1 /*- 2 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD 3 * 4 * Copyright (C) 2001 Julian Elischer <julian@freebsd.org>. 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(s), this list of conditions and the following disclaimer as 12 * the first lines of this file unmodified other than the possible 13 * addition of one or more copyright notices. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice(s), this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 18 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY 19 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED 20 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 21 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY 22 * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES 23 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR 24 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER 25 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH 28 * DAMAGE. 29 */ 30 31 #include "opt_witness.h" 32 #include "opt_hwpmc_hooks.h" 33 34 #include <sys/cdefs.h> 35 __FBSDID("$FreeBSD$"); 36 37 #include <sys/param.h> 38 #include <sys/systm.h> 39 #include <sys/kernel.h> 40 #include <sys/lock.h> 41 #include <sys/msan.h> 42 #include <sys/mutex.h> 43 #include <sys/proc.h> 44 #include <sys/bitstring.h> 45 #include <sys/epoch.h> 46 #include <sys/rangelock.h> 47 #include <sys/resourcevar.h> 48 #include <sys/sdt.h> 49 #include <sys/smp.h> 50 #include <sys/sched.h> 51 #include <sys/sleepqueue.h> 52 #include <sys/selinfo.h> 53 #include <sys/syscallsubr.h> 54 #include <sys/dtrace_bsd.h> 55 #include <sys/sysent.h> 56 #include <sys/turnstile.h> 57 #include <sys/taskqueue.h> 58 #include <sys/ktr.h> 59 #include <sys/rwlock.h> 60 #include <sys/umtxvar.h> 61 #include <sys/vmmeter.h> 62 #include <sys/cpuset.h> 63 #ifdef HWPMC_HOOKS 64 #include <sys/pmckern.h> 65 #endif 66 #include <sys/priv.h> 67 68 #include <security/audit/audit.h> 69 70 #include <vm/pmap.h> 71 #include <vm/vm.h> 72 #include <vm/vm_extern.h> 73 #include <vm/uma.h> 74 #include <vm/vm_phys.h> 75 #include <sys/eventhandler.h> 76 77 /* 78 * Asserts below verify the stability of struct thread and struct proc 79 * layout, as exposed by KBI to modules. On head, the KBI is allowed 80 * to drift, change to the structures must be accompanied by the 81 * assert update. 82 * 83 * On the stable branches after KBI freeze, conditions must not be 84 * violated. Typically new fields are moved to the end of the 85 * structures. 86 */ 87 #ifdef __amd64__ 88 _Static_assert(offsetof(struct thread, td_flags) == 0x108, 89 "struct thread KBI td_flags"); 90 _Static_assert(offsetof(struct thread, td_pflags) == 0x114, 91 "struct thread KBI td_pflags"); 92 _Static_assert(offsetof(struct thread, td_frame) == 0x4b0, 93 "struct thread KBI td_frame"); 94 _Static_assert(offsetof(struct thread, td_emuldata) == 0x6c0, 95 "struct thread KBI td_emuldata"); 96 _Static_assert(offsetof(struct proc, p_flag) == 0xb8, 97 "struct proc KBI p_flag"); 98 _Static_assert(offsetof(struct proc, p_pid) == 0xc4, 99 "struct proc KBI p_pid"); 100 _Static_assert(offsetof(struct proc, p_filemon) == 0x3c8, 101 "struct proc KBI p_filemon"); 102 _Static_assert(offsetof(struct proc, p_comm) == 0x3e4, 103 "struct proc KBI p_comm"); 104 _Static_assert(offsetof(struct proc, p_emuldata) == 0x4c8, 105 "struct proc KBI p_emuldata"); 106 #endif 107 #ifdef __i386__ 108 _Static_assert(offsetof(struct thread, td_flags) == 0x9c, 109 "struct thread KBI td_flags"); 110 _Static_assert(offsetof(struct thread, td_pflags) == 0xa8, 111 "struct thread KBI td_pflags"); 112 _Static_assert(offsetof(struct thread, td_frame) == 0x30c, 113 "struct thread KBI td_frame"); 114 _Static_assert(offsetof(struct thread, td_emuldata) == 0x350, 115 "struct thread KBI td_emuldata"); 116 _Static_assert(offsetof(struct proc, p_flag) == 0x6c, 117 "struct proc KBI p_flag"); 118 _Static_assert(offsetof(struct proc, p_pid) == 0x78, 119 "struct proc KBI p_pid"); 120 _Static_assert(offsetof(struct proc, p_filemon) == 0x270, 121 "struct proc KBI p_filemon"); 122 _Static_assert(offsetof(struct proc, p_comm) == 0x288, 123 "struct proc KBI p_comm"); 124 _Static_assert(offsetof(struct proc, p_emuldata) == 0x314, 125 "struct proc KBI p_emuldata"); 126 #endif 127 128 SDT_PROVIDER_DECLARE(proc); 129 SDT_PROBE_DEFINE(proc, , , lwp__exit); 130 131 /* 132 * thread related storage. 133 */ 134 static uma_zone_t thread_zone; 135 136 struct thread_domain_data { 137 struct thread *tdd_zombies; 138 int tdd_reapticks; 139 } __aligned(CACHE_LINE_SIZE); 140 141 static struct thread_domain_data thread_domain_data[MAXMEMDOM]; 142 143 static struct task thread_reap_task; 144 static struct callout thread_reap_callout; 145 146 static void thread_zombie(struct thread *); 147 static void thread_reap(void); 148 static void thread_reap_all(void); 149 static void thread_reap_task_cb(void *, int); 150 static void thread_reap_callout_cb(void *); 151 static int thread_unsuspend_one(struct thread *td, struct proc *p, 152 bool boundary); 153 static void thread_free_batched(struct thread *td); 154 155 static __exclusive_cache_line struct mtx tid_lock; 156 static bitstr_t *tid_bitmap; 157 158 static MALLOC_DEFINE(M_TIDHASH, "tidhash", "thread hash"); 159 160 static int maxthread; 161 SYSCTL_INT(_kern, OID_AUTO, maxthread, CTLFLAG_RDTUN, 162 &maxthread, 0, "Maximum number of threads"); 163 164 static __exclusive_cache_line int nthreads; 165 166 static LIST_HEAD(tidhashhead, thread) *tidhashtbl; 167 static u_long tidhash; 168 static u_long tidhashlock; 169 static struct rwlock *tidhashtbl_lock; 170 #define TIDHASH(tid) (&tidhashtbl[(tid) & tidhash]) 171 #define TIDHASHLOCK(tid) (&tidhashtbl_lock[(tid) & tidhashlock]) 172 173 EVENTHANDLER_LIST_DEFINE(thread_ctor); 174 EVENTHANDLER_LIST_DEFINE(thread_dtor); 175 EVENTHANDLER_LIST_DEFINE(thread_init); 176 EVENTHANDLER_LIST_DEFINE(thread_fini); 177 178 static bool 179 thread_count_inc_try(void) 180 { 181 int nthreads_new; 182 183 nthreads_new = atomic_fetchadd_int(&nthreads, 1) + 1; 184 if (nthreads_new >= maxthread - 100) { 185 if (priv_check_cred(curthread->td_ucred, PRIV_MAXPROC) != 0 || 186 nthreads_new >= maxthread) { 187 atomic_subtract_int(&nthreads, 1); 188 return (false); 189 } 190 } 191 return (true); 192 } 193 194 static bool 195 thread_count_inc(void) 196 { 197 static struct timeval lastfail; 198 static int curfail; 199 200 thread_reap(); 201 if (thread_count_inc_try()) { 202 return (true); 203 } 204 205 thread_reap_all(); 206 if (thread_count_inc_try()) { 207 return (true); 208 } 209 210 if (ppsratecheck(&lastfail, &curfail, 1)) { 211 printf("maxthread limit exceeded by uid %u " 212 "(pid %d); consider increasing kern.maxthread\n", 213 curthread->td_ucred->cr_ruid, curproc->p_pid); 214 } 215 return (false); 216 } 217 218 static void 219 thread_count_sub(int n) 220 { 221 222 atomic_subtract_int(&nthreads, n); 223 } 224 225 static void 226 thread_count_dec(void) 227 { 228 229 thread_count_sub(1); 230 } 231 232 static lwpid_t 233 tid_alloc(void) 234 { 235 static lwpid_t trytid; 236 lwpid_t tid; 237 238 mtx_lock(&tid_lock); 239 /* 240 * It is an invariant that the bitmap is big enough to hold maxthread 241 * IDs. If we got to this point there has to be at least one free. 242 */ 243 if (trytid >= maxthread) 244 trytid = 0; 245 bit_ffc_at(tid_bitmap, trytid, maxthread, &tid); 246 if (tid == -1) { 247 KASSERT(trytid != 0, ("unexpectedly ran out of IDs")); 248 trytid = 0; 249 bit_ffc_at(tid_bitmap, trytid, maxthread, &tid); 250 KASSERT(tid != -1, ("unexpectedly ran out of IDs")); 251 } 252 bit_set(tid_bitmap, tid); 253 trytid = tid + 1; 254 mtx_unlock(&tid_lock); 255 return (tid + NO_PID); 256 } 257 258 static void 259 tid_free_locked(lwpid_t rtid) 260 { 261 lwpid_t tid; 262 263 mtx_assert(&tid_lock, MA_OWNED); 264 KASSERT(rtid >= NO_PID, 265 ("%s: invalid tid %d\n", __func__, rtid)); 266 tid = rtid - NO_PID; 267 KASSERT(bit_test(tid_bitmap, tid) != 0, 268 ("thread ID %d not allocated\n", rtid)); 269 bit_clear(tid_bitmap, tid); 270 } 271 272 static void 273 tid_free(lwpid_t rtid) 274 { 275 276 mtx_lock(&tid_lock); 277 tid_free_locked(rtid); 278 mtx_unlock(&tid_lock); 279 } 280 281 static void 282 tid_free_batch(lwpid_t *batch, int n) 283 { 284 int i; 285 286 mtx_lock(&tid_lock); 287 for (i = 0; i < n; i++) { 288 tid_free_locked(batch[i]); 289 } 290 mtx_unlock(&tid_lock); 291 } 292 293 /* 294 * Batching for thread reapping. 295 */ 296 struct tidbatch { 297 lwpid_t tab[16]; 298 int n; 299 }; 300 301 static void 302 tidbatch_prep(struct tidbatch *tb) 303 { 304 305 tb->n = 0; 306 } 307 308 static void 309 tidbatch_add(struct tidbatch *tb, struct thread *td) 310 { 311 312 KASSERT(tb->n < nitems(tb->tab), 313 ("%s: count too high %d", __func__, tb->n)); 314 tb->tab[tb->n] = td->td_tid; 315 tb->n++; 316 } 317 318 static void 319 tidbatch_process(struct tidbatch *tb) 320 { 321 322 KASSERT(tb->n <= nitems(tb->tab), 323 ("%s: count too high %d", __func__, tb->n)); 324 if (tb->n == nitems(tb->tab)) { 325 tid_free_batch(tb->tab, tb->n); 326 tb->n = 0; 327 } 328 } 329 330 static void 331 tidbatch_final(struct tidbatch *tb) 332 { 333 334 KASSERT(tb->n <= nitems(tb->tab), 335 ("%s: count too high %d", __func__, tb->n)); 336 if (tb->n != 0) { 337 tid_free_batch(tb->tab, tb->n); 338 } 339 } 340 341 /* 342 * Prepare a thread for use. 343 */ 344 static int 345 thread_ctor(void *mem, int size, void *arg, int flags) 346 { 347 struct thread *td; 348 349 td = (struct thread *)mem; 350 TD_SET_STATE(td, TDS_INACTIVE); 351 td->td_lastcpu = td->td_oncpu = NOCPU; 352 353 /* 354 * Note that td_critnest begins life as 1 because the thread is not 355 * running and is thereby implicitly waiting to be on the receiving 356 * end of a context switch. 357 */ 358 td->td_critnest = 1; 359 td->td_lend_user_pri = PRI_MAX; 360 #ifdef AUDIT 361 audit_thread_alloc(td); 362 #endif 363 #ifdef KDTRACE_HOOKS 364 kdtrace_thread_ctor(td); 365 #endif 366 umtx_thread_alloc(td); 367 MPASS(td->td_sel == NULL); 368 return (0); 369 } 370 371 /* 372 * Reclaim a thread after use. 373 */ 374 static void 375 thread_dtor(void *mem, int size, void *arg) 376 { 377 struct thread *td; 378 379 td = (struct thread *)mem; 380 381 #ifdef INVARIANTS 382 /* Verify that this thread is in a safe state to free. */ 383 switch (TD_GET_STATE(td)) { 384 case TDS_INHIBITED: 385 case TDS_RUNNING: 386 case TDS_CAN_RUN: 387 case TDS_RUNQ: 388 /* 389 * We must never unlink a thread that is in one of 390 * these states, because it is currently active. 391 */ 392 panic("bad state for thread unlinking"); 393 /* NOTREACHED */ 394 case TDS_INACTIVE: 395 break; 396 default: 397 panic("bad thread state"); 398 /* NOTREACHED */ 399 } 400 #endif 401 #ifdef AUDIT 402 audit_thread_free(td); 403 #endif 404 #ifdef KDTRACE_HOOKS 405 kdtrace_thread_dtor(td); 406 #endif 407 /* Free all OSD associated to this thread. */ 408 osd_thread_exit(td); 409 ast_kclear(td); 410 seltdfini(td); 411 } 412 413 /* 414 * Initialize type-stable parts of a thread (when newly created). 415 */ 416 static int 417 thread_init(void *mem, int size, int flags) 418 { 419 struct thread *td; 420 421 td = (struct thread *)mem; 422 423 td->td_allocdomain = vm_phys_domain(vtophys(td)); 424 td->td_sleepqueue = sleepq_alloc(); 425 td->td_turnstile = turnstile_alloc(); 426 td->td_rlqe = NULL; 427 EVENTHANDLER_DIRECT_INVOKE(thread_init, td); 428 umtx_thread_init(td); 429 td->td_kstack = 0; 430 td->td_sel = NULL; 431 return (0); 432 } 433 434 /* 435 * Tear down type-stable parts of a thread (just before being discarded). 436 */ 437 static void 438 thread_fini(void *mem, int size) 439 { 440 struct thread *td; 441 442 td = (struct thread *)mem; 443 EVENTHANDLER_DIRECT_INVOKE(thread_fini, td); 444 rlqentry_free(td->td_rlqe); 445 turnstile_free(td->td_turnstile); 446 sleepq_free(td->td_sleepqueue); 447 umtx_thread_fini(td); 448 MPASS(td->td_sel == NULL); 449 } 450 451 /* 452 * For a newly created process, 453 * link up all the structures and its initial threads etc. 454 * called from: 455 * {arch}/{arch}/machdep.c {arch}_init(), init386() etc. 456 * proc_dtor() (should go away) 457 * proc_init() 458 */ 459 void 460 proc_linkup0(struct proc *p, struct thread *td) 461 { 462 TAILQ_INIT(&p->p_threads); /* all threads in proc */ 463 proc_linkup(p, td); 464 } 465 466 void 467 proc_linkup(struct proc *p, struct thread *td) 468 { 469 470 sigqueue_init(&p->p_sigqueue, p); 471 p->p_ksi = ksiginfo_alloc(1); 472 if (p->p_ksi != NULL) { 473 /* XXX p_ksi may be null if ksiginfo zone is not ready */ 474 p->p_ksi->ksi_flags = KSI_EXT | KSI_INS; 475 } 476 LIST_INIT(&p->p_mqnotifier); 477 p->p_numthreads = 0; 478 thread_link(td, p); 479 } 480 481 static void 482 ast_suspend(struct thread *td, int tda __unused) 483 { 484 struct proc *p; 485 486 p = td->td_proc; 487 /* 488 * We need to check to see if we have to exit or wait due to a 489 * single threading requirement or some other STOP condition. 490 */ 491 PROC_LOCK(p); 492 thread_suspend_check(0); 493 PROC_UNLOCK(p); 494 } 495 496 extern int max_threads_per_proc; 497 498 /* 499 * Initialize global thread allocation resources. 500 */ 501 void 502 threadinit(void) 503 { 504 u_long i; 505 lwpid_t tid0; 506 uint32_t flags; 507 508 /* 509 * Place an upper limit on threads which can be allocated. 510 * 511 * Note that other factors may make the de facto limit much lower. 512 * 513 * Platform limits are somewhat arbitrary but deemed "more than good 514 * enough" for the foreseable future. 515 */ 516 if (maxthread == 0) { 517 #ifdef _LP64 518 maxthread = MIN(maxproc * max_threads_per_proc, 1000000); 519 #else 520 maxthread = MIN(maxproc * max_threads_per_proc, 100000); 521 #endif 522 } 523 524 mtx_init(&tid_lock, "TID lock", NULL, MTX_DEF); 525 tid_bitmap = bit_alloc(maxthread, M_TIDHASH, M_WAITOK); 526 /* 527 * Handle thread0. 528 */ 529 thread_count_inc(); 530 tid0 = tid_alloc(); 531 if (tid0 != THREAD0_TID) 532 panic("tid0 %d != %d\n", tid0, THREAD0_TID); 533 534 flags = UMA_ZONE_NOFREE; 535 #ifdef __aarch64__ 536 /* 537 * Force thread structures to be allocated from the direct map. 538 * Otherwise, superpage promotions and demotions may temporarily 539 * invalidate thread structure mappings. For most dynamically allocated 540 * structures this is not a problem, but translation faults cannot be 541 * handled without accessing curthread. 542 */ 543 flags |= UMA_ZONE_CONTIG; 544 #endif 545 thread_zone = uma_zcreate("THREAD", sched_sizeof_thread(), 546 thread_ctor, thread_dtor, thread_init, thread_fini, 547 32 - 1, flags); 548 tidhashtbl = hashinit(maxproc / 2, M_TIDHASH, &tidhash); 549 tidhashlock = (tidhash + 1) / 64; 550 if (tidhashlock > 0) 551 tidhashlock--; 552 tidhashtbl_lock = malloc(sizeof(*tidhashtbl_lock) * (tidhashlock + 1), 553 M_TIDHASH, M_WAITOK | M_ZERO); 554 for (i = 0; i < tidhashlock + 1; i++) 555 rw_init(&tidhashtbl_lock[i], "tidhash"); 556 557 TASK_INIT(&thread_reap_task, 0, thread_reap_task_cb, NULL); 558 callout_init(&thread_reap_callout, 1); 559 callout_reset(&thread_reap_callout, 5 * hz, 560 thread_reap_callout_cb, NULL); 561 ast_register(TDA_SUSPEND, ASTR_ASTF_REQUIRED, 0, ast_suspend); 562 } 563 564 /* 565 * Place an unused thread on the zombie list. 566 */ 567 void 568 thread_zombie(struct thread *td) 569 { 570 struct thread_domain_data *tdd; 571 struct thread *ztd; 572 573 tdd = &thread_domain_data[td->td_allocdomain]; 574 ztd = atomic_load_ptr(&tdd->tdd_zombies); 575 for (;;) { 576 td->td_zombie = ztd; 577 if (atomic_fcmpset_rel_ptr((uintptr_t *)&tdd->tdd_zombies, 578 (uintptr_t *)&ztd, (uintptr_t)td)) 579 break; 580 continue; 581 } 582 } 583 584 /* 585 * Release a thread that has exited after cpu_throw(). 586 */ 587 void 588 thread_stash(struct thread *td) 589 { 590 atomic_subtract_rel_int(&td->td_proc->p_exitthreads, 1); 591 thread_zombie(td); 592 } 593 594 /* 595 * Reap zombies from passed domain. 596 */ 597 static void 598 thread_reap_domain(struct thread_domain_data *tdd) 599 { 600 struct thread *itd, *ntd; 601 struct tidbatch tidbatch; 602 struct credbatch credbatch; 603 int tdcount; 604 struct plimit *lim; 605 int limcount; 606 607 /* 608 * Reading upfront is pessimal if followed by concurrent atomic_swap, 609 * but most of the time the list is empty. 610 */ 611 if (tdd->tdd_zombies == NULL) 612 return; 613 614 itd = (struct thread *)atomic_swap_ptr((uintptr_t *)&tdd->tdd_zombies, 615 (uintptr_t)NULL); 616 if (itd == NULL) 617 return; 618 619 /* 620 * Multiple CPUs can get here, the race is fine as ticks is only 621 * advisory. 622 */ 623 tdd->tdd_reapticks = ticks; 624 625 tidbatch_prep(&tidbatch); 626 credbatch_prep(&credbatch); 627 tdcount = 0; 628 lim = NULL; 629 limcount = 0; 630 631 while (itd != NULL) { 632 ntd = itd->td_zombie; 633 EVENTHANDLER_DIRECT_INVOKE(thread_dtor, itd); 634 tidbatch_add(&tidbatch, itd); 635 credbatch_add(&credbatch, itd); 636 MPASS(itd->td_limit != NULL); 637 if (lim != itd->td_limit) { 638 if (limcount != 0) { 639 lim_freen(lim, limcount); 640 limcount = 0; 641 } 642 } 643 lim = itd->td_limit; 644 limcount++; 645 thread_free_batched(itd); 646 tidbatch_process(&tidbatch); 647 credbatch_process(&credbatch); 648 tdcount++; 649 if (tdcount == 32) { 650 thread_count_sub(tdcount); 651 tdcount = 0; 652 } 653 itd = ntd; 654 } 655 656 tidbatch_final(&tidbatch); 657 credbatch_final(&credbatch); 658 if (tdcount != 0) { 659 thread_count_sub(tdcount); 660 } 661 MPASS(limcount != 0); 662 lim_freen(lim, limcount); 663 } 664 665 /* 666 * Reap zombies from all domains. 667 */ 668 static void 669 thread_reap_all(void) 670 { 671 struct thread_domain_data *tdd; 672 int i, domain; 673 674 domain = PCPU_GET(domain); 675 for (i = 0; i < vm_ndomains; i++) { 676 tdd = &thread_domain_data[(i + domain) % vm_ndomains]; 677 thread_reap_domain(tdd); 678 } 679 } 680 681 /* 682 * Reap zombies from local domain. 683 */ 684 static void 685 thread_reap(void) 686 { 687 struct thread_domain_data *tdd; 688 int domain; 689 690 domain = PCPU_GET(domain); 691 tdd = &thread_domain_data[domain]; 692 693 thread_reap_domain(tdd); 694 } 695 696 static void 697 thread_reap_task_cb(void *arg __unused, int pending __unused) 698 { 699 700 thread_reap_all(); 701 } 702 703 static void 704 thread_reap_callout_cb(void *arg __unused) 705 { 706 struct thread_domain_data *tdd; 707 int i, cticks, lticks; 708 bool wantreap; 709 710 wantreap = false; 711 cticks = atomic_load_int(&ticks); 712 for (i = 0; i < vm_ndomains; i++) { 713 tdd = &thread_domain_data[i]; 714 lticks = tdd->tdd_reapticks; 715 if (tdd->tdd_zombies != NULL && 716 (u_int)(cticks - lticks) > 5 * hz) { 717 wantreap = true; 718 break; 719 } 720 } 721 722 if (wantreap) 723 taskqueue_enqueue(taskqueue_thread, &thread_reap_task); 724 callout_reset(&thread_reap_callout, 5 * hz, 725 thread_reap_callout_cb, NULL); 726 } 727 728 /* 729 * Calling this function guarantees that any thread that exited before 730 * the call is reaped when the function returns. By 'exited' we mean 731 * a thread removed from the process linkage with thread_unlink(). 732 * Practically this means that caller must lock/unlock corresponding 733 * process lock before the call, to synchronize with thread_exit(). 734 */ 735 void 736 thread_reap_barrier(void) 737 { 738 struct task *t; 739 740 /* 741 * First do context switches to each CPU to ensure that all 742 * PCPU pc_deadthreads are moved to zombie list. 743 */ 744 quiesce_all_cpus("", PDROP); 745 746 /* 747 * Second, fire the task in the same thread as normal 748 * thread_reap() is done, to serialize reaping. 749 */ 750 t = malloc(sizeof(*t), M_TEMP, M_WAITOK); 751 TASK_INIT(t, 0, thread_reap_task_cb, t); 752 taskqueue_enqueue(taskqueue_thread, t); 753 taskqueue_drain(taskqueue_thread, t); 754 free(t, M_TEMP); 755 } 756 757 /* 758 * Allocate a thread. 759 */ 760 struct thread * 761 thread_alloc(int pages) 762 { 763 struct thread *td; 764 lwpid_t tid; 765 766 if (!thread_count_inc()) { 767 return (NULL); 768 } 769 770 tid = tid_alloc(); 771 td = uma_zalloc(thread_zone, M_WAITOK); 772 KASSERT(td->td_kstack == 0, ("thread_alloc got thread with kstack")); 773 if (!vm_thread_new(td, pages)) { 774 uma_zfree(thread_zone, td); 775 tid_free(tid); 776 thread_count_dec(); 777 return (NULL); 778 } 779 td->td_tid = tid; 780 bzero(&td->td_sa.args, sizeof(td->td_sa.args)); 781 kmsan_thread_alloc(td); 782 cpu_thread_alloc(td); 783 EVENTHANDLER_DIRECT_INVOKE(thread_ctor, td); 784 return (td); 785 } 786 787 int 788 thread_alloc_stack(struct thread *td, int pages) 789 { 790 791 KASSERT(td->td_kstack == 0, 792 ("thread_alloc_stack called on a thread with kstack")); 793 if (!vm_thread_new(td, pages)) 794 return (0); 795 cpu_thread_alloc(td); 796 return (1); 797 } 798 799 /* 800 * Deallocate a thread. 801 */ 802 static void 803 thread_free_batched(struct thread *td) 804 { 805 806 lock_profile_thread_exit(td); 807 if (td->td_cpuset) 808 cpuset_rel(td->td_cpuset); 809 td->td_cpuset = NULL; 810 cpu_thread_free(td); 811 if (td->td_kstack != 0) 812 vm_thread_dispose(td); 813 callout_drain(&td->td_slpcallout); 814 /* 815 * Freeing handled by the caller. 816 */ 817 td->td_tid = -1; 818 kmsan_thread_free(td); 819 uma_zfree(thread_zone, td); 820 } 821 822 void 823 thread_free(struct thread *td) 824 { 825 lwpid_t tid; 826 827 EVENTHANDLER_DIRECT_INVOKE(thread_dtor, td); 828 tid = td->td_tid; 829 thread_free_batched(td); 830 tid_free(tid); 831 thread_count_dec(); 832 } 833 834 void 835 thread_cow_get_proc(struct thread *newtd, struct proc *p) 836 { 837 838 PROC_LOCK_ASSERT(p, MA_OWNED); 839 newtd->td_realucred = crcowget(p->p_ucred); 840 newtd->td_ucred = newtd->td_realucred; 841 newtd->td_limit = lim_hold(p->p_limit); 842 newtd->td_cowgen = p->p_cowgen; 843 } 844 845 void 846 thread_cow_get(struct thread *newtd, struct thread *td) 847 { 848 849 MPASS(td->td_realucred == td->td_ucred); 850 newtd->td_realucred = crcowget(td->td_realucred); 851 newtd->td_ucred = newtd->td_realucred; 852 newtd->td_limit = lim_hold(td->td_limit); 853 newtd->td_cowgen = td->td_cowgen; 854 } 855 856 void 857 thread_cow_free(struct thread *td) 858 { 859 860 if (td->td_realucred != NULL) 861 crcowfree(td); 862 if (td->td_limit != NULL) 863 lim_free(td->td_limit); 864 } 865 866 void 867 thread_cow_update(struct thread *td) 868 { 869 struct proc *p; 870 struct ucred *oldcred; 871 struct plimit *oldlimit; 872 873 p = td->td_proc; 874 PROC_LOCK(p); 875 oldcred = crcowsync(); 876 oldlimit = lim_cowsync(); 877 td->td_cowgen = p->p_cowgen; 878 PROC_UNLOCK(p); 879 if (oldcred != NULL) 880 crfree(oldcred); 881 if (oldlimit != NULL) 882 lim_free(oldlimit); 883 } 884 885 void 886 thread_cow_synced(struct thread *td) 887 { 888 struct proc *p; 889 890 p = td->td_proc; 891 PROC_LOCK_ASSERT(p, MA_OWNED); 892 MPASS(td->td_cowgen != p->p_cowgen); 893 MPASS(td->td_ucred == p->p_ucred); 894 MPASS(td->td_limit == p->p_limit); 895 td->td_cowgen = p->p_cowgen; 896 } 897 898 /* 899 * Discard the current thread and exit from its context. 900 * Always called with scheduler locked. 901 * 902 * Because we can't free a thread while we're operating under its context, 903 * push the current thread into our CPU's deadthread holder. This means 904 * we needn't worry about someone else grabbing our context before we 905 * do a cpu_throw(). 906 */ 907 void 908 thread_exit(void) 909 { 910 uint64_t runtime, new_switchtime; 911 struct thread *td; 912 struct thread *td2; 913 struct proc *p; 914 int wakeup_swapper; 915 916 td = curthread; 917 p = td->td_proc; 918 919 PROC_SLOCK_ASSERT(p, MA_OWNED); 920 mtx_assert(&Giant, MA_NOTOWNED); 921 922 PROC_LOCK_ASSERT(p, MA_OWNED); 923 KASSERT(p != NULL, ("thread exiting without a process")); 924 CTR3(KTR_PROC, "thread_exit: thread %p (pid %ld, %s)", td, 925 (long)p->p_pid, td->td_name); 926 SDT_PROBE0(proc, , , lwp__exit); 927 KASSERT(TAILQ_EMPTY(&td->td_sigqueue.sq_list), ("signal pending")); 928 MPASS(td->td_realucred == td->td_ucred); 929 930 /* 931 * drop FPU & debug register state storage, or any other 932 * architecture specific resources that 933 * would not be on a new untouched process. 934 */ 935 cpu_thread_exit(td); 936 937 /* 938 * The last thread is left attached to the process 939 * So that the whole bundle gets recycled. Skip 940 * all this stuff if we never had threads. 941 * EXIT clears all sign of other threads when 942 * it goes to single threading, so the last thread always 943 * takes the short path. 944 */ 945 if (p->p_flag & P_HADTHREADS) { 946 if (p->p_numthreads > 1) { 947 atomic_add_int(&td->td_proc->p_exitthreads, 1); 948 thread_unlink(td); 949 td2 = FIRST_THREAD_IN_PROC(p); 950 sched_exit_thread(td2, td); 951 952 /* 953 * The test below is NOT true if we are the 954 * sole exiting thread. P_STOPPED_SINGLE is unset 955 * in exit1() after it is the only survivor. 956 */ 957 if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) { 958 if (p->p_numthreads == p->p_suspcount) { 959 thread_lock(p->p_singlethread); 960 wakeup_swapper = thread_unsuspend_one( 961 p->p_singlethread, p, false); 962 if (wakeup_swapper) 963 kick_proc0(); 964 } 965 } 966 967 PCPU_SET(deadthread, td); 968 } else { 969 /* 970 * The last thread is exiting.. but not through exit() 971 */ 972 panic ("thread_exit: Last thread exiting on its own"); 973 } 974 } 975 #ifdef HWPMC_HOOKS 976 /* 977 * If this thread is part of a process that is being tracked by hwpmc(4), 978 * inform the module of the thread's impending exit. 979 */ 980 if (PMC_PROC_IS_USING_PMCS(td->td_proc)) { 981 PMC_SWITCH_CONTEXT(td, PMC_FN_CSW_OUT); 982 PMC_CALL_HOOK_UNLOCKED(td, PMC_FN_THR_EXIT, NULL); 983 } else if (PMC_SYSTEM_SAMPLING_ACTIVE()) 984 PMC_CALL_HOOK_UNLOCKED(td, PMC_FN_THR_EXIT_LOG, NULL); 985 #endif 986 PROC_UNLOCK(p); 987 PROC_STATLOCK(p); 988 thread_lock(td); 989 PROC_SUNLOCK(p); 990 991 /* Do the same timestamp bookkeeping that mi_switch() would do. */ 992 new_switchtime = cpu_ticks(); 993 runtime = new_switchtime - PCPU_GET(switchtime); 994 td->td_runtime += runtime; 995 td->td_incruntime += runtime; 996 PCPU_SET(switchtime, new_switchtime); 997 PCPU_SET(switchticks, ticks); 998 VM_CNT_INC(v_swtch); 999 1000 /* Save our resource usage in our process. */ 1001 td->td_ru.ru_nvcsw++; 1002 ruxagg_locked(p, td); 1003 rucollect(&p->p_ru, &td->td_ru); 1004 PROC_STATUNLOCK(p); 1005 1006 TD_SET_STATE(td, TDS_INACTIVE); 1007 #ifdef WITNESS 1008 witness_thread_exit(td); 1009 #endif 1010 CTR1(KTR_PROC, "thread_exit: cpu_throw() thread %p", td); 1011 sched_throw(td); 1012 panic("I'm a teapot!"); 1013 /* NOTREACHED */ 1014 } 1015 1016 /* 1017 * Do any thread specific cleanups that may be needed in wait() 1018 * called with Giant, proc and schedlock not held. 1019 */ 1020 void 1021 thread_wait(struct proc *p) 1022 { 1023 struct thread *td; 1024 1025 mtx_assert(&Giant, MA_NOTOWNED); 1026 KASSERT(p->p_numthreads == 1, ("multiple threads in thread_wait()")); 1027 KASSERT(p->p_exitthreads == 0, ("p_exitthreads leaking")); 1028 td = FIRST_THREAD_IN_PROC(p); 1029 /* Lock the last thread so we spin until it exits cpu_throw(). */ 1030 thread_lock(td); 1031 thread_unlock(td); 1032 lock_profile_thread_exit(td); 1033 cpuset_rel(td->td_cpuset); 1034 td->td_cpuset = NULL; 1035 cpu_thread_clean(td); 1036 thread_cow_free(td); 1037 callout_drain(&td->td_slpcallout); 1038 thread_reap(); /* check for zombie threads etc. */ 1039 } 1040 1041 /* 1042 * Link a thread to a process. 1043 * set up anything that needs to be initialized for it to 1044 * be used by the process. 1045 */ 1046 void 1047 thread_link(struct thread *td, struct proc *p) 1048 { 1049 1050 /* 1051 * XXX This can't be enabled because it's called for proc0 before 1052 * its lock has been created. 1053 * PROC_LOCK_ASSERT(p, MA_OWNED); 1054 */ 1055 TD_SET_STATE(td, TDS_INACTIVE); 1056 td->td_proc = p; 1057 td->td_flags = TDF_INMEM; 1058 1059 LIST_INIT(&td->td_contested); 1060 LIST_INIT(&td->td_lprof[0]); 1061 LIST_INIT(&td->td_lprof[1]); 1062 #ifdef EPOCH_TRACE 1063 SLIST_INIT(&td->td_epochs); 1064 #endif 1065 sigqueue_init(&td->td_sigqueue, p); 1066 callout_init(&td->td_slpcallout, 1); 1067 TAILQ_INSERT_TAIL(&p->p_threads, td, td_plist); 1068 p->p_numthreads++; 1069 } 1070 1071 /* 1072 * Called from: 1073 * thread_exit() 1074 */ 1075 void 1076 thread_unlink(struct thread *td) 1077 { 1078 struct proc *p = td->td_proc; 1079 1080 PROC_LOCK_ASSERT(p, MA_OWNED); 1081 #ifdef EPOCH_TRACE 1082 MPASS(SLIST_EMPTY(&td->td_epochs)); 1083 #endif 1084 1085 TAILQ_REMOVE(&p->p_threads, td, td_plist); 1086 p->p_numthreads--; 1087 /* could clear a few other things here */ 1088 /* Must NOT clear links to proc! */ 1089 } 1090 1091 static int 1092 calc_remaining(struct proc *p, int mode) 1093 { 1094 int remaining; 1095 1096 PROC_LOCK_ASSERT(p, MA_OWNED); 1097 PROC_SLOCK_ASSERT(p, MA_OWNED); 1098 if (mode == SINGLE_EXIT) 1099 remaining = p->p_numthreads; 1100 else if (mode == SINGLE_BOUNDARY) 1101 remaining = p->p_numthreads - p->p_boundary_count; 1102 else if (mode == SINGLE_NO_EXIT || mode == SINGLE_ALLPROC) 1103 remaining = p->p_numthreads - p->p_suspcount; 1104 else 1105 panic("calc_remaining: wrong mode %d", mode); 1106 return (remaining); 1107 } 1108 1109 static int 1110 remain_for_mode(int mode) 1111 { 1112 1113 return (mode == SINGLE_ALLPROC ? 0 : 1); 1114 } 1115 1116 static int 1117 weed_inhib(int mode, struct thread *td2, struct proc *p) 1118 { 1119 int wakeup_swapper; 1120 1121 PROC_LOCK_ASSERT(p, MA_OWNED); 1122 PROC_SLOCK_ASSERT(p, MA_OWNED); 1123 THREAD_LOCK_ASSERT(td2, MA_OWNED); 1124 1125 wakeup_swapper = 0; 1126 1127 /* 1128 * Since the thread lock is dropped by the scheduler we have 1129 * to retry to check for races. 1130 */ 1131 restart: 1132 switch (mode) { 1133 case SINGLE_EXIT: 1134 if (TD_IS_SUSPENDED(td2)) { 1135 wakeup_swapper |= thread_unsuspend_one(td2, p, true); 1136 thread_lock(td2); 1137 goto restart; 1138 } 1139 if (TD_CAN_ABORT(td2)) { 1140 wakeup_swapper |= sleepq_abort(td2, EINTR); 1141 return (wakeup_swapper); 1142 } 1143 break; 1144 case SINGLE_BOUNDARY: 1145 case SINGLE_NO_EXIT: 1146 if (TD_IS_SUSPENDED(td2) && 1147 (td2->td_flags & TDF_BOUNDARY) == 0) { 1148 wakeup_swapper |= thread_unsuspend_one(td2, p, false); 1149 thread_lock(td2); 1150 goto restart; 1151 } 1152 if (TD_CAN_ABORT(td2)) { 1153 wakeup_swapper |= sleepq_abort(td2, ERESTART); 1154 return (wakeup_swapper); 1155 } 1156 break; 1157 case SINGLE_ALLPROC: 1158 /* 1159 * ALLPROC suspend tries to avoid spurious EINTR for 1160 * threads sleeping interruptable, by suspending the 1161 * thread directly, similarly to sig_suspend_threads(). 1162 * Since such sleep is not neccessary performed at the user 1163 * boundary, TDF_ALLPROCSUSP is used to avoid immediate 1164 * un-suspend. 1165 */ 1166 if (TD_IS_SUSPENDED(td2) && (td2->td_flags & 1167 TDF_ALLPROCSUSP) == 0) { 1168 wakeup_swapper |= thread_unsuspend_one(td2, p, false); 1169 thread_lock(td2); 1170 goto restart; 1171 } 1172 if (TD_CAN_ABORT(td2)) { 1173 td2->td_flags |= TDF_ALLPROCSUSP; 1174 wakeup_swapper |= sleepq_abort(td2, ERESTART); 1175 return (wakeup_swapper); 1176 } 1177 break; 1178 default: 1179 break; 1180 } 1181 thread_unlock(td2); 1182 return (wakeup_swapper); 1183 } 1184 1185 /* 1186 * Enforce single-threading. 1187 * 1188 * Returns 1 if the caller must abort (another thread is waiting to 1189 * exit the process or similar). Process is locked! 1190 * Returns 0 when you are successfully the only thread running. 1191 * A process has successfully single threaded in the suspend mode when 1192 * There are no threads in user mode. Threads in the kernel must be 1193 * allowed to continue until they get to the user boundary. They may even 1194 * copy out their return values and data before suspending. They may however be 1195 * accelerated in reaching the user boundary as we will wake up 1196 * any sleeping threads that are interruptable. (PCATCH). 1197 */ 1198 int 1199 thread_single(struct proc *p, int mode) 1200 { 1201 struct thread *td; 1202 struct thread *td2; 1203 int remaining, wakeup_swapper; 1204 1205 td = curthread; 1206 KASSERT(mode == SINGLE_EXIT || mode == SINGLE_BOUNDARY || 1207 mode == SINGLE_ALLPROC || mode == SINGLE_NO_EXIT, 1208 ("invalid mode %d", mode)); 1209 /* 1210 * If allowing non-ALLPROC singlethreading for non-curproc 1211 * callers, calc_remaining() and remain_for_mode() should be 1212 * adjusted to also account for td->td_proc != p. For now 1213 * this is not implemented because it is not used. 1214 */ 1215 KASSERT((mode == SINGLE_ALLPROC && td->td_proc != p) || 1216 (mode != SINGLE_ALLPROC && td->td_proc == p), 1217 ("mode %d proc %p curproc %p", mode, p, td->td_proc)); 1218 mtx_assert(&Giant, MA_NOTOWNED); 1219 PROC_LOCK_ASSERT(p, MA_OWNED); 1220 1221 /* 1222 * Is someone already single threading? 1223 * Or may be singlethreading is not needed at all. 1224 */ 1225 if (mode == SINGLE_ALLPROC) { 1226 while ((p->p_flag & P_STOPPED_SINGLE) != 0) { 1227 if ((p->p_flag2 & P2_WEXIT) != 0) 1228 return (1); 1229 msleep(&p->p_flag, &p->p_mtx, PCATCH, "thrsgl", 0); 1230 } 1231 } else if ((p->p_flag & P_HADTHREADS) == 0) 1232 return (0); 1233 if (p->p_singlethread != NULL && p->p_singlethread != td) 1234 return (1); 1235 1236 if (mode == SINGLE_EXIT) { 1237 p->p_flag |= P_SINGLE_EXIT; 1238 p->p_flag &= ~P_SINGLE_BOUNDARY; 1239 } else { 1240 p->p_flag &= ~P_SINGLE_EXIT; 1241 if (mode == SINGLE_BOUNDARY) 1242 p->p_flag |= P_SINGLE_BOUNDARY; 1243 else 1244 p->p_flag &= ~P_SINGLE_BOUNDARY; 1245 } 1246 if (mode == SINGLE_ALLPROC) { 1247 p->p_flag |= P_TOTAL_STOP; 1248 thread_lock(td); 1249 td->td_flags |= TDF_DOING_SA; 1250 thread_unlock(td); 1251 } 1252 p->p_flag |= P_STOPPED_SINGLE; 1253 PROC_SLOCK(p); 1254 p->p_singlethread = td; 1255 remaining = calc_remaining(p, mode); 1256 while (remaining != remain_for_mode(mode)) { 1257 if (P_SHOULDSTOP(p) != P_STOPPED_SINGLE) 1258 goto stopme; 1259 wakeup_swapper = 0; 1260 FOREACH_THREAD_IN_PROC(p, td2) { 1261 if (td2 == td) 1262 continue; 1263 thread_lock(td2); 1264 ast_sched_locked(td2, TDA_SUSPEND); 1265 if (TD_IS_INHIBITED(td2)) { 1266 wakeup_swapper |= weed_inhib(mode, td2, p); 1267 #ifdef SMP 1268 } else if (TD_IS_RUNNING(td2)) { 1269 forward_signal(td2); 1270 thread_unlock(td2); 1271 #endif 1272 } else 1273 thread_unlock(td2); 1274 } 1275 if (wakeup_swapper) 1276 kick_proc0(); 1277 remaining = calc_remaining(p, mode); 1278 1279 /* 1280 * Maybe we suspended some threads.. was it enough? 1281 */ 1282 if (remaining == remain_for_mode(mode)) 1283 break; 1284 1285 stopme: 1286 /* 1287 * Wake us up when everyone else has suspended. 1288 * In the mean time we suspend as well. 1289 */ 1290 thread_suspend_switch(td, p); 1291 remaining = calc_remaining(p, mode); 1292 } 1293 if (mode == SINGLE_EXIT) { 1294 /* 1295 * Convert the process to an unthreaded process. The 1296 * SINGLE_EXIT is called by exit1() or execve(), in 1297 * both cases other threads must be retired. 1298 */ 1299 KASSERT(p->p_numthreads == 1, ("Unthreading with >1 threads")); 1300 p->p_singlethread = NULL; 1301 p->p_flag &= ~(P_STOPPED_SINGLE | P_SINGLE_EXIT | P_HADTHREADS); 1302 1303 /* 1304 * Wait for any remaining threads to exit cpu_throw(). 1305 */ 1306 while (p->p_exitthreads != 0) { 1307 PROC_SUNLOCK(p); 1308 PROC_UNLOCK(p); 1309 sched_relinquish(td); 1310 PROC_LOCK(p); 1311 PROC_SLOCK(p); 1312 } 1313 } else if (mode == SINGLE_BOUNDARY) { 1314 /* 1315 * Wait until all suspended threads are removed from 1316 * the processors. The thread_suspend_check() 1317 * increments p_boundary_count while it is still 1318 * running, which makes it possible for the execve() 1319 * to destroy vmspace while our other threads are 1320 * still using the address space. 1321 * 1322 * We lock the thread, which is only allowed to 1323 * succeed after context switch code finished using 1324 * the address space. 1325 */ 1326 FOREACH_THREAD_IN_PROC(p, td2) { 1327 if (td2 == td) 1328 continue; 1329 thread_lock(td2); 1330 KASSERT((td2->td_flags & TDF_BOUNDARY) != 0, 1331 ("td %p not on boundary", td2)); 1332 KASSERT(TD_IS_SUSPENDED(td2), 1333 ("td %p is not suspended", td2)); 1334 thread_unlock(td2); 1335 } 1336 } 1337 PROC_SUNLOCK(p); 1338 if (mode == SINGLE_ALLPROC) { 1339 thread_lock(td); 1340 td->td_flags &= ~TDF_DOING_SA; 1341 thread_unlock(td); 1342 } 1343 return (0); 1344 } 1345 1346 bool 1347 thread_suspend_check_needed(void) 1348 { 1349 struct proc *p; 1350 struct thread *td; 1351 1352 td = curthread; 1353 p = td->td_proc; 1354 PROC_LOCK_ASSERT(p, MA_OWNED); 1355 return (P_SHOULDSTOP(p) || ((p->p_flag & P_TRACED) != 0 && 1356 (td->td_dbgflags & TDB_SUSPEND) != 0)); 1357 } 1358 1359 /* 1360 * Called in from locations that can safely check to see 1361 * whether we have to suspend or at least throttle for a 1362 * single-thread event (e.g. fork). 1363 * 1364 * Such locations include userret(). 1365 * If the "return_instead" argument is non zero, the thread must be able to 1366 * accept 0 (caller may continue), or 1 (caller must abort) as a result. 1367 * 1368 * The 'return_instead' argument tells the function if it may do a 1369 * thread_exit() or suspend, or whether the caller must abort and back 1370 * out instead. 1371 * 1372 * If the thread that set the single_threading request has set the 1373 * P_SINGLE_EXIT bit in the process flags then this call will never return 1374 * if 'return_instead' is false, but will exit. 1375 * 1376 * P_SINGLE_EXIT | return_instead == 0| return_instead != 0 1377 *---------------+--------------------+--------------------- 1378 * 0 | returns 0 | returns 0 or 1 1379 * | when ST ends | immediately 1380 *---------------+--------------------+--------------------- 1381 * 1 | thread exits | returns 1 1382 * | | immediately 1383 * 0 = thread_exit() or suspension ok, 1384 * other = return error instead of stopping the thread. 1385 * 1386 * While a full suspension is under effect, even a single threading 1387 * thread would be suspended if it made this call (but it shouldn't). 1388 * This call should only be made from places where 1389 * thread_exit() would be safe as that may be the outcome unless 1390 * return_instead is set. 1391 */ 1392 int 1393 thread_suspend_check(int return_instead) 1394 { 1395 struct thread *td; 1396 struct proc *p; 1397 int wakeup_swapper; 1398 1399 td = curthread; 1400 p = td->td_proc; 1401 mtx_assert(&Giant, MA_NOTOWNED); 1402 PROC_LOCK_ASSERT(p, MA_OWNED); 1403 while (thread_suspend_check_needed()) { 1404 if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) { 1405 KASSERT(p->p_singlethread != NULL, 1406 ("singlethread not set")); 1407 /* 1408 * The only suspension in action is a 1409 * single-threading. Single threader need not stop. 1410 * It is safe to access p->p_singlethread unlocked 1411 * because it can only be set to our address by us. 1412 */ 1413 if (p->p_singlethread == td) 1414 return (0); /* Exempt from stopping. */ 1415 } 1416 if ((p->p_flag & P_SINGLE_EXIT) && return_instead) 1417 return (EINTR); 1418 1419 /* Should we goto user boundary if we didn't come from there? */ 1420 if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE && 1421 (p->p_flag & P_SINGLE_BOUNDARY) && return_instead) 1422 return (ERESTART); 1423 1424 /* 1425 * Ignore suspend requests if they are deferred. 1426 */ 1427 if ((td->td_flags & TDF_SBDRY) != 0) { 1428 KASSERT(return_instead, 1429 ("TDF_SBDRY set for unsafe thread_suspend_check")); 1430 KASSERT((td->td_flags & (TDF_SEINTR | TDF_SERESTART)) != 1431 (TDF_SEINTR | TDF_SERESTART), 1432 ("both TDF_SEINTR and TDF_SERESTART")); 1433 return (TD_SBDRY_INTR(td) ? TD_SBDRY_ERRNO(td) : 0); 1434 } 1435 1436 /* 1437 * If the process is waiting for us to exit, 1438 * this thread should just suicide. 1439 * Assumes that P_SINGLE_EXIT implies P_STOPPED_SINGLE. 1440 */ 1441 if ((p->p_flag & P_SINGLE_EXIT) && (p->p_singlethread != td)) { 1442 PROC_UNLOCK(p); 1443 1444 /* 1445 * Allow Linux emulation layer to do some work 1446 * before thread suicide. 1447 */ 1448 if (__predict_false(p->p_sysent->sv_thread_detach != NULL)) 1449 (p->p_sysent->sv_thread_detach)(td); 1450 umtx_thread_exit(td); 1451 kern_thr_exit(td); 1452 panic("stopped thread did not exit"); 1453 } 1454 1455 PROC_SLOCK(p); 1456 thread_stopped(p); 1457 if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) { 1458 if (p->p_numthreads == p->p_suspcount + 1) { 1459 thread_lock(p->p_singlethread); 1460 wakeup_swapper = thread_unsuspend_one( 1461 p->p_singlethread, p, false); 1462 if (wakeup_swapper) 1463 kick_proc0(); 1464 } 1465 } 1466 PROC_UNLOCK(p); 1467 thread_lock(td); 1468 /* 1469 * When a thread suspends, it just 1470 * gets taken off all queues. 1471 */ 1472 thread_suspend_one(td); 1473 if (return_instead == 0) { 1474 p->p_boundary_count++; 1475 td->td_flags |= TDF_BOUNDARY; 1476 } 1477 PROC_SUNLOCK(p); 1478 mi_switch(SW_INVOL | SWT_SUSPEND); 1479 PROC_LOCK(p); 1480 } 1481 return (0); 1482 } 1483 1484 /* 1485 * Check for possible stops and suspensions while executing a 1486 * casueword or similar transiently failing operation. 1487 * 1488 * The sleep argument controls whether the function can handle a stop 1489 * request itself or it should return ERESTART and the request is 1490 * proceed at the kernel/user boundary in ast. 1491 * 1492 * Typically, when retrying due to casueword(9) failure (rv == 1), we 1493 * should handle the stop requests there, with exception of cases when 1494 * the thread owns a kernel resource, for instance busied the umtx 1495 * key, or when functions return immediately if thread_check_susp() 1496 * returned non-zero. On the other hand, retrying the whole lock 1497 * operation, we better not stop there but delegate the handling to 1498 * ast. 1499 * 1500 * If the request is for thread termination P_SINGLE_EXIT, we cannot 1501 * handle it at all, and simply return EINTR. 1502 */ 1503 int 1504 thread_check_susp(struct thread *td, bool sleep) 1505 { 1506 struct proc *p; 1507 int error; 1508 1509 /* 1510 * The check for TDA_SUSPEND is racy, but it is enough to 1511 * eventually break the lockstep loop. 1512 */ 1513 if (!td_ast_pending(td, TDA_SUSPEND)) 1514 return (0); 1515 error = 0; 1516 p = td->td_proc; 1517 PROC_LOCK(p); 1518 if (p->p_flag & P_SINGLE_EXIT) 1519 error = EINTR; 1520 else if (P_SHOULDSTOP(p) || 1521 ((p->p_flag & P_TRACED) && (td->td_dbgflags & TDB_SUSPEND))) 1522 error = sleep ? thread_suspend_check(0) : ERESTART; 1523 PROC_UNLOCK(p); 1524 return (error); 1525 } 1526 1527 void 1528 thread_suspend_switch(struct thread *td, struct proc *p) 1529 { 1530 1531 KASSERT(!TD_IS_SUSPENDED(td), ("already suspended")); 1532 PROC_LOCK_ASSERT(p, MA_OWNED); 1533 PROC_SLOCK_ASSERT(p, MA_OWNED); 1534 /* 1535 * We implement thread_suspend_one in stages here to avoid 1536 * dropping the proc lock while the thread lock is owned. 1537 */ 1538 if (p == td->td_proc) { 1539 thread_stopped(p); 1540 p->p_suspcount++; 1541 } 1542 PROC_UNLOCK(p); 1543 thread_lock(td); 1544 ast_unsched_locked(td, TDA_SUSPEND); 1545 TD_SET_SUSPENDED(td); 1546 sched_sleep(td, 0); 1547 PROC_SUNLOCK(p); 1548 DROP_GIANT(); 1549 mi_switch(SW_VOL | SWT_SUSPEND); 1550 PICKUP_GIANT(); 1551 PROC_LOCK(p); 1552 PROC_SLOCK(p); 1553 } 1554 1555 void 1556 thread_suspend_one(struct thread *td) 1557 { 1558 struct proc *p; 1559 1560 p = td->td_proc; 1561 PROC_SLOCK_ASSERT(p, MA_OWNED); 1562 THREAD_LOCK_ASSERT(td, MA_OWNED); 1563 KASSERT(!TD_IS_SUSPENDED(td), ("already suspended")); 1564 p->p_suspcount++; 1565 ast_unsched_locked(td, TDA_SUSPEND); 1566 TD_SET_SUSPENDED(td); 1567 sched_sleep(td, 0); 1568 } 1569 1570 static int 1571 thread_unsuspend_one(struct thread *td, struct proc *p, bool boundary) 1572 { 1573 1574 THREAD_LOCK_ASSERT(td, MA_OWNED); 1575 KASSERT(TD_IS_SUSPENDED(td), ("Thread not suspended")); 1576 TD_CLR_SUSPENDED(td); 1577 td->td_flags &= ~TDF_ALLPROCSUSP; 1578 if (td->td_proc == p) { 1579 PROC_SLOCK_ASSERT(p, MA_OWNED); 1580 p->p_suspcount--; 1581 if (boundary && (td->td_flags & TDF_BOUNDARY) != 0) { 1582 td->td_flags &= ~TDF_BOUNDARY; 1583 p->p_boundary_count--; 1584 } 1585 } 1586 return (setrunnable(td, 0)); 1587 } 1588 1589 void 1590 thread_run_flash(struct thread *td) 1591 { 1592 struct proc *p; 1593 1594 p = td->td_proc; 1595 PROC_LOCK_ASSERT(p, MA_OWNED); 1596 1597 if (TD_ON_SLEEPQ(td)) 1598 sleepq_remove_nested(td); 1599 else 1600 thread_lock(td); 1601 1602 THREAD_LOCK_ASSERT(td, MA_OWNED); 1603 KASSERT(TD_IS_SUSPENDED(td), ("Thread not suspended")); 1604 1605 TD_CLR_SUSPENDED(td); 1606 PROC_SLOCK(p); 1607 MPASS(p->p_suspcount > 0); 1608 p->p_suspcount--; 1609 PROC_SUNLOCK(p); 1610 if (setrunnable(td, 0)) 1611 kick_proc0(); 1612 } 1613 1614 /* 1615 * Allow all threads blocked by single threading to continue running. 1616 */ 1617 void 1618 thread_unsuspend(struct proc *p) 1619 { 1620 struct thread *td; 1621 int wakeup_swapper; 1622 1623 PROC_LOCK_ASSERT(p, MA_OWNED); 1624 PROC_SLOCK_ASSERT(p, MA_OWNED); 1625 wakeup_swapper = 0; 1626 if (!P_SHOULDSTOP(p)) { 1627 FOREACH_THREAD_IN_PROC(p, td) { 1628 thread_lock(td); 1629 if (TD_IS_SUSPENDED(td) && (td->td_flags & 1630 TDF_DOING_SA) == 0) { 1631 wakeup_swapper |= thread_unsuspend_one(td, p, 1632 true); 1633 } else 1634 thread_unlock(td); 1635 } 1636 } else if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE && 1637 p->p_numthreads == p->p_suspcount) { 1638 /* 1639 * Stopping everything also did the job for the single 1640 * threading request. Now we've downgraded to single-threaded, 1641 * let it continue. 1642 */ 1643 if (p->p_singlethread->td_proc == p) { 1644 thread_lock(p->p_singlethread); 1645 wakeup_swapper = thread_unsuspend_one( 1646 p->p_singlethread, p, false); 1647 } 1648 } 1649 if (wakeup_swapper) 1650 kick_proc0(); 1651 } 1652 1653 /* 1654 * End the single threading mode.. 1655 */ 1656 void 1657 thread_single_end(struct proc *p, int mode) 1658 { 1659 struct thread *td; 1660 int wakeup_swapper; 1661 1662 KASSERT(mode == SINGLE_EXIT || mode == SINGLE_BOUNDARY || 1663 mode == SINGLE_ALLPROC || mode == SINGLE_NO_EXIT, 1664 ("invalid mode %d", mode)); 1665 PROC_LOCK_ASSERT(p, MA_OWNED); 1666 KASSERT((mode == SINGLE_ALLPROC && (p->p_flag & P_TOTAL_STOP) != 0) || 1667 (mode != SINGLE_ALLPROC && (p->p_flag & P_TOTAL_STOP) == 0), 1668 ("mode %d does not match P_TOTAL_STOP", mode)); 1669 KASSERT(mode == SINGLE_ALLPROC || p->p_singlethread == curthread, 1670 ("thread_single_end from other thread %p %p", 1671 curthread, p->p_singlethread)); 1672 KASSERT(mode != SINGLE_BOUNDARY || 1673 (p->p_flag & P_SINGLE_BOUNDARY) != 0, 1674 ("mis-matched SINGLE_BOUNDARY flags %x", p->p_flag)); 1675 p->p_flag &= ~(P_STOPPED_SINGLE | P_SINGLE_EXIT | P_SINGLE_BOUNDARY | 1676 P_TOTAL_STOP); 1677 PROC_SLOCK(p); 1678 p->p_singlethread = NULL; 1679 wakeup_swapper = 0; 1680 /* 1681 * If there are other threads they may now run, 1682 * unless of course there is a blanket 'stop order' 1683 * on the process. The single threader must be allowed 1684 * to continue however as this is a bad place to stop. 1685 */ 1686 if (p->p_numthreads != remain_for_mode(mode) && !P_SHOULDSTOP(p)) { 1687 FOREACH_THREAD_IN_PROC(p, td) { 1688 thread_lock(td); 1689 if (TD_IS_SUSPENDED(td)) { 1690 wakeup_swapper |= thread_unsuspend_one(td, p, 1691 true); 1692 } else 1693 thread_unlock(td); 1694 } 1695 } 1696 KASSERT(mode != SINGLE_BOUNDARY || p->p_boundary_count == 0, 1697 ("inconsistent boundary count %d", p->p_boundary_count)); 1698 PROC_SUNLOCK(p); 1699 if (wakeup_swapper) 1700 kick_proc0(); 1701 wakeup(&p->p_flag); 1702 } 1703 1704 /* 1705 * Locate a thread by number and return with proc lock held. 1706 * 1707 * thread exit establishes proc -> tidhash lock ordering, but lookup 1708 * takes tidhash first and needs to return locked proc. 1709 * 1710 * The problem is worked around by relying on type-safety of both 1711 * structures and doing the work in 2 steps: 1712 * - tidhash-locked lookup which saves both thread and proc pointers 1713 * - proc-locked verification that the found thread still matches 1714 */ 1715 static bool 1716 tdfind_hash(lwpid_t tid, pid_t pid, struct proc **pp, struct thread **tdp) 1717 { 1718 #define RUN_THRESH 16 1719 struct proc *p; 1720 struct thread *td; 1721 int run; 1722 bool locked; 1723 1724 run = 0; 1725 rw_rlock(TIDHASHLOCK(tid)); 1726 locked = true; 1727 LIST_FOREACH(td, TIDHASH(tid), td_hash) { 1728 if (td->td_tid != tid) { 1729 run++; 1730 continue; 1731 } 1732 p = td->td_proc; 1733 if (pid != -1 && p->p_pid != pid) { 1734 td = NULL; 1735 break; 1736 } 1737 if (run > RUN_THRESH) { 1738 if (rw_try_upgrade(TIDHASHLOCK(tid))) { 1739 LIST_REMOVE(td, td_hash); 1740 LIST_INSERT_HEAD(TIDHASH(td->td_tid), 1741 td, td_hash); 1742 rw_wunlock(TIDHASHLOCK(tid)); 1743 locked = false; 1744 break; 1745 } 1746 } 1747 break; 1748 } 1749 if (locked) 1750 rw_runlock(TIDHASHLOCK(tid)); 1751 if (td == NULL) 1752 return (false); 1753 *pp = p; 1754 *tdp = td; 1755 return (true); 1756 } 1757 1758 struct thread * 1759 tdfind(lwpid_t tid, pid_t pid) 1760 { 1761 struct proc *p; 1762 struct thread *td; 1763 1764 td = curthread; 1765 if (td->td_tid == tid) { 1766 if (pid != -1 && td->td_proc->p_pid != pid) 1767 return (NULL); 1768 PROC_LOCK(td->td_proc); 1769 return (td); 1770 } 1771 1772 for (;;) { 1773 if (!tdfind_hash(tid, pid, &p, &td)) 1774 return (NULL); 1775 PROC_LOCK(p); 1776 if (td->td_tid != tid) { 1777 PROC_UNLOCK(p); 1778 continue; 1779 } 1780 if (td->td_proc != p) { 1781 PROC_UNLOCK(p); 1782 continue; 1783 } 1784 if (p->p_state == PRS_NEW) { 1785 PROC_UNLOCK(p); 1786 return (NULL); 1787 } 1788 return (td); 1789 } 1790 } 1791 1792 void 1793 tidhash_add(struct thread *td) 1794 { 1795 rw_wlock(TIDHASHLOCK(td->td_tid)); 1796 LIST_INSERT_HEAD(TIDHASH(td->td_tid), td, td_hash); 1797 rw_wunlock(TIDHASHLOCK(td->td_tid)); 1798 } 1799 1800 void 1801 tidhash_remove(struct thread *td) 1802 { 1803 1804 rw_wlock(TIDHASHLOCK(td->td_tid)); 1805 LIST_REMOVE(td, td_hash); 1806 rw_wunlock(TIDHASHLOCK(td->td_tid)); 1807 } 1808