1 /*- 2 * Copyright (c) 2006 John Baldwin <jhb@FreeBSD.org> 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 3. Neither the name of the author nor the names of any co-contributors 14 * may be used to endorse or promote products derived from this software 15 * without specific prior written permission. 16 * 17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 20 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 27 * SUCH DAMAGE. 28 */ 29 30 /* 31 * Machine independent bits of reader/writer lock implementation. 32 */ 33 34 #include <sys/cdefs.h> 35 __FBSDID("$FreeBSD$"); 36 37 #include "opt_ddb.h" 38 #include "opt_no_adaptive_rwlocks.h" 39 40 #include <sys/param.h> 41 #include <sys/ktr.h> 42 #include <sys/lock.h> 43 #include <sys/mutex.h> 44 #include <sys/proc.h> 45 #include <sys/rwlock.h> 46 #include <sys/systm.h> 47 #include <sys/turnstile.h> 48 49 #include <machine/cpu.h> 50 51 CTASSERT((RW_RECURSE & LO_CLASSFLAGS) == RW_RECURSE); 52 53 #if defined(SMP) && !defined(NO_ADAPTIVE_RWLOCKS) 54 #define ADAPTIVE_RWLOCKS 55 #endif 56 57 #ifdef DDB 58 #include <ddb/ddb.h> 59 60 static void db_show_rwlock(struct lock_object *lock); 61 #endif 62 static void assert_rw(struct lock_object *lock, int what); 63 static void lock_rw(struct lock_object *lock, int how); 64 static int unlock_rw(struct lock_object *lock); 65 66 struct lock_class lock_class_rw = { 67 .lc_name = "rw", 68 .lc_flags = LC_SLEEPLOCK | LC_RECURSABLE | LC_UPGRADABLE, 69 .lc_assert = assert_rw, 70 #ifdef DDB 71 .lc_ddb_show = db_show_rwlock, 72 #endif 73 .lc_lock = lock_rw, 74 .lc_unlock = unlock_rw, 75 }; 76 77 /* 78 * Return a pointer to the owning thread if the lock is write-locked or 79 * NULL if the lock is unlocked or read-locked. 80 */ 81 #define rw_wowner(rw) \ 82 ((rw)->rw_lock & RW_LOCK_READ ? NULL : \ 83 (struct thread *)RW_OWNER((rw)->rw_lock)) 84 85 /* 86 * Returns if a write owner is recursed. Write ownership is not assured 87 * here and should be previously checked. 88 */ 89 #define rw_recursed(rw) ((rw)->rw_recurse != 0) 90 91 /* 92 * Return true if curthread helds the lock. 93 */ 94 #define rw_wlocked(rw) (rw_wowner((rw)) == curthread) 95 96 /* 97 * Return a pointer to the owning thread for this lock who should receive 98 * any priority lent by threads that block on this lock. Currently this 99 * is identical to rw_wowner(). 100 */ 101 #define rw_owner(rw) rw_wowner(rw) 102 103 #ifndef INVARIANTS 104 #define _rw_assert(rw, what, file, line) 105 #endif 106 107 void 108 assert_rw(struct lock_object *lock, int what) 109 { 110 111 rw_assert((struct rwlock *)lock, what); 112 } 113 114 void 115 lock_rw(struct lock_object *lock, int how) 116 { 117 struct rwlock *rw; 118 119 rw = (struct rwlock *)lock; 120 if (how) 121 rw_wlock(rw); 122 else 123 rw_rlock(rw); 124 } 125 126 int 127 unlock_rw(struct lock_object *lock) 128 { 129 struct rwlock *rw; 130 131 rw = (struct rwlock *)lock; 132 rw_assert(rw, RA_LOCKED | LA_NOTRECURSED); 133 if (rw->rw_lock & RW_LOCK_READ) { 134 rw_runlock(rw); 135 return (0); 136 } else { 137 rw_wunlock(rw); 138 return (1); 139 } 140 } 141 142 void 143 rw_init_flags(struct rwlock *rw, const char *name, int opts) 144 { 145 int flags; 146 147 MPASS((opts & ~(RW_DUPOK | RW_NOPROFILE | RW_NOWITNESS | RW_QUIET | 148 RW_RECURSE)) == 0); 149 150 flags = LO_UPGRADABLE | LO_RECURSABLE; 151 if (opts & RW_DUPOK) 152 flags |= LO_DUPOK; 153 if (opts & RW_NOPROFILE) 154 flags |= LO_NOPROFILE; 155 if (!(opts & RW_NOWITNESS)) 156 flags |= LO_WITNESS; 157 if (opts & RW_QUIET) 158 flags |= LO_QUIET; 159 flags |= opts & RW_RECURSE; 160 161 rw->rw_lock = RW_UNLOCKED; 162 rw->rw_recurse = 0; 163 lock_init(&rw->lock_object, &lock_class_rw, name, NULL, flags); 164 } 165 166 void 167 rw_destroy(struct rwlock *rw) 168 { 169 170 KASSERT(rw->rw_lock == RW_UNLOCKED, ("rw lock not unlocked")); 171 KASSERT(rw->rw_recurse == 0, ("rw lock still recursed")); 172 rw->rw_lock = RW_DESTROYED; 173 lock_destroy(&rw->lock_object); 174 } 175 176 void 177 rw_sysinit(void *arg) 178 { 179 struct rw_args *args = arg; 180 181 rw_init(args->ra_rw, args->ra_desc); 182 } 183 184 int 185 rw_wowned(struct rwlock *rw) 186 { 187 188 return (rw_wowner(rw) == curthread); 189 } 190 191 void 192 _rw_wlock(struct rwlock *rw, const char *file, int line) 193 { 194 195 MPASS(curthread != NULL); 196 KASSERT(rw->rw_lock != RW_DESTROYED, 197 ("rw_wlock() of destroyed rwlock @ %s:%d", file, line)); 198 WITNESS_CHECKORDER(&rw->lock_object, LOP_NEWORDER | LOP_EXCLUSIVE, file, 199 line); 200 __rw_wlock(rw, curthread, file, line); 201 LOCK_LOG_LOCK("WLOCK", &rw->lock_object, 0, rw->rw_recurse, file, line); 202 WITNESS_LOCK(&rw->lock_object, LOP_EXCLUSIVE, file, line); 203 curthread->td_locks++; 204 } 205 206 void 207 _rw_wunlock(struct rwlock *rw, const char *file, int line) 208 { 209 210 MPASS(curthread != NULL); 211 KASSERT(rw->rw_lock != RW_DESTROYED, 212 ("rw_wunlock() of destroyed rwlock @ %s:%d", file, line)); 213 _rw_assert(rw, RA_WLOCKED, file, line); 214 curthread->td_locks--; 215 WITNESS_UNLOCK(&rw->lock_object, LOP_EXCLUSIVE, file, line); 216 LOCK_LOG_LOCK("WUNLOCK", &rw->lock_object, 0, rw->rw_recurse, file, 217 line); 218 if (!rw_recursed(rw)) 219 lock_profile_release_lock(&rw->lock_object); 220 __rw_wunlock(rw, curthread, file, line); 221 } 222 223 void 224 _rw_rlock(struct rwlock *rw, const char *file, int line) 225 { 226 struct turnstile *ts; 227 #ifdef ADAPTIVE_RWLOCKS 228 volatile struct thread *owner; 229 #endif 230 #ifdef LOCK_PROFILING_SHARED 231 uint64_t waittime = 0; 232 int contested = 0; 233 #endif 234 uintptr_t x; 235 236 KASSERT(rw->rw_lock != RW_DESTROYED, 237 ("rw_rlock() of destroyed rwlock @ %s:%d", file, line)); 238 KASSERT(rw_wowner(rw) != curthread, 239 ("%s (%s): wlock already held @ %s:%d", __func__, 240 rw->lock_object.lo_name, file, line)); 241 WITNESS_CHECKORDER(&rw->lock_object, LOP_NEWORDER, file, line); 242 243 /* 244 * Note that we don't make any attempt to try to block read 245 * locks once a writer has blocked on the lock. The reason is 246 * that we currently allow for read locks to recurse and we 247 * don't keep track of all the holders of read locks. Thus, if 248 * we were to block readers once a writer blocked and a reader 249 * tried to recurse on their reader lock after a writer had 250 * blocked we would end up in a deadlock since the reader would 251 * be blocked on the writer, and the writer would be blocked 252 * waiting for the reader to release its original read lock. 253 */ 254 for (;;) { 255 /* 256 * Handle the easy case. If no other thread has a write 257 * lock, then try to bump up the count of read locks. Note 258 * that we have to preserve the current state of the 259 * RW_LOCK_WRITE_WAITERS flag. If we fail to acquire a 260 * read lock, then rw_lock must have changed, so restart 261 * the loop. Note that this handles the case of a 262 * completely unlocked rwlock since such a lock is encoded 263 * as a read lock with no waiters. 264 */ 265 x = rw->rw_lock; 266 if (x & RW_LOCK_READ) { 267 268 /* 269 * The RW_LOCK_READ_WAITERS flag should only be set 270 * if another thread currently holds a write lock, 271 * and in that case RW_LOCK_READ should be clear. 272 */ 273 MPASS((x & RW_LOCK_READ_WAITERS) == 0); 274 if (atomic_cmpset_acq_ptr(&rw->rw_lock, x, 275 x + RW_ONE_READER)) { 276 #ifdef LOCK_PROFILING_SHARED 277 if (RW_READERS(x) == 0) 278 lock_profile_obtain_lock_success( 279 &rw->lock_object, contested, 280 waittime, file, line); 281 #endif 282 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 283 CTR4(KTR_LOCK, 284 "%s: %p succeed %p -> %p", __func__, 285 rw, (void *)x, 286 (void *)(x + RW_ONE_READER)); 287 break; 288 } 289 cpu_spinwait(); 290 continue; 291 } 292 293 #ifdef ADAPTIVE_RWLOCKS 294 /* 295 * If the owner is running on another CPU, spin until 296 * the owner stops running or the state of the lock 297 * changes. 298 */ 299 owner = (struct thread *)RW_OWNER(x); 300 if (TD_IS_RUNNING(owner)) { 301 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 302 CTR3(KTR_LOCK, "%s: spinning on %p held by %p", 303 __func__, rw, owner); 304 #ifdef LOCK_PROFILING_SHARED 305 lock_profile_obtain_lock_failed(&rw->lock_object, 306 &contested, &waittime); 307 #endif 308 while ((struct thread*)RW_OWNER(rw->rw_lock) == owner && 309 TD_IS_RUNNING(owner)) 310 cpu_spinwait(); 311 continue; 312 } 313 #endif 314 315 /* 316 * Okay, now it's the hard case. Some other thread already 317 * has a write lock, so acquire the turnstile lock so we can 318 * begin the process of blocking. 319 */ 320 ts = turnstile_trywait(&rw->lock_object); 321 322 /* 323 * The lock might have been released while we spun, so 324 * recheck its state and restart the loop if there is no 325 * longer a write lock. 326 */ 327 x = rw->rw_lock; 328 if (x & RW_LOCK_READ) { 329 turnstile_cancel(ts); 330 cpu_spinwait(); 331 continue; 332 } 333 334 #ifdef ADAPTIVE_RWLOCKS 335 /* 336 * If the current owner of the lock is executing on another 337 * CPU quit the hard path and try to spin. 338 */ 339 owner = (struct thread *)RW_OWNER(x); 340 if (TD_IS_RUNNING(owner)) { 341 turnstile_cancel(ts); 342 cpu_spinwait(); 343 continue; 344 } 345 #endif 346 347 /* 348 * Ok, it's still a write lock. If the RW_LOCK_READ_WAITERS 349 * flag is already set, then we can go ahead and block. If 350 * it is not set then try to set it. If we fail to set it 351 * drop the turnstile lock and restart the loop. 352 */ 353 if (!(x & RW_LOCK_READ_WAITERS)) { 354 if (!atomic_cmpset_ptr(&rw->rw_lock, x, 355 x | RW_LOCK_READ_WAITERS)) { 356 turnstile_cancel(ts); 357 cpu_spinwait(); 358 continue; 359 } 360 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 361 CTR2(KTR_LOCK, "%s: %p set read waiters flag", 362 __func__, rw); 363 } 364 365 /* 366 * We were unable to acquire the lock and the read waiters 367 * flag is set, so we must block on the turnstile. 368 */ 369 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 370 CTR2(KTR_LOCK, "%s: %p blocking on turnstile", __func__, 371 rw); 372 #ifdef LOCK_PROFILING_SHARED 373 lock_profile_obtain_lock_failed(&rw->lock_object, &contested, 374 &waittime); 375 #endif 376 turnstile_wait(ts, rw_owner(rw), TS_SHARED_QUEUE); 377 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 378 CTR2(KTR_LOCK, "%s: %p resuming from turnstile", 379 __func__, rw); 380 } 381 382 /* 383 * TODO: acquire "owner of record" here. Here be turnstile dragons 384 * however. turnstiles don't like owners changing between calls to 385 * turnstile_wait() currently. 386 */ 387 388 LOCK_LOG_LOCK("RLOCK", &rw->lock_object, 0, 0, file, line); 389 WITNESS_LOCK(&rw->lock_object, 0, file, line); 390 curthread->td_locks++; 391 } 392 393 void 394 _rw_runlock(struct rwlock *rw, const char *file, int line) 395 { 396 struct turnstile *ts; 397 uintptr_t x; 398 399 KASSERT(rw->rw_lock != RW_DESTROYED, 400 ("rw_runlock() of destroyed rwlock @ %s:%d", file, line)); 401 _rw_assert(rw, RA_RLOCKED, file, line); 402 curthread->td_locks--; 403 WITNESS_UNLOCK(&rw->lock_object, 0, file, line); 404 LOCK_LOG_LOCK("RUNLOCK", &rw->lock_object, 0, 0, file, line); 405 406 /* TODO: drop "owner of record" here. */ 407 408 for (;;) { 409 /* 410 * See if there is more than one read lock held. If so, 411 * just drop one and return. 412 */ 413 x = rw->rw_lock; 414 if (RW_READERS(x) > 1) { 415 if (atomic_cmpset_ptr(&rw->rw_lock, x, 416 x - RW_ONE_READER)) { 417 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 418 CTR4(KTR_LOCK, 419 "%s: %p succeeded %p -> %p", 420 __func__, rw, (void *)x, 421 (void *)(x - RW_ONE_READER)); 422 break; 423 } 424 continue; 425 } 426 427 428 /* 429 * We should never have read waiters while at least one 430 * thread holds a read lock. (See note above) 431 */ 432 KASSERT(!(x & RW_LOCK_READ_WAITERS), 433 ("%s: waiting readers", __func__)); 434 #ifdef LOCK_PROFILING_SHARED 435 lock_profile_release_lock(&rw->lock_object); 436 #endif 437 438 /* 439 * If there aren't any waiters for a write lock, then try 440 * to drop it quickly. 441 */ 442 if (!(x & RW_LOCK_WRITE_WAITERS)) { 443 444 /* 445 * There shouldn't be any flags set and we should 446 * be the only read lock. If we fail to release 447 * the single read lock, then another thread might 448 * have just acquired a read lock, so go back up 449 * to the multiple read locks case. 450 */ 451 MPASS(x == RW_READERS_LOCK(1)); 452 if (atomic_cmpset_ptr(&rw->rw_lock, RW_READERS_LOCK(1), 453 RW_UNLOCKED)) { 454 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 455 CTR2(KTR_LOCK, "%s: %p last succeeded", 456 __func__, rw); 457 break; 458 } 459 continue; 460 } 461 462 /* 463 * There should just be one reader with one or more 464 * writers waiting. 465 */ 466 MPASS(x == (RW_READERS_LOCK(1) | RW_LOCK_WRITE_WAITERS)); 467 468 /* 469 * Ok, we know we have a waiting writer and we think we 470 * are the last reader, so grab the turnstile lock. 471 */ 472 turnstile_chain_lock(&rw->lock_object); 473 474 /* 475 * Try to drop our lock leaving the lock in a unlocked 476 * state. 477 * 478 * If you wanted to do explicit lock handoff you'd have to 479 * do it here. You'd also want to use turnstile_signal() 480 * and you'd have to handle the race where a higher 481 * priority thread blocks on the write lock before the 482 * thread you wakeup actually runs and have the new thread 483 * "steal" the lock. For now it's a lot simpler to just 484 * wakeup all of the waiters. 485 * 486 * As above, if we fail, then another thread might have 487 * acquired a read lock, so drop the turnstile lock and 488 * restart. 489 */ 490 if (!atomic_cmpset_ptr(&rw->rw_lock, 491 RW_READERS_LOCK(1) | RW_LOCK_WRITE_WAITERS, RW_UNLOCKED)) { 492 turnstile_chain_unlock(&rw->lock_object); 493 continue; 494 } 495 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 496 CTR2(KTR_LOCK, "%s: %p last succeeded with waiters", 497 __func__, rw); 498 499 /* 500 * Ok. The lock is released and all that's left is to 501 * wake up the waiters. Note that the lock might not be 502 * free anymore, but in that case the writers will just 503 * block again if they run before the new lock holder(s) 504 * release the lock. 505 */ 506 ts = turnstile_lookup(&rw->lock_object); 507 MPASS(ts != NULL); 508 turnstile_broadcast(ts, TS_EXCLUSIVE_QUEUE); 509 turnstile_unpend(ts, TS_SHARED_LOCK); 510 turnstile_chain_unlock(&rw->lock_object); 511 break; 512 } 513 } 514 515 /* 516 * This function is called when we are unable to obtain a write lock on the 517 * first try. This means that at least one other thread holds either a 518 * read or write lock. 519 */ 520 void 521 _rw_wlock_hard(struct rwlock *rw, uintptr_t tid, const char *file, int line) 522 { 523 struct turnstile *ts; 524 #ifdef ADAPTIVE_RWLOCKS 525 volatile struct thread *owner; 526 #endif 527 uint64_t waittime = 0; 528 uintptr_t v; 529 int contested = 0; 530 531 if (rw_wlocked(rw)) { 532 KASSERT(rw->lock_object.lo_flags & RW_RECURSE, 533 ("%s: recursing but non-recursive rw %s @ %s:%d\n", 534 __func__, rw->lock_object.lo_name, file, line)); 535 rw->rw_recurse++; 536 atomic_set_ptr(&rw->rw_lock, RW_LOCK_RECURSED); 537 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 538 CTR2(KTR_LOCK, "%s: %p recursing", __func__, rw); 539 return; 540 } 541 542 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 543 CTR5(KTR_LOCK, "%s: %s contested (lock=%p) at %s:%d", __func__, 544 rw->lock_object.lo_name, (void *)rw->rw_lock, file, line); 545 546 while (!_rw_write_lock(rw, tid)) { 547 #ifdef ADAPTIVE_RWLOCKS 548 /* 549 * If the lock is write locked and the owner is 550 * running on another CPU, spin until the owner stops 551 * running or the state of the lock changes. 552 */ 553 v = rw->rw_lock; 554 owner = (struct thread *)RW_OWNER(v); 555 if (!(v & RW_LOCK_READ) && TD_IS_RUNNING(owner)) { 556 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 557 CTR3(KTR_LOCK, "%s: spinning on %p held by %p", 558 __func__, rw, owner); 559 lock_profile_obtain_lock_failed(&rw->lock_object, 560 &contested, &waittime); 561 while ((struct thread*)RW_OWNER(rw->rw_lock) == owner && 562 TD_IS_RUNNING(owner)) 563 cpu_spinwait(); 564 continue; 565 } 566 #endif 567 568 ts = turnstile_trywait(&rw->lock_object); 569 v = rw->rw_lock; 570 571 /* 572 * If the lock was released while spinning on the 573 * turnstile chain lock, try again. 574 */ 575 if (v == RW_UNLOCKED) { 576 turnstile_cancel(ts); 577 cpu_spinwait(); 578 continue; 579 } 580 581 #ifdef ADAPTIVE_RWLOCKS 582 /* 583 * If the current owner of the lock is executing on another 584 * CPU quit the hard path and try to spin. 585 */ 586 if (!(v & RW_LOCK_READ)) { 587 owner = (struct thread *)RW_OWNER(v); 588 if (TD_IS_RUNNING(owner)) { 589 turnstile_cancel(ts); 590 cpu_spinwait(); 591 continue; 592 } 593 } 594 #endif 595 596 /* 597 * If the lock was released by a writer with both readers 598 * and writers waiting and a reader hasn't woken up and 599 * acquired the lock yet, rw_lock will be set to the 600 * value RW_UNLOCKED | RW_LOCK_WRITE_WAITERS. If we see 601 * that value, try to acquire it once. Note that we have 602 * to preserve the RW_LOCK_WRITE_WAITERS flag as there are 603 * other writers waiting still. If we fail, restart the 604 * loop. 605 */ 606 if (v == (RW_UNLOCKED | RW_LOCK_WRITE_WAITERS)) { 607 if (atomic_cmpset_acq_ptr(&rw->rw_lock, 608 RW_UNLOCKED | RW_LOCK_WRITE_WAITERS, 609 tid | RW_LOCK_WRITE_WAITERS)) { 610 turnstile_claim(ts); 611 CTR2(KTR_LOCK, "%s: %p claimed by new writer", 612 __func__, rw); 613 break; 614 } 615 turnstile_cancel(ts); 616 cpu_spinwait(); 617 continue; 618 } 619 620 /* 621 * If the RW_LOCK_WRITE_WAITERS flag isn't set, then try to 622 * set it. If we fail to set it, then loop back and try 623 * again. 624 */ 625 if (!(v & RW_LOCK_WRITE_WAITERS)) { 626 if (!atomic_cmpset_ptr(&rw->rw_lock, v, 627 v | RW_LOCK_WRITE_WAITERS)) { 628 turnstile_cancel(ts); 629 cpu_spinwait(); 630 continue; 631 } 632 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 633 CTR2(KTR_LOCK, "%s: %p set write waiters flag", 634 __func__, rw); 635 } 636 637 /* 638 * We were unable to acquire the lock and the write waiters 639 * flag is set, so we must block on the turnstile. 640 */ 641 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 642 CTR2(KTR_LOCK, "%s: %p blocking on turnstile", __func__, 643 rw); 644 lock_profile_obtain_lock_failed(&rw->lock_object, &contested, 645 &waittime); 646 turnstile_wait(ts, rw_owner(rw), TS_EXCLUSIVE_QUEUE); 647 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 648 CTR2(KTR_LOCK, "%s: %p resuming from turnstile", 649 __func__, rw); 650 } 651 lock_profile_obtain_lock_success(&rw->lock_object, contested, waittime, 652 file, line); 653 } 654 655 /* 656 * This function is called if the first try at releasing a write lock failed. 657 * This means that one of the 2 waiter bits must be set indicating that at 658 * least one thread is waiting on this lock. 659 */ 660 void 661 _rw_wunlock_hard(struct rwlock *rw, uintptr_t tid, const char *file, int line) 662 { 663 struct turnstile *ts; 664 uintptr_t v; 665 int queue; 666 667 if (rw_wlocked(rw) && rw_recursed(rw)) { 668 if ((--rw->rw_recurse) == 0) 669 atomic_clear_ptr(&rw->rw_lock, RW_LOCK_RECURSED); 670 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 671 CTR2(KTR_LOCK, "%s: %p unrecursing", __func__, rw); 672 return; 673 } 674 675 KASSERT(rw->rw_lock & (RW_LOCK_READ_WAITERS | RW_LOCK_WRITE_WAITERS), 676 ("%s: neither of the waiter flags are set", __func__)); 677 678 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 679 CTR2(KTR_LOCK, "%s: %p contested", __func__, rw); 680 681 turnstile_chain_lock(&rw->lock_object); 682 ts = turnstile_lookup(&rw->lock_object); 683 684 MPASS(ts != NULL); 685 686 /* 687 * Use the same algo as sx locks for now. Prefer waking up shared 688 * waiters if we have any over writers. This is probably not ideal. 689 * 690 * 'v' is the value we are going to write back to rw_lock. If we 691 * have waiters on both queues, we need to preserve the state of 692 * the waiter flag for the queue we don't wake up. For now this is 693 * hardcoded for the algorithm mentioned above. 694 * 695 * In the case of both readers and writers waiting we wakeup the 696 * readers but leave the RW_LOCK_WRITE_WAITERS flag set. If a 697 * new writer comes in before a reader it will claim the lock up 698 * above. There is probably a potential priority inversion in 699 * there that could be worked around either by waking both queues 700 * of waiters or doing some complicated lock handoff gymnastics. 701 */ 702 v = RW_UNLOCKED; 703 if (rw->rw_lock & RW_LOCK_READ_WAITERS) { 704 queue = TS_SHARED_QUEUE; 705 v |= (rw->rw_lock & RW_LOCK_WRITE_WAITERS); 706 } else 707 queue = TS_EXCLUSIVE_QUEUE; 708 709 /* Wake up all waiters for the specific queue. */ 710 if (LOCK_LOG_TEST(&rw->lock_object, 0)) 711 CTR3(KTR_LOCK, "%s: %p waking up %s waiters", __func__, rw, 712 queue == TS_SHARED_QUEUE ? "read" : "write"); 713 turnstile_broadcast(ts, queue); 714 atomic_store_rel_ptr(&rw->rw_lock, v); 715 turnstile_unpend(ts, TS_EXCLUSIVE_LOCK); 716 turnstile_chain_unlock(&rw->lock_object); 717 } 718 719 /* 720 * Attempt to do a non-blocking upgrade from a read lock to a write 721 * lock. This will only succeed if this thread holds a single read 722 * lock. Returns true if the upgrade succeeded and false otherwise. 723 */ 724 int 725 _rw_try_upgrade(struct rwlock *rw, const char *file, int line) 726 { 727 uintptr_t v, tid; 728 struct turnstile *ts; 729 int success; 730 731 KASSERT(rw->rw_lock != RW_DESTROYED, 732 ("rw_try_upgrade() of destroyed rwlock @ %s:%d", file, line)); 733 _rw_assert(rw, RA_RLOCKED, file, line); 734 735 /* 736 * Attempt to switch from one reader to a writer. If there 737 * are any write waiters, then we will have to lock the 738 * turnstile first to prevent races with another writer 739 * calling turnstile_wait() before we have claimed this 740 * turnstile. So, do the simple case of no waiters first. 741 */ 742 tid = (uintptr_t)curthread; 743 if (!(rw->rw_lock & RW_LOCK_WRITE_WAITERS)) { 744 success = atomic_cmpset_ptr(&rw->rw_lock, RW_READERS_LOCK(1), 745 tid); 746 goto out; 747 } 748 749 /* 750 * Ok, we think we have write waiters, so lock the 751 * turnstile. 752 */ 753 ts = turnstile_trywait(&rw->lock_object); 754 755 /* 756 * Try to switch from one reader to a writer again. This time 757 * we honor the current state of the RW_LOCK_WRITE_WAITERS 758 * flag. If we obtain the lock with the flag set, then claim 759 * ownership of the turnstile. 760 */ 761 v = rw->rw_lock & RW_LOCK_WRITE_WAITERS; 762 success = atomic_cmpset_ptr(&rw->rw_lock, RW_READERS_LOCK(1) | v, 763 tid | v); 764 if (success && v) 765 turnstile_claim(ts); 766 else 767 turnstile_cancel(ts); 768 out: 769 LOCK_LOG_TRY("WUPGRADE", &rw->lock_object, 0, success, file, line); 770 if (success) 771 WITNESS_UPGRADE(&rw->lock_object, LOP_EXCLUSIVE | LOP_TRYLOCK, 772 file, line); 773 return (success); 774 } 775 776 /* 777 * Downgrade a write lock into a single read lock. 778 */ 779 void 780 _rw_downgrade(struct rwlock *rw, const char *file, int line) 781 { 782 struct turnstile *ts; 783 uintptr_t tid, v; 784 785 KASSERT(rw->rw_lock != RW_DESTROYED, 786 ("rw_downgrade() of destroyed rwlock @ %s:%d", file, line)); 787 _rw_assert(rw, RA_WLOCKED | RA_NOTRECURSED, file, line); 788 #ifndef INVARIANTS 789 if (rw_recursed(rw)) 790 panic("downgrade of a recursed lock"); 791 #endif 792 793 WITNESS_DOWNGRADE(&rw->lock_object, 0, file, line); 794 795 /* 796 * Convert from a writer to a single reader. First we handle 797 * the easy case with no waiters. If there are any waiters, we 798 * lock the turnstile, "disown" the lock, and awaken any read 799 * waiters. 800 */ 801 tid = (uintptr_t)curthread; 802 if (atomic_cmpset_rel_ptr(&rw->rw_lock, tid, RW_READERS_LOCK(1))) 803 goto out; 804 805 /* 806 * Ok, we think we have waiters, so lock the turnstile so we can 807 * read the waiter flags without any races. 808 */ 809 turnstile_chain_lock(&rw->lock_object); 810 v = rw->rw_lock; 811 MPASS(v & (RW_LOCK_READ_WAITERS | RW_LOCK_WRITE_WAITERS)); 812 813 /* 814 * Downgrade from a write lock while preserving 815 * RW_LOCK_WRITE_WAITERS and give up ownership of the 816 * turnstile. If there are any read waiters, wake them up. 817 */ 818 ts = turnstile_lookup(&rw->lock_object); 819 MPASS(ts != NULL); 820 if (v & RW_LOCK_READ_WAITERS) 821 turnstile_broadcast(ts, TS_SHARED_QUEUE); 822 atomic_store_rel_ptr(&rw->rw_lock, RW_READERS_LOCK(1) | 823 (v & RW_LOCK_WRITE_WAITERS)); 824 if (v & RW_LOCK_READ_WAITERS) 825 turnstile_unpend(ts, TS_EXCLUSIVE_LOCK); 826 else if (ts) 827 turnstile_disown(ts); 828 turnstile_chain_unlock(&rw->lock_object); 829 out: 830 LOCK_LOG_LOCK("WDOWNGRADE", &rw->lock_object, 0, 0, file, line); 831 } 832 833 #ifdef INVARIANT_SUPPORT 834 #ifndef INVARIANTS 835 #undef _rw_assert 836 #endif 837 838 /* 839 * In the non-WITNESS case, rw_assert() can only detect that at least 840 * *some* thread owns an rlock, but it cannot guarantee that *this* 841 * thread owns an rlock. 842 */ 843 void 844 _rw_assert(struct rwlock *rw, int what, const char *file, int line) 845 { 846 847 if (panicstr != NULL) 848 return; 849 switch (what) { 850 case RA_LOCKED: 851 case RA_LOCKED | RA_RECURSED: 852 case RA_LOCKED | RA_NOTRECURSED: 853 case RA_RLOCKED: 854 #ifdef WITNESS 855 witness_assert(&rw->lock_object, what, file, line); 856 #else 857 /* 858 * If some other thread has a write lock or we have one 859 * and are asserting a read lock, fail. Also, if no one 860 * has a lock at all, fail. 861 */ 862 if (rw->rw_lock == RW_UNLOCKED || 863 (!(rw->rw_lock & RW_LOCK_READ) && (what == RA_RLOCKED || 864 rw_wowner(rw) != curthread))) 865 panic("Lock %s not %slocked @ %s:%d\n", 866 rw->lock_object.lo_name, (what == RA_RLOCKED) ? 867 "read " : "", file, line); 868 869 if (!(rw->rw_lock & RW_LOCK_READ)) { 870 if (rw_recursed(rw)) { 871 if (what & RA_NOTRECURSED) 872 panic("Lock %s recursed @ %s:%d\n", 873 rw->lock_object.lo_name, file, 874 line); 875 } else if (what & RA_RECURSED) 876 panic("Lock %s not recursed @ %s:%d\n", 877 rw->lock_object.lo_name, file, line); 878 } 879 #endif 880 break; 881 case RA_WLOCKED: 882 case RA_WLOCKED | RA_RECURSED: 883 case RA_WLOCKED | RA_NOTRECURSED: 884 if (rw_wowner(rw) != curthread) 885 panic("Lock %s not exclusively locked @ %s:%d\n", 886 rw->lock_object.lo_name, file, line); 887 if (rw_recursed(rw)) { 888 if (what & RA_NOTRECURSED) 889 panic("Lock %s recursed @ %s:%d\n", 890 rw->lock_object.lo_name, file, line); 891 } else if (what & RA_RECURSED) 892 panic("Lock %s not recursed @ %s:%d\n", 893 rw->lock_object.lo_name, file, line); 894 break; 895 case RA_UNLOCKED: 896 #ifdef WITNESS 897 witness_assert(&rw->lock_object, what, file, line); 898 #else 899 /* 900 * If we hold a write lock fail. We can't reliably check 901 * to see if we hold a read lock or not. 902 */ 903 if (rw_wowner(rw) == curthread) 904 panic("Lock %s exclusively locked @ %s:%d\n", 905 rw->lock_object.lo_name, file, line); 906 #endif 907 break; 908 default: 909 panic("Unknown rw lock assertion: %d @ %s:%d", what, file, 910 line); 911 } 912 } 913 #endif /* INVARIANT_SUPPORT */ 914 915 #ifdef DDB 916 void 917 db_show_rwlock(struct lock_object *lock) 918 { 919 struct rwlock *rw; 920 struct thread *td; 921 922 rw = (struct rwlock *)lock; 923 924 db_printf(" state: "); 925 if (rw->rw_lock == RW_UNLOCKED) 926 db_printf("UNLOCKED\n"); 927 else if (rw->rw_lock == RW_DESTROYED) { 928 db_printf("DESTROYED\n"); 929 return; 930 } else if (rw->rw_lock & RW_LOCK_READ) 931 db_printf("RLOCK: %ju locks\n", 932 (uintmax_t)(RW_READERS(rw->rw_lock))); 933 else { 934 td = rw_wowner(rw); 935 db_printf("WLOCK: %p (tid %d, pid %d, \"%s\")\n", td, 936 td->td_tid, td->td_proc->p_pid, td->td_name); 937 if (rw_recursed(rw)) 938 db_printf(" recursed: %u\n", rw->rw_recurse); 939 } 940 db_printf(" waiters: "); 941 switch (rw->rw_lock & (RW_LOCK_READ_WAITERS | RW_LOCK_WRITE_WAITERS)) { 942 case RW_LOCK_READ_WAITERS: 943 db_printf("readers\n"); 944 break; 945 case RW_LOCK_WRITE_WAITERS: 946 db_printf("writers\n"); 947 break; 948 case RW_LOCK_READ_WAITERS | RW_LOCK_WRITE_WAITERS: 949 db_printf("readers and writers\n"); 950 break; 951 default: 952 db_printf("none\n"); 953 break; 954 } 955 } 956 957 #endif 958