Merge remote-tracking branch 'remotes/mjt/tags/trivial-patches-fetch' into staging
[qemu.git] / util / thread-pool.c
blobce6cd30193a8da51bd8b3a7f62bbaa9718918284
1 /*
2 * QEMU block layer thread pool
4 * Copyright IBM, Corp. 2008
5 * Copyright Red Hat, Inc. 2012
7 * Authors:
8 * Anthony Liguori <aliguori@us.ibm.com>
9 * Paolo Bonzini <pbonzini@redhat.com>
11 * This work is licensed under the terms of the GNU GPL, version 2. See
12 * the COPYING file in the top-level directory.
14 * Contributions after 2012-01-13 are licensed under the terms of the
15 * GNU GPL, version 2 or (at your option) any later version.
17 #include "qemu/osdep.h"
18 #include "qemu-common.h"
19 #include "qemu/queue.h"
20 #include "qemu/thread.h"
21 #include "qemu/coroutine.h"
22 #include "trace.h"
23 #include "block/thread-pool.h"
24 #include "qemu/main-loop.h"
26 static void do_spawn_thread(ThreadPool *pool);
28 typedef struct ThreadPoolElement ThreadPoolElement;
30 enum ThreadState {
31 THREAD_QUEUED,
32 THREAD_ACTIVE,
33 THREAD_DONE,
36 struct ThreadPoolElement {
37 BlockAIOCB common;
38 ThreadPool *pool;
39 ThreadPoolFunc *func;
40 void *arg;
42 /* Moving state out of THREAD_QUEUED is protected by lock. After
43 * that, only the worker thread can write to it. Reads and writes
44 * of state and ret are ordered with memory barriers.
46 enum ThreadState state;
47 int ret;
49 /* Access to this list is protected by lock. */
50 QTAILQ_ENTRY(ThreadPoolElement) reqs;
52 /* Access to this list is protected by the global mutex. */
53 QLIST_ENTRY(ThreadPoolElement) all;
56 struct ThreadPool {
57 AioContext *ctx;
58 QEMUBH *completion_bh;
59 QemuMutex lock;
60 QemuCond worker_stopped;
61 QemuSemaphore sem;
62 int max_threads;
63 QEMUBH *new_thread_bh;
65 /* The following variables are only accessed from one AioContext. */
66 QLIST_HEAD(, ThreadPoolElement) head;
68 /* The following variables are protected by lock. */
69 QTAILQ_HEAD(, ThreadPoolElement) request_list;
70 int cur_threads;
71 int idle_threads;
72 int new_threads; /* backlog of threads we need to create */
73 int pending_threads; /* threads created but not running yet */
74 bool stopping;
77 static void *worker_thread(void *opaque)
79 ThreadPool *pool = opaque;
81 qemu_mutex_lock(&pool->lock);
82 pool->pending_threads--;
83 do_spawn_thread(pool);
85 while (!pool->stopping) {
86 ThreadPoolElement *req;
87 int ret;
89 do {
90 pool->idle_threads++;
91 qemu_mutex_unlock(&pool->lock);
92 ret = qemu_sem_timedwait(&pool->sem, 10000);
93 qemu_mutex_lock(&pool->lock);
94 pool->idle_threads--;
95 } while (ret == -1 && !QTAILQ_EMPTY(&pool->request_list));
96 if (ret == -1 || pool->stopping) {
97 break;
100 req = QTAILQ_FIRST(&pool->request_list);
101 QTAILQ_REMOVE(&pool->request_list, req, reqs);
102 req->state = THREAD_ACTIVE;
103 qemu_mutex_unlock(&pool->lock);
105 ret = req->func(req->arg);
107 req->ret = ret;
108 /* Write ret before state. */
109 smp_wmb();
110 req->state = THREAD_DONE;
112 qemu_mutex_lock(&pool->lock);
114 qemu_bh_schedule(pool->completion_bh);
117 pool->cur_threads--;
118 qemu_cond_signal(&pool->worker_stopped);
119 qemu_mutex_unlock(&pool->lock);
120 return NULL;
123 static void do_spawn_thread(ThreadPool *pool)
125 QemuThread t;
127 /* Runs with lock taken. */
128 if (!pool->new_threads) {
129 return;
132 pool->new_threads--;
133 pool->pending_threads++;
135 qemu_thread_create(&t, "worker", worker_thread, pool, QEMU_THREAD_DETACHED);
138 static void spawn_thread_bh_fn(void *opaque)
140 ThreadPool *pool = opaque;
142 qemu_mutex_lock(&pool->lock);
143 do_spawn_thread(pool);
144 qemu_mutex_unlock(&pool->lock);
147 static void spawn_thread(ThreadPool *pool)
149 pool->cur_threads++;
150 pool->new_threads++;
151 /* If there are threads being created, they will spawn new workers, so
152 * we don't spend time creating many threads in a loop holding a mutex or
153 * starving the current vcpu.
155 * If there are no idle threads, ask the main thread to create one, so we
156 * inherit the correct affinity instead of the vcpu affinity.
158 if (!pool->pending_threads) {
159 qemu_bh_schedule(pool->new_thread_bh);
163 static void thread_pool_completion_bh(void *opaque)
165 ThreadPool *pool = opaque;
166 ThreadPoolElement *elem, *next;
168 aio_context_acquire(pool->ctx);
169 restart:
170 QLIST_FOREACH_SAFE(elem, &pool->head, all, next) {
171 if (elem->state != THREAD_DONE) {
172 continue;
175 trace_thread_pool_complete(pool, elem, elem->common.opaque,
176 elem->ret);
177 QLIST_REMOVE(elem, all);
179 if (elem->common.cb) {
180 /* Read state before ret. */
181 smp_rmb();
183 /* Schedule ourselves in case elem->common.cb() calls aio_poll() to
184 * wait for another request that completed at the same time.
186 qemu_bh_schedule(pool->completion_bh);
188 aio_context_release(pool->ctx);
189 elem->common.cb(elem->common.opaque, elem->ret);
190 aio_context_acquire(pool->ctx);
191 qemu_aio_unref(elem);
192 goto restart;
193 } else {
194 qemu_aio_unref(elem);
197 aio_context_release(pool->ctx);
200 static void thread_pool_cancel(BlockAIOCB *acb)
202 ThreadPoolElement *elem = (ThreadPoolElement *)acb;
203 ThreadPool *pool = elem->pool;
205 trace_thread_pool_cancel(elem, elem->common.opaque);
207 qemu_mutex_lock(&pool->lock);
208 if (elem->state == THREAD_QUEUED &&
209 /* No thread has yet started working on elem. we can try to "steal"
210 * the item from the worker if we can get a signal from the
211 * semaphore. Because this is non-blocking, we can do it with
212 * the lock taken and ensure that elem will remain THREAD_QUEUED.
214 qemu_sem_timedwait(&pool->sem, 0) == 0) {
215 QTAILQ_REMOVE(&pool->request_list, elem, reqs);
216 qemu_bh_schedule(pool->completion_bh);
218 elem->state = THREAD_DONE;
219 elem->ret = -ECANCELED;
222 qemu_mutex_unlock(&pool->lock);
225 static AioContext *thread_pool_get_aio_context(BlockAIOCB *acb)
227 ThreadPoolElement *elem = (ThreadPoolElement *)acb;
228 ThreadPool *pool = elem->pool;
229 return pool->ctx;
232 static const AIOCBInfo thread_pool_aiocb_info = {
233 .aiocb_size = sizeof(ThreadPoolElement),
234 .cancel_async = thread_pool_cancel,
235 .get_aio_context = thread_pool_get_aio_context,
238 BlockAIOCB *thread_pool_submit_aio(ThreadPool *pool,
239 ThreadPoolFunc *func, void *arg,
240 BlockCompletionFunc *cb, void *opaque)
242 ThreadPoolElement *req;
244 req = qemu_aio_get(&thread_pool_aiocb_info, NULL, cb, opaque);
245 req->func = func;
246 req->arg = arg;
247 req->state = THREAD_QUEUED;
248 req->pool = pool;
250 QLIST_INSERT_HEAD(&pool->head, req, all);
252 trace_thread_pool_submit(pool, req, arg);
254 qemu_mutex_lock(&pool->lock);
255 if (pool->idle_threads == 0 && pool->cur_threads < pool->max_threads) {
256 spawn_thread(pool);
258 QTAILQ_INSERT_TAIL(&pool->request_list, req, reqs);
259 qemu_mutex_unlock(&pool->lock);
260 qemu_sem_post(&pool->sem);
261 return &req->common;
264 typedef struct ThreadPoolCo {
265 Coroutine *co;
266 int ret;
267 } ThreadPoolCo;
269 static void thread_pool_co_cb(void *opaque, int ret)
271 ThreadPoolCo *co = opaque;
273 co->ret = ret;
274 aio_co_wake(co->co);
277 int coroutine_fn thread_pool_submit_co(ThreadPool *pool, ThreadPoolFunc *func,
278 void *arg)
280 ThreadPoolCo tpc = { .co = qemu_coroutine_self(), .ret = -EINPROGRESS };
281 assert(qemu_in_coroutine());
282 thread_pool_submit_aio(pool, func, arg, thread_pool_co_cb, &tpc);
283 qemu_coroutine_yield();
284 return tpc.ret;
287 void thread_pool_submit(ThreadPool *pool, ThreadPoolFunc *func, void *arg)
289 thread_pool_submit_aio(pool, func, arg, NULL, NULL);
292 static void thread_pool_init_one(ThreadPool *pool, AioContext *ctx)
294 if (!ctx) {
295 ctx = qemu_get_aio_context();
298 memset(pool, 0, sizeof(*pool));
299 pool->ctx = ctx;
300 pool->completion_bh = aio_bh_new(ctx, thread_pool_completion_bh, pool);
301 qemu_mutex_init(&pool->lock);
302 qemu_cond_init(&pool->worker_stopped);
303 qemu_sem_init(&pool->sem, 0);
304 pool->max_threads = 64;
305 pool->new_thread_bh = aio_bh_new(ctx, spawn_thread_bh_fn, pool);
307 QLIST_INIT(&pool->head);
308 QTAILQ_INIT(&pool->request_list);
311 ThreadPool *thread_pool_new(AioContext *ctx)
313 ThreadPool *pool = g_new(ThreadPool, 1);
314 thread_pool_init_one(pool, ctx);
315 return pool;
318 void thread_pool_free(ThreadPool *pool)
320 if (!pool) {
321 return;
324 assert(QLIST_EMPTY(&pool->head));
326 qemu_mutex_lock(&pool->lock);
328 /* Stop new threads from spawning */
329 qemu_bh_delete(pool->new_thread_bh);
330 pool->cur_threads -= pool->new_threads;
331 pool->new_threads = 0;
333 /* Wait for worker threads to terminate */
334 pool->stopping = true;
335 while (pool->cur_threads > 0) {
336 qemu_sem_post(&pool->sem);
337 qemu_cond_wait(&pool->worker_stopped, &pool->lock);
340 qemu_mutex_unlock(&pool->lock);
342 qemu_bh_delete(pool->completion_bh);
343 qemu_sem_destroy(&pool->sem);
344 qemu_cond_destroy(&pool->worker_stopped);
345 qemu_mutex_destroy(&pool->lock);
346 g_free(pool);