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. | |
b4fddbbe | 18 | * Continuation Barriers:: Protection from non-local control flow. |
07d83abe | 19 | * Threads:: Multiple threads of execution. |
2567692a | 20 | * Mutexes and Condition Variables:: Synchronization primitives. |
b4fddbbe | 21 | * Blocking:: How to block properly in guile mode. |
2567692a | 22 | * Critical Sections:: Avoiding concurrency and reentries. |
b4fddbbe | 23 | * Fluids and Dynamic States:: Thread-local variables, etc. |
07d83abe MV |
24 | * Futures:: Delayed execution in new threads. |
25 | * Parallel Forms:: Parallel execution of forms. | |
26 | @end menu | |
27 | ||
28 | ||
29 | @node Arbiters | |
30 | @subsection Arbiters | |
07d83abe MV |
31 | @cindex arbiters |
32 | ||
e136aab0 KR |
33 | Arbiters are synchronization objects, they can be used by threads to |
34 | control access to a shared resource. An arbiter can be locked to | |
35 | indicate a resource is in use, and unlocked when done. | |
07d83abe | 36 | |
b4fddbbe MV |
37 | An arbiter is like a light-weight mutex (@pxref{Mutexes and Condition |
38 | Variables}). It uses less memory and may be faster, but there's no | |
39 | way for a thread to block waiting on an arbiter, it can only test and | |
40 | get the status returned. | |
07d83abe MV |
41 | |
42 | @deffn {Scheme Procedure} make-arbiter name | |
43 | @deffnx {C Function} scm_make_arbiter (name) | |
cdf1ad3b MV |
44 | Return an object of type arbiter and name @var{name}. Its |
45 | state is initially unlocked. Arbiters are a way to achieve | |
46 | process synchronization. | |
07d83abe MV |
47 | @end deffn |
48 | ||
49 | @deffn {Scheme Procedure} try-arbiter arb | |
50 | @deffnx {C Function} scm_try_arbiter (arb) | |
cdf1ad3b MV |
51 | @deffnx {C Function} scm_try_arbiter (arb) |
52 | If @var{arb} is unlocked, then lock it and return @code{#t}. | |
53 | If @var{arb} is already locked, then do nothing and return | |
54 | @code{#f}. | |
07d83abe MV |
55 | @end deffn |
56 | ||
57 | @deffn {Scheme Procedure} release-arbiter arb | |
58 | @deffnx {C Function} scm_release_arbiter (arb) | |
e136aab0 KR |
59 | If @var{arb} is locked, then unlock it and return @code{#t}. If |
60 | @var{arb} is already unlocked, then do nothing and return @code{#f}. | |
61 | ||
62 | Typical usage is for the thread which locked an arbiter to later | |
63 | release it, but that's not required, any thread can release it. | |
07d83abe MV |
64 | @end deffn |
65 | ||
66 | ||
67 | @node Asyncs | |
68 | @subsection Asyncs | |
69 | ||
70 | @cindex asyncs | |
71 | @cindex user asyncs | |
72 | @cindex system asyncs | |
73 | ||
74 | Asyncs are a means of deferring the excution of Scheme code until it is | |
75 | safe to do so. | |
76 | ||
77 | Guile provides two kinds of asyncs that share the basic concept but are | |
78 | otherwise quite different: system asyncs and user asyncs. System asyncs | |
79 | are integrated into the core of Guile and are executed automatically | |
80 | when the system is in a state to allow the execution of Scheme code. | |
81 | For example, it is not possible to execute Scheme code in a POSIX signal | |
82 | handler, but such a signal handler can queue a system async to be | |
83 | executed in the near future, when it is safe to do so. | |
84 | ||
85 | System asyncs can also be queued for threads other than the current one. | |
86 | This way, you can cause threads to asynchronously execute arbitrary | |
87 | code. | |
88 | ||
89 | User asyncs offer a convenient means of queueing procedures for future | |
90 | execution and triggering this execution. They will not be executed | |
91 | automatically. | |
92 | ||
93 | @menu | |
94 | * System asyncs:: | |
95 | * User asyncs:: | |
96 | @end menu | |
97 | ||
98 | @node System asyncs | |
99 | @subsubsection System asyncs | |
100 | ||
101 | To cause the future asynchronous execution of a procedure in a given | |
102 | thread, use @code{system-async-mark}. | |
103 | ||
104 | Automatic invocation of system asyncs can be temporarily disabled by | |
105 | calling @code{call-with-blocked-asyncs}. This function works by | |
106 | temporarily increasing the @emph{async blocking level} of the current | |
107 | thread while a given procedure is running. The blocking level starts | |
108 | out at zero, and whenever a safe point is reached, a blocking level | |
109 | greater than zero will prevent the execution of queued asyncs. | |
110 | ||
111 | Analogously, the procedure @code{call-with-unblocked-asyncs} will | |
112 | temporarily decrease the blocking level of the current thread. You | |
113 | can use it when you want to disable asyncs by default and only allow | |
114 | them temporarily. | |
115 | ||
116 | In addition to the C versions of @code{call-with-blocked-asyncs} and | |
117 | @code{call-with-unblocked-asyncs}, C code can use | |
b4fddbbe | 118 | @code{scm_frame_block_asyncs} and @code{scm_frame_unblock_asyncs} |
07d83abe MV |
119 | inside a @dfn{frame} (@pxref{Frames}) to block or unblock system asyncs |
120 | temporarily. | |
121 | ||
122 | @deffn {Scheme Procedure} system-async-mark proc [thread] | |
123 | @deffnx {C Function} scm_system_async_mark (proc) | |
124 | @deffnx {C Function} scm_system_async_mark_for_thread (proc, thread) | |
125 | Mark @var{proc} (a procedure with zero arguments) for future execution | |
126 | in @var{thread}. When @var{proc} has already been marked for | |
127 | @var{thread} but has not been executed yet, this call has no effect. | |
128 | When @var{thread} is omitted, the thread that called | |
129 | @code{system-async-mark} is used. | |
130 | ||
131 | This procedure is not safe to be called from signal handlers. Use | |
132 | @code{scm_sigaction} or @code{scm_sigaction_for_thread} to install | |
133 | signal handlers. | |
134 | @end deffn | |
135 | ||
136 | @c FIXME: The use of @deffnx for scm_c_call_with_blocked_asyncs and | |
137 | @c scm_c_call_with_unblocked_asyncs puts "void" into the function | |
138 | @c index. Would prefer to use @deftypefnx if makeinfo allowed that, | |
139 | @c or a @deftypefn with an empty return type argument if it didn't | |
140 | @c introduce an extra space. | |
141 | ||
142 | @deffn {Scheme Procedure} call-with-blocked-asyncs proc | |
143 | @deffnx {C Function} scm_call_with_blocked_asyncs (proc) | |
144 | @deffnx {C Function} void *scm_c_call_with_blocked_asyncs (void * (*proc) (void *data), void *data) | |
145 | @findex scm_c_call_with_blocked_asyncs | |
146 | Call @var{proc} and block the execution of system asyncs by one level | |
147 | for the current thread while it is running. Return the value returned | |
148 | by @var{proc}. For the first two variants, call @var{proc} with no | |
149 | arguments; for the third, call it with @var{data}. | |
150 | @end deffn | |
151 | ||
152 | @deffn {Scheme Procedure} call-with-unblocked-asyncs proc | |
153 | @deffnx {C Function} scm_call_with_unblocked_asyncs (proc) | |
154 | @deffnx {C Function} void *scm_c_call_with_unblocked_asyncs (void *(*p) (void *d), void *d) | |
155 | @findex scm_c_call_with_unblocked_asyncs | |
156 | Call @var{proc} and unblock the execution of system asyncs by one | |
157 | level for the current thread while it is running. Return the value | |
158 | returned by @var{proc}. For the first two variants, call @var{proc} | |
159 | with no arguments; for the third, call it with @var{data}. | |
160 | @end deffn | |
161 | ||
162 | @deftypefn {C Function} void scm_frame_block_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 blocked by one level. | |
166 | @end deftypefn | |
167 | ||
168 | @deftypefn {C Function} void scm_frame_unblock_asyncs () | |
169 | This function must be used inside a pair of calls to | |
170 | @code{scm_frame_begin} and @code{scm_frame_end} (@pxref{Frames}). | |
171 | During the dynamic extent of the frame, asyncs are unblocked by one | |
172 | level. | |
173 | @end deftypefn | |
174 | ||
175 | @node User asyncs | |
176 | @subsubsection User asyncs | |
177 | ||
178 | A user async is a pair of a thunk (a parameterless procedure) and a | |
179 | mark. Setting the mark on a user async will cause the thunk to be | |
180 | executed when the user async is passed to @code{run-asyncs}. Setting | |
181 | the mark more than once is satisfied by one execution of the thunk. | |
182 | ||
183 | User asyncs are created with @code{async}. They are marked with | |
184 | @code{async-mark}. | |
185 | ||
186 | @deffn {Scheme Procedure} async thunk | |
187 | @deffnx {C Function} scm_async (thunk) | |
188 | Create a new user async for the procedure @var{thunk}. | |
189 | @end deffn | |
190 | ||
191 | @deffn {Scheme Procedure} async-mark a | |
192 | @deffnx {C Function} scm_async_mark (a) | |
193 | Mark the user async @var{a} for future execution. | |
194 | @end deffn | |
195 | ||
196 | @deffn {Scheme Procedure} run-asyncs list_of_a | |
197 | @deffnx {C Function} scm_run_asyncs (list_of_a) | |
198 | Execute all thunks from the marked asyncs of the list @var{list_of_a}. | |
199 | @end deffn | |
200 | ||
b4fddbbe MV |
201 | @node Continuation Barriers |
202 | @subsection Continuation Barriers | |
07d83abe | 203 | |
b4fddbbe MV |
204 | The non-local flow of control caused by continuations might sometimes |
205 | not be wanted. You can use @code{with-continuation-barrier} etc to | |
206 | errect fences that continuations can not pass. | |
07d83abe | 207 | |
673ba2da | 208 | @deffn {Scheme Procedure} with-continuation-barrier proc |
b4fddbbe MV |
209 | @deffnx {C Function} scm_with_continuation_barrier (proc) |
210 | Call @var{proc} and return its result. Do not allow the invocation of | |
211 | continuations that would leave or enter the dynamic extent of the call | |
212 | to @code{with-continuation-barrier}. Such an attempt causes an error | |
213 | to be signaled. | |
07d83abe | 214 | |
b4fddbbe MV |
215 | Throws (such as errors) that are not caught from within @var{proc} are |
216 | caught by @code{with-continuation-barrier}. In that case, a short | |
217 | message is printed to the current error port and @code{#f} is returned. | |
07d83abe | 218 | |
b4fddbbe | 219 | Thus, @code{with-continuation-barrier} returns exactly once. |
07d83abe MV |
220 | @end deffn |
221 | ||
c2110081 | 222 | @deftypefn {C Function} {void *} scm_c_with_continuation_barrier (void *(*func) (void *), void *data) |
b4fddbbe MV |
223 | Like @code{scm_with_continuation_barrier} but call @var{func} on |
224 | @var{data}. When an error is caught, @code{NULL} is returned. | |
225 | @end deftypefn | |
07d83abe MV |
226 | |
227 | @node Threads | |
228 | @subsection Threads | |
229 | @cindex threads | |
230 | @cindex Guile threads | |
231 | @cindex POSIX threads | |
232 | ||
cdf1ad3b MV |
233 | @deffn {Scheme Procedure} all-threads |
234 | @deffnx {C Function} scm_all_threads () | |
235 | Return a list of all threads. | |
236 | @end deffn | |
237 | ||
238 | @deffn {Scheme Procedure} current-thread | |
239 | @deffnx {C Function} scm_current_thread () | |
240 | Return the thread that called this function. | |
241 | @end deffn | |
07d83abe MV |
242 | |
243 | @c begin (texi-doc-string "guile" "call-with-new-thread") | |
b4fddbbe MV |
244 | @deffn {Scheme Procedure} call-with-new-thread thunk handler |
245 | Call @code{thunk} in a new thread and with a new dynamic state, | |
246 | returning the new thread. The procedure @var{thunk} is called via | |
247 | @code{with-continuation-barrier}. | |
07d83abe | 248 | |
b4fddbbe MV |
249 | When @var{handler} is specified, then @var{thunk} is called from |
250 | within a @code{catch} with tag @code{#t} that has @var{handler} as its | |
251 | handler. This catch is established inside the continuation barrier. | |
07d83abe | 252 | |
b4fddbbe MV |
253 | Once @var{thunk} or @var{handler} returns, the return value is made |
254 | the @emph{exit value} of the thread and the thread is terminated. | |
07d83abe MV |
255 | @end deffn |
256 | ||
b4fddbbe MV |
257 | @deftypefn {C Function} SCM scm_spawn_thread (scm_t_catch_body body, void *body_data, scm_t_catch_handler handler, void *handler_data) |
258 | Call @var{body} in a new thread, passing it @var{body_data}, returning | |
259 | the new thread. The function @var{body} is called via | |
260 | @code{scm_c_with_continuation_barrier}. | |
261 | ||
262 | When @var{handler} is non-@code{NULL}, @var{body} is called via | |
263 | @code{scm_internal_catch} with tag @code{SCM_BOOL_T} that has | |
264 | @var{handler} and @var{handler_data} as the handler and its data. This | |
265 | catch is established inside the continuation barrier. | |
266 | ||
267 | Once @var{body} or @var{handler} returns, the return value is made the | |
268 | @emph{exit value} of the thread and the thread is terminated. | |
269 | @end deftypefn | |
270 | ||
07d83abe MV |
271 | @c begin (texi-doc-string "guile" "join-thread") |
272 | @deffn {Scheme Procedure} join-thread thread | |
b4fddbbe MV |
273 | Wait for @var{thread} to terminate and return its exit value. Threads |
274 | that have not been created with @code{call-with-new-thread} or | |
275 | @code{scm_spawn_thread} have an exit value of @code{#f}. | |
07d83abe MV |
276 | @end deffn |
277 | ||
cdf1ad3b MV |
278 | @deffn {Scheme Procedure} thread-exited? thread |
279 | @deffnx {C Function} scm_thread_exited_p (thread) | |
280 | Return @code{#t} iff @var{thread} has exited. | |
281 | @end deffn | |
282 | ||
07d83abe MV |
283 | @c begin (texi-doc-string "guile" "yield") |
284 | @deffn {Scheme Procedure} yield | |
285 | If one or more threads are waiting to execute, calling yield forces an | |
286 | immediate context switch to one of them. Otherwise, yield has no effect. | |
287 | @end deffn | |
288 | ||
07d83abe MV |
289 | Higher level thread procedures are available by loading the |
290 | @code{(ice-9 threads)} module. These provide standardized | |
3cf066df | 291 | thread creation. |
07d83abe MV |
292 | |
293 | @deffn macro make-thread proc [args@dots{}] | |
294 | Apply @var{proc} to @var{args} in a new thread formed by | |
295 | @code{call-with-new-thread} using a default error handler that display | |
b4fddbbe MV |
296 | the error to the current error port. The @var{args@dots{}} |
297 | expressions are evaluated in the new thread. | |
07d83abe MV |
298 | @end deffn |
299 | ||
300 | @deffn macro begin-thread first [rest@dots{}] | |
301 | Evaluate forms @var{first} and @var{rest} in a new thread formed by | |
302 | @code{call-with-new-thread} using a default error handler that display | |
303 | the error to the current error port. | |
304 | @end deffn | |
305 | ||
2567692a MV |
306 | @node Mutexes and Condition Variables |
307 | @subsection Mutexes and Condition Variables | |
308 | @cindex mutex | |
309 | @cindex condition variable | |
310 | ||
311 | A mutex is a thread synchronization object, it can be used by threads | |
312 | to control access to a shared resource. A mutex can be locked to | |
313 | indicate a resource is in use, and other threads can then block on the | |
314 | mutex to wait for the resource (or can just test and do something else | |
315 | if not available). ``Mutex'' is short for ``mutual exclusion''. | |
316 | ||
317 | There are two types of mutexes in Guile, ``standard'' and | |
318 | ``recursive''. They're created by @code{make-mutex} and | |
319 | @code{make-recursive-mutex} respectively, the operation functions are | |
320 | then common to both. | |
321 | ||
322 | Note that for both types of mutex there's no protection against a | |
323 | ``deadly embrace''. For instance if one thread has locked mutex A and | |
324 | is waiting on mutex B, but another thread owns B and is waiting on A, | |
325 | then an endless wait will occur (in the current implementation). | |
326 | Acquiring requisite mutexes in a fixed order (like always A before B) | |
327 | in all threads is one way to avoid such problems. | |
328 | ||
329 | @sp 1 | |
330 | @deffn {Scheme Procedure} make-mutex | |
331 | @deffnx {C Function} scm_make_mutex () | |
332 | Return a new standard mutex. It is initially unlocked. | |
333 | @end deffn | |
334 | ||
335 | @deffn {Scheme Procedure} make-recursive-mutex | |
336 | @deffnx {C Function} scm_make_recursive_mutex () | |
337 | Create a new recursive mutex. It is initialloy unlocked. | |
338 | @end deffn | |
339 | ||
340 | @deffn {Scheme Procedure} lock-mutex mutex | |
341 | @deffnx {C Function} scm_lock_mutex (mutex) | |
342 | Lock @var{mutex}. If the mutex is already locked by another thread | |
343 | then block and return only when @var{mutex} has been acquired. | |
344 | ||
345 | For standard mutexes (@code{make-mutex}), and error is signalled if | |
346 | the thread has itself already locked @var{mutex}. | |
347 | ||
348 | For a recursive mutex (@code{make-recursive-mutex}), if the thread has | |
349 | itself already locked @var{mutex}, then a further @code{lock-mutex} | |
350 | call increments the lock count. An additional @code{unlock-mutex} | |
351 | will be required to finally release. | |
352 | ||
353 | When a system async (@pxref{System asyncs}) is activated for a thread | |
354 | blocked in @code{lock-mutex}, the wait is interrupted and the async is | |
355 | executed. When the async returns, the wait resumes. | |
356 | @end deffn | |
357 | ||
358 | @deftypefn {C Function} void scm_frame_lock_mutex (SCM mutex) | |
359 | Arrange for @var{mutex} to be locked whenever the current frame is | |
360 | entered and to be unlocked when it is exited. | |
361 | @end deftypefn | |
362 | ||
363 | @deffn {Scheme Procedure} try-mutex mx | |
364 | @deffnx {C Function} scm_try_mutex (mx) | |
365 | Try to lock @var{mutex} as per @code{lock-mutex}. If @var{mutex} can | |
366 | be acquired immediately then this is done and the return is @code{#t}. | |
367 | If @var{mutex} is locked by some other thread then nothing is done and | |
368 | the return is @code{#f}. | |
369 | @end deffn | |
370 | ||
371 | @deffn {Scheme Procedure} unlock-mutex mutex | |
372 | @deffnx {C Function} scm_unlock_mutex (mutex) | |
373 | Unlock @var{mutex}. An error is signalled if @var{mutex} is not | |
374 | locked by the calling thread. | |
375 | @end deffn | |
376 | ||
377 | @deffn {Scheme Procedure} make-condition-variable | |
378 | @deffnx {C Function} scm_make_condition_variable () | |
379 | Return a new condition variable. | |
380 | @end deffn | |
381 | ||
382 | @deffn {Scheme Procedure} wait-condition-variable condvar mutex [time] | |
383 | @deffnx {C Function} scm_wait_condition_variable (condvar, mutex, time) | |
384 | Wait until @var{condvar} has been signalled. While waiting, | |
385 | @var{mutex} is atomically unlocked (as with @code{unlock-mutex}) and | |
386 | is locked again when this function returns. When @var{time} is given, | |
387 | it specifies a point in time where the waiting should be aborted. It | |
388 | can be either a integer as returned by @code{current-time} or a pair | |
389 | as returned by @code{gettimeofday}. When the waiting is aborted, | |
390 | @code{#f} is returned. When the condition variable has in fact been | |
391 | signalled, @code{#t} is returned. The mutex is re-locked in any case | |
392 | before @code{wait-condition-variable} returns. | |
393 | ||
394 | When a system async is activated for a thread that is blocked in a | |
395 | call to @code{wait-condition-variable}, the waiting is interrupted, | |
396 | the mutex is locked, and the async is executed. When the async | |
397 | returns, the mutex is unlocked again and the waiting is resumed. When | |
398 | the thread block while re-acquiring the mutex, execution of asyncs is | |
399 | blocked. | |
400 | @end deffn | |
401 | ||
402 | @deffn {Scheme Procedure} signal-condition-variable condvar | |
403 | @deffnx {C Function} scm_signal_condition_variable (condvar) | |
404 | Wake up one thread that is waiting for @var{condvar}. | |
405 | @end deffn | |
406 | ||
407 | @deffn {Scheme Procedure} broadcast-condition-variable condvar | |
408 | @deffnx {C Function} scm_broadcast_condition_variable (condvar) | |
409 | Wake up all threads that are waiting for @var{condvar}. | |
410 | @end deffn | |
411 | ||
412 | @sp 1 | |
413 | The following are higher level operations on mutexes. These are | |
414 | available from | |
415 | ||
416 | @example | |
417 | (use-modules (ice-9 threads)) | |
418 | @end example | |
419 | ||
420 | @deffn macro with-mutex mutex [body@dots{}] | |
421 | Lock @var{mutex}, evaluate the @var{body} forms, then unlock | |
422 | @var{mutex}. The return value is the return from the last @var{body} | |
423 | form. | |
424 | ||
425 | The lock, body and unlock form the branches of a @code{dynamic-wind} | |
426 | (@pxref{Dynamic Wind}), so @var{mutex} is automatically unlocked if an | |
427 | error or new continuation exits @var{body}, and is re-locked if | |
428 | @var{body} is re-entered by a captured continuation. | |
429 | @end deffn | |
430 | ||
431 | @deffn macro monitor body@dots{} | |
432 | Evaluate the @var{body} forms, with a mutex locked so only one thread | |
433 | can execute that code at any one time. The return value is the return | |
434 | from the last @var{body} form. | |
435 | ||
436 | Each @code{monitor} form has its own private mutex and the locking and | |
437 | evaluation is as per @code{with-mutex} above. A standard mutex | |
438 | (@code{make-mutex}) is used, which means @var{body} must not | |
439 | recursively re-enter the @code{monitor} form. | |
440 | ||
441 | The term ``monitor'' comes from operating system theory, where it | |
442 | means a particular bit of code managing access to some resource and | |
443 | which only ever executes on behalf of one process at any one time. | |
444 | @end deffn | |
445 | ||
446 | ||
b4fddbbe MV |
447 | @node Blocking |
448 | @subsection Blocking in Guile Mode | |
07d83abe | 449 | |
b4fddbbe MV |
450 | A thread must not block outside of a libguile function while it is in |
451 | guile mode. The following functions can be used to temporily leave | |
452 | guile mode or to perform some common blocking operations in a supported | |
453 | way. | |
07d83abe | 454 | |
54428bb8 MV |
455 | @deftypefn {C Function} {void *} scm_without_guile (void *(*func) (void *), void *data) |
456 | Leave guile mode, call @var{func} on @var{data}, enter guile mode and | |
457 | return the result of calling @var{func}. | |
07d83abe | 458 | |
b4fddbbe | 459 | While a thread has left guile mode, it must not call any libguile |
54428bb8 MV |
460 | functions except @code{scm_with_guile} or @code{scm_without_guile} and |
461 | must not use any libguile macros. Also, local variables of type | |
462 | @code{SCM} that are allocated while not in guile mode are not | |
463 | protected from the garbage collector. | |
464 | ||
465 | When used from non-guile mode, calling @code{scm_without_guile} is | |
466 | still allowed: it simply calls @var{func}. In that way, you can leave | |
467 | guile mode without having to know whether the current thread is in | |
468 | guile mode or not. | |
07d83abe MV |
469 | @end deftypefn |
470 | ||
b4fddbbe MV |
471 | @deftypefn {C Function} int scm_pthread_mutex_lock (pthread_mutex_t *mutex) |
472 | Like @code{pthread_mutex_lock}, but leaves guile mode while waiting for | |
473 | the mutex. | |
07d83abe MV |
474 | @end deftypefn |
475 | ||
b4fddbbe MV |
476 | @deftypefn {C Function} int scm_pthread_cond_wait (pthread_cond_t *cond, pthread_mutex_t *mutex) |
477 | @deftypefnx {C Function} int scm_pthread_cond_timedwait (pthread_cond_t *cond, pthread_mutex_t *mutex, struct timespec *abstime) | |
478 | Like @code{pthread_cond_wait} and @code{pthread_cond_timedwait}, but | |
479 | leaves guile mode while waiting for the condition variable. | |
07d83abe MV |
480 | @end deftypefn |
481 | ||
b4fddbbe MV |
482 | @deftypefn {C Function} int scm_std_select (int nfds, fd_set *readfds, fd_set *writefds, fd_set *exceptfds, struct timeval *timeout) |
483 | Like @code{select} but leaves guile mode while waiting. Also, the | |
484 | delivery of a system async causes this function to be interrupted with | |
485 | error code @code{EINTR}. | |
07d83abe MV |
486 | @end deftypefn |
487 | ||
b4fddbbe MV |
488 | @deftypefn {C Function} {unsigned int} scm_std_sleep ({unsigned int} seconds) |
489 | Like @code{sleep}, but leaves guile mode while sleeping. Also, the | |
490 | delivery of a system async causes this function to be interrupted. | |
07d83abe MV |
491 | @end deftypefn |
492 | ||
b4fddbbe MV |
493 | @deftypefn {C Function} {unsigned long} scm_std_usleep ({unsigned long} usecs) |
494 | Like @code{usleep}, but leaves guile mode while sleeping. Also, the | |
495 | delivery of a system async causes this function to be interrupted. | |
07d83abe MV |
496 | @end deftypefn |
497 | ||
07d83abe | 498 | |
2567692a MV |
499 | @node Critical Sections |
500 | @subsection Critical Sections | |
501 | ||
502 | @deffn {C Macro} SCM_CRITICAL_SECTION_START | |
503 | @deffnx {C Macro} SCM_CRITICAL_SECTION_END | |
504 | These two macros can be used to delimit a critical section. | |
505 | Syntactically, they are both statements and need to be followed | |
506 | immediately by a semicolon. | |
507 | ||
508 | Executing @code{SCM_CRITICAL_SECTION_START} will lock a recursive | |
509 | mutex and block the executing of system asyncs. Executing | |
510 | @code{SCM_CRITICAL_SECTION_END} will unblock the execution of system | |
511 | asyncs and unlock the mutex. Thus, the code that executes between | |
512 | these two macros can only be executed in one thread at any one time | |
513 | and no system asyncs will run. However, because the mutex is a | |
514 | recursive one, the code might still be reentered by the same thread. | |
515 | You must either allow for this or avoid it, both by careful coding. | |
516 | ||
517 | On the other hand, critical sections delimited with these macros can | |
518 | be nested since the mutex is recursive. | |
519 | ||
520 | You must make sure that for each @code{SCM_CRITICAL_SECTION_START}, | |
521 | the corresponding @code{SCM_CRITICAL_SECTION_END} is always executed. | |
522 | This means that no non-local exit (such as a signalled error) might | |
523 | happen, for example. | |
524 | @end deffn | |
525 | ||
526 | @deftypefn {C Function} void scm_frame_critical_section (SCM mutex) | |
527 | Call @code{scm_frame_lock_mutex} on @var{mutex} and call | |
528 | @code{scm_frame_block_asyncs}. When @var{mutex} is false, a recursive | |
529 | mutex provided by Guile is used instead. | |
530 | ||
531 | The effect of a call to @code{scm_frame_critical_section} is that the | |
532 | current frame (@pxref{Frames}) turns into a critical section. Because | |
533 | of the locked mutex, no second thread can enter it concurrently and | |
534 | because of the blocked asyncs, no system async can reenter it from the | |
535 | current thread. | |
536 | ||
537 | When the current thread reenters the critical section anyway, the kind | |
538 | of @var{mutex} determines what happens: When @var{mutex} is recursive, | |
539 | the reentry is allowed. When it is a normal mutex, an error is | |
540 | signalled. | |
541 | @end deftypefn | |
542 | ||
543 | ||
b4fddbbe MV |
544 | @node Fluids and Dynamic States |
545 | @subsection Fluids and Dynamic States | |
07d83abe MV |
546 | |
547 | @cindex fluids | |
548 | ||
b4fddbbe MV |
549 | A @emph{fluid} is an object that can store one value per @emph{dynamic |
550 | state}. Each thread has a current dynamic state, and when accessing a | |
551 | fluid, this current dynamic state is used to provide the actual value. | |
552 | In this way, fluids can be used for thread local storage, but they are | |
553 | in fact more flexible: dynamic states are objects of their own and can | |
554 | be made current for more than one thread at the same time, or only be | |
555 | made current temporarily, for example. | |
556 | ||
557 | Fluids can also be used to simulate the desirable effects of | |
558 | dynamically scoped variables. Dynamically scoped variables are useful | |
559 | when you want to set a variable to a value during some dynamic extent | |
560 | in the execution of your program and have them revert to their | |
561 | original value when the control flow is outside of this dynamic | |
562 | extent. See the description of @code{with-fluids} below for details. | |
07d83abe MV |
563 | |
564 | New fluids are created with @code{make-fluid} and @code{fluid?} is | |
565 | used for testing whether an object is actually a fluid. The values | |
566 | stored in a fluid can be accessed with @code{fluid-ref} and | |
567 | @code{fluid-set!}. | |
568 | ||
569 | @deffn {Scheme Procedure} make-fluid | |
570 | @deffnx {C Function} scm_make_fluid () | |
571 | Return a newly created fluid. | |
b4fddbbe MV |
572 | Fluids are objects that can hold one |
573 | value per dynamic state. That is, modifications to this value are | |
574 | only visible to code that executes with the same dynamic state as | |
575 | the modifying code. When a new dynamic state is constructed, it | |
576 | inherits the values from its parent. Because each thread normally executes | |
577 | with its own dynamic state, you can use fluids for thread local storage. | |
07d83abe MV |
578 | @end deffn |
579 | ||
580 | @deffn {Scheme Procedure} fluid? obj | |
581 | @deffnx {C Function} scm_fluid_p (obj) | |
582 | Return @code{#t} iff @var{obj} is a fluid; otherwise, return | |
583 | @code{#f}. | |
584 | @end deffn | |
585 | ||
586 | @deffn {Scheme Procedure} fluid-ref fluid | |
587 | @deffnx {C Function} scm_fluid_ref (fluid) | |
588 | Return the value associated with @var{fluid} in the current | |
589 | dynamic root. If @var{fluid} has not been set, then return | |
590 | @code{#f}. | |
591 | @end deffn | |
592 | ||
593 | @deffn {Scheme Procedure} fluid-set! fluid value | |
594 | @deffnx {C Function} scm_fluid_set_x (fluid, value) | |
595 | Set the value associated with @var{fluid} in the current dynamic root. | |
596 | @end deffn | |
597 | ||
598 | @code{with-fluids*} temporarily changes the values of one or more fluids, | |
599 | so that the given procedure and each procedure called by it access the | |
600 | given values. After the procedure returns, the old values are restored. | |
601 | ||
cdf1ad3b MV |
602 | @deffn {Scheme Procedure} with-fluid* fluid value thunk |
603 | @deffnx {C Function} scm_with_fluid (fluid, value, thunk) | |
604 | Set @var{fluid} to @var{value} temporarily, and call @var{thunk}. | |
605 | @var{thunk} must be a procedure with no argument. | |
606 | @end deffn | |
607 | ||
07d83abe MV |
608 | @deffn {Scheme Procedure} with-fluids* fluids values thunk |
609 | @deffnx {C Function} scm_with_fluids (fluids, values, thunk) | |
610 | Set @var{fluids} to @var{values} temporary, and call @var{thunk}. | |
611 | @var{fluids} must be a list of fluids and @var{values} must be the | |
612 | same number of their values to be applied. Each substitution is done | |
613 | in the order given. @var{thunk} must be a procedure with no argument. | |
614 | it is called inside a @code{dynamic-wind} and the fluids are | |
615 | set/restored when control enter or leaves the established dynamic | |
616 | extent. | |
617 | @end deffn | |
618 | ||
619 | @deffn {Scheme Macro} with-fluids ((fluid value) ...) body... | |
620 | Execute @var{body...} while each @var{fluid} is set to the | |
621 | corresponding @var{value}. Both @var{fluid} and @var{value} are | |
622 | evaluated and @var{fluid} must yield a fluid. @var{body...} is | |
623 | executed inside a @code{dynamic-wind} and the fluids are set/restored | |
624 | when control enter or leaves the established dynamic extent. | |
625 | @end deffn | |
626 | ||
627 | @deftypefn {C Function} SCM scm_c_with_fluids (SCM fluids, SCM vals, SCM (*cproc)(void *), void *data) | |
628 | @deftypefnx {C Function} SCM scm_c_with_fluid (SCM fluid, SCM val, SCM (*cproc)(void *), void *data) | |
629 | The function @code{scm_c_with_fluids} is like @code{scm_with_fluids} | |
630 | except that it takes a C function to call instead of a Scheme thunk. | |
631 | ||
632 | The function @code{scm_c_with_fluid} is similar but only allows one | |
633 | fluid to be set instead of a list. | |
634 | @end deftypefn | |
635 | ||
636 | @deftypefn {C Function} void scm_frame_fluid (SCM fluid, SCM val) | |
637 | This function must be used inside a pair of calls to | |
638 | @code{scm_frame_begin} and @code{scm_frame_end} (@pxref{Frames}). | |
639 | During the dynamic extent of the frame, the fluid @var{fluid} is set | |
640 | to @var{val}. | |
641 | ||
642 | More precisely, the value of the fluid is swapped with a `backup' | |
643 | value whenever the frame is entered or left. The backup value is | |
644 | initialized with the @var{val} argument. | |
645 | @end deftypefn | |
646 | ||
b4fddbbe MV |
647 | @deffn {Scheme Procedure} make-dynamic-state [parent] |
648 | @deffnx {C Function} scm_make_dynamic_state (parent) | |
649 | Return a copy of the dynamic state object @var{parent} | |
650 | or of the current dynamic state when @var{parent} is omitted. | |
651 | @end deffn | |
652 | ||
653 | @deffn {Scheme Procedure} dynamic-state? obj | |
654 | @deffnx {C Function} scm_dynamic_state_p (obj) | |
655 | Return @code{#t} if @var{obj} is a dynamic state object; | |
656 | return @code{#f} otherwise. | |
657 | @end deffn | |
658 | ||
659 | @deftypefn {C Procedure} int scm_is_dynamic_state (SCM obj) | |
660 | Return non-zero if @var{obj} is a dynamic state object; | |
661 | return zero otherwise. | |
662 | @end deftypefn | |
663 | ||
664 | @deffn {Scheme Procedure} current-dynamic-state | |
665 | @deffnx {C Function} scm_current_dynamic_state () | |
666 | Return the current dynamic state object. | |
667 | @end deffn | |
668 | ||
669 | @deffn {Scheme Procedure} set-current-dynamic-state state | |
670 | @deffnx {C Function} scm_set_current_dynamic_state (state) | |
671 | Set the current dynamic state object to @var{state} | |
672 | and return the previous current dynamic state object. | |
673 | @end deffn | |
674 | ||
675 | @deffn {Scheme Procedure} with-dynamic-state state proc | |
676 | @deffnx {C Function} scm_with_dynamic_state (state, proc) | |
677 | Call @var{proc} while @var{state} is the current dynamic | |
678 | state object. | |
679 | @end deffn | |
680 | ||
681 | @deftypefn {C Procedure} void scm_frame_current_dynamic_state (SCM state) | |
682 | Set the current dynamic state to @var{state} for the dynamic extent of | |
683 | the current frame. | |
684 | @end deftypefn | |
685 | ||
c2110081 | 686 | @deftypefn {C Procedure} {void *} scm_c_with_dynamic_state (SCM state, void *(*func)(void *), void *data) |
b4fddbbe MV |
687 | Like @code{scm_with_dynamic_state}, but call @var{func} with |
688 | @var{data}. | |
689 | @end deftypefn | |
690 | ||
07d83abe MV |
691 | @node Futures |
692 | @subsection Futures | |
693 | @cindex futures | |
694 | ||
695 | Futures are a convenient way to run a calculation in a new thread, and | |
696 | only wait for the result when it's actually needed. | |
697 | ||
698 | Futures are similar to promises (@pxref{Delayed Evaluation}), in that | |
699 | they allow mainline code to continue immediately. But @code{delay} | |
700 | doesn't evaluate at all until forced, whereas @code{future} starts | |
701 | immediately in a new thread. | |
702 | ||
703 | @deffn {syntax} future expr | |
704 | Begin evaluating @var{expr} in a new thread, and return a ``future'' | |
705 | object representing the calculation. | |
706 | @end deffn | |
707 | ||
708 | @deffn {Scheme Procedure} make-future thunk | |
709 | @deffnx {C Function} scm_make_future (thunk) | |
710 | Begin evaluating the call @code{(@var{thunk})} in a new thread, and | |
711 | return a ``future'' object representing the calculation. | |
712 | @end deffn | |
713 | ||
714 | @deffn {Scheme Procedure} future-ref f | |
715 | @deffnx {C Function} scm_future_ref (f) | |
716 | Return the value computed by the future @var{f}. If @var{f} has not | |
717 | yet finished executing then wait for it to do so. | |
718 | @end deffn | |
719 | ||
720 | ||
721 | @node Parallel Forms | |
722 | @subsection Parallel forms | |
723 | @cindex parallel forms | |
724 | ||
725 | The functions described in this section are available from | |
726 | ||
727 | @example | |
728 | (use-modules (ice-9 threads)) | |
729 | @end example | |
730 | ||
731 | @deffn syntax parallel expr1 @dots{} exprN | |
af1323c5 | 732 | Evaluate each @var{expr} expression in parallel, each in its own thread. |
07d83abe MV |
733 | Return the results as a set of @var{N} multiple values |
734 | (@pxref{Multiple Values}). | |
735 | @end deffn | |
736 | ||
737 | @deffn syntax letpar ((var1 expr1) @dots{} (varN exprN)) body@dots{} | |
af1323c5 | 738 | Evaluate each @var{expr} in parallel, each in its own thread, then bind |
07d83abe MV |
739 | the results to the corresponding @var{var} variables and evaluate |
740 | @var{body}. | |
741 | ||
742 | @code{letpar} is like @code{let} (@pxref{Local Bindings}), but all the | |
743 | expressions for the bindings are evaluated in parallel. | |
744 | @end deffn | |
745 | ||
746 | @deffn {Scheme Procedure} par-map proc lst1 @dots{} lstN | |
747 | @deffnx {Scheme Procedure} par-for-each proc lst1 @dots{} lstN | |
748 | Call @var{proc} on the elements of the given lists. @code{par-map} | |
749 | returns a list comprising the return values from @var{proc}. | |
750 | @code{par-for-each} returns an unspecified value, but waits for all | |
751 | calls to complete. | |
752 | ||
753 | The @var{proc} calls are @code{(@var{proc} @var{elem1} @dots{} | |
754 | @var{elemN})}, where each @var{elem} is from the corresponding | |
755 | @var{lst}. Each @var{lst} must be the same length. The calls are | |
af1323c5 | 756 | made in parallel, each in its own thread. |
07d83abe MV |
757 | |
758 | These functions are like @code{map} and @code{for-each} (@pxref{List | |
759 | Mapping}), but make their @var{proc} calls in parallel. | |
760 | @end deffn | |
761 | ||
762 | @deffn {Scheme Procedure} n-par-map n proc lst1 @dots{} lstN | |
763 | @deffnx {Scheme Procedure} n-par-for-each n proc lst1 @dots{} lstN | |
764 | Call @var{proc} on the elements of the given lists, in the same way as | |
765 | @code{par-map} and @code{par-for-each} above, but use no more than | |
af1323c5 | 766 | @var{n} threads at any one time. The order in which calls are |
07d83abe MV |
767 | initiated within that threads limit is unspecified. |
768 | ||
769 | These functions are good for controlling resource consumption if | |
770 | @var{proc} calls might be costly, or if there are many to be made. On | |
771 | a dual-CPU system for instance @math{@var{n}=4} might be enough to | |
772 | keep the CPUs utilized, and not consume too much memory. | |
773 | @end deffn | |
774 | ||
775 | @deffn {Scheme Procedure} n-for-each-par-map n sproc pproc lst1 @dots{} lstN | |
776 | Apply @var{pproc} to the elements of the given lists, and apply | |
777 | @var{sproc} to each result returned by @var{pproc}. The final return | |
778 | value is unspecified, but all calls will have been completed before | |
779 | returning. | |
780 | ||
781 | The calls made are @code{(@var{sproc} (@var{pproc} @var{elem1} @dots{} | |
782 | @var{elemN}))}, where each @var{elem} is from the corresponding | |
783 | @var{lst}. Each @var{lst} must have the same number of elements. | |
784 | ||
af1323c5 KR |
785 | The @var{pproc} calls are made in parallel, in separate threads. No more |
786 | than @var{n} threads are used at any one time. The order in which | |
07d83abe MV |
787 | @var{pproc} calls are initiated within that limit is unspecified. |
788 | ||
789 | The @var{sproc} calls are made serially, in list element order, one at | |
790 | a time. @var{pproc} calls on later elements may execute in parallel | |
791 | with the @var{sproc} calls. Exactly which thread makes each | |
792 | @var{sproc} call is unspecified. | |
793 | ||
794 | This function is designed for individual calculations that can be done | |
795 | in parallel, but with results needing to be handled serially, for | |
796 | instance to write them to a file. The @var{n} limit on threads | |
797 | controls system resource usage when there are many calculations or | |
798 | when they might be costly. | |
799 | ||
800 | It will be seen that @code{n-for-each-par-map} is like a combination | |
801 | of @code{n-par-map} and @code{for-each}, | |
802 | ||
803 | @example | |
af1323c5 | 804 | (for-each sproc (n-par-map n pproc lst1 ... lstN)) |
07d83abe MV |
805 | @end example |
806 | ||
807 | @noindent | |
808 | But the actual implementation is more efficient since each @var{sproc} | |
809 | call, in turn, can be initiated once the relevant @var{pproc} call has | |
810 | completed, it doesn't need to wait for all to finish. | |
811 | @end deffn | |
812 | ||
813 | ||
3cf066df | 814 | |
07d83abe MV |
815 | @c Local Variables: |
816 | @c TeX-master: "guile.texi" | |
817 | @c End: |