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
48 @subsection Making, Joining, and Yielding Threads
50 @include fun-sb-thread-make-thread.texinfo
51 @include fun-sb-thread-thread-yield.texinfo
52 @include fun-sb-thread-join-thread.texinfo
54 @subsection Asynchronous Operations
56 @include fun-sb-thread-interrupt-thread.texinfo
57 @include fun-sb-thread-terminate-thread.texinfo
59 @subsection Miscellaneous Operations
61 @include fun-sb-thread-symbol-value-in-thread.texinfo
63 @subsection Error Conditions
65 @include condition-sb-thread-thread-error.texinfo
66 @include fun-sb-thread-thread-error-thread.texinfo
68 @c @include condition-sb-thread-symbol-value-in-thread-error.texinfo
69 @include condition-sb-thread-interrupt-thread-error.texinfo
70 @include condition-sb-thread-join-thread-error.texinfo
72 @node Special Variables
73 @comment node-name, next, previous, up
74 @section Special Variables
76 The interaction of special variables with multiple threads is mostly
77 as one would expect, with behaviour very similar to other
82 global special values are visible across all threads;
84 bindings (e.g. using LET) are local to the thread;
86 threads do not inherit dynamic bindings from the parent thread
89 The last point means that
94 (sb-thread:make-thread (lambda () (print *x*))))
97 prints @code{0} and not @code{1} as of 0.9.6.
99 @node Atomic Operations
100 @comment node-name, next, previous, up
101 @section Atomic Operations
103 SBCL provides a few special purpose atomic operations, particularly
104 useful for implementing lockless algorithms.
106 @include macro-sb-ext-atomic-decf.texinfo
107 @include macro-sb-ext-atomic-incf.texinfo
108 @include macro-sb-ext-compare-and-swap.texinfo
111 @comment node-name, next, previous, up
112 @section Mutex Support
114 Mutexes are used for controlling access to a shared resource. One
115 thread is allowed to hold the mutex, others which attempt to take it
116 will be made to wait until it's free. Threads are woken in the order
117 that they go to sleep.
119 There isn't a timeout on mutex acquisition, but the usual WITH-TIMEOUT
120 macro (which throws a TIMEOUT condition after n seconds) can be used
121 if you want a bounded wait.
124 (defpackage :demo (:use "CL" "SB-THREAD" "SB-EXT"))
128 (defvar *a-mutex* (make-mutex :name "my lock"))
131 (format t "Thread ~A running ~%" *current-thread*)
132 (with-mutex (*a-mutex*)
133 (format t "Thread ~A got the lock~%" *current-thread*)
135 (format t "Thread ~A dropped lock, dying now~%" *current-thread*))
137 (make-thread #'thread-fn)
138 (make-thread #'thread-fn)
141 @include struct-sb-thread-mutex.texinfo
142 @include fun-sb-thread-make-mutex.texinfo
143 @include fun-sb-thread-mutex-name.texinfo
144 @include fun-sb-thread-mutex-value.texinfo
145 @include fun-sb-thread-get-mutex.texinfo
146 @include fun-sb-thread-release-mutex.texinfo
147 @include macro-sb-thread-with-mutex.texinfo
148 @include macro-sb-thread-with-recursive-lock.texinfo
151 @comment node-name, next, previous, up
154 described here should be considered
155 experimental, subject to API changes without notice.
157 @include struct-sb-thread-semaphore.texinfo
158 @include fun-sb-thread-make-semaphore.texinfo
159 @include fun-sb-thread-semaphore-count.texinfo
160 @include fun-sb-thread-semaphore-name.texinfo
161 @include fun-sb-thread-signal-semaphore.texinfo
162 @include fun-sb-thread-try-semaphore.texinfo
163 @include fun-sb-thread-wait-on-semaphore.texinfo
165 @node Waitqueue/condition variables
166 @comment node-name, next, previous, up
167 @section Waitqueue/condition variables
169 These are based on the POSIX condition variable design, hence the
170 annoyingly CL-conflicting name. For use when you want to check a
171 condition and sleep until it's true. For example: you have a shared
172 queue, a writer process checking ``queue is empty'' and one or more
173 readers that need to know when ``queue is not empty''. It sounds
174 simple, but is astonishingly easy to deadlock if another process runs
175 when you weren't expecting it to.
177 There are three components:
181 the condition itself (not represented in code)
184 the condition variable (a.k.a waitqueue) which proxies for it
187 a lock to hold while testing the condition
190 Important stuff to be aware of:
194 when calling condition-wait, you must hold the mutex. condition-wait
195 will drop the mutex while it waits, and obtain it again before
196 returning for whatever reason;
199 likewise, you must be holding the mutex around calls to
203 a process may return from condition-wait in several circumstances: it
204 is not guaranteed that the underlying condition has become true. You
205 must check that the resource is ready for whatever you want to do to
211 (defvar *buffer-queue* (make-waitqueue))
212 (defvar *buffer-lock* (make-mutex :name "buffer lock"))
214 (defvar *buffer* (list nil))
217 (with-mutex (*buffer-lock*)
219 (condition-wait *buffer-queue* *buffer-lock*)
221 (unless *buffer* (return))
222 (let ((head (car *buffer*)))
223 (setf *buffer* (cdr *buffer*))
224 (format t "reader ~A woke, read ~A~%"
225 *current-thread* head))))))
230 (with-mutex (*buffer-lock*)
233 (+ (char-code #\A) (random 26)))))))
234 (setf *buffer* (cons el *buffer*)))
235 (condition-notify *buffer-queue*))))
237 (make-thread #'writer)
238 (make-thread #'reader)
239 (make-thread #'reader)
242 @include struct-sb-thread-waitqueue.texinfo
243 @include fun-sb-thread-make-waitqueue.texinfo
244 @include fun-sb-thread-waitqueue-name.texinfo
245 @include fun-sb-thread-condition-wait.texinfo
246 @include fun-sb-thread-condition-notify.texinfo
247 @include fun-sb-thread-condition-broadcast.texinfo
250 @comment node-name, next, previous, up
253 These are based on the Linux kernel barrier design, which is in turn
254 based on the Alpha CPU memory model. They are presently implemented for
255 x86, x86-64, and PPC systems, and behave as compiler barriers on all
258 In addition to explicit use of the @code{sb-thread:barrier} macro, the
259 following functions and macros also serve as @code{:memory} barriers:
263 @code{sb-ext:atomic-decf} and @code{sb-ext:atomic-incf}.
265 @code{sb-ext:compare-and-swap}.
267 @code{sb-thread:get-mutex}, @code{sb-thread:release-mutex},
268 @code{sb-thread:with-mutex} and @code{sb-thread:with-recursive-lock}.
270 @code{sb-thread:signal-semaphore}, @code{sb-thread:try-semaphore} and
271 @code{sb-thread:wait-on-semaphore}.
273 @code{sb-thread:condition-wait}, @code{sb-thread:condition-notify} and
274 @code{sb-thread:condition-broadcast}.
277 @include macro-sb-thread-barrier.texinfo
279 @node Sessions/Debugging
280 @comment node-name, next, previous, up
281 @section Sessions/Debugging
283 If the user has multiple views onto the same Lisp image (for example,
284 using multiple terminals, or a windowing system, or network access)
285 they are typically set up as multiple @dfn{sessions} such that each
286 view has its own collection of foreground/background/stopped threads.
287 A thread which wishes to create a new session can use
288 @code{sb-thread:with-new-session} to remove itself from the current
289 session (which it shares with its parent and siblings) and create a
291 # See also @code{sb-thread:make-listener-thread}.
293 Within a single session, threads arbitrate between themselves for the
294 user's attention. A thread may be in one of three notional states:
295 foreground, background, or stopped. When a background process
296 attempts to print a repl prompt or to enter the debugger, it will stop
297 and print a message saying that it has stopped. The user at his
298 leisure may switch to that thread to find out what it needs. If a
299 background thread enters the debugger, selecting any restart will put
300 it back into the background before it resumes. Arbitration for the
301 input stream is managed by calls to @code{sb-thread:get-foreground}
302 (which may block) and @code{sb-thread:release-foreground}.
304 @code{sb-ext:quit} terminates all threads in the current session, but
305 leaves other sessions running.
307 @node Foreign threads
308 @comment node-name, next, previous, up
309 @section Foreign threads
311 Direct calls to @code{pthread_create} (instead of @code{MAKE-THREAD})
312 create threads that SBCL is not aware of, these are called foreign
313 threads. Currently, it is not possible to run Lisp code in such
314 threads. This means that the Lisp side signal handlers cannot work.
315 The best solution is to start foreign threads with signals blocked,
316 but since third party libraries may create threads, it is not always
317 feasible to do so. As a workaround, upon receiving a signal in a
318 foreign thread, SBCL changes the thread's sigmask to block all signals
319 that it wants to handle and resends the signal to the current process
320 which should land in a thread that does not block it, that is, a Lisp
323 The resignalling trick cannot work for synchronously triggered signals
324 (SIGSEGV and co), take care not to trigger any. Resignalling for
325 synchronously triggered signals in foreign threads is subject to
326 @code{--lose-on-corruption}, see @ref{Runtime Options}.
328 @node Implementation (Linux x86/x86-64)
329 @comment node-name, next, previous, up
330 @section Implementation (Linux x86/x86-64)
332 Threading is implemented using pthreads and some Linux specific bits
335 On x86 the per-thread local bindings for special variables is achieved
336 using the %fs segment register to point to a per-thread storage area.
337 This may cause interesting results if you link to foreign code that
338 expects threading or creates new threads, and the thread library in
339 question uses %fs in an incompatible way. On x86-64 the r12 register
342 Queues require the @code{sys_futex()} system call to be available:
343 this is the reason for the NPTL requirement. We test at runtime that
344 this system call exists.
346 Garbage collection is done with the existing Conservative Generational
347 GC. Allocation is done in small (typically 8k) regions: each thread
348 has its own region so this involves no stopping. However, when a
349 region fills, a lock must be obtained while another is allocated, and
350 when a collection is required, all processes are stopped. This is
351 achieved by sending them signals, which may make for interesting
352 behaviour if they are interrupted in system calls. The streams
353 interface is believed to handle the required system call restarting
354 correctly, but this may be a consideration when making other blocking
355 calls e.g. from foreign library code.
357 Large amounts of the SBCL library have not been inspected for
358 thread-safety. Some of the obviously unsafe areas have large locks
359 around them, so compilation and fasl loading, for example, cannot be
360 parallelized. Work is ongoing in this area.
362 A new thread by default is created in the same POSIX process group and
363 session as the thread it was created by. This has an impact on
364 keyboard interrupt handling: pressing your terminal's intr key
365 (typically @kbd{Control-C}) will interrupt all processes in the
366 foreground process group, including Lisp threads that SBCL considers
367 to be notionally `background'. This is undesirable, so background
368 threads are set to ignore the SIGINT signal.
370 @code{sb-thread:make-listener-thread} in addition to creating a new
371 Lisp session makes a new POSIX session, so that pressing
372 @kbd{Control-C} in one window will not interrupt another listener -
373 this has been found to be embarrassing.