… | |
… | |
40 | points in your program, so locking and parallel access are rarely an |
40 | points in your program, so locking and parallel access are rarely an |
41 | issue, making thread programming much safer and easier than using other |
41 | issue, making thread programming much safer and easier than using other |
42 | thread models. |
42 | thread models. |
43 | |
43 | |
44 | Unlike the so-called "Perl threads" (which are not actually real threads |
44 | Unlike the so-called "Perl threads" (which are not actually real threads |
45 | but only the windows process emulation ported to unix), Coro provides a |
45 | but only the windows process emulation (see section of same name for |
|
|
46 | more details) ported to UNIX, and as such act as processes), Coro |
46 | full shared address space, which makes communication between threads |
47 | provides a full shared address space, which makes communication between |
47 | very easy. And threads are fast, too: disabling the Windows process |
48 | threads very easy. And coro threads are fast, too: disabling the Windows |
48 | emulation code in your perl and using Coro can easily result in a two to |
49 | process emulation code in your perl and using Coro can easily result in |
49 | four times speed increase for your programs. |
50 | a two to four times speed increase for your programs. A parallel matrix |
|
|
51 | multiplication benchmark (very communication-intensive) runs over 300 |
|
|
52 | times faster on a single core than perls pseudo-threads on a quad core |
|
|
53 | using all four cores. |
50 | |
54 | |
51 | Coro achieves that by supporting multiple running interpreters that share |
55 | Coro achieves that by supporting multiple running interpreters that share |
52 | data, which is especially useful to code pseudo-parallel processes and |
56 | data, which is especially useful to code pseudo-parallel processes and |
53 | for event-based programming, such as multiple HTTP-GET requests running |
57 | for event-based programming, such as multiple HTTP-GET requests running |
54 | concurrently. See L<Coro::AnyEvent> to learn more on how to integrate Coro |
58 | concurrently. See L<Coro::AnyEvent> to learn more on how to integrate Coro |
55 | into an event-based environment. |
59 | into an event-based environment. |
56 | |
60 | |
57 | In this module, a thread is defined as "callchain + lexical variables + |
61 | In this module, a thread is defined as "callchain + lexical variables + |
58 | @_ + $_ + $@ + $/ + C stack), that is, a thread has its own callchain, |
62 | some package variables + C stack), that is, a thread has its own callchain, |
59 | its own set of lexicals and its own set of perls most important global |
63 | its own set of lexicals and its own set of perls most important global |
60 | variables (see L<Coro::State> for more configuration and background info). |
64 | variables (see L<Coro::State> for more configuration and background info). |
61 | |
65 | |
62 | See also the C<SEE ALSO> section at the end of this document - the Coro |
66 | See also the C<SEE ALSO> section at the end of this document - the Coro |
63 | module family is quite large. |
67 | module family is quite large. |
64 | |
68 | |
|
|
69 | =head1 CORO THREAD LIFE CYCLE |
|
|
70 | |
|
|
71 | During the long and exciting (or not) life of a coro thread, it goes |
|
|
72 | through a number of states: |
|
|
73 | |
|
|
74 | =over 4 |
|
|
75 | |
|
|
76 | =item 1. Creation |
|
|
77 | |
|
|
78 | The first thing in the life of a coro thread is it's creation - |
|
|
79 | obviously. The typical way to create a thread is to call the C<async |
|
|
80 | BLOCK> function: |
|
|
81 | |
|
|
82 | async { |
|
|
83 | # thread code goes here |
|
|
84 | }; |
|
|
85 | |
|
|
86 | You can also pass arguments, which are put in C<@_>: |
|
|
87 | |
|
|
88 | async { |
|
|
89 | print $_[1]; # prints 2 |
|
|
90 | } 1, 2, 3; |
|
|
91 | |
|
|
92 | This creates a new coro thread and puts it into the ready queue, meaning |
|
|
93 | it will run as soon as the CPU is free for it. |
|
|
94 | |
|
|
95 | C<async> will return a coro object - you can store this for future |
|
|
96 | reference or ignore it, the thread itself will keep a reference to it's |
|
|
97 | thread object - threads are alive on their own. |
|
|
98 | |
|
|
99 | Another way to create a thread is to call the C<new> constructor with a |
|
|
100 | code-reference: |
|
|
101 | |
|
|
102 | new Coro sub { |
|
|
103 | # thread code goes here |
|
|
104 | }, @optional_arguments; |
|
|
105 | |
|
|
106 | This is quite similar to calling C<async>, but the important difference is |
|
|
107 | that the new thread is not put into the ready queue, so the thread will |
|
|
108 | not run until somebody puts it there. C<async> is, therefore, identical to |
|
|
109 | this sequence: |
|
|
110 | |
|
|
111 | my $coro = new Coro sub { |
|
|
112 | # thread code goes here |
|
|
113 | }; |
|
|
114 | $coro->ready; |
|
|
115 | return $coro; |
|
|
116 | |
|
|
117 | =item 2. Startup |
|
|
118 | |
|
|
119 | When a new coro thread is created, only a copy of the code reference |
|
|
120 | and the arguments are stored, no extra memory for stacks and so on is |
|
|
121 | allocated, keeping the coro thread in a low-memory state. |
|
|
122 | |
|
|
123 | Only when it actually starts executing will all the resources be finally |
|
|
124 | allocated. |
|
|
125 | |
|
|
126 | The optional arguments specified at coro creation are available in C<@_>, |
|
|
127 | similar to function calls. |
|
|
128 | |
|
|
129 | =item 3. Running / Blocking |
|
|
130 | |
|
|
131 | A lot can happen after the coro thread has started running. Quite usually, |
|
|
132 | it will not run to the end in one go (because you could use a function |
|
|
133 | instead), but it will give up the CPU regularly because it waits for |
|
|
134 | external events. |
|
|
135 | |
|
|
136 | As long as a coro thread runs, it's coro object is available in the global |
|
|
137 | variable C<$Coro::current>. |
|
|
138 | |
|
|
139 | The low-level way to give up the CPU is to call the scheduler, which |
|
|
140 | selects a new coro thread to run: |
|
|
141 | |
|
|
142 | Coro::schedule; |
|
|
143 | |
|
|
144 | Since running threads are not in the ready queue, calling the scheduler |
|
|
145 | without doing anything else will block the coro thread forever - you need |
|
|
146 | to arrange either for the coro to put woken up (readied) by some other |
|
|
147 | event or some other thread, or you can put it into the ready queue before |
|
|
148 | scheduling: |
|
|
149 | |
|
|
150 | # this is exactly what Coro::cede does |
|
|
151 | $Coro::current->ready; |
|
|
152 | Coro::schedule; |
|
|
153 | |
|
|
154 | All the higher-level synchronisation methods (Coro::Semaphore, |
|
|
155 | Coro::rouse_*...) are actually implemented via C<< ->ready >> and C<< |
|
|
156 | Coro::schedule >>. |
|
|
157 | |
|
|
158 | While the coro thread is running it also might get assigned a C-level |
|
|
159 | thread, or the C-level thread might be unassigned from it, as the Coro |
|
|
160 | runtime wishes. A C-level thread needs to be assigned when your perl |
|
|
161 | thread calls into some C-level function and that function in turn calls |
|
|
162 | perl and perl then wants to switch coroutines. This happens most often |
|
|
163 | when you run an event loop and block in the callback, or when perl |
|
|
164 | itself calls some function such as C<AUTOLOAD> or methods via the C<tie> |
|
|
165 | mechanism. |
|
|
166 | |
|
|
167 | =item 4. Termination |
|
|
168 | |
|
|
169 | Many threads actually terminate after some time. There are a number of |
|
|
170 | ways to terminate a coro thread, the simplest is returning from the |
|
|
171 | top-level code reference: |
|
|
172 | |
|
|
173 | async { |
|
|
174 | # after returning from here, the coro thread is terminated |
|
|
175 | }; |
|
|
176 | |
|
|
177 | async { |
|
|
178 | return if 0.5 < rand; # terminate a little earlier, maybe |
|
|
179 | print "got a chance to print this\n"; |
|
|
180 | # or here |
|
|
181 | }; |
|
|
182 | |
|
|
183 | Any values returned from the coroutine can be recovered using C<< ->join |
|
|
184 | >>: |
|
|
185 | |
|
|
186 | my $coro = async { |
|
|
187 | "hello, world\n" # return a string |
|
|
188 | }; |
|
|
189 | |
|
|
190 | my $hello_world = $coro->join; |
|
|
191 | |
|
|
192 | print $hello_world; |
|
|
193 | |
|
|
194 | Another way to terminate is to call C<< Coro::terminate >>, which at any |
|
|
195 | subroutine call nesting level: |
|
|
196 | |
|
|
197 | async { |
|
|
198 | Coro::terminate "return value 1", "return value 2"; |
|
|
199 | }; |
|
|
200 | |
|
|
201 | And yet another way is to C<< ->cancel >> (or C<< ->safe_cancel >>) the |
|
|
202 | coro thread from another thread: |
|
|
203 | |
|
|
204 | my $coro = async { |
|
|
205 | exit 1; |
|
|
206 | }; |
|
|
207 | |
|
|
208 | $coro->cancel; # also accepts values for ->join to retrieve |
|
|
209 | |
|
|
210 | Cancellation I<can> be dangerous - it's a bit like calling C<exit> without |
|
|
211 | actually exiting, and might leave C libraries and XS modules in a weird |
|
|
212 | state. Unlike other thread implementations, however, Coro is exceptionally |
|
|
213 | safe with regards to cancellation, as perl will always be in a consistent |
|
|
214 | state, and for those cases where you want to do truly marvellous things |
|
|
215 | with your coro while it is being cancelled - that is, make sure all |
|
|
216 | cleanup code is executed from the thread being cancelled - there is even a |
|
|
217 | C<< ->safe_cancel >> method. |
|
|
218 | |
|
|
219 | So, cancelling a thread that runs in an XS event loop might not be the |
|
|
220 | best idea, but any other combination that deals with perl only (cancelling |
|
|
221 | when a thread is in a C<tie> method or an C<AUTOLOAD> for example) is |
|
|
222 | safe. |
|
|
223 | |
|
|
224 | =item 5. Cleanup |
|
|
225 | |
|
|
226 | Threads will allocate various resources. Most but not all will be returned |
|
|
227 | when a thread terminates, during clean-up. |
|
|
228 | |
|
|
229 | Cleanup is quite similar to throwing an uncaught exception: perl will |
|
|
230 | work it's way up through all subroutine calls and blocks. On it's way, it |
|
|
231 | will release all C<my> variables, undo all C<local>'s and free any other |
|
|
232 | resources truly local to the thread. |
|
|
233 | |
|
|
234 | So, a common way to free resources is to keep them referenced only by my |
|
|
235 | variables: |
|
|
236 | |
|
|
237 | async { |
|
|
238 | my $big_cache = new Cache ...; |
|
|
239 | }; |
|
|
240 | |
|
|
241 | If there are no other references, then the C<$big_cache> object will be |
|
|
242 | freed when the thread terminates, regardless of how it does so. |
|
|
243 | |
|
|
244 | What it does C<NOT> do is unlock any Coro::Semaphores or similar |
|
|
245 | resources, but that's where the C<guard> methods come in handy: |
|
|
246 | |
|
|
247 | my $sem = new Coro::Semaphore; |
|
|
248 | |
|
|
249 | async { |
|
|
250 | my $lock_guard = $sem->guard; |
|
|
251 | # if we reutrn, or die or get cancelled, here, |
|
|
252 | # then the semaphore will be "up"ed. |
|
|
253 | }; |
|
|
254 | |
|
|
255 | The C<Guard::guard> function comes in handy for any custom cleanup you |
|
|
256 | might want to do: |
|
|
257 | |
|
|
258 | async { |
|
|
259 | my $window = new Gtk2::Window "toplevel"; |
|
|
260 | # The window will not be cleaned up automatically, even when $window |
|
|
261 | # gets freed, so use a guard to ensure it's destruction |
|
|
262 | # in case of an error: |
|
|
263 | my $window_guard = Guard::guard { $window->destroy }; |
|
|
264 | |
|
|
265 | # we are safe here |
|
|
266 | }; |
|
|
267 | |
|
|
268 | Last not least, C<local> can often be handy, too, e.g. when temporarily |
|
|
269 | replacing the coro thread description: |
|
|
270 | |
|
|
271 | sub myfunction { |
|
|
272 | local $Coro::current->{desc} = "inside myfunction(@_)"; |
|
|
273 | |
|
|
274 | # if we return or die here, the description will be restored |
|
|
275 | } |
|
|
276 | |
|
|
277 | =item 6. Viva La Zombie Muerte |
|
|
278 | |
|
|
279 | Even after a thread has terminated and cleaned up it's resources, the coro |
|
|
280 | object still is there and stores the return values of the thread. Only in |
|
|
281 | this state will the coro object be "reference counted" in the normal perl |
|
|
282 | sense: the thread code keeps a reference to it when it is active, but not |
|
|
283 | after it has terminated. |
|
|
284 | |
|
|
285 | The means the coro object gets freed automatically when the thread has |
|
|
286 | terminated and cleaned up and there arenot other references. |
|
|
287 | |
|
|
288 | If there are, the coro object will stay around, and you can call C<< |
|
|
289 | ->join >> as many times as you wish to retrieve the result values: |
|
|
290 | |
|
|
291 | async { |
|
|
292 | print "hi\n"; |
|
|
293 | 1 |
|
|
294 | }; |
|
|
295 | |
|
|
296 | # run the async above, and free everything before returning |
|
|
297 | # from Coro::cede: |
|
|
298 | Coro::cede; |
|
|
299 | |
|
|
300 | { |
|
|
301 | my $coro = async { |
|
|
302 | print "hi\n"; |
|
|
303 | 1 |
|
|
304 | }; |
|
|
305 | |
|
|
306 | # run the async above, and clean up, but do not free the coro |
|
|
307 | # object: |
|
|
308 | Coro::cede; |
|
|
309 | |
|
|
310 | # optionally retrieve the result values |
|
|
311 | my @results = $coro->join; |
|
|
312 | |
|
|
313 | # now $coro goes out of scope, and presumably gets freed |
|
|
314 | }; |
|
|
315 | |
|
|
316 | =back |
|
|
317 | |
65 | =cut |
318 | =cut |
66 | |
319 | |
67 | package Coro; |
320 | package Coro; |
68 | |
321 | |
69 | use strict qw(vars subs); |
322 | use common::sense; |
70 | no warnings "uninitialized"; |
323 | |
|
|
324 | use Carp (); |
71 | |
325 | |
72 | use Guard (); |
326 | use Guard (); |
73 | |
327 | |
74 | use Coro::State; |
328 | use Coro::State; |
75 | |
329 | |
… | |
… | |
77 | |
331 | |
78 | our $idle; # idle handler |
332 | our $idle; # idle handler |
79 | our $main; # main coro |
333 | our $main; # main coro |
80 | our $current; # current coro |
334 | our $current; # current coro |
81 | |
335 | |
82 | our $VERSION = 5.13; |
336 | our $VERSION = 5.372; |
83 | |
337 | |
84 | our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub); |
338 | our @EXPORT = qw(async async_pool cede schedule terminate current unblock_sub rouse_cb rouse_wait); |
85 | our %EXPORT_TAGS = ( |
339 | our %EXPORT_TAGS = ( |
86 | prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)], |
340 | prio => [qw(PRIO_MAX PRIO_HIGH PRIO_NORMAL PRIO_LOW PRIO_IDLE PRIO_MIN)], |
87 | ); |
341 | ); |
88 | our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready)); |
342 | our @EXPORT_OK = (@{$EXPORT_TAGS{prio}}, qw(nready)); |
89 | |
343 | |
… | |
… | |
120 | |
374 | |
121 | This variable is mainly useful to integrate Coro into event loops. It is |
375 | This variable is mainly useful to integrate Coro into event loops. It is |
122 | usually better to rely on L<Coro::AnyEvent> or L<Coro::EV>, as this is |
376 | usually better to rely on L<Coro::AnyEvent> or L<Coro::EV>, as this is |
123 | pretty low-level functionality. |
377 | pretty low-level functionality. |
124 | |
378 | |
125 | This variable stores either a Coro object or a callback. |
379 | This variable stores a Coro object that is put into the ready queue when |
|
|
380 | there are no other ready threads (without invoking any ready hooks). |
126 | |
381 | |
127 | If it is a callback, the it is called whenever the scheduler finds no |
382 | The default implementation dies with "FATAL: deadlock detected.", followed |
128 | ready coros to run. The default implementation prints "FATAL: |
383 | by a thread listing, because the program has no other way to continue. |
129 | deadlock detected" and exits, because the program has no other way to |
|
|
130 | continue. |
|
|
131 | |
|
|
132 | If it is a coro object, then this object will be readied (without |
|
|
133 | invoking any ready hooks, however) when the scheduler finds no other ready |
|
|
134 | coros to run. |
|
|
135 | |
384 | |
136 | This hook is overwritten by modules such as C<Coro::EV> and |
385 | This hook is overwritten by modules such as C<Coro::EV> and |
137 | C<Coro::AnyEvent> to wait on an external event that hopefully wake up a |
386 | C<Coro::AnyEvent> to wait on an external event that hopefully wakes up a |
138 | coro so the scheduler can run it. |
387 | coro so the scheduler can run it. |
139 | |
388 | |
140 | Note that the callback I<must not>, under any circumstances, block |
|
|
141 | the current coro. Normally, this is achieved by having an "idle |
|
|
142 | coro" that calls the event loop and then blocks again, and then |
|
|
143 | readying that coro in the idle handler, or by simply placing the idle |
|
|
144 | coro in this variable. |
|
|
145 | |
|
|
146 | See L<Coro::Event> or L<Coro::AnyEvent> for examples of using this |
389 | See L<Coro::EV> or L<Coro::AnyEvent> for examples of using this technique. |
147 | technique. |
|
|
148 | |
390 | |
149 | Please note that if your callback recursively invokes perl (e.g. for event |
|
|
150 | handlers), then it must be prepared to be called recursively itself. |
|
|
151 | |
|
|
152 | =cut |
391 | =cut |
153 | |
392 | |
154 | $idle = sub { |
393 | # ||= because other modules could have provided their own by now |
155 | require Carp; |
394 | $idle ||= new Coro sub { |
156 | Carp::croak ("FATAL: deadlock detected"); |
395 | require Coro::Debug; |
|
|
396 | die "FATAL: deadlock detected.\n" |
|
|
397 | . Coro::Debug::ps_listing (); |
157 | }; |
398 | }; |
158 | |
399 | |
159 | # this coro is necessary because a coro |
400 | # this coro is necessary because a coro |
160 | # cannot destroy itself. |
401 | # cannot destroy itself. |
161 | our @destroy; |
402 | our @destroy; |
162 | our $manager; |
403 | our $manager; |
163 | |
404 | |
164 | $manager = new Coro sub { |
405 | $manager = new Coro sub { |
165 | while () { |
406 | while () { |
166 | Coro::State::cancel shift @destroy |
407 | _destroy shift @destroy |
167 | while @destroy; |
408 | while @destroy; |
168 | |
409 | |
169 | &schedule; |
410 | &schedule; |
170 | } |
411 | } |
171 | }; |
412 | }; |
… | |
… | |
203 | Example: Create a new coro that just prints its arguments. |
444 | Example: Create a new coro that just prints its arguments. |
204 | |
445 | |
205 | async { |
446 | async { |
206 | print "@_\n"; |
447 | print "@_\n"; |
207 | } 1,2,3,4; |
448 | } 1,2,3,4; |
208 | |
|
|
209 | =cut |
|
|
210 | |
|
|
211 | sub async(&@) { |
|
|
212 | my $coro = new Coro @_; |
|
|
213 | $coro->ready; |
|
|
214 | $coro |
|
|
215 | } |
|
|
216 | |
449 | |
217 | =item async_pool { ... } [@args...] |
450 | =item async_pool { ... } [@args...] |
218 | |
451 | |
219 | Similar to C<async>, but uses a coro pool, so you should not call |
452 | Similar to C<async>, but uses a coro pool, so you should not call |
220 | terminate or join on it (although you are allowed to), and you get a |
453 | terminate or join on it (although you are allowed to), and you get a |
… | |
… | |
277 | =item schedule |
510 | =item schedule |
278 | |
511 | |
279 | Calls the scheduler. The scheduler will find the next coro that is |
512 | Calls the scheduler. The scheduler will find the next coro that is |
280 | to be run from the ready queue and switches to it. The next coro |
513 | to be run from the ready queue and switches to it. The next coro |
281 | to be run is simply the one with the highest priority that is longest |
514 | to be run is simply the one with the highest priority that is longest |
282 | in its ready queue. If there is no coro ready, it will clal the |
515 | in its ready queue. If there is no coro ready, it will call the |
283 | C<$Coro::idle> hook. |
516 | C<$Coro::idle> hook. |
284 | |
517 | |
285 | Please note that the current coro will I<not> be put into the ready |
518 | Please note that the current coro will I<not> be put into the ready |
286 | queue, so calling this function usually means you will never be called |
519 | queue, so calling this function usually means you will never be called |
287 | again unless something else (e.g. an event handler) calls C<< ->ready >>, |
520 | again unless something else (e.g. an event handler) calls C<< ->ready >>, |
… | |
… | |
313 | coro, regardless of priority. This is useful sometimes to ensure |
546 | coro, regardless of priority. This is useful sometimes to ensure |
314 | progress is made. |
547 | progress is made. |
315 | |
548 | |
316 | =item terminate [arg...] |
549 | =item terminate [arg...] |
317 | |
550 | |
318 | Terminates the current coro with the given status values (see L<cancel>). |
551 | Terminates the current coro with the given status values (see |
|
|
552 | L<cancel>). The values will not be copied, but referenced directly. |
319 | |
553 | |
320 | =item Coro::on_enter BLOCK, Coro::on_leave BLOCK |
554 | =item Coro::on_enter BLOCK, Coro::on_leave BLOCK |
321 | |
555 | |
322 | These function install enter and leave winders in the current scope. The |
556 | These function install enter and leave winders in the current scope. The |
323 | enter block will be executed when on_enter is called and whenever the |
557 | enter block will be executed when on_enter is called and whenever the |
… | |
… | |
335 | |
569 | |
336 | These functions implement the same concept as C<dynamic-wind> in scheme |
570 | These functions implement the same concept as C<dynamic-wind> in scheme |
337 | does, and are useful when you want to localise some resource to a specific |
571 | does, and are useful when you want to localise some resource to a specific |
338 | coro. |
572 | coro. |
339 | |
573 | |
340 | They slow down coro switching considerably for coros that use |
574 | They slow down thread switching considerably for coros that use them |
341 | them (But coro switching is still reasonably fast if the handlers are |
575 | (about 40% for a BLOCK with a single assignment, so thread switching is |
342 | fast). |
576 | still reasonably fast if the handlers are fast). |
343 | |
577 | |
344 | These functions are best understood by an example: The following function |
578 | These functions are best understood by an example: The following function |
345 | will change the current timezone to "Antarctica/South_Pole", which |
579 | will change the current timezone to "Antarctica/South_Pole", which |
346 | requires a call to C<tzset>, but by using C<on_enter> and C<on_leave>, |
580 | requires a call to C<tzset>, but by using C<on_enter> and C<on_leave>, |
347 | which remember/change the current timezone and restore the previous |
581 | which remember/change the current timezone and restore the previous |
348 | value, respectively, the timezone is only changes for the coro that |
582 | value, respectively, the timezone is only changed for the coro that |
349 | installed those handlers. |
583 | installed those handlers. |
350 | |
584 | |
351 | use POSIX qw(tzset); |
585 | use POSIX qw(tzset); |
352 | |
586 | |
353 | async { |
587 | async { |
… | |
… | |
370 | }; |
604 | }; |
371 | |
605 | |
372 | This can be used to localise about any resource (locale, uid, current |
606 | This can be used to localise about any resource (locale, uid, current |
373 | working directory etc.) to a block, despite the existance of other |
607 | working directory etc.) to a block, despite the existance of other |
374 | coros. |
608 | coros. |
|
|
609 | |
|
|
610 | Another interesting example implements time-sliced multitasking using |
|
|
611 | interval timers (this could obviously be optimised, but does the job): |
|
|
612 | |
|
|
613 | # "timeslice" the given block |
|
|
614 | sub timeslice(&) { |
|
|
615 | use Time::HiRes (); |
|
|
616 | |
|
|
617 | Coro::on_enter { |
|
|
618 | # on entering the thread, we set an VTALRM handler to cede |
|
|
619 | $SIG{VTALRM} = sub { cede }; |
|
|
620 | # and then start the interval timer |
|
|
621 | Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0.01, 0.01; |
|
|
622 | }; |
|
|
623 | Coro::on_leave { |
|
|
624 | # on leaving the thread, we stop the interval timer again |
|
|
625 | Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0, 0; |
|
|
626 | }; |
|
|
627 | |
|
|
628 | &{+shift}; |
|
|
629 | } |
|
|
630 | |
|
|
631 | # use like this: |
|
|
632 | timeslice { |
|
|
633 | # The following is an endless loop that would normally |
|
|
634 | # monopolise the process. Since it runs in a timesliced |
|
|
635 | # environment, it will regularly cede to other threads. |
|
|
636 | while () { } |
|
|
637 | }; |
|
|
638 | |
375 | |
639 | |
376 | =item killall |
640 | =item killall |
377 | |
641 | |
378 | Kills/terminates/cancels all coros except the currently running one. |
642 | Kills/terminates/cancels all coros except the currently running one. |
379 | |
643 | |
… | |
… | |
424 | |
688 | |
425 | This ensures that the scheduler will resume this coro automatically |
689 | This ensures that the scheduler will resume this coro automatically |
426 | once all the coro of higher priority and all coro of the same |
690 | once all the coro of higher priority and all coro of the same |
427 | priority that were put into the ready queue earlier have been resumed. |
691 | priority that were put into the ready queue earlier have been resumed. |
428 | |
692 | |
|
|
693 | =item $coro->suspend |
|
|
694 | |
|
|
695 | Suspends the specified coro. A suspended coro works just like any other |
|
|
696 | coro, except that the scheduler will not select a suspended coro for |
|
|
697 | execution. |
|
|
698 | |
|
|
699 | Suspending a coro can be useful when you want to keep the coro from |
|
|
700 | running, but you don't want to destroy it, or when you want to temporarily |
|
|
701 | freeze a coro (e.g. for debugging) to resume it later. |
|
|
702 | |
|
|
703 | A scenario for the former would be to suspend all (other) coros after a |
|
|
704 | fork and keep them alive, so their destructors aren't called, but new |
|
|
705 | coros can be created. |
|
|
706 | |
|
|
707 | =item $coro->resume |
|
|
708 | |
|
|
709 | If the specified coro was suspended, it will be resumed. Note that when |
|
|
710 | the coro was in the ready queue when it was suspended, it might have been |
|
|
711 | unreadied by the scheduler, so an activation might have been lost. |
|
|
712 | |
|
|
713 | To avoid this, it is best to put a suspended coro into the ready queue |
|
|
714 | unconditionally, as every synchronisation mechanism must protect itself |
|
|
715 | against spurious wakeups, and the one in the Coro family certainly do |
|
|
716 | that. |
|
|
717 | |
429 | =item $is_ready = $coro->is_ready |
718 | =item $is_ready = $coro->is_ready |
430 | |
719 | |
431 | Returns true iff the Coro object is in the ready queue. Unless the Coro |
720 | Returns true iff the Coro object is in the ready queue. Unless the Coro |
432 | object gets destroyed, it will eventually be scheduled by the scheduler. |
721 | object gets destroyed, it will eventually be scheduled by the scheduler. |
433 | |
722 | |
… | |
… | |
442 | Returns true iff this Coro object has been suspended. Suspended Coros will |
731 | Returns true iff this Coro object has been suspended. Suspended Coros will |
443 | not ever be scheduled. |
732 | not ever be scheduled. |
444 | |
733 | |
445 | =item $coro->cancel (arg...) |
734 | =item $coro->cancel (arg...) |
446 | |
735 | |
447 | Terminates the given Coro and makes it return the given arguments as |
736 | Terminates the given Coro thread and makes it return the given arguments as |
448 | status (default: the empty list). Never returns if the Coro is the |
737 | status (default: an empty list). Never returns if the Coro is the |
449 | current Coro. |
738 | current Coro. |
450 | |
739 | |
451 | =cut |
740 | This is a rather brutal way to free a coro, with some limitations - if |
|
|
741 | the thread is inside a C callback that doesn't expect to be canceled, |
|
|
742 | bad things can happen, or if the cancelled thread insists on running |
|
|
743 | complicated cleanup handlers that rely on it'S thread context, things will |
|
|
744 | not work. |
452 | |
745 | |
453 | sub cancel { |
746 | Any cleanup code being run (e.g. from C<guard> blocks) will be run without |
454 | my $self = shift; |
747 | a thread context, and is not allowed to switch to other threads. On the |
|
|
748 | plus side, C<< ->cancel >> will always clean up the thread, no matter |
|
|
749 | what. If your cleanup code is complex or you want to avoid cancelling a |
|
|
750 | C-thread that doesn't know how to clean up itself, it can be better to C<< |
|
|
751 | ->throw >> an exception, or use C<< ->safe_cancel >>. |
455 | |
752 | |
456 | if ($current == $self) { |
753 | The arguments to C<< ->cancel >> are not copied, but instead will |
457 | terminate @_; |
754 | be referenced directly (e.g. if you pass C<$var> and after the call |
458 | } else { |
755 | change that variable, then you might change the return values passed to |
459 | $self->{_status} = [@_]; |
756 | e.g. C<join>, so don't do that). |
460 | Coro::State::cancel $self; |
757 | |
|
|
758 | The resources of the Coro are usually freed (or destructed) before this |
|
|
759 | call returns, but this can be delayed for an indefinite amount of time, as |
|
|
760 | in some cases the manager thread has to run first to actually destruct the |
|
|
761 | Coro object. |
|
|
762 | |
|
|
763 | =item $coro->safe_cancel ($arg...) |
|
|
764 | |
|
|
765 | Works mostly like C<< ->cancel >>, but is inherently "safer", and |
|
|
766 | consequently, can fail with an exception in cases the thread is not in a |
|
|
767 | cancellable state. |
|
|
768 | |
|
|
769 | This method works a bit like throwing an exception that cannot be caught |
|
|
770 | - specifically, it will clean up the thread from within itself, so |
|
|
771 | all cleanup handlers (e.g. C<guard> blocks) are run with full thread |
|
|
772 | context and can block if they wish. The downside is that there is no |
|
|
773 | guarantee that the thread can be cancelled when you call this method, and |
|
|
774 | therefore, it might fail. It is also considerably slower than C<cancel> or |
|
|
775 | C<terminate>. |
|
|
776 | |
|
|
777 | A thread is in a safe-cancellable state if it either hasn't been run yet, |
|
|
778 | or it has no C context attached and is inside an SLF function. |
|
|
779 | |
|
|
780 | The latter two basically mean that the thread isn't currently inside a |
|
|
781 | perl callback called from some C function (usually via some XS modules) |
|
|
782 | and isn't currently executing inside some C function itself (via Coro's XS |
|
|
783 | API). |
|
|
784 | |
|
|
785 | This call returns true when it could cancel the thread, or croaks with an |
|
|
786 | error otherwise (i.e. it either returns true or doesn't return at all). |
|
|
787 | |
|
|
788 | Why the weird interface? Well, there are two common models on how and |
|
|
789 | when to cancel things. In the first, you have the expectation that your |
|
|
790 | coro thread can be cancelled when you want to cancel it - if the thread |
|
|
791 | isn't cancellable, this would be a bug somewhere, so C<< ->safe_cancel >> |
|
|
792 | croaks to notify of the bug. |
|
|
793 | |
|
|
794 | In the second model you sometimes want to ask nicely to cancel a thread, |
|
|
795 | but if it's not a good time, well, then don't cancel. This can be done |
|
|
796 | relatively easy like this: |
|
|
797 | |
|
|
798 | if (! eval { $coro->safe_cancel }) { |
|
|
799 | warn "unable to cancel thread: $@"; |
461 | } |
800 | } |
462 | } |
801 | |
|
|
802 | However, what you never should do is first try to cancel "safely" and |
|
|
803 | if that fails, cancel the "hard" way with C<< ->cancel >>. That makes |
|
|
804 | no sense: either you rely on being able to execute cleanup code in your |
|
|
805 | thread context, or you don't. If you do, then C<< ->safe_cancel >> is the |
|
|
806 | only way, and if you don't, then C<< ->cancel >> is always faster and more |
|
|
807 | direct. |
463 | |
808 | |
464 | =item $coro->schedule_to |
809 | =item $coro->schedule_to |
465 | |
810 | |
466 | Puts the current coro to sleep (like C<Coro::schedule>), but instead |
811 | Puts the current coro to sleep (like C<Coro::schedule>), but instead |
467 | of continuing with the next coro from the ready queue, always switch to |
812 | of continuing with the next coro from the ready queue, always switch to |
… | |
… | |
486 | inside the coro at the next convenient point in time. Otherwise |
831 | inside the coro at the next convenient point in time. Otherwise |
487 | clears the exception object. |
832 | clears the exception object. |
488 | |
833 | |
489 | Coro will check for the exception each time a schedule-like-function |
834 | Coro will check for the exception each time a schedule-like-function |
490 | returns, i.e. after each C<schedule>, C<cede>, C<< Coro::Semaphore->down |
835 | returns, i.e. after each C<schedule>, C<cede>, C<< Coro::Semaphore->down |
491 | >>, C<< Coro::Handle->readable >> and so on. Most of these functions |
836 | >>, C<< Coro::Handle->readable >> and so on. Most of those functions (all |
492 | detect this case and return early in case an exception is pending. |
837 | that are part of Coro itself) detect this case and return early in case an |
|
|
838 | exception is pending. |
493 | |
839 | |
494 | The exception object will be thrown "as is" with the specified scalar in |
840 | The exception object will be thrown "as is" with the specified scalar in |
495 | C<$@>, i.e. if it is a string, no line number or newline will be appended |
841 | C<$@>, i.e. if it is a string, no line number or newline will be appended |
496 | (unlike with C<die>). |
842 | (unlike with C<die>). |
497 | |
843 | |
498 | This can be used as a softer means than C<cancel> to ask a coro to |
844 | This can be used as a softer means than either C<cancel> or C<safe_cancel |
499 | end itself, although there is no guarantee that the exception will lead to |
845 | >to ask a coro to end itself, although there is no guarantee that the |
500 | termination, and if the exception isn't caught it might well end the whole |
846 | exception will lead to termination, and if the exception isn't caught it |
501 | program. |
847 | might well end the whole program. |
502 | |
848 | |
503 | You might also think of C<throw> as being the moral equivalent of |
849 | You might also think of C<throw> as being the moral equivalent of |
504 | C<kill>ing a coro with a signal (in this case, a scalar). |
850 | C<kill>ing a coro with a signal (in this case, a scalar). |
505 | |
851 | |
506 | =item $coro->join |
852 | =item $coro->join |
507 | |
853 | |
508 | Wait until the coro terminates and return any values given to the |
854 | Wait until the coro terminates and return any values given to the |
509 | C<terminate> or C<cancel> functions. C<join> can be called concurrently |
855 | C<terminate> or C<cancel> functions. C<join> can be called concurrently |
510 | from multiple coro, and all will be resumed and given the status |
856 | from multiple threads, and all will be resumed and given the status |
511 | return once the C<$coro> terminates. |
857 | return once the C<$coro> terminates. |
512 | |
858 | |
513 | =cut |
859 | =cut |
514 | |
860 | |
515 | sub join { |
861 | sub join { |
… | |
… | |
524 | }; |
870 | }; |
525 | |
871 | |
526 | &schedule while $current; |
872 | &schedule while $current; |
527 | } |
873 | } |
528 | |
874 | |
529 | wantarray ? @{$self->{_status}} : $self->{_status}[0]; |
875 | wantarray ? @{$self->{_status}} : $self->{_status}[0] |
530 | } |
876 | } |
531 | |
877 | |
532 | =item $coro->on_destroy (\&cb) |
878 | =item $coro->on_destroy (\&cb) |
533 | |
879 | |
534 | Registers a callback that is called when this coro gets destroyed, |
880 | Registers a callback that is called when this coro thread gets destroyed, |
535 | but before it is joined. The callback gets passed the terminate arguments, |
881 | that is, after it's resources have been freed but before it is joined. The |
|
|
882 | callback gets passed the terminate/cancel arguments, if any, and I<must |
536 | if any, and I<must not> die, under any circumstances. |
883 | not> die, under any circumstances. |
|
|
884 | |
|
|
885 | There can be any number of C<on_destroy> callbacks per coro, and there is |
|
|
886 | no way currently to remove a callback once added. |
537 | |
887 | |
538 | =cut |
888 | =cut |
539 | |
889 | |
540 | sub on_destroy { |
890 | sub on_destroy { |
541 | my ($self, $cb) = @_; |
891 | my ($self, $cb) = @_; |
… | |
… | |
544 | } |
894 | } |
545 | |
895 | |
546 | =item $oldprio = $coro->prio ($newprio) |
896 | =item $oldprio = $coro->prio ($newprio) |
547 | |
897 | |
548 | Sets (or gets, if the argument is missing) the priority of the |
898 | Sets (or gets, if the argument is missing) the priority of the |
549 | coro. Higher priority coro get run before lower priority |
899 | coro thread. Higher priority coro get run before lower priority |
550 | coro. Priorities are small signed integers (currently -4 .. +3), |
900 | coros. Priorities are small signed integers (currently -4 .. +3), |
551 | that you can refer to using PRIO_xxx constants (use the import tag :prio |
901 | that you can refer to using PRIO_xxx constants (use the import tag :prio |
552 | to get then): |
902 | to get then): |
553 | |
903 | |
554 | PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN |
904 | PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN |
555 | 3 > 1 > 0 > -1 > -3 > -4 |
905 | 3 > 1 > 0 > -1 > -3 > -4 |
556 | |
906 | |
557 | # set priority to HIGH |
907 | # set priority to HIGH |
558 | current->prio (PRIO_HIGH); |
908 | current->prio (PRIO_HIGH); |
559 | |
909 | |
560 | The idle coro ($Coro::idle) always has a lower priority than any |
910 | The idle coro thread ($Coro::idle) always has a lower priority than any |
561 | existing coro. |
911 | existing coro. |
562 | |
912 | |
563 | Changing the priority of the current coro will take effect immediately, |
913 | Changing the priority of the current coro will take effect immediately, |
564 | but changing the priority of coro in the ready queue (but not |
914 | but changing the priority of a coro in the ready queue (but not running) |
565 | running) will only take effect after the next schedule (of that |
915 | will only take effect after the next schedule (of that coro). This is a |
566 | coro). This is a bug that will be fixed in some future version. |
916 | bug that will be fixed in some future version. |
567 | |
917 | |
568 | =item $newprio = $coro->nice ($change) |
918 | =item $newprio = $coro->nice ($change) |
569 | |
919 | |
570 | Similar to C<prio>, but subtract the given value from the priority (i.e. |
920 | Similar to C<prio>, but subtract the given value from the priority (i.e. |
571 | higher values mean lower priority, just as in unix). |
921 | higher values mean lower priority, just as in UNIX's nice command). |
572 | |
922 | |
573 | =item $olddesc = $coro->desc ($newdesc) |
923 | =item $olddesc = $coro->desc ($newdesc) |
574 | |
924 | |
575 | Sets (or gets in case the argument is missing) the description for this |
925 | Sets (or gets in case the argument is missing) the description for this |
576 | coro. This is just a free-form string you can associate with a |
926 | coro thread. This is just a free-form string you can associate with a |
577 | coro. |
927 | coro. |
578 | |
928 | |
579 | This method simply sets the C<< $coro->{desc} >> member to the given |
929 | This method simply sets the C<< $coro->{desc} >> member to the given |
580 | string. You can modify this member directly if you wish. |
930 | string. You can modify this member directly if you wish, and in fact, this |
|
|
931 | is often preferred to indicate major processing states that cna then be |
|
|
932 | seen for example in a L<Coro::Debug> session: |
|
|
933 | |
|
|
934 | sub my_long_function { |
|
|
935 | local $Coro::current->{desc} = "now in my_long_function"; |
|
|
936 | ... |
|
|
937 | $Coro::current->{desc} = "my_long_function: phase 1"; |
|
|
938 | ... |
|
|
939 | $Coro::current->{desc} = "my_long_function: phase 2"; |
|
|
940 | ... |
|
|
941 | } |
581 | |
942 | |
582 | =cut |
943 | =cut |
583 | |
944 | |
584 | sub desc { |
945 | sub desc { |
585 | my $old = $_[0]{desc}; |
946 | my $old = $_[0]{desc}; |
… | |
… | |
622 | returning a new coderef. Unblocking means that calling the new coderef |
983 | returning a new coderef. Unblocking means that calling the new coderef |
623 | will return immediately without blocking, returning nothing, while the |
984 | will return immediately without blocking, returning nothing, while the |
624 | original code ref will be called (with parameters) from within another |
985 | original code ref will be called (with parameters) from within another |
625 | coro. |
986 | coro. |
626 | |
987 | |
627 | The reason this function exists is that many event libraries (such as the |
988 | The reason this function exists is that many event libraries (such as |
628 | venerable L<Event|Event> module) are not thread-safe (a weaker form |
989 | the venerable L<Event|Event> module) are not thread-safe (a weaker form |
629 | of reentrancy). This means you must not block within event callbacks, |
990 | of reentrancy). This means you must not block within event callbacks, |
630 | otherwise you might suffer from crashes or worse. The only event library |
991 | otherwise you might suffer from crashes or worse. The only event library |
631 | currently known that is safe to use without C<unblock_sub> is L<EV>. |
992 | currently known that is safe to use without C<unblock_sub> is L<EV> (but |
|
|
993 | you might still run into deadlocks if all event loops are blocked). |
|
|
994 | |
|
|
995 | Coro will try to catch you when you block in the event loop |
|
|
996 | ("FATAL:$Coro::IDLE blocked itself"), but this is just best effort and |
|
|
997 | only works when you do not run your own event loop. |
632 | |
998 | |
633 | This function allows your callbacks to block by executing them in another |
999 | This function allows your callbacks to block by executing them in another |
634 | coro where it is safe to block. One example where blocking is handy |
1000 | coro where it is safe to block. One example where blocking is handy |
635 | is when you use the L<Coro::AIO|Coro::AIO> functions to save results to |
1001 | is when you use the L<Coro::AIO|Coro::AIO> functions to save results to |
636 | disk, for example. |
1002 | disk, for example. |
… | |
… | |
678 | unshift @unblock_queue, [$cb, @_]; |
1044 | unshift @unblock_queue, [$cb, @_]; |
679 | $unblock_scheduler->ready; |
1045 | $unblock_scheduler->ready; |
680 | } |
1046 | } |
681 | } |
1047 | } |
682 | |
1048 | |
683 | =item $cb = Coro::rouse_cb |
1049 | =item $cb = rouse_cb |
684 | |
1050 | |
685 | Create and return a "rouse callback". That's a code reference that, |
1051 | Create and return a "rouse callback". That's a code reference that, |
686 | when called, will remember a copy of its arguments and notify the owner |
1052 | when called, will remember a copy of its arguments and notify the owner |
687 | coro of the callback. |
1053 | coro of the callback. |
688 | |
1054 | |
689 | See the next function. |
1055 | See the next function. |
690 | |
1056 | |
691 | =item @args = Coro::rouse_wait [$cb] |
1057 | =item @args = rouse_wait [$cb] |
692 | |
1058 | |
693 | Wait for the specified rouse callback (or the last one that was created in |
1059 | Wait for the specified rouse callback (or the last one that was created in |
694 | this coro). |
1060 | this coro). |
695 | |
1061 | |
696 | As soon as the callback is invoked (or when the callback was invoked |
1062 | As soon as the callback is invoked (or when the callback was invoked |
697 | before C<rouse_wait>), it will return the arguments originally passed to |
1063 | before C<rouse_wait>), it will return the arguments originally passed to |
698 | the rouse callback. |
1064 | the rouse callback. In scalar context, that means you get the I<last> |
|
|
1065 | argument, just as if C<rouse_wait> had a C<return ($a1, $a2, $a3...)> |
|
|
1066 | statement at the end. |
699 | |
1067 | |
700 | See the section B<HOW TO WAIT FOR A CALLBACK> for an actual usage example. |
1068 | See the section B<HOW TO WAIT FOR A CALLBACK> for an actual usage example. |
701 | |
1069 | |
702 | =back |
1070 | =back |
703 | |
1071 | |
704 | =cut |
1072 | =cut |
|
|
1073 | |
|
|
1074 | for my $module (qw(Channel RWLock Semaphore SemaphoreSet Signal Specific)) { |
|
|
1075 | my $old = defined &{"Coro::$module\::new"} && \&{"Coro::$module\::new"}; |
|
|
1076 | |
|
|
1077 | *{"Coro::$module\::new"} = sub { |
|
|
1078 | require "Coro/$module.pm"; |
|
|
1079 | |
|
|
1080 | # some modules have their new predefined in State.xs, some don't |
|
|
1081 | *{"Coro::$module\::new"} = $old |
|
|
1082 | if $old; |
|
|
1083 | |
|
|
1084 | goto &{"Coro::$module\::new"}; |
|
|
1085 | }; |
|
|
1086 | } |
705 | |
1087 | |
706 | 1; |
1088 | 1; |
707 | |
1089 | |
708 | =head1 HOW TO WAIT FOR A CALLBACK |
1090 | =head1 HOW TO WAIT FOR A CALLBACK |
709 | |
1091 | |
… | |
… | |
791 | the windows process emulation enabled under unix roughly halves perl |
1173 | the windows process emulation enabled under unix roughly halves perl |
792 | performance, even when not used. |
1174 | performance, even when not used. |
793 | |
1175 | |
794 | =item coro switching is not signal safe |
1176 | =item coro switching is not signal safe |
795 | |
1177 | |
796 | You must not switch to another coro from within a signal handler |
1178 | You must not switch to another coro from within a signal handler (only |
797 | (only relevant with %SIG - most event libraries provide safe signals). |
1179 | relevant with %SIG - most event libraries provide safe signals), I<unless> |
|
|
1180 | you are sure you are not interrupting a Coro function. |
798 | |
1181 | |
799 | That means you I<MUST NOT> call any function that might "block" the |
1182 | That means you I<MUST NOT> call any function that might "block" the |
800 | current coro - C<cede>, C<schedule> C<< Coro::Semaphore->down >> or |
1183 | current coro - C<cede>, C<schedule> C<< Coro::Semaphore->down >> or |
801 | anything that calls those. Everything else, including calling C<ready>, |
1184 | anything that calls those. Everything else, including calling C<ready>, |
802 | works. |
1185 | works. |
803 | |
1186 | |
804 | =back |
1187 | =back |
805 | |
1188 | |
806 | |
1189 | |
|
|
1190 | =head1 WINDOWS PROCESS EMULATION |
|
|
1191 | |
|
|
1192 | A great many people seem to be confused about ithreads (for example, Chip |
|
|
1193 | Salzenberg called me unintelligent, incapable, stupid and gullible, |
|
|
1194 | while in the same mail making rather confused statements about perl |
|
|
1195 | ithreads (for example, that memory or files would be shared), showing his |
|
|
1196 | lack of understanding of this area - if it is hard to understand for Chip, |
|
|
1197 | it is probably not obvious to everybody). |
|
|
1198 | |
|
|
1199 | What follows is an ultra-condensed version of my talk about threads in |
|
|
1200 | scripting languages given on the perl workshop 2009: |
|
|
1201 | |
|
|
1202 | The so-called "ithreads" were originally implemented for two reasons: |
|
|
1203 | first, to (badly) emulate unix processes on native win32 perls, and |
|
|
1204 | secondly, to replace the older, real thread model ("5.005-threads"). |
|
|
1205 | |
|
|
1206 | It does that by using threads instead of OS processes. The difference |
|
|
1207 | between processes and threads is that threads share memory (and other |
|
|
1208 | state, such as files) between threads within a single process, while |
|
|
1209 | processes do not share anything (at least not semantically). That |
|
|
1210 | means that modifications done by one thread are seen by others, while |
|
|
1211 | modifications by one process are not seen by other processes. |
|
|
1212 | |
|
|
1213 | The "ithreads" work exactly like that: when creating a new ithreads |
|
|
1214 | process, all state is copied (memory is copied physically, files and code |
|
|
1215 | is copied logically). Afterwards, it isolates all modifications. On UNIX, |
|
|
1216 | the same behaviour can be achieved by using operating system processes, |
|
|
1217 | except that UNIX typically uses hardware built into the system to do this |
|
|
1218 | efficiently, while the windows process emulation emulates this hardware in |
|
|
1219 | software (rather efficiently, but of course it is still much slower than |
|
|
1220 | dedicated hardware). |
|
|
1221 | |
|
|
1222 | As mentioned before, loading code, modifying code, modifying data |
|
|
1223 | structures and so on is only visible in the ithreads process doing the |
|
|
1224 | modification, not in other ithread processes within the same OS process. |
|
|
1225 | |
|
|
1226 | This is why "ithreads" do not implement threads for perl at all, only |
|
|
1227 | processes. What makes it so bad is that on non-windows platforms, you can |
|
|
1228 | actually take advantage of custom hardware for this purpose (as evidenced |
|
|
1229 | by the forks module, which gives you the (i-) threads API, just much |
|
|
1230 | faster). |
|
|
1231 | |
|
|
1232 | Sharing data is in the i-threads model is done by transfering data |
|
|
1233 | structures between threads using copying semantics, which is very slow - |
|
|
1234 | shared data simply does not exist. Benchmarks using i-threads which are |
|
|
1235 | communication-intensive show extremely bad behaviour with i-threads (in |
|
|
1236 | fact, so bad that Coro, which cannot take direct advantage of multiple |
|
|
1237 | CPUs, is often orders of magnitude faster because it shares data using |
|
|
1238 | real threads, refer to my talk for details). |
|
|
1239 | |
|
|
1240 | As summary, i-threads *use* threads to implement processes, while |
|
|
1241 | the compatible forks module *uses* processes to emulate, uhm, |
|
|
1242 | processes. I-threads slow down every perl program when enabled, and |
|
|
1243 | outside of windows, serve no (or little) practical purpose, but |
|
|
1244 | disadvantages every single-threaded Perl program. |
|
|
1245 | |
|
|
1246 | This is the reason that I try to avoid the name "ithreads", as it is |
|
|
1247 | misleading as it implies that it implements some kind of thread model for |
|
|
1248 | perl, and prefer the name "windows process emulation", which describes the |
|
|
1249 | actual use and behaviour of it much better. |
|
|
1250 | |
807 | =head1 SEE ALSO |
1251 | =head1 SEE ALSO |
808 | |
1252 | |
809 | Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>. |
1253 | Event-Loop integration: L<Coro::AnyEvent>, L<Coro::EV>, L<Coro::Event>. |
810 | |
1254 | |
811 | Debugging: L<Coro::Debug>. |
1255 | Debugging: L<Coro::Debug>. |