1 /*- 2 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD 3 * 4 * Copyright (c) 2000 Doug Rabson 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 26 * SUCH DAMAGE. 27 */ 28 29 #include <sys/cdefs.h> 30 __FBSDID("$FreeBSD$"); 31 32 #include <sys/param.h> 33 #include <sys/systm.h> 34 #include <sys/bus.h> 35 #include <sys/cpuset.h> 36 #include <sys/interrupt.h> 37 #include <sys/kernel.h> 38 #include <sys/kthread.h> 39 #include <sys/libkern.h> 40 #include <sys/limits.h> 41 #include <sys/lock.h> 42 #include <sys/malloc.h> 43 #include <sys/mutex.h> 44 #include <sys/proc.h> 45 #include <sys/epoch.h> 46 #include <sys/sched.h> 47 #include <sys/smp.h> 48 #include <sys/taskqueue.h> 49 #include <sys/unistd.h> 50 #include <machine/stdarg.h> 51 52 static MALLOC_DEFINE(M_TASKQUEUE, "taskqueue", "Task Queues"); 53 static void *taskqueue_giant_ih; 54 static void *taskqueue_ih; 55 static void taskqueue_fast_enqueue(void *); 56 static void taskqueue_swi_enqueue(void *); 57 static void taskqueue_swi_giant_enqueue(void *); 58 59 struct taskqueue_busy { 60 struct task *tb_running; 61 u_int tb_seq; 62 LIST_ENTRY(taskqueue_busy) tb_link; 63 }; 64 65 struct taskqueue { 66 STAILQ_HEAD(, task) tq_queue; 67 LIST_HEAD(, taskqueue_busy) tq_active; 68 struct task *tq_hint; 69 u_int tq_seq; 70 int tq_callouts; 71 struct mtx_padalign tq_mutex; 72 taskqueue_enqueue_fn tq_enqueue; 73 void *tq_context; 74 char *tq_name; 75 struct thread **tq_threads; 76 int tq_tcount; 77 int tq_spin; 78 int tq_flags; 79 taskqueue_callback_fn tq_callbacks[TASKQUEUE_NUM_CALLBACKS]; 80 void *tq_cb_contexts[TASKQUEUE_NUM_CALLBACKS]; 81 }; 82 83 #define TQ_FLAGS_ACTIVE (1 << 0) 84 #define TQ_FLAGS_BLOCKED (1 << 1) 85 #define TQ_FLAGS_UNLOCKED_ENQUEUE (1 << 2) 86 87 #define DT_CALLOUT_ARMED (1 << 0) 88 #define DT_DRAIN_IN_PROGRESS (1 << 1) 89 90 #define TQ_LOCK(tq) \ 91 do { \ 92 if ((tq)->tq_spin) \ 93 mtx_lock_spin(&(tq)->tq_mutex); \ 94 else \ 95 mtx_lock(&(tq)->tq_mutex); \ 96 } while (0) 97 #define TQ_ASSERT_LOCKED(tq) mtx_assert(&(tq)->tq_mutex, MA_OWNED) 98 99 #define TQ_UNLOCK(tq) \ 100 do { \ 101 if ((tq)->tq_spin) \ 102 mtx_unlock_spin(&(tq)->tq_mutex); \ 103 else \ 104 mtx_unlock(&(tq)->tq_mutex); \ 105 } while (0) 106 #define TQ_ASSERT_UNLOCKED(tq) mtx_assert(&(tq)->tq_mutex, MA_NOTOWNED) 107 108 void 109 _timeout_task_init(struct taskqueue *queue, struct timeout_task *timeout_task, 110 int priority, task_fn_t func, void *context) 111 { 112 113 TASK_INIT(&timeout_task->t, priority, func, context); 114 callout_init_mtx(&timeout_task->c, &queue->tq_mutex, 115 CALLOUT_RETURNUNLOCKED); 116 timeout_task->q = queue; 117 timeout_task->f = 0; 118 } 119 120 static __inline int 121 TQ_SLEEP(struct taskqueue *tq, void *p, const char *wm) 122 { 123 if (tq->tq_spin) 124 return (msleep_spin(p, (struct mtx *)&tq->tq_mutex, wm, 0)); 125 return (msleep(p, &tq->tq_mutex, 0, wm, 0)); 126 } 127 128 static struct taskqueue * 129 _taskqueue_create(const char *name, int mflags, 130 taskqueue_enqueue_fn enqueue, void *context, 131 int mtxflags, const char *mtxname __unused) 132 { 133 struct taskqueue *queue; 134 char *tq_name; 135 136 tq_name = malloc(TASKQUEUE_NAMELEN, M_TASKQUEUE, mflags | M_ZERO); 137 if (tq_name == NULL) 138 return (NULL); 139 140 queue = malloc(sizeof(struct taskqueue), M_TASKQUEUE, mflags | M_ZERO); 141 if (queue == NULL) { 142 free(tq_name, M_TASKQUEUE); 143 return (NULL); 144 } 145 146 snprintf(tq_name, TASKQUEUE_NAMELEN, "%s", (name) ? name : "taskqueue"); 147 148 STAILQ_INIT(&queue->tq_queue); 149 LIST_INIT(&queue->tq_active); 150 queue->tq_enqueue = enqueue; 151 queue->tq_context = context; 152 queue->tq_name = tq_name; 153 queue->tq_spin = (mtxflags & MTX_SPIN) != 0; 154 queue->tq_flags |= TQ_FLAGS_ACTIVE; 155 if (enqueue == taskqueue_fast_enqueue || 156 enqueue == taskqueue_swi_enqueue || 157 enqueue == taskqueue_swi_giant_enqueue || 158 enqueue == taskqueue_thread_enqueue) 159 queue->tq_flags |= TQ_FLAGS_UNLOCKED_ENQUEUE; 160 mtx_init(&queue->tq_mutex, tq_name, NULL, mtxflags); 161 162 return (queue); 163 } 164 165 struct taskqueue * 166 taskqueue_create(const char *name, int mflags, 167 taskqueue_enqueue_fn enqueue, void *context) 168 { 169 170 return _taskqueue_create(name, mflags, enqueue, context, 171 MTX_DEF, name); 172 } 173 174 void 175 taskqueue_set_callback(struct taskqueue *queue, 176 enum taskqueue_callback_type cb_type, taskqueue_callback_fn callback, 177 void *context) 178 { 179 180 KASSERT(((cb_type >= TASKQUEUE_CALLBACK_TYPE_MIN) && 181 (cb_type <= TASKQUEUE_CALLBACK_TYPE_MAX)), 182 ("Callback type %d not valid, must be %d-%d", cb_type, 183 TASKQUEUE_CALLBACK_TYPE_MIN, TASKQUEUE_CALLBACK_TYPE_MAX)); 184 KASSERT((queue->tq_callbacks[cb_type] == NULL), 185 ("Re-initialization of taskqueue callback?")); 186 187 queue->tq_callbacks[cb_type] = callback; 188 queue->tq_cb_contexts[cb_type] = context; 189 } 190 191 /* 192 * Signal a taskqueue thread to terminate. 193 */ 194 static void 195 taskqueue_terminate(struct thread **pp, struct taskqueue *tq) 196 { 197 198 while (tq->tq_tcount > 0 || tq->tq_callouts > 0) { 199 wakeup(tq); 200 TQ_SLEEP(tq, pp, "tq_destroy"); 201 } 202 } 203 204 void 205 taskqueue_free(struct taskqueue *queue) 206 { 207 208 TQ_LOCK(queue); 209 queue->tq_flags &= ~TQ_FLAGS_ACTIVE; 210 taskqueue_terminate(queue->tq_threads, queue); 211 KASSERT(LIST_EMPTY(&queue->tq_active), ("Tasks still running?")); 212 KASSERT(queue->tq_callouts == 0, ("Armed timeout tasks")); 213 mtx_destroy(&queue->tq_mutex); 214 free(queue->tq_threads, M_TASKQUEUE); 215 free(queue->tq_name, M_TASKQUEUE); 216 free(queue, M_TASKQUEUE); 217 } 218 219 static int 220 taskqueue_enqueue_locked(struct taskqueue *queue, struct task *task) 221 { 222 struct task *ins; 223 struct task *prev; 224 225 KASSERT(task->ta_func != NULL, ("enqueueing task with NULL func")); 226 /* 227 * Count multiple enqueues. 228 */ 229 if (task->ta_pending) { 230 if (task->ta_pending < USHRT_MAX) 231 task->ta_pending++; 232 TQ_UNLOCK(queue); 233 return (0); 234 } 235 236 /* 237 * Optimise cases when all tasks use small set of priorities. 238 * In case of only one priority we always insert at the end. 239 * In case of two tq_hint typically gives the insertion point. 240 * In case of more then two tq_hint should halve the search. 241 */ 242 prev = STAILQ_LAST(&queue->tq_queue, task, ta_link); 243 if (!prev || prev->ta_priority >= task->ta_priority) { 244 STAILQ_INSERT_TAIL(&queue->tq_queue, task, ta_link); 245 } else { 246 prev = queue->tq_hint; 247 if (prev && prev->ta_priority >= task->ta_priority) { 248 ins = STAILQ_NEXT(prev, ta_link); 249 } else { 250 prev = NULL; 251 ins = STAILQ_FIRST(&queue->tq_queue); 252 } 253 for (; ins; prev = ins, ins = STAILQ_NEXT(ins, ta_link)) 254 if (ins->ta_priority < task->ta_priority) 255 break; 256 257 if (prev) { 258 STAILQ_INSERT_AFTER(&queue->tq_queue, prev, task, ta_link); 259 queue->tq_hint = task; 260 } else 261 STAILQ_INSERT_HEAD(&queue->tq_queue, task, ta_link); 262 } 263 264 task->ta_pending = 1; 265 if ((queue->tq_flags & TQ_FLAGS_UNLOCKED_ENQUEUE) != 0) 266 TQ_UNLOCK(queue); 267 if ((queue->tq_flags & TQ_FLAGS_BLOCKED) == 0) 268 queue->tq_enqueue(queue->tq_context); 269 if ((queue->tq_flags & TQ_FLAGS_UNLOCKED_ENQUEUE) == 0) 270 TQ_UNLOCK(queue); 271 272 /* Return with lock released. */ 273 return (0); 274 } 275 276 int 277 taskqueue_enqueue(struct taskqueue *queue, struct task *task) 278 { 279 int res; 280 281 TQ_LOCK(queue); 282 res = taskqueue_enqueue_locked(queue, task); 283 /* The lock is released inside. */ 284 285 return (res); 286 } 287 288 static void 289 taskqueue_timeout_func(void *arg) 290 { 291 struct taskqueue *queue; 292 struct timeout_task *timeout_task; 293 294 timeout_task = arg; 295 queue = timeout_task->q; 296 KASSERT((timeout_task->f & DT_CALLOUT_ARMED) != 0, ("Stray timeout")); 297 timeout_task->f &= ~DT_CALLOUT_ARMED; 298 queue->tq_callouts--; 299 taskqueue_enqueue_locked(timeout_task->q, &timeout_task->t); 300 /* The lock is released inside. */ 301 } 302 303 int 304 taskqueue_enqueue_timeout_sbt(struct taskqueue *queue, 305 struct timeout_task *timeout_task, sbintime_t sbt, sbintime_t pr, int flags) 306 { 307 int res; 308 309 TQ_LOCK(queue); 310 KASSERT(timeout_task->q == NULL || timeout_task->q == queue, 311 ("Migrated queue")); 312 timeout_task->q = queue; 313 res = timeout_task->t.ta_pending; 314 if (timeout_task->f & DT_DRAIN_IN_PROGRESS) { 315 /* Do nothing */ 316 TQ_UNLOCK(queue); 317 res = -1; 318 } else if (sbt == 0) { 319 taskqueue_enqueue_locked(queue, &timeout_task->t); 320 /* The lock is released inside. */ 321 } else { 322 if ((timeout_task->f & DT_CALLOUT_ARMED) != 0) { 323 res++; 324 } else { 325 queue->tq_callouts++; 326 timeout_task->f |= DT_CALLOUT_ARMED; 327 if (sbt < 0) 328 sbt = -sbt; /* Ignore overflow. */ 329 } 330 if (sbt > 0) { 331 if (queue->tq_spin) 332 flags |= C_DIRECT_EXEC; 333 callout_reset_sbt(&timeout_task->c, sbt, pr, 334 taskqueue_timeout_func, timeout_task, flags); 335 } 336 TQ_UNLOCK(queue); 337 } 338 return (res); 339 } 340 341 int 342 taskqueue_enqueue_timeout(struct taskqueue *queue, 343 struct timeout_task *ttask, int ticks) 344 { 345 346 return (taskqueue_enqueue_timeout_sbt(queue, ttask, ticks * tick_sbt, 347 0, C_HARDCLOCK)); 348 } 349 350 static void 351 taskqueue_task_nop_fn(void *context, int pending) 352 { 353 } 354 355 /* 356 * Block until all currently queued tasks in this taskqueue 357 * have begun execution. Tasks queued during execution of 358 * this function are ignored. 359 */ 360 static int 361 taskqueue_drain_tq_queue(struct taskqueue *queue) 362 { 363 struct task t_barrier; 364 365 if (STAILQ_EMPTY(&queue->tq_queue)) 366 return (0); 367 368 /* 369 * Enqueue our barrier after all current tasks, but with 370 * the highest priority so that newly queued tasks cannot 371 * pass it. Because of the high priority, we can not use 372 * taskqueue_enqueue_locked directly (which drops the lock 373 * anyway) so just insert it at tail while we have the 374 * queue lock. 375 */ 376 TASK_INIT(&t_barrier, UCHAR_MAX, taskqueue_task_nop_fn, &t_barrier); 377 STAILQ_INSERT_TAIL(&queue->tq_queue, &t_barrier, ta_link); 378 queue->tq_hint = &t_barrier; 379 t_barrier.ta_pending = 1; 380 381 /* 382 * Once the barrier has executed, all previously queued tasks 383 * have completed or are currently executing. 384 */ 385 while (t_barrier.ta_pending != 0) 386 TQ_SLEEP(queue, &t_barrier, "tq_qdrain"); 387 return (1); 388 } 389 390 /* 391 * Block until all currently executing tasks for this taskqueue 392 * complete. Tasks that begin execution during the execution 393 * of this function are ignored. 394 */ 395 static int 396 taskqueue_drain_tq_active(struct taskqueue *queue) 397 { 398 struct taskqueue_busy *tb; 399 u_int seq; 400 401 if (LIST_EMPTY(&queue->tq_active)) 402 return (0); 403 404 /* Block taskq_terminate().*/ 405 queue->tq_callouts++; 406 407 /* Wait for any active task with sequence from the past. */ 408 seq = queue->tq_seq; 409 restart: 410 LIST_FOREACH(tb, &queue->tq_active, tb_link) { 411 if ((int)(tb->tb_seq - seq) <= 0) { 412 TQ_SLEEP(queue, tb->tb_running, "tq_adrain"); 413 goto restart; 414 } 415 } 416 417 /* Release taskqueue_terminate(). */ 418 queue->tq_callouts--; 419 if ((queue->tq_flags & TQ_FLAGS_ACTIVE) == 0) 420 wakeup_one(queue->tq_threads); 421 return (1); 422 } 423 424 void 425 taskqueue_block(struct taskqueue *queue) 426 { 427 428 TQ_LOCK(queue); 429 queue->tq_flags |= TQ_FLAGS_BLOCKED; 430 TQ_UNLOCK(queue); 431 } 432 433 void 434 taskqueue_unblock(struct taskqueue *queue) 435 { 436 437 TQ_LOCK(queue); 438 queue->tq_flags &= ~TQ_FLAGS_BLOCKED; 439 if (!STAILQ_EMPTY(&queue->tq_queue)) 440 queue->tq_enqueue(queue->tq_context); 441 TQ_UNLOCK(queue); 442 } 443 444 static void 445 taskqueue_run_locked(struct taskqueue *queue) 446 { 447 struct epoch_tracker et; 448 struct taskqueue_busy tb; 449 struct task *task; 450 bool in_net_epoch; 451 int pending; 452 453 KASSERT(queue != NULL, ("tq is NULL")); 454 TQ_ASSERT_LOCKED(queue); 455 tb.tb_running = NULL; 456 LIST_INSERT_HEAD(&queue->tq_active, &tb, tb_link); 457 in_net_epoch = false; 458 459 while ((task = STAILQ_FIRST(&queue->tq_queue)) != NULL) { 460 STAILQ_REMOVE_HEAD(&queue->tq_queue, ta_link); 461 if (queue->tq_hint == task) 462 queue->tq_hint = NULL; 463 pending = task->ta_pending; 464 task->ta_pending = 0; 465 tb.tb_running = task; 466 tb.tb_seq = ++queue->tq_seq; 467 TQ_UNLOCK(queue); 468 469 KASSERT(task->ta_func != NULL, ("task->ta_func is NULL")); 470 if (!in_net_epoch && TASK_IS_NET(task)) { 471 in_net_epoch = true; 472 NET_EPOCH_ENTER(et); 473 } else if (in_net_epoch && !TASK_IS_NET(task)) { 474 NET_EPOCH_EXIT(et); 475 in_net_epoch = false; 476 } 477 task->ta_func(task->ta_context, pending); 478 479 TQ_LOCK(queue); 480 wakeup(task); 481 } 482 if (in_net_epoch) 483 NET_EPOCH_EXIT(et); 484 LIST_REMOVE(&tb, tb_link); 485 } 486 487 void 488 taskqueue_run(struct taskqueue *queue) 489 { 490 491 TQ_LOCK(queue); 492 taskqueue_run_locked(queue); 493 TQ_UNLOCK(queue); 494 } 495 496 static int 497 task_is_running(struct taskqueue *queue, struct task *task) 498 { 499 struct taskqueue_busy *tb; 500 501 TQ_ASSERT_LOCKED(queue); 502 LIST_FOREACH(tb, &queue->tq_active, tb_link) { 503 if (tb->tb_running == task) 504 return (1); 505 } 506 return (0); 507 } 508 509 /* 510 * Only use this function in single threaded contexts. It returns 511 * non-zero if the given task is either pending or running. Else the 512 * task is idle and can be queued again or freed. 513 */ 514 int 515 taskqueue_poll_is_busy(struct taskqueue *queue, struct task *task) 516 { 517 int retval; 518 519 TQ_LOCK(queue); 520 retval = task->ta_pending > 0 || task_is_running(queue, task); 521 TQ_UNLOCK(queue); 522 523 return (retval); 524 } 525 526 static int 527 taskqueue_cancel_locked(struct taskqueue *queue, struct task *task, 528 u_int *pendp) 529 { 530 531 if (task->ta_pending > 0) { 532 STAILQ_REMOVE(&queue->tq_queue, task, task, ta_link); 533 if (queue->tq_hint == task) 534 queue->tq_hint = NULL; 535 } 536 if (pendp != NULL) 537 *pendp = task->ta_pending; 538 task->ta_pending = 0; 539 return (task_is_running(queue, task) ? EBUSY : 0); 540 } 541 542 int 543 taskqueue_cancel(struct taskqueue *queue, struct task *task, u_int *pendp) 544 { 545 int error; 546 547 TQ_LOCK(queue); 548 error = taskqueue_cancel_locked(queue, task, pendp); 549 TQ_UNLOCK(queue); 550 551 return (error); 552 } 553 554 int 555 taskqueue_cancel_timeout(struct taskqueue *queue, 556 struct timeout_task *timeout_task, u_int *pendp) 557 { 558 u_int pending, pending1; 559 int error; 560 561 TQ_LOCK(queue); 562 pending = !!(callout_stop(&timeout_task->c) > 0); 563 error = taskqueue_cancel_locked(queue, &timeout_task->t, &pending1); 564 if ((timeout_task->f & DT_CALLOUT_ARMED) != 0) { 565 timeout_task->f &= ~DT_CALLOUT_ARMED; 566 queue->tq_callouts--; 567 } 568 TQ_UNLOCK(queue); 569 570 if (pendp != NULL) 571 *pendp = pending + pending1; 572 return (error); 573 } 574 575 void 576 taskqueue_drain(struct taskqueue *queue, struct task *task) 577 { 578 579 if (!queue->tq_spin) 580 WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL, __func__); 581 582 TQ_LOCK(queue); 583 while (task->ta_pending != 0 || task_is_running(queue, task)) 584 TQ_SLEEP(queue, task, "tq_drain"); 585 TQ_UNLOCK(queue); 586 } 587 588 void 589 taskqueue_drain_all(struct taskqueue *queue) 590 { 591 592 if (!queue->tq_spin) 593 WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL, __func__); 594 595 TQ_LOCK(queue); 596 (void)taskqueue_drain_tq_queue(queue); 597 (void)taskqueue_drain_tq_active(queue); 598 TQ_UNLOCK(queue); 599 } 600 601 void 602 taskqueue_drain_timeout(struct taskqueue *queue, 603 struct timeout_task *timeout_task) 604 { 605 606 /* 607 * Set flag to prevent timer from re-starting during drain: 608 */ 609 TQ_LOCK(queue); 610 KASSERT((timeout_task->f & DT_DRAIN_IN_PROGRESS) == 0, 611 ("Drain already in progress")); 612 timeout_task->f |= DT_DRAIN_IN_PROGRESS; 613 TQ_UNLOCK(queue); 614 615 callout_drain(&timeout_task->c); 616 taskqueue_drain(queue, &timeout_task->t); 617 618 /* 619 * Clear flag to allow timer to re-start: 620 */ 621 TQ_LOCK(queue); 622 timeout_task->f &= ~DT_DRAIN_IN_PROGRESS; 623 TQ_UNLOCK(queue); 624 } 625 626 void 627 taskqueue_quiesce(struct taskqueue *queue) 628 { 629 int ret; 630 631 TQ_LOCK(queue); 632 do { 633 ret = taskqueue_drain_tq_queue(queue); 634 if (ret == 0) 635 ret = taskqueue_drain_tq_active(queue); 636 } while (ret != 0); 637 TQ_UNLOCK(queue); 638 } 639 640 static void 641 taskqueue_swi_enqueue(void *context) 642 { 643 swi_sched(taskqueue_ih, 0); 644 } 645 646 static void 647 taskqueue_swi_run(void *dummy) 648 { 649 taskqueue_run(taskqueue_swi); 650 } 651 652 static void 653 taskqueue_swi_giant_enqueue(void *context) 654 { 655 swi_sched(taskqueue_giant_ih, 0); 656 } 657 658 static void 659 taskqueue_swi_giant_run(void *dummy) 660 { 661 taskqueue_run(taskqueue_swi_giant); 662 } 663 664 static int 665 _taskqueue_start_threads(struct taskqueue **tqp, int count, int pri, 666 cpuset_t *mask, struct proc *p, const char *name, va_list ap) 667 { 668 char ktname[MAXCOMLEN + 1]; 669 struct thread *td; 670 struct taskqueue *tq; 671 int i, error; 672 673 if (count <= 0) 674 return (EINVAL); 675 676 vsnprintf(ktname, sizeof(ktname), name, ap); 677 tq = *tqp; 678 679 tq->tq_threads = malloc(sizeof(struct thread *) * count, M_TASKQUEUE, 680 M_NOWAIT | M_ZERO); 681 if (tq->tq_threads == NULL) { 682 printf("%s: no memory for %s threads\n", __func__, ktname); 683 return (ENOMEM); 684 } 685 686 for (i = 0; i < count; i++) { 687 if (count == 1) 688 error = kthread_add(taskqueue_thread_loop, tqp, p, 689 &tq->tq_threads[i], RFSTOPPED, 0, "%s", ktname); 690 else 691 error = kthread_add(taskqueue_thread_loop, tqp, p, 692 &tq->tq_threads[i], RFSTOPPED, 0, 693 "%s_%d", ktname, i); 694 if (error) { 695 /* should be ok to continue, taskqueue_free will dtrt */ 696 printf("%s: kthread_add(%s): error %d", __func__, 697 ktname, error); 698 tq->tq_threads[i] = NULL; /* paranoid */ 699 } else 700 tq->tq_tcount++; 701 } 702 if (tq->tq_tcount == 0) { 703 free(tq->tq_threads, M_TASKQUEUE); 704 tq->tq_threads = NULL; 705 return (ENOMEM); 706 } 707 for (i = 0; i < count; i++) { 708 if (tq->tq_threads[i] == NULL) 709 continue; 710 td = tq->tq_threads[i]; 711 if (mask) { 712 error = cpuset_setthread(td->td_tid, mask); 713 /* 714 * Failing to pin is rarely an actual fatal error; 715 * it'll just affect performance. 716 */ 717 if (error) 718 printf("%s: curthread=%llu: can't pin; " 719 "error=%d\n", 720 __func__, 721 (unsigned long long) td->td_tid, 722 error); 723 } 724 thread_lock(td); 725 sched_prio(td, pri); 726 sched_add(td, SRQ_BORING); 727 } 728 729 return (0); 730 } 731 732 int 733 taskqueue_start_threads(struct taskqueue **tqp, int count, int pri, 734 const char *name, ...) 735 { 736 va_list ap; 737 int error; 738 739 va_start(ap, name); 740 error = _taskqueue_start_threads(tqp, count, pri, NULL, NULL, name, ap); 741 va_end(ap); 742 return (error); 743 } 744 745 int 746 taskqueue_start_threads_in_proc(struct taskqueue **tqp, int count, int pri, 747 struct proc *proc, const char *name, ...) 748 { 749 va_list ap; 750 int error; 751 752 va_start(ap, name); 753 error = _taskqueue_start_threads(tqp, count, pri, NULL, proc, name, ap); 754 va_end(ap); 755 return (error); 756 } 757 758 int 759 taskqueue_start_threads_cpuset(struct taskqueue **tqp, int count, int pri, 760 cpuset_t *mask, const char *name, ...) 761 { 762 va_list ap; 763 int error; 764 765 va_start(ap, name); 766 error = _taskqueue_start_threads(tqp, count, pri, mask, NULL, name, ap); 767 va_end(ap); 768 return (error); 769 } 770 771 static inline void 772 taskqueue_run_callback(struct taskqueue *tq, 773 enum taskqueue_callback_type cb_type) 774 { 775 taskqueue_callback_fn tq_callback; 776 777 TQ_ASSERT_UNLOCKED(tq); 778 tq_callback = tq->tq_callbacks[cb_type]; 779 if (tq_callback != NULL) 780 tq_callback(tq->tq_cb_contexts[cb_type]); 781 } 782 783 void 784 taskqueue_thread_loop(void *arg) 785 { 786 struct taskqueue **tqp, *tq; 787 788 tqp = arg; 789 tq = *tqp; 790 taskqueue_run_callback(tq, TASKQUEUE_CALLBACK_TYPE_INIT); 791 TQ_LOCK(tq); 792 while ((tq->tq_flags & TQ_FLAGS_ACTIVE) != 0) { 793 /* XXX ? */ 794 taskqueue_run_locked(tq); 795 /* 796 * Because taskqueue_run() can drop tq_mutex, we need to 797 * check if the TQ_FLAGS_ACTIVE flag wasn't removed in the 798 * meantime, which means we missed a wakeup. 799 */ 800 if ((tq->tq_flags & TQ_FLAGS_ACTIVE) == 0) 801 break; 802 TQ_SLEEP(tq, tq, "-"); 803 } 804 taskqueue_run_locked(tq); 805 /* 806 * This thread is on its way out, so just drop the lock temporarily 807 * in order to call the shutdown callback. This allows the callback 808 * to look at the taskqueue, even just before it dies. 809 */ 810 TQ_UNLOCK(tq); 811 taskqueue_run_callback(tq, TASKQUEUE_CALLBACK_TYPE_SHUTDOWN); 812 TQ_LOCK(tq); 813 814 /* rendezvous with thread that asked us to terminate */ 815 tq->tq_tcount--; 816 wakeup_one(tq->tq_threads); 817 TQ_UNLOCK(tq); 818 kthread_exit(); 819 } 820 821 void 822 taskqueue_thread_enqueue(void *context) 823 { 824 struct taskqueue **tqp, *tq; 825 826 tqp = context; 827 tq = *tqp; 828 wakeup_any(tq); 829 } 830 831 TASKQUEUE_DEFINE(swi, taskqueue_swi_enqueue, NULL, 832 swi_add(NULL, "task queue", taskqueue_swi_run, NULL, SWI_TQ, 833 INTR_MPSAFE, &taskqueue_ih)); 834 835 TASKQUEUE_DEFINE(swi_giant, taskqueue_swi_giant_enqueue, NULL, 836 swi_add(NULL, "Giant taskq", taskqueue_swi_giant_run, 837 NULL, SWI_TQ_GIANT, 0, &taskqueue_giant_ih)); 838 839 TASKQUEUE_DEFINE_THREAD(thread); 840 841 struct taskqueue * 842 taskqueue_create_fast(const char *name, int mflags, 843 taskqueue_enqueue_fn enqueue, void *context) 844 { 845 return _taskqueue_create(name, mflags, enqueue, context, 846 MTX_SPIN, "fast_taskqueue"); 847 } 848 849 static void *taskqueue_fast_ih; 850 851 static void 852 taskqueue_fast_enqueue(void *context) 853 { 854 swi_sched(taskqueue_fast_ih, 0); 855 } 856 857 static void 858 taskqueue_fast_run(void *dummy) 859 { 860 taskqueue_run(taskqueue_fast); 861 } 862 863 TASKQUEUE_FAST_DEFINE(fast, taskqueue_fast_enqueue, NULL, 864 swi_add(NULL, "fast taskq", taskqueue_fast_run, NULL, 865 SWI_TQ_FAST, INTR_MPSAFE, &taskqueue_fast_ih)); 866 867 int 868 taskqueue_member(struct taskqueue *queue, struct thread *td) 869 { 870 int i, j, ret = 0; 871 872 for (i = 0, j = 0; ; i++) { 873 if (queue->tq_threads[i] == NULL) 874 continue; 875 if (queue->tq_threads[i] == td) { 876 ret = 1; 877 break; 878 } 879 if (++j >= queue->tq_tcount) 880 break; 881 } 882 return (ret); 883 } 884