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();
460
	exit_files(current);
L
Linus Torvalds 已提交
461 462 463
	current->files = init_task.files;
	atomic_inc(&current->files->count);

464
	reparent_to_kthreadd();
L
Linus Torvalds 已提交
465 466 467 468
}

EXPORT_SYMBOL(daemonize);

469 470
#ifdef CONFIG_MM_OWNER
/*
471
 * A task is exiting.   If it owned this mm, find a new owner for the mm.
472 473 474 475 476 477
 */
void mm_update_next_owner(struct mm_struct *mm)
{
	struct task_struct *c, *g, *p = current;

retry:
478 479 480 481 482
	/*
	 * If the exiting or execing task is not the owner, it's
	 * someone else's problem.
	 */
	if (mm->owner != p)
483
		return;
484 485 486 487 488 489 490 491 492
	/*
	 * 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;
	}
493 494 495 496 497 498 499 500 501 502 503 504 505

	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
	 */
506
	list_for_each_entry(c, &p->real_parent->children, sibling) {
507 508 509 510 511 512 513 514 515 516 517 518 519 520
		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);
521 522 523
	/*
	 * We found no owner yet mm_users > 1: this implies that we are
	 * most likely racing with swapoff (try_to_unuse()) or /proc or
524
	 * ptrace or page migration (get_task_mm()).  Mark owner as NULL.
525 526
	 */
	mm->owner = NULL;
527 528 529 530 531 532 533 534 535 536
	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);
537 538 539 540 541
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
542 543 544 545 546 547 548 549 550 551 552
	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
553 554 555 556
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
557
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
558 559
{
	struct mm_struct *mm = tsk->mm;
560
	struct core_state *core_state;
L
Linus Torvalds 已提交
561

562
	mm_release(tsk, mm);
L
Linus Torvalds 已提交
563 564
	if (!mm)
		return;
565
	sync_mm_rss(mm);
L
Linus Torvalds 已提交
566 567
	/*
	 * Serialize with any possible pending coredump.
568
	 * We must hold mmap_sem around checking core_state
L
Linus Torvalds 已提交
569
	 * and clearing tsk->mm.  The core-inducing thread
570
	 * will increment ->nr_threads for each thread in the
L
Linus Torvalds 已提交
571 572 573
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
574 575 576
	core_state = mm->core_state;
	if (core_state) {
		struct core_thread self;
L
Linus Torvalds 已提交
577 578
		up_read(&mm->mmap_sem);

579 580 581 582 583 584 585 586
		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
587

O
Oleg Nesterov 已提交
588 589 590 591 592 593 594
		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
595 596 597
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
598
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
599 600 601 602 603 604
	/* 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);
605
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
606 607 608 609
	mmput(mm);
}

/*
610 611 612 613 614
 * 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
615
 */
616
static struct task_struct *find_new_reaper(struct task_struct *father)
617 618
	__releases(&tasklist_lock)
	__acquires(&tasklist_lock)
L
Linus Torvalds 已提交
619
{
620 621
	struct pid_namespace *pid_ns = task_active_pid_ns(father);
	struct task_struct *thread;
L
Linus Torvalds 已提交
622

623 624 625 626 627 628 629 630
	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
631

632 633
	if (unlikely(pid_ns->child_reaper == father)) {
		write_unlock_irq(&tasklist_lock);
634 635 636 637 638
		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
639

640 641
		zap_pid_ns_processes(pid_ns);
		write_lock_irq(&tasklist_lock);
642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664
	} 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
665
	}
666

667 668 669
	return pid_ns->child_reaper;
}

670 671 672
/*
* Any that need to be release_task'd are put on the @dead list.
 */
673
static void reparent_leader(struct task_struct *father, struct task_struct *p,
674 675 676 677
				struct list_head *dead)
{
	list_move_tail(&p->sibling, &p->real_parent->children);

678
	if (p->exit_state == EXIT_DEAD)
679 680 681 682 683 684 685 686 687 688 689 690
		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
691
	if (!p->ptrace &&
692
	    p->exit_state == EXIT_ZOMBIE && thread_group_empty(p)) {
693
		if (do_notify_parent(p, p->exit_signal)) {
694 695 696 697 698 699 700 701
			p->exit_state = EXIT_DEAD;
			list_move_tail(&p->sibling, dead);
		}
	}

	kill_orphaned_pgrp(p, father);
}

702
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
703
{
704
	struct task_struct *p, *n, *reaper;
705
	LIST_HEAD(dead_children);
706 707

	write_lock_irq(&tasklist_lock);
708 709 710 711 712
	/*
	 * Note that exit_ptrace() and find_new_reaper() might
	 * drop tasklist_lock and reacquire it.
	 */
	exit_ptrace(father);
713
	reaper = find_new_reaper(father);
R
Roland McGrath 已提交
714

715
	list_for_each_entry_safe(p, n, &father->children, sibling) {
716 717 718 719
		struct task_struct *t = p;
		do {
			t->real_parent = reaper;
			if (t->parent == father) {
T
Tejun Heo 已提交
720
				BUG_ON(t->ptrace);
721 722 723 724 725 726 727
				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
728
	}
729
	write_unlock_irq(&tasklist_lock);
730

731 732
	BUG_ON(!list_empty(&father->children));

733 734
	list_for_each_entry_safe(p, n, &dead_children, sibling) {
		list_del_init(&p->sibling);
735 736
		release_task(p);
	}
L
Linus Torvalds 已提交
737 738 739 740 741 742
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
743
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
744
{
745
	bool autoreap;
L
Linus Torvalds 已提交
746 747 748 749 750 751 752 753 754

	/*
	 * 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)
	 */
755
	forget_original_parent(tsk);
756
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
757

758
	write_lock_irq(&tasklist_lock);
759 760
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
761

O
Oleg Nesterov 已提交
762 763 764 765 766 767 768 769 770 771 772 773
	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
774

775
	tsk->exit_state = autoreap ? EXIT_DEAD : EXIT_ZOMBIE;
L
Linus Torvalds 已提交
776

777 778
	/* mt-exec, de_thread() is waiting for group leader */
	if (unlikely(tsk->signal->notify_count < 0))
779
		wake_up_process(tsk->signal->group_exit_task);
L
Linus Torvalds 已提交
780 781 782
	write_unlock_irq(&tasklist_lock);

	/* If the process is dead, release it - nobody will wait for it */
783
	if (autoreap)
L
Linus Torvalds 已提交
784 785 786
		release_task(tsk);
}

787 788 789 790 791 792 793
#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;

794
	free = stack_not_used(current);
795 796 797 798 799 800

	if (free >= lowest_to_date)
		return;

	spin_lock(&low_water_lock);
	if (free < lowest_to_date) {
801 802 803
		printk(KERN_WARNING "%s (%d) used greatest stack depth: "
				"%lu bytes left\n",
				current->comm, task_pid_nr(current), free);
804 805 806 807 808 809 810 811
		lowest_to_date = free;
	}
	spin_unlock(&low_water_lock);
}
#else
static inline void check_stack_usage(void) {}
#endif

812
void do_exit(long code)
L
Linus Torvalds 已提交
813 814 815 816 817 818
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

819
	WARN_ON(blk_needs_flush_plug(tsk));
820

L
Linus Torvalds 已提交
821 822 823 824 825
	if (unlikely(in_interrupt()))
		panic("Aiee, killing interrupt handler!");
	if (unlikely(!tsk->pid))
		panic("Attempted to kill the idle task!");

826 827 828 829 830 831 832 833 834
	/*
	 * 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
835
	ptrace_event(PTRACE_EVENT_EXIT, code);
L
Linus Torvalds 已提交
836

837 838
	validate_creds_for_do_exit(tsk);

839 840 841 842 843 844 845
	/*
	 * 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");
846 847 848 849 850 851 852 853 854 855
		/*
		 * 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;
856 857 858 859
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

O
Oleg Nesterov 已提交
860
	exit_signals(tsk);  /* sets PF_EXITING */
861 862
	/*
	 * tsk->flags are checked in the futex code to protect against
863
	 * an exiting task cleaning up the robust pi futexes.
864
	 */
865
	smp_mb();
866
	raw_spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
867 868 869

	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
870
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
871 872 873
				preempt_count());

	acct_update_integrals(tsk);
874 875 876
	/* sync mm's RSS info before statistics gathering */
	if (tsk->mm)
		sync_mm_rss(tsk->mm);
L
Linus Torvalds 已提交
877
	group_dead = atomic_dec_and_test(&tsk->signal->live);
878
	if (group_dead) {
879
		hrtimer_cancel(&tsk->signal->real_timer);
880
		exit_itimers(tsk->signal);
J
Jiri Pirko 已提交
881 882
		if (tsk->mm)
			setmax_mm_hiwater_rss(&tsk->signal->maxrss, tsk->mm);
883
	}
884
	acct_collect(code, group_dead);
M
Miloslav Trmac 已提交
885 886
	if (group_dead)
		tty_audit_exit();
887
	audit_free(tsk);
888

889
	tsk->exit_code = code;
890
	taskstats_exit(tsk, group_dead);
891

L
Linus Torvalds 已提交
892 893
	exit_mm(tsk);

894
	if (group_dead)
895
		acct_process();
896 897
	trace_sched_process_exit(tsk);

L
Linus Torvalds 已提交
898
	exit_sem(tsk);
899
	exit_shm(tsk);
900 901
	exit_files(tsk);
	exit_fs(tsk);
902
	exit_task_work(tsk);
903
	check_stack_usage();
L
Linus Torvalds 已提交
904
	exit_thread();
905 906 907 908 909 910 911 912 913

	/*
	 * 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);

914
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
915

916
	if (group_dead)
L
Linus Torvalds 已提交
917 918
		disassociate_ctty(1);

A
Al Viro 已提交
919
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
920

M
Matt Helsley 已提交
921
	proc_exit_connector(tsk);
922

923 924 925
	/*
	 * FIXME: do that only when needed, using sched_exit tracepoint
	 */
926
	ptrace_put_breakpoints(tsk);
927

928
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
929
#ifdef CONFIG_NUMA
930
	task_lock(tsk);
931
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
932
	tsk->mempolicy = NULL;
933
	task_unlock(tsk);
L
Linus Torvalds 已提交
934
#endif
935
#ifdef CONFIG_FUTEX
936 937
	if (unlikely(current->pi_state_cache))
		kfree(current->pi_state_cache);
938
#endif
939
	/*
940
	 * Make sure we are holding no locks:
941
	 */
942
	debug_check_no_locks_held(tsk);
943 944 945 946 947 948
	/*
	 * 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
949

950
	if (tsk->io_context)
951
		exit_io_context(tsk);
952

953 954 955
	if (tsk->splice_pipe)
		__free_pipe_info(tsk->splice_pipe);

956 957
	validate_creds_for_do_exit(tsk);

C
Coywolf Qi Hunt 已提交
958
	preempt_disable();
959 960
	if (tsk->nr_dirtied)
		__this_cpu_add(dirty_throttle_leaks, tsk->nr_dirtied);
961
	exit_rcu();
962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977

	/*
	 * 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
978
	/* causes final put_task_struct in finish_task_switch(). */
979
	tsk->state = TASK_DEAD;
980
	tsk->flags |= PF_NOFREEZE;	/* tell freezer to ignore us */
L
Linus Torvalds 已提交
981 982 983
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
984 985
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
986 987
}

988 989
EXPORT_SYMBOL_GPL(do_exit);

990
void complete_and_exit(struct completion *comp, long code)
L
Linus Torvalds 已提交
991 992 993
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
994

L
Linus Torvalds 已提交
995 996 997 998 999
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

1000
SYSCALL_DEFINE1(exit, int, error_code)
L
Linus Torvalds 已提交
1001 1002 1003 1004 1005 1006 1007 1008
{
	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).
 */
1009
void
L
Linus Torvalds 已提交
1010 1011
do_group_exit(int exit_code)
{
1012 1013
	struct signal_struct *sig = current->signal;

L
Linus Torvalds 已提交
1014 1015
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

1016 1017
	if (signal_group_exit(sig))
		exit_code = sig->group_exit_code;
L
Linus Torvalds 已提交
1018 1019 1020
	else if (!thread_group_empty(current)) {
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1021
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1022 1023 1024 1025
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1026
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040
			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.
 */
1041
SYSCALL_DEFINE1(exit_group, int, error_code)
L
Linus Torvalds 已提交
1042 1043
{
	do_group_exit((error_code & 0xff) << 8);
1044 1045
	/* NOTREACHED */
	return 0;
L
Linus Torvalds 已提交
1046 1047
}

1048 1049 1050
struct wait_opts {
	enum pid_type		wo_type;
	int			wo_flags;
1051
	struct pid		*wo_pid;
1052 1053 1054 1055 1056

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

1057
	wait_queue_t		child_wait;
1058 1059 1060
	int			notask_error;
};

1061 1062
static inline
struct pid *task_pid_type(struct task_struct *task, enum pid_type type)
E
Eric W. Biederman 已提交
1063
{
1064 1065 1066
	if (type != PIDTYPE_PID)
		task = task->group_leader;
	return task->pids[type].pid;
E
Eric W. Biederman 已提交
1067 1068
}

1069
static int eligible_pid(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1070
{
1071 1072 1073
	return	wo->wo_type == PIDTYPE_MAX ||
		task_pid_type(p, wo->wo_type) == wo->wo_pid;
}
L
Linus Torvalds 已提交
1074

1075 1076 1077 1078
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
1079 1080 1081 1082 1083
	/* 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.) */
1084 1085
	if (((p->exit_signal != SIGCHLD) ^ !!(wo->wo_flags & __WCLONE))
	    && !(wo->wo_flags & __WALL))
L
Linus Torvalds 已提交
1086 1087
		return 0;

1088
	return 1;
L
Linus Torvalds 已提交
1089 1090
}

1091 1092
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
1093
{
1094 1095 1096
	struct siginfo __user *infop;
	int retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
1097

L
Linus Torvalds 已提交
1098
	put_task_struct(p);
1099
	infop = wo->wo_info;
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
	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
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124
	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.
 */
1125
static int wait_task_zombie(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1126 1127
{
	unsigned long state;
1128
	int retval, status, traced;
1129
	pid_t pid = task_pid_vnr(p);
1130
	uid_t uid = from_kuid_munged(current_user_ns(), task_uid(p));
1131
	struct siginfo __user *infop;
L
Linus Torvalds 已提交
1132

1133
	if (!likely(wo->wo_flags & WEXITED))
R
Roland McGrath 已提交
1134 1135
		return 0;

1136
	if (unlikely(wo->wo_flags & WNOWAIT)) {
L
Linus Torvalds 已提交
1137
		int exit_code = p->exit_code;
1138
		int why;
L
Linus Torvalds 已提交
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148

		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;
		}
1149
		return wait_noreap_copyout(wo, p, pid, uid, why, status);
L
Linus Torvalds 已提交
1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161
	}

	/*
	 * 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;
	}

1162
	traced = ptrace_reparented(p);
1163 1164
	/*
	 * It can be ptraced but not reparented, check
O
Oleg Nesterov 已提交
1165
	 * thread_group_leader() to filter out sub-threads.
1166
	 */
O
Oleg Nesterov 已提交
1167
	if (likely(!traced) && thread_group_leader(p)) {
1168 1169
		struct signal_struct *psig;
		struct signal_struct *sig;
J
Jiri Pirko 已提交
1170
		unsigned long maxrss;
1171
		cputime_t tgutime, tgstime;
1172

L
Linus Torvalds 已提交
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183
		/*
		 * 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
1184
		 * need to protect the access to parent->signal fields,
L
Linus Torvalds 已提交
1185 1186
		 * as other threads in the parent group can be right
		 * here reaping other children at the same time.
1187 1188 1189 1190
		 *
		 * We use thread_group_times() to get times for the thread
		 * group, which consolidates times for all threads in the
		 * group including the group leader.
L
Linus Torvalds 已提交
1191
		 */
1192
		thread_group_times(p, &tgutime, &tgstime);
1193 1194
		spin_lock_irq(&p->real_parent->sighand->siglock);
		psig = p->real_parent->signal;
1195
		sig = p->signal;
1196 1197 1198
		psig->cutime += tgutime + sig->cutime;
		psig->cstime += tgstime + sig->cstime;
		psig->cgtime += p->gtime + sig->gtime + sig->cgtime;
1199 1200 1201 1202 1203 1204 1205 1206
		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;
1207 1208 1209 1210 1211 1212
		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
1213 1214 1215
		maxrss = max(sig->maxrss, sig->cmaxrss);
		if (psig->cmaxrss < maxrss)
			psig->cmaxrss = maxrss;
1216 1217
		task_io_accounting_add(&psig->ioac, &p->ioac);
		task_io_accounting_add(&psig->ioac, &sig->ioac);
1218
		spin_unlock_irq(&p->real_parent->sighand->siglock);
L
Linus Torvalds 已提交
1219 1220 1221 1222 1223 1224 1225 1226
	}

	/*
	 * 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);

1227 1228
	retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
L
Linus Torvalds 已提交
1229 1230
	status = (p->signal->flags & SIGNAL_GROUP_EXIT)
		? p->signal->group_exit_code : p->exit_code;
1231 1232 1233 1234
	if (!retval && wo->wo_stat)
		retval = put_user(status, wo->wo_stat);

	infop = wo->wo_info;
L
Linus Torvalds 已提交
1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
	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)
1254
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1255
	if (!retval && infop)
1256
		retval = put_user(uid, &infop->si_uid);
1257
	if (!retval)
1258
		retval = pid;
1259 1260

	if (traced) {
L
Linus Torvalds 已提交
1261
		write_lock_irq(&tasklist_lock);
1262 1263 1264
		/* We dropped tasklist, ptracer could die and untrace */
		ptrace_unlink(p);
		/*
1265 1266
		 * If this is not a sub-thread, notify the parent.
		 * If parent wants a zombie, don't release it now.
1267
		 */
1268 1269 1270 1271
		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
1272 1273 1274 1275 1276
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1277

L
Linus Torvalds 已提交
1278 1279 1280
	return retval;
}

1281 1282 1283
static int *task_stopped_code(struct task_struct *p, bool ptrace)
{
	if (ptrace) {
T
Tejun Heo 已提交
1284 1285
		if (task_is_stopped_or_traced(p) &&
		    !(p->jobctl & JOBCTL_LISTENING))
1286 1287 1288 1289 1290 1291 1292 1293
			return &p->exit_code;
	} else {
		if (p->signal->flags & SIGNAL_STOP_STOPPED)
			return &p->signal->group_exit_code;
	}
	return NULL;
}

1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
/**
 * 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
1311
 */
1312 1313
static int wait_task_stopped(struct wait_opts *wo,
				int ptrace, struct task_struct *p)
L
Linus Torvalds 已提交
1314
{
1315
	struct siginfo __user *infop;
1316
	int retval, exit_code, *p_code, why;
1317
	uid_t uid = 0; /* unneeded, required by compiler */
1318
	pid_t pid;
L
Linus Torvalds 已提交
1319

1320 1321 1322
	/*
	 * Traditionally we see ptrace'd stopped tasks regardless of options.
	 */
1323
	if (!ptrace && !(wo->wo_flags & WUNTRACED))
R
Roland McGrath 已提交
1324 1325
		return 0;

1326 1327 1328
	if (!task_stopped_code(p, ptrace))
		return 0;

1329 1330 1331
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

1332 1333
	p_code = task_stopped_code(p, ptrace);
	if (unlikely(!p_code))
1334 1335
		goto unlock_sig;

1336
	exit_code = *p_code;
1337 1338 1339
	if (!exit_code)
		goto unlock_sig;

1340
	if (!unlikely(wo->wo_flags & WNOWAIT))
1341
		*p_code = 0;
1342

1343
	uid = from_kuid_munged(current_user_ns(), task_uid(p));
1344 1345 1346
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
		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);
1357
	pid = task_pid_vnr(p);
R
Roland McGrath 已提交
1358
	why = ptrace ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1359 1360
	read_unlock(&tasklist_lock);

1361 1362 1363 1364 1365 1366 1367
	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
1368

1369
	infop = wo->wo_info;
L
Linus Torvalds 已提交
1370 1371 1372 1373 1374
	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
1375
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1376 1377 1378
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1379
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1380
	if (!retval && infop)
1381
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1382
	if (!retval)
1383
		retval = pid;
L
Linus Torvalds 已提交
1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395
	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.
 */
1396
static int wait_task_continued(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1397 1398 1399 1400 1401
{
	int retval;
	pid_t pid;
	uid_t uid;

1402
	if (!unlikely(wo->wo_flags & WCONTINUED))
R
Roland McGrath 已提交
1403 1404
		return 0;

L
Linus Torvalds 已提交
1405 1406 1407 1408 1409 1410 1411 1412 1413
	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;
	}
1414
	if (!unlikely(wo->wo_flags & WNOWAIT))
L
Linus Torvalds 已提交
1415
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
1416
	uid = from_kuid_munged(current_user_ns(), task_uid(p));
L
Linus Torvalds 已提交
1417 1418
	spin_unlock_irq(&p->sighand->siglock);

1419
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1420 1421 1422
	get_task_struct(p);
	read_unlock(&tasklist_lock);

1423 1424 1425
	if (!wo->wo_info) {
		retval = wo->wo_rusage
			? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
L
Linus Torvalds 已提交
1426
		put_task_struct(p);
1427 1428
		if (!retval && wo->wo_stat)
			retval = put_user(0xffff, wo->wo_stat);
L
Linus Torvalds 已提交
1429
		if (!retval)
1430
			retval = pid;
L
Linus Torvalds 已提交
1431
	} else {
1432 1433
		retval = wait_noreap_copyout(wo, p, pid, uid,
					     CLD_CONTINUED, SIGCONT);
L
Linus Torvalds 已提交
1434 1435 1436 1437 1438 1439
		BUG_ON(retval == 0);
	}

	return retval;
}

R
Roland McGrath 已提交
1440 1441 1442
/*
 * Consider @p for a wait by @parent.
 *
1443
 * -ECHILD should be in ->notask_error before the first call.
R
Roland McGrath 已提交
1444 1445
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue;
1446
 * then ->notask_error is 0 if @p is an eligible child,
1447
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1448
 */
1449 1450
static int wait_consider_task(struct wait_opts *wo, int ptrace,
				struct task_struct *p)
R
Roland McGrath 已提交
1451
{
1452
	int ret = eligible_child(wo, p);
1453
	if (!ret)
R
Roland McGrath 已提交
1454 1455
		return ret;

1456
	ret = security_task_wait(p);
1457 1458 1459 1460 1461 1462 1463 1464
	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.
		 */
1465 1466
		if (wo->notask_error)
			wo->notask_error = ret;
1467
		return 0;
1468 1469
	}

1470
	/* dead body doesn't have much to contribute */
1471 1472 1473 1474 1475 1476 1477
	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;
1478
		return 0;
1479
	}
1480

1481 1482
	/* slay zombie? */
	if (p->exit_state == EXIT_ZOMBIE) {
R
Roland McGrath 已提交
1483
		/*
1484 1485 1486
		 * 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
1487
		 */
T
Tejun Heo 已提交
1488
		if (likely(!ptrace) && unlikely(p->ptrace)) {
1489 1490 1491 1492
			/* it will become visible, clear notask_error */
			wo->notask_error = 0;
			return 0;
		}
R
Roland McGrath 已提交
1493

1494 1495 1496
		/* 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
1497

R
Roland McGrath 已提交
1498
		/*
1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520
		 * 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 {
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530
		/*
		 * 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.
		 */
1531
		if (likely(!ptrace) && p->ptrace && !ptrace_reparented(p))
1532 1533
			return 0;

1534 1535 1536
		/*
		 * @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
1537
		 */
1538
		wo->notask_error = 0;
R
Roland McGrath 已提交
1539 1540
	}

R
Roland McGrath 已提交
1541
	/*
1542 1543
	 * 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
1544
	 */
1545 1546 1547
	ret = wait_task_stopped(wo, ptrace, p);
	if (ret)
		return ret;
R
Roland McGrath 已提交
1548 1549

	/*
1550 1551 1552
	 * 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
1553
	 */
1554
	return wait_task_continued(wo, p);
R
Roland McGrath 已提交
1555 1556 1557 1558 1559
}

/*
 * Do the work of do_wait() for one thread in the group, @tsk.
 *
1560
 * -ECHILD should be in ->notask_error before the first call.
R
Roland McGrath 已提交
1561 1562
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue; then
1563
 * ->notask_error is 0 if there were any eligible children,
1564
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1565
 */
1566
static int do_wait_thread(struct wait_opts *wo, struct task_struct *tsk)
R
Roland McGrath 已提交
1567 1568 1569 1570
{
	struct task_struct *p;

	list_for_each_entry(p, &tsk->children, sibling) {
1571 1572 1573
		int ret = wait_consider_task(wo, 0, p);
		if (ret)
			return ret;
R
Roland McGrath 已提交
1574 1575 1576 1577 1578
	}

	return 0;
}

1579
static int ptrace_do_wait(struct wait_opts *wo, struct task_struct *tsk)
R
Roland McGrath 已提交
1580 1581 1582
{
	struct task_struct *p;

R
Roland McGrath 已提交
1583
	list_for_each_entry(p, &tsk->ptraced, ptrace_entry) {
1584
		int ret = wait_consider_task(wo, 1, p);
R
Roland McGrath 已提交
1585
		if (ret)
R
Roland McGrath 已提交
1586 1587 1588 1589 1590 1591
			return ret;
	}

	return 0;
}

1592 1593 1594 1595 1596 1597 1598
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;

1599
	if (!eligible_pid(wo, p))
1600 1601
		return 0;

1602 1603 1604
	if ((wo->wo_flags & __WNOTHREAD) && wait->private != p->parent)
		return 0;

1605 1606 1607
	return default_wake_function(wait, mode, sync, key);
}

1608 1609
void __wake_up_parent(struct task_struct *p, struct task_struct *parent)
{
1610 1611
	__wake_up_sync_key(&parent->signal->wait_chldexit,
				TASK_INTERRUPTIBLE, 1, p);
1612 1613
}

1614
static long do_wait(struct wait_opts *wo)
L
Linus Torvalds 已提交
1615 1616
{
	struct task_struct *tsk;
R
Roland McGrath 已提交
1617
	int retval;
L
Linus Torvalds 已提交
1618

1619
	trace_sched_process_wait(wo->wo_pid);
1620

1621 1622 1623
	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
1624
repeat:
R
Roland McGrath 已提交
1625 1626
	/*
	 * If there is nothing that can match our critiera just get out.
1627 1628 1629
	 * 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
1630
	 */
1631
	wo->notask_error = -ECHILD;
1632 1633
	if ((wo->wo_type < PIDTYPE_MAX) &&
	   (!wo->wo_pid || hlist_empty(&wo->wo_pid->tasks[wo->wo_type])))
1634
		goto notask;
E
Eric W. Biederman 已提交
1635

1636
	set_current_state(TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
1637 1638 1639
	read_lock(&tasklist_lock);
	tsk = current;
	do {
1640 1641 1642
		retval = do_wait_thread(wo, tsk);
		if (retval)
			goto end;
1643

1644 1645
		retval = ptrace_do_wait(wo, tsk);
		if (retval)
R
Roland McGrath 已提交
1646 1647
			goto end;

1648
		if (wo->wo_flags & __WNOTHREAD)
L
Linus Torvalds 已提交
1649
			break;
1650
	} while_each_thread(current, tsk);
L
Linus Torvalds 已提交
1651
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1652

1653
notask:
1654 1655
	retval = wo->notask_error;
	if (!retval && !(wo->wo_flags & WNOHANG)) {
L
Linus Torvalds 已提交
1656
		retval = -ERESTARTSYS;
R
Roland McGrath 已提交
1657 1658 1659 1660
		if (!signal_pending(current)) {
			schedule();
			goto repeat;
		}
L
Linus Torvalds 已提交
1661 1662
	}
end:
1663
	__set_current_state(TASK_RUNNING);
1664
	remove_wait_queue(&current->signal->wait_chldexit, &wo->child_wait);
L
Linus Torvalds 已提交
1665 1666 1667
	return retval;
}

1668 1669
SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *,
		infop, int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1670
{
1671
	struct wait_opts wo;
E
Eric W. Biederman 已提交
1672 1673
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1674 1675 1676 1677 1678 1679 1680 1681 1682
	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
1683
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1684 1685
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1686 1687
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1688 1689 1690
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1691 1692
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1693 1694 1695 1696 1697 1698
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1699 1700
	if (type < PIDTYPE_MAX)
		pid = find_get_pid(upid);
1701 1702 1703 1704 1705 1706 1707 1708

	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);
1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731

	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
1732
	put_pid(pid);
L
Linus Torvalds 已提交
1733 1734

	/* avoid REGPARM breakage on x86: */
1735
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1736 1737 1738
	return ret;
}

1739 1740
SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr,
		int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1741
{
1742
	struct wait_opts wo;
E
Eric W. Biederman 已提交
1743 1744
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1745 1746 1747 1748 1749
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1750 1751 1752 1753 1754 1755 1756 1757

	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;
1758
		pid = get_task_pid(current, PIDTYPE_PGID);
E
Eric W. Biederman 已提交
1759 1760 1761 1762 1763
	} else /* upid > 0 */ {
		type = PIDTYPE_PID;
		pid = find_get_pid(upid);
	}

1764 1765 1766 1767 1768 1769 1770
	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
1771
	put_pid(pid);
L
Linus Torvalds 已提交
1772 1773

	/* avoid REGPARM breakage on x86: */
1774
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1775 1776 1777 1778 1779 1780 1781 1782 1783
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
1784
SYSCALL_DEFINE3(waitpid, pid_t, pid, int __user *, stat_addr, int, options)
L
Linus Torvalds 已提交
1785 1786 1787 1788 1789
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif