xref: /freebsd/sys/kern/kern_thread.c (revision f4b37ed0f8b307b1f3f0f630ca725d68f1dff30d)
1 /*-
2  * Copyright (C) 2001 Julian Elischer <julian@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_witness.h"
30 #include "opt_hwpmc_hooks.h"
31 
32 #include <sys/cdefs.h>
33 __FBSDID("$FreeBSD$");
34 
35 #include <sys/param.h>
36 #include <sys/systm.h>
37 #include <sys/kernel.h>
38 #include <sys/lock.h>
39 #include <sys/mutex.h>
40 #include <sys/proc.h>
41 #include <sys/rangelock.h>
42 #include <sys/resourcevar.h>
43 #include <sys/sdt.h>
44 #include <sys/smp.h>
45 #include <sys/sched.h>
46 #include <sys/sleepqueue.h>
47 #include <sys/selinfo.h>
48 #include <sys/sysent.h>
49 #include <sys/turnstile.h>
50 #include <sys/ktr.h>
51 #include <sys/rwlock.h>
52 #include <sys/umtx.h>
53 #include <sys/cpuset.h>
54 #ifdef	HWPMC_HOOKS
55 #include <sys/pmckern.h>
56 #endif
57 
58 #include <security/audit/audit.h>
59 
60 #include <vm/vm.h>
61 #include <vm/vm_extern.h>
62 #include <vm/uma.h>
63 #include <vm/vm_domain.h>
64 #include <sys/eventhandler.h>
65 
66 SDT_PROVIDER_DECLARE(proc);
67 SDT_PROBE_DEFINE(proc, , , lwp__exit);
68 
69 /*
70  * thread related storage.
71  */
72 static uma_zone_t thread_zone;
73 
74 TAILQ_HEAD(, thread) zombie_threads = TAILQ_HEAD_INITIALIZER(zombie_threads);
75 static struct mtx zombie_lock;
76 MTX_SYSINIT(zombie_lock, &zombie_lock, "zombie lock", MTX_SPIN);
77 
78 static void thread_zombie(struct thread *);
79 static int thread_unsuspend_one(struct thread *td, struct proc *p,
80     bool boundary);
81 
82 #define TID_BUFFER_SIZE	1024
83 
84 struct mtx tid_lock;
85 static struct unrhdr *tid_unrhdr;
86 static lwpid_t tid_buffer[TID_BUFFER_SIZE];
87 static int tid_head, tid_tail;
88 static MALLOC_DEFINE(M_TIDHASH, "tidhash", "thread hash");
89 
90 struct	tidhashhead *tidhashtbl;
91 u_long	tidhash;
92 struct	rwlock tidhash_lock;
93 
94 static lwpid_t
95 tid_alloc(void)
96 {
97 	lwpid_t	tid;
98 
99 	tid = alloc_unr(tid_unrhdr);
100 	if (tid != -1)
101 		return (tid);
102 	mtx_lock(&tid_lock);
103 	if (tid_head == tid_tail) {
104 		mtx_unlock(&tid_lock);
105 		return (-1);
106 	}
107 	tid = tid_buffer[tid_head];
108 	tid_head = (tid_head + 1) % TID_BUFFER_SIZE;
109 	mtx_unlock(&tid_lock);
110 	return (tid);
111 }
112 
113 static void
114 tid_free(lwpid_t tid)
115 {
116 	lwpid_t tmp_tid = -1;
117 
118 	mtx_lock(&tid_lock);
119 	if ((tid_tail + 1) % TID_BUFFER_SIZE == tid_head) {
120 		tmp_tid = tid_buffer[tid_head];
121 		tid_head = (tid_head + 1) % TID_BUFFER_SIZE;
122 	}
123 	tid_buffer[tid_tail] = tid;
124 	tid_tail = (tid_tail + 1) % TID_BUFFER_SIZE;
125 	mtx_unlock(&tid_lock);
126 	if (tmp_tid != -1)
127 		free_unr(tid_unrhdr, tmp_tid);
128 }
129 
130 /*
131  * Prepare a thread for use.
132  */
133 static int
134 thread_ctor(void *mem, int size, void *arg, int flags)
135 {
136 	struct thread	*td;
137 
138 	td = (struct thread *)mem;
139 	td->td_state = TDS_INACTIVE;
140 	td->td_oncpu = NOCPU;
141 
142 	td->td_tid = tid_alloc();
143 
144 	/*
145 	 * Note that td_critnest begins life as 1 because the thread is not
146 	 * running and is thereby implicitly waiting to be on the receiving
147 	 * end of a context switch.
148 	 */
149 	td->td_critnest = 1;
150 	td->td_lend_user_pri = PRI_MAX;
151 	EVENTHANDLER_INVOKE(thread_ctor, td);
152 #ifdef AUDIT
153 	audit_thread_alloc(td);
154 #endif
155 	umtx_thread_alloc(td);
156 	return (0);
157 }
158 
159 /*
160  * Reclaim a thread after use.
161  */
162 static void
163 thread_dtor(void *mem, int size, void *arg)
164 {
165 	struct thread *td;
166 
167 	td = (struct thread *)mem;
168 
169 #ifdef INVARIANTS
170 	/* Verify that this thread is in a safe state to free. */
171 	switch (td->td_state) {
172 	case TDS_INHIBITED:
173 	case TDS_RUNNING:
174 	case TDS_CAN_RUN:
175 	case TDS_RUNQ:
176 		/*
177 		 * We must never unlink a thread that is in one of
178 		 * these states, because it is currently active.
179 		 */
180 		panic("bad state for thread unlinking");
181 		/* NOTREACHED */
182 	case TDS_INACTIVE:
183 		break;
184 	default:
185 		panic("bad thread state");
186 		/* NOTREACHED */
187 	}
188 #endif
189 #ifdef AUDIT
190 	audit_thread_free(td);
191 #endif
192 	/* Free all OSD associated to this thread. */
193 	osd_thread_exit(td);
194 
195 	EVENTHANDLER_INVOKE(thread_dtor, td);
196 	tid_free(td->td_tid);
197 }
198 
199 /*
200  * Initialize type-stable parts of a thread (when newly created).
201  */
202 static int
203 thread_init(void *mem, int size, int flags)
204 {
205 	struct thread *td;
206 
207 	td = (struct thread *)mem;
208 
209 	td->td_sleepqueue = sleepq_alloc();
210 	td->td_turnstile = turnstile_alloc();
211 	td->td_rlqe = NULL;
212 	EVENTHANDLER_INVOKE(thread_init, td);
213 	td->td_sched = (struct td_sched *)&td[1];
214 	umtx_thread_init(td);
215 	td->td_kstack = 0;
216 	td->td_sel = NULL;
217 	return (0);
218 }
219 
220 /*
221  * Tear down type-stable parts of a thread (just before being discarded).
222  */
223 static void
224 thread_fini(void *mem, int size)
225 {
226 	struct thread *td;
227 
228 	td = (struct thread *)mem;
229 	EVENTHANDLER_INVOKE(thread_fini, td);
230 	rlqentry_free(td->td_rlqe);
231 	turnstile_free(td->td_turnstile);
232 	sleepq_free(td->td_sleepqueue);
233 	umtx_thread_fini(td);
234 	seltdfini(td);
235 }
236 
237 /*
238  * For a newly created process,
239  * link up all the structures and its initial threads etc.
240  * called from:
241  * {arch}/{arch}/machdep.c   {arch}_init(), init386() etc.
242  * proc_dtor() (should go away)
243  * proc_init()
244  */
245 void
246 proc_linkup0(struct proc *p, struct thread *td)
247 {
248 	TAILQ_INIT(&p->p_threads);	     /* all threads in proc */
249 	proc_linkup(p, td);
250 }
251 
252 void
253 proc_linkup(struct proc *p, struct thread *td)
254 {
255 
256 	sigqueue_init(&p->p_sigqueue, p);
257 	p->p_ksi = ksiginfo_alloc(1);
258 	if (p->p_ksi != NULL) {
259 		/* XXX p_ksi may be null if ksiginfo zone is not ready */
260 		p->p_ksi->ksi_flags = KSI_EXT | KSI_INS;
261 	}
262 	LIST_INIT(&p->p_mqnotifier);
263 	p->p_numthreads = 0;
264 	thread_link(td, p);
265 }
266 
267 /*
268  * Initialize global thread allocation resources.
269  */
270 void
271 threadinit(void)
272 {
273 
274 	mtx_init(&tid_lock, "TID lock", NULL, MTX_DEF);
275 
276 	/*
277 	 * pid_max cannot be greater than PID_MAX.
278 	 * leave one number for thread0.
279 	 */
280 	tid_unrhdr = new_unrhdr(PID_MAX + 2, INT_MAX, &tid_lock);
281 
282 	thread_zone = uma_zcreate("THREAD", sched_sizeof_thread(),
283 	    thread_ctor, thread_dtor, thread_init, thread_fini,
284 	    16 - 1, 0);
285 	tidhashtbl = hashinit(maxproc / 2, M_TIDHASH, &tidhash);
286 	rw_init(&tidhash_lock, "tidhash");
287 }
288 
289 /*
290  * Place an unused thread on the zombie list.
291  * Use the slpq as that must be unused by now.
292  */
293 void
294 thread_zombie(struct thread *td)
295 {
296 	mtx_lock_spin(&zombie_lock);
297 	TAILQ_INSERT_HEAD(&zombie_threads, td, td_slpq);
298 	mtx_unlock_spin(&zombie_lock);
299 }
300 
301 /*
302  * Release a thread that has exited after cpu_throw().
303  */
304 void
305 thread_stash(struct thread *td)
306 {
307 	atomic_subtract_rel_int(&td->td_proc->p_exitthreads, 1);
308 	thread_zombie(td);
309 }
310 
311 /*
312  * Reap zombie resources.
313  */
314 void
315 thread_reap(void)
316 {
317 	struct thread *td_first, *td_next;
318 
319 	/*
320 	 * Don't even bother to lock if none at this instant,
321 	 * we really don't care about the next instant..
322 	 */
323 	if (!TAILQ_EMPTY(&zombie_threads)) {
324 		mtx_lock_spin(&zombie_lock);
325 		td_first = TAILQ_FIRST(&zombie_threads);
326 		if (td_first)
327 			TAILQ_INIT(&zombie_threads);
328 		mtx_unlock_spin(&zombie_lock);
329 		while (td_first) {
330 			td_next = TAILQ_NEXT(td_first, td_slpq);
331 			thread_cow_free(td_first);
332 			thread_free(td_first);
333 			td_first = td_next;
334 		}
335 	}
336 }
337 
338 /*
339  * Allocate a thread.
340  */
341 struct thread *
342 thread_alloc(int pages)
343 {
344 	struct thread *td;
345 
346 	thread_reap(); /* check if any zombies to get */
347 
348 	td = (struct thread *)uma_zalloc(thread_zone, M_WAITOK);
349 	KASSERT(td->td_kstack == 0, ("thread_alloc got thread with kstack"));
350 	if (!vm_thread_new(td, pages)) {
351 		uma_zfree(thread_zone, td);
352 		return (NULL);
353 	}
354 	cpu_thread_alloc(td);
355 	vm_domain_policy_init(&td->td_vm_dom_policy);
356 	return (td);
357 }
358 
359 int
360 thread_alloc_stack(struct thread *td, int pages)
361 {
362 
363 	KASSERT(td->td_kstack == 0,
364 	    ("thread_alloc_stack called on a thread with kstack"));
365 	if (!vm_thread_new(td, pages))
366 		return (0);
367 	cpu_thread_alloc(td);
368 	return (1);
369 }
370 
371 /*
372  * Deallocate a thread.
373  */
374 void
375 thread_free(struct thread *td)
376 {
377 
378 	lock_profile_thread_exit(td);
379 	if (td->td_cpuset)
380 		cpuset_rel(td->td_cpuset);
381 	td->td_cpuset = NULL;
382 	cpu_thread_free(td);
383 	if (td->td_kstack != 0)
384 		vm_thread_dispose(td);
385 	vm_domain_policy_cleanup(&td->td_vm_dom_policy);
386 	uma_zfree(thread_zone, td);
387 }
388 
389 void
390 thread_cow_get_proc(struct thread *newtd, struct proc *p)
391 {
392 
393 	PROC_LOCK_ASSERT(p, MA_OWNED);
394 	newtd->td_ucred = crhold(p->p_ucred);
395 	newtd->td_limit = lim_hold(p->p_limit);
396 	newtd->td_cowgen = p->p_cowgen;
397 }
398 
399 void
400 thread_cow_get(struct thread *newtd, struct thread *td)
401 {
402 
403 	newtd->td_ucred = crhold(td->td_ucred);
404 	newtd->td_limit = lim_hold(td->td_limit);
405 	newtd->td_cowgen = td->td_cowgen;
406 }
407 
408 void
409 thread_cow_free(struct thread *td)
410 {
411 
412 	if (td->td_ucred != NULL)
413 		crfree(td->td_ucred);
414 	if (td->td_limit != NULL)
415 		lim_free(td->td_limit);
416 }
417 
418 void
419 thread_cow_update(struct thread *td)
420 {
421 	struct proc *p;
422 	struct ucred *oldcred;
423 	struct plimit *oldlimit;
424 
425 	p = td->td_proc;
426 	oldcred = NULL;
427 	oldlimit = NULL;
428 	PROC_LOCK(p);
429 	if (td->td_ucred != p->p_ucred) {
430 		oldcred = td->td_ucred;
431 		td->td_ucred = crhold(p->p_ucred);
432 	}
433 	if (td->td_limit != p->p_limit) {
434 		oldlimit = td->td_limit;
435 		td->td_limit = lim_hold(p->p_limit);
436 	}
437 	td->td_cowgen = p->p_cowgen;
438 	PROC_UNLOCK(p);
439 	if (oldcred != NULL)
440 		crfree(oldcred);
441 	if (oldlimit != NULL)
442 		lim_free(oldlimit);
443 }
444 
445 /*
446  * Discard the current thread and exit from its context.
447  * Always called with scheduler locked.
448  *
449  * Because we can't free a thread while we're operating under its context,
450  * push the current thread into our CPU's deadthread holder. This means
451  * we needn't worry about someone else grabbing our context before we
452  * do a cpu_throw().
453  */
454 void
455 thread_exit(void)
456 {
457 	uint64_t runtime, new_switchtime;
458 	struct thread *td;
459 	struct thread *td2;
460 	struct proc *p;
461 	int wakeup_swapper;
462 
463 	td = curthread;
464 	p = td->td_proc;
465 
466 	PROC_SLOCK_ASSERT(p, MA_OWNED);
467 	mtx_assert(&Giant, MA_NOTOWNED);
468 
469 	PROC_LOCK_ASSERT(p, MA_OWNED);
470 	KASSERT(p != NULL, ("thread exiting without a process"));
471 	CTR3(KTR_PROC, "thread_exit: thread %p (pid %ld, %s)", td,
472 	    (long)p->p_pid, td->td_name);
473 	KASSERT(TAILQ_EMPTY(&td->td_sigqueue.sq_list), ("signal pending"));
474 
475 #ifdef AUDIT
476 	AUDIT_SYSCALL_EXIT(0, td);
477 #endif
478 	/*
479 	 * drop FPU & debug register state storage, or any other
480 	 * architecture specific resources that
481 	 * would not be on a new untouched process.
482 	 */
483 	cpu_thread_exit(td);	/* XXXSMP */
484 
485 	/*
486 	 * The last thread is left attached to the process
487 	 * So that the whole bundle gets recycled. Skip
488 	 * all this stuff if we never had threads.
489 	 * EXIT clears all sign of other threads when
490 	 * it goes to single threading, so the last thread always
491 	 * takes the short path.
492 	 */
493 	if (p->p_flag & P_HADTHREADS) {
494 		if (p->p_numthreads > 1) {
495 			atomic_add_int(&td->td_proc->p_exitthreads, 1);
496 			thread_unlink(td);
497 			td2 = FIRST_THREAD_IN_PROC(p);
498 			sched_exit_thread(td2, td);
499 
500 			/*
501 			 * The test below is NOT true if we are the
502 			 * sole exiting thread. P_STOPPED_SINGLE is unset
503 			 * in exit1() after it is the only survivor.
504 			 */
505 			if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) {
506 				if (p->p_numthreads == p->p_suspcount) {
507 					thread_lock(p->p_singlethread);
508 					wakeup_swapper = thread_unsuspend_one(
509 						p->p_singlethread, p, false);
510 					thread_unlock(p->p_singlethread);
511 					if (wakeup_swapper)
512 						kick_proc0();
513 				}
514 			}
515 
516 			PCPU_SET(deadthread, td);
517 		} else {
518 			/*
519 			 * The last thread is exiting.. but not through exit()
520 			 */
521 			panic ("thread_exit: Last thread exiting on its own");
522 		}
523 	}
524 #ifdef	HWPMC_HOOKS
525 	/*
526 	 * If this thread is part of a process that is being tracked by hwpmc(4),
527 	 * inform the module of the thread's impending exit.
528 	 */
529 	if (PMC_PROC_IS_USING_PMCS(td->td_proc))
530 		PMC_SWITCH_CONTEXT(td, PMC_FN_CSW_OUT);
531 #endif
532 	PROC_UNLOCK(p);
533 	PROC_STATLOCK(p);
534 	thread_lock(td);
535 	PROC_SUNLOCK(p);
536 
537 	/* Do the same timestamp bookkeeping that mi_switch() would do. */
538 	new_switchtime = cpu_ticks();
539 	runtime = new_switchtime - PCPU_GET(switchtime);
540 	td->td_runtime += runtime;
541 	td->td_incruntime += runtime;
542 	PCPU_SET(switchtime, new_switchtime);
543 	PCPU_SET(switchticks, ticks);
544 	PCPU_INC(cnt.v_swtch);
545 
546 	/* Save our resource usage in our process. */
547 	td->td_ru.ru_nvcsw++;
548 	ruxagg(p, td);
549 	rucollect(&p->p_ru, &td->td_ru);
550 	PROC_STATUNLOCK(p);
551 
552 	td->td_state = TDS_INACTIVE;
553 #ifdef WITNESS
554 	witness_thread_exit(td);
555 #endif
556 	CTR1(KTR_PROC, "thread_exit: cpu_throw() thread %p", td);
557 	sched_throw(td);
558 	panic("I'm a teapot!");
559 	/* NOTREACHED */
560 }
561 
562 /*
563  * Do any thread specific cleanups that may be needed in wait()
564  * called with Giant, proc and schedlock not held.
565  */
566 void
567 thread_wait(struct proc *p)
568 {
569 	struct thread *td;
570 
571 	mtx_assert(&Giant, MA_NOTOWNED);
572 	KASSERT(p->p_numthreads == 1, ("multiple threads in thread_wait()"));
573 	KASSERT(p->p_exitthreads == 0, ("p_exitthreads leaking"));
574 	td = FIRST_THREAD_IN_PROC(p);
575 	/* Lock the last thread so we spin until it exits cpu_throw(). */
576 	thread_lock(td);
577 	thread_unlock(td);
578 	lock_profile_thread_exit(td);
579 	cpuset_rel(td->td_cpuset);
580 	td->td_cpuset = NULL;
581 	cpu_thread_clean(td);
582 	thread_cow_free(td);
583 	thread_reap();	/* check for zombie threads etc. */
584 }
585 
586 /*
587  * Link a thread to a process.
588  * set up anything that needs to be initialized for it to
589  * be used by the process.
590  */
591 void
592 thread_link(struct thread *td, struct proc *p)
593 {
594 
595 	/*
596 	 * XXX This can't be enabled because it's called for proc0 before
597 	 * its lock has been created.
598 	 * PROC_LOCK_ASSERT(p, MA_OWNED);
599 	 */
600 	td->td_state    = TDS_INACTIVE;
601 	td->td_proc     = p;
602 	td->td_flags    = TDF_INMEM;
603 
604 	LIST_INIT(&td->td_contested);
605 	LIST_INIT(&td->td_lprof[0]);
606 	LIST_INIT(&td->td_lprof[1]);
607 	sigqueue_init(&td->td_sigqueue, p);
608 	callout_init(&td->td_slpcallout, 1);
609 	TAILQ_INSERT_TAIL(&p->p_threads, td, td_plist);
610 	p->p_numthreads++;
611 }
612 
613 /*
614  * Called from:
615  *  thread_exit()
616  */
617 void
618 thread_unlink(struct thread *td)
619 {
620 	struct proc *p = td->td_proc;
621 
622 	PROC_LOCK_ASSERT(p, MA_OWNED);
623 	TAILQ_REMOVE(&p->p_threads, td, td_plist);
624 	p->p_numthreads--;
625 	/* could clear a few other things here */
626 	/* Must  NOT clear links to proc! */
627 }
628 
629 static int
630 calc_remaining(struct proc *p, int mode)
631 {
632 	int remaining;
633 
634 	PROC_LOCK_ASSERT(p, MA_OWNED);
635 	PROC_SLOCK_ASSERT(p, MA_OWNED);
636 	if (mode == SINGLE_EXIT)
637 		remaining = p->p_numthreads;
638 	else if (mode == SINGLE_BOUNDARY)
639 		remaining = p->p_numthreads - p->p_boundary_count;
640 	else if (mode == SINGLE_NO_EXIT || mode == SINGLE_ALLPROC)
641 		remaining = p->p_numthreads - p->p_suspcount;
642 	else
643 		panic("calc_remaining: wrong mode %d", mode);
644 	return (remaining);
645 }
646 
647 static int
648 remain_for_mode(int mode)
649 {
650 
651 	return (mode == SINGLE_ALLPROC ? 0 : 1);
652 }
653 
654 static int
655 weed_inhib(int mode, struct thread *td2, struct proc *p)
656 {
657 	int wakeup_swapper;
658 
659 	PROC_LOCK_ASSERT(p, MA_OWNED);
660 	PROC_SLOCK_ASSERT(p, MA_OWNED);
661 	THREAD_LOCK_ASSERT(td2, MA_OWNED);
662 
663 	wakeup_swapper = 0;
664 	switch (mode) {
665 	case SINGLE_EXIT:
666 		if (TD_IS_SUSPENDED(td2))
667 			wakeup_swapper |= thread_unsuspend_one(td2, p, true);
668 		if (TD_ON_SLEEPQ(td2) && (td2->td_flags & TDF_SINTR) != 0)
669 			wakeup_swapper |= sleepq_abort(td2, EINTR);
670 		break;
671 	case SINGLE_BOUNDARY:
672 		if (TD_IS_SUSPENDED(td2) && (td2->td_flags & TDF_BOUNDARY) == 0)
673 			wakeup_swapper |= thread_unsuspend_one(td2, p, false);
674 		if (TD_ON_SLEEPQ(td2) && (td2->td_flags & TDF_SINTR) != 0)
675 			wakeup_swapper |= sleepq_abort(td2, ERESTART);
676 		break;
677 	case SINGLE_NO_EXIT:
678 		if (TD_IS_SUSPENDED(td2) && (td2->td_flags & TDF_BOUNDARY) == 0)
679 			wakeup_swapper |= thread_unsuspend_one(td2, p, false);
680 		if (TD_ON_SLEEPQ(td2) && (td2->td_flags & TDF_SINTR) != 0)
681 			wakeup_swapper |= sleepq_abort(td2, ERESTART);
682 		break;
683 	case SINGLE_ALLPROC:
684 		/*
685 		 * ALLPROC suspend tries to avoid spurious EINTR for
686 		 * threads sleeping interruptable, by suspending the
687 		 * thread directly, similarly to sig_suspend_threads().
688 		 * Since such sleep is not performed at the user
689 		 * boundary, TDF_BOUNDARY flag is not set, and TDF_ALLPROCSUSP
690 		 * is used to avoid immediate un-suspend.
691 		 */
692 		if (TD_IS_SUSPENDED(td2) && (td2->td_flags & (TDF_BOUNDARY |
693 		    TDF_ALLPROCSUSP)) == 0)
694 			wakeup_swapper |= thread_unsuspend_one(td2, p, false);
695 		if (TD_ON_SLEEPQ(td2) && (td2->td_flags & TDF_SINTR) != 0) {
696 			if ((td2->td_flags & TDF_SBDRY) == 0) {
697 				thread_suspend_one(td2);
698 				td2->td_flags |= TDF_ALLPROCSUSP;
699 			} else {
700 				wakeup_swapper |= sleepq_abort(td2, ERESTART);
701 			}
702 		}
703 		break;
704 	}
705 	return (wakeup_swapper);
706 }
707 
708 /*
709  * Enforce single-threading.
710  *
711  * Returns 1 if the caller must abort (another thread is waiting to
712  * exit the process or similar). Process is locked!
713  * Returns 0 when you are successfully the only thread running.
714  * A process has successfully single threaded in the suspend mode when
715  * There are no threads in user mode. Threads in the kernel must be
716  * allowed to continue until they get to the user boundary. They may even
717  * copy out their return values and data before suspending. They may however be
718  * accelerated in reaching the user boundary as we will wake up
719  * any sleeping threads that are interruptable. (PCATCH).
720  */
721 int
722 thread_single(struct proc *p, int mode)
723 {
724 	struct thread *td;
725 	struct thread *td2;
726 	int remaining, wakeup_swapper;
727 
728 	td = curthread;
729 	KASSERT(mode == SINGLE_EXIT || mode == SINGLE_BOUNDARY ||
730 	    mode == SINGLE_ALLPROC || mode == SINGLE_NO_EXIT,
731 	    ("invalid mode %d", mode));
732 	/*
733 	 * If allowing non-ALLPROC singlethreading for non-curproc
734 	 * callers, calc_remaining() and remain_for_mode() should be
735 	 * adjusted to also account for td->td_proc != p.  For now
736 	 * this is not implemented because it is not used.
737 	 */
738 	KASSERT((mode == SINGLE_ALLPROC && td->td_proc != p) ||
739 	    (mode != SINGLE_ALLPROC && td->td_proc == p),
740 	    ("mode %d proc %p curproc %p", mode, p, td->td_proc));
741 	mtx_assert(&Giant, MA_NOTOWNED);
742 	PROC_LOCK_ASSERT(p, MA_OWNED);
743 
744 	if ((p->p_flag & P_HADTHREADS) == 0 && mode != SINGLE_ALLPROC)
745 		return (0);
746 
747 	/* Is someone already single threading? */
748 	if (p->p_singlethread != NULL && p->p_singlethread != td)
749 		return (1);
750 
751 	if (mode == SINGLE_EXIT) {
752 		p->p_flag |= P_SINGLE_EXIT;
753 		p->p_flag &= ~P_SINGLE_BOUNDARY;
754 	} else {
755 		p->p_flag &= ~P_SINGLE_EXIT;
756 		if (mode == SINGLE_BOUNDARY)
757 			p->p_flag |= P_SINGLE_BOUNDARY;
758 		else
759 			p->p_flag &= ~P_SINGLE_BOUNDARY;
760 	}
761 	if (mode == SINGLE_ALLPROC)
762 		p->p_flag |= P_TOTAL_STOP;
763 	p->p_flag |= P_STOPPED_SINGLE;
764 	PROC_SLOCK(p);
765 	p->p_singlethread = td;
766 	remaining = calc_remaining(p, mode);
767 	while (remaining != remain_for_mode(mode)) {
768 		if (P_SHOULDSTOP(p) != P_STOPPED_SINGLE)
769 			goto stopme;
770 		wakeup_swapper = 0;
771 		FOREACH_THREAD_IN_PROC(p, td2) {
772 			if (td2 == td)
773 				continue;
774 			thread_lock(td2);
775 			td2->td_flags |= TDF_ASTPENDING | TDF_NEEDSUSPCHK;
776 			if (TD_IS_INHIBITED(td2)) {
777 				wakeup_swapper |= weed_inhib(mode, td2, p);
778 #ifdef SMP
779 			} else if (TD_IS_RUNNING(td2) && td != td2) {
780 				forward_signal(td2);
781 #endif
782 			}
783 			thread_unlock(td2);
784 		}
785 		if (wakeup_swapper)
786 			kick_proc0();
787 		remaining = calc_remaining(p, mode);
788 
789 		/*
790 		 * Maybe we suspended some threads.. was it enough?
791 		 */
792 		if (remaining == remain_for_mode(mode))
793 			break;
794 
795 stopme:
796 		/*
797 		 * Wake us up when everyone else has suspended.
798 		 * In the mean time we suspend as well.
799 		 */
800 		thread_suspend_switch(td, p);
801 		remaining = calc_remaining(p, mode);
802 	}
803 	if (mode == SINGLE_EXIT) {
804 		/*
805 		 * Convert the process to an unthreaded process.  The
806 		 * SINGLE_EXIT is called by exit1() or execve(), in
807 		 * both cases other threads must be retired.
808 		 */
809 		KASSERT(p->p_numthreads == 1, ("Unthreading with >1 threads"));
810 		p->p_singlethread = NULL;
811 		p->p_flag &= ~(P_STOPPED_SINGLE | P_SINGLE_EXIT | P_HADTHREADS);
812 
813 		/*
814 		 * Wait for any remaining threads to exit cpu_throw().
815 		 */
816 		while (p->p_exitthreads != 0) {
817 			PROC_SUNLOCK(p);
818 			PROC_UNLOCK(p);
819 			sched_relinquish(td);
820 			PROC_LOCK(p);
821 			PROC_SLOCK(p);
822 		}
823 	} else if (mode == SINGLE_BOUNDARY) {
824 		/*
825 		 * Wait until all suspended threads are removed from
826 		 * the processors.  The thread_suspend_check()
827 		 * increments p_boundary_count while it is still
828 		 * running, which makes it possible for the execve()
829 		 * to destroy vmspace while our other threads are
830 		 * still using the address space.
831 		 *
832 		 * We lock the thread, which is only allowed to
833 		 * succeed after context switch code finished using
834 		 * the address space.
835 		 */
836 		FOREACH_THREAD_IN_PROC(p, td2) {
837 			if (td2 == td)
838 				continue;
839 			thread_lock(td2);
840 			KASSERT((td2->td_flags & TDF_BOUNDARY) != 0,
841 			    ("td %p not on boundary", td2));
842 			KASSERT(TD_IS_SUSPENDED(td2),
843 			    ("td %p is not suspended", td2));
844 			thread_unlock(td2);
845 		}
846 	}
847 	PROC_SUNLOCK(p);
848 	return (0);
849 }
850 
851 bool
852 thread_suspend_check_needed(void)
853 {
854 	struct proc *p;
855 	struct thread *td;
856 
857 	td = curthread;
858 	p = td->td_proc;
859 	PROC_LOCK_ASSERT(p, MA_OWNED);
860 	return (P_SHOULDSTOP(p) || ((p->p_flag & P_TRACED) != 0 &&
861 	    (td->td_dbgflags & TDB_SUSPEND) != 0));
862 }
863 
864 /*
865  * Called in from locations that can safely check to see
866  * whether we have to suspend or at least throttle for a
867  * single-thread event (e.g. fork).
868  *
869  * Such locations include userret().
870  * If the "return_instead" argument is non zero, the thread must be able to
871  * accept 0 (caller may continue), or 1 (caller must abort) as a result.
872  *
873  * The 'return_instead' argument tells the function if it may do a
874  * thread_exit() or suspend, or whether the caller must abort and back
875  * out instead.
876  *
877  * If the thread that set the single_threading request has set the
878  * P_SINGLE_EXIT bit in the process flags then this call will never return
879  * if 'return_instead' is false, but will exit.
880  *
881  * P_SINGLE_EXIT | return_instead == 0| return_instead != 0
882  *---------------+--------------------+---------------------
883  *       0       | returns 0          |   returns 0 or 1
884  *               | when ST ends       |   immediately
885  *---------------+--------------------+---------------------
886  *       1       | thread exits       |   returns 1
887  *               |                    |  immediately
888  * 0 = thread_exit() or suspension ok,
889  * other = return error instead of stopping the thread.
890  *
891  * While a full suspension is under effect, even a single threading
892  * thread would be suspended if it made this call (but it shouldn't).
893  * This call should only be made from places where
894  * thread_exit() would be safe as that may be the outcome unless
895  * return_instead is set.
896  */
897 int
898 thread_suspend_check(int return_instead)
899 {
900 	struct thread *td;
901 	struct proc *p;
902 	int wakeup_swapper;
903 
904 	td = curthread;
905 	p = td->td_proc;
906 	mtx_assert(&Giant, MA_NOTOWNED);
907 	PROC_LOCK_ASSERT(p, MA_OWNED);
908 	while (thread_suspend_check_needed()) {
909 		if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) {
910 			KASSERT(p->p_singlethread != NULL,
911 			    ("singlethread not set"));
912 			/*
913 			 * The only suspension in action is a
914 			 * single-threading. Single threader need not stop.
915 			 * XXX Should be safe to access unlocked
916 			 * as it can only be set to be true by us.
917 			 */
918 			if (p->p_singlethread == td)
919 				return (0);	/* Exempt from stopping. */
920 		}
921 		if ((p->p_flag & P_SINGLE_EXIT) && return_instead)
922 			return (EINTR);
923 
924 		/* Should we goto user boundary if we didn't come from there? */
925 		if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE &&
926 		    (p->p_flag & P_SINGLE_BOUNDARY) && return_instead)
927 			return (ERESTART);
928 
929 		/*
930 		 * Ignore suspend requests if they are deferred.
931 		 */
932 		if ((td->td_flags & TDF_SBDRY) != 0) {
933 			KASSERT(return_instead,
934 			    ("TDF_SBDRY set for unsafe thread_suspend_check"));
935 			return (0);
936 		}
937 
938 		/*
939 		 * If the process is waiting for us to exit,
940 		 * this thread should just suicide.
941 		 * Assumes that P_SINGLE_EXIT implies P_STOPPED_SINGLE.
942 		 */
943 		if ((p->p_flag & P_SINGLE_EXIT) && (p->p_singlethread != td)) {
944 			PROC_UNLOCK(p);
945 			tidhash_remove(td);
946 
947 			/*
948 			 * Allow Linux emulation layer to do some work
949 			 * before thread suicide.
950 			 */
951 			if (__predict_false(p->p_sysent->sv_thread_detach != NULL))
952 				(p->p_sysent->sv_thread_detach)(td);
953 
954 			PROC_LOCK(p);
955 			tdsigcleanup(td);
956 			umtx_thread_exit(td);
957 			PROC_SLOCK(p);
958 			thread_stopped(p);
959 			thread_exit();
960 		}
961 
962 		PROC_SLOCK(p);
963 		thread_stopped(p);
964 		if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) {
965 			if (p->p_numthreads == p->p_suspcount + 1) {
966 				thread_lock(p->p_singlethread);
967 				wakeup_swapper = thread_unsuspend_one(
968 				    p->p_singlethread, p, false);
969 				thread_unlock(p->p_singlethread);
970 				if (wakeup_swapper)
971 					kick_proc0();
972 			}
973 		}
974 		PROC_UNLOCK(p);
975 		thread_lock(td);
976 		/*
977 		 * When a thread suspends, it just
978 		 * gets taken off all queues.
979 		 */
980 		thread_suspend_one(td);
981 		if (return_instead == 0) {
982 			p->p_boundary_count++;
983 			td->td_flags |= TDF_BOUNDARY;
984 		}
985 		PROC_SUNLOCK(p);
986 		mi_switch(SW_INVOL | SWT_SUSPEND, NULL);
987 		thread_unlock(td);
988 		PROC_LOCK(p);
989 	}
990 	return (0);
991 }
992 
993 void
994 thread_suspend_switch(struct thread *td, struct proc *p)
995 {
996 
997 	KASSERT(!TD_IS_SUSPENDED(td), ("already suspended"));
998 	PROC_LOCK_ASSERT(p, MA_OWNED);
999 	PROC_SLOCK_ASSERT(p, MA_OWNED);
1000 	/*
1001 	 * We implement thread_suspend_one in stages here to avoid
1002 	 * dropping the proc lock while the thread lock is owned.
1003 	 */
1004 	if (p == td->td_proc) {
1005 		thread_stopped(p);
1006 		p->p_suspcount++;
1007 	}
1008 	PROC_UNLOCK(p);
1009 	thread_lock(td);
1010 	td->td_flags &= ~TDF_NEEDSUSPCHK;
1011 	TD_SET_SUSPENDED(td);
1012 	sched_sleep(td, 0);
1013 	PROC_SUNLOCK(p);
1014 	DROP_GIANT();
1015 	mi_switch(SW_VOL | SWT_SUSPEND, NULL);
1016 	thread_unlock(td);
1017 	PICKUP_GIANT();
1018 	PROC_LOCK(p);
1019 	PROC_SLOCK(p);
1020 }
1021 
1022 void
1023 thread_suspend_one(struct thread *td)
1024 {
1025 	struct proc *p;
1026 
1027 	p = td->td_proc;
1028 	PROC_SLOCK_ASSERT(p, MA_OWNED);
1029 	THREAD_LOCK_ASSERT(td, MA_OWNED);
1030 	KASSERT(!TD_IS_SUSPENDED(td), ("already suspended"));
1031 	p->p_suspcount++;
1032 	td->td_flags &= ~TDF_NEEDSUSPCHK;
1033 	TD_SET_SUSPENDED(td);
1034 	sched_sleep(td, 0);
1035 }
1036 
1037 static int
1038 thread_unsuspend_one(struct thread *td, struct proc *p, bool boundary)
1039 {
1040 
1041 	THREAD_LOCK_ASSERT(td, MA_OWNED);
1042 	KASSERT(TD_IS_SUSPENDED(td), ("Thread not suspended"));
1043 	TD_CLR_SUSPENDED(td);
1044 	td->td_flags &= ~TDF_ALLPROCSUSP;
1045 	if (td->td_proc == p) {
1046 		PROC_SLOCK_ASSERT(p, MA_OWNED);
1047 		p->p_suspcount--;
1048 		if (boundary && (td->td_flags & TDF_BOUNDARY) != 0) {
1049 			td->td_flags &= ~TDF_BOUNDARY;
1050 			p->p_boundary_count--;
1051 		}
1052 	}
1053 	return (setrunnable(td));
1054 }
1055 
1056 /*
1057  * Allow all threads blocked by single threading to continue running.
1058  */
1059 void
1060 thread_unsuspend(struct proc *p)
1061 {
1062 	struct thread *td;
1063 	int wakeup_swapper;
1064 
1065 	PROC_LOCK_ASSERT(p, MA_OWNED);
1066 	PROC_SLOCK_ASSERT(p, MA_OWNED);
1067 	wakeup_swapper = 0;
1068 	if (!P_SHOULDSTOP(p)) {
1069                 FOREACH_THREAD_IN_PROC(p, td) {
1070 			thread_lock(td);
1071 			if (TD_IS_SUSPENDED(td)) {
1072 				wakeup_swapper |= thread_unsuspend_one(td, p,
1073 				    true);
1074 			}
1075 			thread_unlock(td);
1076 		}
1077 	} else if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE &&
1078 	    p->p_numthreads == p->p_suspcount) {
1079 		/*
1080 		 * Stopping everything also did the job for the single
1081 		 * threading request. Now we've downgraded to single-threaded,
1082 		 * let it continue.
1083 		 */
1084 		if (p->p_singlethread->td_proc == p) {
1085 			thread_lock(p->p_singlethread);
1086 			wakeup_swapper = thread_unsuspend_one(
1087 			    p->p_singlethread, p, false);
1088 			thread_unlock(p->p_singlethread);
1089 		}
1090 	}
1091 	if (wakeup_swapper)
1092 		kick_proc0();
1093 }
1094 
1095 /*
1096  * End the single threading mode..
1097  */
1098 void
1099 thread_single_end(struct proc *p, int mode)
1100 {
1101 	struct thread *td;
1102 	int wakeup_swapper;
1103 
1104 	KASSERT(mode == SINGLE_EXIT || mode == SINGLE_BOUNDARY ||
1105 	    mode == SINGLE_ALLPROC || mode == SINGLE_NO_EXIT,
1106 	    ("invalid mode %d", mode));
1107 	PROC_LOCK_ASSERT(p, MA_OWNED);
1108 	KASSERT((mode == SINGLE_ALLPROC && (p->p_flag & P_TOTAL_STOP) != 0) ||
1109 	    (mode != SINGLE_ALLPROC && (p->p_flag & P_TOTAL_STOP) == 0),
1110 	    ("mode %d does not match P_TOTAL_STOP", mode));
1111 	KASSERT(mode == SINGLE_ALLPROC || p->p_singlethread == curthread,
1112 	    ("thread_single_end from other thread %p %p",
1113 	    curthread, p->p_singlethread));
1114 	KASSERT(mode != SINGLE_BOUNDARY ||
1115 	    (p->p_flag & P_SINGLE_BOUNDARY) != 0,
1116 	    ("mis-matched SINGLE_BOUNDARY flags %x", p->p_flag));
1117 	p->p_flag &= ~(P_STOPPED_SINGLE | P_SINGLE_EXIT | P_SINGLE_BOUNDARY |
1118 	    P_TOTAL_STOP);
1119 	PROC_SLOCK(p);
1120 	p->p_singlethread = NULL;
1121 	wakeup_swapper = 0;
1122 	/*
1123 	 * If there are other threads they may now run,
1124 	 * unless of course there is a blanket 'stop order'
1125 	 * on the process. The single threader must be allowed
1126 	 * to continue however as this is a bad place to stop.
1127 	 */
1128 	if (p->p_numthreads != remain_for_mode(mode) && !P_SHOULDSTOP(p)) {
1129                 FOREACH_THREAD_IN_PROC(p, td) {
1130 			thread_lock(td);
1131 			if (TD_IS_SUSPENDED(td)) {
1132 				wakeup_swapper |= thread_unsuspend_one(td, p,
1133 				    mode == SINGLE_BOUNDARY);
1134 			}
1135 			thread_unlock(td);
1136 		}
1137 	}
1138 	KASSERT(mode != SINGLE_BOUNDARY || p->p_boundary_count == 0,
1139 	    ("inconsistent boundary count %d", p->p_boundary_count));
1140 	PROC_SUNLOCK(p);
1141 	if (wakeup_swapper)
1142 		kick_proc0();
1143 }
1144 
1145 struct thread *
1146 thread_find(struct proc *p, lwpid_t tid)
1147 {
1148 	struct thread *td;
1149 
1150 	PROC_LOCK_ASSERT(p, MA_OWNED);
1151 	FOREACH_THREAD_IN_PROC(p, td) {
1152 		if (td->td_tid == tid)
1153 			break;
1154 	}
1155 	return (td);
1156 }
1157 
1158 /* Locate a thread by number; return with proc lock held. */
1159 struct thread *
1160 tdfind(lwpid_t tid, pid_t pid)
1161 {
1162 #define RUN_THRESH	16
1163 	struct thread *td;
1164 	int run = 0;
1165 
1166 	rw_rlock(&tidhash_lock);
1167 	LIST_FOREACH(td, TIDHASH(tid), td_hash) {
1168 		if (td->td_tid == tid) {
1169 			if (pid != -1 && td->td_proc->p_pid != pid) {
1170 				td = NULL;
1171 				break;
1172 			}
1173 			PROC_LOCK(td->td_proc);
1174 			if (td->td_proc->p_state == PRS_NEW) {
1175 				PROC_UNLOCK(td->td_proc);
1176 				td = NULL;
1177 				break;
1178 			}
1179 			if (run > RUN_THRESH) {
1180 				if (rw_try_upgrade(&tidhash_lock)) {
1181 					LIST_REMOVE(td, td_hash);
1182 					LIST_INSERT_HEAD(TIDHASH(td->td_tid),
1183 						td, td_hash);
1184 					rw_wunlock(&tidhash_lock);
1185 					return (td);
1186 				}
1187 			}
1188 			break;
1189 		}
1190 		run++;
1191 	}
1192 	rw_runlock(&tidhash_lock);
1193 	return (td);
1194 }
1195 
1196 void
1197 tidhash_add(struct thread *td)
1198 {
1199 	rw_wlock(&tidhash_lock);
1200 	LIST_INSERT_HEAD(TIDHASH(td->td_tid), td, td_hash);
1201 	rw_wunlock(&tidhash_lock);
1202 }
1203 
1204 void
1205 tidhash_remove(struct thread *td)
1206 {
1207 	rw_wlock(&tidhash_lock);
1208 	LIST_REMOVE(td, td_hash);
1209 	rw_wunlock(&tidhash_lock);
1210 }
1211