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

Comparing AnyEvent-Fork/Fork.pm (file contents):
Revision 1.24 by root, Sat Apr 6 08:32:23 2013 UTC vs.
Revision 1.67 by root, Thu May 12 16:54:01 2016 UTC

…		…
27		27
28	Special care has been taken to make this module useful from other modules,	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>	29	while still supporting specialised environments such as L<App::Staticperl>
30	or L<PAR::Packer>.	30	or L<PAR::Packer>.
31		31
32	=head1 WHAT THIS MODULE IS NOT	32	=head2 WHAT THIS MODULE IS NOT
33		33
34	This module only creates processes and lets you pass file handles and	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 -	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	36	there is no back channel from the process back to you, and there is no RPC
37	or message passing going on.	37	or message passing going on.
38		38
39	If you need some form of RPC, you can either implement it yourself	39	If you need some form of RPC, you could use the L<AnyEvent::Fork::RPC>
40	in whatever way you like, use some message-passing module such	40	companion module, which adds simple RPC/job queueing to a process created
41	as L<AnyEvent::MP>, some pipe such as L<AnyEvent::ZeroMQ>, use	41	by this module.
42	L<AnyEvent::Handle> on both sides to send e.g. JSON or Storable messages,	42
43	and so on.	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 or not being able to use event processing, GUI
		62	toolkits or 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. Exec'ing a new process, in contrast, means modules
		95	and data files might need to be loaded again, at extra CPU and memory
		96	cost.
		97
		98	But when forking, you still create a copy of your data structures - if
		99	the program frees them and replaces them by new data, the child processes
		100	will retain the old version even if it isn't used, which can suddenly and
		101	unexpectedly increase memory usage when freeing memory.
		102
		103	For example, L<Gtk2::CV> is an image viewer optimised for large
		104	directories (millions of pictures). It also forks subprocesses for
		105	thumbnail generation, which inherit the data structure that stores all
		106	file information. If the user changes the directory, it gets freed in
		107	the main process, leaving a copy in the thumbnailer processes. This can
		108	lead to many times the memory usage that would actually be required. The
		109	solution is to fork early (and being unable to dynamically generate more
		110	subprocesses or do this from a module)... or to use L<AnyEvent:Fork>.
		111
		112	There is a trade-off between more sharing with fork (which can be good or
		113	bad), and no sharing with exec.
		114
		115	This module allows the main program to do a controlled fork, and allows
		116	modules to exec processes safely at any time. When creating a custom
		117	process pool you can take advantage of data sharing via fork without
		118	risking to share large dynamic data structures that will blow up child
		119	memory usage.
		120
		121	In other words, this module puts you into control over what is being
		122	shared and what isn't, at all times.
		123
		124	=item Exec'ing a new perl process might be difficult.
		125
		126	For example, it is not easy to find the correct path to the perl
		127	interpreter - C<$^X> might not be a perl interpreter at all. Worse, there
		128	might not even be a perl binary installed on the system.
		129
		130	This module tries hard to identify the correct path to the perl
		131	interpreter. With a cooperative main program, exec'ing the interpreter
		132	might not even be necessary, but even without help from the main program,
		133	it will still work when used from a module.
		134
		135	=item Exec'ing a new perl process might be slow, as all necessary modules
		136	have to be loaded from disk again, with no guarantees of success.
		137
		138	Long running processes might run into problems when perl is upgraded
		139	and modules are no longer loadable because they refer to a different
		140	perl version, or parts of a distribution are newer than the ones already
		141	loaded.
		142
		143	This module supports creating pre-initialised perl processes to be used as
		144	a template for new processes at a later time, e.g. for use in a process
		145	pool.
		146
		147	=item Forking might be impossible when a program is running.
		148
		149	For example, POSIX makes it almost impossible to fork from a
		150	multi-threaded program while doing anything useful in the child - in
		151	fact, if your perl program uses POSIX threads (even indirectly via
		152	e.g. L<IO::AIO> or L<threads>), you cannot call fork on the perl level
		153	anymore without risking memory corruption or worse on a number of
		154	operating systems.
		155
		156	This module can safely fork helper processes at any time, by calling
		157	fork+exec in C, in a POSIX-compatible way (via L<Proc::FastSpawn>).
		158
		159	=item Parallel processing with fork might be inconvenient or difficult
		160	to implement. Modules might not work in both parent and child.
		161
		162	For example, when a program uses an event loop and creates watchers it
		163	becomes very hard to use the event loop from a child program, as the
		164	watchers already exist but are only meaningful in the parent. Worse, a
		165	module might want to use such a module, not knowing whether another module
		166	or the main program also does, leading to problems.
		167
		168	Apart from event loops, graphical toolkits also commonly fall into the
		169	"unsafe module" category, or just about anything that communicates with
		170	the external world, such as network libraries and file I/O modules, which
		171	usually don't like being copied and then allowed to continue in two
		172	processes.
		173
		174	With this module only the main program is allowed to create new processes
		175	by forking (because only the main program can know when it is still safe
		176	to do so) - all other processes are created via fork+exec, which makes it
		177	possible to use modules such as event loops or window interfaces safely.
		178
		179	=back
44		180
45	=head1 EXAMPLES	181	=head1 EXAMPLES
		182
		183	This is where the wall of text ends and code speaks.
46		184
47	=head2 Create a single new process, tell it to run your worker function.	185	=head2 Create a single new process, tell it to run your worker function.
48		186
49	AnyEvent::Fork	187	AnyEvent::Fork
50	->new	188	->new
…		…
54		192
55	# now $master_filehandle is connected to the	193	# now $master_filehandle is connected to the
56	# $slave_filehandle in the new process.	194	# $slave_filehandle in the new process.
57	});	195	});
58		196
59	# MyModule::worker might look like this	197	C<MyModule> might look like this:
		198
		199	package MyModule;
		200
60	sub MyModule::worker {	201	sub worker {
61	my ($slave_filehandle) = @_;	202	my ($slave_filehandle) = @_;
62		203
63	# now $slave_filehandle is connected to the $master_filehandle	204	# now $slave_filehandle is connected to the $master_filehandle
64	# in the original prorcess. have fun!	205	# in the original process. have fun!
65	}	206	}
66		207
67	=head2 Create a pool of server processes all accepting on the same socket.	208	=head2 Create a pool of server processes all accepting on the same socket.
68		209
69	# create listener socket	210	# create listener socket
…		…
84	}	225	}
85		226
86	# now do other things - maybe use the filehandle provided by run	227	# now do other things - maybe use the filehandle provided by run
87	# to wait for the processes to die. or whatever.	228	# to wait for the processes to die. or whatever.
88		229
89	# My::Server::run might look like this	230	C<My::Server> might look like this:
90	sub My::Server::run {	231
		232	package My::Server;
		233
		234	sub run {
91	my ($slave, $listener, $id) = @_;	235	my ($slave, $listener, $id) = @_;
92		236
93	close $slave; # we do not use the socket, so close it to save resources	237	close $slave; # we do not use the socket, so close it to save resources
94		238
95	# we could go ballistic and use e.g. AnyEvent here, or IO::AIO,	239	# we could go ballistic and use e.g. AnyEvent here, or IO::AIO,
…		…
99	}	243	}
100	}	244	}
101		245
102	=head2 use AnyEvent::Fork as a faster fork+exec	246	=head2 use AnyEvent::Fork as a faster fork+exec
103		247
104	This runs /bin/echo hi, with stdout redirected to /tmp/log and stderr to	248	This runs C</bin/echo hi>, with standard output redirected to F</tmp/log>
105	the communications socket. It is usually faster than fork+exec, but still	249	and standard error redirected to the communications socket. It is usually
106	let's you prepare the environment.	250	faster than fork+exec, but still lets you prepare the environment.
107		251
108	open my $output, ">/tmp/log" or die "$!";	252	open my $output, ">/tmp/log" or die "$!";
109		253
110	AnyEvent::Fork	254	AnyEvent::Fork
111	->new	255	->new
112	->eval ('	256	->eval ('
		257	# compile a helper function for later use
113	sub run {	258	sub run {
114	my ($fh, $output, @cmd) = @_;	259	my ($fh, $output, @cmd) = @_;
115		260
116	# perl will clear close-on-exec on STDOUT/STDERR	261	# perl will clear close-on-exec on STDOUT/STDERR
117	open STDOUT, ">&", $output or die;	262	open STDOUT, ">&", $output or die;
…		…
124	->send_arg ("/bin/echo", "hi")	269	->send_arg ("/bin/echo", "hi")
125	->run ("run", my $cv = AE::cv);	270	->run ("run", my $cv = AE::cv);
126		271
127	my $stderr = $cv->recv;	272	my $stderr = $cv->recv;
128		273
129	=head1 PROBLEM STATEMENT	274	=head2 For stingy users: put the worker code into a C<DATA> section.
130		275
131	There are two ways to implement parallel processing on UNIX like operating	276	When you want to be stingy with files, you can put your code into the
132	systems - fork and process, and fork+exec and process. They have different	277	C<DATA> section of your module (or program):
133	advantages and disadvantages that I describe below, together with how this
134	module tries to mitigate the disadvantages.
135		278
136	=over 4	279	use AnyEvent::Fork;
137		280
138	=item Forking from a big process can be very slow (a 5GB process needs	281	AnyEvent::Fork
139	0.05s to fork on my 3.6GHz amd64 GNU/Linux box for example). This overhead	282	->new
140	is often shared with exec (because you have to fork first), but in some	283	->eval (do { local $/; <DATA> })
141	circumstances (e.g. when vfork is used), fork+exec can be much faster.	284	->run ("doit", sub { ... });
142		285
143	This module can help here by telling a small(er) helper process to fork,	286	__DATA__
144	or fork+exec instead.
145		287
146	=item Forking usually creates a copy-on-write copy of the parent	288	sub doit {
147	process. Memory (for example, modules or data files that have been	289	... do something!
148	will not take additional memory). When exec'ing a new process, modules	290	}
149	and data files might need to be loaded again, at extra CPU and memory
150	cost. Likewise when forking, all data structures are copied as well - if
151	the program frees them and replaces them by new data, the child processes
152	will retain the memory even if it isn't used.
153		291
154	This module allows the main program to do a controlled fork, and allows	292	=head2 For stingy standalone programs: do not rely on external files at
155	modules to exec processes safely at any time. When creating a custom	293	all.
156	process pool you can take advantage of data sharing via fork without
157	risking to share large dynamic data structures that will blow up child
158	memory usage.
159		294
160	=item Exec'ing a new perl process might be difficult and slow. For	295	For single-file scripts it can be inconvenient to rely on external
161	example, it is not easy to find the correct path to the perl interpreter,	296	files - even when using a C<DATA> section, you still need to C<exec> an
162	and all modules have to be loaded from disk again. Long running processes	297	external perl interpreter, which might not be available when using
163	might run into problems when perl is upgraded for example.	298	L<App::Staticperl>, L<Urlader> or L<PAR::Packer> for example.
164		299
165	This module supports creating pre-initialised perl processes to be used	300	Two modules help here - L<AnyEvent::Fork::Early> forks a template process
166	as template, and also tries hard to identify the correct path to the perl	301	for all further calls to C<new_exec>, and L<AnyEvent::Fork::Template>
167	interpreter. With a cooperative main program, exec'ing the interpreter	302	forks the main program as a template process.
168	might not even be necessary.
169		303
170	=item Forking might be impossible when a program is running. For example,	304	Here is how your main program should look like:
171	POSIX makes it almost impossible to fork from a multi-threaded program and
172	do anything useful in the child - strictly speaking, if your perl program
173	uses posix threads (even indirectly via e.g. L<IO::AIO> or L<threads>),
174	you cannot call fork on the perl level anymore, at all.
175		305
176	This module can safely fork helper processes at any time, by calling	306	#! perl
177	fork+exec in C, in a POSIX-compatible way.
178		307
179	=item Parallel processing with fork might be inconvenient or difficult	308	# optional, as the very first thing.
180	to implement. For example, when a program uses an event loop and creates	309	# in case modules want to create their own processes.
181	watchers it becomes very hard to use the event loop from a child	310	use AnyEvent::Fork::Early;
182	program, as the watchers already exist but are only meaningful in the
183	parent. Worse, a module might want to use such a system, not knowing
184	whether another module or the main program also does, leading to problems.
185		311
186	This module only lets the main program create pools by forking (because	312	# next, load all modules you need in your template process
187	only the main program can know when it is still safe to do so) - all other	313	use Example::My::Module
188	pools are created by fork+exec, after which such modules can again be	314	use Example::Whatever;
189	loaded.
190		315
191	=back	316	# next, put your run function definition and anything else you
		317	# need, but do not use code outside of BEGIN blocks.
		318	sub worker_run {
		319	my ($fh, @args) = @_;
		320	...
		321	}
		322
		323	# now preserve everything so far as AnyEvent::Fork object
		324	# in $TEMPLATE.
		325	use AnyEvent::Fork::Template;
		326
		327	# do not put code outside of BEGIN blocks until here
		328
		329	# now use the $TEMPLATE process in any way you like
		330
		331	# for example: create 10 worker processes
		332	my @worker;
		333	my $cv = AE::cv;
		334	for (1..10) {
		335	$cv->begin;
		336	$TEMPLATE->fork->send_arg ($_)->run ("worker_run", sub {
		337	push @worker, shift;
		338	$cv->end;
		339	});
		340	}
		341	$cv->recv;
192		342
193	=head1 CONCEPTS	343	=head1 CONCEPTS
194		344
195	This module can create new processes either by executing a new perl	345	This module can create new processes either by executing a new perl
196	process, or by forking from an existing "template" process.	346	process, or by forking from an existing "template" process.
		347
		348	All these processes are called "child processes" (whether they are direct
		349	children or not), while the process that manages them is called the
		350	"parent process".
197		351
198	Each such process comes with its own file handle that can be used to	352	Each such process comes with its own file handle that can be used to
199	communicate with it (it's actually a socket - one end in the new process,	353	communicate with it (it's actually a socket - one end in the new process,
200	one end in the main process), and among the things you can do in it are	354	one end in the main process), and among the things you can do in it are
201	load modules, fork new processes, send file handles to it, and execute	355	load modules, fork new processes, send file handles to it, and execute
…		…
275	my ($fork_fh) = @_;	429	my ($fork_fh) = @_;
276	});	430	});
277		431
278	=back	432	=back
279		433
280	=head1 FUNCTIONS	434	=head1 THE C<AnyEvent::Fork> CLASS
		435
		436	This module exports nothing, and only implements a single class -
		437	C<AnyEvent::Fork>.
		438
		439	There are two class constructors that both create new processes - C<new>
		440	and C<new_exec>. The C<fork> method creates a new process by forking an
		441	existing one and could be considered a third constructor.
		442
		443	Most of the remaining methods deal with preparing the new process, by
		444	loading code, evaluating code and sending data to the new process. They
		445	usually return the process object, so you can chain method calls.
		446
		447	If a process object is destroyed before calling its C<run> method, then
		448	the process simply exits. After C<run> is called, all responsibility is
		449	passed to the specified function.
		450
		451	As long as there is any outstanding work to be done, process objects
		452	resist being destroyed, so there is no reason to store them unless you
		453	need them later - configure and forget works just fine.
281		454
282	=over 4	455	=over 4
283		456
284	=cut	457	=cut
285		458
…		…
292	use AnyEvent;	465	use AnyEvent;
293	use AnyEvent::Util ();	466	use AnyEvent::Util ();
294		467
295	use IO::FDPass;	468	use IO::FDPass;
296		469
297	our $VERSION = 0.5;	470	our $VERSION = 1.3;
298
299	our $PERL; # the path to the perl interpreter, deduces with various forms of magic
300
301	=item my $pool = new AnyEvent::Fork key => value...
302
303	Create a new process pool. The following named parameters are supported:
304
305	=over 4
306
307	=back
308
309	=cut
310		471
311	# the early fork template process	472	# the early fork template process
312	our $EARLY;	473	our $EARLY;
313		474
314	# the empty template process	475	# the empty template process
315	our $TEMPLATE;	476	our $TEMPLATE;
		477
		478	sub QUEUE() { 0 }
		479	sub FH() { 1 }
		480	sub WW() { 2 }
		481	sub PID() { 3 }
		482	sub CB() { 4 }
		483
		484	sub _new {
		485	my ($self, $fh, $pid) = @_;
		486
		487	AnyEvent::Util::fh_nonblocking $fh, 1;
		488
		489	$self = bless [
		490	[], # write queue - strings or fd's
		491	$fh,
		492	undef, # AE watcher
		493	$pid,
		494	], $self;
		495
		496	$self
		497	}
316		498
317	sub _cmd {	499	sub _cmd {
318	my $self = shift;	500	my $self = shift;
319		501
320	# ideally, we would want to use "a (w/a)*" as format string, but perl	502	# ideally, we would want to use "a (w/a)*" as format string, but perl
321	# versions from at least 5.8.9 to 5.16.3 are all buggy and can't unpack	503	# versions from at least 5.8.9 to 5.16.3 are all buggy and can't unpack
322	# it.	504	# it.
323	push @{ $self->[2] }, pack "a L/a*", $_[0], $_[1];	505	push @{ $self->[QUEUE] }, pack "a L/a*", $_[0], $_[1];
324		506
325	$self->[3] \|\|= AE::io $self->[1], 1, sub {	507	$self->[WW] \|\|= AE::io $self->[FH], 1, sub {
326	do {	508	do {
327	# send the next "thing" in the queue - either a reference to an fh,	509	# send the next "thing" in the queue - either a reference to an fh,
328	# or a plain string.	510	# or a plain string.
329		511
330	if (ref $self->[2][0]) {	512	if (ref $self->[QUEUE][0]) {
331	# send fh	513	# send fh
332	unless (IO::FDPass::send fileno $self->[1], fileno ${ $self->[2][0] }) {	514	unless (IO::FDPass::send fileno $self->[FH], fileno ${ $self->[QUEUE][0] }) {
333	return if $! == Errno::EAGAIN \|\| $! == Errno::EWOULDBLOCK;	515	return if $! == Errno::EAGAIN \|\| $! == Errno::EWOULDBLOCK;
334	undef $self->[3];	516	undef $self->[WW];
335	die "AnyEvent::Fork: file descriptor send failure: $!";	517	die "AnyEvent::Fork: file descriptor send failure: $!";
336	}	518	}
337		519
338	shift @{ $self->[2] };	520	shift @{ $self->[QUEUE] };
339		521
340	} else {	522	} else {
341	# send string	523	# send string
342	my $len = syswrite $self->[1], $self->[2][0];	524	my $len = syswrite $self->[FH], $self->[QUEUE][0];
343		525
344	unless ($len) {	526	unless ($len) {
345	return if $! == Errno::EAGAIN \|\| $! == Errno::EWOULDBLOCK;	527	return if $! == Errno::EAGAIN \|\| $! == Errno::EWOULDBLOCK;
346	undef $self->[3];	528	undef $self->[WW];
347	die "AnyEvent::Fork: command write failure: $!";	529	die "AnyEvent::Fork: command write failure: $!";
348	}	530	}
349		531
350	substr $self->[2][0], 0, $len, "";	532	substr $self->[QUEUE][0], 0, $len, "";
351	shift @{ $self->[2] } unless length $self->[2][0];	533	shift @{ $self->[QUEUE] } unless length $self->[QUEUE][0];
352	}	534	}
353	} while @{ $self->[2] };	535	} while @{ $self->[QUEUE] };
354		536
355	# everything written	537	# everything written
356	undef $self->[3];	538	undef $self->[WW];
357		539
358	# invoke run callback, if any	540	# invoke run callback, if any
359	$self->[4]->($self->[1]) if $self->[4];	541	if ($self->[CB]) {
		542	$self->[CB]->($self->[FH]);
		543	@$self = ();
		544	}
360	};	545	};
361		546
362	() # make sure we don't leak the watcher	547	() # make sure we don't leak the watcher
363	}
364
365	sub _new {
366	my ($self, $fh, $pid) = @_;
367
368	AnyEvent::Util::fh_nonblocking $fh, 1;
369
370	$self = bless [
371	$pid,
372	$fh,
373	[], # write queue - strings or fd's
374	undef, # AE watcher
375	], $self;
376
377	$self
378	}	548	}
379		549
380	# fork template from current process, used by AnyEvent::Fork::Early/Template	550	# fork template from current process, used by AnyEvent::Fork::Early/Template
381	sub _new_fork {	551	sub _new_fork {
382	my ($fh, $slave) = AnyEvent::Util::portable_socketpair;	552	my ($fh, $slave) = AnyEvent::Util::portable_socketpair;
…		…
387	if ($pid eq 0) {	557	if ($pid eq 0) {
388	require AnyEvent::Fork::Serve;	558	require AnyEvent::Fork::Serve;
389	$AnyEvent::Fork::Serve::OWNER = $parent;	559	$AnyEvent::Fork::Serve::OWNER = $parent;
390	close $fh;	560	close $fh;
391	$0 = "$_[1] of $parent";	561	$0 = "$_[1] of $parent";
392	$SIG{CHLD} = 'IGNORE';
393	AnyEvent::Fork::Serve::serve ($slave);	562	AnyEvent::Fork::Serve::serve ($slave);
394	exit 0;	563	exit 0;
395	} elsif (!$pid) {	564	} elsif (!$pid) {
396	die "AnyEvent::Fork::Early/Template: unable to fork template process: $!";	565	die "AnyEvent::Fork::Early/Template: unable to fork template process: $!";
397	}	566	}
…		…
404	Create a new "empty" perl interpreter process and returns its process	573	Create a new "empty" perl interpreter process and returns its process
405	object for further manipulation.	574	object for further manipulation.
406		575
407	The new process is forked from a template process that is kept around	576	The new process is forked from a template process that is kept around
408	for this purpose. When it doesn't exist yet, it is created by a call to	577	for this purpose. When it doesn't exist yet, it is created by a call to
409	C<new_exec> and kept around for future calls.	578	C<new_exec> first and then stays around for future calls.
410
411	When the process object is destroyed, it will release the file handle
412	that connects it with the new process. When the new process has not yet
413	called C<run>, then the process will exit. Otherwise, what happens depends
414	entirely on the code that is executed.
415		579
416	=cut	580	=cut
417		581
418	sub new {	582	sub new {
419	my $class = shift;	583	my $class = shift;
…		…
456		620
457	You should use C<new> whenever possible, except when having a template	621	You should use C<new> whenever possible, except when having a template
458	process around is unacceptable.	622	process around is unacceptable.
459		623
460	The path to the perl interpreter is divined using various methods - first	624	The path to the perl interpreter is divined using various methods - first
461	C<$^X> is investigated to see if the path ends with something that sounds	625	C<$^X> is investigated to see if the path ends with something that looks
462	as if it were the perl interpreter. Failing this, the module falls back to	626	as if it were the perl interpreter. Failing this, the module falls back to
463	using C<$Config::Config{perlpath}>.	627	using C<$Config::Config{perlpath}>.
464		628
		629	The path to perl can also be overriden by setting the global variable
		630	C<$AnyEvent::Fork::PERL> - it's value will be used for all subsequent
		631	invocations.
		632
465	=cut	633	=cut
		634
		635	our $PERL;
466		636
467	sub new_exec {	637	sub new_exec {
468	my ($self) = @_;	638	my ($self) = @_;
469		639
470	return $EARLY->fork	640	return $EARLY->fork
471	if $EARLY;	641	if $EARLY;
472		642
		643	unless (defined $PERL) {
473	# first find path of perl	644	# first find path of perl
474	my $perl = $;	645	my $perl = $^X;
475		646
476	# first we try $^X, but the path must be absolute (always on win32), and end in sth.	647	# first we try $^X, but the path must be absolute (always on win32), and end in sth.
477	# that looks like perl. this obviously only works for posix and win32	648	# that looks like perl. this obviously only works for posix and win32
478	unless (	649	unless (
479	($^O eq "MSWin32" \|\| $perl =~ m%^/%)	650	($^O eq "MSWin32" \|\| $perl =~ m%^/%)
480	&& $perl =~ m%[/\\]perl(?:[0-9]+(\.[0-9]+)+)?(\.exe)?$%i	651	&& $perl =~ m%[/\\]perl(?:[0-9]+(\.[0-9]+)+)?(\.exe)?$%i
481	) {	652	) {
482	# if it doesn't look perlish enough, try Config	653	# if it doesn't look perlish enough, try Config
483	require Config;	654	require Config;
484	$perl = $Config::Config{perlpath};	655	$perl = $Config::Config{perlpath};
485	$perl =~ s/(?:\Q$Config::Config{_exe}\E)?$/$Config::Config{_exe}/;	656	$perl =~ s/(?:\Q$Config::Config{_exe}\E)?$/$Config::Config{_exe}/;
		657	}
		658
		659	$PERL = $perl;
486	}	660	}
487		661
488	require Proc::FastSpawn;	662	require Proc::FastSpawn;
489		663
490	my ($fh, $slave) = AnyEvent::Util::portable_socketpair;	664	my ($fh, $slave) = AnyEvent::Util::portable_socketpair;
…		…
498	#local $ENV{PERL5LIB} = join ":", grep !ref, @INC;	672	#local $ENV{PERL5LIB} = join ":", grep !ref, @INC;
499	my %env = %ENV;	673	my %env = %ENV;
500	$env{PERL5LIB} = join +($^O eq "MSWin32" ? ";" : ":"), grep !ref, @INC;	674	$env{PERL5LIB} = join +($^O eq "MSWin32" ? ";" : ":"), grep !ref, @INC;
501		675
502	my $pid = Proc::FastSpawn::spawn (	676	my $pid = Proc::FastSpawn::spawn (
503	$perl,	677	$PERL,
504	["perl", "-MAnyEvent::Fork::Serve", "-e", "AnyEvent::Fork::Serve::me", fileno $slave, $$],	678	[$PERL, "-MAnyEvent::Fork::Serve", "-e", "AnyEvent::Fork::Serve::me", fileno $slave, $$],
505	[map "$_=$env{$_}", keys %env],	679	[map "$_=$env{$_}", keys %env],
506	) or die "unable to spawn AnyEvent::Fork server: $!";	680	) or die "unable to spawn AnyEvent::Fork server: $!";
507		681
508	$self->_new ($fh, $pid)	682	$self->_new ($fh, $pid)
509	}	683	}
510		684
511	=item $pid = $proc->pid	685	=item $pid = $proc->pid
512		686
513	Returns the process id of the process I<iff it is a direct child of the	687	Returns the process id of the process I<iff it is a direct child of the
514	process> running AnyEvent::Fork, and C<undef> otherwise.	688	process running AnyEvent::Fork>, and C<undef> otherwise. As a general
		689	rule (that you cannot rely upon), processes created via C<new_exec>,
		690	L<AnyEvent::Fork::Early> or L<AnyEvent::Fork::Template> are direct
		691	children, while all other processes are not.
515		692
516	Normally, only processes created via C<< AnyEvent::Fork->new_exec >> and	693	Or in other words, you do not normally have to take care of zombies for
517	L<AnyEvent::Fork::Template> are direct children, and you are responsible	694	processes created via C<new>, but when in doubt, or zombies are a problem,
518	to clean up their zombies when they die.	695	you need to check whether a process is a diretc child by calling this
519		696	method, and possibly creating a child watcher or reap it manually.
520	All other processes are not direct children, and will be cleaned up by
521	AnyEvent::Fork.
522		697
523	=cut	698	=cut
524		699
525	sub pid {	700	sub pid {
526	$_[0][0]	701	$_[0][PID]
527	}	702	}
528		703
529	=item $proc = $proc->eval ($perlcode, @args)	704	=item $proc = $proc->eval ($perlcode, @args)
530		705
531	Evaluates the given C<$perlcode> as ... perl code, while setting C<@_> to	706	Evaluates the given C<$perlcode> as ... Perl code, while setting C<@_> to
532	the strings specified by C<@args>, in the "main" package.	707	the strings specified by C<@args>, in the "main" package.
533		708
534	This call is meant to do any custom initialisation that might be required	709	This call is meant to do any custom initialisation that might be required
535	(for example, the C<require> method uses it). It's not supposed to be used	710	(for example, the C<require> method uses it). It's not supposed to be used
536	to completely take over the process, use C<run> for that.	711	to completely take over the process, use C<run> for that.
537		712
538	The code will usually be executed after this call returns, and there is no	713	The code will usually be executed after this call returns, and there is no
539	way to pass anything back to the calling process. Any evaluation errors	714	way to pass anything back to the calling process. Any evaluation errors
540	will be reported to stderr and cause the process to exit.	715	will be reported to stderr and cause the process to exit.
541		716
542	If you want to execute some code to take over the process (see the	717	If you want to execute some code (that isn't in a module) to take over the
543	"fork+exec" example in the SYNOPSIS), you should compile a function via	718	process, you should compile a function via C<eval> first, and then call
544	C<eval> first, and then call it via C<run>. This also gives you access to	719	it via C<run>. This also gives you access to any arguments passed via the
545	any arguments passed via the C<send_xxx> methods, such as file handles.	720	C<send_xxx> methods, such as file handles. See the L<use AnyEvent::Fork as
		721	a faster fork+exec> example to see it in action.
546		722
547	Returns the process object for easy chaining of method calls.	723	Returns the process object for easy chaining of method calls.
		724
		725	It's common to want to call an iniitalisation function with some
		726	arguments. Make sure you actually pass C<@_> to that function (for example
		727	by using C<&name> syntax), and do not just specify a function name:
		728
		729	$proc->eval ('&MyModule::init', $string1, $string2);
548		730
549	=cut	731	=cut
550		732
551	sub eval {	733	sub eval {
552	my ($self, $code, @args) = @_;	734	my ($self, $code, @args) = @_;
…		…
576	=item $proc = $proc->send_fh ($handle, ...)	758	=item $proc = $proc->send_fh ($handle, ...)
577		759
578	Send one or more file handles (I<not> file descriptors) to the process,	760	Send one or more file handles (I<not> file descriptors) to the process,
579	to prepare a call to C<run>.	761	to prepare a call to C<run>.
580		762
581	The process object keeps a reference to the handles until this is done,	763	The process object keeps a reference to the handles until they have
582	so you must not explicitly close the handles. This is most easily	764	been passed over to the process, so you must not explicitly close the
583	accomplished by simply not storing the file handles anywhere after passing	765	handles. This is most easily accomplished by simply not storing the file
584	them to this method.	766	handles anywhere after passing them to this method - when AnyEvent::Fork
		767	is finished using them, perl will automatically close them.
585		768
586	Returns the process object for easy chaining of method calls.	769	Returns the process object for easy chaining of method calls.
587		770
588	Example: pass a file handle to a process, and release it without	771	Example: pass a file handle to a process, and release it without
589	closing. It will be closed automatically when it is no longer used.	772	closing. It will be closed automatically when it is no longer used.
…		…
596	sub send_fh {	779	sub send_fh {
597	my ($self, @fh) = @_;	780	my ($self, @fh) = @_;
598		781
599	for my $fh (@fh) {	782	for my $fh (@fh) {
600	$self->_cmd ("h");	783	$self->_cmd ("h");
601	push @{ $self->[2] }, \$fh;	784	push @{ $self->[QUEUE] }, \$fh;
602	}	785	}
603		786
604	$self	787	$self
605	}	788	}
606		789
607	=item $proc = $proc->send_arg ($string, ...)	790	=item $proc = $proc->send_arg ($string, ...)
608		791
609	Send one or more argument strings to the process, to prepare a call to	792	Send one or more argument strings to the process, to prepare a call to
610	C<run>. The strings can be any octet string.	793	C<run>. The strings can be any octet strings.
611		794
612	The protocol is optimised to pass a moderate number of relatively short	795	The protocol is optimised to pass a moderate number of relatively short
613	strings - while you can pass up to 4GB of data in one go, this is more	796	strings - while you can pass up to 4GB of data in one go, this is more
614	meant to pass some ID information or other startup info, not big chunks of	797	meant to pass some ID information or other startup info, not big chunks of
615	data.	798	data.
…		…
631	Enter the function specified by the function name in C<$func> in the	814	Enter the function specified by the function name in C<$func> in the
632	process. The function is called with the communication socket as first	815	process. The function is called with the communication socket as first
633	argument, followed by all file handles and string arguments sent earlier	816	argument, followed by all file handles and string arguments sent earlier
634	via C<send_fh> and C<send_arg> methods, in the order they were called.	817	via C<send_fh> and C<send_arg> methods, in the order they were called.
635		818
		819	The process object becomes unusable on return from this function - any
		820	further method calls result in undefined behaviour.
		821
636	The function name should be fully qualified, but if it isn't, it will be	822	The function name should be fully qualified, but if it isn't, it will be
637	looked up in the main package.	823	looked up in the C<main> package.
638		824
639	If the called function returns, doesn't exist, or any error occurs, the	825	If the called function returns, doesn't exist, or any error occurs, the
640	process exits.	826	process exits.
641		827
642	Preparing the process is done in the background - when all commands have	828	Preparing the process is done in the background - when all commands have
643	been sent, the callback is invoked with the local communications socket	829	been sent, the callback is invoked with the local communications socket
644	as argument. At this point you can start using the socket in any way you	830	as argument. At this point you can start using the socket in any way you
645	like.	831	like.
646		832
647	The process object becomes unusable on return from this function - any
648	further method calls result in undefined behaviour.
649
650	If the communication socket isn't used, it should be closed on both sides,	833	If the communication socket isn't used, it should be closed on both sides,
651	to save on kernel memory.	834	to save on kernel memory.
652		835
653	The socket is non-blocking in the parent, and blocking in the newly	836	The socket is non-blocking in the parent, and blocking in the newly
654	created process. The close-on-exec flag is set in both.	837	created process. The close-on-exec flag is set in both.
655		838
656	Even if not used otherwise, the socket can be a good indicator for the	839	Even if not used otherwise, the socket can be a good indicator for the
657	existence of the process - if the other process exits, you get a readable	840	existence of the process - if the other process exits, you get a readable
658	event on it, because exiting the process closes the socket (if it didn't	841	event on it, because exiting the process closes the socket (if it didn't
659	create any children using fork).	842	create any children using fork).
		843
		844	=over 4
		845
		846	=item Compatibility to L<AnyEvent::Fork::Remote>
		847
		848	If you want to write code that works with both this module and
		849	L<AnyEvent::Fork::Remote>, you need to write your code so that it assumes
		850	there are two file handles for communications, which might not be unix
		851	domain sockets. The C<run> function should start like this:
		852
		853	sub run {
		854	my ($rfh, @args) = @_; # @args is your normal arguments
		855	my $wfh = fileno $rfh ? $rfh : *STDOUT;
		856
		857	# now use $rfh for reading and $wfh for writing
		858	}
		859
		860	This checks whether the passed file handle is, in fact, the process
		861	C<STDIN> handle. If it is, then the function was invoked visa
		862	L<AnyEvent::Fork::Remote>, so STDIN should be used for reading and
		863	C<STDOUT> should be used for writing.
		864
		865	In all other cases, the function was called via this module, and there is
		866	only one file handle that should be sued for reading and writing.
		867
		868	=back
660		869
661	Example: create a template for a process pool, pass a few strings, some	870	Example: create a template for a process pool, pass a few strings, some
662	file handles, then fork, pass one more string, and run some code.	871	file handles, then fork, pass one more string, and run some code.
663		872
664	my $pool = AnyEvent::Fork	873	my $pool = AnyEvent::Fork
…		…
692	=cut	901	=cut
693		902
694	sub run {	903	sub run {
695	my ($self, $func, $cb) = @_;	904	my ($self, $func, $cb) = @_;
696		905
697	$self->[4] = $cb;	906	$self->[CB] = $cb;
698	$self->_cmd (r => $func);	907	$self->_cmd (r => $func);
		908	}
		909
		910	=back
		911
		912	=head2 EXPERIMENTAL METHODS
		913
		914	These methods might go away completely or change behaviour, at any time.
		915
		916	=over 4
		917
		918	=item $proc->to_fh ($cb->($fh)) # EXPERIMENTAL, MIGHT BE REMOVED
		919
		920	Flushes all commands out to the process and then calls the callback with
		921	the communications socket.
		922
		923	The process object becomes unusable on return from this function - any
		924	further method calls result in undefined behaviour.
		925
		926	The point of this method is to give you a file handle that you can pass
		927	to another process. In that other process, you can call C<new_from_fh
		928	AnyEvent::Fork $fh> to create a new C<AnyEvent::Fork> object from it,
		929	thereby effectively passing a fork object to another process.
		930
		931	=cut
		932
		933	sub to_fh {
		934	my ($self, $cb) = @_;
		935
		936	$self->[CB] = $cb;
		937
		938	unless ($self->[WW]) {
		939	$self->[CB]->($self->[FH]);
		940	@$self = ();
		941	}
		942	}
		943
		944	=item new_from_fh AnyEvent::Fork $fh # EXPERIMENTAL, MIGHT BE REMOVED
		945
		946	Takes a file handle originally rceeived by the C<to_fh> method and creates
		947	a new C<AnyEvent:Fork> object. The child process itself will not change in
		948	any way, i.e. it will keep all the modifications done to it before calling
		949	C<to_fh>.
		950
		951	The new object is very much like the original object, except that the
		952	C<pid> method will return C<undef> even if the process is a direct child.
		953
		954	=cut
		955
		956	sub new_from_fh {
		957	my ($class, $fh) = @_;
		958
		959	$class->_new ($fh)
699	}	960	}
700		961
701	=back	962	=back
702		963
703	=head1 PERFORMANCE	964	=head1 PERFORMANCE
…		…
713		974
714	2079 new processes per second, using manual socketpair + fork	975	2079 new processes per second, using manual socketpair + fork
715		976
716	Then I did the same thing, but instead of calling fork, I called	977	Then I did the same thing, but instead of calling fork, I called
717	AnyEvent::Fork->new->run ("CORE::exit") and then again waited for the	978	AnyEvent::Fork->new->run ("CORE::exit") and then again waited for the
718	socket form the child to close on exit. This does the same thing as manual	979	socket from the child to close on exit. This does the same thing as manual
719	socket pair + fork, except that what is forked is the template process	980	socket pair + fork, except that what is forked is the template process
720	(2440kB), and the socket needs to be passed to the server at the other end	981	(2440kB), and the socket needs to be passed to the server at the other end
721	of the socket first.	982	of the socket first.
722		983
723	2307 new processes per second, using AnyEvent::Fork->new	984	2307 new processes per second, using AnyEvent::Fork->new
…		…
728	479 vfork+execs per second, using AnyEvent::Fork->new_exec	989	479 vfork+execs per second, using AnyEvent::Fork->new_exec
729		990
730	So how can C<< AnyEvent->new >> be faster than a standard fork, even	991	So how can C<< AnyEvent->new >> be faster than a standard fork, even
731	though it uses the same operations, but adds a lot of overhead?	992	though it uses the same operations, but adds a lot of overhead?
732		993
733	The difference is simply the process size: forking the 6MB process takes	994	The difference is simply the process size: forking the 5MB process takes
734	so much longer than forking the 2.5MB template process that the overhead	995	so much longer than forking the 2.5MB template process that the extra
735	introduced is canceled out.	996	overhead is canceled out.
736		997
737	If the benchmark process grows, the normal fork becomes even slower:	998	If the benchmark process grows, the normal fork becomes even slower:
738		999
739	1340 new processes, manual fork in a 20MB process	1000	1340 new processes, manual fork of a 20MB process
740	731 new processes, manual fork in a 200MB process	1001	731 new processes, manual fork of a 200MB process
741	235 new processes, manual fork in a 2000MB process	1002	235 new processes, manual fork of a 2000MB process
742		1003
743	What that means (to me) is that I can use this module without having a	1004	What that means (to me) is that I can use this module without having a bad
744	very bad conscience because of the extra overhead required to start new	1005	conscience because of the extra overhead required to start new processes.
745	processes.
746		1006
747	=head1 TYPICAL PROBLEMS	1007	=head1 TYPICAL PROBLEMS
748		1008
749	This section lists typical problems that remain. I hope by recognising	1009	This section lists typical problems that remain. I hope by recognising
750	them, most can be avoided.	1010	them, most can be avoided.
751		1011
752	=over 4	1012	=over 4
753		1013
754	=item "leaked" file descriptors for exec'ed processes	1014	=item leaked file descriptors for exec'ed processes
755		1015
756	POSIX systems inherit file descriptors by default when exec'ing a new	1016	POSIX systems inherit file descriptors by default when exec'ing a new
757	process. While perl itself laudably sets the close-on-exec flags on new	1017	process. While perl itself laudably sets the close-on-exec flags on new
758	file handles, most C libraries don't care, and even if all cared, it's	1018	file handles, most C libraries don't care, and even if all cared, it's
759	often not possible to set the flag in a race-free manner.	1019	often not possible to set the flag in a race-free manner.
…		…
779	libraries or the code that leaks those file descriptors.	1039	libraries or the code that leaks those file descriptors.
780		1040
781	Fortunately, most of these leaked descriptors do no harm, other than	1041	Fortunately, most of these leaked descriptors do no harm, other than
782	sitting on some resources.	1042	sitting on some resources.
783		1043
784	=item "leaked" file descriptors for fork'ed processes	1044	=item leaked file descriptors for fork'ed processes
785		1045
786	Normally, L<AnyEvent::Fork> does start new processes by exec'ing them,	1046	Normally, L<AnyEvent::Fork> does start new processes by exec'ing them,
787	which closes file descriptors not marked for being inherited.	1047	which closes file descriptors not marked for being inherited.
788		1048
789	However, L<AnyEvent::Fork::Early> and L<AnyEvent::Fork::Template> offer	1049	However, L<AnyEvent::Fork::Early> and L<AnyEvent::Fork::Template> offer
…		…
798		1058
799	The solution is to either not load these modules before use'ing	1059	The solution is to either not load these modules before use'ing
800	L<AnyEvent::Fork::Early> or L<AnyEvent::Fork::Template>, or to delay	1060	L<AnyEvent::Fork::Early> or L<AnyEvent::Fork::Template>, or to delay
801	initialising them, for example, by calling C<init Gtk2> manually.	1061	initialising them, for example, by calling C<init Gtk2> manually.
802		1062
803	=item exit runs destructors	1063	=item exiting calls object destructors
804		1064
805	This only applies to users of Lc<AnyEvent::Fork:Early> and	1065	This only applies to users of L<AnyEvent::Fork:Early> and
806	L<AnyEvent::Fork::Template>.	1066	L<AnyEvent::Fork::Template>, or when initialising code creates objects
		1067	that reference external resources.
807		1068
808	When a process created by AnyEvent::Fork exits, it might do so by calling	1069	When a process created by AnyEvent::Fork exits, it might do so by calling
809	exit, or simply letting perl reach the end of the program. At which point	1070	exit, or simply letting perl reach the end of the program. At which point
810	Perl runs all destructors.	1071	Perl runs all destructors.
811		1072
…		…
830	to make it so, mostly due to the bloody broken perl that nobody seems to	1091	to make it so, mostly due to the bloody broken perl that nobody seems to
831	care about. The fork emulation is a bad joke - I have yet to see something	1092	care about. The fork emulation is a bad joke - I have yet to see something
832	useful that you can do with it without running into memory corruption	1093	useful that you can do with it without running into memory corruption
833	issues or other braindamage. Hrrrr.	1094	issues or other braindamage. Hrrrr.
834		1095
835	Cygwin perl is not supported at the moment, as it should implement fd	1096	Since fork is endlessly broken on win32 perls (it doesn't even remotely
836	passing, but doesn't, and rolling my own is hard, as cygwin doesn't	1097	work within it's documented limits) and quite obviously it's not getting
837	support enough functionality to do it.	1098	improved any time soon, the best way to proceed on windows would be to
		1099	always use C<new_exec> and thus never rely on perl's fork "emulation".
		1100
		1101	Cygwin perl is not supported at the moment due to some hilarious
		1102	shortcomings of its API - see L<IO::FDPoll> for more details. If you never
		1103	use C<send_fh> and always use C<new_exec> to create processes, it should
		1104	work though.
		1105
		1106	=head1 USING AnyEvent::Fork IN SUBPROCESSES
		1107
		1108	AnyEvent::Fork itself cannot generally be used in subprocesses. As long as
		1109	only one process ever forks new processes, sharing the template processes
		1110	is possible (you could use a pipe as a lock by writing a byte into it to
		1111	unlock, and reading the byte to lock for example)
		1112
		1113	To make concurrent calls possible after fork, you should get rid of the
		1114	template and early fork processes. AnyEvent::Fork will create a new
		1115	template process as needed.
		1116
		1117	undef $AnyEvent::Fork::EARLY;
		1118	undef $AnyEvent::Fork::TEMPLATE;
		1119
		1120	It doesn't matter whether you get rid of them in the parent or child after
		1121	a fork.
838		1122
839	=head1 SEE ALSO	1123	=head1 SEE ALSO
840		1124
841	L<AnyEvent::Fork::Early> (to avoid executing a perl interpreter),	1125	L<AnyEvent::Fork::Early>, to avoid executing a perl interpreter at all
		1126	(part of this distribution).
		1127
842	L<AnyEvent::Fork::Template> (to create a process by forking the main	1128	L<AnyEvent::Fork::Template>, to create a process by forking the main
843	program at a convenient time).	1129	program at a convenient time (part of this distribution).
844		1130
845	=head1 AUTHOR	1131	L<AnyEvent::Fork::Remote>, for another way to create processes that is
		1132	mostly compatible to this module and modules building on top of it, but
		1133	works better with remote processes.
		1134
		1135	L<AnyEvent::Fork::RPC>, for simple RPC to child processes (on CPAN).
		1136
		1137	L<AnyEvent::Fork::Pool>, for simple worker process pool (on CPAN).
		1138
		1139	=head1 AUTHOR AND CONTACT INFORMATION
846		1140
847	Marc Lehmann <schmorp@schmorp.de>	1141	Marc Lehmann <schmorp@schmorp.de>
848	http://home.schmorp.de/	1142	http://software.schmorp.de/pkg/AnyEvent-Fork
849		1143
850	=cut	1144	=cut
851		1145
852	1	1146	1
853		1147

Diff Legend

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

Comparing AnyEvent-Fork/Fork.pm (file contents): Revision 1.24 by root, Sat Apr 6 08:32:23 2013 UTC vs. Revision 1.67 by root, Thu May 12 16:54:01 2016 UTC

Diff Legend

Comparing AnyEvent-Fork/Fork.pm (file contents):
Revision 1.24 by root, Sat Apr 6 08:32:23 2013 UTC vs.
Revision 1.67 by root, Thu May 12 16:54:01 2016 UTC