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

Comparing AnyEvent-Fork/Fork.pm (file contents):
Revision 1.17 by root, Fri Apr 5 23:42:24 2013 UTC vs.
Revision 1.53 by root, Fri Apr 26 15:44:44 2013 UTC

…		…
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use AnyEvent::Fork;	7	use AnyEvent::Fork;
8		8
9	##################################################################	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	And if you need some automatic process pool management on top of
		44	L<AnyEvent::Fork::RPC>, you can look at the L<AnyEvent::Fork::Pool>
		45	companion module.
		46
		47	Or you can implement it yourself in whatever way you like: use some
		48	message-passing module such as L<AnyEvent::MP>, some pipe such as
		49	L<AnyEvent::ZeroMQ>, use L<AnyEvent::Handle> on both sides to send
		50	e.g. JSON or Storable messages, and so on.
		51
		52	=head2 COMPARISON TO OTHER MODULES
		53
		54	There is an abundance of modules on CPAN that do "something fork", such as
		55	L<Parallel::ForkManager>, L<AnyEvent::ForkManager>, L<AnyEvent::Worker>
		56	or L<AnyEvent::Subprocess>. There are modules that implement their own
		57	process management, such as L<AnyEvent::DBI>.
		58
		59	The problems that all these modules try to solve are real, however, none
		60	of them (from what I have seen) tackle the very real problems of unwanted
		61	memory sharing, efficiency, not being able to use event processing or
		62	similar modules in the processes they create.
		63
		64	This module doesn't try to replace any of them - instead it tries to solve
		65	the problem of creating processes with a minimum of fuss and overhead (and
		66	also luxury). Ideally, most of these would use AnyEvent::Fork internally,
		67	except they were written before AnyEvent:Fork was available, so obviously
		68	had to roll their own.
		69
		70	=head2 PROBLEM STATEMENT
		71
		72	There are two traditional ways to implement parallel processing on UNIX
		73	like operating systems - fork and process, and fork+exec and process. They
		74	have different advantages and disadvantages that I describe below,
		75	together with how this module tries to mitigate the disadvantages.
		76
		77	=over 4
		78
		79	=item Forking from a big process can be very slow.
		80
		81	A 5GB process needs 0.05s to fork on my 3.6GHz amd64 GNU/Linux box. This
		82	overhead is often shared with exec (because you have to fork first), but
		83	in some circumstances (e.g. when vfork is used), fork+exec can be much
		84	faster.
		85
		86	This module can help here by telling a small(er) helper process to fork,
		87	which is faster then forking the main process, and also uses vfork where
		88	possible. This gives the speed of vfork, with the flexibility of fork.
		89
		90	=item Forking usually creates a copy-on-write copy of the parent
		91	process.
		92
		93	For example, modules or data files that are loaded will not use additional
		94	memory after a fork. When exec'ing a new process, modules and data files
		95	might need to be loaded again, at extra CPU and memory cost. But when
		96	forking, literally all data structures are copied - if the program frees
		97	them and replaces them by new data, the child processes will retain the
		98	old version even if it isn't used, which can suddenly and unexpectedly
		99	increase memory usage when freeing memory.
		100
		101	The trade-off is between more sharing with fork (which can be good or
		102	bad), and no sharing with exec.
		103
		104	This module allows the main program to do a controlled fork, and allows
		105	modules to exec processes safely at any time. When creating a custom
		106	process pool you can take advantage of data sharing via fork without
		107	risking to share large dynamic data structures that will blow up child
		108	memory usage.
		109
		110	In other words, this module puts you into control over what is being
		111	shared and what isn't, at all times.
		112
		113	=item Exec'ing a new perl process might be difficult.
		114
		115	For example, it is not easy to find the correct path to the perl
		116	interpreter - C<$^X> might not be a perl interpreter at all.
		117
		118	This module tries hard to identify the correct path to the perl
		119	interpreter. With a cooperative main program, exec'ing the interpreter
		120	might not even be necessary, but even without help from the main program,
		121	it will still work when used from a module.
		122
		123	=item Exec'ing a new perl process might be slow, as all necessary modules
		124	have to be loaded from disk again, with no guarantees of success.
		125
		126	Long running processes might run into problems when perl is upgraded
		127	and modules are no longer loadable because they refer to a different
		128	perl version, or parts of a distribution are newer than the ones already
		129	loaded.
		130
		131	This module supports creating pre-initialised perl processes to be used as
		132	a template for new processes.
		133
		134	=item Forking might be impossible when a program is running.
		135
		136	For example, POSIX makes it almost impossible to fork from a
		137	multi-threaded program while doing anything useful in the child - in
		138	fact, if your perl program uses POSIX threads (even indirectly via
		139	e.g. L<IO::AIO> or L<threads>), you cannot call fork on the perl level
		140	anymore without risking corruption issues on a number of operating
		141	systems.
		142
		143	This module can safely fork helper processes at any time, by calling
		144	fork+exec in C, in a POSIX-compatible way (via L<Proc::FastSpawn>).
		145
		146	=item Parallel processing with fork might be inconvenient or difficult
		147	to implement. Modules might not work in both parent and child.
		148
		149	For example, when a program uses an event loop and creates watchers it
		150	becomes very hard to use the event loop from a child program, as the
		151	watchers already exist but are only meaningful in the parent. Worse, a
		152	module might want to use such a module, not knowing whether another module
		153	or the main program also does, leading to problems.
		154
		155	Apart from event loops, graphical toolkits also commonly fall into the
		156	"unsafe module" category, or just about anything that communicates with
		157	the external world, such as network libraries and file I/O modules, which
		158	usually don't like being copied and then allowed to continue in two
		159	processes.
		160
		161	With this module only the main program is allowed to create new processes
		162	by forking (because only the main program can know when it is still safe
		163	to do so) - all other processes are created via fork+exec, which makes it
		164	possible to use modules such as event loops or window interfaces safely.
		165
		166	=back
		167
		168	=head1 EXAMPLES
		169
10	# create a single new process, tell it to run your worker function	170	=head2 Create a single new process, tell it to run your worker function.
11		171
12	AnyEvent::Fork	172	AnyEvent::Fork
13	->new	173	->new
14	->require ("MyModule")	174	->require ("MyModule")
15	->run ("MyModule::worker, sub {	175	->run ("MyModule::worker, sub {
…		…
17		177
18	# now $master_filehandle is connected to the	178	# now $master_filehandle is connected to the
19	# $slave_filehandle in the new process.	179	# $slave_filehandle in the new process.
20	});	180	});
21		181
22	# MyModule::worker might look like this	182	C<MyModule> might look like this:
		183
		184	package MyModule;
		185
23	sub MyModule::worker {	186	sub worker {
24	my ($slave_filehandle) = @_;	187	my ($slave_filehandle) = @_;
25		188
26	# now $slave_filehandle is connected to the $master_filehandle	189	# now $slave_filehandle is connected to the $master_filehandle
27	# in the original prorcess. have fun!	190	# in the original prorcess. have fun!
28	}	191	}
29		192
30	##################################################################
31	# create a pool of server processes all accepting on the same socket	193	=head2 Create a pool of server processes all accepting on the same socket.
32		194
33	# create listener socket	195	# create listener socket
34	my $listener = ...;	196	my $listener = ...;
35		197
36	# create a pool template, initialise it and give it the socket	198	# create a pool template, initialise it and give it the socket
…		…
48	}	210	}
49		211
50	# now do other things - maybe use the filehandle provided by run	212	# now do other things - maybe use the filehandle provided by run
51	# to wait for the processes to die. or whatever.	213	# to wait for the processes to die. or whatever.
52		214
53	# My::Server::run might look like this	215	C<My::Server> might look like this:
54	sub My::Server::run {	216
		217	package My::Server;
		218
		219	sub run {
55	my ($slave, $listener, $id) = @_;	220	my ($slave, $listener, $id) = @_;
56		221
57	close $slave; # we do not use the socket, so close it to save resources	222	close $slave; # we do not use the socket, so close it to save resources
58		223
59	# we could go ballistic and use e.g. AnyEvent here, or IO::AIO,	224	# we could go ballistic and use e.g. AnyEvent here, or IO::AIO,
…		…
61	while (my $socket = $listener->accept) {	226	while (my $socket = $listener->accept) {
62	# do sth. with new socket	227	# do sth. with new socket
63	}	228	}
64	}	229	}
65		230
66	=head1 DESCRIPTION	231	=head2 use AnyEvent::Fork as a faster fork+exec
67		232
68	This module allows you to create new processes, without actually forking	233	This runs C</bin/echo hi>, with standard output redirected to F</tmp/log>
69	them from your current process (avoiding the problems of forking), but	234	and standard error redirected to the communications socket. It is usually
70	preserving most of the advantages of fork.	235	faster than fork+exec, but still lets you prepare the environment.
71		236
72	It can be used to create new worker processes or new independent	237	open my $output, ">/tmp/log" or die "$!";
73	subprocesses for short- and long-running jobs, process pools (e.g. for use
74	in pre-forked servers) but also to spawn new external processes (such as
75	CGI scripts from a web server), which can be faster (and more well behaved)
76	than using fork+exec in big processes.
77		238
78	Special care has been taken to make this module useful from other modules,	239	AnyEvent::Fork
79	while still supporting specialised environments such as L<App::Staticperl>	240	->new
80	or L<PAR::Packer>.	241	->eval ('
		242	# compile a helper function for later use
		243	sub run {
		244	my ($fh, $output, @cmd) = @_;
81		245
82	=head1 WHAT THIS MODULE IS NOT	246	# perl will clear close-on-exec on STDOUT/STDERR
		247	open STDOUT, ">&", $output or die;
		248	open STDERR, ">&", $fh or die;
83		249
84	This module only creates processes and lets you pass file handles and	250	exec @cmd;
85	strings to it, and run perl code. It does not implement any kind of RPC -	251	}
86	there is no back channel from the process back to you, and there is no RPC	252	')
87	or message passing going on.	253	->send_fh ($output)
		254	->send_arg ("/bin/echo", "hi")
		255	->run ("run", my $cv = AE::cv);
88		256
89	If you need some form of RPC, you can either implement it yourself	257	my $stderr = $cv->recv;
90	in whatever way you like, use some message-passing module such
91	as L<AnyEvent::MP>, some pipe such as L<AnyEvent::ZeroMQ>, use
92	L<AnyEvent::Handle> on both sides to send e.g. JSON or Storable messages,
93	and so on.
94		258
95	=head1 PROBLEM STATEMENT	259	=head2 For stingy users: put the worker code into a C<DATA> section.
96		260
97	There are two ways to implement parallel processing on UNIX like operating	261	When you want to be stingy with files, you cna put your code into the
98	systems - fork and process, and fork+exec and process. They have different	262	C<DATA> section of your module (or program):
99	advantages and disadvantages that I describe below, together with how this
100	module tries to mitigate the disadvantages.
101		263
102	=over 4	264	use AnyEvent::Fork;
103		265
104	=item Forking from a big process can be very slow (a 5GB process needs	266	AnyEvent::Fork
105	0.05s to fork on my 3.6GHz amd64 GNU/Linux box for example). This overhead	267	->new
106	is often shared with exec (because you have to fork first), but in some	268	->eval (do { local $/; <DATA> })
107	circumstances (e.g. when vfork is used), fork+exec can be much faster.	269	->run ("doit", sub { ... });
108		270
109	This module can help here by telling a small(er) helper process to fork,	271	__DATA__
110	or fork+exec instead.
111		272
112	=item Forking usually creates a copy-on-write copy of the parent	273	sub doit {
113	process. Memory (for example, modules or data files that have been	274	... do something!
114	will not take additional memory). When exec'ing a new process, modules	275	}
115	and data files might need to be loaded again, at extra CPU and memory
116	cost. Likewise when forking, all data structures are copied as well - if
117	the program frees them and replaces them by new data, the child processes
118	will retain the memory even if it isn't used.
119		276
120	This module allows the main program to do a controlled fork, and allows	277	=head2 For stingy standalone programs: do not rely on external files at
121	modules to exec processes safely at any time. When creating a custom	278	all.
122	process pool you can take advantage of data sharing via fork without
123	risking to share large dynamic data structures that will blow up child
124	memory usage.
125		279
126	=item Exec'ing a new perl process might be difficult and slow. For	280	For single-file scripts it can be inconvenient to rely on external
127	example, it is not easy to find the correct path to the perl interpreter,	281	files - even when using < C<DATA> section, you still need to C<exec>
128	and all modules have to be loaded from disk again. Long running processes	282	an external perl interpreter, which might not be available when using
129	might run into problems when perl is upgraded for example.	283	L<App::Staticperl>, L<Urlader> or L<PAR::Packer> for example.
130		284
131	This module supports creating pre-initialised perl processes to be used	285	Two modules help here - L<AnyEvent::Fork::Early> forks a template process
132	as template, and also tries hard to identify the correct path to the perl	286	for all further calls to C<new_exec>, and L<AnyEvent::Fork::Template>
133	interpreter. With a cooperative main program, exec'ing the interpreter	287	forks the main program as a template process.
134	might not even be necessary.
135		288
136	=item Forking might be impossible when a program is running. For example,	289	Here is how your main program should look like:
137	POSIX makes it almost impossible to fork from a multi-threaded program and
138	do anything useful in the child - strictly speaking, if your perl program
139	uses posix threads (even indirectly via e.g. L<IO::AIO> or L<threads>),
140	you cannot call fork on the perl level anymore, at all.
141		290
142	This module can safely fork helper processes at any time, by calling	291	#! perl
143	fork+exec in C, in a POSIX-compatible way.
144		292
145	=item Parallel processing with fork might be inconvenient or difficult	293	# optional, as the very first thing.
146	to implement. For example, when a program uses an event loop and creates	294	# in case modules want to create their own processes.
147	watchers it becomes very hard to use the event loop from a child	295	use AnyEvent::Fork::Early;
148	program, as the watchers already exist but are only meaningful in the
149	parent. Worse, a module might want to use such a system, not knowing
150	whether another module or the main program also does, leading to problems.
151		296
152	This module only lets the main program create pools by forking (because	297	# next, load all modules you need in your template process
153	only the main program can know when it is still safe to do so) - all other	298	use Example::My::Module
154	pools are created by fork+exec, after which such modules can again be	299	use Example::Whatever;
155	loaded.
156		300
157	=back	301	# next, put your run function definition and anything else you
		302	# need, but do not use code outside of BEGIN blocks.
		303	sub worker_run {
		304	my ($fh, @args) = @_;
		305	...
		306	}
		307
		308	# now preserve everything so far as AnyEvent::Fork object
		309	# in §TEMPLATE.
		310	use AnyEvent::Fork::Template;
		311
		312	# do not put code outside of BEGIN blocks until here
		313
		314	# now use the $TEMPLATE process in any way you like
		315
		316	# for example: create 10 worker processes
		317	my @worker;
		318	my $cv = AE::cv;
		319	for (1..10) {
		320	$cv->begin;
		321	$TEMPLATE->fork->send_arg ($_)->run ("worker_run", sub {
		322	push @worker, shift;
		323	$cv->end;
		324	});
		325	}
		326	$cv->recv;
158		327
159	=head1 CONCEPTS	328	=head1 CONCEPTS
160		329
161	This module can create new processes either by executing a new perl	330	This module can create new processes either by executing a new perl
162	process, or by forking from an existing "template" process.	331	process, or by forking from an existing "template" process.
		332
		333	All these processes are called "child processes" (whether they are direct
		334	children or not), while the process that manages them is called the
		335	"parent process".
163		336
164	Each such process comes with its own file handle that can be used to	337	Each such process comes with its own file handle that can be used to
165	communicate with it (it's actually a socket - one end in the new process,	338	communicate with it (it's actually a socket - one end in the new process,
166	one end in the main process), and among the things you can do in it are	339	one end in the main process), and among the things you can do in it are
167	load modules, fork new processes, send file handles to it, and execute	340	load modules, fork new processes, send file handles to it, and execute
…		…
241	my ($fork_fh) = @_;	414	my ($fork_fh) = @_;
242	});	415	});
243		416
244	=back	417	=back
245		418
246	=head1 FUNCTIONS	419	=head1 THE C<AnyEvent::Fork> CLASS
		420
		421	This module exports nothing, and only implements a single class -
		422	C<AnyEvent::Fork>.
		423
		424	There are two class constructors that both create new processes - C<new>
		425	and C<new_exec>. The C<fork> method creates a new process by forking an
		426	existing one and could be considered a third constructor.
		427
		428	Most of the remaining methods deal with preparing the new process, by
		429	loading code, evaluating code and sending data to the new process. They
		430	usually return the process object, so you can chain method calls.
		431
		432	If a process object is destroyed before calling its C<run> method, then
		433	the process simply exits. After C<run> is called, all responsibility is
		434	passed to the specified function.
		435
		436	As long as there is any outstanding work to be done, process objects
		437	resist being destroyed, so there is no reason to store them unless you
		438	need them later - configure and forget works just fine.
247		439
248	=over 4	440	=over 4
249		441
250	=cut	442	=cut
251		443
252	package AnyEvent::Fork;	444	package AnyEvent::Fork;
253		445
254	use common::sense;	446	use common::sense;
255		447
256	use Socket ();	448	use Errno ();
257		449
258	use AnyEvent;	450	use AnyEvent;
259	use AnyEvent::Util ();	451	use AnyEvent::Util ();
260		452
261	use IO::FDPass;	453	use IO::FDPass;
262		454
263	our $VERSION = 0.2;	455	our $VERSION = '1.0';
264
265	our $PERL; # the path to the perl interpreter, deduces with various forms of magic
266
267	=item my $pool = new AnyEvent::Fork key => value...
268
269	Create a new process pool. The following named parameters are supported:
270
271	=over 4
272
273	=back
274
275	=cut
276		456
277	# the early fork template process	457	# the early fork template process
278	our $EARLY;	458	our $EARLY;
279		459
280	# the empty template process	460	# the empty template process
281	our $TEMPLATE;	461	our $TEMPLATE;
282		462
		463	sub QUEUE() { 0 }
		464	sub FH() { 1 }
		465	sub WW() { 2 }
		466	sub PID() { 3 }
		467	sub CB() { 4 }
		468
		469	sub _new {
		470	my ($self, $fh, $pid) = @_;
		471
		472	AnyEvent::Util::fh_nonblocking $fh, 1;
		473
		474	$self = bless [
		475	[], # write queue - strings or fd's
		476	$fh,
		477	undef, # AE watcher
		478	$pid,
		479	], $self;
		480
		481	$self
		482	}
		483
283	sub _cmd {	484	sub _cmd {
284	my $self = shift;	485	my $self = shift;
285		486
286	#TODO: maybe append the packet to any existing string command already in the queue
287
288	# ideally, we would want to use "a (w/a)*" as format string, but perl versions	487	# ideally, we would want to use "a (w/a)*" as format string, but perl
289	# from at least 5.8.9 to 5.16.3 are all buggy and can't unpack it.	488	# versions from at least 5.8.9 to 5.16.3 are all buggy and can't unpack
290	push @{ $self->[2] }, pack "L/a", pack "(w/a)*", @_;	489	# it.
		490	push @{ $self->[QUEUE] }, pack "a L/a*", $_[0], $_[1];
291		491
292	$self->[3] \|\|= AE::io $self->[1], 1, sub {	492	$self->[WW] \|\|= AE::io $self->[FH], 1, sub {
		493	do {
293	# send the next "thing" in the queue - either a reference to an fh,	494	# send the next "thing" in the queue - either a reference to an fh,
294	# or a plain string.	495	# or a plain string.
295		496
296	if (ref $self->[2][0]) {	497	if (ref $self->[QUEUE][0]) {
297	# send fh	498	# send fh
298	IO::FDPass::send fileno $self->[1], fileno ${ $self->[2][0] }	499	unless (IO::FDPass::send fileno $self->[FH], fileno ${ $self->[QUEUE][0] }) {
		500	return if $! == Errno::EAGAIN \|\| $! == Errno::EWOULDBLOCK;
		501	undef $self->[WW];
		502	die "AnyEvent::Fork: file descriptor send failure: $!";
		503	}
		504
299	and shift @{ $self->[2] };	505	shift @{ $self->[QUEUE] };
300		506
301	} else {	507	} else {
302	# send string	508	# send string
303	my $len = syswrite $self->[1], $self->[2][0]	509	my $len = syswrite $self->[FH], $self->[QUEUE][0];
		510
		511	unless ($len) {
		512	return if $! == Errno::EAGAIN \|\| $! == Errno::EWOULDBLOCK;
		513	undef $self->[WW];
304	or do { undef $self->[3]; die "AnyEvent::Fork: command write failure: $!" };	514	die "AnyEvent::Fork: command write failure: $!";
		515	}
305		516
306	substr $self->[2][0], 0, $len, "";	517	substr $self->[QUEUE][0], 0, $len, "";
307	shift @{ $self->[2] } unless length $self->[2][0];	518	shift @{ $self->[QUEUE] } unless length $self->[QUEUE][0];
308	}	519	}
		520	} while @{ $self->[QUEUE] };
309		521
310	unless (@{ $self->[2] }) {	522	# everything written
311	undef $self->[3];	523	undef $self->[WW];
		524
312	# invoke run callback	525	# invoke run callback, if any
		526	if ($self->[CB]) {
313	$self->[0]->($self->[1]) if $self->[0];	527	$self->[CB]->($self->[FH]);
		528	@$self = ();
314	}	529	}
315	};	530	};
316		531
317	() # make sure we don't leak the watcher	532	() # make sure we don't leak the watcher
318	}
319
320	sub _new {
321	my ($self, $fh) = @_;
322
323	AnyEvent::Util::fh_nonblocking $fh, 1;
324
325	$self = bless [
326	undef, # run callback
327	$fh,
328	[], # write queue - strings or fd's
329	undef, # AE watcher
330	], $self;
331
332	$self
333	}	533	}
334		534
335	# fork template from current process, used by AnyEvent::Fork::Early/Template	535	# fork template from current process, used by AnyEvent::Fork::Early/Template
336	sub _new_fork {	536	sub _new_fork {
337	my ($fh, $slave) = AnyEvent::Util::portable_socketpair;	537	my ($fh, $slave) = AnyEvent::Util::portable_socketpair;
…		…
342	if ($pid eq 0) {	542	if ($pid eq 0) {
343	require AnyEvent::Fork::Serve;	543	require AnyEvent::Fork::Serve;
344	$AnyEvent::Fork::Serve::OWNER = $parent;	544	$AnyEvent::Fork::Serve::OWNER = $parent;
345	close $fh;	545	close $fh;
346	$0 = "$_[1] of $parent";	546	$0 = "$_[1] of $parent";
347	$SIG{CHLD} = 'IGNORE';
348	AnyEvent::Fork::Serve::serve ($slave);	547	AnyEvent::Fork::Serve::serve ($slave);
349	exit 0;	548	exit 0;
350	} elsif (!$pid) {	549	} elsif (!$pid) {
351	die "AnyEvent::Fork::Early/Template: unable to fork template process: $!";	550	die "AnyEvent::Fork::Early/Template: unable to fork template process: $!";
352	}	551	}
353		552
354	AnyEvent::Fork->_new ($fh)	553	AnyEvent::Fork->_new ($fh, $pid)
355	}	554	}
356		555
357	=item my $proc = new AnyEvent::Fork	556	=item my $proc = new AnyEvent::Fork
358		557
359	Create a new "empty" perl interpreter process and returns its process	558	Create a new "empty" perl interpreter process and returns its process
360	object for further manipulation.	559	object for further manipulation.
361		560
362	The new process is forked from a template process that is kept around	561	The new process is forked from a template process that is kept around
363	for this purpose. When it doesn't exist yet, it is created by a call to	562	for this purpose. When it doesn't exist yet, it is created by a call to
364	C<new_exec> and kept around for future calls.	563	C<new_exec> first and then stays around for future calls.
365
366	When the process object is destroyed, it will release the file handle
367	that connects it with the new process. When the new process has not yet
368	called C<run>, then the process will exit. Otherwise, what happens depends
369	entirely on the code that is executed.
370		564
371	=cut	565	=cut
372		566
373	sub new {	567	sub new {
374	my $class = shift;	568	my $class = shift;
…		…
452	# quick. also doesn't work in win32. of course. what did you expect	646	# quick. also doesn't work in win32. of course. what did you expect
453	#local $ENV{PERL5LIB} = join ":", grep !ref, @INC;	647	#local $ENV{PERL5LIB} = join ":", grep !ref, @INC;
454	my %env = %ENV;	648	my %env = %ENV;
455	$env{PERL5LIB} = join +($^O eq "MSWin32" ? ";" : ":"), grep !ref, @INC;	649	$env{PERL5LIB} = join +($^O eq "MSWin32" ? ";" : ":"), grep !ref, @INC;
456		650
457	Proc::FastSpawn::spawn (	651	my $pid = Proc::FastSpawn::spawn (
458	$perl,	652	$perl,
459	["perl", "-MAnyEvent::Fork::Serve", "-e", "AnyEvent::Fork::Serve::me", fileno $slave, $$],	653	["perl", "-MAnyEvent::Fork::Serve", "-e", "AnyEvent::Fork::Serve::me", fileno $slave, $$],
460	[map "$_=$env{$_}", keys %env],	654	[map "$_=$env{$_}", keys %env],
461	) or die "unable to spawn AnyEvent::Fork server: $!";	655	) or die "unable to spawn AnyEvent::Fork server: $!";
462		656
463	$self->_new ($fh)	657	$self->_new ($fh, $pid)
		658	}
		659
		660	=item $pid = $proc->pid
		661
		662	Returns the process id of the process I<iff it is a direct child of the
		663	process running AnyEvent::Fork>, and C<undef> otherwise.
		664
		665	Normally, only processes created via C<< AnyEvent::Fork->new_exec >> and
		666	L<AnyEvent::Fork::Template> are direct children, and you are responsible
		667	to clean up their zombies when they die.
		668
		669	All other processes are not direct children, and will be cleaned up by
		670	AnyEvent::Fork itself.
		671
		672	=cut
		673
		674	sub pid {
		675	$_[0][PID]
464	}	676	}
465		677
466	=item $proc = $proc->eval ($perlcode, @args)	678	=item $proc = $proc->eval ($perlcode, @args)
467		679
468	Evaluates the given C<$perlcode> as ... perl code, while setting C<@_> to	680	Evaluates the given C<$perlcode> as ... Perl code, while setting C<@_> to
469	the strings specified by C<@args>.	681	the strings specified by C<@args>, in the "main" package.
470		682
471	This call is meant to do any custom initialisation that might be required	683	This call is meant to do any custom initialisation that might be required
472	(for example, the C<require> method uses it). It's not supposed to be used	684	(for example, the C<require> method uses it). It's not supposed to be used
473	to completely take over the process, use C<run> for that.	685	to completely take over the process, use C<run> for that.
474		686
475	The code will usually be executed after this call returns, and there is no	687	The code will usually be executed after this call returns, and there is no
476	way to pass anything back to the calling process. Any evaluation errors	688	way to pass anything back to the calling process. Any evaluation errors
477	will be reported to stderr and cause the process to exit.	689	will be reported to stderr and cause the process to exit.
478		690
		691	If you want to execute some code (that isn't in a module) to take over the
		692	process, you should compile a function via C<eval> first, and then call
		693	it via C<run>. This also gives you access to any arguments passed via the
		694	C<send_xxx> methods, such as file handles. See the L<use AnyEvent::Fork as
		695	a faster fork+exec> example to see it in action.
		696
479	Returns the process object for easy chaining of method calls.	697	Returns the process object for easy chaining of method calls.
480		698
481	=cut	699	=cut
482		700
483	sub eval {	701	sub eval {
484	my ($self, $code, @args) = @_;	702	my ($self, $code, @args) = @_;
485		703
486	$self->_cmd (e => $code, @args);	704	$self->_cmd (e => pack "(w/a)", $code, @args);
487		705
488	$self	706	$self
489	}	707	}
490		708
491	=item $proc = $proc->require ($module, ...)	709	=item $proc = $proc->require ($module, ...)
…		…
508	=item $proc = $proc->send_fh ($handle, ...)	726	=item $proc = $proc->send_fh ($handle, ...)
509		727
510	Send one or more file handles (I<not> file descriptors) to the process,	728	Send one or more file handles (I<not> file descriptors) to the process,
511	to prepare a call to C<run>.	729	to prepare a call to C<run>.
512		730
513	The process object keeps a reference to the handles until this is done,	731	The process object keeps a reference to the handles until they have
514	so you must not explicitly close the handles. This is most easily	732	been passed over to the process, so you must not explicitly close the
515	accomplished by simply not storing the file handles anywhere after passing	733	handles. This is most easily accomplished by simply not storing the file
516	them to this method.	734	handles anywhere after passing them to this method - when AnyEvent::Fork
		735	is finished using them, perl will automatically close them.
517		736
518	Returns the process object for easy chaining of method calls.	737	Returns the process object for easy chaining of method calls.
519		738
520	Example: pass a file handle to a process, and release it without	739	Example: pass a file handle to a process, and release it without
521	closing. It will be closed automatically when it is no longer used.	740	closing. It will be closed automatically when it is no longer used.
…		…
528	sub send_fh {	747	sub send_fh {
529	my ($self, @fh) = @_;	748	my ($self, @fh) = @_;
530		749
531	for my $fh (@fh) {	750	for my $fh (@fh) {
532	$self->_cmd ("h");	751	$self->_cmd ("h");
533	push @{ $self->[2] }, \$fh;	752	push @{ $self->[QUEUE] }, \$fh;
534	}	753	}
535		754
536	$self	755	$self
537	}	756	}
538		757
539	=item $proc = $proc->send_arg ($string, ...)	758	=item $proc = $proc->send_arg ($string, ...)
540		759
541	Send one or more argument strings to the process, to prepare a call to	760	Send one or more argument strings to the process, to prepare a call to
542	C<run>. The strings can be any octet string.	761	C<run>. The strings can be any octet strings.
		762
		763	The protocol is optimised to pass a moderate number of relatively short
		764	strings - while you can pass up to 4GB of data in one go, this is more
		765	meant to pass some ID information or other startup info, not big chunks of
		766	data.
543		767
544	Returns the process object for easy chaining of method calls.	768	Returns the process object for easy chaining of method calls.
545		769
546	=cut	770	=cut
547		771
548	sub send_arg {	772	sub send_arg {
549	my ($self, @arg) = @_;	773	my ($self, @arg) = @_;
550		774
551	$self->_cmd (a => @arg);	775	$self->_cmd (a => pack "(w/a)", @arg);
552		776
553	$self	777	$self
554	}	778	}
555		779
556	=item $proc->run ($func, $cb->($fh))	780	=item $proc->run ($func, $cb->($fh))
557		781
558	Enter the function specified by the fully qualified name in C<$func> in	782	Enter the function specified by the function name in C<$func> in the
559	the process. The function is called with the communication socket as first	783	process. The function is called with the communication socket as first
560	argument, followed by all file handles and string arguments sent earlier	784	argument, followed by all file handles and string arguments sent earlier
561	via C<send_fh> and C<send_arg> methods, in the order they were called.	785	via C<send_fh> and C<send_arg> methods, in the order they were called.
562		786
563	If the called function returns, the process exits.
564
565	Preparing the process can take time - when the process is ready, the
566	callback is invoked with the local communications socket as argument.
567
568	The process object becomes unusable on return from this function.	787	The process object becomes unusable on return from this function - any
		788	further method calls result in undefined behaviour.
		789
		790	The function name should be fully qualified, but if it isn't, it will be
		791	looked up in the C<main> package.
		792
		793	If the called function returns, doesn't exist, or any error occurs, the
		794	process exits.
		795
		796	Preparing the process is done in the background - when all commands have
		797	been sent, the callback is invoked with the local communications socket
		798	as argument. At this point you can start using the socket in any way you
		799	like.
569		800
570	If the communication socket isn't used, it should be closed on both sides,	801	If the communication socket isn't used, it should be closed on both sides,
571	to save on kernel memory.	802	to save on kernel memory.
572		803
573	The socket is non-blocking in the parent, and blocking in the newly	804	The socket is non-blocking in the parent, and blocking in the newly
574	created process. The close-on-exec flag is set on both. Even if not used	805	created process. The close-on-exec flag is set in both.
		806
575	otherwise, the socket can be a good indicator for the existence of the	807	Even if not used otherwise, the socket can be a good indicator for the
576	process - if the other process exits, you get a readable event on it,	808	existence of the process - if the other process exits, you get a readable
577	because exiting the process closes the socket (if it didn't create any	809	event on it, because exiting the process closes the socket (if it didn't
578	children using fork).	810	create any children using fork).
579		811
580	Example: create a template for a process pool, pass a few strings, some	812	Example: create a template for a process pool, pass a few strings, some
581	file handles, then fork, pass one more string, and run some code.	813	file handles, then fork, pass one more string, and run some code.
582		814
583	my $pool = AnyEvent::Fork	815	my $pool = AnyEvent::Fork
…		…
591	->send_arg ("str3")	823	->send_arg ("str3")
592	->run ("Some::function", sub {	824	->run ("Some::function", sub {
593	my ($fh) = @_;	825	my ($fh) = @_;
594		826
595	# fh is nonblocking, but we trust that the OS can accept these	827	# fh is nonblocking, but we trust that the OS can accept these
596	# extra 3 octets anyway.	828	# few octets anyway.
597	syswrite $fh, "hi #$_\n";	829	syswrite $fh, "hi #$_\n";
598		830
599	# $fh is being closed here, as we don't store it anywhere	831	# $fh is being closed here, as we don't store it anywhere
600	});	832	});
601	}	833	}
…		…
603	# Some::function might look like this - all parameters passed before fork	835	# Some::function might look like this - all parameters passed before fork
604	# and after will be passed, in order, after the communications socket.	836	# and after will be passed, in order, after the communications socket.
605	sub Some::function {	837	sub Some::function {
606	my ($fh, $str1, $str2, $fh1, $fh2, $str3) = @_;	838	my ($fh, $str1, $str2, $fh1, $fh2, $str3) = @_;
607		839
608	print scalar <$fh>; # prints "hi 1\n" and "hi 2\n"	840	print scalar <$fh>; # prints "hi #1\n" and "hi #2\n" in any order
609	}	841	}
610		842
611	=cut	843	=cut
612		844
613	sub run {	845	sub run {
614	my ($self, $func, $cb) = @_;	846	my ($self, $func, $cb) = @_;
615		847
616	$self->[0] = $cb;	848	$self->[CB] = $cb;
617	$self->_cmd (r => $func);	849	$self->_cmd (r => $func);
		850	}
		851
		852	=back
		853
		854	=head2 ADVANCED METHODS
		855
		856	=over 4
		857
		858	=item new_from_stdio AnyEvent::Fork $fh
		859
		860	Assume that you have a perl interpreter running (without any special
		861	options or a program) somewhere and it has it's STDIN and STDOUT connected
		862	to the C<$fh> somehow. I.e. exactly the state perl is in when you start it
		863	without any arguments:
		864
		865	perl
		866
		867	Then you can create an C<AnyEvent::Fork> object out of this perl
		868	interpreter with this constructor.
		869
		870	When the usefulness of this isn't immediately clear, imagine you manage to
		871	run a perl interpreter remotely (F<ssh remotemachine perl>), then you can
		872	manage it mostly like a local C<AnyEvent::Fork> child.
		873
		874	This works without any module support, i.e. the remote F<perl> does not
		875	need to have any special modules installed.
		876
		877	There are a number of limitations though: C<send_fh> will only work if the
		878	L<IO::FDPass> module is loadable by the remote perl and the two processes
		879	are connected in a way that let's L<IO::FDPass> do it's work.
		880
		881	This will therefore not work over a network conenction. From this follows
		882	that C<fork> will also not work under these circumstances, as it relies on
		883	C<send_fh> internally.
		884
		885	=cut
		886
		887	sub new_from_stdio {
		888	my ($class, $fh) = @_;
		889
		890	my $self = $class->_new ($fh);
		891
		892	# send startup code
		893	push @{ $self->[QUEUE] },
		894	(do "AnyEvent/Fork/serve.pl")
		895	. <<'EOF';
		896	{
		897	open my $fh, "+<&0"
		898	or die "AnyEvent::Fork::Serve::stdio: unable to open communications socket: $!\n";
		899	open STDIN , ">&2";
		900	open STDOUT, ">&2";
		901
		902	$OWNER = "another process";
		903	$0 = "AnyEvent::Fork/stdio of $OWNER";
		904
		905	@_ = $fh;
		906	}
		907
		908	&serve;
		909	__END__
		910	EOF
		911
		912	# the data is only sent when the user requests additional things, which
		913	# is likely early enough for our purposes.
		914
		915	$self
		916	}
		917
		918	=back
		919
		920	=head2 EXPERIMENTAL METHODS
		921
		922	These methods might go away completely or change behaviour, a any time.
		923
		924	=over 4
		925
		926	=item $proc->to_fh ($cb->($fh)) # EXPERIMENTAL, MIGHT BE REMOVED
		927
		928	Flushes all commands out to the process and then calls the callback with
		929	the communications socket.
		930
		931	The process object becomes unusable on return from this function - any
		932	further method calls result in undefined behaviour.
		933
		934	The point of this method is to give you a file handle thta you cna pass
		935	to another process. In that other process, you can call C<new_from_fh
		936	AnyEvent::Fork> to create a new C<AnyEvent::Fork> object from it, thereby
		937	effectively passing a fork object to another process.
		938
		939	=cut
		940
		941	sub to_fh {
		942	my ($self, $cb) = @_;
		943
		944	$self->[CB] = $cb;
		945
		946	unless ($self->[WW]) {
		947	$self->[CB]->($self->[FH]);
		948	@$self = ();
		949	}
		950	}
		951
		952	=item new_from_fh AnyEvent::Fork $fh # EXPERIMENTAL, MIGHT BE REMOVED
		953
		954	Takes a file handle originally rceeived by the C<to_fh> method and creates
		955	a new C<AnyEvent:Fork> object. The child process itself will not change in
		956	any way, i.e. it will keep all the modifications done to it before calling
		957	C<to_fh>.
		958
		959	The new object is very much like the original object, except that the
		960	C<pid> method will return C<undef> even if the process is a direct child.
		961
		962	=cut
		963
		964	sub new_from_fh {
		965	my ($class, $fh) = @_;
		966
		967	$class->_new ($fh)
618	}	968	}
619		969
620	=back	970	=back
621		971
622	=head1 PERFORMANCE	972	=head1 PERFORMANCE
623		973
624	Now for some unscientific benchmark numbers (all done on an amd64	974	Now for some unscientific benchmark numbers (all done on an amd64
625	GNU/Linux box). These are intended to give you an idea of the relative	975	GNU/Linux box). These are intended to give you an idea of the relative
626	performance you can expect.	976	performance you can expect, they are not meant to be absolute performance
		977	numbers.
627		978
628	OK, so, I ran a simple benchmark that creates a socket pair, forks, calls	979	OK, so, I ran a simple benchmark that creates a socket pair, forks, calls
629	exit in the child and waits for the socket to close in the parent. I did	980	exit in the child and waits for the socket to close in the parent. I did
630	load AnyEvent, EV and AnyEvent::Fork, for a total process size of 6312kB.	981	load AnyEvent, EV and AnyEvent::Fork, for a total process size of 5100kB.
631		982
632	2079 new processes per second, using socketpair + fork manually	983	2079 new processes per second, using manual socketpair + fork
633		984
634	Then I did the same thing, but instead of calling fork, I called	985	Then I did the same thing, but instead of calling fork, I called
635	AnyEvent::Fork->new->run ("CORE::exit") and then again waited for the	986	AnyEvent::Fork->new->run ("CORE::exit") and then again waited for the
636	socket form the child to close on exit. This does the same thing as manual	987	socket from the child to close on exit. This does the same thing as manual
637	socket pair + fork, except that what is forked is the template process	988	socket pair + fork, except that what is forked is the template process
638	(2440kB), and the socket needs to be passed to the server at the other end	989	(2440kB), and the socket needs to be passed to the server at the other end
639	of the socket first.	990	of the socket first.
640		991
641	2307 new processes per second, using AnyEvent::Fork->new	992	2307 new processes per second, using AnyEvent::Fork->new
…		…
646	479 vfork+execs per second, using AnyEvent::Fork->new_exec	997	479 vfork+execs per second, using AnyEvent::Fork->new_exec
647		998
648	So how can C<< AnyEvent->new >> be faster than a standard fork, even	999	So how can C<< AnyEvent->new >> be faster than a standard fork, even
649	though it uses the same operations, but adds a lot of overhead?	1000	though it uses the same operations, but adds a lot of overhead?
650		1001
651	The difference is simply the process size: forking the 6MB process takes	1002	The difference is simply the process size: forking the 5MB process takes
652	so much longer than forking the 2.5MB template process that the overhead	1003	so much longer than forking the 2.5MB template process that the extra
653	introduced is canceled out.	1004	overhead is canceled out.
654		1005
655	If the benchmark process grows, the normal fork becomes even slower:	1006	If the benchmark process grows, the normal fork becomes even slower:
656		1007
657	1340 new processes, manual fork in a 20MB process	1008	1340 new processes, manual fork of a 20MB process
658	731 new processes, manual fork in a 200MB process	1009	731 new processes, manual fork of a 200MB process
659	235 new processes, manual fork in a 2000MB process	1010	235 new processes, manual fork of a 2000MB process
660		1011
661	What that means (to me) is that I can use this module without having a	1012	What that means (to me) is that I can use this module without having a bad
662	very bad conscience because of the extra overhead required to start new	1013	conscience because of the extra overhead required to start new processes.
663	processes.
664		1014
665	=head1 TYPICAL PROBLEMS	1015	=head1 TYPICAL PROBLEMS
666		1016
667	This section lists typical problems that remain. I hope by recognising	1017	This section lists typical problems that remain. I hope by recognising
668	them, most can be avoided.	1018	them, most can be avoided.
669		1019
670	=over 4	1020	=over 4
671		1021
672	=item exit runs destructors
673
674	=item "leaked" file descriptors for exec'ed processes	1022	=item leaked file descriptors for exec'ed processes
675		1023
676	POSIX systems inherit file descriptors by default when exec'ing a new	1024	POSIX systems inherit file descriptors by default when exec'ing a new
677	process. While perl itself laudably sets the close-on-exec flags on new	1025	process. While perl itself laudably sets the close-on-exec flags on new
678	file handles, most C libraries don't care, and even if all cared, it's	1026	file handles, most C libraries don't care, and even if all cared, it's
679	often not possible to set the flag in a race-free manner.	1027	often not possible to set the flag in a race-free manner.
…		…
699	libraries or the code that leaks those file descriptors.	1047	libraries or the code that leaks those file descriptors.
700		1048
701	Fortunately, most of these leaked descriptors do no harm, other than	1049	Fortunately, most of these leaked descriptors do no harm, other than
702	sitting on some resources.	1050	sitting on some resources.
703		1051
704	=item "leaked" file descriptors for fork'ed processes	1052	=item leaked file descriptors for fork'ed processes
705		1053
706	Normally, L<AnyEvent::Fork> does start new processes by exec'ing them,	1054	Normally, L<AnyEvent::Fork> does start new processes by exec'ing them,
707	which closes file descriptors not marked for being inherited.	1055	which closes file descriptors not marked for being inherited.
708		1056
709	However, L<AnyEvent::Fork::Early> and L<AnyEvent::Fork::Template> offer	1057	However, L<AnyEvent::Fork::Early> and L<AnyEvent::Fork::Template> offer
…		…
717	trouble with a fork.	1065	trouble with a fork.
718		1066
719	The solution is to either not load these modules before use'ing	1067	The solution is to either not load these modules before use'ing
720	L<AnyEvent::Fork::Early> or L<AnyEvent::Fork::Template>, or to delay	1068	L<AnyEvent::Fork::Early> or L<AnyEvent::Fork::Template>, or to delay
721	initialising them, for example, by calling C<init Gtk2> manually.	1069	initialising them, for example, by calling C<init Gtk2> manually.
		1070
		1071	=item exiting calls object destructors
		1072
		1073	This only applies to users of L<AnyEvent::Fork:Early> and
		1074	L<AnyEvent::Fork::Template>, or when initialising code creates objects
		1075	that reference external resources.
		1076
		1077	When a process created by AnyEvent::Fork exits, it might do so by calling
		1078	exit, or simply letting perl reach the end of the program. At which point
		1079	Perl runs all destructors.
		1080
		1081	Not all destructors are fork-safe - for example, an object that represents
		1082	the connection to an X display might tell the X server to free resources,
		1083	which is inconvenient when the "real" object in the parent still needs to
		1084	use them.
		1085
		1086	This is obviously not a problem for L<AnyEvent::Fork::Early>, as you used
		1087	it as the very first thing, right?
		1088
		1089	It is a problem for L<AnyEvent::Fork::Template> though - and the solution
		1090	is to not create objects with nontrivial destructors that might have an
		1091	effect outside of Perl.
722		1092
723	=back	1093	=back
724		1094
725	=head1 PORTABILITY NOTES	1095	=head1 PORTABILITY NOTES
726		1096
…		…
729	to make it so, mostly due to the bloody broken perl that nobody seems to	1099	to make it so, mostly due to the bloody broken perl that nobody seems to
730	care about. The fork emulation is a bad joke - I have yet to see something	1100	care about. The fork emulation is a bad joke - I have yet to see something
731	useful that you can do with it without running into memory corruption	1101	useful that you can do with it without running into memory corruption
732	issues or other braindamage. Hrrrr.	1102	issues or other braindamage. Hrrrr.
733		1103
734	Cygwin perl is not supported at the moment, as it should implement fd	1104	Since fork is endlessly broken on win32 perls (it doesn't even remotely
735	passing, but doesn't, and rolling my own is hard, as cygwin doesn't	1105	work within it's documented limits) and quite obviously it's not getting
736	support enough functionality to do it.	1106	improved any time soon, the best way to proceed on windows would be to
		1107	always use C<new_exec> and thus never rely on perl's fork "emulation".
		1108
		1109	Cygwin perl is not supported at the moment due to some hilarious
		1110	shortcomings of its API - see L<IO::FDPoll> for more details. If you never
		1111	use C<send_fh> and always use C<new_exec> to create processes, it should
		1112	work though.
737		1113
738	=head1 SEE ALSO	1114	=head1 SEE ALSO
739		1115
740	L<AnyEvent::Fork::Early> (to avoid executing a perl interpreter),	1116	L<AnyEvent::Fork::Early>, to avoid executing a perl interpreter at all
		1117	(part of this distribution).
		1118
741	L<AnyEvent::Fork::Template> (to create a process by forking the main	1119	L<AnyEvent::Fork::Template>, to create a process by forking the main
742	program at a convenient time).	1120	program at a convenient time (part of this distribution).
743		1121
744	=head1 AUTHOR	1122	L<AnyEvent::Fork::RPC>, for simple RPC to child processes (on CPAN).
		1123
		1124	L<AnyEvent::Fork::Pool>, for simple worker process pool (on CPAN).
		1125
		1126	=head1 AUTHOR AND CONTACT INFORMATION
745		1127
746	Marc Lehmann <schmorp@schmorp.de>	1128	Marc Lehmann <schmorp@schmorp.de>
747	http://home.schmorp.de/	1129	http://software.schmorp.de/pkg/AnyEvent-Fork
748		1130
749	=cut	1131	=cut
750		1132
751	1	1133	1
752		1134

Diff Legend

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

Comparing AnyEvent-Fork/Fork.pm (file contents): Revision 1.17 by root, Fri Apr 5 23:42:24 2013 UTC vs. Revision 1.53 by root, Fri Apr 26 15:44:44 2013 UTC

Diff Legend

Comparing AnyEvent-Fork/Fork.pm (file contents):
Revision 1.17 by root, Fri Apr 5 23:42:24 2013 UTC vs.
Revision 1.53 by root, Fri Apr 26 15:44:44 2013 UTC