Commit | Line | Data |
---|---|---|
07d83abe MV |
1 | @c -*-texinfo-*- |
2 | @c This is part of the GNU Guile Reference Manual. | |
3 | @c Copyright (C) 1996, 1997, 2000, 2001, 2002, 2003, 2004 | |
4 | @c Free Software Foundation, Inc. | |
5 | @c See the file guile.texi for copying conditions. | |
6 | ||
7 | @page | |
8 | @node Scheduling | |
9 | @section Threads, Mutexes, Asyncs and Dynamic Roots | |
10 | ||
11 | [FIXME: This is pasted in from Tom Lord's original guile.texi chapter | |
12 | plus the Cygnus programmer's manual; it should be *very* carefully | |
13 | reviewed and largely reorganized.] | |
14 | ||
15 | @menu | |
16 | * Arbiters:: Synchronization primitives. | |
17 | * Asyncs:: Asynchronous procedure invocation. | |
18 | * Dynamic Roots:: Root frames of execution. | |
19 | * Threads:: Multiple threads of execution. | |
20 | * Fluids:: Thread-local variables. | |
21 | * Futures:: Delayed execution in new threads. | |
22 | * Parallel Forms:: Parallel execution of forms. | |
23 | @end menu | |
24 | ||
25 | ||
26 | @node Arbiters | |
27 | @subsection Arbiters | |
07d83abe MV |
28 | @cindex arbiters |
29 | ||
e136aab0 KR |
30 | Arbiters are synchronization objects, they can be used by threads to |
31 | control access to a shared resource. An arbiter can be locked to | |
32 | indicate a resource is in use, and unlocked when done. | |
07d83abe | 33 | |
e136aab0 KR |
34 | An arbiter is like a light-weight mutex (@pxref{Low level thread |
35 | primitives}). It uses less memory and may be a little faster, but | |
36 | there's no way for a thread to block waiting on an arbiter, it can | |
37 | only test and get the status returned. | |
07d83abe MV |
38 | |
39 | @deffn {Scheme Procedure} make-arbiter name | |
40 | @deffnx {C Function} scm_make_arbiter (name) | |
e136aab0 KR |
41 | Return an arbiter object, initially unlocked. Currently @var{name} is |
42 | only used for diagnostic output. | |
07d83abe MV |
43 | @end deffn |
44 | ||
45 | @deffn {Scheme Procedure} try-arbiter arb | |
46 | @deffnx {C Function} scm_try_arbiter (arb) | |
e136aab0 KR |
47 | If @var{arb} is unlocked, then lock it and return @code{#t}. If |
48 | @var{arb} is already locked, then do nothing and return @code{#f}. | |
07d83abe MV |
49 | @end deffn |
50 | ||
51 | @deffn {Scheme Procedure} release-arbiter arb | |
52 | @deffnx {C Function} scm_release_arbiter (arb) | |
e136aab0 KR |
53 | If @var{arb} is locked, then unlock it and return @code{#t}. If |
54 | @var{arb} is already unlocked, then do nothing and return @code{#f}. | |
55 | ||
56 | Typical usage is for the thread which locked an arbiter to later | |
57 | release it, but that's not required, any thread can release it. | |
07d83abe MV |
58 | @end deffn |
59 | ||
60 | ||
61 | @node Asyncs | |
62 | @subsection Asyncs | |
63 | ||
64 | @cindex asyncs | |
65 | @cindex user asyncs | |
66 | @cindex system asyncs | |
67 | ||
68 | Asyncs are a means of deferring the excution of Scheme code until it is | |
69 | safe to do so. | |
70 | ||
71 | Guile provides two kinds of asyncs that share the basic concept but are | |
72 | otherwise quite different: system asyncs and user asyncs. System asyncs | |
73 | are integrated into the core of Guile and are executed automatically | |
74 | when the system is in a state to allow the execution of Scheme code. | |
75 | For example, it is not possible to execute Scheme code in a POSIX signal | |
76 | handler, but such a signal handler can queue a system async to be | |
77 | executed in the near future, when it is safe to do so. | |
78 | ||
79 | System asyncs can also be queued for threads other than the current one. | |
80 | This way, you can cause threads to asynchronously execute arbitrary | |
81 | code. | |
82 | ||
83 | User asyncs offer a convenient means of queueing procedures for future | |
84 | execution and triggering this execution. They will not be executed | |
85 | automatically. | |
86 | ||
87 | @menu | |
88 | * System asyncs:: | |
89 | * User asyncs:: | |
90 | @end menu | |
91 | ||
92 | @node System asyncs | |
93 | @subsubsection System asyncs | |
94 | ||
95 | To cause the future asynchronous execution of a procedure in a given | |
96 | thread, use @code{system-async-mark}. | |
97 | ||
98 | Automatic invocation of system asyncs can be temporarily disabled by | |
99 | calling @code{call-with-blocked-asyncs}. This function works by | |
100 | temporarily increasing the @emph{async blocking level} of the current | |
101 | thread while a given procedure is running. The blocking level starts | |
102 | out at zero, and whenever a safe point is reached, a blocking level | |
103 | greater than zero will prevent the execution of queued asyncs. | |
104 | ||
105 | Analogously, the procedure @code{call-with-unblocked-asyncs} will | |
106 | temporarily decrease the blocking level of the current thread. You | |
107 | can use it when you want to disable asyncs by default and only allow | |
108 | them temporarily. | |
109 | ||
110 | In addition to the C versions of @code{call-with-blocked-asyncs} and | |
111 | @code{call-with-unblocked-asyncs}, C code can use | |
112 | @code{scm_with_blocked_asyncs} and @code{scm_with_unblocked_asyncs} | |
113 | inside a @dfn{frame} (@pxref{Frames}) to block or unblock system asyncs | |
114 | temporarily. | |
115 | ||
116 | @deffn {Scheme Procedure} system-async-mark proc [thread] | |
117 | @deffnx {C Function} scm_system_async_mark (proc) | |
118 | @deffnx {C Function} scm_system_async_mark_for_thread (proc, thread) | |
119 | Mark @var{proc} (a procedure with zero arguments) for future execution | |
120 | in @var{thread}. When @var{proc} has already been marked for | |
121 | @var{thread} but has not been executed yet, this call has no effect. | |
122 | When @var{thread} is omitted, the thread that called | |
123 | @code{system-async-mark} is used. | |
124 | ||
125 | This procedure is not safe to be called from signal handlers. Use | |
126 | @code{scm_sigaction} or @code{scm_sigaction_for_thread} to install | |
127 | signal handlers. | |
128 | @end deffn | |
129 | ||
130 | @c FIXME: The use of @deffnx for scm_c_call_with_blocked_asyncs and | |
131 | @c scm_c_call_with_unblocked_asyncs puts "void" into the function | |
132 | @c index. Would prefer to use @deftypefnx if makeinfo allowed that, | |
133 | @c or a @deftypefn with an empty return type argument if it didn't | |
134 | @c introduce an extra space. | |
135 | ||
136 | @deffn {Scheme Procedure} call-with-blocked-asyncs proc | |
137 | @deffnx {C Function} scm_call_with_blocked_asyncs (proc) | |
138 | @deffnx {C Function} void *scm_c_call_with_blocked_asyncs (void * (*proc) (void *data), void *data) | |
139 | @findex scm_c_call_with_blocked_asyncs | |
140 | Call @var{proc} and block the execution of system asyncs by one level | |
141 | for the current thread while it is running. Return the value returned | |
142 | by @var{proc}. For the first two variants, call @var{proc} with no | |
143 | arguments; for the third, call it with @var{data}. | |
144 | @end deffn | |
145 | ||
146 | @deffn {Scheme Procedure} call-with-unblocked-asyncs proc | |
147 | @deffnx {C Function} scm_call_with_unblocked_asyncs (proc) | |
148 | @deffnx {C Function} void *scm_c_call_with_unblocked_asyncs (void *(*p) (void *d), void *d) | |
149 | @findex scm_c_call_with_unblocked_asyncs | |
150 | Call @var{proc} and unblock the execution of system asyncs by one | |
151 | level for the current thread while it is running. Return the value | |
152 | returned by @var{proc}. For the first two variants, call @var{proc} | |
153 | with no arguments; for the third, call it with @var{data}. | |
154 | @end deffn | |
155 | ||
156 | @deftypefn {C Function} void scm_frame_block_asyncs () | |
157 | This function must be used inside a pair of calls to | |
158 | @code{scm_frame_begin} and @code{scm_frame_end} (@pxref{Frames}). | |
159 | During the dynamic extent of the frame, asyncs are blocked by one level. | |
160 | @end deftypefn | |
161 | ||
162 | @deftypefn {C Function} void scm_frame_unblock_asyncs () | |
163 | This function must be used inside a pair of calls to | |
164 | @code{scm_frame_begin} and @code{scm_frame_end} (@pxref{Frames}). | |
165 | During the dynamic extent of the frame, asyncs are unblocked by one | |
166 | level. | |
167 | @end deftypefn | |
168 | ||
169 | @node User asyncs | |
170 | @subsubsection User asyncs | |
171 | ||
172 | A user async is a pair of a thunk (a parameterless procedure) and a | |
173 | mark. Setting the mark on a user async will cause the thunk to be | |
174 | executed when the user async is passed to @code{run-asyncs}. Setting | |
175 | the mark more than once is satisfied by one execution of the thunk. | |
176 | ||
177 | User asyncs are created with @code{async}. They are marked with | |
178 | @code{async-mark}. | |
179 | ||
180 | @deffn {Scheme Procedure} async thunk | |
181 | @deffnx {C Function} scm_async (thunk) | |
182 | Create a new user async for the procedure @var{thunk}. | |
183 | @end deffn | |
184 | ||
185 | @deffn {Scheme Procedure} async-mark a | |
186 | @deffnx {C Function} scm_async_mark (a) | |
187 | Mark the user async @var{a} for future execution. | |
188 | @end deffn | |
189 | ||
190 | @deffn {Scheme Procedure} run-asyncs list_of_a | |
191 | @deffnx {C Function} scm_run_asyncs (list_of_a) | |
192 | Execute all thunks from the marked asyncs of the list @var{list_of_a}. | |
193 | @end deffn | |
194 | ||
195 | ||
196 | @node Dynamic Roots | |
197 | @subsection Dynamic Roots | |
198 | @cindex dynamic roots | |
199 | ||
200 | A @dfn{dynamic root} is a root frame of Scheme evaluation. | |
201 | The top-level repl, for example, is an instance of a dynamic root. | |
202 | ||
203 | Each dynamic root has its own chain of dynamic-wind information. Each | |
204 | has its own set of continuations, jump-buffers, and pending CATCH | |
205 | statements which are inaccessible from the dynamic scope of any | |
206 | other dynamic root. | |
207 | ||
208 | In a thread-based system, each thread has its own dynamic root. Therefore, | |
209 | continuations created by one thread may not be invoked by another. | |
210 | ||
211 | Even in a single-threaded system, it is sometimes useful to create a new | |
212 | dynamic root. For example, if you want to apply a procedure, but to | |
213 | not allow that procedure to capture the current continuation, calling | |
214 | the procedure under a new dynamic root will do the job. | |
215 | ||
216 | @deffn {Scheme Procedure} call-with-dynamic-root thunk handler | |
217 | @deffnx {C Function} scm_call_with_dynamic_root (thunk, handler) | |
218 | Evaluate @code{(thunk)} in a new dynamic context, returning its value. | |
219 | ||
220 | If an error occurs during evaluation, apply @var{handler} to the | |
221 | arguments to the throw, just as @code{throw} would. If this happens, | |
222 | @var{handler} is called outside the scope of the new root -- it is | |
223 | called in the same dynamic context in which | |
224 | @code{call-with-dynamic-root} was evaluated. | |
225 | ||
226 | If @var{thunk} captures a continuation, the continuation is rooted at | |
227 | the call to @var{thunk}. In particular, the call to | |
228 | @code{call-with-dynamic-root} is not captured. Therefore, | |
229 | @code{call-with-dynamic-root} always returns at most one time. | |
230 | ||
231 | Before calling @var{thunk}, the dynamic-wind chain is un-wound back to | |
232 | the root and a new chain started for @var{thunk}. Therefore, this call | |
233 | may not do what you expect: | |
234 | ||
235 | @lisp | |
236 | ;; Almost certainly a bug: | |
237 | (with-output-to-port | |
238 | some-port | |
239 | ||
240 | (lambda () | |
241 | (call-with-dynamic-root | |
242 | (lambda () | |
243 | (display 'fnord) | |
244 | (newline)) | |
245 | (lambda (errcode) errcode)))) | |
246 | @end lisp | |
247 | ||
248 | The problem is, on what port will @samp{fnord} be displayed? You | |
249 | might expect that because of the @code{with-output-to-port} that | |
250 | it will be displayed on the port bound to @code{some-port}. But it | |
251 | probably won't -- before evaluating the thunk, dynamic winds are | |
252 | unwound, including those created by @code{with-output-to-port}. | |
253 | So, the standard output port will have been re-set to its default value | |
254 | before @code{display} is evaluated. | |
255 | ||
256 | (This function was added to Guile mostly to help calls to functions in C | |
257 | libraries that can not tolerate non-local exits or calls that return | |
258 | multiple times. If such functions call back to the interpreter, it should | |
259 | be under a new dynamic root.) | |
260 | @end deffn | |
261 | ||
262 | ||
263 | @deffn {Scheme Procedure} dynamic-root | |
264 | @deffnx {C Function} scm_dynamic_root () | |
265 | Return an object representing the current dynamic root. | |
266 | ||
267 | These objects are only useful for comparison using @code{eq?}. | |
268 | They are currently represented as numbers, but your code should | |
269 | in no way depend on this. | |
270 | @end deffn | |
271 | ||
272 | @c begin (scm-doc-string "boot-9.scm" "quit") | |
273 | @deffn {Scheme Procedure} quit [exit_val] | |
274 | Throw back to the error handler of the current dynamic root. | |
275 | ||
276 | If integer @var{exit_val} is specified and if Guile is being used | |
277 | stand-alone and if quit is called from the initial dynamic-root, | |
278 | @var{exit_val} becomes the exit status of the Guile process and the | |
279 | process exits. | |
280 | @end deffn | |
281 | ||
282 | When Guile is run interactively, errors are caught from within the | |
283 | read-eval-print loop. An error message will be printed and @code{abort} | |
284 | called. A default set of signal handlers is installed, e.g., to allow | |
285 | user interrupt of the interpreter. | |
286 | ||
287 | It is possible to switch to a "batch mode", in which the interpreter | |
288 | will terminate after an error and in which all signals cause their | |
289 | default actions. Switching to batch mode causes any handlers installed | |
290 | from Scheme code to be removed. An example of where this is useful is | |
291 | after forking a new process intended to run non-interactively. | |
292 | ||
293 | @c begin (scm-doc-string "boot-9.scm" "batch-mode?") | |
294 | @deffn {Scheme Procedure} batch-mode? | |
295 | Returns a boolean indicating whether the interpreter is in batch mode. | |
296 | @end deffn | |
297 | ||
298 | @c begin (scm-doc-string "boot-9.scm" "set-batch-mode?!") | |
299 | @deffn {Scheme Procedure} set-batch-mode?! arg | |
300 | If @var{arg} is true, switches the interpreter to batch mode. | |
301 | The @code{#f} case has not been implemented. | |
302 | @end deffn | |
303 | ||
304 | @node Threads | |
305 | @subsection Threads | |
306 | @cindex threads | |
307 | @cindex Guile threads | |
308 | @cindex POSIX threads | |
309 | ||
310 | Guile threads are implemented using POSIX threads, they run | |
311 | pre-emptively and concurrently through both Scheme code and system | |
312 | calls. The only exception is for garbage collection, where all | |
313 | threads must rendezvous. | |
314 | ||
315 | @menu | |
316 | * Low level thread primitives:: | |
317 | * Higher level thread procedures:: | |
318 | * C level thread interface:: | |
319 | @end menu | |
320 | ||
321 | ||
322 | @node Low level thread primitives | |
323 | @subsubsection Low level thread primitives | |
324 | ||
325 | @c NJFIXME no current mechanism for making sure that these docstrings | |
326 | @c are in sync. | |
327 | ||
328 | @c begin (texi-doc-string "guile" "call-with-new-thread") | |
329 | @deffn {Scheme Procedure} call-with-new-thread thunk error-handler | |
330 | Evaluate @code{(thunk)} in a new thread, and new dynamic context, | |
331 | returning a new thread object representing the thread. | |
332 | ||
333 | If an error occurs during evaluation, call error-handler, passing it | |
334 | an error code. If this happens, the error-handler is called outside | |
335 | the scope of the new root -- it is called in the same dynamic context | |
336 | in which with-new-thread was evaluated, but not in the caller's | |
337 | thread. | |
338 | ||
339 | All the evaluation rules for dynamic roots apply to threads. | |
340 | @end deffn | |
341 | ||
342 | @c begin (texi-doc-string "guile" "join-thread") | |
343 | @deffn {Scheme Procedure} join-thread thread | |
344 | Suspend execution of the calling thread until the target @var{thread} | |
345 | terminates, unless the target @var{thread} has already terminated. | |
346 | @end deffn | |
347 | ||
348 | @c begin (texi-doc-string "guile" "yield") | |
349 | @deffn {Scheme Procedure} yield | |
350 | If one or more threads are waiting to execute, calling yield forces an | |
351 | immediate context switch to one of them. Otherwise, yield has no effect. | |
352 | @end deffn | |
353 | ||
354 | @c begin (texi-doc-string "guile" "make-mutex") | |
355 | @deffn {Scheme Procedure} make-mutex | |
356 | Create a new mutex object. | |
357 | @end deffn | |
358 | ||
359 | @c begin (texi-doc-string "guile" "lock-mutex") | |
360 | @deffn {Scheme Procedure} lock-mutex mutex | |
361 | Lock @var{mutex}. If the mutex is already locked, the calling thread | |
362 | blocks until the mutex becomes available. The function returns when | |
363 | the calling thread owns the lock on @var{mutex}. Locking a mutex that | |
364 | a thread already owns will succeed right away and will not block the | |
365 | thread. That is, Guile's mutexes are @emph{recursive}. | |
366 | ||
367 | When a system async is activated for a thread that is blocked in a | |
368 | call to @code{lock-mutex}, the waiting is interrupted and the async is | |
369 | executed. When the async returns, the waiting is resumed. | |
370 | @end deffn | |
371 | ||
372 | @deffn {Scheme Procedure} try-mutex mutex | |
373 | Try to lock @var{mutex}. If the mutex is already locked by someone | |
374 | else, return @code{#f}. Else lock the mutex and return @code{#t}. | |
375 | @end deffn | |
376 | ||
377 | @c begin (texi-doc-string "guile" "unlock-mutex") | |
378 | @deffn {Scheme Procedure} unlock-mutex mutex | |
379 | Unlocks @var{mutex} if the calling thread owns the lock on | |
380 | @var{mutex}. Calling unlock-mutex on a mutex not owned by the current | |
381 | thread results in undefined behaviour. Once a mutex has been unlocked, | |
382 | one thread blocked on @var{mutex} is awakened and grabs the mutex | |
383 | lock. Every call to @code{lock-mutex} by this thread must be matched | |
384 | with a call to @code{unlock-mutex}. Only the last call to | |
385 | @code{unlock-mutex} will actually unlock the mutex. | |
386 | @end deffn | |
387 | ||
388 | @c begin (texi-doc-string "guile" "make-condition-variable") | |
389 | @deffn {Scheme Procedure} make-condition-variable | |
390 | Make a new condition variable. | |
391 | @end deffn | |
392 | ||
393 | @c begin (texi-doc-string "guile" "wait-condition-variable") | |
394 | @deffn {Scheme Procedure} wait-condition-variable cond-var mutex [time] | |
395 | Wait until @var{cond-var} has been signalled. While waiting, | |
396 | @var{mutex} is atomically unlocked (as with @code{unlock-mutex}) and | |
397 | is locked again when this function returns. When @var{time} is given, | |
398 | it specifies a point in time where the waiting should be aborted. It | |
399 | can be either a integer as returned by @code{current-time} or a pair | |
400 | as returned by @code{gettimeofday}. When the waiting is aborted, | |
401 | @code{#f} is returned. When the condition variable has in fact been | |
402 | signalled, @code{#t} is returned. The mutex is re-locked in any case | |
403 | before @code{wait-condition-variable} returns. | |
404 | ||
405 | When a system async is activated for a thread that is blocked in a | |
406 | call to @code{wait-condition-variable}, the waiting is interrupted, | |
407 | the mutex is locked, and the async is executed. When the async | |
408 | returns, the mutex is unlocked again and the waiting is resumed. | |
409 | @end deffn | |
410 | ||
411 | @c begin (texi-doc-string "guile" "signal-condition-variable") | |
412 | @deffn {Scheme Procedure} signal-condition-variable cond-var | |
413 | Wake up one thread that is waiting for @var{cv}. | |
414 | @end deffn | |
415 | ||
416 | @c begin (texi-doc-string "guile" "broadcast-condition-variable") | |
417 | @deffn {Scheme Procedure} broadcast-condition-variable cond-var | |
418 | Wake up all threads that are waiting for @var{cv}. | |
419 | @end deffn | |
420 | ||
421 | @node Higher level thread procedures | |
422 | @subsubsection Higher level thread procedures | |
423 | ||
424 | @c new by ttn, needs review | |
425 | ||
426 | Higher level thread procedures are available by loading the | |
427 | @code{(ice-9 threads)} module. These provide standardized | |
428 | thread creation and mutex interaction. | |
429 | ||
430 | @deffn macro make-thread proc [args@dots{}] | |
431 | Apply @var{proc} to @var{args} in a new thread formed by | |
432 | @code{call-with-new-thread} using a default error handler that display | |
433 | the error to the current error port. | |
434 | @end deffn | |
435 | ||
436 | @deffn macro begin-thread first [rest@dots{}] | |
437 | Evaluate forms @var{first} and @var{rest} in a new thread formed by | |
438 | @code{call-with-new-thread} using a default error handler that display | |
439 | the error to the current error port. | |
440 | @end deffn | |
441 | ||
442 | @deffn macro with-mutex m [body@dots{}] | |
443 | Lock mutex @var{m}, evaluate @var{body}, and then unlock @var{m}. | |
444 | These sub-operations form the branches of a @code{dynamic-wind}. | |
445 | @end deffn | |
446 | ||
447 | @deffn macro monitor body@dots{} | |
448 | Evaluate @var{body}, with a mutex locked so only one thread can | |
449 | execute that code at any one time. Each @code{monitor} form has its | |
450 | own private mutex and the locking is done as per @code{with-mutex} | |
451 | above. The return value is the return from the last form in | |
452 | @var{body}. | |
453 | ||
454 | The term ``monitor'' comes from operating system theory, where it | |
455 | means a particular bit of code managing access to some resource and | |
456 | which only ever executes on behalf of one process at any one time. | |
457 | @end deffn | |
458 | ||
459 | @node C level thread interface | |
460 | @subsubsection C level thread interface | |
461 | ||
462 | You can create and manage threads, mutexes, and condition variables | |
463 | with the C versions of the primitives above. For example, you can | |
464 | create a mutex with @code{scm_make_mutex} and lock it with | |
465 | @code{scm_lock_mutex}. In addition to these primitives there is also | |
466 | a second set of primitives for threading related things. These | |
467 | functions and data types are only available from C and can not be | |
468 | mixed with the first set from above. However, they might be more | |
469 | efficient and can be used in situations where Scheme data types are | |
470 | not allowed or are inconvenient to use. | |
471 | ||
472 | Furthermore, they are the primitives that Guile relies on for its own | |
473 | higher level threads. By reimplementing them, you can adapt Guile to | |
474 | different low-level thread implementations. | |
475 | ||
476 | C code in a thread must call a libguile function periodically. When | |
477 | one thread finds garbage collection is required, it waits for all | |
478 | threads to rendezvous before doing that GC. Such a rendezvous is | |
479 | checked within libguile functions. If C code wants to sleep or block | |
480 | in a thread it should use one of the libguile functions provided. | |
481 | ||
482 | Only threads created by Guile can use the libguile functions. Threads | |
483 | created directly with say @code{pthread_create} are unknown to Guile | |
484 | and they cannot call libguile. The stack in such foreign threads is | |
485 | not scanned during GC, so @code{SCM} values generally cannot be held | |
486 | there. | |
487 | ||
488 | @c FIXME: | |
489 | @c | |
490 | @c Describe SCM_TICK which can be called if no other libguile | |
491 | @c function is being used by a C function. | |
492 | @c | |
493 | @c Describe "Guile mode", which a thread can enter and exit. There | |
494 | @c are no functions for doing this yet. | |
495 | @c | |
496 | @c When in guile mode a thread can call libguile, is subject to the | |
497 | @c tick rule, and its stack is scanned. When not in guile mode it | |
498 | @c cannot call libguile, it doesn't have to tick, and its stack is | |
499 | @c not scanned. The strange guile control flow things like | |
500 | @c exceptions, continuations and asyncs only occur when in guile | |
501 | @c mode. | |
502 | @c | |
503 | @c When guile mode is exited, the portion of the stack allocated | |
504 | @c while it was in guile mode is still scanned. This portion may not | |
505 | @c be modified when outside guile mode. The stack ends up | |
506 | @c partitioned into alternating guile and non-guile regions. | |
507 | @c | |
508 | @c Leaving guile mode is convenient when running an extended | |
509 | @c calculation not involving guile, since one doesn't need to worry | |
510 | @c about SCM_TICK calls. | |
511 | ||
512 | ||
513 | @deftp {C Data Type} scm_t_thread | |
514 | This data type represents a thread, to be used with scm_thread_create, | |
515 | etc. | |
516 | @end deftp | |
517 | ||
518 | @deftypefn {C Function} int scm_thread_create (scm_t_thread *t, void (*proc)(void *), void *data) | |
519 | Create a new thread that will start by calling @var{proc}, passing it | |
520 | @var{data}. A handle for the new thread is stored in @var{t}, which | |
521 | must be non-NULL. The thread terminated when @var{proc} returns. | |
522 | When the thread has not been detached, its handle remains valid after | |
523 | is has terminated so that it can be used with @var{scm_thread_join}, | |
524 | for example. When it has been detached, the handle becomes invalid as | |
525 | soon as the thread terminates. | |
526 | @end deftypefn | |
527 | ||
528 | @deftypefn {C Function} void scm_thread_detach (scm_t_thread t) | |
529 | Detach the thread @var{t}. See @code{scm_thread_create}. | |
530 | @end deftypefn | |
531 | ||
532 | @deftypefn {C Function} void scm_thread_join (scm_t_thread t) | |
533 | Wait for thread @var{t} to terminate. The thread must not have been | |
534 | detached at the time that @code{scm_thread_join} is called, but it | |
535 | might have been detached by the time it terminates. | |
536 | @end deftypefn | |
537 | ||
538 | @deftypefn {C Function} scm_t_thread scm_thread_self () | |
539 | Return the handle of the calling thread. | |
540 | @end deftypefn | |
541 | ||
542 | @deftp {C Data Type} scm_t_mutex | |
543 | This data type represents a mutex, to be used with scm_mutex_init, | |
544 | etc. | |
545 | @end deftp | |
546 | ||
547 | @deftypefn {C Function} void scm_mutex_init (scm_t_mutex *m) | |
548 | Initialize the mutex structure pointed to by @var{m}. | |
549 | @end deftypefn | |
550 | ||
551 | @deftypefn {C Function} void scm_mutex_destroy (scm_t_mutex *m) | |
552 | Deallocate all resources associated with @var{m}. | |
553 | @end deftypefn | |
554 | ||
555 | @deftypefn {C Function} void scm_mutex_lock (scm_t_mutex *m) | |
556 | Lock the mutex @var{m}. When it is already locked by a different | |
557 | thread, wait until it becomes available. Locking a mutex that is | |
558 | already locked by the current threads is not allowd and results in | |
559 | undefined behavior. The mutices are not guaranteed to be fair. That | |
560 | is, a thread that attempts a lock after yourself might be granted it | |
561 | before you. | |
562 | @end deftypefn | |
563 | ||
564 | @deftypefn {C Function} int scm_mutex_trylock (scm_t_mutex *m) | |
565 | Lock @var{m} as with @code{scm_mutex_lock} but don't wait when this | |
566 | does succeed immediately. Returns non-zero when the mutex could in | |
567 | fact be locked , and zero when it is already locked by some other | |
568 | thread. | |
569 | @end deftypefn | |
570 | ||
571 | @deftypefn {C Function} void scm_mutex_unlock (scm_t_mutex *m) | |
572 | Unlock the mutex @var{m}. The mutex must have been locked by the | |
573 | current thread, else the behavior is undefined. | |
574 | @end deftypefn | |
575 | ||
576 | @deftp {C Data Type} scm_t_cond | |
577 | This data type represents a condition variable, to be used with | |
578 | scm_cond_init, etc. | |
579 | @end deftp | |
580 | ||
581 | @deftypefn {C Function} void scm_cond_init (scm_t_cond *c) | |
582 | Initialize the mutex structure pointed to by @var{c}. | |
583 | @end deftypefn | |
584 | ||
585 | @deftypefn {C Function} void scm_cond_destroy (scm_t_cond *c) | |
586 | Deallocate all resources associated with @var{c}. | |
587 | @end deftypefn | |
588 | ||
589 | @deftypefn {C Function} void scm_cond_wait (scm_t_cond *c, scm_t_mutex *m) | |
590 | Wait for @var{c} to be signalled. While waiting @var{m} is unlocked | |
591 | and locked again before @code{scm_cond_wait} returns. | |
592 | @end deftypefn | |
593 | ||
594 | @deftypefn {C Function} void scm_cond_timedwait (scm_t_cond *c, scm_t_mutex *m, timespec *abstime) | |
595 | Wait for @var{c} to be signalled as with @code{scm_cond_wait} but | |
596 | don't wait longer than the point in time specified by @var{abstime}. | |
597 | when the waiting is aborted, zero is returned; non-zero else. | |
598 | @end deftypefn | |
599 | ||
600 | @deftypefn {C Function} void scm_cond_signal (scm_t_cond *c) | |
601 | Signal the condition variable @var{c}. When one or more threads are | |
602 | waiting for it to be signalled, select one arbitrarily and let its | |
603 | wait succeed. | |
604 | @end deftypefn | |
605 | ||
606 | @deftypefn {C Function} void scm_cond_broadcast (scm_t_cond *c) | |
607 | Signal the condition variable @var{c}. When there are threads waiting | |
608 | for it to be signalled, wake them all up and make all their waits | |
609 | succeed. | |
610 | @end deftypefn | |
611 | ||
612 | @deftp {C Type} scm_t_key | |
613 | This type represents a key for a thread-specific value. | |
614 | @end deftp | |
615 | ||
616 | @deftypefn {C Function} void scm_key_create (scm_t_key *keyp) | |
617 | Create a new key for a thread-specific value. Each thread has its own | |
618 | value associated to such a handle. The new handle is stored into | |
619 | @var{keyp}, which must be non-NULL. | |
620 | @end deftypefn | |
621 | ||
622 | @deftypefn {C Function} void scm_key_delete (scm_t_key key) | |
623 | This function makes @var{key} invalid as a key for thread-specific data. | |
624 | @end deftypefn | |
625 | ||
626 | @deftypefn {C Function} void scm_key_setspecific (scm_t_key key, const void *value) | |
627 | Associate @var{value} with @var{key} in the calling thread. | |
628 | @end deftypefn | |
629 | ||
630 | @deftypefn {C Function} int scm_key_getspecific (scm_t_key key) | |
631 | Return the value currently associated with @var{key} in the calling | |
632 | thread. When @code{scm_key_setspecific} has not yet been called in | |
633 | this thread with this key, @code{NULL} is returned. | |
634 | @end deftypefn | |
635 | ||
636 | @deftypefn {C Function} int scm_thread_select (...) | |
637 | This function does the same thing as the system's @code{select} | |
638 | function, but in a way that is friendly to the thread implementation. | |
639 | You should call it in preference to the system @code{select}. | |
640 | @end deftypefn | |
641 | ||
642 | @node Fluids | |
643 | @subsection Fluids | |
644 | ||
645 | @cindex fluids | |
646 | ||
647 | Fluids are objects to store values in. They have a few properties | |
648 | which make them useful in certain situations: Fluids can have one | |
649 | value per dynamic root (@pxref{Dynamic Roots}), so that changes to the | |
650 | value in a fluid are only visible in the same dynamic root. Since | |
651 | threads are executed in separate dynamic roots, fluids can be used for | |
652 | thread local storage (@pxref{Threads}). | |
653 | ||
654 | Fluids can be used to simulate the desirable effects of dynamically | |
655 | scoped variables. Dynamically scoped variables are useful when you | |
656 | want to set a variable to a value during some dynamic extent in the | |
657 | execution of your program and have them revert to their original value | |
658 | when the control flow is outside of this dynamic extent. See the | |
659 | description of @code{with-fluids} below for details. | |
660 | ||
661 | New fluids are created with @code{make-fluid} and @code{fluid?} is | |
662 | used for testing whether an object is actually a fluid. The values | |
663 | stored in a fluid can be accessed with @code{fluid-ref} and | |
664 | @code{fluid-set!}. | |
665 | ||
666 | @deffn {Scheme Procedure} make-fluid | |
667 | @deffnx {C Function} scm_make_fluid () | |
668 | Return a newly created fluid. | |
669 | Fluids are objects of a certain type (a smob) that can hold one SCM | |
670 | value per dynamic root. That is, modifications to this value are | |
671 | only visible to code that executes within the same dynamic root as | |
672 | the modifying code. When a new dynamic root is constructed, it | |
673 | inherits the values from its parent. Because each thread executes | |
674 | in its own dynamic root, you can use fluids for thread local storage. | |
675 | @end deffn | |
676 | ||
677 | @deffn {Scheme Procedure} fluid? obj | |
678 | @deffnx {C Function} scm_fluid_p (obj) | |
679 | Return @code{#t} iff @var{obj} is a fluid; otherwise, return | |
680 | @code{#f}. | |
681 | @end deffn | |
682 | ||
683 | @deffn {Scheme Procedure} fluid-ref fluid | |
684 | @deffnx {C Function} scm_fluid_ref (fluid) | |
685 | Return the value associated with @var{fluid} in the current | |
686 | dynamic root. If @var{fluid} has not been set, then return | |
687 | @code{#f}. | |
688 | @end deffn | |
689 | ||
690 | @deffn {Scheme Procedure} fluid-set! fluid value | |
691 | @deffnx {C Function} scm_fluid_set_x (fluid, value) | |
692 | Set the value associated with @var{fluid} in the current dynamic root. | |
693 | @end deffn | |
694 | ||
695 | @code{with-fluids*} temporarily changes the values of one or more fluids, | |
696 | so that the given procedure and each procedure called by it access the | |
697 | given values. After the procedure returns, the old values are restored. | |
698 | ||
699 | @deffn {Scheme Procedure} with-fluids* fluids values thunk | |
700 | @deffnx {C Function} scm_with_fluids (fluids, values, thunk) | |
701 | Set @var{fluids} to @var{values} temporary, and call @var{thunk}. | |
702 | @var{fluids} must be a list of fluids and @var{values} must be the | |
703 | same number of their values to be applied. Each substitution is done | |
704 | in the order given. @var{thunk} must be a procedure with no argument. | |
705 | it is called inside a @code{dynamic-wind} and the fluids are | |
706 | set/restored when control enter or leaves the established dynamic | |
707 | extent. | |
708 | @end deffn | |
709 | ||
710 | @deffn {Scheme Macro} with-fluids ((fluid value) ...) body... | |
711 | Execute @var{body...} while each @var{fluid} is set to the | |
712 | corresponding @var{value}. Both @var{fluid} and @var{value} are | |
713 | evaluated and @var{fluid} must yield a fluid. @var{body...} is | |
714 | executed inside a @code{dynamic-wind} and the fluids are set/restored | |
715 | when control enter or leaves the established dynamic extent. | |
716 | @end deffn | |
717 | ||
718 | @deftypefn {C Function} SCM scm_c_with_fluids (SCM fluids, SCM vals, SCM (*cproc)(void *), void *data) | |
719 | @deftypefnx {C Function} SCM scm_c_with_fluid (SCM fluid, SCM val, SCM (*cproc)(void *), void *data) | |
720 | The function @code{scm_c_with_fluids} is like @code{scm_with_fluids} | |
721 | except that it takes a C function to call instead of a Scheme thunk. | |
722 | ||
723 | The function @code{scm_c_with_fluid} is similar but only allows one | |
724 | fluid to be set instead of a list. | |
725 | @end deftypefn | |
726 | ||
727 | @deftypefn {C Function} void scm_frame_fluid (SCM fluid, SCM val) | |
728 | This function must be used inside a pair of calls to | |
729 | @code{scm_frame_begin} and @code{scm_frame_end} (@pxref{Frames}). | |
730 | During the dynamic extent of the frame, the fluid @var{fluid} is set | |
731 | to @var{val}. | |
732 | ||
733 | More precisely, the value of the fluid is swapped with a `backup' | |
734 | value whenever the frame is entered or left. The backup value is | |
735 | initialized with the @var{val} argument. | |
736 | @end deftypefn | |
737 | ||
738 | @node Futures | |
739 | @subsection Futures | |
740 | @cindex futures | |
741 | ||
742 | Futures are a convenient way to run a calculation in a new thread, and | |
743 | only wait for the result when it's actually needed. | |
744 | ||
745 | Futures are similar to promises (@pxref{Delayed Evaluation}), in that | |
746 | they allow mainline code to continue immediately. But @code{delay} | |
747 | doesn't evaluate at all until forced, whereas @code{future} starts | |
748 | immediately in a new thread. | |
749 | ||
750 | @deffn {syntax} future expr | |
751 | Begin evaluating @var{expr} in a new thread, and return a ``future'' | |
752 | object representing the calculation. | |
753 | @end deffn | |
754 | ||
755 | @deffn {Scheme Procedure} make-future thunk | |
756 | @deffnx {C Function} scm_make_future (thunk) | |
757 | Begin evaluating the call @code{(@var{thunk})} in a new thread, and | |
758 | return a ``future'' object representing the calculation. | |
759 | @end deffn | |
760 | ||
761 | @deffn {Scheme Procedure} future-ref f | |
762 | @deffnx {C Function} scm_future_ref (f) | |
763 | Return the value computed by the future @var{f}. If @var{f} has not | |
764 | yet finished executing then wait for it to do so. | |
765 | @end deffn | |
766 | ||
767 | ||
768 | @node Parallel Forms | |
769 | @subsection Parallel forms | |
770 | @cindex parallel forms | |
771 | ||
772 | The functions described in this section are available from | |
773 | ||
774 | @example | |
775 | (use-modules (ice-9 threads)) | |
776 | @end example | |
777 | ||
778 | @deffn syntax parallel expr1 @dots{} exprN | |
779 | Evaluate each @var{expr} expression in parallel, each in a new thread. | |
780 | Return the results as a set of @var{N} multiple values | |
781 | (@pxref{Multiple Values}). | |
782 | @end deffn | |
783 | ||
784 | @deffn syntax letpar ((var1 expr1) @dots{} (varN exprN)) body@dots{} | |
785 | Evaluate each @var{expr} in parallel, each in a new thread, then bind | |
786 | the results to the corresponding @var{var} variables and evaluate | |
787 | @var{body}. | |
788 | ||
789 | @code{letpar} is like @code{let} (@pxref{Local Bindings}), but all the | |
790 | expressions for the bindings are evaluated in parallel. | |
791 | @end deffn | |
792 | ||
793 | @deffn {Scheme Procedure} par-map proc lst1 @dots{} lstN | |
794 | @deffnx {Scheme Procedure} par-for-each proc lst1 @dots{} lstN | |
795 | Call @var{proc} on the elements of the given lists. @code{par-map} | |
796 | returns a list comprising the return values from @var{proc}. | |
797 | @code{par-for-each} returns an unspecified value, but waits for all | |
798 | calls to complete. | |
799 | ||
800 | The @var{proc} calls are @code{(@var{proc} @var{elem1} @dots{} | |
801 | @var{elemN})}, where each @var{elem} is from the corresponding | |
802 | @var{lst}. Each @var{lst} must be the same length. The calls are | |
803 | made in parallel, each in a new thread. | |
804 | ||
805 | These functions are like @code{map} and @code{for-each} (@pxref{List | |
806 | Mapping}), but make their @var{proc} calls in parallel. | |
807 | @end deffn | |
808 | ||
809 | @deffn {Scheme Procedure} n-par-map n proc lst1 @dots{} lstN | |
810 | @deffnx {Scheme Procedure} n-par-for-each n proc lst1 @dots{} lstN | |
811 | Call @var{proc} on the elements of the given lists, in the same way as | |
812 | @code{par-map} and @code{par-for-each} above, but use no more than | |
813 | @var{n} new threads at any one time. The order in which calls are | |
814 | initiated within that threads limit is unspecified. | |
815 | ||
816 | These functions are good for controlling resource consumption if | |
817 | @var{proc} calls might be costly, or if there are many to be made. On | |
818 | a dual-CPU system for instance @math{@var{n}=4} might be enough to | |
819 | keep the CPUs utilized, and not consume too much memory. | |
820 | @end deffn | |
821 | ||
822 | @deffn {Scheme Procedure} n-for-each-par-map n sproc pproc lst1 @dots{} lstN | |
823 | Apply @var{pproc} to the elements of the given lists, and apply | |
824 | @var{sproc} to each result returned by @var{pproc}. The final return | |
825 | value is unspecified, but all calls will have been completed before | |
826 | returning. | |
827 | ||
828 | The calls made are @code{(@var{sproc} (@var{pproc} @var{elem1} @dots{} | |
829 | @var{elemN}))}, where each @var{elem} is from the corresponding | |
830 | @var{lst}. Each @var{lst} must have the same number of elements. | |
831 | ||
832 | The @var{pproc} calls are made in parallel, in new threads. No more | |
833 | than @var{n} new threads are used at any one time. The order in which | |
834 | @var{pproc} calls are initiated within that limit is unspecified. | |
835 | ||
836 | The @var{sproc} calls are made serially, in list element order, one at | |
837 | a time. @var{pproc} calls on later elements may execute in parallel | |
838 | with the @var{sproc} calls. Exactly which thread makes each | |
839 | @var{sproc} call is unspecified. | |
840 | ||
841 | This function is designed for individual calculations that can be done | |
842 | in parallel, but with results needing to be handled serially, for | |
843 | instance to write them to a file. The @var{n} limit on threads | |
844 | controls system resource usage when there are many calculations or | |
845 | when they might be costly. | |
846 | ||
847 | It will be seen that @code{n-for-each-par-map} is like a combination | |
848 | of @code{n-par-map} and @code{for-each}, | |
849 | ||
850 | @example | |
851 | (for-each sproc (n-par-map pproc lst1 ... lstN)) | |
852 | @end example | |
853 | ||
854 | @noindent | |
855 | But the actual implementation is more efficient since each @var{sproc} | |
856 | call, in turn, can be initiated once the relevant @var{pproc} call has | |
857 | completed, it doesn't need to wait for all to finish. | |
858 | @end deffn | |
859 | ||
860 | ||
861 | @c Local Variables: | |
862 | @c TeX-master: "guile.texi" | |
863 | @c End: |