1 /*- 2 * Copyright (c) 2008 Attilio Rao <attilio@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(s), this list of conditions and the following disclaimer as 10 * the first lines of this file unmodified other than the possible 11 * addition of one or more copyright notices. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice(s), 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 COPYRIGHT HOLDER(S) ``AS IS'' AND ANY 17 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED 18 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 19 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY 20 * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES 21 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR 22 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER 23 * 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 SUCH 26 * DAMAGE. 27 */ 28 29 #include "opt_adaptive_lockmgrs.h" 30 #include "opt_ddb.h" 31 #include "opt_hwpmc_hooks.h" 32 33 #include <sys/cdefs.h> 34 __FBSDID("$FreeBSD$"); 35 36 #include <sys/param.h> 37 #include <sys/kdb.h> 38 #include <sys/ktr.h> 39 #include <sys/lock.h> 40 #include <sys/lock_profile.h> 41 #include <sys/lockmgr.h> 42 #include <sys/mutex.h> 43 #include <sys/proc.h> 44 #include <sys/sleepqueue.h> 45 #ifdef DEBUG_LOCKS 46 #include <sys/stack.h> 47 #endif 48 #include <sys/sysctl.h> 49 #include <sys/systm.h> 50 51 #include <machine/cpu.h> 52 53 #ifdef DDB 54 #include <ddb/ddb.h> 55 #endif 56 57 #ifdef HWPMC_HOOKS 58 #include <sys/pmckern.h> 59 PMC_SOFT_DECLARE( , , lock, failed); 60 #endif 61 62 CTASSERT(((LK_ADAPTIVE | LK_NOSHARE) & LO_CLASSFLAGS) == 63 (LK_ADAPTIVE | LK_NOSHARE)); 64 CTASSERT(LK_UNLOCKED == (LK_UNLOCKED & 65 ~(LK_ALL_WAITERS | LK_EXCLUSIVE_SPINNERS))); 66 67 #define SQ_EXCLUSIVE_QUEUE 0 68 #define SQ_SHARED_QUEUE 1 69 70 #ifndef INVARIANTS 71 #define _lockmgr_assert(lk, what, file, line) 72 #define TD_LOCKS_INC(td) 73 #define TD_LOCKS_DEC(td) 74 #else 75 #define TD_LOCKS_INC(td) ((td)->td_locks++) 76 #define TD_LOCKS_DEC(td) ((td)->td_locks--) 77 #endif 78 #define TD_SLOCKS_INC(td) ((td)->td_lk_slocks++) 79 #define TD_SLOCKS_DEC(td) ((td)->td_lk_slocks--) 80 81 #ifndef DEBUG_LOCKS 82 #define STACK_PRINT(lk) 83 #define STACK_SAVE(lk) 84 #define STACK_ZERO(lk) 85 #else 86 #define STACK_PRINT(lk) stack_print_ddb(&(lk)->lk_stack) 87 #define STACK_SAVE(lk) stack_save(&(lk)->lk_stack) 88 #define STACK_ZERO(lk) stack_zero(&(lk)->lk_stack) 89 #endif 90 91 #define LOCK_LOG2(lk, string, arg1, arg2) \ 92 if (LOCK_LOG_TEST(&(lk)->lock_object, 0)) \ 93 CTR2(KTR_LOCK, (string), (arg1), (arg2)) 94 #define LOCK_LOG3(lk, string, arg1, arg2, arg3) \ 95 if (LOCK_LOG_TEST(&(lk)->lock_object, 0)) \ 96 CTR3(KTR_LOCK, (string), (arg1), (arg2), (arg3)) 97 98 #define GIANT_DECLARE \ 99 int _i = 0; \ 100 WITNESS_SAVE_DECL(Giant) 101 #define GIANT_RESTORE() do { \ 102 if (_i > 0) { \ 103 while (_i--) \ 104 mtx_lock(&Giant); \ 105 WITNESS_RESTORE(&Giant.lock_object, Giant); \ 106 } \ 107 } while (0) 108 #define GIANT_SAVE() do { \ 109 if (mtx_owned(&Giant)) { \ 110 WITNESS_SAVE(&Giant.lock_object, Giant); \ 111 while (mtx_owned(&Giant)) { \ 112 _i++; \ 113 mtx_unlock(&Giant); \ 114 } \ 115 } \ 116 } while (0) 117 118 #define LK_CAN_SHARE(x, flags) \ 119 (((x) & LK_SHARE) && \ 120 (((x) & (LK_EXCLUSIVE_WAITERS | LK_EXCLUSIVE_SPINNERS)) == 0 || \ 121 (curthread->td_lk_slocks != 0 && !(flags & LK_NODDLKTREAT)) || \ 122 (curthread->td_pflags & TDP_DEADLKTREAT))) 123 #define LK_TRYOP(x) \ 124 ((x) & LK_NOWAIT) 125 126 #define LK_CAN_WITNESS(x) \ 127 (((x) & LK_NOWITNESS) == 0 && !LK_TRYOP(x)) 128 #define LK_TRYWIT(x) \ 129 (LK_TRYOP(x) ? LOP_TRYLOCK : 0) 130 131 #define LK_CAN_ADAPT(lk, f) \ 132 (((lk)->lock_object.lo_flags & LK_ADAPTIVE) != 0 && \ 133 ((f) & LK_SLEEPFAIL) == 0) 134 135 #define lockmgr_disowned(lk) \ 136 (((lk)->lk_lock & ~(LK_FLAGMASK & ~LK_SHARE)) == LK_KERNPROC) 137 138 #define lockmgr_xlocked(lk) \ 139 (((lk)->lk_lock & ~(LK_FLAGMASK & ~LK_SHARE)) == (uintptr_t)curthread) 140 141 static void assert_lockmgr(const struct lock_object *lock, int how); 142 #ifdef DDB 143 static void db_show_lockmgr(const struct lock_object *lock); 144 #endif 145 static void lock_lockmgr(struct lock_object *lock, uintptr_t how); 146 #ifdef KDTRACE_HOOKS 147 static int owner_lockmgr(const struct lock_object *lock, 148 struct thread **owner); 149 #endif 150 static uintptr_t unlock_lockmgr(struct lock_object *lock); 151 152 struct lock_class lock_class_lockmgr = { 153 .lc_name = "lockmgr", 154 .lc_flags = LC_RECURSABLE | LC_SLEEPABLE | LC_SLEEPLOCK | LC_UPGRADABLE, 155 .lc_assert = assert_lockmgr, 156 #ifdef DDB 157 .lc_ddb_show = db_show_lockmgr, 158 #endif 159 .lc_lock = lock_lockmgr, 160 .lc_unlock = unlock_lockmgr, 161 #ifdef KDTRACE_HOOKS 162 .lc_owner = owner_lockmgr, 163 #endif 164 }; 165 166 #ifdef ADAPTIVE_LOCKMGRS 167 static u_int alk_retries = 10; 168 static u_int alk_loops = 10000; 169 static SYSCTL_NODE(_debug, OID_AUTO, lockmgr, CTLFLAG_RD, NULL, 170 "lockmgr debugging"); 171 SYSCTL_UINT(_debug_lockmgr, OID_AUTO, retries, CTLFLAG_RW, &alk_retries, 0, ""); 172 SYSCTL_UINT(_debug_lockmgr, OID_AUTO, loops, CTLFLAG_RW, &alk_loops, 0, ""); 173 #endif 174 175 static __inline struct thread * 176 lockmgr_xholder(const struct lock *lk) 177 { 178 uintptr_t x; 179 180 x = lk->lk_lock; 181 return ((x & LK_SHARE) ? NULL : (struct thread *)LK_HOLDER(x)); 182 } 183 184 /* 185 * It assumes sleepq_lock held and returns with this one unheld. 186 * It also assumes the generic interlock is sane and previously checked. 187 * If LK_INTERLOCK is specified the interlock is not reacquired after the 188 * sleep. 189 */ 190 static __inline int 191 sleeplk(struct lock *lk, u_int flags, struct lock_object *ilk, 192 const char *wmesg, int pri, int timo, int queue) 193 { 194 GIANT_DECLARE; 195 struct lock_class *class; 196 int catch, error; 197 198 class = (flags & LK_INTERLOCK) ? LOCK_CLASS(ilk) : NULL; 199 catch = pri & PCATCH; 200 pri &= PRIMASK; 201 error = 0; 202 203 LOCK_LOG3(lk, "%s: %p blocking on the %s sleepqueue", __func__, lk, 204 (queue == SQ_EXCLUSIVE_QUEUE) ? "exclusive" : "shared"); 205 206 if (flags & LK_INTERLOCK) 207 class->lc_unlock(ilk); 208 if (queue == SQ_EXCLUSIVE_QUEUE && (flags & LK_SLEEPFAIL) != 0) 209 lk->lk_exslpfail++; 210 GIANT_SAVE(); 211 sleepq_add(&lk->lock_object, NULL, wmesg, SLEEPQ_LK | (catch ? 212 SLEEPQ_INTERRUPTIBLE : 0), queue); 213 if ((flags & LK_TIMELOCK) && timo) 214 sleepq_set_timeout(&lk->lock_object, timo); 215 216 /* 217 * Decisional switch for real sleeping. 218 */ 219 if ((flags & LK_TIMELOCK) && timo && catch) 220 error = sleepq_timedwait_sig(&lk->lock_object, pri); 221 else if ((flags & LK_TIMELOCK) && timo) 222 error = sleepq_timedwait(&lk->lock_object, pri); 223 else if (catch) 224 error = sleepq_wait_sig(&lk->lock_object, pri); 225 else 226 sleepq_wait(&lk->lock_object, pri); 227 GIANT_RESTORE(); 228 if ((flags & LK_SLEEPFAIL) && error == 0) 229 error = ENOLCK; 230 231 return (error); 232 } 233 234 static __inline int 235 wakeupshlk(struct lock *lk, const char *file, int line) 236 { 237 uintptr_t v, x; 238 u_int realexslp; 239 int queue, wakeup_swapper; 240 241 WITNESS_UNLOCK(&lk->lock_object, 0, file, line); 242 LOCK_LOG_LOCK("SUNLOCK", &lk->lock_object, 0, 0, file, line); 243 244 wakeup_swapper = 0; 245 for (;;) { 246 x = lk->lk_lock; 247 248 /* 249 * If there is more than one shared lock held, just drop one 250 * and return. 251 */ 252 if (LK_SHARERS(x) > 1) { 253 if (atomic_cmpset_rel_ptr(&lk->lk_lock, x, 254 x - LK_ONE_SHARER)) 255 break; 256 continue; 257 } 258 259 /* 260 * If there are not waiters on the exclusive queue, drop the 261 * lock quickly. 262 */ 263 if ((x & LK_ALL_WAITERS) == 0) { 264 MPASS((x & ~LK_EXCLUSIVE_SPINNERS) == 265 LK_SHARERS_LOCK(1)); 266 if (atomic_cmpset_rel_ptr(&lk->lk_lock, x, LK_UNLOCKED)) 267 break; 268 continue; 269 } 270 271 /* 272 * We should have a sharer with waiters, so enter the hard 273 * path in order to handle wakeups correctly. 274 */ 275 sleepq_lock(&lk->lock_object); 276 x = lk->lk_lock & (LK_ALL_WAITERS | LK_EXCLUSIVE_SPINNERS); 277 v = LK_UNLOCKED; 278 279 /* 280 * If the lock has exclusive waiters, give them preference in 281 * order to avoid deadlock with shared runners up. 282 * If interruptible sleeps left the exclusive queue empty 283 * avoid a starvation for the threads sleeping on the shared 284 * queue by giving them precedence and cleaning up the 285 * exclusive waiters bit anyway. 286 * Please note that lk_exslpfail count may be lying about 287 * the real number of waiters with the LK_SLEEPFAIL flag on 288 * because they may be used in conjuction with interruptible 289 * sleeps so lk_exslpfail might be considered an 'upper limit' 290 * bound, including the edge cases. 291 */ 292 realexslp = sleepq_sleepcnt(&lk->lock_object, 293 SQ_EXCLUSIVE_QUEUE); 294 if ((x & LK_EXCLUSIVE_WAITERS) != 0 && realexslp != 0) { 295 if (lk->lk_exslpfail < realexslp) { 296 lk->lk_exslpfail = 0; 297 queue = SQ_EXCLUSIVE_QUEUE; 298 v |= (x & LK_SHARED_WAITERS); 299 } else { 300 lk->lk_exslpfail = 0; 301 LOCK_LOG2(lk, 302 "%s: %p has only LK_SLEEPFAIL sleepers", 303 __func__, lk); 304 LOCK_LOG2(lk, 305 "%s: %p waking up threads on the exclusive queue", 306 __func__, lk); 307 wakeup_swapper = 308 sleepq_broadcast(&lk->lock_object, 309 SLEEPQ_LK, 0, SQ_EXCLUSIVE_QUEUE); 310 queue = SQ_SHARED_QUEUE; 311 } 312 313 } else { 314 315 /* 316 * Exclusive waiters sleeping with LK_SLEEPFAIL on 317 * and using interruptible sleeps/timeout may have 318 * left spourious lk_exslpfail counts on, so clean 319 * it up anyway. 320 */ 321 lk->lk_exslpfail = 0; 322 queue = SQ_SHARED_QUEUE; 323 } 324 325 if (!atomic_cmpset_rel_ptr(&lk->lk_lock, LK_SHARERS_LOCK(1) | x, 326 v)) { 327 sleepq_release(&lk->lock_object); 328 continue; 329 } 330 LOCK_LOG3(lk, "%s: %p waking up threads on the %s queue", 331 __func__, lk, queue == SQ_SHARED_QUEUE ? "shared" : 332 "exclusive"); 333 wakeup_swapper |= sleepq_broadcast(&lk->lock_object, SLEEPQ_LK, 334 0, queue); 335 sleepq_release(&lk->lock_object); 336 break; 337 } 338 339 lock_profile_release_lock(&lk->lock_object); 340 TD_LOCKS_DEC(curthread); 341 TD_SLOCKS_DEC(curthread); 342 return (wakeup_swapper); 343 } 344 345 static void 346 assert_lockmgr(const struct lock_object *lock, int what) 347 { 348 349 panic("lockmgr locks do not support assertions"); 350 } 351 352 static void 353 lock_lockmgr(struct lock_object *lock, uintptr_t how) 354 { 355 356 panic("lockmgr locks do not support sleep interlocking"); 357 } 358 359 static uintptr_t 360 unlock_lockmgr(struct lock_object *lock) 361 { 362 363 panic("lockmgr locks do not support sleep interlocking"); 364 } 365 366 #ifdef KDTRACE_HOOKS 367 static int 368 owner_lockmgr(const struct lock_object *lock, struct thread **owner) 369 { 370 371 panic("lockmgr locks do not support owner inquiring"); 372 } 373 #endif 374 375 void 376 lockinit(struct lock *lk, int pri, const char *wmesg, int timo, int flags) 377 { 378 int iflags; 379 380 MPASS((flags & ~LK_INIT_MASK) == 0); 381 ASSERT_ATOMIC_LOAD_PTR(lk->lk_lock, 382 ("%s: lockmgr not aligned for %s: %p", __func__, wmesg, 383 &lk->lk_lock)); 384 385 iflags = LO_SLEEPABLE | LO_UPGRADABLE; 386 if (flags & LK_CANRECURSE) 387 iflags |= LO_RECURSABLE; 388 if ((flags & LK_NODUP) == 0) 389 iflags |= LO_DUPOK; 390 if (flags & LK_NOPROFILE) 391 iflags |= LO_NOPROFILE; 392 if ((flags & LK_NOWITNESS) == 0) 393 iflags |= LO_WITNESS; 394 if (flags & LK_QUIET) 395 iflags |= LO_QUIET; 396 if (flags & LK_IS_VNODE) 397 iflags |= LO_IS_VNODE; 398 iflags |= flags & (LK_ADAPTIVE | LK_NOSHARE); 399 400 lock_init(&lk->lock_object, &lock_class_lockmgr, wmesg, NULL, iflags); 401 lk->lk_lock = LK_UNLOCKED; 402 lk->lk_recurse = 0; 403 lk->lk_exslpfail = 0; 404 lk->lk_timo = timo; 405 lk->lk_pri = pri; 406 STACK_ZERO(lk); 407 } 408 409 /* 410 * XXX: Gross hacks to manipulate external lock flags after 411 * initialization. Used for certain vnode and buf locks. 412 */ 413 void 414 lockallowshare(struct lock *lk) 415 { 416 417 lockmgr_assert(lk, KA_XLOCKED); 418 lk->lock_object.lo_flags &= ~LK_NOSHARE; 419 } 420 421 void 422 lockdisableshare(struct lock *lk) 423 { 424 425 lockmgr_assert(lk, KA_XLOCKED); 426 lk->lock_object.lo_flags |= LK_NOSHARE; 427 } 428 429 void 430 lockallowrecurse(struct lock *lk) 431 { 432 433 lockmgr_assert(lk, KA_XLOCKED); 434 lk->lock_object.lo_flags |= LO_RECURSABLE; 435 } 436 437 void 438 lockdisablerecurse(struct lock *lk) 439 { 440 441 lockmgr_assert(lk, KA_XLOCKED); 442 lk->lock_object.lo_flags &= ~LO_RECURSABLE; 443 } 444 445 void 446 lockdestroy(struct lock *lk) 447 { 448 449 KASSERT(lk->lk_lock == LK_UNLOCKED, ("lockmgr still held")); 450 KASSERT(lk->lk_recurse == 0, ("lockmgr still recursed")); 451 KASSERT(lk->lk_exslpfail == 0, ("lockmgr still exclusive waiters")); 452 lock_destroy(&lk->lock_object); 453 } 454 455 int 456 __lockmgr_args(struct lock *lk, u_int flags, struct lock_object *ilk, 457 const char *wmesg, int pri, int timo, const char *file, int line) 458 { 459 GIANT_DECLARE; 460 struct lock_class *class; 461 const char *iwmesg; 462 uintptr_t tid, v, x; 463 u_int op, realexslp; 464 int error, ipri, itimo, queue, wakeup_swapper; 465 #ifdef LOCK_PROFILING 466 uint64_t waittime = 0; 467 int contested = 0; 468 #endif 469 #ifdef ADAPTIVE_LOCKMGRS 470 volatile struct thread *owner; 471 u_int i, spintries = 0; 472 #endif 473 474 error = 0; 475 tid = (uintptr_t)curthread; 476 op = (flags & LK_TYPE_MASK); 477 iwmesg = (wmesg == LK_WMESG_DEFAULT) ? lk->lock_object.lo_name : wmesg; 478 ipri = (pri == LK_PRIO_DEFAULT) ? lk->lk_pri : pri; 479 itimo = (timo == LK_TIMO_DEFAULT) ? lk->lk_timo : timo; 480 481 MPASS((flags & ~LK_TOTAL_MASK) == 0); 482 KASSERT((op & (op - 1)) == 0, 483 ("%s: Invalid requested operation @ %s:%d", __func__, file, line)); 484 KASSERT((flags & (LK_NOWAIT | LK_SLEEPFAIL)) == 0 || 485 (op != LK_DOWNGRADE && op != LK_RELEASE), 486 ("%s: Invalid flags in regard of the operation desired @ %s:%d", 487 __func__, file, line)); 488 KASSERT((flags & LK_INTERLOCK) == 0 || ilk != NULL, 489 ("%s: LK_INTERLOCK passed without valid interlock @ %s:%d", 490 __func__, file, line)); 491 KASSERT(kdb_active != 0 || !TD_IS_IDLETHREAD(curthread), 492 ("%s: idle thread %p on lockmgr %s @ %s:%d", __func__, curthread, 493 lk->lock_object.lo_name, file, line)); 494 495 class = (flags & LK_INTERLOCK) ? LOCK_CLASS(ilk) : NULL; 496 if (panicstr != NULL) { 497 if (flags & LK_INTERLOCK) 498 class->lc_unlock(ilk); 499 return (0); 500 } 501 502 if (lk->lock_object.lo_flags & LK_NOSHARE) { 503 switch (op) { 504 case LK_SHARED: 505 op = LK_EXCLUSIVE; 506 break; 507 case LK_UPGRADE: 508 case LK_TRYUPGRADE: 509 case LK_DOWNGRADE: 510 _lockmgr_assert(lk, KA_XLOCKED | KA_NOTRECURSED, 511 file, line); 512 if (flags & LK_INTERLOCK) 513 class->lc_unlock(ilk); 514 return (0); 515 } 516 } 517 518 wakeup_swapper = 0; 519 switch (op) { 520 case LK_SHARED: 521 if (LK_CAN_WITNESS(flags)) 522 WITNESS_CHECKORDER(&lk->lock_object, LOP_NEWORDER, 523 file, line, flags & LK_INTERLOCK ? ilk : NULL); 524 for (;;) { 525 x = lk->lk_lock; 526 527 /* 528 * If no other thread has an exclusive lock, or 529 * no exclusive waiter is present, bump the count of 530 * sharers. Since we have to preserve the state of 531 * waiters, if we fail to acquire the shared lock 532 * loop back and retry. 533 */ 534 if (LK_CAN_SHARE(x, flags)) { 535 if (atomic_cmpset_acq_ptr(&lk->lk_lock, x, 536 x + LK_ONE_SHARER)) 537 break; 538 continue; 539 } 540 #ifdef HWPMC_HOOKS 541 PMC_SOFT_CALL( , , lock, failed); 542 #endif 543 lock_profile_obtain_lock_failed(&lk->lock_object, 544 &contested, &waittime); 545 546 /* 547 * If the lock is already held by curthread in 548 * exclusive way avoid a deadlock. 549 */ 550 if (LK_HOLDER(x) == tid) { 551 LOCK_LOG2(lk, 552 "%s: %p already held in exclusive mode", 553 __func__, lk); 554 error = EDEADLK; 555 break; 556 } 557 558 /* 559 * If the lock is expected to not sleep just give up 560 * and return. 561 */ 562 if (LK_TRYOP(flags)) { 563 LOCK_LOG2(lk, "%s: %p fails the try operation", 564 __func__, lk); 565 error = EBUSY; 566 break; 567 } 568 569 #ifdef ADAPTIVE_LOCKMGRS 570 /* 571 * If the owner is running on another CPU, spin until 572 * the owner stops running or the state of the lock 573 * changes. We need a double-state handle here 574 * because for a failed acquisition the lock can be 575 * either held in exclusive mode or shared mode 576 * (for the writer starvation avoidance technique). 577 */ 578 if (LK_CAN_ADAPT(lk, flags) && (x & LK_SHARE) == 0 && 579 LK_HOLDER(x) != LK_KERNPROC) { 580 owner = (struct thread *)LK_HOLDER(x); 581 if (LOCK_LOG_TEST(&lk->lock_object, 0)) 582 CTR3(KTR_LOCK, 583 "%s: spinning on %p held by %p", 584 __func__, lk, owner); 585 KTR_STATE1(KTR_SCHED, "thread", 586 sched_tdname(td), "spinning", 587 "lockname:\"%s\"", lk->lock_object.lo_name); 588 589 /* 590 * If we are holding also an interlock drop it 591 * in order to avoid a deadlock if the lockmgr 592 * owner is adaptively spinning on the 593 * interlock itself. 594 */ 595 if (flags & LK_INTERLOCK) { 596 class->lc_unlock(ilk); 597 flags &= ~LK_INTERLOCK; 598 } 599 GIANT_SAVE(); 600 while (LK_HOLDER(lk->lk_lock) == 601 (uintptr_t)owner && TD_IS_RUNNING(owner)) 602 cpu_spinwait(); 603 KTR_STATE0(KTR_SCHED, "thread", 604 sched_tdname(td), "running"); 605 GIANT_RESTORE(); 606 continue; 607 } else if (LK_CAN_ADAPT(lk, flags) && 608 (x & LK_SHARE) != 0 && LK_SHARERS(x) && 609 spintries < alk_retries) { 610 KTR_STATE1(KTR_SCHED, "thread", 611 sched_tdname(td), "spinning", 612 "lockname:\"%s\"", lk->lock_object.lo_name); 613 if (flags & LK_INTERLOCK) { 614 class->lc_unlock(ilk); 615 flags &= ~LK_INTERLOCK; 616 } 617 GIANT_SAVE(); 618 spintries++; 619 for (i = 0; i < alk_loops; i++) { 620 if (LOCK_LOG_TEST(&lk->lock_object, 0)) 621 CTR4(KTR_LOCK, 622 "%s: shared spinning on %p with %u and %u", 623 __func__, lk, spintries, i); 624 x = lk->lk_lock; 625 if ((x & LK_SHARE) == 0 || 626 LK_CAN_SHARE(x, flags) != 0) 627 break; 628 cpu_spinwait(); 629 } 630 KTR_STATE0(KTR_SCHED, "thread", 631 sched_tdname(td), "running"); 632 GIANT_RESTORE(); 633 if (i != alk_loops) 634 continue; 635 } 636 #endif 637 638 /* 639 * Acquire the sleepqueue chain lock because we 640 * probabilly will need to manipulate waiters flags. 641 */ 642 sleepq_lock(&lk->lock_object); 643 x = lk->lk_lock; 644 645 /* 646 * if the lock can be acquired in shared mode, try 647 * again. 648 */ 649 if (LK_CAN_SHARE(x, flags)) { 650 sleepq_release(&lk->lock_object); 651 continue; 652 } 653 654 #ifdef ADAPTIVE_LOCKMGRS 655 /* 656 * The current lock owner might have started executing 657 * on another CPU (or the lock could have changed 658 * owner) while we were waiting on the turnstile 659 * chain lock. If so, drop the turnstile lock and try 660 * again. 661 */ 662 if (LK_CAN_ADAPT(lk, flags) && (x & LK_SHARE) == 0 && 663 LK_HOLDER(x) != LK_KERNPROC) { 664 owner = (struct thread *)LK_HOLDER(x); 665 if (TD_IS_RUNNING(owner)) { 666 sleepq_release(&lk->lock_object); 667 continue; 668 } 669 } 670 #endif 671 672 /* 673 * Try to set the LK_SHARED_WAITERS flag. If we fail, 674 * loop back and retry. 675 */ 676 if ((x & LK_SHARED_WAITERS) == 0) { 677 if (!atomic_cmpset_acq_ptr(&lk->lk_lock, x, 678 x | LK_SHARED_WAITERS)) { 679 sleepq_release(&lk->lock_object); 680 continue; 681 } 682 LOCK_LOG2(lk, "%s: %p set shared waiters flag", 683 __func__, lk); 684 } 685 686 /* 687 * As far as we have been unable to acquire the 688 * shared lock and the shared waiters flag is set, 689 * we will sleep. 690 */ 691 error = sleeplk(lk, flags, ilk, iwmesg, ipri, itimo, 692 SQ_SHARED_QUEUE); 693 flags &= ~LK_INTERLOCK; 694 if (error) { 695 LOCK_LOG3(lk, 696 "%s: interrupted sleep for %p with %d", 697 __func__, lk, error); 698 break; 699 } 700 LOCK_LOG2(lk, "%s: %p resuming from the sleep queue", 701 __func__, lk); 702 } 703 if (error == 0) { 704 lock_profile_obtain_lock_success(&lk->lock_object, 705 contested, waittime, file, line); 706 LOCK_LOG_LOCK("SLOCK", &lk->lock_object, 0, 0, file, 707 line); 708 WITNESS_LOCK(&lk->lock_object, LK_TRYWIT(flags), file, 709 line); 710 TD_LOCKS_INC(curthread); 711 TD_SLOCKS_INC(curthread); 712 STACK_SAVE(lk); 713 } 714 break; 715 case LK_UPGRADE: 716 case LK_TRYUPGRADE: 717 _lockmgr_assert(lk, KA_SLOCKED, file, line); 718 v = lk->lk_lock; 719 x = v & LK_ALL_WAITERS; 720 v &= LK_EXCLUSIVE_SPINNERS; 721 722 /* 723 * Try to switch from one shared lock to an exclusive one. 724 * We need to preserve waiters flags during the operation. 725 */ 726 if (atomic_cmpset_ptr(&lk->lk_lock, LK_SHARERS_LOCK(1) | x | v, 727 tid | x)) { 728 LOCK_LOG_LOCK("XUPGRADE", &lk->lock_object, 0, 0, file, 729 line); 730 WITNESS_UPGRADE(&lk->lock_object, LOP_EXCLUSIVE | 731 LK_TRYWIT(flags), file, line); 732 TD_SLOCKS_DEC(curthread); 733 break; 734 } 735 736 /* 737 * In LK_TRYUPGRADE mode, do not drop the lock, 738 * returning EBUSY instead. 739 */ 740 if (op == LK_TRYUPGRADE) { 741 LOCK_LOG2(lk, "%s: %p failed the nowait upgrade", 742 __func__, lk); 743 error = EBUSY; 744 break; 745 } 746 747 /* 748 * We have been unable to succeed in upgrading, so just 749 * give up the shared lock. 750 */ 751 wakeup_swapper |= wakeupshlk(lk, file, line); 752 753 /* FALLTHROUGH */ 754 case LK_EXCLUSIVE: 755 if (LK_CAN_WITNESS(flags)) 756 WITNESS_CHECKORDER(&lk->lock_object, LOP_NEWORDER | 757 LOP_EXCLUSIVE, file, line, flags & LK_INTERLOCK ? 758 ilk : NULL); 759 760 /* 761 * If curthread already holds the lock and this one is 762 * allowed to recurse, simply recurse on it. 763 */ 764 if (lockmgr_xlocked(lk)) { 765 if ((flags & LK_CANRECURSE) == 0 && 766 (lk->lock_object.lo_flags & LO_RECURSABLE) == 0) { 767 768 /* 769 * If the lock is expected to not panic just 770 * give up and return. 771 */ 772 if (LK_TRYOP(flags)) { 773 LOCK_LOG2(lk, 774 "%s: %p fails the try operation", 775 __func__, lk); 776 error = EBUSY; 777 break; 778 } 779 if (flags & LK_INTERLOCK) 780 class->lc_unlock(ilk); 781 panic("%s: recursing on non recursive lockmgr %s @ %s:%d\n", 782 __func__, iwmesg, file, line); 783 } 784 lk->lk_recurse++; 785 LOCK_LOG2(lk, "%s: %p recursing", __func__, lk); 786 LOCK_LOG_LOCK("XLOCK", &lk->lock_object, 0, 787 lk->lk_recurse, file, line); 788 WITNESS_LOCK(&lk->lock_object, LOP_EXCLUSIVE | 789 LK_TRYWIT(flags), file, line); 790 TD_LOCKS_INC(curthread); 791 break; 792 } 793 794 while (!atomic_cmpset_acq_ptr(&lk->lk_lock, LK_UNLOCKED, 795 tid)) { 796 #ifdef HWPMC_HOOKS 797 PMC_SOFT_CALL( , , lock, failed); 798 #endif 799 lock_profile_obtain_lock_failed(&lk->lock_object, 800 &contested, &waittime); 801 802 /* 803 * If the lock is expected to not sleep just give up 804 * and return. 805 */ 806 if (LK_TRYOP(flags)) { 807 LOCK_LOG2(lk, "%s: %p fails the try operation", 808 __func__, lk); 809 error = EBUSY; 810 break; 811 } 812 813 #ifdef ADAPTIVE_LOCKMGRS 814 /* 815 * If the owner is running on another CPU, spin until 816 * the owner stops running or the state of the lock 817 * changes. 818 */ 819 x = lk->lk_lock; 820 if (LK_CAN_ADAPT(lk, flags) && (x & LK_SHARE) == 0 && 821 LK_HOLDER(x) != LK_KERNPROC) { 822 owner = (struct thread *)LK_HOLDER(x); 823 if (LOCK_LOG_TEST(&lk->lock_object, 0)) 824 CTR3(KTR_LOCK, 825 "%s: spinning on %p held by %p", 826 __func__, lk, owner); 827 KTR_STATE1(KTR_SCHED, "thread", 828 sched_tdname(td), "spinning", 829 "lockname:\"%s\"", lk->lock_object.lo_name); 830 831 /* 832 * If we are holding also an interlock drop it 833 * in order to avoid a deadlock if the lockmgr 834 * owner is adaptively spinning on the 835 * interlock itself. 836 */ 837 if (flags & LK_INTERLOCK) { 838 class->lc_unlock(ilk); 839 flags &= ~LK_INTERLOCK; 840 } 841 GIANT_SAVE(); 842 while (LK_HOLDER(lk->lk_lock) == 843 (uintptr_t)owner && TD_IS_RUNNING(owner)) 844 cpu_spinwait(); 845 KTR_STATE0(KTR_SCHED, "thread", 846 sched_tdname(td), "running"); 847 GIANT_RESTORE(); 848 continue; 849 } else if (LK_CAN_ADAPT(lk, flags) && 850 (x & LK_SHARE) != 0 && LK_SHARERS(x) && 851 spintries < alk_retries) { 852 if ((x & LK_EXCLUSIVE_SPINNERS) == 0 && 853 !atomic_cmpset_ptr(&lk->lk_lock, x, 854 x | LK_EXCLUSIVE_SPINNERS)) 855 continue; 856 KTR_STATE1(KTR_SCHED, "thread", 857 sched_tdname(td), "spinning", 858 "lockname:\"%s\"", lk->lock_object.lo_name); 859 if (flags & LK_INTERLOCK) { 860 class->lc_unlock(ilk); 861 flags &= ~LK_INTERLOCK; 862 } 863 GIANT_SAVE(); 864 spintries++; 865 for (i = 0; i < alk_loops; i++) { 866 if (LOCK_LOG_TEST(&lk->lock_object, 0)) 867 CTR4(KTR_LOCK, 868 "%s: shared spinning on %p with %u and %u", 869 __func__, lk, spintries, i); 870 if ((lk->lk_lock & 871 LK_EXCLUSIVE_SPINNERS) == 0) 872 break; 873 cpu_spinwait(); 874 } 875 KTR_STATE0(KTR_SCHED, "thread", 876 sched_tdname(td), "running"); 877 GIANT_RESTORE(); 878 if (i != alk_loops) 879 continue; 880 } 881 #endif 882 883 /* 884 * Acquire the sleepqueue chain lock because we 885 * probabilly will need to manipulate waiters flags. 886 */ 887 sleepq_lock(&lk->lock_object); 888 x = lk->lk_lock; 889 890 /* 891 * if the lock has been released while we spun on 892 * the sleepqueue chain lock just try again. 893 */ 894 if (x == LK_UNLOCKED) { 895 sleepq_release(&lk->lock_object); 896 continue; 897 } 898 899 #ifdef ADAPTIVE_LOCKMGRS 900 /* 901 * The current lock owner might have started executing 902 * on another CPU (or the lock could have changed 903 * owner) while we were waiting on the turnstile 904 * chain lock. If so, drop the turnstile lock and try 905 * again. 906 */ 907 if (LK_CAN_ADAPT(lk, flags) && (x & LK_SHARE) == 0 && 908 LK_HOLDER(x) != LK_KERNPROC) { 909 owner = (struct thread *)LK_HOLDER(x); 910 if (TD_IS_RUNNING(owner)) { 911 sleepq_release(&lk->lock_object); 912 continue; 913 } 914 } 915 #endif 916 917 /* 918 * The lock can be in the state where there is a 919 * pending queue of waiters, but still no owner. 920 * This happens when the lock is contested and an 921 * owner is going to claim the lock. 922 * If curthread is the one successfully acquiring it 923 * claim lock ownership and return, preserving waiters 924 * flags. 925 */ 926 v = x & (LK_ALL_WAITERS | LK_EXCLUSIVE_SPINNERS); 927 if ((x & ~v) == LK_UNLOCKED) { 928 v &= ~LK_EXCLUSIVE_SPINNERS; 929 if (atomic_cmpset_acq_ptr(&lk->lk_lock, x, 930 tid | v)) { 931 sleepq_release(&lk->lock_object); 932 LOCK_LOG2(lk, 933 "%s: %p claimed by a new writer", 934 __func__, lk); 935 break; 936 } 937 sleepq_release(&lk->lock_object); 938 continue; 939 } 940 941 /* 942 * Try to set the LK_EXCLUSIVE_WAITERS flag. If we 943 * fail, loop back and retry. 944 */ 945 if ((x & LK_EXCLUSIVE_WAITERS) == 0) { 946 if (!atomic_cmpset_ptr(&lk->lk_lock, x, 947 x | LK_EXCLUSIVE_WAITERS)) { 948 sleepq_release(&lk->lock_object); 949 continue; 950 } 951 LOCK_LOG2(lk, "%s: %p set excl waiters flag", 952 __func__, lk); 953 } 954 955 /* 956 * As far as we have been unable to acquire the 957 * exclusive lock and the exclusive waiters flag 958 * is set, we will sleep. 959 */ 960 error = sleeplk(lk, flags, ilk, iwmesg, ipri, itimo, 961 SQ_EXCLUSIVE_QUEUE); 962 flags &= ~LK_INTERLOCK; 963 if (error) { 964 LOCK_LOG3(lk, 965 "%s: interrupted sleep for %p with %d", 966 __func__, lk, error); 967 break; 968 } 969 LOCK_LOG2(lk, "%s: %p resuming from the sleep queue", 970 __func__, lk); 971 } 972 if (error == 0) { 973 lock_profile_obtain_lock_success(&lk->lock_object, 974 contested, waittime, file, line); 975 LOCK_LOG_LOCK("XLOCK", &lk->lock_object, 0, 976 lk->lk_recurse, file, line); 977 WITNESS_LOCK(&lk->lock_object, LOP_EXCLUSIVE | 978 LK_TRYWIT(flags), file, line); 979 TD_LOCKS_INC(curthread); 980 STACK_SAVE(lk); 981 } 982 break; 983 case LK_DOWNGRADE: 984 _lockmgr_assert(lk, KA_XLOCKED, file, line); 985 LOCK_LOG_LOCK("XDOWNGRADE", &lk->lock_object, 0, 0, file, line); 986 WITNESS_DOWNGRADE(&lk->lock_object, 0, file, line); 987 988 /* 989 * Panic if the lock is recursed. 990 */ 991 if (lockmgr_xlocked(lk) && lockmgr_recursed(lk)) { 992 if (flags & LK_INTERLOCK) 993 class->lc_unlock(ilk); 994 panic("%s: downgrade a recursed lockmgr %s @ %s:%d\n", 995 __func__, iwmesg, file, line); 996 } 997 TD_SLOCKS_INC(curthread); 998 999 /* 1000 * In order to preserve waiters flags, just spin. 1001 */ 1002 for (;;) { 1003 x = lk->lk_lock; 1004 MPASS((x & LK_EXCLUSIVE_SPINNERS) == 0); 1005 x &= LK_ALL_WAITERS; 1006 if (atomic_cmpset_rel_ptr(&lk->lk_lock, tid | x, 1007 LK_SHARERS_LOCK(1) | x)) 1008 break; 1009 cpu_spinwait(); 1010 } 1011 break; 1012 case LK_RELEASE: 1013 _lockmgr_assert(lk, KA_LOCKED, file, line); 1014 x = lk->lk_lock; 1015 1016 if ((x & LK_SHARE) == 0) { 1017 1018 /* 1019 * As first option, treact the lock as if it has not 1020 * any waiter. 1021 * Fix-up the tid var if the lock has been disowned. 1022 */ 1023 if (LK_HOLDER(x) == LK_KERNPROC) 1024 tid = LK_KERNPROC; 1025 else { 1026 WITNESS_UNLOCK(&lk->lock_object, LOP_EXCLUSIVE, 1027 file, line); 1028 TD_LOCKS_DEC(curthread); 1029 } 1030 LOCK_LOG_LOCK("XUNLOCK", &lk->lock_object, 0, 1031 lk->lk_recurse, file, line); 1032 1033 /* 1034 * The lock is held in exclusive mode. 1035 * If the lock is recursed also, then unrecurse it. 1036 */ 1037 if (lockmgr_xlocked(lk) && lockmgr_recursed(lk)) { 1038 LOCK_LOG2(lk, "%s: %p unrecursing", __func__, 1039 lk); 1040 lk->lk_recurse--; 1041 break; 1042 } 1043 if (tid != LK_KERNPROC) 1044 lock_profile_release_lock(&lk->lock_object); 1045 1046 if (atomic_cmpset_rel_ptr(&lk->lk_lock, tid, 1047 LK_UNLOCKED)) 1048 break; 1049 1050 sleepq_lock(&lk->lock_object); 1051 x = lk->lk_lock; 1052 v = LK_UNLOCKED; 1053 1054 /* 1055 * If the lock has exclusive waiters, give them 1056 * preference in order to avoid deadlock with 1057 * shared runners up. 1058 * If interruptible sleeps left the exclusive queue 1059 * empty avoid a starvation for the threads sleeping 1060 * on the shared queue by giving them precedence 1061 * and cleaning up the exclusive waiters bit anyway. 1062 * Please note that lk_exslpfail count may be lying 1063 * about the real number of waiters with the 1064 * LK_SLEEPFAIL flag on because they may be used in 1065 * conjuction with interruptible sleeps so 1066 * lk_exslpfail might be considered an 'upper limit' 1067 * bound, including the edge cases. 1068 */ 1069 MPASS((x & LK_EXCLUSIVE_SPINNERS) == 0); 1070 realexslp = sleepq_sleepcnt(&lk->lock_object, 1071 SQ_EXCLUSIVE_QUEUE); 1072 if ((x & LK_EXCLUSIVE_WAITERS) != 0 && realexslp != 0) { 1073 if (lk->lk_exslpfail < realexslp) { 1074 lk->lk_exslpfail = 0; 1075 queue = SQ_EXCLUSIVE_QUEUE; 1076 v |= (x & LK_SHARED_WAITERS); 1077 } else { 1078 lk->lk_exslpfail = 0; 1079 LOCK_LOG2(lk, 1080 "%s: %p has only LK_SLEEPFAIL sleepers", 1081 __func__, lk); 1082 LOCK_LOG2(lk, 1083 "%s: %p waking up threads on the exclusive queue", 1084 __func__, lk); 1085 wakeup_swapper = 1086 sleepq_broadcast(&lk->lock_object, 1087 SLEEPQ_LK, 0, SQ_EXCLUSIVE_QUEUE); 1088 queue = SQ_SHARED_QUEUE; 1089 } 1090 } else { 1091 1092 /* 1093 * Exclusive waiters sleeping with LK_SLEEPFAIL 1094 * on and using interruptible sleeps/timeout 1095 * may have left spourious lk_exslpfail counts 1096 * on, so clean it up anyway. 1097 */ 1098 lk->lk_exslpfail = 0; 1099 queue = SQ_SHARED_QUEUE; 1100 } 1101 1102 LOCK_LOG3(lk, 1103 "%s: %p waking up threads on the %s queue", 1104 __func__, lk, queue == SQ_SHARED_QUEUE ? "shared" : 1105 "exclusive"); 1106 atomic_store_rel_ptr(&lk->lk_lock, v); 1107 wakeup_swapper |= sleepq_broadcast(&lk->lock_object, 1108 SLEEPQ_LK, 0, queue); 1109 sleepq_release(&lk->lock_object); 1110 break; 1111 } else 1112 wakeup_swapper = wakeupshlk(lk, file, line); 1113 break; 1114 case LK_DRAIN: 1115 if (LK_CAN_WITNESS(flags)) 1116 WITNESS_CHECKORDER(&lk->lock_object, LOP_NEWORDER | 1117 LOP_EXCLUSIVE, file, line, flags & LK_INTERLOCK ? 1118 ilk : NULL); 1119 1120 /* 1121 * Trying to drain a lock we already own will result in a 1122 * deadlock. 1123 */ 1124 if (lockmgr_xlocked(lk)) { 1125 if (flags & LK_INTERLOCK) 1126 class->lc_unlock(ilk); 1127 panic("%s: draining %s with the lock held @ %s:%d\n", 1128 __func__, iwmesg, file, line); 1129 } 1130 1131 while (!atomic_cmpset_acq_ptr(&lk->lk_lock, LK_UNLOCKED, tid)) { 1132 #ifdef HWPMC_HOOKS 1133 PMC_SOFT_CALL( , , lock, failed); 1134 #endif 1135 lock_profile_obtain_lock_failed(&lk->lock_object, 1136 &contested, &waittime); 1137 1138 /* 1139 * If the lock is expected to not sleep just give up 1140 * and return. 1141 */ 1142 if (LK_TRYOP(flags)) { 1143 LOCK_LOG2(lk, "%s: %p fails the try operation", 1144 __func__, lk); 1145 error = EBUSY; 1146 break; 1147 } 1148 1149 /* 1150 * Acquire the sleepqueue chain lock because we 1151 * probabilly will need to manipulate waiters flags. 1152 */ 1153 sleepq_lock(&lk->lock_object); 1154 x = lk->lk_lock; 1155 1156 /* 1157 * if the lock has been released while we spun on 1158 * the sleepqueue chain lock just try again. 1159 */ 1160 if (x == LK_UNLOCKED) { 1161 sleepq_release(&lk->lock_object); 1162 continue; 1163 } 1164 1165 v = x & (LK_ALL_WAITERS | LK_EXCLUSIVE_SPINNERS); 1166 if ((x & ~v) == LK_UNLOCKED) { 1167 v = (x & ~LK_EXCLUSIVE_SPINNERS); 1168 1169 /* 1170 * If interruptible sleeps left the exclusive 1171 * queue empty avoid a starvation for the 1172 * threads sleeping on the shared queue by 1173 * giving them precedence and cleaning up the 1174 * exclusive waiters bit anyway. 1175 * Please note that lk_exslpfail count may be 1176 * lying about the real number of waiters with 1177 * the LK_SLEEPFAIL flag on because they may 1178 * be used in conjuction with interruptible 1179 * sleeps so lk_exslpfail might be considered 1180 * an 'upper limit' bound, including the edge 1181 * cases. 1182 */ 1183 if (v & LK_EXCLUSIVE_WAITERS) { 1184 queue = SQ_EXCLUSIVE_QUEUE; 1185 v &= ~LK_EXCLUSIVE_WAITERS; 1186 } else { 1187 1188 /* 1189 * Exclusive waiters sleeping with 1190 * LK_SLEEPFAIL on and using 1191 * interruptible sleeps/timeout may 1192 * have left spourious lk_exslpfail 1193 * counts on, so clean it up anyway. 1194 */ 1195 MPASS(v & LK_SHARED_WAITERS); 1196 lk->lk_exslpfail = 0; 1197 queue = SQ_SHARED_QUEUE; 1198 v &= ~LK_SHARED_WAITERS; 1199 } 1200 if (queue == SQ_EXCLUSIVE_QUEUE) { 1201 realexslp = 1202 sleepq_sleepcnt(&lk->lock_object, 1203 SQ_EXCLUSIVE_QUEUE); 1204 if (lk->lk_exslpfail >= realexslp) { 1205 lk->lk_exslpfail = 0; 1206 queue = SQ_SHARED_QUEUE; 1207 v &= ~LK_SHARED_WAITERS; 1208 if (realexslp != 0) { 1209 LOCK_LOG2(lk, 1210 "%s: %p has only LK_SLEEPFAIL sleepers", 1211 __func__, lk); 1212 LOCK_LOG2(lk, 1213 "%s: %p waking up threads on the exclusive queue", 1214 __func__, lk); 1215 wakeup_swapper = 1216 sleepq_broadcast( 1217 &lk->lock_object, 1218 SLEEPQ_LK, 0, 1219 SQ_EXCLUSIVE_QUEUE); 1220 } 1221 } else 1222 lk->lk_exslpfail = 0; 1223 } 1224 if (!atomic_cmpset_ptr(&lk->lk_lock, x, v)) { 1225 sleepq_release(&lk->lock_object); 1226 continue; 1227 } 1228 LOCK_LOG3(lk, 1229 "%s: %p waking up all threads on the %s queue", 1230 __func__, lk, queue == SQ_SHARED_QUEUE ? 1231 "shared" : "exclusive"); 1232 wakeup_swapper |= sleepq_broadcast( 1233 &lk->lock_object, SLEEPQ_LK, 0, queue); 1234 1235 /* 1236 * If shared waiters have been woken up we need 1237 * to wait for one of them to acquire the lock 1238 * before to set the exclusive waiters in 1239 * order to avoid a deadlock. 1240 */ 1241 if (queue == SQ_SHARED_QUEUE) { 1242 for (v = lk->lk_lock; 1243 (v & LK_SHARE) && !LK_SHARERS(v); 1244 v = lk->lk_lock) 1245 cpu_spinwait(); 1246 } 1247 } 1248 1249 /* 1250 * Try to set the LK_EXCLUSIVE_WAITERS flag. If we 1251 * fail, loop back and retry. 1252 */ 1253 if ((x & LK_EXCLUSIVE_WAITERS) == 0) { 1254 if (!atomic_cmpset_ptr(&lk->lk_lock, x, 1255 x | LK_EXCLUSIVE_WAITERS)) { 1256 sleepq_release(&lk->lock_object); 1257 continue; 1258 } 1259 LOCK_LOG2(lk, "%s: %p set drain waiters flag", 1260 __func__, lk); 1261 } 1262 1263 /* 1264 * As far as we have been unable to acquire the 1265 * exclusive lock and the exclusive waiters flag 1266 * is set, we will sleep. 1267 */ 1268 if (flags & LK_INTERLOCK) { 1269 class->lc_unlock(ilk); 1270 flags &= ~LK_INTERLOCK; 1271 } 1272 GIANT_SAVE(); 1273 sleepq_add(&lk->lock_object, NULL, iwmesg, SLEEPQ_LK, 1274 SQ_EXCLUSIVE_QUEUE); 1275 sleepq_wait(&lk->lock_object, ipri & PRIMASK); 1276 GIANT_RESTORE(); 1277 LOCK_LOG2(lk, "%s: %p resuming from the sleep queue", 1278 __func__, lk); 1279 } 1280 1281 if (error == 0) { 1282 lock_profile_obtain_lock_success(&lk->lock_object, 1283 contested, waittime, file, line); 1284 LOCK_LOG_LOCK("DRAIN", &lk->lock_object, 0, 1285 lk->lk_recurse, file, line); 1286 WITNESS_LOCK(&lk->lock_object, LOP_EXCLUSIVE | 1287 LK_TRYWIT(flags), file, line); 1288 TD_LOCKS_INC(curthread); 1289 STACK_SAVE(lk); 1290 } 1291 break; 1292 default: 1293 if (flags & LK_INTERLOCK) 1294 class->lc_unlock(ilk); 1295 panic("%s: unknown lockmgr request 0x%x\n", __func__, op); 1296 } 1297 1298 if (flags & LK_INTERLOCK) 1299 class->lc_unlock(ilk); 1300 if (wakeup_swapper) 1301 kick_proc0(); 1302 1303 return (error); 1304 } 1305 1306 void 1307 _lockmgr_disown(struct lock *lk, const char *file, int line) 1308 { 1309 uintptr_t tid, x; 1310 1311 if (SCHEDULER_STOPPED()) 1312 return; 1313 1314 tid = (uintptr_t)curthread; 1315 _lockmgr_assert(lk, KA_XLOCKED, file, line); 1316 1317 /* 1318 * Panic if the lock is recursed. 1319 */ 1320 if (lockmgr_xlocked(lk) && lockmgr_recursed(lk)) 1321 panic("%s: disown a recursed lockmgr @ %s:%d\n", 1322 __func__, file, line); 1323 1324 /* 1325 * If the owner is already LK_KERNPROC just skip the whole operation. 1326 */ 1327 if (LK_HOLDER(lk->lk_lock) != tid) 1328 return; 1329 lock_profile_release_lock(&lk->lock_object); 1330 LOCK_LOG_LOCK("XDISOWN", &lk->lock_object, 0, 0, file, line); 1331 WITNESS_UNLOCK(&lk->lock_object, LOP_EXCLUSIVE, file, line); 1332 TD_LOCKS_DEC(curthread); 1333 STACK_SAVE(lk); 1334 1335 /* 1336 * In order to preserve waiters flags, just spin. 1337 */ 1338 for (;;) { 1339 x = lk->lk_lock; 1340 MPASS((x & LK_EXCLUSIVE_SPINNERS) == 0); 1341 x &= LK_ALL_WAITERS; 1342 if (atomic_cmpset_rel_ptr(&lk->lk_lock, tid | x, 1343 LK_KERNPROC | x)) 1344 return; 1345 cpu_spinwait(); 1346 } 1347 } 1348 1349 void 1350 lockmgr_printinfo(const struct lock *lk) 1351 { 1352 struct thread *td; 1353 uintptr_t x; 1354 1355 if (lk->lk_lock == LK_UNLOCKED) 1356 printf("lock type %s: UNLOCKED\n", lk->lock_object.lo_name); 1357 else if (lk->lk_lock & LK_SHARE) 1358 printf("lock type %s: SHARED (count %ju)\n", 1359 lk->lock_object.lo_name, 1360 (uintmax_t)LK_SHARERS(lk->lk_lock)); 1361 else { 1362 td = lockmgr_xholder(lk); 1363 if (td == (struct thread *)LK_KERNPROC) 1364 printf("lock type %s: EXCL by KERNPROC\n", 1365 lk->lock_object.lo_name); 1366 else 1367 printf("lock type %s: EXCL by thread %p " 1368 "(pid %d, %s, tid %d)\n", lk->lock_object.lo_name, 1369 td, td->td_proc->p_pid, td->td_proc->p_comm, 1370 td->td_tid); 1371 } 1372 1373 x = lk->lk_lock; 1374 if (x & LK_EXCLUSIVE_WAITERS) 1375 printf(" with exclusive waiters pending\n"); 1376 if (x & LK_SHARED_WAITERS) 1377 printf(" with shared waiters pending\n"); 1378 if (x & LK_EXCLUSIVE_SPINNERS) 1379 printf(" with exclusive spinners pending\n"); 1380 1381 STACK_PRINT(lk); 1382 } 1383 1384 int 1385 lockstatus(const struct lock *lk) 1386 { 1387 uintptr_t v, x; 1388 int ret; 1389 1390 ret = LK_SHARED; 1391 x = lk->lk_lock; 1392 v = LK_HOLDER(x); 1393 1394 if ((x & LK_SHARE) == 0) { 1395 if (v == (uintptr_t)curthread || v == LK_KERNPROC) 1396 ret = LK_EXCLUSIVE; 1397 else 1398 ret = LK_EXCLOTHER; 1399 } else if (x == LK_UNLOCKED) 1400 ret = 0; 1401 1402 return (ret); 1403 } 1404 1405 #ifdef INVARIANT_SUPPORT 1406 1407 FEATURE(invariant_support, 1408 "Support for modules compiled with INVARIANTS option"); 1409 1410 #ifndef INVARIANTS 1411 #undef _lockmgr_assert 1412 #endif 1413 1414 void 1415 _lockmgr_assert(const struct lock *lk, int what, const char *file, int line) 1416 { 1417 int slocked = 0; 1418 1419 if (panicstr != NULL) 1420 return; 1421 switch (what) { 1422 case KA_SLOCKED: 1423 case KA_SLOCKED | KA_NOTRECURSED: 1424 case KA_SLOCKED | KA_RECURSED: 1425 slocked = 1; 1426 case KA_LOCKED: 1427 case KA_LOCKED | KA_NOTRECURSED: 1428 case KA_LOCKED | KA_RECURSED: 1429 #ifdef WITNESS 1430 1431 /* 1432 * We cannot trust WITNESS if the lock is held in exclusive 1433 * mode and a call to lockmgr_disown() happened. 1434 * Workaround this skipping the check if the lock is held in 1435 * exclusive mode even for the KA_LOCKED case. 1436 */ 1437 if (slocked || (lk->lk_lock & LK_SHARE)) { 1438 witness_assert(&lk->lock_object, what, file, line); 1439 break; 1440 } 1441 #endif 1442 if (lk->lk_lock == LK_UNLOCKED || 1443 ((lk->lk_lock & LK_SHARE) == 0 && (slocked || 1444 (!lockmgr_xlocked(lk) && !lockmgr_disowned(lk))))) 1445 panic("Lock %s not %slocked @ %s:%d\n", 1446 lk->lock_object.lo_name, slocked ? "share" : "", 1447 file, line); 1448 1449 if ((lk->lk_lock & LK_SHARE) == 0) { 1450 if (lockmgr_recursed(lk)) { 1451 if (what & KA_NOTRECURSED) 1452 panic("Lock %s recursed @ %s:%d\n", 1453 lk->lock_object.lo_name, file, 1454 line); 1455 } else if (what & KA_RECURSED) 1456 panic("Lock %s not recursed @ %s:%d\n", 1457 lk->lock_object.lo_name, file, line); 1458 } 1459 break; 1460 case KA_XLOCKED: 1461 case KA_XLOCKED | KA_NOTRECURSED: 1462 case KA_XLOCKED | KA_RECURSED: 1463 if (!lockmgr_xlocked(lk) && !lockmgr_disowned(lk)) 1464 panic("Lock %s not exclusively locked @ %s:%d\n", 1465 lk->lock_object.lo_name, file, line); 1466 if (lockmgr_recursed(lk)) { 1467 if (what & KA_NOTRECURSED) 1468 panic("Lock %s recursed @ %s:%d\n", 1469 lk->lock_object.lo_name, file, line); 1470 } else if (what & KA_RECURSED) 1471 panic("Lock %s not recursed @ %s:%d\n", 1472 lk->lock_object.lo_name, file, line); 1473 break; 1474 case KA_UNLOCKED: 1475 if (lockmgr_xlocked(lk) || lockmgr_disowned(lk)) 1476 panic("Lock %s exclusively locked @ %s:%d\n", 1477 lk->lock_object.lo_name, file, line); 1478 break; 1479 default: 1480 panic("Unknown lockmgr assertion: %d @ %s:%d\n", what, file, 1481 line); 1482 } 1483 } 1484 #endif 1485 1486 #ifdef DDB 1487 int 1488 lockmgr_chain(struct thread *td, struct thread **ownerp) 1489 { 1490 struct lock *lk; 1491 1492 lk = td->td_wchan; 1493 1494 if (LOCK_CLASS(&lk->lock_object) != &lock_class_lockmgr) 1495 return (0); 1496 db_printf("blocked on lockmgr %s", lk->lock_object.lo_name); 1497 if (lk->lk_lock & LK_SHARE) 1498 db_printf("SHARED (count %ju)\n", 1499 (uintmax_t)LK_SHARERS(lk->lk_lock)); 1500 else 1501 db_printf("EXCL\n"); 1502 *ownerp = lockmgr_xholder(lk); 1503 1504 return (1); 1505 } 1506 1507 static void 1508 db_show_lockmgr(const struct lock_object *lock) 1509 { 1510 struct thread *td; 1511 const struct lock *lk; 1512 1513 lk = (const struct lock *)lock; 1514 1515 db_printf(" state: "); 1516 if (lk->lk_lock == LK_UNLOCKED) 1517 db_printf("UNLOCKED\n"); 1518 else if (lk->lk_lock & LK_SHARE) 1519 db_printf("SLOCK: %ju\n", (uintmax_t)LK_SHARERS(lk->lk_lock)); 1520 else { 1521 td = lockmgr_xholder(lk); 1522 if (td == (struct thread *)LK_KERNPROC) 1523 db_printf("XLOCK: LK_KERNPROC\n"); 1524 else 1525 db_printf("XLOCK: %p (tid %d, pid %d, \"%s\")\n", td, 1526 td->td_tid, td->td_proc->p_pid, 1527 td->td_proc->p_comm); 1528 if (lockmgr_recursed(lk)) 1529 db_printf(" recursed: %d\n", lk->lk_recurse); 1530 } 1531 db_printf(" waiters: "); 1532 switch (lk->lk_lock & LK_ALL_WAITERS) { 1533 case LK_SHARED_WAITERS: 1534 db_printf("shared\n"); 1535 break; 1536 case LK_EXCLUSIVE_WAITERS: 1537 db_printf("exclusive\n"); 1538 break; 1539 case LK_ALL_WAITERS: 1540 db_printf("shared and exclusive\n"); 1541 break; 1542 default: 1543 db_printf("none\n"); 1544 } 1545 db_printf(" spinners: "); 1546 if (lk->lk_lock & LK_EXCLUSIVE_SPINNERS) 1547 db_printf("exclusive\n"); 1548 else 1549 db_printf("none\n"); 1550 } 1551 #endif 1552