2 @comment node-name, next, previous, up
5 SBCL supports a fairly low-level threading interface that maps onto
6 the host operating system's concept of threads or lightweight
7 processes. This means that threads may take advantage of hardware
8 multiprocessing on machines that have more than one CPU, but it does
9 not allow Lisp control of the scheduler. This is found in the
12 Threads are part of the default build on x86[-64] Linux only.
14 They are also experimentally supported on: x86[-64] Darwin (Mac OS X),
15 x86[-64] FreeBSD, x86 SunOS (Solaris), and PPC Linux. On these platforms
16 threads must be explicitly enabled at build-time, see @file{INSTALL} for
25 * Waitqueue/condition variables::
27 * Sessions/Debugging::
29 * Implementation (Linux x86/x86-64)::
32 @node Threading basics
33 @comment node-name, next, previous, up
34 @section Threading basics
37 (make-thread (lambda () (write-line "Hello, world")))
40 @subsection Thread Objects
42 @include struct-sb-thread-thread.texinfo
43 @include var-sb-thread-star-current-thread-star.texinfo
44 @include fun-sb-thread-list-all-threads.texinfo
45 @include fun-sb-thread-thread-alive-p.texinfo
46 @include fun-sb-thread-thread-name.texinfo
47 @include fun-sb-thread-main-thread-p.texinfo
48 @include fun-sb-thread-main-thread.texinfo
50 @subsection Making, Returning From, Joining, and Yielding Threads
52 @include fun-sb-thread-make-thread.texinfo
53 @include macro-sb-thread-return-from-thread.texinfo
54 @include fun-sb-thread-abort-thread.texinfo
55 @include fun-sb-thread-join-thread.texinfo
56 @include fun-sb-thread-thread-yield.texinfo
58 @subsection Asynchronous Operations
60 @include fun-sb-thread-interrupt-thread.texinfo
61 @include fun-sb-thread-terminate-thread.texinfo
63 @subsection Miscellaneous Operations
65 @include fun-sb-thread-symbol-value-in-thread.texinfo
67 @subsection Error Conditions
69 @include condition-sb-thread-thread-error.texinfo
70 @include fun-sb-thread-thread-error-thread.texinfo
72 @c @include condition-sb-thread-symbol-value-in-thread-error.texinfo
73 @include condition-sb-thread-interrupt-thread-error.texinfo
74 @include condition-sb-thread-join-thread-error.texinfo
76 @node Special Variables
77 @comment node-name, next, previous, up
78 @section Special Variables
80 The interaction of special variables with multiple threads is mostly
81 as one would expect, with behaviour very similar to other
86 global special values are visible across all threads;
88 bindings (e.g. using LET) are local to the thread;
90 threads do not inherit dynamic bindings from the parent thread
93 The last point means that
98 (sb-thread:make-thread (lambda () (print *x*))))
101 prints @code{0} and not @code{1} as of 0.9.6.
103 @node Atomic Operations
104 @comment node-name, next, previous, up
105 @section Atomic Operations
107 Following atomic operations are particularly useful for implementing
110 @include macro-sb-ext-atomic-decf.texinfo
111 @include macro-sb-ext-atomic-incf.texinfo
112 @include macro-sb-ext-atomic-update.texinfo
113 @include macro-sb-ext-compare-and-swap.texinfo
115 @unnumberedsubsec CAS Protocol
117 Our @code{compare-and-swap} is user-extensible using a protocol
118 similar to @code{setf}, allowing users to add CAS support to new
119 places via eg. @code{defcas}.
121 At the same time, new atomic operations can be built on top of CAS
122 using @code{get-cas-expansion}. See @code{atomic-update} for an
125 @include macro-sb-ext-cas.texinfo
126 @include macro-sb-ext-define-cas-expander.texinfo
127 @include macro-sb-ext-defcas.texinfo
128 @include fun-sb-ext-get-cas-expansion.texinfo
131 @comment node-name, next, previous, up
132 @section Mutex Support
134 Mutexes are used for controlling access to a shared resource. One
135 thread is allowed to hold the mutex, others which attempt to take it
136 will be made to wait until it's free. Threads are woken in the order
137 that they go to sleep.
139 There isn't a timeout on mutex acquisition, but the usual WITH-TIMEOUT
140 macro (which throws a TIMEOUT condition after n seconds) can be used
141 if you want a bounded wait.
144 (defpackage :demo (:use "CL" "SB-THREAD" "SB-EXT"))
148 (defvar *a-mutex* (make-mutex :name "my lock"))
151 (format t "Thread ~A running ~%" *current-thread*)
152 (with-mutex (*a-mutex*)
153 (format t "Thread ~A got the lock~%" *current-thread*)
155 (format t "Thread ~A dropped lock, dying now~%" *current-thread*))
157 (make-thread #'thread-fn)
158 (make-thread #'thread-fn)
161 @include struct-sb-thread-mutex.texinfo
162 @include fun-sb-thread-make-mutex.texinfo
163 @include fun-sb-thread-mutex-name.texinfo
164 @include fun-sb-thread-mutex-value.texinfo
165 @include fun-sb-thread-grab-mutex.texinfo
166 @include fun-sb-thread-release-mutex.texinfo
167 @include macro-sb-thread-with-mutex.texinfo
168 @include macro-sb-thread-with-recursive-lock.texinfo
169 @include fun-sb-thread-get-mutex.texinfo
172 @comment node-name, next, previous, up
175 Semaphores are among other things useful for keeping track of a
176 countable resource, eg. messages in a queue, and sleep when the
177 resource is exhausted.
179 @include struct-sb-thread-semaphore.texinfo
180 @include fun-sb-thread-make-semaphore.texinfo
181 @include fun-sb-thread-signal-semaphore.texinfo
182 @include fun-sb-thread-wait-on-semaphore.texinfo
183 @include fun-sb-thread-try-semaphore.texinfo
184 @include fun-sb-thread-semaphore-count.texinfo
185 @include fun-sb-thread-semaphore-name.texinfo
187 @include struct-sb-thread-semaphore-notification.texinfo
188 @include fun-sb-thread-make-semaphore-notification.texinfo
189 @include fun-sb-thread-semaphore-notification-status.texinfo
190 @include fun-sb-thread-clear-semaphore-notification.texinfo
192 @node Waitqueue/condition variables
193 @comment node-name, next, previous, up
194 @section Waitqueue/condition variables
196 These are based on the POSIX condition variable design, hence the
197 annoyingly CL-conflicting name. For use when you want to check a
198 condition and sleep until it's true. For example: you have a shared
199 queue, a writer process checking ``queue is empty'' and one or more
200 readers that need to know when ``queue is not empty''. It sounds
201 simple, but is astonishingly easy to deadlock if another process runs
202 when you weren't expecting it to.
204 There are three components:
208 the condition itself (not represented in code)
211 the condition variable (a.k.a waitqueue) which proxies for it
214 a lock to hold while testing the condition
217 Important stuff to be aware of:
221 when calling condition-wait, you must hold the mutex. condition-wait
222 will drop the mutex while it waits, and obtain it again before
223 returning for whatever reason;
226 likewise, you must be holding the mutex around calls to
230 a process may return from condition-wait in several circumstances: it
231 is not guaranteed that the underlying condition has become true. You
232 must check that the resource is ready for whatever you want to do to
238 (defvar *buffer-queue* (make-waitqueue))
239 (defvar *buffer-lock* (make-mutex :name "buffer lock"))
241 (defvar *buffer* (list nil))
244 (with-mutex (*buffer-lock*)
246 (condition-wait *buffer-queue* *buffer-lock*)
248 (unless *buffer* (return))
249 (let ((head (car *buffer*)))
250 (setf *buffer* (cdr *buffer*))
251 (format t "reader ~A woke, read ~A~%"
252 *current-thread* head))))))
257 (with-mutex (*buffer-lock*)
260 (+ (char-code #\A) (random 26)))))))
261 (setf *buffer* (cons el *buffer*)))
262 (condition-notify *buffer-queue*))))
264 (make-thread #'writer)
265 (make-thread #'reader)
266 (make-thread #'reader)
269 @include struct-sb-thread-waitqueue.texinfo
270 @include fun-sb-thread-make-waitqueue.texinfo
271 @include fun-sb-thread-waitqueue-name.texinfo
272 @include fun-sb-thread-condition-wait.texinfo
273 @include fun-sb-thread-condition-notify.texinfo
274 @include fun-sb-thread-condition-broadcast.texinfo
277 @comment node-name, next, previous, up
280 These are based on the Linux kernel barrier design, which is in turn
281 based on the Alpha CPU memory model. They are presently implemented for
282 x86, x86-64, and PPC systems, and behave as compiler barriers on all
285 In addition to explicit use of the @code{sb-thread:barrier} macro, the
286 following functions and macros also serve as @code{:memory} barriers:
290 @code{sb-ext:atomic-decf} and @code{sb-ext:atomic-incf}.
292 @code{sb-ext:compare-and-swap}.
294 @code{sb-thread:get-mutex}, @code{sb-thread:release-mutex},
295 @code{sb-thread:with-mutex} and @code{sb-thread:with-recursive-lock}.
297 @code{sb-thread:signal-semaphore}, @code{sb-thread:try-semaphore} and
298 @code{sb-thread:wait-on-semaphore}.
300 @code{sb-thread:condition-wait}, @code{sb-thread:condition-notify} and
301 @code{sb-thread:condition-broadcast}.
304 @include macro-sb-thread-barrier.texinfo
306 @node Sessions/Debugging
307 @comment node-name, next, previous, up
308 @section Sessions/Debugging
310 If the user has multiple views onto the same Lisp image (for example,
311 using multiple terminals, or a windowing system, or network access)
312 they are typically set up as multiple @dfn{sessions} such that each
313 view has its own collection of foreground/background/stopped threads.
314 A thread which wishes to create a new session can use
315 @code{sb-thread:with-new-session} to remove itself from the current
316 session (which it shares with its parent and siblings) and create a
318 # See also @code{sb-thread:make-listener-thread}.
320 Within a single session, threads arbitrate between themselves for the
321 user's attention. A thread may be in one of three notional states:
322 foreground, background, or stopped. When a background process
323 attempts to print a repl prompt or to enter the debugger, it will stop
324 and print a message saying that it has stopped. The user at his
325 leisure may switch to that thread to find out what it needs. If a
326 background thread enters the debugger, selecting any restart will put
327 it back into the background before it resumes. Arbitration for the
328 input stream is managed by calls to @code{sb-thread:get-foreground}
329 (which may block) and @code{sb-thread:release-foreground}.
331 @node Foreign threads
332 @comment node-name, next, previous, up
333 @section Foreign threads
335 Direct calls to @code{pthread_create} (instead of @code{MAKE-THREAD})
336 create threads that SBCL is not aware of, these are called foreign
337 threads. Currently, it is not possible to run Lisp code in such
338 threads. This means that the Lisp side signal handlers cannot work.
339 The best solution is to start foreign threads with signals blocked,
340 but since third party libraries may create threads, it is not always
341 feasible to do so. As a workaround, upon receiving a signal in a
342 foreign thread, SBCL changes the thread's sigmask to block all signals
343 that it wants to handle and resends the signal to the current process
344 which should land in a thread that does not block it, that is, a Lisp
347 The resignalling trick cannot work for synchronously triggered signals
348 (SIGSEGV and co), take care not to trigger any. Resignalling for
349 synchronously triggered signals in foreign threads is subject to
350 @code{--lose-on-corruption}, see @ref{Runtime Options}.
352 @node Implementation (Linux x86/x86-64)
353 @comment node-name, next, previous, up
354 @section Implementation (Linux x86/x86-64)
356 Threading is implemented using pthreads and some Linux specific bits
359 On x86 the per-thread local bindings for special variables is achieved
360 using the %fs segment register to point to a per-thread storage area.
361 This may cause interesting results if you link to foreign code that
362 expects threading or creates new threads, and the thread library in
363 question uses %fs in an incompatible way. On x86-64 the r12 register
366 Queues require the @code{sys_futex()} system call to be available:
367 this is the reason for the NPTL requirement. We test at runtime that
368 this system call exists.
370 Garbage collection is done with the existing Conservative Generational
371 GC. Allocation is done in small (typically 8k) regions: each thread
372 has its own region so this involves no stopping. However, when a
373 region fills, a lock must be obtained while another is allocated, and
374 when a collection is required, all processes are stopped. This is
375 achieved by sending them signals, which may make for interesting
376 behaviour if they are interrupted in system calls. The streams
377 interface is believed to handle the required system call restarting
378 correctly, but this may be a consideration when making other blocking
379 calls e.g. from foreign library code.
381 Large amounts of the SBCL library have not been inspected for
382 thread-safety. Some of the obviously unsafe areas have large locks
383 around them, so compilation and fasl loading, for example, cannot be
384 parallelized. Work is ongoing in this area.
386 A new thread by default is created in the same POSIX process group and
387 session as the thread it was created by. This has an impact on
388 keyboard interrupt handling: pressing your terminal's intr key
389 (typically @kbd{Control-C}) will interrupt all processes in the
390 foreground process group, including Lisp threads that SBCL considers
391 to be notionally `background'. This is undesirable, so background
392 threads are set to ignore the SIGINT signal.
394 @code{sb-thread:make-listener-thread} in addition to creating a new
395 Lisp session makes a new POSIX session, so that pressing
396 @kbd{Control-C} in one window will not interrupt another listener -
397 this has been found to be embarrassing.