1 | NAME |
1 | NAME |
2 | Coro - the only real threads in perl |
2 | Coro - the only real threads in perl |
3 | |
3 | |
4 | SYNOPSIS |
4 | SYNOPSIS |
5 | use Coro; |
5 | use Coro; |
6 | |
6 | |
7 | async { |
7 | async { |
8 | # some asynchronous thread of execution |
8 | # some asynchronous thread of execution |
9 | print "2\n"; |
9 | print "2\n"; |
10 | cede; # yield back to main |
10 | cede; # yield back to main |
11 | print "4\n"; |
11 | print "4\n"; |
12 | }; |
12 | }; |
13 | print "1\n"; |
13 | print "1\n"; |
14 | cede; # yield to coroutine |
14 | cede; # yield to coro |
15 | print "3\n"; |
15 | print "3\n"; |
16 | cede; # and again |
16 | cede; # and again |
17 | |
17 | |
18 | # use locking |
18 | # use locking |
19 | use Coro::Semaphore; |
19 | use Coro::Semaphore; |
20 | my $lock = new Coro::Semaphore; |
20 | my $lock = new Coro::Semaphore; |
21 | my $locked; |
21 | my $locked; |
22 | |
22 | |
23 | $lock->down; |
23 | $lock->down; |
24 | $locked = 1; |
24 | $locked = 1; |
25 | $lock->up; |
25 | $lock->up; |
26 | |
26 | |
27 | DESCRIPTION |
27 | DESCRIPTION |
28 | For a tutorial-style introduction, please read the Coro::Intro manpage. |
28 | For a tutorial-style introduction, please read the Coro::Intro manpage. |
29 | This manpage mainly contains reference information. |
29 | This manpage mainly contains reference information. |
30 | |
30 | |
31 | This module collection manages continuations in general, most often in |
31 | This module collection manages continuations in general, most often in |
32 | the form of cooperative threads (also called coroutines in the |
32 | the form of cooperative threads (also called coros, or simply "coro" in |
33 | documentation). They are similar to kernel threads but don't (in |
33 | the documentation). They are similar to kernel threads but don't (in |
34 | general) run in parallel at the same time even on SMP machines. The |
34 | general) run in parallel at the same time even on SMP machines. The |
35 | specific flavor of thread offered by this module also guarantees you |
35 | specific flavor of thread offered by this module also guarantees you |
36 | that it will not switch between threads unless necessary, at |
36 | that it will not switch between threads unless necessary, at |
37 | easily-identified points in your program, so locking and parallel access |
37 | easily-identified points in your program, so locking and parallel access |
38 | are rarely an issue, making thread programming much safer and easier |
38 | are rarely an issue, making thread programming much safer and easier |
39 | than using other thread models. |
39 | than using other thread models. |
40 | |
40 | |
41 | Unlike the so-called "Perl threads" (which are not actually real threads |
41 | Unlike the so-called "Perl threads" (which are not actually real threads |
42 | but only the windows process emulation ported to unix), Coro provides a |
42 | but only the windows process emulation (see section of same name for |
|
|
43 | more details) ported to UNIX, and as such act as processes), Coro |
43 | full shared address space, which makes communication between threads |
44 | provides a full shared address space, which makes communication between |
44 | very easy. And threads are fast, too: disabling the Windows process |
45 | threads very easy. And coro threads are fast, too: disabling the Windows |
45 | emulation code in your perl and using Coro can easily result in a two to |
46 | process emulation code in your perl and using Coro can easily result in |
46 | four times speed increase for your programs. |
47 | a two to four times speed increase for your programs. A parallel matrix |
|
|
48 | multiplication benchmark (very communication-intensive) runs over 300 |
|
|
49 | times faster on a single core than perls pseudo-threads on a quad core |
|
|
50 | using all four cores. |
47 | |
51 | |
48 | Coro achieves that by supporting multiple running interpreters that |
52 | Coro achieves that by supporting multiple running interpreters that |
49 | share data, which is especially useful to code pseudo-parallel processes |
53 | share data, which is especially useful to code pseudo-parallel processes |
50 | and for event-based programming, such as multiple HTTP-GET requests |
54 | and for event-based programming, such as multiple HTTP-GET requests |
51 | running concurrently. See Coro::AnyEvent to learn more on how to |
55 | running concurrently. See Coro::AnyEvent to learn more on how to |
52 | integrate Coro into an event-based environment. |
56 | integrate Coro into an event-based environment. |
53 | |
57 | |
54 | In this module, a thread is defined as "callchain + lexical variables + |
58 | In this module, a thread is defined as "callchain + lexical variables + |
55 | @_ + $_ + $@ + $/ + C stack), that is, a thread has its own callchain, |
59 | some package variables + C stack), that is, a thread has its own |
56 | its own set of lexicals and its own set of perls most important global |
60 | callchain, its own set of lexicals and its own set of perls most |
57 | variables (see Coro::State for more configuration and background info). |
61 | important global variables (see Coro::State for more configuration and |
|
|
62 | background info). |
58 | |
63 | |
59 | See also the "SEE ALSO" section at the end of this document - the Coro |
64 | See also the "SEE ALSO" section at the end of this document - the Coro |
60 | module family is quite large. |
65 | module family is quite large. |
61 | |
66 | |
|
|
67 | CORO THREAD LIFE CYCLE |
|
|
68 | During the long and exciting (or not) life of a coro thread, it goes |
|
|
69 | through a number of states: |
|
|
70 | |
|
|
71 | 1. Creation |
|
|
72 | The first thing in the life of a coro thread is it's creation - |
|
|
73 | obviously. The typical way to create a thread is to call the "async |
|
|
74 | BLOCK" function: |
|
|
75 | |
|
|
76 | async { |
|
|
77 | # thread code goes here |
|
|
78 | }; |
|
|
79 | |
|
|
80 | You can also pass arguments, which are put in @_: |
|
|
81 | |
|
|
82 | async { |
|
|
83 | print $_[1]; # prints 2 |
|
|
84 | } 1, 2, 3; |
|
|
85 | |
|
|
86 | This creates a new coro thread and puts it into the ready queue, |
|
|
87 | meaning it will run as soon as the CPU is free for it. |
|
|
88 | |
|
|
89 | "async" will return a coro object - you can store this for future |
|
|
90 | reference or ignore it, the thread itself will keep a reference to |
|
|
91 | it's thread object - threads are alive on their own. |
|
|
92 | |
|
|
93 | Another way to create a thread is to call the "new" constructor with |
|
|
94 | a code-reference: |
|
|
95 | |
|
|
96 | new Coro sub { |
|
|
97 | # thread code goes here |
|
|
98 | }, @optional_arguments; |
|
|
99 | |
|
|
100 | This is quite similar to calling "async", but the important |
|
|
101 | difference is that the new thread is not put into the ready queue, |
|
|
102 | so the thread will not run until somebody puts it there. "async" is, |
|
|
103 | therefore, identical to this sequence: |
|
|
104 | |
|
|
105 | my $coro = new Coro sub { |
|
|
106 | # thread code goes here |
|
|
107 | }; |
|
|
108 | $coro->ready; |
|
|
109 | return $coro; |
|
|
110 | |
|
|
111 | 2. Startup |
|
|
112 | When a new coro thread is created, only a copy of the code reference |
|
|
113 | and the arguments are stored, no extra memory for stacks and so on |
|
|
114 | is allocated, keeping the coro thread in a low-memory state. |
|
|
115 | |
|
|
116 | Only when it actually starts executing will all the resources be |
|
|
117 | finally allocated. |
|
|
118 | |
|
|
119 | The optional arguments specified at coro creation are available in |
|
|
120 | @_, similar to function calls. |
|
|
121 | |
|
|
122 | 3. Running / Blocking |
|
|
123 | A lot can happen after the coro thread has started running. Quite |
|
|
124 | usually, it will not run to the end in one go (because you could use |
|
|
125 | a function instead), but it will give up the CPU regularly because |
|
|
126 | it waits for external events. |
|
|
127 | |
|
|
128 | As long as a coro thread runs, it's coro object is available in the |
|
|
129 | global variable $Coro::current. |
|
|
130 | |
|
|
131 | The low-level way to give up the CPU is to call the scheduler, which |
|
|
132 | selects a new coro thread to run: |
|
|
133 | |
|
|
134 | Coro::schedule; |
|
|
135 | |
|
|
136 | Since running threads are not in the ready queue, calling the |
|
|
137 | scheduler without doing anything else will block the coro thread |
|
|
138 | forever - you need to arrange either for the coro to put woken up |
|
|
139 | (readied) by some other event or some other thread, or you can put |
|
|
140 | it into the ready queue before scheduling: |
|
|
141 | |
|
|
142 | # this is exactly what Coro::cede does |
|
|
143 | $Coro::current->ready; |
|
|
144 | Coro::schedule; |
|
|
145 | |
|
|
146 | All the higher-level synchronisation methods (Coro::Semaphore, |
|
|
147 | Coro::rouse_*...) are actually implemented via "->ready" and |
|
|
148 | "Coro::schedule". |
|
|
149 | |
|
|
150 | While the coro thread is running it also might get assigned a |
|
|
151 | C-level thread, or the C-level thread might be unassigned from it, |
|
|
152 | as the Coro runtime wishes. A C-level thread needs to be assigned |
|
|
153 | when your perl thread calls into some C-level function and that |
|
|
154 | function in turn calls perl and perl then wants to switch |
|
|
155 | coroutines. This happens most often when you run an event loop and |
|
|
156 | block in the callback, or when perl itself calls some function such |
|
|
157 | as "AUTOLOAD" or methods via the "tie" mechanism. |
|
|
158 | |
|
|
159 | 4. Termination |
|
|
160 | Many threads actually terminate after some time. There are a number |
|
|
161 | of ways to terminate a coro thread, the simplest is returning from |
|
|
162 | the top-level code reference: |
|
|
163 | |
|
|
164 | async { |
|
|
165 | # after returning from here, the coro thread is terminated |
|
|
166 | }; |
|
|
167 | |
|
|
168 | async { |
|
|
169 | return if 0.5 < rand; # terminate a little earlier, maybe |
|
|
170 | print "got a chance to print this\n"; |
|
|
171 | # or here |
|
|
172 | }; |
|
|
173 | |
|
|
174 | Any values returned from the coroutine can be recovered using |
|
|
175 | "->join": |
|
|
176 | |
|
|
177 | my $coro = async { |
|
|
178 | "hello, world\n" # return a string |
|
|
179 | }; |
|
|
180 | |
|
|
181 | my $hello_world = $coro->join; |
|
|
182 | |
|
|
183 | print $hello_world; |
|
|
184 | |
|
|
185 | Another way to terminate is to call "Coro::terminate", which at any |
|
|
186 | subroutine call nesting level: |
|
|
187 | |
|
|
188 | async { |
|
|
189 | Coro::terminate "return value 1", "return value 2"; |
|
|
190 | }; |
|
|
191 | |
|
|
192 | And yet another way is to "->cancel" the coro thread from another |
|
|
193 | thread: |
|
|
194 | |
|
|
195 | my $coro = async { |
|
|
196 | exit 1; |
|
|
197 | }; |
|
|
198 | |
|
|
199 | $coro->cancel; # an also accept values for ->join to retrieve |
|
|
200 | |
|
|
201 | Cancellation *can* be dangerous - it's a bit like calling "exit" |
|
|
202 | without actually exiting, and might leave C libraries and XS modules |
|
|
203 | in a weird state. Unlike other thread implementations, however, Coro |
|
|
204 | is exceptionally safe with regards to cancellation, as perl will |
|
|
205 | always be in a consistent state. |
|
|
206 | |
|
|
207 | So, cancelling a thread that runs in an XS event loop might not be |
|
|
208 | the best idea, but any other combination that deals with perl only |
|
|
209 | (cancelling when a thread is in a "tie" method or an "AUTOLOAD" for |
|
|
210 | example) is safe. |
|
|
211 | |
|
|
212 | 5. Cleanup |
|
|
213 | Threads will allocate various resources. Most but not all will be |
|
|
214 | returned when a thread terminates, during clean-up. |
|
|
215 | |
|
|
216 | Cleanup is quite similar to throwing an uncaught exception: perl |
|
|
217 | will work it's way up through all subroutine calls and blocks. On |
|
|
218 | it's way, it will release all "my" variables, undo all "local"'s and |
|
|
219 | free any other resources truly local to the thread. |
|
|
220 | |
|
|
221 | So, a common way to free resources is to keep them referenced only |
|
|
222 | by my variables: |
|
|
223 | |
|
|
224 | async { |
|
|
225 | my $big_cache = new Cache ...; |
|
|
226 | }; |
|
|
227 | |
|
|
228 | If there are no other references, then the $big_cache object will be |
|
|
229 | freed when the thread terminates, regardless of how it does so. |
|
|
230 | |
|
|
231 | What it does "NOT" do is unlock any Coro::Semaphores or similar |
|
|
232 | resources, but that's where the "guard" methods come in handy: |
|
|
233 | |
|
|
234 | my $sem = new Coro::Semaphore; |
|
|
235 | |
|
|
236 | async { |
|
|
237 | my $lock_guard = $sem->guard; |
|
|
238 | # if we reutrn, or die or get cancelled, here, |
|
|
239 | # then the semaphore will be "up"ed. |
|
|
240 | }; |
|
|
241 | |
|
|
242 | The "Guard::guard" function comes in handy for any custom cleanup |
|
|
243 | you might want to do: |
|
|
244 | |
|
|
245 | async { |
|
|
246 | my $window = new Gtk2::Window "toplevel"; |
|
|
247 | # The window will not be cleaned up automatically, even when $window |
|
|
248 | # gets freed, so use a guard to ensure it's destruction |
|
|
249 | # in case of an error: |
|
|
250 | my $window_guard = Guard::guard { $window->destroy }; |
|
|
251 | |
|
|
252 | # we are safe here |
|
|
253 | }; |
|
|
254 | |
|
|
255 | Last not least, "local" can often be handy, too, e.g. when |
|
|
256 | temporarily replacing the coro thread description: |
|
|
257 | |
|
|
258 | sub myfunction { |
|
|
259 | local $Coro::current->{desc} = "inside myfunction(@_)"; |
|
|
260 | |
|
|
261 | # if we return or die here, the description will be restored |
|
|
262 | } |
|
|
263 | |
|
|
264 | 6. Viva La Zombie Muerte |
|
|
265 | Even after a thread has terminated and cleaned up it's resources, |
|
|
266 | the coro object still is there and stores the return values of the |
|
|
267 | thread. Only in this state will the coro object be "reference |
|
|
268 | counted" in the normal perl sense: the thread code keeps a reference |
|
|
269 | to it when it is active, but not after it has terminated. |
|
|
270 | |
|
|
271 | The means the coro object gets freed automatically when the thread |
|
|
272 | has terminated and cleaned up and there arenot other references. |
|
|
273 | |
|
|
274 | If there are, the coro object will stay around, and you can call |
|
|
275 | "->join" as many times as you wish to retrieve the result values: |
|
|
276 | |
|
|
277 | async { |
|
|
278 | print "hi\n"; |
|
|
279 | 1 |
|
|
280 | }; |
|
|
281 | |
|
|
282 | # run the async above, and free everything before returning |
|
|
283 | # from Coro::cede: |
|
|
284 | Coro::cede; |
|
|
285 | |
|
|
286 | { |
|
|
287 | my $coro = async { |
|
|
288 | print "hi\n"; |
|
|
289 | 1 |
|
|
290 | }; |
|
|
291 | |
|
|
292 | # run the async above, and clean up, but do not free the coro |
|
|
293 | # object: |
|
|
294 | Coro::cede; |
|
|
295 | |
|
|
296 | # optionally retrieve the result values |
|
|
297 | my @results = $coro->join; |
|
|
298 | |
|
|
299 | # now $coro goes out of scope, and presumably gets freed |
|
|
300 | }; |
|
|
301 | |
62 | GLOBAL VARIABLES |
302 | GLOBAL VARIABLES |
63 | $Coro::main |
303 | $Coro::main |
64 | This variable stores the coroutine object that represents the main |
304 | This variable stores the Coro object that represents the main |
65 | program. While you cna "ready" it and do most other things you can |
305 | program. While you cna "ready" it and do most other things you can |
66 | do to coroutines, it is mainly useful to compare again |
306 | do to coro, it is mainly useful to compare again $Coro::current, to |
67 | $Coro::current, to see whether you are running in the main program |
307 | see whether you are running in the main program or not. |
68 | or not. |
|
|
69 | |
308 | |
70 | $Coro::current |
309 | $Coro::current |
71 | The coroutine object representing the current coroutine (the last |
310 | The Coro object representing the current coro (the last coro that |
72 | coroutine that the Coro scheduler switched to). The initial value is |
311 | the Coro scheduler switched to). The initial value is $Coro::main |
73 | $Coro::main (of course). |
312 | (of course). |
74 | |
313 | |
75 | This variable is strictly *read-only*. You can take copies of the |
314 | This variable is strictly *read-only*. You can take copies of the |
76 | value stored in it and use it as any other coroutine object, but you |
315 | value stored in it and use it as any other Coro object, but you must |
77 | must not otherwise modify the variable itself. |
316 | not otherwise modify the variable itself. |
78 | |
317 | |
79 | $Coro::idle |
318 | $Coro::idle |
80 | This variable is mainly useful to integrate Coro into event loops. |
319 | This variable is mainly useful to integrate Coro into event loops. |
81 | It is usually better to rely on Coro::AnyEvent or Coro::EV, as this |
320 | It is usually better to rely on Coro::AnyEvent or Coro::EV, as this |
82 | is pretty low-level functionality. |
321 | is pretty low-level functionality. |
83 | |
322 | |
84 | This variable stores either a coroutine or a callback. |
323 | This variable stores a Coro object that is put into the ready queue |
|
|
324 | when there are no other ready threads (without invoking any ready |
|
|
325 | hooks). |
85 | |
326 | |
86 | If it is a callback, the it is called whenever the scheduler finds |
327 | The default implementation dies with "FATAL: deadlock detected.", |
87 | no ready coroutines to run. The default implementation prints |
328 | followed by a thread listing, because the program has no other way |
88 | "FATAL: deadlock detected" and exits, because the program has no |
|
|
89 | other way to continue. |
329 | to continue. |
90 | |
|
|
91 | If it is a coroutine object, then this object will be readied |
|
|
92 | (without invoking any ready hooks, however) when the scheduler finds |
|
|
93 | no other ready coroutines to run. |
|
|
94 | |
330 | |
95 | This hook is overwritten by modules such as "Coro::EV" and |
331 | This hook is overwritten by modules such as "Coro::EV" and |
96 | "Coro::AnyEvent" to wait on an external event that hopefully wake up |
332 | "Coro::AnyEvent" to wait on an external event that hopefully wakes |
97 | a coroutine so the scheduler can run it. |
333 | up a coro so the scheduler can run it. |
98 | |
334 | |
99 | Note that the callback *must not*, under any circumstances, block |
|
|
100 | the current coroutine. Normally, this is achieved by having an "idle |
|
|
101 | coroutine" that calls the event loop and then blocks again, and then |
|
|
102 | readying that coroutine in the idle handler, or by simply placing |
|
|
103 | the idle coroutine in this variable. |
|
|
104 | |
|
|
105 | See Coro::Event or Coro::AnyEvent for examples of using this |
335 | See Coro::EV or Coro::AnyEvent for examples of using this technique. |
106 | technique. |
|
|
107 | |
336 | |
108 | Please note that if your callback recursively invokes perl (e.g. for |
|
|
109 | event handlers), then it must be prepared to be called recursively |
|
|
110 | itself. |
|
|
111 | |
|
|
112 | SIMPLE COROUTINE CREATION |
337 | SIMPLE CORO CREATION |
113 | async { ... } [@args...] |
338 | async { ... } [@args...] |
114 | Create a new coroutine and return its coroutine object (usually |
339 | Create a new coro and return its Coro object (usually unused). The |
115 | unused). The coroutine will be put into the ready queue, so it will |
340 | coro will be put into the ready queue, so it will start running |
116 | start running automatically on the next scheduler run. |
341 | automatically on the next scheduler run. |
117 | |
342 | |
118 | The first argument is a codeblock/closure that should be executed in |
343 | The first argument is a codeblock/closure that should be executed in |
119 | the coroutine. When it returns argument returns the coroutine is |
344 | the coro. When it returns argument returns the coro is automatically |
120 | automatically terminated. |
345 | terminated. |
121 | |
346 | |
122 | The remaining arguments are passed as arguments to the closure. |
347 | The remaining arguments are passed as arguments to the closure. |
123 | |
348 | |
124 | See the "Coro::State::new" constructor for info about the coroutine |
349 | See the "Coro::State::new" constructor for info about the coro |
125 | environment in which coroutines are executed. |
350 | environment in which coro are executed. |
126 | |
351 | |
127 | Calling "exit" in a coroutine will do the same as calling exit |
352 | Calling "exit" in a coro will do the same as calling exit outside |
128 | outside the coroutine. Likewise, when the coroutine dies, the |
353 | the coro. Likewise, when the coro dies, the program will exit, just |
129 | program will exit, just as it would in the main program. |
354 | as it would in the main program. |
130 | |
355 | |
131 | If you do not want that, you can provide a default "die" handler, or |
356 | If you do not want that, you can provide a default "die" handler, or |
132 | simply avoid dieing (by use of "eval"). |
357 | simply avoid dieing (by use of "eval"). |
133 | |
358 | |
134 | Example: Create a new coroutine that just prints its arguments. |
359 | Example: Create a new coro that just prints its arguments. |
135 | |
360 | |
136 | async { |
361 | async { |
137 | print "@_\n"; |
362 | print "@_\n"; |
138 | } 1,2,3,4; |
363 | } 1,2,3,4; |
139 | |
364 | |
140 | async_pool { ... } [@args...] |
365 | async_pool { ... } [@args...] |
141 | Similar to "async", but uses a coroutine pool, so you should not |
366 | Similar to "async", but uses a coro pool, so you should not call |
142 | call terminate or join on it (although you are allowed to), and you |
367 | terminate or join on it (although you are allowed to), and you get a |
143 | get a coroutine that might have executed other code already (which |
368 | coro that might have executed other code already (which can be good |
144 | can be good or bad :). |
369 | or bad :). |
145 | |
370 | |
146 | On the plus side, this function is about twice as fast as creating |
371 | On the plus side, this function is about twice as fast as creating |
147 | (and destroying) a completely new coroutine, so if you need a lot of |
372 | (and destroying) a completely new coro, so if you need a lot of |
148 | generic coroutines in quick successsion, use "async_pool", not |
373 | generic coros in quick successsion, use "async_pool", not "async". |
149 | "async". |
|
|
150 | |
374 | |
151 | The code block is executed in an "eval" context and a warning will |
375 | The code block is executed in an "eval" context and a warning will |
152 | be issued in case of an exception instead of terminating the |
376 | be issued in case of an exception instead of terminating the |
153 | program, as "async" does. As the coroutine is being reused, stuff |
377 | program, as "async" does. As the coro is being reused, stuff like |
154 | like "on_destroy" will not work in the expected way, unless you call |
378 | "on_destroy" will not work in the expected way, unless you call |
155 | terminate or cancel, which somehow defeats the purpose of pooling |
379 | terminate or cancel, which somehow defeats the purpose of pooling |
156 | (but is fine in the exceptional case). |
380 | (but is fine in the exceptional case). |
157 | |
381 | |
158 | The priority will be reset to 0 after each run, tracing will be |
382 | The priority will be reset to 0 after each run, tracing will be |
159 | disabled, the description will be reset and the default output |
383 | disabled, the description will be reset and the default output |
160 | filehandle gets restored, so you can change all these. Otherwise the |
384 | filehandle gets restored, so you can change all these. Otherwise the |
161 | coroutine will be re-used "as-is": most notably if you change other |
385 | coro will be re-used "as-is": most notably if you change other |
162 | per-coroutine global stuff such as $/ you *must needs* revert that |
386 | per-coro global stuff such as $/ you *must needs* revert that |
163 | change, which is most simply done by using local as in: "local $/". |
387 | change, which is most simply done by using local as in: "local $/". |
164 | |
388 | |
165 | The idle pool size is limited to 8 idle coroutines (this can be |
389 | The idle pool size is limited to 8 idle coros (this can be adjusted |
166 | adjusted by changing $Coro::POOL_SIZE), but there can be as many |
390 | by changing $Coro::POOL_SIZE), but there can be as many non-idle |
167 | non-idle coros as required. |
391 | coros as required. |
168 | |
392 | |
169 | If you are concerned about pooled coroutines growing a lot because a |
393 | If you are concerned about pooled coros growing a lot because a |
170 | single "async_pool" used a lot of stackspace you can e.g. |
394 | single "async_pool" used a lot of stackspace you can e.g. |
171 | "async_pool { terminate }" once per second or so to slowly replenish |
395 | "async_pool { terminate }" once per second or so to slowly replenish |
172 | the pool. In addition to that, when the stacks used by a handler |
396 | the pool. In addition to that, when the stacks used by a handler |
173 | grows larger than 32kb (adjustable via $Coro::POOL_RSS) it will also |
397 | grows larger than 32kb (adjustable via $Coro::POOL_RSS) it will also |
174 | be destroyed. |
398 | be destroyed. |
175 | |
399 | |
176 | STATIC METHODS |
400 | STATIC METHODS |
177 | Static methods are actually functions that implicitly operate on the |
401 | Static methods are actually functions that implicitly operate on the |
178 | current coroutine. |
402 | current coro. |
179 | |
403 | |
180 | schedule |
404 | schedule |
181 | Calls the scheduler. The scheduler will find the next coroutine that |
405 | Calls the scheduler. The scheduler will find the next coro that is |
182 | is to be run from the ready queue and switches to it. The next |
406 | to be run from the ready queue and switches to it. The next coro to |
183 | coroutine to be run is simply the one with the highest priority that |
407 | be run is simply the one with the highest priority that is longest |
184 | is longest in its ready queue. If there is no coroutine ready, it |
408 | in its ready queue. If there is no coro ready, it will call the |
185 | will clal the $Coro::idle hook. |
409 | $Coro::idle hook. |
186 | |
410 | |
187 | Please note that the current coroutine will *not* be put into the |
411 | Please note that the current coro will *not* be put into the ready |
188 | ready queue, so calling this function usually means you will never |
412 | queue, so calling this function usually means you will never be |
189 | be called again unless something else (e.g. an event handler) calls |
413 | called again unless something else (e.g. an event handler) calls |
190 | "->ready", thus waking you up. |
414 | "->ready", thus waking you up. |
191 | |
415 | |
192 | This makes "schedule" *the* generic method to use to block the |
416 | This makes "schedule" *the* generic method to use to block the |
193 | current coroutine and wait for events: first you remember the |
417 | current coro and wait for events: first you remember the current |
194 | current coroutine in a variable, then arrange for some callback of |
418 | coro in a variable, then arrange for some callback of yours to call |
195 | yours to call "->ready" on that once some event happens, and last |
419 | "->ready" on that once some event happens, and last you call |
196 | you call "schedule" to put yourself to sleep. Note that a lot of |
420 | "schedule" to put yourself to sleep. Note that a lot of things can |
197 | things can wake your coroutine up, so you need to check whether the |
421 | wake your coro up, so you need to check whether the event indeed |
198 | event indeed happened, e.g. by storing the status in a variable. |
422 | happened, e.g. by storing the status in a variable. |
199 | |
423 | |
200 | See HOW TO WAIT FOR A CALLBACK, below, for some ways to wait for |
424 | See HOW TO WAIT FOR A CALLBACK, below, for some ways to wait for |
201 | callbacks. |
425 | callbacks. |
202 | |
426 | |
203 | cede |
427 | cede |
204 | "Cede" to other coroutines. This function puts the current coroutine |
428 | "Cede" to other coros. This function puts the current coro into the |
205 | into the ready queue and calls "schedule", which has the effect of |
429 | ready queue and calls "schedule", which has the effect of giving up |
206 | giving up the current "timeslice" to other coroutines of the same or |
430 | the current "timeslice" to other coros of the same or higher |
207 | higher priority. Once your coroutine gets its turn again it will |
431 | priority. Once your coro gets its turn again it will automatically |
208 | automatically be resumed. |
432 | be resumed. |
209 | |
433 | |
210 | This function is often called "yield" in other languages. |
434 | This function is often called "yield" in other languages. |
211 | |
435 | |
212 | Coro::cede_notself |
436 | Coro::cede_notself |
213 | Works like cede, but is not exported by default and will cede to |
437 | Works like cede, but is not exported by default and will cede to |
214 | *any* coroutine, regardless of priority. This is useful sometimes to |
438 | *any* coro, regardless of priority. This is useful sometimes to |
215 | ensure progress is made. |
439 | ensure progress is made. |
216 | |
440 | |
217 | terminate [arg...] |
441 | terminate [arg...] |
218 | Terminates the current coroutine with the given status values (see |
442 | Terminates the current coro with the given status values (see |
219 | cancel). |
443 | cancel). |
220 | |
444 | |
|
|
445 | Coro::on_enter BLOCK, Coro::on_leave BLOCK |
|
|
446 | These function install enter and leave winders in the current scope. |
|
|
447 | The enter block will be executed when on_enter is called and |
|
|
448 | whenever the current coro is re-entered by the scheduler, while the |
|
|
449 | leave block is executed whenever the current coro is blocked by the |
|
|
450 | scheduler, and also when the containing scope is exited (by whatever |
|
|
451 | means, be it exit, die, last etc.). |
|
|
452 | |
|
|
453 | *Neither invoking the scheduler, nor exceptions, are allowed within |
|
|
454 | those BLOCKs*. That means: do not even think about calling "die" |
|
|
455 | without an eval, and do not even think of entering the scheduler in |
|
|
456 | any way. |
|
|
457 | |
|
|
458 | Since both BLOCKs are tied to the current scope, they will |
|
|
459 | automatically be removed when the current scope exits. |
|
|
460 | |
|
|
461 | These functions implement the same concept as "dynamic-wind" in |
|
|
462 | scheme does, and are useful when you want to localise some resource |
|
|
463 | to a specific coro. |
|
|
464 | |
|
|
465 | They slow down thread switching considerably for coros that use them |
|
|
466 | (about 40% for a BLOCK with a single assignment, so thread switching |
|
|
467 | is still reasonably fast if the handlers are fast). |
|
|
468 | |
|
|
469 | These functions are best understood by an example: The following |
|
|
470 | function will change the current timezone to |
|
|
471 | "Antarctica/South_Pole", which requires a call to "tzset", but by |
|
|
472 | using "on_enter" and "on_leave", which remember/change the current |
|
|
473 | timezone and restore the previous value, respectively, the timezone |
|
|
474 | is only changed for the coro that installed those handlers. |
|
|
475 | |
|
|
476 | use POSIX qw(tzset); |
|
|
477 | |
|
|
478 | async { |
|
|
479 | my $old_tz; # store outside TZ value here |
|
|
480 | |
|
|
481 | Coro::on_enter { |
|
|
482 | $old_tz = $ENV{TZ}; # remember the old value |
|
|
483 | |
|
|
484 | $ENV{TZ} = "Antarctica/South_Pole"; |
|
|
485 | tzset; # enable new value |
|
|
486 | }; |
|
|
487 | |
|
|
488 | Coro::on_leave { |
|
|
489 | $ENV{TZ} = $old_tz; |
|
|
490 | tzset; # restore old value |
|
|
491 | }; |
|
|
492 | |
|
|
493 | # at this place, the timezone is Antarctica/South_Pole, |
|
|
494 | # without disturbing the TZ of any other coro. |
|
|
495 | }; |
|
|
496 | |
|
|
497 | This can be used to localise about any resource (locale, uid, |
|
|
498 | current working directory etc.) to a block, despite the existance of |
|
|
499 | other coros. |
|
|
500 | |
|
|
501 | Another interesting example implements time-sliced multitasking |
|
|
502 | using interval timers (this could obviously be optimised, but does |
|
|
503 | the job): |
|
|
504 | |
|
|
505 | # "timeslice" the given block |
|
|
506 | sub timeslice(&) { |
|
|
507 | use Time::HiRes (); |
|
|
508 | |
|
|
509 | Coro::on_enter { |
|
|
510 | # on entering the thread, we set an VTALRM handler to cede |
|
|
511 | $SIG{VTALRM} = sub { cede }; |
|
|
512 | # and then start the interval timer |
|
|
513 | Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0.01, 0.01; |
|
|
514 | }; |
|
|
515 | Coro::on_leave { |
|
|
516 | # on leaving the thread, we stop the interval timer again |
|
|
517 | Time::HiRes::setitimer &Time::HiRes::ITIMER_VIRTUAL, 0, 0; |
|
|
518 | }; |
|
|
519 | |
|
|
520 | &{+shift}; |
|
|
521 | } |
|
|
522 | |
|
|
523 | # use like this: |
|
|
524 | timeslice { |
|
|
525 | # The following is an endless loop that would normally |
|
|
526 | # monopolise the process. Since it runs in a timesliced |
|
|
527 | # environment, it will regularly cede to other threads. |
|
|
528 | while () { } |
|
|
529 | }; |
|
|
530 | |
221 | killall |
531 | killall |
222 | Kills/terminates/cancels all coroutines except the currently running |
532 | Kills/terminates/cancels all coros except the currently running one. |
223 | one. This is useful after a fork, either in the child or the parent, |
|
|
224 | as usually only one of them should inherit the running coroutines. |
|
|
225 | |
533 | |
226 | Note that while this will try to free some of the main programs |
534 | Note that while this will try to free some of the main interpreter |
227 | resources, you cannot free all of them, so if a coroutine that is |
535 | resources if the calling coro isn't the main coro, but one cannot |
228 | not the main program calls this function, there will be some |
536 | free all of them, so if a coro that is not the main coro calls this |
229 | one-time resource leak. |
537 | function, there will be some one-time resource leak. |
230 | |
538 | |
231 | COROUTINE OBJECT METHODS |
539 | CORO OBJECT METHODS |
232 | These are the methods you can call on coroutine objects (or to create |
540 | These are the methods you can call on coro objects (or to create them). |
233 | them). |
|
|
234 | |
541 | |
235 | new Coro \&sub [, @args...] |
542 | new Coro \&sub [, @args...] |
236 | Create a new coroutine and return it. When the sub returns, the |
543 | Create a new coro and return it. When the sub returns, the coro |
237 | coroutine automatically terminates as if "terminate" with the |
544 | automatically terminates as if "terminate" with the returned values |
238 | returned values were called. To make the coroutine run you must |
545 | were called. To make the coro run you must first put it into the |
239 | first put it into the ready queue by calling the ready method. |
546 | ready queue by calling the ready method. |
240 | |
547 | |
241 | See "async" and "Coro::State::new" for additional info about the |
548 | See "async" and "Coro::State::new" for additional info about the |
242 | coroutine environment. |
549 | coro environment. |
243 | |
550 | |
244 | $success = $coroutine->ready |
551 | $success = $coro->ready |
245 | Put the given coroutine into the end of its ready queue (there is |
552 | Put the given coro into the end of its ready queue (there is one |
246 | one queue for each priority) and return true. If the coroutine is |
553 | queue for each priority) and return true. If the coro is already in |
247 | already in the ready queue, do nothing and return false. |
554 | the ready queue, do nothing and return false. |
248 | |
555 | |
249 | This ensures that the scheduler will resume this coroutine |
556 | This ensures that the scheduler will resume this coro automatically |
250 | automatically once all the coroutines of higher priority and all |
557 | once all the coro of higher priority and all coro of the same |
251 | coroutines of the same priority that were put into the ready queue |
558 | priority that were put into the ready queue earlier have been |
252 | earlier have been resumed. |
559 | resumed. |
253 | |
560 | |
|
|
561 | $coro->suspend |
|
|
562 | Suspends the specified coro. A suspended coro works just like any |
|
|
563 | other coro, except that the scheduler will not select a suspended |
|
|
564 | coro for execution. |
|
|
565 | |
|
|
566 | Suspending a coro can be useful when you want to keep the coro from |
|
|
567 | running, but you don't want to destroy it, or when you want to |
|
|
568 | temporarily freeze a coro (e.g. for debugging) to resume it later. |
|
|
569 | |
|
|
570 | A scenario for the former would be to suspend all (other) coros |
|
|
571 | after a fork and keep them alive, so their destructors aren't |
|
|
572 | called, but new coros can be created. |
|
|
573 | |
|
|
574 | $coro->resume |
|
|
575 | If the specified coro was suspended, it will be resumed. Note that |
|
|
576 | when the coro was in the ready queue when it was suspended, it might |
|
|
577 | have been unreadied by the scheduler, so an activation might have |
|
|
578 | been lost. |
|
|
579 | |
|
|
580 | To avoid this, it is best to put a suspended coro into the ready |
|
|
581 | queue unconditionally, as every synchronisation mechanism must |
|
|
582 | protect itself against spurious wakeups, and the one in the Coro |
|
|
583 | family certainly do that. |
|
|
584 | |
254 | $is_ready = $coroutine->is_ready |
585 | $is_ready = $coro->is_ready |
255 | Return whether the coroutine is currently the ready queue or not, |
586 | Returns true iff the Coro object is in the ready queue. Unless the |
|
|
587 | Coro object gets destroyed, it will eventually be scheduled by the |
|
|
588 | scheduler. |
256 | |
589 | |
|
|
590 | $is_running = $coro->is_running |
|
|
591 | Returns true iff the Coro object is currently running. Only one Coro |
|
|
592 | object can ever be in the running state (but it currently is |
|
|
593 | possible to have multiple running Coro::States). |
|
|
594 | |
|
|
595 | $is_suspended = $coro->is_suspended |
|
|
596 | Returns true iff this Coro object has been suspended. Suspended |
|
|
597 | Coros will not ever be scheduled. |
|
|
598 | |
257 | $coroutine->cancel (arg...) |
599 | $coro->cancel (arg...) |
258 | Terminates the given coroutine and makes it return the given |
600 | Terminates the given Coro and makes it return the given arguments as |
259 | arguments as status (default: the empty list). Never returns if the |
601 | status (default: the empty list). Never returns if the Coro is the |
260 | coroutine is the current coroutine. |
602 | current Coro. |
261 | |
603 | |
262 | $coroutine->schedule_to |
604 | $coro->schedule_to |
263 | Puts the current coroutine to sleep (like "Coro::schedule"), but |
605 | Puts the current coro to sleep (like "Coro::schedule"), but instead |
264 | instead of continuing with the next coro from the ready queue, |
606 | of continuing with the next coro from the ready queue, always switch |
265 | always switch to the given coroutine object (regardless of priority |
607 | to the given coro object (regardless of priority etc.). The |
266 | etc.). The readyness state of that coroutine isn't changed. |
608 | readyness state of that coro isn't changed. |
267 | |
609 | |
268 | This is an advanced method for special cases - I'd love to hear |
610 | This is an advanced method for special cases - I'd love to hear |
269 | about any uses for this one. |
611 | about any uses for this one. |
270 | |
612 | |
271 | $coroutine->cede_to |
613 | $coro->cede_to |
272 | Like "schedule_to", but puts the current coroutine into the ready |
614 | Like "schedule_to", but puts the current coro into the ready queue. |
273 | queue. This has the effect of temporarily switching to the given |
615 | This has the effect of temporarily switching to the given coro, and |
274 | coroutine, and continuing some time later. |
616 | continuing some time later. |
275 | |
617 | |
276 | This is an advanced method for special cases - I'd love to hear |
618 | This is an advanced method for special cases - I'd love to hear |
277 | about any uses for this one. |
619 | about any uses for this one. |
278 | |
620 | |
279 | $coroutine->throw ([$scalar]) |
621 | $coro->throw ([$scalar]) |
280 | If $throw is specified and defined, it will be thrown as an |
622 | If $throw is specified and defined, it will be thrown as an |
281 | exception inside the coroutine at the next convenient point in time. |
623 | exception inside the coro at the next convenient point in time. |
282 | Otherwise clears the exception object. |
624 | Otherwise clears the exception object. |
283 | |
625 | |
284 | Coro will check for the exception each time a schedule-like-function |
626 | Coro will check for the exception each time a schedule-like-function |
285 | returns, i.e. after each "schedule", "cede", |
627 | returns, i.e. after each "schedule", "cede", |
286 | "Coro::Semaphore->down", "Coro::Handle->readable" and so on. Most of |
628 | "Coro::Semaphore->down", "Coro::Handle->readable" and so on. Most of |
… | |
… | |
289 | |
631 | |
290 | The exception object will be thrown "as is" with the specified |
632 | The exception object will be thrown "as is" with the specified |
291 | scalar in $@, i.e. if it is a string, no line number or newline will |
633 | scalar in $@, i.e. if it is a string, no line number or newline will |
292 | be appended (unlike with "die"). |
634 | be appended (unlike with "die"). |
293 | |
635 | |
294 | This can be used as a softer means than "cancel" to ask a coroutine |
636 | This can be used as a softer means than "cancel" to ask a coro to |
295 | to end itself, although there is no guarantee that the exception |
637 | end itself, although there is no guarantee that the exception will |
296 | will lead to termination, and if the exception isn't caught it might |
638 | lead to termination, and if the exception isn't caught it might well |
297 | well end the whole program. |
639 | end the whole program. |
298 | |
640 | |
299 | You might also think of "throw" as being the moral equivalent of |
641 | You might also think of "throw" as being the moral equivalent of |
300 | "kill"ing a coroutine with a signal (in this case, a scalar). |
642 | "kill"ing a coro with a signal (in this case, a scalar). |
301 | |
643 | |
302 | $coroutine->join |
644 | $coro->join |
303 | Wait until the coroutine terminates and return any values given to |
645 | Wait until the coro terminates and return any values given to the |
304 | the "terminate" or "cancel" functions. "join" can be called |
646 | "terminate" or "cancel" functions. "join" can be called concurrently |
305 | concurrently from multiple coroutines, and all will be resumed and |
647 | from multiple coro, and all will be resumed and given the status |
306 | given the status return once the $coroutine terminates. |
648 | return once the $coro terminates. |
307 | |
649 | |
308 | $coroutine->on_destroy (\&cb) |
650 | $coro->on_destroy (\&cb) |
309 | Registers a callback that is called when this coroutine gets |
651 | Registers a callback that is called when this coro thread gets |
310 | destroyed, but before it is joined. The callback gets passed the |
652 | destroyed, but before it is joined. The callback gets passed the |
311 | terminate arguments, if any, and *must not* die, under any |
653 | terminate arguments, if any, and *must not* die, under any |
312 | circumstances. |
654 | circumstances. |
313 | |
655 | |
|
|
656 | There can be any number of "on_destroy" callbacks per coro. |
|
|
657 | |
314 | $oldprio = $coroutine->prio ($newprio) |
658 | $oldprio = $coro->prio ($newprio) |
315 | Sets (or gets, if the argument is missing) the priority of the |
659 | Sets (or gets, if the argument is missing) the priority of the coro |
316 | coroutine. Higher priority coroutines get run before lower priority |
660 | thread. Higher priority coro get run before lower priority coros. |
317 | coroutines. Priorities are small signed integers (currently -4 .. |
661 | Priorities are small signed integers (currently -4 .. +3), that you |
318 | +3), that you can refer to using PRIO_xxx constants (use the import |
662 | can refer to using PRIO_xxx constants (use the import tag :prio to |
319 | tag :prio to get then): |
663 | get then): |
320 | |
664 | |
321 | PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN |
665 | PRIO_MAX > PRIO_HIGH > PRIO_NORMAL > PRIO_LOW > PRIO_IDLE > PRIO_MIN |
322 | 3 > 1 > 0 > -1 > -3 > -4 |
666 | 3 > 1 > 0 > -1 > -3 > -4 |
323 | |
667 | |
324 | # set priority to HIGH |
668 | # set priority to HIGH |
325 | current->prio(PRIO_HIGH); |
669 | current->prio (PRIO_HIGH); |
326 | |
670 | |
327 | The idle coroutine ($Coro::idle) always has a lower priority than |
671 | The idle coro thread ($Coro::idle) always has a lower priority than |
328 | any existing coroutine. |
672 | any existing coro. |
329 | |
673 | |
330 | Changing the priority of the current coroutine will take effect |
674 | Changing the priority of the current coro will take effect |
331 | immediately, but changing the priority of coroutines in the ready |
675 | immediately, but changing the priority of a coro in the ready queue |
332 | queue (but not running) will only take effect after the next |
676 | (but not running) will only take effect after the next schedule (of |
333 | schedule (of that coroutine). This is a bug that will be fixed in |
677 | that coro). This is a bug that will be fixed in some future version. |
334 | some future version. |
|
|
335 | |
678 | |
336 | $newprio = $coroutine->nice ($change) |
679 | $newprio = $coro->nice ($change) |
337 | Similar to "prio", but subtract the given value from the priority |
680 | Similar to "prio", but subtract the given value from the priority |
338 | (i.e. higher values mean lower priority, just as in unix). |
681 | (i.e. higher values mean lower priority, just as in UNIX's nice |
|
|
682 | command). |
339 | |
683 | |
340 | $olddesc = $coroutine->desc ($newdesc) |
684 | $olddesc = $coro->desc ($newdesc) |
341 | Sets (or gets in case the argument is missing) the description for |
685 | Sets (or gets in case the argument is missing) the description for |
342 | this coroutine. This is just a free-form string you can associate |
686 | this coro thread. This is just a free-form string you can associate |
343 | with a coroutine. |
687 | with a coro. |
344 | |
688 | |
345 | This method simply sets the "$coroutine->{desc}" member to the given |
689 | This method simply sets the "$coro->{desc}" member to the given |
346 | string. You can modify this member directly if you wish. |
690 | string. You can modify this member directly if you wish, and in |
|
|
691 | fact, this is often preferred to indicate major processing states |
|
|
692 | that cna then be seen for example in a Coro::Debug session: |
|
|
693 | |
|
|
694 | sub my_long_function { |
|
|
695 | local $Coro::current->{desc} = "now in my_long_function"; |
|
|
696 | ... |
|
|
697 | $Coro::current->{desc} = "my_long_function: phase 1"; |
|
|
698 | ... |
|
|
699 | $Coro::current->{desc} = "my_long_function: phase 2"; |
|
|
700 | ... |
|
|
701 | } |
347 | |
702 | |
348 | GLOBAL FUNCTIONS |
703 | GLOBAL FUNCTIONS |
349 | Coro::nready |
704 | Coro::nready |
350 | Returns the number of coroutines that are currently in the ready |
705 | Returns the number of coro that are currently in the ready state, |
351 | state, i.e. that can be switched to by calling "schedule" directory |
706 | i.e. that can be switched to by calling "schedule" directory or |
352 | or indirectly. The value 0 means that the only runnable coroutine is |
707 | indirectly. The value 0 means that the only runnable coro is the |
353 | the currently running one, so "cede" would have no effect, and |
708 | currently running one, so "cede" would have no effect, and |
354 | "schedule" would cause a deadlock unless there is an idle handler |
709 | "schedule" would cause a deadlock unless there is an idle handler |
355 | that wakes up some coroutines. |
710 | that wakes up some coro. |
356 | |
711 | |
357 | my $guard = Coro::guard { ... } |
712 | my $guard = Coro::guard { ... } |
358 | This creates and returns a guard object. Nothing happens until the |
713 | This function still exists, but is deprecated. Please use the |
359 | object gets destroyed, in which case the codeblock given as argument |
714 | "Guard::guard" function instead. |
360 | will be executed. This is useful to free locks or other resources in |
|
|
361 | case of a runtime error or when the coroutine gets canceled, as in |
|
|
362 | both cases the guard block will be executed. The guard object |
|
|
363 | supports only one method, "->cancel", which will keep the codeblock |
|
|
364 | from being executed. |
|
|
365 | |
|
|
366 | Example: set some flag and clear it again when the coroutine gets |
|
|
367 | canceled or the function returns: |
|
|
368 | |
|
|
369 | sub do_something { |
|
|
370 | my $guard = Coro::guard { $busy = 0 }; |
|
|
371 | $busy = 1; |
|
|
372 | |
|
|
373 | # do something that requires $busy to be true |
|
|
374 | } |
|
|
375 | |
715 | |
376 | unblock_sub { ... } |
716 | unblock_sub { ... } |
377 | This utility function takes a BLOCK or code reference and "unblocks" |
717 | This utility function takes a BLOCK or code reference and "unblocks" |
378 | it, returning a new coderef. Unblocking means that calling the new |
718 | it, returning a new coderef. Unblocking means that calling the new |
379 | coderef will return immediately without blocking, returning nothing, |
719 | coderef will return immediately without blocking, returning nothing, |
380 | while the original code ref will be called (with parameters) from |
720 | while the original code ref will be called (with parameters) from |
381 | within another coroutine. |
721 | within another coro. |
382 | |
722 | |
383 | The reason this function exists is that many event libraries (such |
723 | The reason this function exists is that many event libraries (such |
384 | as the venerable Event module) are not coroutine-safe (a weaker form |
724 | as the venerable Event module) are not thread-safe (a weaker form of |
385 | of reentrancy). This means you must not block within event |
725 | reentrancy). This means you must not block within event callbacks, |
386 | callbacks, otherwise you might suffer from crashes or worse. The |
726 | otherwise you might suffer from crashes or worse. The only event |
387 | only event library currently known that is safe to use without |
727 | library currently known that is safe to use without "unblock_sub" is |
388 | "unblock_sub" is EV. |
728 | EV (but you might still run into deadlocks if all event loops are |
|
|
729 | blocked). |
|
|
730 | |
|
|
731 | Coro will try to catch you when you block in the event loop |
|
|
732 | ("FATAL:$Coro::IDLE blocked itself"), but this is just best effort |
|
|
733 | and only works when you do not run your own event loop. |
389 | |
734 | |
390 | This function allows your callbacks to block by executing them in |
735 | This function allows your callbacks to block by executing them in |
391 | another coroutine where it is safe to block. One example where |
736 | another coro where it is safe to block. One example where blocking |
392 | blocking is handy is when you use the Coro::AIO functions to save |
737 | is handy is when you use the Coro::AIO functions to save results to |
393 | results to disk, for example. |
738 | disk, for example. |
394 | |
739 | |
395 | In short: simply use "unblock_sub { ... }" instead of "sub { ... }" |
740 | In short: simply use "unblock_sub { ... }" instead of "sub { ... }" |
396 | when creating event callbacks that want to block. |
741 | when creating event callbacks that want to block. |
397 | |
742 | |
398 | If your handler does not plan to block (e.g. simply sends a message |
743 | If your handler does not plan to block (e.g. simply sends a message |
399 | to another coroutine, or puts some other coroutine into the ready |
744 | to another coro, or puts some other coro into the ready queue), |
400 | queue), there is no reason to use "unblock_sub". |
745 | there is no reason to use "unblock_sub". |
401 | |
746 | |
402 | Note that you also need to use "unblock_sub" for any other callbacks |
747 | Note that you also need to use "unblock_sub" for any other callbacks |
403 | that are indirectly executed by any C-based event loop. For example, |
748 | that are indirectly executed by any C-based event loop. For example, |
404 | when you use a module that uses AnyEvent (and you use |
749 | when you use a module that uses AnyEvent (and you use |
405 | Coro::AnyEvent) and it provides callbacks that are the result of |
750 | Coro::AnyEvent) and it provides callbacks that are the result of |
406 | some event callback, then you must not block either, or use |
751 | some event callback, then you must not block either, or use |
407 | "unblock_sub". |
752 | "unblock_sub". |
408 | |
753 | |
409 | $cb = Coro::rouse_cb |
754 | $cb = rouse_cb |
410 | Create and return a "rouse callback". That's a code reference that, |
755 | Create and return a "rouse callback". That's a code reference that, |
411 | when called, will remember a copy of its arguments and notify the |
756 | when called, will remember a copy of its arguments and notify the |
412 | owner coroutine of the callback. |
757 | owner coro of the callback. |
413 | |
758 | |
414 | See the next function. |
759 | See the next function. |
415 | |
760 | |
416 | @args = Coro::rouse_wait [$cb] |
761 | @args = rouse_wait [$cb] |
417 | Wait for the specified rouse callback (or the last one that was |
762 | Wait for the specified rouse callback (or the last one that was |
418 | created in this coroutine). |
763 | created in this coro). |
419 | |
764 | |
420 | As soon as the callback is invoked (or when the callback was invoked |
765 | As soon as the callback is invoked (or when the callback was invoked |
421 | before "rouse_wait"), it will return the arguments originally passed |
766 | before "rouse_wait"), it will return the arguments originally passed |
422 | to the rouse callback. |
767 | to the rouse callback. In scalar context, that means you get the |
|
|
768 | *last* argument, just as if "rouse_wait" had a "return ($a1, $a2, |
|
|
769 | $a3...)" statement at the end. |
423 | |
770 | |
424 | See the section HOW TO WAIT FOR A CALLBACK for an actual usage |
771 | See the section HOW TO WAIT FOR A CALLBACK for an actual usage |
425 | example. |
772 | example. |
426 | |
773 | |
427 | HOW TO WAIT FOR A CALLBACK |
774 | HOW TO WAIT FOR A CALLBACK |
428 | It is very common for a coroutine to wait for some callback to be |
775 | It is very common for a coro to wait for some callback to be called. |
429 | called. This occurs naturally when you use coroutines in an otherwise |
776 | This occurs naturally when you use coro in an otherwise event-based |
430 | event-based program, or when you use event-based libraries. |
777 | program, or when you use event-based libraries. |
431 | |
778 | |
432 | These typically register a callback for some event, and call that |
779 | These typically register a callback for some event, and call that |
433 | callback when the event occured. In a coroutine, however, you typically |
780 | callback when the event occured. In a coro, however, you typically want |
434 | want to just wait for the event, simplyifying things. |
781 | to just wait for the event, simplyifying things. |
435 | |
782 | |
436 | For example "AnyEvent->child" registers a callback to be called when a |
783 | For example "AnyEvent->child" registers a callback to be called when a |
437 | specific child has exited: |
784 | specific child has exited: |
438 | |
785 | |
439 | my $child_watcher = AnyEvent->child (pid => $pid, cb => sub { ... }); |
786 | my $child_watcher = AnyEvent->child (pid => $pid, cb => sub { ... }); |
440 | |
787 | |
441 | But from withina coroutine, you often just want to write this: |
788 | But from within a coro, you often just want to write this: |
442 | |
789 | |
443 | my $status = wait_for_child $pid; |
790 | my $status = wait_for_child $pid; |
444 | |
791 | |
445 | Coro offers two functions specifically designed to make this easy, |
792 | Coro offers two functions specifically designed to make this easy, |
446 | "Coro::rouse_cb" and "Coro::rouse_wait". |
793 | "Coro::rouse_cb" and "Coro::rouse_wait". |
447 | |
794 | |
448 | The first function, "rouse_cb", generates and returns a callback that, |
795 | The first function, "rouse_cb", generates and returns a callback that, |
449 | when invoked, will save its arguments and notify the coroutine that |
796 | when invoked, will save its arguments and notify the coro that created |
450 | created the callback. |
797 | the callback. |
451 | |
798 | |
452 | The second function, "rouse_wait", waits for the callback to be called |
799 | The second function, "rouse_wait", waits for the callback to be called |
453 | (by calling "schedule" to go to sleep) and returns the arguments |
800 | (by calling "schedule" to go to sleep) and returns the arguments |
454 | originally passed to the callback. |
801 | originally passed to the callback. |
455 | |
802 | |
… | |
… | |
469 | you can roll your own, using "schedule": |
816 | you can roll your own, using "schedule": |
470 | |
817 | |
471 | sub wait_for_child($) { |
818 | sub wait_for_child($) { |
472 | my ($pid) = @_; |
819 | my ($pid) = @_; |
473 | |
820 | |
474 | # store the current coroutine in $current, |
821 | # store the current coro in $current, |
475 | # and provide result variables for the closure passed to ->child |
822 | # and provide result variables for the closure passed to ->child |
476 | my $current = $Coro::current; |
823 | my $current = $Coro::current; |
477 | my ($done, $rstatus); |
824 | my ($done, $rstatus); |
478 | |
825 | |
479 | # pass a closure to ->child |
826 | # pass a closure to ->child |
… | |
… | |
490 | |
837 | |
491 | BUGS/LIMITATIONS |
838 | BUGS/LIMITATIONS |
492 | fork with pthread backend |
839 | fork with pthread backend |
493 | When Coro is compiled using the pthread backend (which isn't |
840 | When Coro is compiled using the pthread backend (which isn't |
494 | recommended but required on many BSDs as their libcs are completely |
841 | recommended but required on many BSDs as their libcs are completely |
495 | broken), then coroutines will not survive a fork. There is no known |
842 | broken), then coro will not survive a fork. There is no known |
496 | workaround except to fix your libc and use a saner backend. |
843 | workaround except to fix your libc and use a saner backend. |
497 | |
844 | |
498 | perl process emulation ("threads") |
845 | perl process emulation ("threads") |
499 | This module is not perl-pseudo-thread-safe. You should only ever use |
846 | This module is not perl-pseudo-thread-safe. You should only ever use |
500 | this module from the first thread (this requirement might be removed |
847 | this module from the first thread (this requirement might be removed |
501 | in the future to allow per-thread schedulers, but Coro::State does |
848 | in the future to allow per-thread schedulers, but Coro::State does |
502 | not yet allow this). I recommend disabling thread support and using |
849 | not yet allow this). I recommend disabling thread support and using |
503 | processes, as having the windows process emulation enabled under |
850 | processes, as having the windows process emulation enabled under |
504 | unix roughly halves perl performance, even when not used. |
851 | unix roughly halves perl performance, even when not used. |
505 | |
852 | |
506 | coroutine switching not signal safe |
853 | coro switching is not signal safe |
507 | You must not switch to another coroutine from within a signal |
854 | You must not switch to another coro from within a signal handler |
508 | handler (only relevant with %SIG - most event libraries provide safe |
855 | (only relevant with %SIG - most event libraries provide safe |
509 | signals). |
856 | signals), *unless* you are sure you are not interrupting a Coro |
|
|
857 | function. |
510 | |
858 | |
511 | That means you *MUST NOT* call any function that might "block" the |
859 | That means you *MUST NOT* call any function that might "block" the |
512 | current coroutine - "cede", "schedule" "Coro::Semaphore->down" or |
860 | current coro - "cede", "schedule" "Coro::Semaphore->down" or |
513 | anything that calls those. Everything else, including calling |
861 | anything that calls those. Everything else, including calling |
514 | "ready", works. |
862 | "ready", works. |
|
|
863 | |
|
|
864 | WINDOWS PROCESS EMULATION |
|
|
865 | A great many people seem to be confused about ithreads (for example, |
|
|
866 | Chip Salzenberg called me unintelligent, incapable, stupid and gullible, |
|
|
867 | while in the same mail making rather confused statements about perl |
|
|
868 | ithreads (for example, that memory or files would be shared), showing |
|
|
869 | his lack of understanding of this area - if it is hard to understand for |
|
|
870 | Chip, it is probably not obvious to everybody). |
|
|
871 | |
|
|
872 | What follows is an ultra-condensed version of my talk about threads in |
|
|
873 | scripting languages given on the perl workshop 2009: |
|
|
874 | |
|
|
875 | The so-called "ithreads" were originally implemented for two reasons: |
|
|
876 | first, to (badly) emulate unix processes on native win32 perls, and |
|
|
877 | secondly, to replace the older, real thread model ("5.005-threads"). |
|
|
878 | |
|
|
879 | It does that by using threads instead of OS processes. The difference |
|
|
880 | between processes and threads is that threads share memory (and other |
|
|
881 | state, such as files) between threads within a single process, while |
|
|
882 | processes do not share anything (at least not semantically). That means |
|
|
883 | that modifications done by one thread are seen by others, while |
|
|
884 | modifications by one process are not seen by other processes. |
|
|
885 | |
|
|
886 | The "ithreads" work exactly like that: when creating a new ithreads |
|
|
887 | process, all state is copied (memory is copied physically, files and |
|
|
888 | code is copied logically). Afterwards, it isolates all modifications. On |
|
|
889 | UNIX, the same behaviour can be achieved by using operating system |
|
|
890 | processes, except that UNIX typically uses hardware built into the |
|
|
891 | system to do this efficiently, while the windows process emulation |
|
|
892 | emulates this hardware in software (rather efficiently, but of course it |
|
|
893 | is still much slower than dedicated hardware). |
|
|
894 | |
|
|
895 | As mentioned before, loading code, modifying code, modifying data |
|
|
896 | structures and so on is only visible in the ithreads process doing the |
|
|
897 | modification, not in other ithread processes within the same OS process. |
|
|
898 | |
|
|
899 | This is why "ithreads" do not implement threads for perl at all, only |
|
|
900 | processes. What makes it so bad is that on non-windows platforms, you |
|
|
901 | can actually take advantage of custom hardware for this purpose (as |
|
|
902 | evidenced by the forks module, which gives you the (i-) threads API, |
|
|
903 | just much faster). |
|
|
904 | |
|
|
905 | Sharing data is in the i-threads model is done by transfering data |
|
|
906 | structures between threads using copying semantics, which is very slow - |
|
|
907 | shared data simply does not exist. Benchmarks using i-threads which are |
|
|
908 | communication-intensive show extremely bad behaviour with i-threads (in |
|
|
909 | fact, so bad that Coro, which cannot take direct advantage of multiple |
|
|
910 | CPUs, is often orders of magnitude faster because it shares data using |
|
|
911 | real threads, refer to my talk for details). |
|
|
912 | |
|
|
913 | As summary, i-threads *use* threads to implement processes, while the |
|
|
914 | compatible forks module *uses* processes to emulate, uhm, processes. |
|
|
915 | I-threads slow down every perl program when enabled, and outside of |
|
|
916 | windows, serve no (or little) practical purpose, but disadvantages every |
|
|
917 | single-threaded Perl program. |
|
|
918 | |
|
|
919 | This is the reason that I try to avoid the name "ithreads", as it is |
|
|
920 | misleading as it implies that it implements some kind of thread model |
|
|
921 | for perl, and prefer the name "windows process emulation", which |
|
|
922 | describes the actual use and behaviour of it much better. |
515 | |
923 | |
516 | SEE ALSO |
924 | SEE ALSO |
517 | Event-Loop integration: Coro::AnyEvent, Coro::EV, Coro::Event. |
925 | Event-Loop integration: Coro::AnyEvent, Coro::EV, Coro::Event. |
518 | |
926 | |
519 | Debugging: Coro::Debug. |
927 | Debugging: Coro::Debug. |