exit.c 45.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10
/*
 *  linux/kernel/exit.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/module.h>
11
#include <linux/capability.h>
L
Linus Torvalds 已提交
12 13 14
#include <linux/completion.h>
#include <linux/personality.h>
#include <linux/tty.h>
15
#include <linux/iocontext.h>
L
Linus Torvalds 已提交
16 17 18 19
#include <linux/key.h>
#include <linux/security.h>
#include <linux/cpu.h>
#include <linux/acct.h>
20
#include <linux/tsacct_kern.h>
L
Linus Torvalds 已提交
21
#include <linux/file.h>
A
Al Viro 已提交
22
#include <linux/fdtable.h>
L
Linus Torvalds 已提交
23
#include <linux/binfmts.h>
S
Serge E. Hallyn 已提交
24
#include <linux/nsproxy.h>
25
#include <linux/pid_namespace.h>
L
Linus Torvalds 已提交
26 27 28 29
#include <linux/ptrace.h>
#include <linux/profile.h>
#include <linux/mount.h>
#include <linux/proc_fs.h>
30
#include <linux/kthread.h>
L
Linus Torvalds 已提交
31
#include <linux/mempolicy.h>
32
#include <linux/taskstats_kern.h>
33
#include <linux/delayacct.h>
34
#include <linux/freezer.h>
35
#include <linux/cgroup.h>
L
Linus Torvalds 已提交
36
#include <linux/syscalls.h>
37
#include <linux/signal.h>
38
#include <linux/posix-timers.h>
M
Matt Helsley 已提交
39
#include <linux/cn_proc.h>
40
#include <linux/mutex.h>
41
#include <linux/futex.h>
42
#include <linux/pipe_fs_i.h>
43
#include <linux/audit.h> /* for audit_free() */
A
Adrian Bunk 已提交
44
#include <linux/resource.h>
45
#include <linux/blkdev.h>
46
#include <linux/task_io_accounting_ops.h>
R
Roland McGrath 已提交
47
#include <linux/tracehook.h>
48
#include <linux/fs_struct.h>
D
David Howells 已提交
49
#include <linux/init_task.h>
50
#include <linux/perf_event.h>
51
#include <trace/events/sched.h>
52
#include <linux/hw_breakpoint.h>
Y
Ying Han 已提交
53
#include <linux/oom.h>
54
#include <linux/writeback.h>
A
Al Viro 已提交
55
#include <linux/shm.h>
L
Linus Torvalds 已提交
56 57 58 59 60 61

#include <asm/uaccess.h>
#include <asm/unistd.h>
#include <asm/pgtable.h>
#include <asm/mmu_context.h>

A
Adrian Bunk 已提交
62 63
static void exit_mm(struct task_struct * tsk);

64
static void __unhash_process(struct task_struct *p, bool group_dead)
L
Linus Torvalds 已提交
65 66
{
	nr_threads--;
67
	detach_pid(p, PIDTYPE_PID);
68
	if (group_dead) {
L
Linus Torvalds 已提交
69 70
		detach_pid(p, PIDTYPE_PGID);
		detach_pid(p, PIDTYPE_SID);
71

72
		list_del_rcu(&p->tasks);
73
		list_del_init(&p->sibling);
74
		__this_cpu_dec(process_counts);
75 76 77 78 79 80 81
		/*
		 * If we are the last child process in a pid namespace to be
		 * reaped, notify the reaper sleeping zap_pid_ns_processes().
		 */
		if (IS_ENABLED(CONFIG_PID_NS)) {
			struct task_struct *parent = p->real_parent;

82
			if ((task_active_pid_ns(parent)->child_reaper == parent) &&
83 84 85 86
			    list_empty(&parent->children) &&
			    (parent->flags & PF_EXITING))
				wake_up_process(parent);
		}
L
Linus Torvalds 已提交
87
	}
O
Oleg Nesterov 已提交
88
	list_del_rcu(&p->thread_group);
L
Linus Torvalds 已提交
89 90
}

91 92 93 94 95 96
/*
 * This function expects the tasklist_lock write-locked.
 */
static void __exit_signal(struct task_struct *tsk)
{
	struct signal_struct *sig = tsk->signal;
97
	bool group_dead = thread_group_leader(tsk);
98
	struct sighand_struct *sighand;
99
	struct tty_struct *uninitialized_var(tty);
100

101
	sighand = rcu_dereference_check(tsk->sighand,
102
					lockdep_tasklist_lock_is_held());
103 104 105
	spin_lock(&sighand->siglock);

	posix_cpu_timers_exit(tsk);
106
	if (group_dead) {
107
		posix_cpu_timers_exit_group(tsk);
108 109
		tty = sig->tty;
		sig->tty = NULL;
110
	} else {
111 112 113 114 115 116 117 118
		/*
		 * This can only happen if the caller is de_thread().
		 * FIXME: this is the temporary hack, we should teach
		 * posix-cpu-timers to handle this case correctly.
		 */
		if (unlikely(has_group_leader_pid(tsk)))
			posix_cpu_timers_exit_group(tsk);

119 120 121 122
		/*
		 * If there is any task waiting for the group exit
		 * then notify it:
		 */
123
		if (sig->notify_count > 0 && !--sig->notify_count)
124
			wake_up_process(sig->group_exit_task);
125

126 127 128 129 130 131 132 133 134 135 136 137
		if (tsk == sig->curr_target)
			sig->curr_target = next_thread(tsk);
		/*
		 * Accumulate here the counters for all threads but the
		 * group leader as they die, so they can be added into
		 * the process-wide totals when those are taken.
		 * The group leader stays around as a zombie as long
		 * as there are other threads.  When it gets reaped,
		 * the exit.c code will add its counts into these totals.
		 * We won't ever get here for the group leader, since it
		 * will have been the last reference on the signal_struct.
		 */
138 139 140
		sig->utime += tsk->utime;
		sig->stime += tsk->stime;
		sig->gtime += tsk->gtime;
141 142 143 144
		sig->min_flt += tsk->min_flt;
		sig->maj_flt += tsk->maj_flt;
		sig->nvcsw += tsk->nvcsw;
		sig->nivcsw += tsk->nivcsw;
145 146
		sig->inblock += task_io_get_inblock(tsk);
		sig->oublock += task_io_get_oublock(tsk);
147
		task_io_accounting_add(&sig->ioac, &tsk->ioac);
148
		sig->sum_sched_runtime += tsk->se.sum_exec_runtime;
149 150
	}

151
	sig->nr_threads--;
152
	__unhash_process(tsk, group_dead);
153

154 155 156 157 158
	/*
	 * Do this under ->siglock, we can race with another thread
	 * doing sigqueue_free() if we have SIGQUEUE_PREALLOC signals.
	 */
	flush_sigqueue(&tsk->pending);
159
	tsk->sighand = NULL;
160 161
	spin_unlock(&sighand->siglock);

162
	__cleanup_sighand(sighand);
163
	clear_tsk_thread_flag(tsk,TIF_SIGPENDING);
164
	if (group_dead) {
165
		flush_sigqueue(&sig->shared_pending);
166
		tty_kref_put(tty);
167 168 169
	}
}

170 171
static void delayed_put_task_struct(struct rcu_head *rhp)
{
172 173
	struct task_struct *tsk = container_of(rhp, struct task_struct, rcu);

174
	perf_event_delayed_put(tsk);
175 176
	trace_sched_process_free(tsk);
	put_task_struct(tsk);
177 178
}

R
Roland McGrath 已提交
179

L
Linus Torvalds 已提交
180 181
void release_task(struct task_struct * p)
{
182
	struct task_struct *leader;
L
Linus Torvalds 已提交
183
	int zap_leader;
184
repeat:
185
	/* don't need to get the RCU readlock here - the process is dead and
186 187
	 * can't be modifying its own credentials. But shut RCU-lockdep up */
	rcu_read_lock();
188
	atomic_dec(&__task_cred(p)->user->processes);
189
	rcu_read_unlock();
190

191
	proc_flush_task(p);
192

L
Linus Torvalds 已提交
193
	write_lock_irq(&tasklist_lock);
T
Tejun Heo 已提交
194
	ptrace_release_task(p);
L
Linus Torvalds 已提交
195
	__exit_signal(p);
196

L
Linus Torvalds 已提交
197 198 199 200 201 202 203 204 205 206 207 208
	/*
	 * If we are the last non-leader member of the thread
	 * group, and the leader is zombie, then notify the
	 * group leader's parent process. (if it wants notification.)
	 */
	zap_leader = 0;
	leader = p->group_leader;
	if (leader != p && thread_group_empty(leader) && leader->exit_state == EXIT_ZOMBIE) {
		/*
		 * If we were the last child thread and the leader has
		 * exited already, and the leader's parent ignores SIGCHLD,
		 * then we are the one who should release the leader.
R
Roland McGrath 已提交
209
		 */
210
		zap_leader = do_notify_parent(leader, leader->exit_signal);
R
Roland McGrath 已提交
211 212
		if (zap_leader)
			leader->exit_state = EXIT_DEAD;
L
Linus Torvalds 已提交
213 214 215 216
	}

	write_unlock_irq(&tasklist_lock);
	release_thread(p);
217
	call_rcu(&p->rcu, delayed_put_task_struct);
L
Linus Torvalds 已提交
218 219 220 221 222 223 224 225 226 227

	p = leader;
	if (unlikely(zap_leader))
		goto repeat;
}

/*
 * This checks not only the pgrp, but falls back on the pid if no
 * satisfactory pgrp is found. I dunno - gdb doesn't work correctly
 * without this...
228 229
 *
 * The caller must hold rcu lock or the tasklist lock.
L
Linus Torvalds 已提交
230
 */
231
struct pid *session_of_pgrp(struct pid *pgrp)
L
Linus Torvalds 已提交
232 233
{
	struct task_struct *p;
234
	struct pid *sid = NULL;
235

236
	p = pid_task(pgrp, PIDTYPE_PGID);
237
	if (p == NULL)
238
		p = pid_task(pgrp, PIDTYPE_PID);
239
	if (p != NULL)
240
		sid = task_session(p);
241

L
Linus Torvalds 已提交
242 243 244 245 246 247 248 249 250 251 252
	return sid;
}

/*
 * Determine if a process group is "orphaned", according to the POSIX
 * definition in 2.2.2.52.  Orphaned process groups are not to be affected
 * by terminal-generated stop signals.  Newly orphaned process groups are
 * to receive a SIGHUP and a SIGCONT.
 *
 * "I ask you, have you ever known what it is to be an orphan?"
 */
253
static int will_become_orphaned_pgrp(struct pid *pgrp, struct task_struct *ignored_task)
L
Linus Torvalds 已提交
254 255 256
{
	struct task_struct *p;

257
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
258 259 260
		if ((p == ignored_task) ||
		    (p->exit_state && thread_group_empty(p)) ||
		    is_global_init(p->real_parent))
L
Linus Torvalds 已提交
261
			continue;
262

263
		if (task_pgrp(p->real_parent) != pgrp &&
264 265
		    task_session(p->real_parent) == task_session(p))
			return 0;
266
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
267 268

	return 1;
L
Linus Torvalds 已提交
269 270
}

271
int is_current_pgrp_orphaned(void)
L
Linus Torvalds 已提交
272 273 274 275
{
	int retval;

	read_lock(&tasklist_lock);
276
	retval = will_become_orphaned_pgrp(task_pgrp(current), NULL);
L
Linus Torvalds 已提交
277 278 279 280 281
	read_unlock(&tasklist_lock);

	return retval;
}

282
static bool has_stopped_jobs(struct pid *pgrp)
L
Linus Torvalds 已提交
283 284 285
{
	struct task_struct *p;

286
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
287 288
		if (p->signal->flags & SIGNAL_STOP_STOPPED)
			return true;
289
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
290 291

	return false;
L
Linus Torvalds 已提交
292 293
}

294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
/*
 * Check to see if any process groups have become orphaned as
 * a result of our exiting, and if they have any stopped jobs,
 * send them a SIGHUP and then a SIGCONT. (POSIX 3.2.2.2)
 */
static void
kill_orphaned_pgrp(struct task_struct *tsk, struct task_struct *parent)
{
	struct pid *pgrp = task_pgrp(tsk);
	struct task_struct *ignored_task = tsk;

	if (!parent)
		 /* exit: our father is in a different pgrp than
		  * we are and we were the only connection outside.
		  */
		parent = tsk->real_parent;
	else
		/* reparent: our child is in a different pgrp than
		 * we are, and it was the only connection outside.
		 */
		ignored_task = NULL;

	if (task_pgrp(parent) != pgrp &&
	    task_session(parent) == task_session(tsk) &&
	    will_become_orphaned_pgrp(pgrp, ignored_task) &&
	    has_stopped_jobs(pgrp)) {
		__kill_pgrp_info(SIGHUP, SEND_SIG_PRIV, pgrp);
		__kill_pgrp_info(SIGCONT, SEND_SIG_PRIV, pgrp);
	}
}

L
Linus Torvalds 已提交
325
/**
326
 * reparent_to_kthreadd - Reparent the calling kernel thread to kthreadd
L
Linus Torvalds 已提交
327 328
 *
 * If a kernel thread is launched as a result of a system call, or if
329 330
 * it ever exits, it should generally reparent itself to kthreadd so it
 * isn't in the way of other processes and is correctly cleaned up on exit.
L
Linus Torvalds 已提交
331 332 333 334
 *
 * The various task state such as scheduling policy and priority may have
 * been inherited from a user process, so we reset them to sane values here.
 *
335
 * NOTE that reparent_to_kthreadd() gives the caller full capabilities.
L
Linus Torvalds 已提交
336
 */
337
static void reparent_to_kthreadd(void)
L
Linus Torvalds 已提交
338 339 340 341 342
{
	write_lock_irq(&tasklist_lock);

	ptrace_unlink(current);
	/* Reparent to init */
343
	current->real_parent = current->parent = kthreadd_task;
R
Roland McGrath 已提交
344
	list_move_tail(&current->sibling, &current->real_parent->children);
L
Linus Torvalds 已提交
345 346 347 348

	/* Set the exit signal to SIGCHLD so we signal init on exit */
	current->exit_signal = SIGCHLD;

349
	if (task_nice(current) < 0)
L
Linus Torvalds 已提交
350 351 352 353 354 355
		set_user_nice(current, 0);
	/* cpus_allowed? */
	/* rt_priority? */
	/* signals? */
	memcpy(current->signal->rlim, init_task.signal->rlim,
	       sizeof(current->signal->rlim));
D
David Howells 已提交
356 357 358

	atomic_inc(&init_cred.usage);
	commit_creds(&init_cred);
L
Linus Torvalds 已提交
359 360 361
	write_unlock_irq(&tasklist_lock);
}

362
void __set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
363
{
364
	struct task_struct *curr = current->group_leader;
L
Linus Torvalds 已提交
365

366
	if (task_session(curr) != pid)
367
		change_pid(curr, PIDTYPE_SID, pid);
368 369

	if (task_pgrp(curr) != pid)
370
		change_pid(curr, PIDTYPE_PGID, pid);
L
Linus Torvalds 已提交
371 372
}

373
static void set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
374 375
{
	write_lock_irq(&tasklist_lock);
376
	__set_special_pids(pid);
L
Linus Torvalds 已提交
377 378 379 380
	write_unlock_irq(&tasklist_lock);
}

/*
381 382
 * Let kernel threads use this to say that they allow a certain signal.
 * Must not be used if kthread was cloned with CLONE_SIGHAND.
L
Linus Torvalds 已提交
383 384 385
 */
int allow_signal(int sig)
{
386
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
387 388 389
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
390
	/* This is only needed for daemonize()'ed kthreads */
L
Linus Torvalds 已提交
391
	sigdelset(&current->blocked, sig);
392 393 394 395 396 397
	/*
	 * Kernel threads handle their own signals. Let the signal code
	 * know it'll be handled, so that they don't get converted to
	 * SIGKILL or just silently dropped.
	 */
	current->sighand->action[(sig)-1].sa.sa_handler = (void __user *)2;
L
Linus Torvalds 已提交
398 399 400 401 402 403 404 405 406
	recalc_sigpending();
	spin_unlock_irq(&current->sighand->siglock);
	return 0;
}

EXPORT_SYMBOL(allow_signal);

int disallow_signal(int sig)
{
407
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
408 409 410
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
411
	current->sighand->action[(sig)-1].sa.sa_handler = SIG_IGN;
L
Linus Torvalds 已提交
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
	recalc_sigpending();
	spin_unlock_irq(&current->sighand->siglock);
	return 0;
}

EXPORT_SYMBOL(disallow_signal);

/*
 *	Put all the gunge required to become a kernel thread without
 *	attached user resources in one place where it belongs.
 */

void daemonize(const char *name, ...)
{
	va_list args;
	sigset_t blocked;

	va_start(args, name);
	vsnprintf(current->comm, sizeof(current->comm), name, args);
	va_end(args);

	/*
	 * If we were started as result of loading a module, close all of the
	 * user space pages.  We don't need them, and if we didn't close them
	 * they would be locked into memory.
	 */
	exit_mm(current);
439
	/*
440
	 * We don't want to get frozen, in case system-wide hibernation
441 442
	 * or suspend transition begins right now.
	 */
O
Oleg Nesterov 已提交
443
	current->flags |= (PF_NOFREEZE | PF_KTHREAD);
L
Linus Torvalds 已提交
444

445 446 447 448
	if (current->nsproxy != &init_nsproxy) {
		get_nsproxy(&init_nsproxy);
		switch_task_namespaces(current, &init_nsproxy);
	}
449
	set_special_pids(&init_struct_pid);
450
	proc_clear_tty(current);
L
Linus Torvalds 已提交
451 452 453 454 455 456 457 458

	/* Block and flush all signals */
	sigfillset(&blocked);
	sigprocmask(SIG_BLOCK, &blocked, NULL);
	flush_signals(current);

	/* Become as one with the init task */

459
	daemonize_fs_struct();
A
Al Viro 已提交
460
	daemonize_descriptors();
L
Linus Torvalds 已提交
461

462
	reparent_to_kthreadd();
L
Linus Torvalds 已提交
463 464 465 466
}

EXPORT_SYMBOL(daemonize);

467 468
#ifdef CONFIG_MM_OWNER
/*
469
 * A task is exiting.   If it owned this mm, find a new owner for the mm.
470 471 472 473 474 475
 */
void mm_update_next_owner(struct mm_struct *mm)
{
	struct task_struct *c, *g, *p = current;

retry:
476 477 478 479 480
	/*
	 * If the exiting or execing task is not the owner, it's
	 * someone else's problem.
	 */
	if (mm->owner != p)
481
		return;
482 483 484 485 486 487 488 489 490
	/*
	 * The current owner is exiting/execing and there are no other
	 * candidates.  Do not leave the mm pointing to a possibly
	 * freed task structure.
	 */
	if (atomic_read(&mm->mm_users) <= 1) {
		mm->owner = NULL;
		return;
	}
491 492 493 494 495 496 497 498 499 500 501 502 503

	read_lock(&tasklist_lock);
	/*
	 * Search in the children
	 */
	list_for_each_entry(c, &p->children, sibling) {
		if (c->mm == mm)
			goto assign_new_owner;
	}

	/*
	 * Search in the siblings
	 */
504
	list_for_each_entry(c, &p->real_parent->children, sibling) {
505 506 507 508 509 510 511 512 513 514 515 516 517 518
		if (c->mm == mm)
			goto assign_new_owner;
	}

	/*
	 * Search through everything else. We should not get
	 * here often
	 */
	do_each_thread(g, c) {
		if (c->mm == mm)
			goto assign_new_owner;
	} while_each_thread(g, c);

	read_unlock(&tasklist_lock);
519 520 521
	/*
	 * We found no owner yet mm_users > 1: this implies that we are
	 * most likely racing with swapoff (try_to_unuse()) or /proc or
522
	 * ptrace or page migration (get_task_mm()).  Mark owner as NULL.
523 524
	 */
	mm->owner = NULL;
525 526 527 528 529 530 531 532 533 534
	return;

assign_new_owner:
	BUG_ON(c == p);
	get_task_struct(c);
	/*
	 * The task_lock protects c->mm from changing.
	 * We always want mm->owner->mm == mm
	 */
	task_lock(c);
535 536 537 538 539
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
540 541 542 543 544 545 546 547 548 549 550
	if (c->mm != mm) {
		task_unlock(c);
		put_task_struct(c);
		goto retry;
	}
	mm->owner = c;
	task_unlock(c);
	put_task_struct(c);
}
#endif /* CONFIG_MM_OWNER */

L
Linus Torvalds 已提交
551 552 553 554
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
555
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
556 557
{
	struct mm_struct *mm = tsk->mm;
558
	struct core_state *core_state;
L
Linus Torvalds 已提交
559

560
	mm_release(tsk, mm);
L
Linus Torvalds 已提交
561 562
	if (!mm)
		return;
563
	sync_mm_rss(mm);
L
Linus Torvalds 已提交
564 565
	/*
	 * Serialize with any possible pending coredump.
566
	 * We must hold mmap_sem around checking core_state
L
Linus Torvalds 已提交
567
	 * and clearing tsk->mm.  The core-inducing thread
568
	 * will increment ->nr_threads for each thread in the
L
Linus Torvalds 已提交
569 570 571
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
572 573 574
	core_state = mm->core_state;
	if (core_state) {
		struct core_thread self;
L
Linus Torvalds 已提交
575 576
		up_read(&mm->mmap_sem);

577 578 579 580 581 582 583 584
		self.task = tsk;
		self.next = xchg(&core_state->dumper.next, &self);
		/*
		 * Implies mb(), the result of xchg() must be visible
		 * to core_state->dumper.
		 */
		if (atomic_dec_and_test(&core_state->nr_threads))
			complete(&core_state->startup);
L
Linus Torvalds 已提交
585

O
Oleg Nesterov 已提交
586 587 588 589 590 591 592
		for (;;) {
			set_task_state(tsk, TASK_UNINTERRUPTIBLE);
			if (!self.task) /* see coredump_finish() */
				break;
			schedule();
		}
		__set_task_state(tsk, TASK_RUNNING);
L
Linus Torvalds 已提交
593 594 595
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
596
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
597 598 599 600 601 602
	/* more a memory barrier than a real lock */
	task_lock(tsk);
	tsk->mm = NULL;
	up_read(&mm->mmap_sem);
	enter_lazy_tlb(mm, current);
	task_unlock(tsk);
603
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
604 605 606 607
	mmput(mm);
}

/*
608 609 610 611 612
 * When we die, we re-parent all our children, and try to:
 * 1. give them to another thread in our thread group, if such a member exists
 * 2. give it to the first ancestor process which prctl'd itself as a
 *    child_subreaper for its children (like a service manager)
 * 3. give it to the init process (PID 1) in our pid namespace
L
Linus Torvalds 已提交
613
 */
614
static struct task_struct *find_new_reaper(struct task_struct *father)
615 616
	__releases(&tasklist_lock)
	__acquires(&tasklist_lock)
L
Linus Torvalds 已提交
617
{
618 619
	struct pid_namespace *pid_ns = task_active_pid_ns(father);
	struct task_struct *thread;
L
Linus Torvalds 已提交
620

621 622 623 624 625 626 627 628
	thread = father;
	while_each_thread(father, thread) {
		if (thread->flags & PF_EXITING)
			continue;
		if (unlikely(pid_ns->child_reaper == father))
			pid_ns->child_reaper = thread;
		return thread;
	}
L
Linus Torvalds 已提交
629

630 631
	if (unlikely(pid_ns->child_reaper == father)) {
		write_unlock_irq(&tasklist_lock);
632 633 634 635 636
		if (unlikely(pid_ns == &init_pid_ns)) {
			panic("Attempted to kill init! exitcode=0x%08x\n",
				father->signal->group_exit_code ?:
					father->exit_code);
		}
L
Linus Torvalds 已提交
637

638 639
		zap_pid_ns_processes(pid_ns);
		write_lock_irq(&tasklist_lock);
640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662
	} else if (father->signal->has_child_subreaper) {
		struct task_struct *reaper;

		/*
		 * Find the first ancestor marked as child_subreaper.
		 * Note that the code below checks same_thread_group(reaper,
		 * pid_ns->child_reaper).  This is what we need to DTRT in a
		 * PID namespace. However we still need the check above, see
		 * http://marc.info/?l=linux-kernel&m=131385460420380
		 */
		for (reaper = father->real_parent;
		     reaper != &init_task;
		     reaper = reaper->real_parent) {
			if (same_thread_group(reaper, pid_ns->child_reaper))
				break;
			if (!reaper->signal->is_child_subreaper)
				continue;
			thread = reaper;
			do {
				if (!(thread->flags & PF_EXITING))
					return reaper;
			} while_each_thread(reaper, thread);
		}
L
Linus Torvalds 已提交
663
	}
664

665 666 667
	return pid_ns->child_reaper;
}

668 669 670
/*
* Any that need to be release_task'd are put on the @dead list.
 */
671
static void reparent_leader(struct task_struct *father, struct task_struct *p,
672 673 674 675
				struct list_head *dead)
{
	list_move_tail(&p->sibling, &p->real_parent->children);

676
	if (p->exit_state == EXIT_DEAD)
677 678 679 680 681 682 683 684 685 686 687 688
		return;
	/*
	 * If this is a threaded reparent there is no need to
	 * notify anyone anything has happened.
	 */
	if (same_thread_group(p->real_parent, father))
		return;

	/* We don't want people slaying init.  */
	p->exit_signal = SIGCHLD;

	/* If it has exited notify the new parent about this child's death. */
T
Tejun Heo 已提交
689
	if (!p->ptrace &&
690
	    p->exit_state == EXIT_ZOMBIE && thread_group_empty(p)) {
691
		if (do_notify_parent(p, p->exit_signal)) {
692 693 694 695 696 697 698 699
			p->exit_state = EXIT_DEAD;
			list_move_tail(&p->sibling, dead);
		}
	}

	kill_orphaned_pgrp(p, father);
}

700
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
701
{
702
	struct task_struct *p, *n, *reaper;
703
	LIST_HEAD(dead_children);
704 705

	write_lock_irq(&tasklist_lock);
706 707 708 709 710
	/*
	 * Note that exit_ptrace() and find_new_reaper() might
	 * drop tasklist_lock and reacquire it.
	 */
	exit_ptrace(father);
711
	reaper = find_new_reaper(father);
R
Roland McGrath 已提交
712

713
	list_for_each_entry_safe(p, n, &father->children, sibling) {
714 715 716 717
		struct task_struct *t = p;
		do {
			t->real_parent = reaper;
			if (t->parent == father) {
T
Tejun Heo 已提交
718
				BUG_ON(t->ptrace);
719 720 721 722 723 724 725
				t->parent = t->real_parent;
			}
			if (t->pdeath_signal)
				group_send_sig_info(t->pdeath_signal,
						    SEND_SIG_NOINFO, t);
		} while_each_thread(p, t);
		reparent_leader(father, p, &dead_children);
L
Linus Torvalds 已提交
726
	}
727
	write_unlock_irq(&tasklist_lock);
728

729 730
	BUG_ON(!list_empty(&father->children));

731 732
	list_for_each_entry_safe(p, n, &dead_children, sibling) {
		list_del_init(&p->sibling);
733 734
		release_task(p);
	}
L
Linus Torvalds 已提交
735 736 737 738 739 740
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
741
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
742
{
743
	bool autoreap;
L
Linus Torvalds 已提交
744 745 746 747 748 749 750 751 752

	/*
	 * This does two things:
	 *
  	 * A.  Make init inherit all the child processes
	 * B.  Check to see if any process groups have become orphaned
	 *	as a result of our exiting, and if they have any stopped
	 *	jobs, send them a SIGHUP and then a SIGCONT.  (POSIX 3.2.2.2)
	 */
753
	forget_original_parent(tsk);
754
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
755

756
	write_lock_irq(&tasklist_lock);
757 758
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
759

O
Oleg Nesterov 已提交
760 761 762 763 764 765 766 767 768 769 770 771
	if (unlikely(tsk->ptrace)) {
		int sig = thread_group_leader(tsk) &&
				thread_group_empty(tsk) &&
				!ptrace_reparented(tsk) ?
			tsk->exit_signal : SIGCHLD;
		autoreap = do_notify_parent(tsk, sig);
	} else if (thread_group_leader(tsk)) {
		autoreap = thread_group_empty(tsk) &&
			do_notify_parent(tsk, tsk->exit_signal);
	} else {
		autoreap = true;
	}
L
Linus Torvalds 已提交
772

773
	tsk->exit_state = autoreap ? EXIT_DEAD : EXIT_ZOMBIE;
L
Linus Torvalds 已提交
774

775 776
	/* mt-exec, de_thread() is waiting for group leader */
	if (unlikely(tsk->signal->notify_count < 0))
777
		wake_up_process(tsk->signal->group_exit_task);
L
Linus Torvalds 已提交
778 779 780
	write_unlock_irq(&tasklist_lock);

	/* If the process is dead, release it - nobody will wait for it */
781
	if (autoreap)
L
Linus Torvalds 已提交
782 783 784
		release_task(tsk);
}

785 786 787 788 789 790 791
#ifdef CONFIG_DEBUG_STACK_USAGE
static void check_stack_usage(void)
{
	static DEFINE_SPINLOCK(low_water_lock);
	static int lowest_to_date = THREAD_SIZE;
	unsigned long free;

792
	free = stack_not_used(current);
793 794 795 796 797 798

	if (free >= lowest_to_date)
		return;

	spin_lock(&low_water_lock);
	if (free < lowest_to_date) {
799 800 801
		printk(KERN_WARNING "%s (%d) used greatest stack depth: "
				"%lu bytes left\n",
				current->comm, task_pid_nr(current), free);
802 803 804 805 806 807 808 809
		lowest_to_date = free;
	}
	spin_unlock(&low_water_lock);
}
#else
static inline void check_stack_usage(void) {}
#endif

810
void do_exit(long code)
L
Linus Torvalds 已提交
811 812 813 814 815 816
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

817
	WARN_ON(blk_needs_flush_plug(tsk));
818

L
Linus Torvalds 已提交
819 820 821 822 823
	if (unlikely(in_interrupt()))
		panic("Aiee, killing interrupt handler!");
	if (unlikely(!tsk->pid))
		panic("Attempted to kill the idle task!");

824 825 826 827 828 829 830 831 832
	/*
	 * If do_exit is called because this processes oopsed, it's possible
	 * that get_fs() was left as KERNEL_DS, so reset it to USER_DS before
	 * continuing. Amongst other possible reasons, this is to prevent
	 * mm_release()->clear_child_tid() from writing to a user-controlled
	 * kernel address.
	 */
	set_fs(USER_DS);

T
Tejun Heo 已提交
833
	ptrace_event(PTRACE_EVENT_EXIT, code);
L
Linus Torvalds 已提交
834

835 836
	validate_creds_for_do_exit(tsk);

837 838 839 840 841 842 843
	/*
	 * We're taking recursive faults here in do_exit. Safest is to just
	 * leave this task alone and wait for reboot.
	 */
	if (unlikely(tsk->flags & PF_EXITING)) {
		printk(KERN_ALERT
			"Fixing recursive fault but reboot is needed!\n");
844 845 846 847 848 849 850 851 852 853
		/*
		 * We can do this unlocked here. The futex code uses
		 * this flag just to verify whether the pi state
		 * cleanup has been done or not. In the worst case it
		 * loops once more. We pretend that the cleanup was
		 * done as there is no way to return. Either the
		 * OWNER_DIED bit is set by now or we push the blocked
		 * task into the wait for ever nirwana as well.
		 */
		tsk->flags |= PF_EXITPIDONE;
854 855 856 857
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

O
Oleg Nesterov 已提交
858
	exit_signals(tsk);  /* sets PF_EXITING */
859 860
	/*
	 * tsk->flags are checked in the futex code to protect against
861
	 * an exiting task cleaning up the robust pi futexes.
862
	 */
863
	smp_mb();
864
	raw_spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
865 866 867

	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
868
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
869 870 871
				preempt_count());

	acct_update_integrals(tsk);
872 873 874
	/* sync mm's RSS info before statistics gathering */
	if (tsk->mm)
		sync_mm_rss(tsk->mm);
L
Linus Torvalds 已提交
875
	group_dead = atomic_dec_and_test(&tsk->signal->live);
876
	if (group_dead) {
877
		hrtimer_cancel(&tsk->signal->real_timer);
878
		exit_itimers(tsk->signal);
J
Jiri Pirko 已提交
879 880
		if (tsk->mm)
			setmax_mm_hiwater_rss(&tsk->signal->maxrss, tsk->mm);
881
	}
882
	acct_collect(code, group_dead);
M
Miloslav Trmac 已提交
883 884
	if (group_dead)
		tty_audit_exit();
885
	audit_free(tsk);
886

887
	tsk->exit_code = code;
888
	taskstats_exit(tsk, group_dead);
889

L
Linus Torvalds 已提交
890 891
	exit_mm(tsk);

892
	if (group_dead)
893
		acct_process();
894 895
	trace_sched_process_exit(tsk);

L
Linus Torvalds 已提交
896
	exit_sem(tsk);
897
	exit_shm(tsk);
898 899
	exit_files(tsk);
	exit_fs(tsk);
900
	exit_task_work(tsk);
901
	check_stack_usage();
L
Linus Torvalds 已提交
902
	exit_thread();
903 904 905 906 907 908 909 910 911

	/*
	 * Flush inherited counters to the parent - before the parent
	 * gets woken up by child-exit notifications.
	 *
	 * because of cgroup mode, must be called before cgroup_exit()
	 */
	perf_event_exit_task(tsk);

912
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
913

914
	if (group_dead)
L
Linus Torvalds 已提交
915 916
		disassociate_ctty(1);

A
Al Viro 已提交
917
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
918

M
Matt Helsley 已提交
919
	proc_exit_connector(tsk);
920

921 922 923
	/*
	 * FIXME: do that only when needed, using sched_exit tracepoint
	 */
924
	ptrace_put_breakpoints(tsk);
925

926
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
927
#ifdef CONFIG_NUMA
928
	task_lock(tsk);
929
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
930
	tsk->mempolicy = NULL;
931
	task_unlock(tsk);
L
Linus Torvalds 已提交
932
#endif
933
#ifdef CONFIG_FUTEX
934 935
	if (unlikely(current->pi_state_cache))
		kfree(current->pi_state_cache);
936
#endif
937
	/*
938
	 * Make sure we are holding no locks:
939
	 */
940
	debug_check_no_locks_held(tsk);
941 942 943 944 945 946
	/*
	 * We can do this unlocked here. The futex code uses this flag
	 * just to verify whether the pi state cleanup has been done
	 * or not. In the worst case it loops once more.
	 */
	tsk->flags |= PF_EXITPIDONE;
L
Linus Torvalds 已提交
947

948
	if (tsk->io_context)
949
		exit_io_context(tsk);
950

951 952 953
	if (tsk->splice_pipe)
		__free_pipe_info(tsk->splice_pipe);

954 955 956
	if (tsk->task_frag.page)
		put_page(tsk->task_frag.page);

957 958
	validate_creds_for_do_exit(tsk);

C
Coywolf Qi Hunt 已提交
959
	preempt_disable();
960 961
	if (tsk->nr_dirtied)
		__this_cpu_add(dirty_throttle_leaks, tsk->nr_dirtied);
962
	exit_rcu();
963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978

	/*
	 * The setting of TASK_RUNNING by try_to_wake_up() may be delayed
	 * when the following two conditions become true.
	 *   - There is race condition of mmap_sem (It is acquired by
	 *     exit_mm()), and
	 *   - SMI occurs before setting TASK_RUNINNG.
	 *     (or hypervisor of virtual machine switches to other guest)
	 *  As a result, we may become TASK_RUNNING after becoming TASK_DEAD
	 *
	 * To avoid it, we have to wait for releasing tsk->pi_lock which
	 * is held by try_to_wake_up()
	 */
	smp_mb();
	raw_spin_unlock_wait(&tsk->pi_lock);

O
Oleg Nesterov 已提交
979
	/* causes final put_task_struct in finish_task_switch(). */
980
	tsk->state = TASK_DEAD;
981
	tsk->flags |= PF_NOFREEZE;	/* tell freezer to ignore us */
L
Linus Torvalds 已提交
982 983 984
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
985 986
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
987 988
}

989 990
EXPORT_SYMBOL_GPL(do_exit);

991
void complete_and_exit(struct completion *comp, long code)
L
Linus Torvalds 已提交
992 993 994
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
995

L
Linus Torvalds 已提交
996 997 998 999 1000
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

1001
SYSCALL_DEFINE1(exit, int, error_code)
L
Linus Torvalds 已提交
1002 1003 1004 1005 1006 1007 1008 1009
{
	do_exit((error_code&0xff)<<8);
}

/*
 * Take down every thread in the group.  This is called by fatal signals
 * as well as by sys_exit_group (below).
 */
1010
void
L
Linus Torvalds 已提交
1011 1012
do_group_exit(int exit_code)
{
1013 1014
	struct signal_struct *sig = current->signal;

L
Linus Torvalds 已提交
1015 1016
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

1017 1018
	if (signal_group_exit(sig))
		exit_code = sig->group_exit_code;
L
Linus Torvalds 已提交
1019 1020 1021
	else if (!thread_group_empty(current)) {
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1022
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1023 1024 1025 1026
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1027
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
			zap_other_threads(current);
		}
		spin_unlock_irq(&sighand->siglock);
	}

	do_exit(exit_code);
	/* NOTREACHED */
}

/*
 * this kills every thread in the thread group. Note that any externally
 * wait4()-ing process will get the correct exit code - even if this
 * thread is not the thread group leader.
 */
1042
SYSCALL_DEFINE1(exit_group, int, error_code)
L
Linus Torvalds 已提交
1043 1044
{
	do_group_exit((error_code & 0xff) << 8);
1045 1046
	/* NOTREACHED */
	return 0;
L
Linus Torvalds 已提交
1047 1048
}

1049 1050 1051
struct wait_opts {
	enum pid_type		wo_type;
	int			wo_flags;
1052
	struct pid		*wo_pid;
1053 1054 1055 1056 1057

	struct siginfo __user	*wo_info;
	int __user		*wo_stat;
	struct rusage __user	*wo_rusage;

1058
	wait_queue_t		child_wait;
1059 1060 1061
	int			notask_error;
};

1062 1063
static inline
struct pid *task_pid_type(struct task_struct *task, enum pid_type type)
E
Eric W. Biederman 已提交
1064
{
1065 1066 1067
	if (type != PIDTYPE_PID)
		task = task->group_leader;
	return task->pids[type].pid;
E
Eric W. Biederman 已提交
1068 1069
}

1070
static int eligible_pid(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1071
{
1072 1073 1074
	return	wo->wo_type == PIDTYPE_MAX ||
		task_pid_type(p, wo->wo_type) == wo->wo_pid;
}
L
Linus Torvalds 已提交
1075

1076 1077 1078 1079
static int eligible_child(struct wait_opts *wo, struct task_struct *p)
{
	if (!eligible_pid(wo, p))
		return 0;
L
Linus Torvalds 已提交
1080 1081 1082 1083 1084
	/* Wait for all children (clone and not) if __WALL is set;
	 * otherwise, wait for clone children *only* if __WCLONE is
	 * set; otherwise, wait for non-clone children *only*.  (Note:
	 * A "clone" child here is one that reports to its parent
	 * using a signal other than SIGCHLD.) */
1085 1086
	if (((p->exit_signal != SIGCHLD) ^ !!(wo->wo_flags & __WCLONE))
	    && !(wo->wo_flags & __WALL))
L
Linus Torvalds 已提交
1087 1088
		return 0;

1089
	return 1;
L
Linus Torvalds 已提交
1090 1091
}

1092 1093
static int wait_noreap_copyout(struct wait_opts *wo, struct task_struct *p,
				pid_t pid, uid_t uid, int why, int status)
L
Linus Torvalds 已提交
1094
{
1095 1096 1097
	struct siginfo __user *infop;
	int retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
1098

L
Linus Torvalds 已提交
1099
	put_task_struct(p);
1100
	infop = wo->wo_info;
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
	if (infop) {
		if (!retval)
			retval = put_user(SIGCHLD, &infop->si_signo);
		if (!retval)
			retval = put_user(0, &infop->si_errno);
		if (!retval)
			retval = put_user((short)why, &infop->si_code);
		if (!retval)
			retval = put_user(pid, &infop->si_pid);
		if (!retval)
			retval = put_user(uid, &infop->si_uid);
		if (!retval)
			retval = put_user(status, &infop->si_status);
	}
L
Linus Torvalds 已提交
1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
	if (!retval)
		retval = pid;
	return retval;
}

/*
 * Handle sys_wait4 work for one task in state EXIT_ZOMBIE.  We hold
 * read_lock(&tasklist_lock) on entry.  If we return zero, we still hold
 * the lock and this task is uninteresting.  If we return nonzero, we have
 * released the lock and the system call should return.
 */
1126
static int wait_task_zombie(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1127 1128
{
	unsigned long state;
1129
	int retval, status, traced;
1130
	pid_t pid = task_pid_vnr(p);
1131
	uid_t uid = from_kuid_munged(current_user_ns(), task_uid(p));
1132
	struct siginfo __user *infop;
L
Linus Torvalds 已提交
1133

1134
	if (!likely(wo->wo_flags & WEXITED))
R
Roland McGrath 已提交
1135 1136
		return 0;

1137
	if (unlikely(wo->wo_flags & WNOWAIT)) {
L
Linus Torvalds 已提交
1138
		int exit_code = p->exit_code;
1139
		int why;
L
Linus Torvalds 已提交
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149

		get_task_struct(p);
		read_unlock(&tasklist_lock);
		if ((exit_code & 0x7f) == 0) {
			why = CLD_EXITED;
			status = exit_code >> 8;
		} else {
			why = (exit_code & 0x80) ? CLD_DUMPED : CLD_KILLED;
			status = exit_code & 0x7f;
		}
1150
		return wait_noreap_copyout(wo, p, pid, uid, why, status);
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
	}

	/*
	 * Try to move the task's state to DEAD
	 * only one thread is allowed to do this:
	 */
	state = xchg(&p->exit_state, EXIT_DEAD);
	if (state != EXIT_ZOMBIE) {
		BUG_ON(state != EXIT_DEAD);
		return 0;
	}

1163
	traced = ptrace_reparented(p);
1164 1165
	/*
	 * It can be ptraced but not reparented, check
O
Oleg Nesterov 已提交
1166
	 * thread_group_leader() to filter out sub-threads.
1167
	 */
O
Oleg Nesterov 已提交
1168
	if (likely(!traced) && thread_group_leader(p)) {
1169 1170
		struct signal_struct *psig;
		struct signal_struct *sig;
J
Jiri Pirko 已提交
1171
		unsigned long maxrss;
1172
		cputime_t tgutime, tgstime;
1173

L
Linus Torvalds 已提交
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
		/*
		 * The resource counters for the group leader are in its
		 * own task_struct.  Those for dead threads in the group
		 * are in its signal_struct, as are those for the child
		 * processes it has previously reaped.  All these
		 * accumulate in the parent's signal_struct c* fields.
		 *
		 * We don't bother to take a lock here to protect these
		 * p->signal fields, because they are only touched by
		 * __exit_signal, which runs with tasklist_lock
		 * write-locked anyway, and so is excluded here.  We do
1185
		 * need to protect the access to parent->signal fields,
L
Linus Torvalds 已提交
1186 1187
		 * as other threads in the parent group can be right
		 * here reaping other children at the same time.
1188
		 *
1189
		 * We use thread_group_cputime_adjusted() to get times for the thread
1190 1191
		 * group, which consolidates times for all threads in the
		 * group including the group leader.
L
Linus Torvalds 已提交
1192
		 */
1193
		thread_group_cputime_adjusted(p, &tgutime, &tgstime);
1194 1195
		spin_lock_irq(&p->real_parent->sighand->siglock);
		psig = p->real_parent->signal;
1196
		sig = p->signal;
1197 1198 1199
		psig->cutime += tgutime + sig->cutime;
		psig->cstime += tgstime + sig->cstime;
		psig->cgtime += p->gtime + sig->gtime + sig->cgtime;
1200 1201 1202 1203 1204 1205 1206 1207
		psig->cmin_flt +=
			p->min_flt + sig->min_flt + sig->cmin_flt;
		psig->cmaj_flt +=
			p->maj_flt + sig->maj_flt + sig->cmaj_flt;
		psig->cnvcsw +=
			p->nvcsw + sig->nvcsw + sig->cnvcsw;
		psig->cnivcsw +=
			p->nivcsw + sig->nivcsw + sig->cnivcsw;
1208 1209 1210 1211 1212 1213
		psig->cinblock +=
			task_io_get_inblock(p) +
			sig->inblock + sig->cinblock;
		psig->coublock +=
			task_io_get_oublock(p) +
			sig->oublock + sig->coublock;
J
Jiri Pirko 已提交
1214 1215 1216
		maxrss = max(sig->maxrss, sig->cmaxrss);
		if (psig->cmaxrss < maxrss)
			psig->cmaxrss = maxrss;
1217 1218
		task_io_accounting_add(&psig->ioac, &p->ioac);
		task_io_accounting_add(&psig->ioac, &sig->ioac);
1219
		spin_unlock_irq(&p->real_parent->sighand->siglock);
L
Linus Torvalds 已提交
1220 1221 1222 1223 1224 1225 1226 1227
	}

	/*
	 * Now we are sure this task is interesting, and no other
	 * thread can reap it because we set its state to EXIT_DEAD.
	 */
	read_unlock(&tasklist_lock);

1228 1229
	retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
L
Linus Torvalds 已提交
1230 1231
	status = (p->signal->flags & SIGNAL_GROUP_EXIT)
		? p->signal->group_exit_code : p->exit_code;
1232 1233 1234 1235
	if (!retval && wo->wo_stat)
		retval = put_user(status, wo->wo_stat);

	infop = wo->wo_info;
L
Linus Torvalds 已提交
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
	if (!retval && infop)
		retval = put_user(SIGCHLD, &infop->si_signo);
	if (!retval && infop)
		retval = put_user(0, &infop->si_errno);
	if (!retval && infop) {
		int why;

		if ((status & 0x7f) == 0) {
			why = CLD_EXITED;
			status >>= 8;
		} else {
			why = (status & 0x80) ? CLD_DUMPED : CLD_KILLED;
			status &= 0x7f;
		}
		retval = put_user((short)why, &infop->si_code);
		if (!retval)
			retval = put_user(status, &infop->si_status);
	}
	if (!retval && infop)
1255
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1256
	if (!retval && infop)
1257
		retval = put_user(uid, &infop->si_uid);
1258
	if (!retval)
1259
		retval = pid;
1260 1261

	if (traced) {
L
Linus Torvalds 已提交
1262
		write_lock_irq(&tasklist_lock);
1263 1264 1265
		/* We dropped tasklist, ptracer could die and untrace */
		ptrace_unlink(p);
		/*
1266 1267
		 * If this is not a sub-thread, notify the parent.
		 * If parent wants a zombie, don't release it now.
1268
		 */
1269 1270 1271 1272
		if (thread_group_leader(p) &&
		    !do_notify_parent(p, p->exit_signal)) {
			p->exit_state = EXIT_ZOMBIE;
			p = NULL;
L
Linus Torvalds 已提交
1273 1274 1275 1276 1277
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1278

L
Linus Torvalds 已提交
1279 1280 1281
	return retval;
}

1282 1283 1284
static int *task_stopped_code(struct task_struct *p, bool ptrace)
{
	if (ptrace) {
T
Tejun Heo 已提交
1285 1286
		if (task_is_stopped_or_traced(p) &&
		    !(p->jobctl & JOBCTL_LISTENING))
1287 1288 1289 1290 1291 1292 1293 1294
			return &p->exit_code;
	} else {
		if (p->signal->flags & SIGNAL_STOP_STOPPED)
			return &p->signal->group_exit_code;
	}
	return NULL;
}

1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311
/**
 * wait_task_stopped - Wait for %TASK_STOPPED or %TASK_TRACED
 * @wo: wait options
 * @ptrace: is the wait for ptrace
 * @p: task to wait for
 *
 * Handle sys_wait4() work for %p in state %TASK_STOPPED or %TASK_TRACED.
 *
 * CONTEXT:
 * read_lock(&tasklist_lock), which is released if return value is
 * non-zero.  Also, grabs and releases @p->sighand->siglock.
 *
 * RETURNS:
 * 0 if wait condition didn't exist and search for other wait conditions
 * should continue.  Non-zero return, -errno on failure and @p's pid on
 * success, implies that tasklist_lock is released and wait condition
 * search should terminate.
L
Linus Torvalds 已提交
1312
 */
1313 1314
static int wait_task_stopped(struct wait_opts *wo,
				int ptrace, struct task_struct *p)
L
Linus Torvalds 已提交
1315
{
1316
	struct siginfo __user *infop;
1317
	int retval, exit_code, *p_code, why;
1318
	uid_t uid = 0; /* unneeded, required by compiler */
1319
	pid_t pid;
L
Linus Torvalds 已提交
1320

1321 1322 1323
	/*
	 * Traditionally we see ptrace'd stopped tasks regardless of options.
	 */
1324
	if (!ptrace && !(wo->wo_flags & WUNTRACED))
R
Roland McGrath 已提交
1325 1326
		return 0;

1327 1328 1329
	if (!task_stopped_code(p, ptrace))
		return 0;

1330 1331 1332
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

1333 1334
	p_code = task_stopped_code(p, ptrace);
	if (unlikely(!p_code))
1335 1336
		goto unlock_sig;

1337
	exit_code = *p_code;
1338 1339 1340
	if (!exit_code)
		goto unlock_sig;

1341
	if (!unlikely(wo->wo_flags & WNOWAIT))
1342
		*p_code = 0;
1343

1344
	uid = from_kuid_munged(current_user_ns(), task_uid(p));
1345 1346 1347
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1348 1349 1350 1351 1352 1353 1354 1355 1356 1357
		return 0;

	/*
	 * Now we are pretty sure this task is interesting.
	 * Make sure it doesn't get reaped out from under us while we
	 * give up the lock and then examine it below.  We don't want to
	 * keep holding onto the tasklist_lock while we call getrusage and
	 * possibly take page faults for user memory.
	 */
	get_task_struct(p);
1358
	pid = task_pid_vnr(p);
R
Roland McGrath 已提交
1359
	why = ptrace ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1360 1361
	read_unlock(&tasklist_lock);

1362 1363 1364 1365 1366 1367 1368
	if (unlikely(wo->wo_flags & WNOWAIT))
		return wait_noreap_copyout(wo, p, pid, uid, why, exit_code);

	retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
	if (!retval && wo->wo_stat)
		retval = put_user((exit_code << 8) | 0x7f, wo->wo_stat);
L
Linus Torvalds 已提交
1369

1370
	infop = wo->wo_info;
L
Linus Torvalds 已提交
1371 1372 1373 1374 1375
	if (!retval && infop)
		retval = put_user(SIGCHLD, &infop->si_signo);
	if (!retval && infop)
		retval = put_user(0, &infop->si_errno);
	if (!retval && infop)
R
Roland McGrath 已提交
1376
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1377 1378 1379
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1380
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1381
	if (!retval && infop)
1382
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1383
	if (!retval)
1384
		retval = pid;
L
Linus Torvalds 已提交
1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
	put_task_struct(p);

	BUG_ON(!retval);
	return retval;
}

/*
 * Handle do_wait work for one task in a live, non-stopped state.
 * read_lock(&tasklist_lock) on entry.  If we return zero, we still hold
 * the lock and this task is uninteresting.  If we return nonzero, we have
 * released the lock and the system call should return.
 */
1397
static int wait_task_continued(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1398 1399 1400 1401 1402
{
	int retval;
	pid_t pid;
	uid_t uid;

1403
	if (!unlikely(wo->wo_flags & WCONTINUED))
R
Roland McGrath 已提交
1404 1405
		return 0;

L
Linus Torvalds 已提交
1406 1407 1408 1409 1410 1411 1412 1413 1414
	if (!(p->signal->flags & SIGNAL_STOP_CONTINUED))
		return 0;

	spin_lock_irq(&p->sighand->siglock);
	/* Re-check with the lock held.  */
	if (!(p->signal->flags & SIGNAL_STOP_CONTINUED)) {
		spin_unlock_irq(&p->sighand->siglock);
		return 0;
	}
1415
	if (!unlikely(wo->wo_flags & WNOWAIT))
L
Linus Torvalds 已提交
1416
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
1417
	uid = from_kuid_munged(current_user_ns(), task_uid(p));
L
Linus Torvalds 已提交
1418 1419
	spin_unlock_irq(&p->sighand->siglock);

1420
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1421 1422 1423
	get_task_struct(p);
	read_unlock(&tasklist_lock);

1424 1425 1426
	if (!wo->wo_info) {
		retval = wo->wo_rusage
			? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
L
Linus Torvalds 已提交
1427
		put_task_struct(p);
1428 1429
		if (!retval && wo->wo_stat)
			retval = put_user(0xffff, wo->wo_stat);
L
Linus Torvalds 已提交
1430
		if (!retval)
1431
			retval = pid;
L
Linus Torvalds 已提交
1432
	} else {
1433 1434
		retval = wait_noreap_copyout(wo, p, pid, uid,
					     CLD_CONTINUED, SIGCONT);
L
Linus Torvalds 已提交
1435 1436 1437 1438 1439 1440
		BUG_ON(retval == 0);
	}

	return retval;
}

R
Roland McGrath 已提交
1441 1442 1443
/*
 * Consider @p for a wait by @parent.
 *
1444
 * -ECHILD should be in ->notask_error before the first call.
R
Roland McGrath 已提交
1445 1446
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue;
1447
 * then ->notask_error is 0 if @p is an eligible child,
1448
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1449
 */
1450 1451
static int wait_consider_task(struct wait_opts *wo, int ptrace,
				struct task_struct *p)
R
Roland McGrath 已提交
1452
{
1453
	int ret = eligible_child(wo, p);
1454
	if (!ret)
R
Roland McGrath 已提交
1455 1456
		return ret;

1457
	ret = security_task_wait(p);
1458 1459 1460 1461 1462 1463 1464 1465
	if (unlikely(ret < 0)) {
		/*
		 * If we have not yet seen any eligible child,
		 * then let this error code replace -ECHILD.
		 * A permission error will give the user a clue
		 * to look for security policy problems, rather
		 * than for mysterious wait bugs.
		 */
1466 1467
		if (wo->notask_error)
			wo->notask_error = ret;
1468
		return 0;
1469 1470
	}

1471
	/* dead body doesn't have much to contribute */
1472 1473 1474 1475 1476 1477 1478
	if (unlikely(p->exit_state == EXIT_DEAD)) {
		/*
		 * But do not ignore this task until the tracer does
		 * wait_task_zombie()->do_notify_parent().
		 */
		if (likely(!ptrace) && unlikely(ptrace_reparented(p)))
			wo->notask_error = 0;
1479
		return 0;
1480
	}
1481

1482 1483
	/* slay zombie? */
	if (p->exit_state == EXIT_ZOMBIE) {
R
Roland McGrath 已提交
1484
		/*
1485 1486 1487
		 * A zombie ptracee is only visible to its ptracer.
		 * Notification and reaping will be cascaded to the real
		 * parent when the ptracer detaches.
R
Roland McGrath 已提交
1488
		 */
T
Tejun Heo 已提交
1489
		if (likely(!ptrace) && unlikely(p->ptrace)) {
1490 1491 1492 1493
			/* it will become visible, clear notask_error */
			wo->notask_error = 0;
			return 0;
		}
R
Roland McGrath 已提交
1494

1495 1496 1497
		/* we don't reap group leaders with subthreads */
		if (!delay_group_leader(p))
			return wait_task_zombie(wo, p);
R
Roland McGrath 已提交
1498

R
Roland McGrath 已提交
1499
		/*
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521
		 * Allow access to stopped/continued state via zombie by
		 * falling through.  Clearing of notask_error is complex.
		 *
		 * When !@ptrace:
		 *
		 * If WEXITED is set, notask_error should naturally be
		 * cleared.  If not, subset of WSTOPPED|WCONTINUED is set,
		 * so, if there are live subthreads, there are events to
		 * wait for.  If all subthreads are dead, it's still safe
		 * to clear - this function will be called again in finite
		 * amount time once all the subthreads are released and
		 * will then return without clearing.
		 *
		 * When @ptrace:
		 *
		 * Stopped state is per-task and thus can't change once the
		 * target task dies.  Only continued and exited can happen.
		 * Clear notask_error if WCONTINUED | WEXITED.
		 */
		if (likely(!ptrace) || (wo->wo_flags & (WCONTINUED | WEXITED)))
			wo->notask_error = 0;
	} else {
1522 1523 1524 1525 1526 1527 1528 1529 1530 1531
		/*
		 * If @p is ptraced by a task in its real parent's group,
		 * hide group stop/continued state when looking at @p as
		 * the real parent; otherwise, a single stop can be
		 * reported twice as group and ptrace stops.
		 *
		 * If a ptracer wants to distinguish the two events for its
		 * own children, it should create a separate process which
		 * takes the role of real parent.
		 */
1532
		if (likely(!ptrace) && p->ptrace && !ptrace_reparented(p))
1533 1534
			return 0;

1535 1536 1537
		/*
		 * @p is alive and it's gonna stop, continue or exit, so
		 * there always is something to wait for.
R
Roland McGrath 已提交
1538
		 */
1539
		wo->notask_error = 0;
R
Roland McGrath 已提交
1540 1541
	}

R
Roland McGrath 已提交
1542
	/*
1543 1544
	 * Wait for stopped.  Depending on @ptrace, different stopped state
	 * is used and the two don't interact with each other.
R
Roland McGrath 已提交
1545
	 */
1546 1547 1548
	ret = wait_task_stopped(wo, ptrace, p);
	if (ret)
		return ret;
R
Roland McGrath 已提交
1549 1550

	/*
1551 1552 1553
	 * Wait for continued.  There's only one continued state and the
	 * ptracer can consume it which can confuse the real parent.  Don't
	 * use WCONTINUED from ptracer.  You don't need or want it.
R
Roland McGrath 已提交
1554
	 */
1555
	return wait_task_continued(wo, p);
R
Roland McGrath 已提交
1556 1557 1558 1559 1560
}

/*
 * Do the work of do_wait() for one thread in the group, @tsk.
 *
1561
 * -ECHILD should be in ->notask_error before the first call.
R
Roland McGrath 已提交
1562 1563
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue; then
1564
 * ->notask_error is 0 if there were any eligible children,
1565
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1566
 */
1567
static int do_wait_thread(struct wait_opts *wo, struct task_struct *tsk)
R
Roland McGrath 已提交
1568 1569 1570 1571
{
	struct task_struct *p;

	list_for_each_entry(p, &tsk->children, sibling) {
1572 1573 1574
		int ret = wait_consider_task(wo, 0, p);
		if (ret)
			return ret;
R
Roland McGrath 已提交
1575 1576 1577 1578 1579
	}

	return 0;
}

1580
static int ptrace_do_wait(struct wait_opts *wo, struct task_struct *tsk)
R
Roland McGrath 已提交
1581 1582 1583
{
	struct task_struct *p;

R
Roland McGrath 已提交
1584
	list_for_each_entry(p, &tsk->ptraced, ptrace_entry) {
1585
		int ret = wait_consider_task(wo, 1, p);
R
Roland McGrath 已提交
1586
		if (ret)
R
Roland McGrath 已提交
1587 1588 1589 1590 1591 1592
			return ret;
	}

	return 0;
}

1593 1594 1595 1596 1597 1598 1599
static int child_wait_callback(wait_queue_t *wait, unsigned mode,
				int sync, void *key)
{
	struct wait_opts *wo = container_of(wait, struct wait_opts,
						child_wait);
	struct task_struct *p = key;

1600
	if (!eligible_pid(wo, p))
1601 1602
		return 0;

1603 1604 1605
	if ((wo->wo_flags & __WNOTHREAD) && wait->private != p->parent)
		return 0;

1606 1607 1608
	return default_wake_function(wait, mode, sync, key);
}

1609 1610
void __wake_up_parent(struct task_struct *p, struct task_struct *parent)
{
1611 1612
	__wake_up_sync_key(&parent->signal->wait_chldexit,
				TASK_INTERRUPTIBLE, 1, p);
1613 1614
}

1615
static long do_wait(struct wait_opts *wo)
L
Linus Torvalds 已提交
1616 1617
{
	struct task_struct *tsk;
R
Roland McGrath 已提交
1618
	int retval;
L
Linus Torvalds 已提交
1619

1620
	trace_sched_process_wait(wo->wo_pid);
1621

1622 1623 1624
	init_waitqueue_func_entry(&wo->child_wait, child_wait_callback);
	wo->child_wait.private = current;
	add_wait_queue(&current->signal->wait_chldexit, &wo->child_wait);
L
Linus Torvalds 已提交
1625
repeat:
R
Roland McGrath 已提交
1626 1627
	/*
	 * If there is nothing that can match our critiera just get out.
1628 1629 1630
	 * We will clear ->notask_error to zero if we see any child that
	 * might later match our criteria, even if we are not able to reap
	 * it yet.
R
Roland McGrath 已提交
1631
	 */
1632
	wo->notask_error = -ECHILD;
1633 1634
	if ((wo->wo_type < PIDTYPE_MAX) &&
	   (!wo->wo_pid || hlist_empty(&wo->wo_pid->tasks[wo->wo_type])))
1635
		goto notask;
E
Eric W. Biederman 已提交
1636

1637
	set_current_state(TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
1638 1639 1640
	read_lock(&tasklist_lock);
	tsk = current;
	do {
1641 1642 1643
		retval = do_wait_thread(wo, tsk);
		if (retval)
			goto end;
1644

1645 1646
		retval = ptrace_do_wait(wo, tsk);
		if (retval)
R
Roland McGrath 已提交
1647 1648
			goto end;

1649
		if (wo->wo_flags & __WNOTHREAD)
L
Linus Torvalds 已提交
1650
			break;
1651
	} while_each_thread(current, tsk);
L
Linus Torvalds 已提交
1652
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1653

1654
notask:
1655 1656
	retval = wo->notask_error;
	if (!retval && !(wo->wo_flags & WNOHANG)) {
L
Linus Torvalds 已提交
1657
		retval = -ERESTARTSYS;
R
Roland McGrath 已提交
1658 1659 1660 1661
		if (!signal_pending(current)) {
			schedule();
			goto repeat;
		}
L
Linus Torvalds 已提交
1662 1663
	}
end:
1664
	__set_current_state(TASK_RUNNING);
1665
	remove_wait_queue(&current->signal->wait_chldexit, &wo->child_wait);
L
Linus Torvalds 已提交
1666 1667 1668
	return retval;
}

1669 1670
SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *,
		infop, int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1671
{
1672
	struct wait_opts wo;
E
Eric W. Biederman 已提交
1673 1674
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1675 1676 1677 1678 1679 1680 1681 1682 1683
	long ret;

	if (options & ~(WNOHANG|WNOWAIT|WEXITED|WSTOPPED|WCONTINUED))
		return -EINVAL;
	if (!(options & (WEXITED|WSTOPPED|WCONTINUED)))
		return -EINVAL;

	switch (which) {
	case P_ALL:
E
Eric W. Biederman 已提交
1684
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1685 1686
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1687 1688
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1689 1690 1691
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1692 1693
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1694 1695 1696 1697 1698 1699
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1700 1701
	if (type < PIDTYPE_MAX)
		pid = find_get_pid(upid);
1702 1703 1704 1705 1706 1707 1708 1709

	wo.wo_type	= type;
	wo.wo_pid	= pid;
	wo.wo_flags	= options;
	wo.wo_info	= infop;
	wo.wo_stat	= NULL;
	wo.wo_rusage	= ru;
	ret = do_wait(&wo);
1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732

	if (ret > 0) {
		ret = 0;
	} else if (infop) {
		/*
		 * For a WNOHANG return, clear out all the fields
		 * we would set so the user can easily tell the
		 * difference.
		 */
		if (!ret)
			ret = put_user(0, &infop->si_signo);
		if (!ret)
			ret = put_user(0, &infop->si_errno);
		if (!ret)
			ret = put_user(0, &infop->si_code);
		if (!ret)
			ret = put_user(0, &infop->si_pid);
		if (!ret)
			ret = put_user(0, &infop->si_uid);
		if (!ret)
			ret = put_user(0, &infop->si_status);
	}

E
Eric W. Biederman 已提交
1733
	put_pid(pid);
L
Linus Torvalds 已提交
1734 1735

	/* avoid REGPARM breakage on x86: */
1736
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1737 1738 1739
	return ret;
}

1740 1741
SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr,
		int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1742
{
1743
	struct wait_opts wo;
E
Eric W. Biederman 已提交
1744 1745
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1746 1747 1748 1749 1750
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1751 1752 1753 1754 1755 1756 1757 1758

	if (upid == -1)
		type = PIDTYPE_MAX;
	else if (upid < 0) {
		type = PIDTYPE_PGID;
		pid = find_get_pid(-upid);
	} else if (upid == 0) {
		type = PIDTYPE_PGID;
1759
		pid = get_task_pid(current, PIDTYPE_PGID);
E
Eric W. Biederman 已提交
1760 1761 1762 1763 1764
	} else /* upid > 0 */ {
		type = PIDTYPE_PID;
		pid = find_get_pid(upid);
	}

1765 1766 1767 1768 1769 1770 1771
	wo.wo_type	= type;
	wo.wo_pid	= pid;
	wo.wo_flags	= options | WEXITED;
	wo.wo_info	= NULL;
	wo.wo_stat	= stat_addr;
	wo.wo_rusage	= ru;
	ret = do_wait(&wo);
E
Eric W. Biederman 已提交
1772
	put_pid(pid);
L
Linus Torvalds 已提交
1773 1774

	/* avoid REGPARM breakage on x86: */
1775
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1776 1777 1778 1779 1780 1781 1782 1783 1784
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
1785
SYSCALL_DEFINE3(waitpid, pid_t, pid, int __user *, stat_addr, int, options)
L
Linus Torvalds 已提交
1786 1787 1788 1789 1790
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif