listener->accept) {
65	# do sth. with new socket	223	# do sth. with new socket
66	}	224	}
67	}	225	}
68		226
69	=head1 DESCRIPTION	227	=head2 use AnyEvent::Fork as a faster fork+exec
70		228
71	This module allows you to create new processes, without actually forking	229	This runs C</bin/echo hi>, with standard output redirected to F</tmp/log>
72	them from your current process (avoiding the problems of forking), but	230	and standard error redirected to the communications socket. It is usually
73	preserving most of the advantages of fork.	231	faster than fork+exec, but still lets you prepare the environment.
74		232
75	It can be used to create new worker processes or new independent	233	open my $output, ">/tmp/log" or die "$!";
76	subprocesses for short- and long-running jobs, process pools (e.g. for use
77	in pre-forked servers) but also to spawn new external processes (such as
78	CGI scripts from a webserver), which can be faster (and more well behaved)
79	than using fork+exec in big processes.
80		234
81	Special care has been taken to make this module useful from other modules,	235	AnyEvent::Fork
82	while still supporting specialised environments such as L<App::Staticperl>	236	->new
83	or L<PAR::Packer>.	237	->eval ('
		238	# compile a helper function for later use
		239	sub run {
		240	my ($fh, $output, @cmd) = @_;
84		241
85	=head1 PROBLEM STATEMENT	242	# perl will clear close-on-exec on STDOUT/STDERR
		243	open STDOUT, ">&", $output or die;
		244	open STDERR, ">&", $fh or die;
86		245
87	There are two ways to implement parallel processing on UNIX like operating	246	exec @cmd;
88	systems - fork and process, and fork+exec and process. They have different	247	}
89	advantages and disadvantages that I describe below, together with how this	248	')
90	module tries to mitigate the disadvantages.	249	->send_fh ($output)
		250	->send_arg ("/bin/echo", "hi")
		251	->run ("run", my $cv = AE::cv);
91		252
92	=over 4	253	my $stderr = $cv->recv;
93
94	=item Forking from a big process can be very slow (a 5GB process needs
95	0.05s to fork on my 3.6GHz amd64 GNU/Linux box for example). This overhead
96	is often shared with exec (because you have to fork first), but in some
97	circumstances (e.g. when vfork is used), fork+exec can be much faster.
98
99	This module can help here by telling a small(er) helper process to fork,
100	or fork+exec instead.
101
102	=item Forking usually creates a copy-on-write copy of the parent
103	process. Memory (for example, modules or data files that have been
104	will not take additional memory). When exec'ing a new process, modules
105	and data files might need to be loaded again, at extra cpu and memory
106	cost. Likewise when forking, all data structures are copied as well - if
107	the program frees them and replaces them by new data, the child processes
108	will retain the memory even if it isn't used.
109
110	This module allows the main program to do a controlled fork, and allows
111	modules to exec processes safely at any time. When creating a custom
112	process pool you can take advantage of data sharing via fork without
113	risking to share large dynamic data structures that will blow up child
114	memory usage.
115
116	=item Exec'ing a new perl process might be difficult and slow. For
117	example, it is not easy to find the correct path to the perl interpreter,
118	and all modules have to be loaded from disk again. Long running processes
119	might run into problems when perl is upgraded for example.
120
121	This module supports creating pre-initialised perl processes to be used
122	as template, and also tries hard to identify the correct path to the perl
123	interpreter. With a cooperative main program, exec'ing the interpreter
124	might not even be necessary.
125
126	=item Forking might be impossible when a program is running. For example,
127	POSIX makes it almost impossible to fork from a multithreaded program and
128	do anything useful in the child - strictly speaking, if your perl program
129	uses posix threads (even indirectly via e.g. L<IO::AIO> or L<threads>),
130	you cannot call fork on the perl level anymore, at all.
131
132	This module can safely fork helper processes at any time, by caling
133	fork+exec in C, in a POSIX-compatible way.
134
135	=item Parallel processing with fork might be inconvenient or difficult
136	to implement. For example, when a program uses an event loop and creates
137	watchers it becomes very hard to use the event loop from a child
138	program, as the watchers already exist but are only meaningful in the
139	parent. Worse, a module might want to use such a system, not knowing
140	whether another module or the main program also does, leading to problems.
141
142	This module only lets the main program create pools by forking (because
143	only the main program can know when it is still safe to do so) - all other
144	pools are created by fork+exec, after which such modules can again be
145	loaded.
146
147	=back
148		254
149	=head1 CONCEPTS	255	=head1 CONCEPTS
150		256
151	This module can create new processes either by executing a new perl	257	This module can create new processes either by executing a new perl
152	process, or by forking from an existing "template" process.	258	process, or by forking from an existing "template" process.
		259
		260	All these processes are called "child processes" (whether they are direct
		261	children or not), while the process that manages them is called the
		262	"parent process".
153		263
154	Each such process comes with its own file handle that can be used to	264	Each such process comes with its own file handle that can be used to
155	communicate with it (it's actually a socket - one end in the new process,	265	communicate with it (it's actually a socket - one end in the new process,
156	one end in the main process), and among the things you can do in it are	266	one end in the main process), and among the things you can do in it are
157	load modules, fork new processes, send file handles to it, and execute	267	load modules, fork new processes, send file handles to it, and execute
…		…
169	needed the first time. Forking from this process shares the memory used	279	needed the first time. Forking from this process shares the memory used
170	for the perl interpreter with the new process, but loading modules takes	280	for the perl interpreter with the new process, but loading modules takes
171	time, and the memory is not shared with anything else.	281	time, and the memory is not shared with anything else.
172		282
173	This is ideal for when you only need one extra process of a kind, with the	283	This is ideal for when you only need one extra process of a kind, with the
174	option of starting and stipping it on demand.	284	option of starting and stopping it on demand.
175		285
176	Example:	286	Example:
177		287
178	AnyEvent::Fork	288	AnyEvent::Fork
179	->new	289	->new
…		…
194	modules you loaded) is shared between the processes, and each new process	304	modules you loaded) is shared between the processes, and each new process
195	consumes relatively little memory of its own.	305	consumes relatively little memory of its own.
196		306
197	The disadvantage of this approach is that you need to create a template	307	The disadvantage of this approach is that you need to create a template
198	process for the sole purpose of forking new processes from it, but if you	308	process for the sole purpose of forking new processes from it, but if you
199	only need a fixed number of proceses you can create them, and then destroy	309	only need a fixed number of processes you can create them, and then destroy
200	the template process.	310	the template process.
201		311
202	Example:	312	Example:
203		313
204	my $template = AnyEvent::Fork->new->require ("Some::Module");	314	my $template = AnyEvent::Fork->new->require ("Some::Module");
…		…
231	my ($fork_fh) = @_;	341	my ($fork_fh) = @_;
232	});	342	});
233		343
234	=back	344	=back
235		345
236	=head1 FUNCTIONS	346	=head1 THE C<AnyEvent::Fork> CLASS
		347
		348	This module exports nothing, and only implements a single class -
		349	C<AnyEvent::Fork>.
		350
		351	There are two class constructors that both create new processes - C<new>
		352	and C<new_exec>. The C<fork> method creates a new process by forking an
		353	existing one and could be considered a third constructor.
		354
		355	Most of the remaining methods deal with preparing the new process, by
		356	loading code, evaluating code and sending data to the new process. They
		357	usually return the process object, so you can chain method calls.
		358
		359	If a process object is destroyed before calling its C<run> method, then
		360	the process simply exits. After C<run> is called, all responsibility is
		361	passed to the specified function.
		362
		363	As long as there is any outstanding work to be done, process objects
		364	resist being destroyed, so there is no reason to store them unless you
		365	need them later - configure and forget works just fine.
237		366
238	=over 4	367	=over 4
239		368
240	=cut	369	=cut
241		370
242	package AnyEvent::Fork;	371	package AnyEvent::Fork;
243		372
244	use common::sense;	373	use common::sense;
245		374
246	use Socket ();	375	use Errno ();
247		376
248	use AnyEvent;	377	use AnyEvent;
249	use AnyEvent::Fork::Util;
250	use AnyEvent::Util ();	378	use AnyEvent::Util ();
251		379
252	our $PERL; # the path to the perl interpreter, deduces with various forms of magic	380	use IO::FDPass;
253		381
254	=item my $pool = new AnyEvent::Fork key => value...	382	our $VERSION = 0.7;
255
256	Create a new process pool. The following named parameters are supported:
257
258	=over 4
259
260	=back
261
262	=cut
263		383
264	# the early fork template process	384	# the early fork template process
265	our $EARLY;	385	our $EARLY;
266		386
267	# the empty template process	387	# the empty template process
268	our $TEMPLATE;	388	our $TEMPLATE;
269		389
		390	sub QUEUE() { 0 }
		391	sub FH() { 1 }
		392	sub WW() { 2 }
		393	sub PID() { 3 }
		394	sub CB() { 4 }
		395
		396	sub _new {
		397	my ($self, $fh, $pid) = @_;
		398
		399	AnyEvent::Util::fh_nonblocking $fh, 1;
		400
		401	$self = bless [
		402	[], # write queue - strings or fd's
		403	$fh,
		404	undef, # AE watcher
		405	$pid,
		406	], $self;
		407
		408	$self
		409	}
		410
270	sub _cmd {	411	sub _cmd {
271	my $self = shift;	412	my $self = shift;
272		413
273	#TODO: maybe append the packet to any existing string command already in the queue
274
275	# ideally, we would want to use "a (w/a)*" as format string, but perl versions	414	# ideally, we would want to use "a (w/a)*" as format string, but perl
276	# from at least 5.8.9 to 5.16.3 are all buggy and can't unpack it.	415	# versions from at least 5.8.9 to 5.16.3 are all buggy and can't unpack
277	push @{ $self->[2] }, pack "N/a", pack "(w/a)*", @_;	416	# it.
		417	push @{ $self->[QUEUE] }, pack "a L/a*", $_[0], $_[1];
278		418
279	$self->[3] \|\|= AE::io $self->[1], 1, sub {	419	$self->[WW] \|\|= AE::io $self->[FH], 1, sub {
		420	do {
280	# send the next "thing" in the queue - either a reference to an fh,	421	# send the next "thing" in the queue - either a reference to an fh,
281	# or a plain string.	422	# or a plain string.
282		423
283	if (ref $self->[2][0]) {	424	if (ref $self->[QUEUE][0]) {
284	# send fh	425	# send fh
285	AnyEvent::Fork::Util::fd_send fileno $self->[1], fileno ${ $self->[2][0] }	426	unless (IO::FDPass::send fileno $self->[FH], fileno ${ $self->[QUEUE][0] }) {
		427	return if $! == Errno::EAGAIN \|\| $! == Errno::EWOULDBLOCK;
		428	undef $self->[WW];
		429	die "AnyEvent::Fork: file descriptor send failure: $!";
		430	}
		431
286	and shift @{ $self->[2] };	432	shift @{ $self->[QUEUE] };
287		433
288	} else {	434	} else {
289	# send string	435	# send string
290	my $len = syswrite $self->[1], $self->[2][0]	436	my $len = syswrite $self->[FH], $self->[QUEUE][0];
		437
		438	unless ($len) {
		439	return if $! == Errno::EAGAIN \|\| $! == Errno::EWOULDBLOCK;
		440	undef $self->[3];
291	or do { undef $self->[3]; die "AnyEvent::Fork: command write failure: $!" };	441	die "AnyEvent::Fork: command write failure: $!";
		442	}
292		443
293	substr $self->[2][0], 0, $len, "";	444	substr $self->[QUEUE][0], 0, $len, "";
294	shift @{ $self->[2] } unless length $self->[2][0];	445	shift @{ $self->[QUEUE] } unless length $self->[QUEUE][0];
295	}	446	}
		447	} while @{ $self->[QUEUE] };
296		448
297	unless (@{ $self->[2] }) {	449	# everything written
298	undef $self->[3];	450	undef $self->[WW];
		451
299	# invoke run callback	452	# invoke run callback, if any
		453	if ($self->[CB]) {
300	$self->[0]->($self->[1]) if $self->[0];	454	$self->[CB]->($self->[FH]);
		455	@$self = ();
301	}	456	}
302	};	457	};
303	}
304		458
305	sub _new {	459	() # make sure we don't leak the watcher
306	my ($self, $fh) = @_;
307
308	AnyEvent::Util::fh_nonblocking $fh, 1;
309
310	$self = bless [
311	undef, # run callback
312	$fh,
313	[], # write queue - strings or fd's
314	undef, # AE watcher
315	], $self;
316
317	$self
318	}	460	}
319		461
320	# fork template from current process, used by AnyEvent::Fork::Early/Template	462	# fork template from current process, used by AnyEvent::Fork::Early/Template
321	sub _new_fork {	463	sub _new_fork {
322	my ($fh, $slave) = AnyEvent::Util::portable_socketpair;	464	my ($fh, $slave) = AnyEvent::Util::portable_socketpair;
…		…
328	require AnyEvent::Fork::Serve;	470	require AnyEvent::Fork::Serve;
329	$AnyEvent::Fork::Serve::OWNER = $parent;	471	$AnyEvent::Fork::Serve::OWNER = $parent;
330	close $fh;	472	close $fh;
331	$0 = "$_[1] of $parent";	473	$0 = "$_[1] of $parent";
332	AnyEvent::Fork::Serve::serve ($slave);	474	AnyEvent::Fork::Serve::serve ($slave);
333	AnyEvent::Fork::Util::_exit 0;	475	exit 0;
334	} elsif (!$pid) {	476	} elsif (!$pid) {
335	die "AnyEvent::Fork::Early/Template: unable to fork template process: $!";	477	die "AnyEvent::Fork::Early/Template: unable to fork template process: $!";
336	}	478	}
337		479
338	AnyEvent::Fork->_new ($fh)	480	AnyEvent::Fork->_new ($fh, $pid)
339	}	481	}
340		482
341	=item my $proc = new AnyEvent::Fork	483	=item my $proc = new AnyEvent::Fork
342		484
343	Create a new "empty" perl interpreter process and returns its process	485	Create a new "empty" perl interpreter process and returns its process
344	object for further manipulation.	486	object for further manipulation.
345		487
346	The new process is forked from a template process that is kept around	488	The new process is forked from a template process that is kept around
347	for this purpose. When it doesn't exist yet, it is created by a call to	489	for this purpose. When it doesn't exist yet, it is created by a call to
348	C<new_exec> and kept around for future calls.	490	C<new_exec> first and then stays around for future calls.
349
350	When the process object is destroyed, it will release the file handle
351	that connects it with the new process. When the new process has not yet
352	called C<run>, then the process will exit. Otherwise, what happens depends
353	entirely on the code that is executed.
354		491
355	=cut	492	=cut
356		493
357	sub new {	494	sub new {
358	my $class = shift;	495	my $class = shift;
…		…
394	reduces the amount of memory sharing that is possible, and is also slower.	531	reduces the amount of memory sharing that is possible, and is also slower.
395		532
396	You should use C<new> whenever possible, except when having a template	533	You should use C<new> whenever possible, except when having a template
397	process around is unacceptable.	534	process around is unacceptable.
398		535
399	The path to the perl interpreter is divined usign various methods - first	536	The path to the perl interpreter is divined using various methods - first
400	C<$^X> is investigated to see if the path ends with something that sounds	537	C<$^X> is investigated to see if the path ends with something that sounds
401	as if it were the perl interpreter. Failing this, the module falls back to	538	as if it were the perl interpreter. Failing this, the module falls back to
402	using C<$Config::Config{perlpath}>.	539	using C<$Config::Config{perlpath}>.
403		540
404	=cut	541	=cut
…		…
413	my $perl = $;	550	my $perl = $;
414		551
415	# first we try $^X, but the path must be absolute (always on win32), and end in sth.	552	# first we try $^X, but the path must be absolute (always on win32), and end in sth.
416	# that looks like perl. this obviously only works for posix and win32	553	# that looks like perl. this obviously only works for posix and win32
417	unless (	554	unless (
418	(AnyEvent::Fork::Util::WIN32 \|\| $perl =~ m%^/%)	555	($^O eq "MSWin32" \|\| $perl =~ m%^/%)
419	&& $perl =~ m%[/\\]perl(?:[0-9]+(\.[0-9]+)+)?(\.exe)?$%i	556	&& $perl =~ m%[/\\]perl(?:[0-9]+(\.[0-9]+)+)?(\.exe)?$%i
420	) {	557	) {
421	# if it doesn't look perlish enough, try Config	558	# if it doesn't look perlish enough, try Config
422	require Config;	559	require Config;
423	$perl = $Config::Config{perlpath};	560	$perl = $Config::Config{perlpath};
…		…
434	Proc::FastSpawn::fd_inherit (fileno $fh, 0);	571	Proc::FastSpawn::fd_inherit (fileno $fh, 0);
435		572
436	# quick. also doesn't work in win32. of course. what did you expect	573	# quick. also doesn't work in win32. of course. what did you expect
437	#local $ENV{PERL5LIB} = join ":", grep !ref, @INC;	574	#local $ENV{PERL5LIB} = join ":", grep !ref, @INC;
438	my %env = %ENV;	575	my %env = %ENV;
439	$env{PERL5LIB} = join +(AnyEvent::Fork::Util::WIN32 ? ";" : ":"), grep !ref, @INC;	576	$env{PERL5LIB} = join +($^O eq "MSWin32" ? ";" : ":"), grep !ref, @INC;
440		577
441	Proc::FastSpawn::spawn (	578	my $pid = Proc::FastSpawn::spawn (
442	$perl,	579	$perl,
443	["perl", "-MAnyEvent::Fork::Serve", "-e", "AnyEvent::Fork::Serve::me", fileno $slave, $$],	580	["perl", "-MAnyEvent::Fork::Serve", "-e", "AnyEvent::Fork::Serve::me", fileno $slave, $$],
444	[map "$_=$env{$_}", keys %env],	581	[map "$_=$env{$_}", keys %env],
445	) or die "unable to spawn AnyEvent::Fork server: $!";	582	) or die "unable to spawn AnyEvent::Fork server: $!";
446		583
447	$self->_new ($fh)	584	$self->_new ($fh, $pid)
		585	}
		586
		587	=item $pid = $proc->pid
		588
		589	Returns the process id of the process I<iff it is a direct child of the
		590	process running AnyEvent::Fork>, and C<undef> otherwise.
		591
		592	Normally, only processes created via C<< AnyEvent::Fork->new_exec >> and
		593	L<AnyEvent::Fork::Template> are direct children, and you are responsible
		594	to clean up their zombies when they die.
		595
		596	All other processes are not direct children, and will be cleaned up by
		597	AnyEvent::Fork itself.
		598
		599	=cut
		600
		601	sub pid {
		602	$_[0][PID]
448	}	603	}
449		604
450	=item $proc = $proc->eval ($perlcode, @args)	605	=item $proc = $proc->eval ($perlcode, @args)
451		606
452	Evaluates the given C<$perlcode> as ... perl code, while setting C<@_> to	607	Evaluates the given C<$perlcode> as ... Perl code, while setting C<@_> to
453	the strings specified by C<@args>.	608	the strings specified by C<@args>, in the "main" package.
454		609
455	This call is meant to do any custom initialisation that might be required	610	This call is meant to do any custom initialisation that might be required
456	(for example, the C<require> method uses it). It's not supposed to be used	611	(for example, the C<require> method uses it). It's not supposed to be used
457	to completely take over the process, use C<run> for that.	612	to completely take over the process, use C<run> for that.
458		613
459	The code will usually be executed after this call returns, and there is no	614	The code will usually be executed after this call returns, and there is no
460	way to pass anything back to the calling process. Any evaluation errors	615	way to pass anything back to the calling process. Any evaluation errors
461	will be reported to stderr and cause the process to exit.	616	will be reported to stderr and cause the process to exit.
462		617
		618	If you want to execute some code (that isn't in a module) to take over the
		619	process, you should compile a function via C<eval> first, and then call
		620	it via C<run>. This also gives you access to any arguments passed via the
		621	C<send_xxx> methods, such as file handles. See the L<use AnyEvent::Fork as
		622	a faster fork+exec> example to see it in action.
		623
463	Returns the process object for easy chaining of method calls.	624	Returns the process object for easy chaining of method calls.
464		625
465	=cut	626	=cut
466		627
467	sub eval {	628	sub eval {
468	my ($self, $code, @args) = @_;	629	my ($self, $code, @args) = @_;
469		630
470	$self->_cmd (e => $code, @args);	631	$self->_cmd (e => pack "(w/a)", $code, @args);
471		632
472	$self	633	$self
473	}	634	}
474		635
475	=item $proc = $proc->require ($module, ...)	636	=item $proc = $proc->require ($module, ...)
…		…
492	=item $proc = $proc->send_fh ($handle, ...)	653	=item $proc = $proc->send_fh ($handle, ...)
493		654
494	Send one or more file handles (I<not> file descriptors) to the process,	655	Send one or more file handles (I<not> file descriptors) to the process,
495	to prepare a call to C<run>.	656	to prepare a call to C<run>.
496		657
497	The process object keeps a reference to the handles until this is done,	658	The process object keeps a reference to the handles until they have
498	so you must not explicitly close the handles. This is most easily	659	been passed over to the process, so you must not explicitly close the
499	accomplished by simply not storing the file handles anywhere after passing	660	handles. This is most easily accomplished by simply not storing the file
500	them to this method.	661	handles anywhere after passing them to this method - when AnyEvent::Fork
		662	is finished using them, perl will automatically close them.
501		663
502	Returns the process object for easy chaining of method calls.	664	Returns the process object for easy chaining of method calls.
503		665
504	Example: pass an fh to a process, and release it without closing. it will	666	Example: pass a file handle to a process, and release it without
505	be closed automatically when it is no longer used.	667	closing. It will be closed automatically when it is no longer used.
506		668
507	$proc->send_fh ($my_fh);	669	$proc->send_fh ($my_fh);
508	undef $my_fh; # free the reference if you want, but DO NOT CLOSE IT	670	undef $my_fh; # free the reference if you want, but DO NOT CLOSE IT
509		671
510	=cut	672	=cut
…		…
512	sub send_fh {	674	sub send_fh {
513	my ($self, @fh) = @_;	675	my ($self, @fh) = @_;
514		676
515	for my $fh (@fh) {	677	for my $fh (@fh) {
516	$self->_cmd ("h");	678	$self->_cmd ("h");
517	push @{ $self->[2] }, \$fh;	679	push @{ $self->[QUEUE] }, \$fh;
518	}	680	}
519		681
520	$self	682	$self
521	}	683	}
522		684
523	=item $proc = $proc->send_arg ($string, ...)	685	=item $proc = $proc->send_arg ($string, ...)
524		686
525	Send one or more argument strings to the process, to prepare a call to	687	Send one or more argument strings to the process, to prepare a call to
526	C<run>. The strings can be any octet string.	688	C<run>. The strings can be any octet strings.
527		689
		690	The protocol is optimised to pass a moderate number of relatively short
		691	strings - while you can pass up to 4GB of data in one go, this is more
		692	meant to pass some ID information or other startup info, not big chunks of
		693	data.
		694
528	Returns the process object for easy chaining of emthod calls.	695	Returns the process object for easy chaining of method calls.
529		696
530	=cut	697	=cut
531		698
532	sub send_arg {	699	sub send_arg {
533	my ($self, @arg) = @_;	700	my ($self, @arg) = @_;
534		701
535	$self->_cmd (a => @arg);	702	$self->_cmd (a => pack "(w/a)", @arg);
536		703
537	$self	704	$self
538	}	705	}
539		706
540	=item $proc->run ($func, $cb->($fh))	707	=item $proc->run ($func, $cb->($fh))
541		708
542	Enter the function specified by the fully qualified name in C<$func> in	709	Enter the function specified by the function name in C<$func> in the
543	the process. The function is called with the communication socket as first	710	process. The function is called with the communication socket as first
544	argument, followed by all file handles and string arguments sent earlier	711	argument, followed by all file handles and string arguments sent earlier
545	via C<send_fh> and C<send_arg> methods, in the order they were called.	712	via C<send_fh> and C<send_arg> methods, in the order they were called.
546		713
547	If the called function returns, the process exits.
548
549	Preparing the process can take time - when the process is ready, the
550	callback is invoked with the local communications socket as argument.
551
552	The process object becomes unusable on return from this function.	714	The process object becomes unusable on return from this function - any
		715	further method calls result in undefined behaviour.
		716
		717	The function name should be fully qualified, but if it isn't, it will be
		718	looked up in the C<main> package.
		719
		720	If the called function returns, doesn't exist, or any error occurs, the
		721	process exits.
		722
		723	Preparing the process is done in the background - when all commands have
		724	been sent, the callback is invoked with the local communications socket
		725	as argument. At this point you can start using the socket in any way you
		726	like.
553		727
554	If the communication socket isn't used, it should be closed on both sides,	728	If the communication socket isn't used, it should be closed on both sides,
555	to save on kernel memory.	729	to save on kernel memory.
556		730
557	The socket is non-blocking in the parent, and blocking in the newly	731	The socket is non-blocking in the parent, and blocking in the newly
558	created process. The close-on-exec flag is set on both. Even if not used	732	created process. The close-on-exec flag is set in both.
		733
559	otherwise, the socket can be a good indicator for the existance of the	734	Even if not used otherwise, the socket can be a good indicator for the
560	process - if the other process exits, you get a readable event on it,	735	existence of the process - if the other process exits, you get a readable
561	because exiting the process closes the socket (if it didn't create any	736	event on it, because exiting the process closes the socket (if it didn't
562	children using fork).	737	create any children using fork).
563		738
564	Example: create a template for a process pool, pass a few strings, some	739	Example: create a template for a process pool, pass a few strings, some
565	file handles, then fork, pass one more string, and run some code.	740	file handles, then fork, pass one more string, and run some code.
566		741
567	my $pool = AnyEvent::Fork	742	my $pool = AnyEvent::Fork
…		…
575	->send_arg ("str3")	750	->send_arg ("str3")
576	->run ("Some::function", sub {	751	->run ("Some::function", sub {
577	my ($fh) = @_;	752	my ($fh) = @_;
578		753
579	# fh is nonblocking, but we trust that the OS can accept these	754	# fh is nonblocking, but we trust that the OS can accept these
580	# extra 3 octets anyway.	755	# few octets anyway.
581	syswrite $fh, "hi #$_\n";	756	syswrite $fh, "hi #$_\n";
582		757
583	# $fh is being closed here, as we don't store it anywhere	758	# $fh is being closed here, as we don't store it anywhere
584	});	759	});
585	}	760	}
…		…
587	# Some::function might look like this - all parameters passed before fork	762	# Some::function might look like this - all parameters passed before fork
588	# and after will be passed, in order, after the communications socket.	763	# and after will be passed, in order, after the communications socket.
589	sub Some::function {	764	sub Some::function {
590	my ($fh, $str1, $str2, $fh1, $fh2, $str3) = @_;	765	my ($fh, $str1, $str2, $fh1, $fh2, $str3) = @_;
591		766
592	print scalar <$fh>; # prints "hi 1\n" and "hi 2\n"	767	print scalar <$fh>; # prints "hi #1\n" and "hi #2\n" in any order
593	}	768	}
594		769
595	=cut	770	=cut
596		771
597	sub run {	772	sub run {
598	my ($self, $func, $cb) = @_;	773	my ($self, $func, $cb) = @_;
599		774
600	$self->[0] = $cb;	775	$self->[CB] = $cb;
601	$self->_cmd (r => $func);	776	$self->_cmd (r => $func);
602	}	777	}
		778
		779	=item $proc->to_fh ($cb->($fh))
		780
		781	Flushes all commands out to the process and then calls the callback with
		782	the communications socket.
		783
		784	The process object becomes unusable on return from this function - any
		785	further method calls result in undefined behaviour.
		786
		787	The point of this method is to give you a file handle thta you cna pass
		788	to another process. In that other process, you can call C<new_from_fh
		789	AnyEvent::Fork::RPC> to create a new C<AnyEvent::Fork> object from it,
		790	thereby effectively passing a fork object to another process.
		791
		792	=cut
		793
		794	sub to_fh {
		795	my ($self, $cb) = @_;
		796
		797	$self->[CB] = $cb;
		798
		799	unless ($self->[WW]) {
		800	$self->[CB]->($self->[FH]);
		801	@$self = ();
		802	}
		803	}
		804
		805	=item new_from_fh AnyEvent::Fork $fh
		806
		807	Takes a file handle originally rceeived by the C<to_fh> method and creates
		808	a new C<AnyEvent:Fork> object. The child process itself will not change in
		809	any way, i.e. it will keep all the modifications done to it before calling
		810	C<to_fh>.
		811
		812	The new object is very much like the original object, except that the
		813	C<pid> method will return C<undef> even if the process is a direct child.
		814
		815	=cut
		816
		817	sub new_from_fh {
		818	my ($class, $fh) = @_;
		819
		820	$class->_new ($fh)
		821	}
		822
		823	=back
		824
		825	=head1 PERFORMANCE
		826
		827	Now for some unscientific benchmark numbers (all done on an amd64
		828	GNU/Linux box). These are intended to give you an idea of the relative
		829	performance you can expect, they are not meant to be absolute performance
		830	numbers.
		831
		832	OK, so, I ran a simple benchmark that creates a socket pair, forks, calls
		833	exit in the child and waits for the socket to close in the parent. I did
		834	load AnyEvent, EV and AnyEvent::Fork, for a total process size of 5100kB.
		835
		836	2079 new processes per second, using manual socketpair + fork
		837
		838	Then I did the same thing, but instead of calling fork, I called
		839	AnyEvent::Fork->new->run ("CORE::exit") and then again waited for the
		840	socket from the child to close on exit. This does the same thing as manual
		841	socket pair + fork, except that what is forked is the template process
		842	(2440kB), and the socket needs to be passed to the server at the other end
		843	of the socket first.
		844
		845	2307 new processes per second, using AnyEvent::Fork->new
		846
		847	And finally, using C<new_exec> instead C<new>, using vforks+execs to exec
		848	a new perl interpreter and compile the small server each time, I get:
		849
		850	479 vfork+execs per second, using AnyEvent::Fork->new_exec
		851
		852	So how can C<< AnyEvent->new >> be faster than a standard fork, even
		853	though it uses the same operations, but adds a lot of overhead?
		854
		855	The difference is simply the process size: forking the 5MB process takes
		856	so much longer than forking the 2.5MB template process that the extra
		857	overhead is canceled out.
		858
		859	If the benchmark process grows, the normal fork becomes even slower:
		860
		861	1340 new processes, manual fork of a 20MB process
		862	731 new processes, manual fork of a 200MB process
		863	235 new processes, manual fork of a 2000MB process
		864
		865	What that means (to me) is that I can use this module without having a bad
		866	conscience because of the extra overhead required to start new processes.
		867
		868	=head1 TYPICAL PROBLEMS
		869
		870	This section lists typical problems that remain. I hope by recognising
		871	them, most can be avoided.
		872
		873	=over 4
		874
		875	=item leaked file descriptors for exec'ed processes
		876
		877	POSIX systems inherit file descriptors by default when exec'ing a new
		878	process. While perl itself laudably sets the close-on-exec flags on new
		879	file handles, most C libraries don't care, and even if all cared, it's
		880	often not possible to set the flag in a race-free manner.
		881
		882	That means some file descriptors can leak through. And since it isn't
		883	possible to know which file descriptors are "good" and "necessary" (or
		884	even to know which file descriptors are open), there is no good way to
		885	close the ones that might harm.
		886
		887	As an example of what "harm" can be done consider a web server that
		888	accepts connections and afterwards some module uses AnyEvent::Fork for the
		889	first time, causing it to fork and exec a new process, which might inherit
		890	the network socket. When the server closes the socket, it is still open
		891	in the child (which doesn't even know that) and the client might conclude
		892	that the connection is still fine.
		893
		894	For the main program, there are multiple remedies available -
		895	L<AnyEvent::Fork::Early> is one, creating a process early and not using
		896	C<new_exec> is another, as in both cases, the first process can be exec'ed
		897	well before many random file descriptors are open.
		898
		899	In general, the solution for these kind of problems is to fix the
		900	libraries or the code that leaks those file descriptors.
		901
		902	Fortunately, most of these leaked descriptors do no harm, other than
		903	sitting on some resources.
		904
		905	=item leaked file descriptors for fork'ed processes
		906
		907	Normally, L<AnyEvent::Fork> does start new processes by exec'ing them,
		908	which closes file descriptors not marked for being inherited.
		909
		910	However, L<AnyEvent::Fork::Early> and L<AnyEvent::Fork::Template> offer
		911	a way to create these processes by forking, and this leaks more file
		912	descriptors than exec'ing them, as there is no way to mark descriptors as
		913	"close on fork".
		914
		915	An example would be modules like L<EV>, L<IO::AIO> or L<Gtk2>. Both create
		916	pipes for internal uses, and L<Gtk2> might open a connection to the X
		917	server. L<EV> and L<IO::AIO> can deal with fork, but Gtk2 might have
		918	trouble with a fork.
		919
		920	The solution is to either not load these modules before use'ing
		921	L<AnyEvent::Fork::Early> or L<AnyEvent::Fork::Template>, or to delay
		922	initialising them, for example, by calling C<init Gtk2> manually.
		923
		924	=item exiting calls object destructors
		925
		926	This only applies to users of L<AnyEvent::Fork:Early> and
		927	L<AnyEvent::Fork::Template>, or when initialising code creates objects
		928	that reference external resources.
		929
		930	When a process created by AnyEvent::Fork exits, it might do so by calling
		931	exit, or simply letting perl reach the end of the program. At which point
		932	Perl runs all destructors.
		933
		934	Not all destructors are fork-safe - for example, an object that represents
		935	the connection to an X display might tell the X server to free resources,
		936	which is inconvenient when the "real" object in the parent still needs to
		937	use them.
		938
		939	This is obviously not a problem for L<AnyEvent::Fork::Early>, as you used
		940	it as the very first thing, right?
		941
		942	It is a problem for L<AnyEvent::Fork::Template> though - and the solution
		943	is to not create objects with nontrivial destructors that might have an
		944	effect outside of Perl.
603		945
604	=back	946	=back
605		947
606	=head1 PORTABILITY NOTES	948	=head1 PORTABILITY NOTES
607		949
608	Native win32 perls are somewhat supported (AnyEvent::Fork::Early is a nop,	950	Native win32 perls are somewhat supported (AnyEvent::Fork::Early is a nop,
609	and ::Template is not going to work), and it cost a lot of blood and sweat	951	and ::Template is not going to work), and it cost a lot of blood and sweat
610	to make it so, mostly due to the bloody broken perl that nobody seems to	952	to make it so, mostly due to the bloody broken perl that nobody seems to
611	care about. The fork emulation is a bad joke - I have yet to see something	953	care about. The fork emulation is a bad joke - I have yet to see something
612	useful that you cna do with it without running into memory corruption	954	useful that you can do with it without running into memory corruption
613	issues or other braindamage. Hrrrr.	955	issues or other braindamage. Hrrrr.
614		956
615	Cygwin perl is not supported at the moment, as it should implement fd	957	Since fork is endlessly broken on win32 perls (it doesn't even remotely
616	passing, but doesn't, and rolling my own is hard, as cygwin doesn't	958	work within it's documented limits) and quite obviously it's not getting
617	support enough functionality to do it.	959	improved any time soon, the best way to proceed on windows would be to
		960	always use C<new_exec> and thus never rely on perl's fork "emulation".
618		961
619	=head1 AUTHOR	962	Cygwin perl is not supported at the moment due to some hilarious
		963	shortcomings of its API - see L<IO::FDPoll> for more details. If you never
		964	use C<send_fh> and always use C<new_exec> to create processes, it should
		965	work though.
		966
		967	=head1 SEE ALSO
		968
		969	L<AnyEvent::Fork::Early>, to avoid executing a perl interpreter at all
		970	(part of this distribution).
		971
		972	L<AnyEvent::Fork::Template>, to create a process by forking the main
		973	program at a convenient time (part of this distribution).
		974
		975	L<AnyEvent::Fork::RPC>, for simple RPC to child processes (on CPAN).
		976
		977	=head1 AUTHOR AND CONTACT INFORMATION
620		978
621	Marc Lehmann <schmorp@schmorp.de>	979	Marc Lehmann <schmorp@schmorp.de>
622	http://home.schmorp.de/	980	http://software.schmorp.de/pkg/AnyEvent-Fork
623		981
624	=cut	982	=cut
625		983
626	1	984	1
627		985

Diff Legend

-–
+Removed lines
-+
+Added lines
-<
+Changed lines
->
+Changed lines

Comparing AnyEvent-Fork/Fork.pm (file contents): Revision 1.10 by root, Thu Apr 4 06:09:15 2013 UTC vs. Revision 1.49 by root, Fri Apr 19 12:56:53 2013 UTC

Diff Legend

Comparing AnyEvent-Fork/Fork.pm (file contents):
Revision 1.10 by root, Thu Apr 4 06:09:15 2013 UTC vs.
Revision 1.49 by root, Fri Apr 19 12:56:53 2013 UTC