exit.c 46.7 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>
L
Linus Torvalds 已提交
54 55 56 57 58 59

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

A
Adrian Bunk 已提交
60 61
static void exit_mm(struct task_struct * tsk);

62
static void __unhash_process(struct task_struct *p, bool group_dead)
L
Linus Torvalds 已提交
63 64 65
{
	nr_threads--;
	detach_pid(p, PIDTYPE_PID);
66
	if (group_dead) {
L
Linus Torvalds 已提交
67 68
		detach_pid(p, PIDTYPE_PGID);
		detach_pid(p, PIDTYPE_SID);
69

70
		list_del_rcu(&p->tasks);
71
		list_del_init(&p->sibling);
72
		__this_cpu_dec(process_counts);
L
Linus Torvalds 已提交
73
	}
O
Oleg Nesterov 已提交
74
	list_del_rcu(&p->thread_group);
L
Linus Torvalds 已提交
75 76
}

77 78 79 80 81 82
/*
 * This function expects the tasklist_lock write-locked.
 */
static void __exit_signal(struct task_struct *tsk)
{
	struct signal_struct *sig = tsk->signal;
83
	bool group_dead = thread_group_leader(tsk);
84
	struct sighand_struct *sighand;
85
	struct tty_struct *uninitialized_var(tty);
86

87 88
	sighand = rcu_dereference_check(tsk->sighand,
					rcu_read_lock_held() ||
89
					lockdep_tasklist_lock_is_held());
90 91 92
	spin_lock(&sighand->siglock);

	posix_cpu_timers_exit(tsk);
93
	if (group_dead) {
94
		posix_cpu_timers_exit_group(tsk);
95 96
		tty = sig->tty;
		sig->tty = NULL;
97
	} else {
98 99 100 101 102 103 104 105
		/*
		 * 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);

106 107 108 109
		/*
		 * If there is any task waiting for the group exit
		 * then notify it:
		 */
110
		if (sig->notify_count > 0 && !--sig->notify_count)
111
			wake_up_process(sig->group_exit_task);
112

113 114 115 116 117 118 119 120 121 122 123 124
		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.
		 */
125 126
		sig->utime = cputime_add(sig->utime, tsk->utime);
		sig->stime = cputime_add(sig->stime, tsk->stime);
H
Hidetoshi Seto 已提交
127
		sig->gtime = cputime_add(sig->gtime, tsk->gtime);
128 129 130 131
		sig->min_flt += tsk->min_flt;
		sig->maj_flt += tsk->maj_flt;
		sig->nvcsw += tsk->nvcsw;
		sig->nivcsw += tsk->nivcsw;
132 133
		sig->inblock += task_io_get_inblock(tsk);
		sig->oublock += task_io_get_oublock(tsk);
134
		task_io_accounting_add(&sig->ioac, &tsk->ioac);
135
		sig->sum_sched_runtime += tsk->se.sum_exec_runtime;
136 137
	}

138
	sig->nr_threads--;
139
	__unhash_process(tsk, group_dead);
140

141 142 143 144 145
	/*
	 * Do this under ->siglock, we can race with another thread
	 * doing sigqueue_free() if we have SIGQUEUE_PREALLOC signals.
	 */
	flush_sigqueue(&tsk->pending);
146
	tsk->sighand = NULL;
147 148
	spin_unlock(&sighand->siglock);

149
	__cleanup_sighand(sighand);
150
	clear_tsk_thread_flag(tsk,TIF_SIGPENDING);
151
	if (group_dead) {
152
		flush_sigqueue(&sig->shared_pending);
153
		tty_kref_put(tty);
154 155 156
	}
}

157 158
static void delayed_put_task_struct(struct rcu_head *rhp)
{
159 160
	struct task_struct *tsk = container_of(rhp, struct task_struct, rcu);

161
	perf_event_delayed_put(tsk);
162 163
	trace_sched_process_free(tsk);
	put_task_struct(tsk);
164 165
}

R
Roland McGrath 已提交
166

L
Linus Torvalds 已提交
167 168
void release_task(struct task_struct * p)
{
169
	struct task_struct *leader;
L
Linus Torvalds 已提交
170
	int zap_leader;
171
repeat:
R
Roland McGrath 已提交
172
	tracehook_prepare_release_task(p);
173
	/* don't need to get the RCU readlock here - the process is dead and
174 175
	 * can't be modifying its own credentials. But shut RCU-lockdep up */
	rcu_read_lock();
176
	atomic_dec(&__task_cred(p)->user->processes);
177
	rcu_read_unlock();
178

179
	proc_flush_task(p);
180

L
Linus Torvalds 已提交
181
	write_lock_irq(&tasklist_lock);
R
Roland McGrath 已提交
182
	tracehook_finish_release_task(p);
L
Linus Torvalds 已提交
183
	__exit_signal(p);
184

L
Linus Torvalds 已提交
185 186 187 188 189 190 191 192
	/*
	 * 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) {
193
		BUG_ON(task_detached(leader));
L
Linus Torvalds 已提交
194 195 196 197 198 199 200 201 202
		do_notify_parent(leader, leader->exit_signal);
		/*
		 * 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.
		 *
		 * do_notify_parent() will have marked it self-reaping in
		 * that case.
		 */
203
		zap_leader = task_detached(leader);
R
Roland McGrath 已提交
204 205 206 207 208 209 210

		/*
		 * This maintains the invariant that release_task()
		 * only runs on a task in EXIT_DEAD, just for sanity.
		 */
		if (zap_leader)
			leader->exit_state = EXIT_DEAD;
L
Linus Torvalds 已提交
211 212 213 214
	}

	write_unlock_irq(&tasklist_lock);
	release_thread(p);
215
	call_rcu(&p->rcu, delayed_put_task_struct);
L
Linus Torvalds 已提交
216 217 218 219 220 221 222 223 224 225

	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...
226 227
 *
 * The caller must hold rcu lock or the tasklist lock.
L
Linus Torvalds 已提交
228
 */
229
struct pid *session_of_pgrp(struct pid *pgrp)
L
Linus Torvalds 已提交
230 231
{
	struct task_struct *p;
232
	struct pid *sid = NULL;
233

234
	p = pid_task(pgrp, PIDTYPE_PGID);
235
	if (p == NULL)
236
		p = pid_task(pgrp, PIDTYPE_PID);
237
	if (p != NULL)
238
		sid = task_session(p);
239

L
Linus Torvalds 已提交
240 241 242 243 244 245 246 247 248 249 250
	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?"
 */
251
static int will_become_orphaned_pgrp(struct pid *pgrp, struct task_struct *ignored_task)
L
Linus Torvalds 已提交
252 253 254
{
	struct task_struct *p;

255
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
256 257 258
		if ((p == ignored_task) ||
		    (p->exit_state && thread_group_empty(p)) ||
		    is_global_init(p->real_parent))
L
Linus Torvalds 已提交
259
			continue;
260

261
		if (task_pgrp(p->real_parent) != pgrp &&
262 263
		    task_session(p->real_parent) == task_session(p))
			return 0;
264
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
265 266

	return 1;
L
Linus Torvalds 已提交
267 268
}

269
int is_current_pgrp_orphaned(void)
L
Linus Torvalds 已提交
270 271 272 273
{
	int retval;

	read_lock(&tasklist_lock);
274
	retval = will_become_orphaned_pgrp(task_pgrp(current), NULL);
L
Linus Torvalds 已提交
275 276 277 278 279
	read_unlock(&tasklist_lock);

	return retval;
}

280
static int has_stopped_jobs(struct pid *pgrp)
L
Linus Torvalds 已提交
281 282 283 284
{
	int retval = 0;
	struct task_struct *p;

285
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
M
Matthew Wilcox 已提交
286
		if (!task_is_stopped(p))
L
Linus Torvalds 已提交
287 288 289
			continue;
		retval = 1;
		break;
290
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
L
Linus Torvalds 已提交
291 292 293
	return retval;
}

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);
439 440 441 442
	/*
	 * We don't want to have TIF_FREEZE set if the system-wide hibernation
	 * 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
static void close_files(struct files_struct * files)
L
Linus Torvalds 已提交
470 471
{
	int i, j;
472
	struct fdtable *fdt;
L
Linus Torvalds 已提交
473 474

	j = 0;
475 476 477 478

	/*
	 * It is safe to dereference the fd table without RCU or
	 * ->file_lock because this is the last reference to the
479
	 * files structure.  But use RCU to shut RCU-lockdep up.
480
	 */
481
	rcu_read_lock();
482
	fdt = files_fdtable(files);
483
	rcu_read_unlock();
L
Linus Torvalds 已提交
484 485 486
	for (;;) {
		unsigned long set;
		i = j * __NFDBITS;
487
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
488
			break;
489
		set = fdt->open_fds->fds_bits[j++];
L
Linus Torvalds 已提交
490 491
		while (set) {
			if (set & 1) {
492
				struct file * file = xchg(&fdt->fd[i], NULL);
493
				if (file) {
L
Linus Torvalds 已提交
494
					filp_close(file, files);
495 496
					cond_resched();
				}
L
Linus Torvalds 已提交
497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516
			}
			i++;
			set >>= 1;
		}
	}
}

struct files_struct *get_files_struct(struct task_struct *task)
{
	struct files_struct *files;

	task_lock(task);
	files = task->files;
	if (files)
		atomic_inc(&files->count);
	task_unlock(task);

	return files;
}

517
void put_files_struct(struct files_struct *files)
L
Linus Torvalds 已提交
518
{
519 520
	struct fdtable *fdt;

L
Linus Torvalds 已提交
521 522 523 524
	if (atomic_dec_and_test(&files->count)) {
		close_files(files);
		/*
		 * Free the fd and fdset arrays if we expanded them.
525 526 527
		 * If the fdtable was embedded, pass files for freeing
		 * at the end of the RCU grace period. Otherwise,
		 * you can free files immediately.
L
Linus Torvalds 已提交
528
		 */
529
		rcu_read_lock();
530
		fdt = files_fdtable(files);
531
		if (fdt != &files->fdtab)
532
			kmem_cache_free(files_cachep, files);
533
		free_fdtable(fdt);
534
		rcu_read_unlock();
L
Linus Torvalds 已提交
535 536 537
	}
}

538
void reset_files_struct(struct files_struct *files)
539
{
540
	struct task_struct *tsk = current;
541 542 543 544 545 546 547 548 549
	struct files_struct *old;

	old = tsk->files;
	task_lock(tsk);
	tsk->files = files;
	task_unlock(tsk);
	put_files_struct(old);
}

550
void exit_files(struct task_struct *tsk)
L
Linus Torvalds 已提交
551 552 553 554 555 556 557 558 559 560 561
{
	struct files_struct * files = tsk->files;

	if (files) {
		task_lock(tsk);
		tsk->files = NULL;
		task_unlock(tsk);
		put_files_struct(files);
	}
}

562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599
#ifdef CONFIG_MM_OWNER
/*
 * Task p is exiting and it owned mm, lets find a new owner for it
 */
static inline int
mm_need_new_owner(struct mm_struct *mm, struct task_struct *p)
{
	/*
	 * If there are other users of the mm and the owner (us) is exiting
	 * we need to find a new owner to take on the responsibility.
	 */
	if (atomic_read(&mm->mm_users) <= 1)
		return 0;
	if (mm->owner != p)
		return 0;
	return 1;
}

void mm_update_next_owner(struct mm_struct *mm)
{
	struct task_struct *c, *g, *p = current;

retry:
	if (!mm_need_new_owner(mm, p))
		return;

	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
	 */
600
	list_for_each_entry(c, &p->real_parent->children, sibling) {
601 602 603 604 605 606 607 608 609 610 611 612 613 614
		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);
615 616 617
	/*
	 * We found no owner yet mm_users > 1: this implies that we are
	 * most likely racing with swapoff (try_to_unuse()) or /proc or
618
	 * ptrace or page migration (get_task_mm()).  Mark owner as NULL.
619 620
	 */
	mm->owner = NULL;
621 622 623 624 625 626 627 628 629 630
	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);
631 632 633 634 635
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
636 637 638 639 640 641 642 643 644 645 646
	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
647 648 649 650
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
651
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
652 653
{
	struct mm_struct *mm = tsk->mm;
654
	struct core_state *core_state;
L
Linus Torvalds 已提交
655 656 657 658 659 660

	mm_release(tsk, mm);
	if (!mm)
		return;
	/*
	 * Serialize with any possible pending coredump.
661
	 * We must hold mmap_sem around checking core_state
L
Linus Torvalds 已提交
662
	 * and clearing tsk->mm.  The core-inducing thread
663
	 * will increment ->nr_threads for each thread in the
L
Linus Torvalds 已提交
664 665 666
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
667 668 669
	core_state = mm->core_state;
	if (core_state) {
		struct core_thread self;
L
Linus Torvalds 已提交
670 671
		up_read(&mm->mmap_sem);

672 673 674 675 676 677 678 679
		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
680

O
Oleg Nesterov 已提交
681 682 683 684 685 686 687
		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
688 689 690
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
691
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
692 693 694 695 696
	/* more a memory barrier than a real lock */
	task_lock(tsk);
	tsk->mm = NULL;
	up_read(&mm->mmap_sem);
	enter_lazy_tlb(mm, current);
697 698
	/* We don't want this task to be frozen prematurely */
	clear_freeze_flag(tsk);
Y
Ying Han 已提交
699 700
	if (tsk->signal->oom_score_adj == OOM_SCORE_ADJ_MIN)
		atomic_dec(&mm->oom_disable_count);
L
Linus Torvalds 已提交
701
	task_unlock(tsk);
702
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
703 704 705 706 707 708 709
	mmput(mm);
}

/*
 * When we die, we re-parent all our children.
 * Try to give them to another thread in our thread
 * group, and if no such member exists, give it to
710 711
 * the child reaper process (ie "init") in our pid
 * space.
L
Linus Torvalds 已提交
712
 */
713
static struct task_struct *find_new_reaper(struct task_struct *father)
714 715
	__releases(&tasklist_lock)
	__acquires(&tasklist_lock)
L
Linus Torvalds 已提交
716
{
717 718
	struct pid_namespace *pid_ns = task_active_pid_ns(father);
	struct task_struct *thread;
L
Linus Torvalds 已提交
719

720 721 722 723 724 725 726 727
	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
728

729 730 731 732
	if (unlikely(pid_ns->child_reaper == father)) {
		write_unlock_irq(&tasklist_lock);
		if (unlikely(pid_ns == &init_pid_ns))
			panic("Attempted to kill init!");
L
Linus Torvalds 已提交
733

734 735
		zap_pid_ns_processes(pid_ns);
		write_lock_irq(&tasklist_lock);
L
Linus Torvalds 已提交
736
		/*
737 738 739
		 * We can not clear ->child_reaper or leave it alone.
		 * There may by stealth EXIT_DEAD tasks on ->children,
		 * forget_original_parent() must move them somewhere.
L
Linus Torvalds 已提交
740
		 */
741
		pid_ns->child_reaper = init_pid_ns.child_reaper;
L
Linus Torvalds 已提交
742
	}
743

744 745 746
	return pid_ns->child_reaper;
}

747 748 749
/*
* Any that need to be release_task'd are put on the @dead list.
 */
750
static void reparent_leader(struct task_struct *father, struct task_struct *p,
751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767
				struct list_head *dead)
{
	list_move_tail(&p->sibling, &p->real_parent->children);

	if (task_detached(p))
		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. */
768
	if (!task_ptrace(p) &&
769 770 771 772 773 774 775 776 777 778 779
	    p->exit_state == EXIT_ZOMBIE && thread_group_empty(p)) {
		do_notify_parent(p, p->exit_signal);
		if (task_detached(p)) {
			p->exit_state = EXIT_DEAD;
			list_move_tail(&p->sibling, dead);
		}
	}

	kill_orphaned_pgrp(p, father);
}

780
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
781
{
782
	struct task_struct *p, *n, *reaper;
783
	LIST_HEAD(dead_children);
784 785

	write_lock_irq(&tasklist_lock);
786 787 788 789 790
	/*
	 * Note that exit_ptrace() and find_new_reaper() might
	 * drop tasklist_lock and reacquire it.
	 */
	exit_ptrace(father);
791
	reaper = find_new_reaper(father);
R
Roland McGrath 已提交
792

793
	list_for_each_entry_safe(p, n, &father->children, sibling) {
794 795 796 797 798 799 800 801 802 803 804 805
		struct task_struct *t = p;
		do {
			t->real_parent = reaper;
			if (t->parent == father) {
				BUG_ON(task_ptrace(t));
				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
806
	}
807
	write_unlock_irq(&tasklist_lock);
808

809 810
	BUG_ON(!list_empty(&father->children));

811 812
	list_for_each_entry_safe(p, n, &dead_children, sibling) {
		list_del_init(&p->sibling);
813 814
		release_task(p);
	}
L
Linus Torvalds 已提交
815 816 817 818 819 820
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
821
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
822
{
R
Roland McGrath 已提交
823 824
	int signal;
	void *cookie;
L
Linus Torvalds 已提交
825 826 827 828 829 830 831 832 833

	/*
	 * 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)
	 */
834
	forget_original_parent(tsk);
835
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
836

837
	write_lock_irq(&tasklist_lock);
838 839
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
840

841
	/* Let father know we died
L
Linus Torvalds 已提交
842 843
	 *
	 * Thread signals are configurable, but you aren't going to use
L
Lucas De Marchi 已提交
844
	 * that to send signals to arbitrary processes.
L
Linus Torvalds 已提交
845 846 847 848 849 850 851 852 853 854
	 * That stops right now.
	 *
	 * If the parent exec id doesn't match the exec id we saved
	 * when we started then we know the parent has changed security
	 * domain.
	 *
	 * If our self_exec id doesn't match our parent_exec_id then
	 * we have changed execution domain as these two values started
	 * the same after a fork.
	 */
855
	if (tsk->exit_signal != SIGCHLD && !task_detached(tsk) &&
856
	    (tsk->parent_exec_id != tsk->real_parent->self_exec_id ||
857
	     tsk->self_exec_id != tsk->parent_exec_id))
L
Linus Torvalds 已提交
858 859
		tsk->exit_signal = SIGCHLD;

R
Roland McGrath 已提交
860
	signal = tracehook_notify_death(tsk, &cookie, group_dead);
861
	if (signal >= 0)
R
Roland McGrath 已提交
862
		signal = do_notify_parent(tsk, signal);
L
Linus Torvalds 已提交
863

864
	tsk->exit_state = signal == DEATH_REAP ? EXIT_DEAD : EXIT_ZOMBIE;
L
Linus Torvalds 已提交
865

866 867
	/* mt-exec, de_thread() is waiting for group leader */
	if (unlikely(tsk->signal->notify_count < 0))
868
		wake_up_process(tsk->signal->group_exit_task);
L
Linus Torvalds 已提交
869 870
	write_unlock_irq(&tasklist_lock);

R
Roland McGrath 已提交
871 872
	tracehook_report_death(tsk, signal, cookie, group_dead);

L
Linus Torvalds 已提交
873
	/* If the process is dead, release it - nobody will wait for it */
874
	if (signal == DEATH_REAP)
L
Linus Torvalds 已提交
875 876 877
		release_task(tsk);
}

878 879 880 881 882 883 884
#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;

885
	free = stack_not_used(current);
886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902

	if (free >= lowest_to_date)
		return;

	spin_lock(&low_water_lock);
	if (free < lowest_to_date) {
		printk(KERN_WARNING "%s used greatest stack depth: %lu bytes "
				"left\n",
				current->comm, free);
		lowest_to_date = free;
	}
	spin_unlock(&low_water_lock);
}
#else
static inline void check_stack_usage(void) {}
#endif

903
NORET_TYPE void do_exit(long code)
L
Linus Torvalds 已提交
904 905 906 907 908 909
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

910
	WARN_ON(atomic_read(&tsk->fs_excl));
911
	WARN_ON(blk_needs_flush_plug(tsk));
912

L
Linus Torvalds 已提交
913 914 915 916 917
	if (unlikely(in_interrupt()))
		panic("Aiee, killing interrupt handler!");
	if (unlikely(!tsk->pid))
		panic("Attempted to kill the idle task!");

918 919 920 921 922 923 924 925 926
	/*
	 * 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);

R
Roland McGrath 已提交
927
	tracehook_report_exit(&code);
L
Linus Torvalds 已提交
928

929 930
	validate_creds_for_do_exit(tsk);

931 932 933 934 935 936 937
	/*
	 * 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");
938 939 940 941 942 943 944 945 946 947
		/*
		 * 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;
948 949 950 951
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

952 953
	exit_irq_thread();

O
Oleg Nesterov 已提交
954
	exit_signals(tsk);  /* sets PF_EXITING */
955 956 957 958
	/*
	 * tsk->flags are checked in the futex code to protect against
	 * an exiting task cleaning up the robust pi futexes.
	 */
959
	smp_mb();
960
	raw_spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
961 962 963

	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
964
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
965 966 967
				preempt_count());

	acct_update_integrals(tsk);
968
	/* sync mm's RSS info before statistics gathering */
969 970
	if (tsk->mm)
		sync_mm_rss(tsk, tsk->mm);
L
Linus Torvalds 已提交
971
	group_dead = atomic_dec_and_test(&tsk->signal->live);
972
	if (group_dead) {
973
		hrtimer_cancel(&tsk->signal->real_timer);
974
		exit_itimers(tsk->signal);
J
Jiri Pirko 已提交
975 976
		if (tsk->mm)
			setmax_mm_hiwater_rss(&tsk->signal->maxrss, tsk->mm);
977
	}
978
	acct_collect(code, group_dead);
M
Miloslav Trmac 已提交
979 980
	if (group_dead)
		tty_audit_exit();
981 982
	if (unlikely(tsk->audit_context))
		audit_free(tsk);
983

984
	tsk->exit_code = code;
985
	taskstats_exit(tsk, group_dead);
986

L
Linus Torvalds 已提交
987 988
	exit_mm(tsk);

989
	if (group_dead)
990
		acct_process();
991 992
	trace_sched_process_exit(tsk);

L
Linus Torvalds 已提交
993
	exit_sem(tsk);
994 995
	exit_files(tsk);
	exit_fs(tsk);
996
	check_stack_usage();
L
Linus Torvalds 已提交
997
	exit_thread();
998 999 1000 1001 1002 1003 1004 1005 1006

	/*
	 * 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);

1007
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
1008

1009
	if (group_dead)
L
Linus Torvalds 已提交
1010 1011
		disassociate_ctty(1);

A
Al Viro 已提交
1012
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
1013

M
Matt Helsley 已提交
1014
	proc_exit_connector(tsk);
1015

1016 1017 1018
	/*
	 * FIXME: do that only when needed, using sched_exit tracepoint
	 */
1019
	ptrace_put_breakpoints(tsk);
1020

1021
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
1022
#ifdef CONFIG_NUMA
1023
	task_lock(tsk);
1024
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
1025
	tsk->mempolicy = NULL;
1026
	task_unlock(tsk);
L
Linus Torvalds 已提交
1027
#endif
1028
#ifdef CONFIG_FUTEX
1029 1030
	if (unlikely(current->pi_state_cache))
		kfree(current->pi_state_cache);
1031
#endif
1032
	/*
1033
	 * Make sure we are holding no locks:
1034
	 */
1035
	debug_check_no_locks_held(tsk);
1036 1037 1038 1039 1040 1041
	/*
	 * 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
1042

1043
	if (tsk->io_context)
1044
		exit_io_context(tsk);
1045

1046 1047 1048
	if (tsk->splice_pipe)
		__free_pipe_info(tsk->splice_pipe);

1049 1050
	validate_creds_for_do_exit(tsk);

C
Coywolf Qi Hunt 已提交
1051
	preempt_disable();
1052
	exit_rcu();
O
Oleg Nesterov 已提交
1053
	/* causes final put_task_struct in finish_task_switch(). */
1054
	tsk->state = TASK_DEAD;
L
Linus Torvalds 已提交
1055 1056 1057
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
1058 1059
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
1060 1061
}

1062 1063
EXPORT_SYMBOL_GPL(do_exit);

L
Linus Torvalds 已提交
1064 1065 1066 1067
NORET_TYPE void complete_and_exit(struct completion *comp, long code)
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
1068

L
Linus Torvalds 已提交
1069 1070 1071 1072 1073
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

1074
SYSCALL_DEFINE1(exit, int, error_code)
L
Linus Torvalds 已提交
1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085
{
	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).
 */
NORET_TYPE void
do_group_exit(int exit_code)
{
1086 1087
	struct signal_struct *sig = current->signal;

L
Linus Torvalds 已提交
1088 1089
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

1090 1091
	if (signal_group_exit(sig))
		exit_code = sig->group_exit_code;
L
Linus Torvalds 已提交
1092 1093 1094
	else if (!thread_group_empty(current)) {
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1095
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1096 1097 1098 1099
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1100
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114
			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.
 */
1115
SYSCALL_DEFINE1(exit_group, int, error_code)
L
Linus Torvalds 已提交
1116 1117
{
	do_group_exit((error_code & 0xff) << 8);
1118 1119
	/* NOTREACHED */
	return 0;
L
Linus Torvalds 已提交
1120 1121
}

1122 1123 1124
struct wait_opts {
	enum pid_type		wo_type;
	int			wo_flags;
1125
	struct pid		*wo_pid;
1126 1127 1128 1129 1130

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

1131
	wait_queue_t		child_wait;
1132 1133 1134
	int			notask_error;
};

1135 1136
static inline
struct pid *task_pid_type(struct task_struct *task, enum pid_type type)
E
Eric W. Biederman 已提交
1137
{
1138 1139 1140
	if (type != PIDTYPE_PID)
		task = task->group_leader;
	return task->pids[type].pid;
E
Eric W. Biederman 已提交
1141 1142
}

1143
static int eligible_pid(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1144
{
1145 1146 1147
	return	wo->wo_type == PIDTYPE_MAX ||
		task_pid_type(p, wo->wo_type) == wo->wo_pid;
}
L
Linus Torvalds 已提交
1148

1149 1150 1151 1152
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
1153 1154 1155 1156 1157
	/* 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.) */
1158 1159
	if (((p->exit_signal != SIGCHLD) ^ !!(wo->wo_flags & __WCLONE))
	    && !(wo->wo_flags & __WALL))
L
Linus Torvalds 已提交
1160 1161
		return 0;

1162
	return 1;
L
Linus Torvalds 已提交
1163 1164
}

1165 1166
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
1167
{
1168 1169 1170
	struct siginfo __user *infop;
	int retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
1171

L
Linus Torvalds 已提交
1172
	put_task_struct(p);
1173
	infop = wo->wo_info;
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187
	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
1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198
	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.
 */
1199
static int wait_task_zombie(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1200 1201
{
	unsigned long state;
1202
	int retval, status, traced;
1203
	pid_t pid = task_pid_vnr(p);
1204
	uid_t uid = __task_cred(p)->uid;
1205
	struct siginfo __user *infop;
L
Linus Torvalds 已提交
1206

1207
	if (!likely(wo->wo_flags & WEXITED))
R
Roland McGrath 已提交
1208 1209
		return 0;

1210
	if (unlikely(wo->wo_flags & WNOWAIT)) {
L
Linus Torvalds 已提交
1211
		int exit_code = p->exit_code;
1212
		int why;
L
Linus Torvalds 已提交
1213 1214 1215 1216 1217 1218 1219 1220 1221 1222

		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;
		}
1223
		return wait_noreap_copyout(wo, p, pid, uid, why, status);
L
Linus Torvalds 已提交
1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235
	}

	/*
	 * 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;
	}

1236
	traced = ptrace_reparented(p);
1237 1238 1239 1240 1241
	/*
	 * It can be ptraced but not reparented, check
	 * !task_detached() to filter out sub-threads.
	 */
	if (likely(!traced) && likely(!task_detached(p))) {
1242 1243
		struct signal_struct *psig;
		struct signal_struct *sig;
J
Jiri Pirko 已提交
1244
		unsigned long maxrss;
1245
		cputime_t tgutime, tgstime;
1246

L
Linus Torvalds 已提交
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257
		/*
		 * 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
1258
		 * need to protect the access to parent->signal fields,
L
Linus Torvalds 已提交
1259 1260
		 * as other threads in the parent group can be right
		 * here reaping other children at the same time.
1261 1262 1263 1264
		 *
		 * 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
1265
		 */
1266
		thread_group_times(p, &tgutime, &tgstime);
1267 1268
		spin_lock_irq(&p->real_parent->sighand->siglock);
		psig = p->real_parent->signal;
1269 1270 1271
		sig = p->signal;
		psig->cutime =
			cputime_add(psig->cutime,
1272 1273
			cputime_add(tgutime,
				    sig->cutime));
1274 1275
		psig->cstime =
			cputime_add(psig->cstime,
1276 1277
			cputime_add(tgstime,
				    sig->cstime));
1278 1279 1280 1281 1282
		psig->cgtime =
			cputime_add(psig->cgtime,
			cputime_add(p->gtime,
			cputime_add(sig->gtime,
				    sig->cgtime)));
1283 1284 1285 1286 1287 1288 1289 1290
		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;
1291 1292 1293 1294 1295 1296
		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
1297 1298 1299
		maxrss = max(sig->maxrss, sig->cmaxrss);
		if (psig->cmaxrss < maxrss)
			psig->cmaxrss = maxrss;
1300 1301
		task_io_accounting_add(&psig->ioac, &p->ioac);
		task_io_accounting_add(&psig->ioac, &sig->ioac);
1302
		spin_unlock_irq(&p->real_parent->sighand->siglock);
L
Linus Torvalds 已提交
1303 1304 1305 1306 1307 1308 1309 1310
	}

	/*
	 * 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);

1311 1312
	retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
L
Linus Torvalds 已提交
1313 1314
	status = (p->signal->flags & SIGNAL_GROUP_EXIT)
		? p->signal->group_exit_code : p->exit_code;
1315 1316 1317 1318
	if (!retval && wo->wo_stat)
		retval = put_user(status, wo->wo_stat);

	infop = wo->wo_info;
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
	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)
1338
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1339
	if (!retval && infop)
1340
		retval = put_user(uid, &infop->si_uid);
1341
	if (!retval)
1342
		retval = pid;
1343 1344

	if (traced) {
L
Linus Torvalds 已提交
1345
		write_lock_irq(&tasklist_lock);
1346 1347 1348 1349 1350 1351 1352
		/* We dropped tasklist, ptracer could die and untrace */
		ptrace_unlink(p);
		/*
		 * If this is not a detached task, notify the parent.
		 * If it's still not detached after that, don't release
		 * it now.
		 */
1353
		if (!task_detached(p)) {
1354
			do_notify_parent(p, p->exit_signal);
1355
			if (!task_detached(p)) {
1356 1357
				p->exit_state = EXIT_ZOMBIE;
				p = NULL;
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362 1363
			}
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1364

L
Linus Torvalds 已提交
1365 1366 1367
	return retval;
}

1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379
static int *task_stopped_code(struct task_struct *p, bool ptrace)
{
	if (ptrace) {
		if (task_is_stopped_or_traced(p))
			return &p->exit_code;
	} else {
		if (p->signal->flags & SIGNAL_STOP_STOPPED)
			return &p->signal->group_exit_code;
	}
	return NULL;
}

1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
/**
 * 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
1397
 */
1398 1399
static int wait_task_stopped(struct wait_opts *wo,
				int ptrace, struct task_struct *p)
L
Linus Torvalds 已提交
1400
{
1401
	struct siginfo __user *infop;
1402
	int retval, exit_code, *p_code, why;
1403
	uid_t uid = 0; /* unneeded, required by compiler */
1404
	pid_t pid;
L
Linus Torvalds 已提交
1405

1406 1407 1408
	/*
	 * Traditionally we see ptrace'd stopped tasks regardless of options.
	 */
1409
	if (!ptrace && !(wo->wo_flags & WUNTRACED))
R
Roland McGrath 已提交
1410 1411
		return 0;

1412 1413 1414
	if (!task_stopped_code(p, ptrace))
		return 0;

1415 1416 1417
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

1418 1419
	p_code = task_stopped_code(p, ptrace);
	if (unlikely(!p_code))
1420 1421
		goto unlock_sig;

1422
	exit_code = *p_code;
1423 1424 1425
	if (!exit_code)
		goto unlock_sig;

1426
	if (!unlikely(wo->wo_flags & WNOWAIT))
1427
		*p_code = 0;
1428

1429
	uid = task_uid(p);
1430 1431 1432
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1433 1434 1435 1436 1437 1438 1439 1440 1441 1442
		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);
1443
	pid = task_pid_vnr(p);
R
Roland McGrath 已提交
1444
	why = ptrace ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1445 1446
	read_unlock(&tasklist_lock);

1447 1448 1449 1450 1451 1452 1453
	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
1454

1455
	infop = wo->wo_info;
L
Linus Torvalds 已提交
1456 1457 1458 1459 1460
	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
1461
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1462 1463 1464
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1465
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1466
	if (!retval && infop)
1467
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1468
	if (!retval)
1469
		retval = pid;
L
Linus Torvalds 已提交
1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481
	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.
 */
1482
static int wait_task_continued(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1483 1484 1485 1486 1487
{
	int retval;
	pid_t pid;
	uid_t uid;

1488
	if (!unlikely(wo->wo_flags & WCONTINUED))
R
Roland McGrath 已提交
1489 1490
		return 0;

L
Linus Torvalds 已提交
1491 1492 1493 1494 1495 1496 1497 1498 1499
	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;
	}
1500
	if (!unlikely(wo->wo_flags & WNOWAIT))
L
Linus Torvalds 已提交
1501
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
1502
	uid = task_uid(p);
L
Linus Torvalds 已提交
1503 1504
	spin_unlock_irq(&p->sighand->siglock);

1505
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1506 1507 1508
	get_task_struct(p);
	read_unlock(&tasklist_lock);

1509 1510 1511
	if (!wo->wo_info) {
		retval = wo->wo_rusage
			? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
L
Linus Torvalds 已提交
1512
		put_task_struct(p);
1513 1514
		if (!retval && wo->wo_stat)
			retval = put_user(0xffff, wo->wo_stat);
L
Linus Torvalds 已提交
1515
		if (!retval)
1516
			retval = pid;
L
Linus Torvalds 已提交
1517
	} else {
1518 1519
		retval = wait_noreap_copyout(wo, p, pid, uid,
					     CLD_CONTINUED, SIGCONT);
L
Linus Torvalds 已提交
1520 1521 1522 1523 1524 1525
		BUG_ON(retval == 0);
	}

	return retval;
}

R
Roland McGrath 已提交
1526 1527 1528
/*
 * Consider @p for a wait by @parent.
 *
1529
 * -ECHILD should be in ->notask_error before the first call.
R
Roland McGrath 已提交
1530 1531
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue;
1532
 * then ->notask_error is 0 if @p is an eligible child,
1533
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1534
 */
1535 1536
static int wait_consider_task(struct wait_opts *wo, int ptrace,
				struct task_struct *p)
R
Roland McGrath 已提交
1537
{
1538
	int ret = eligible_child(wo, p);
1539
	if (!ret)
R
Roland McGrath 已提交
1540 1541
		return ret;

1542
	ret = security_task_wait(p);
1543 1544 1545 1546 1547 1548 1549 1550
	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.
		 */
1551 1552
		if (wo->notask_error)
			wo->notask_error = ret;
1553
		return 0;
1554 1555
	}

1556 1557 1558 1559
	/* dead body doesn't have much to contribute */
	if (p->exit_state == EXIT_DEAD)
		return 0;

1560 1561
	/* slay zombie? */
	if (p->exit_state == EXIT_ZOMBIE) {
R
Roland McGrath 已提交
1562
		/*
1563 1564 1565
		 * 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
1566
		 */
1567 1568 1569 1570 1571
		if (likely(!ptrace) && unlikely(task_ptrace(p))) {
			/* it will become visible, clear notask_error */
			wo->notask_error = 0;
			return 0;
		}
R
Roland McGrath 已提交
1572

1573 1574 1575
		/* 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
1576

R
Roland McGrath 已提交
1577
		/*
1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599
		 * 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 {
1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613
		/*
		 * 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.
		 */
		if (likely(!ptrace) && task_ptrace(p) &&
		    same_thread_group(p->parent, p->real_parent))
			return 0;

1614 1615 1616
		/*
		 * @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
1617
		 */
1618
		wo->notask_error = 0;
R
Roland McGrath 已提交
1619 1620
	}

R
Roland McGrath 已提交
1621
	/*
1622 1623
	 * 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
1624
	 */
1625 1626 1627
	ret = wait_task_stopped(wo, ptrace, p);
	if (ret)
		return ret;
R
Roland McGrath 已提交
1628 1629

	/*
1630 1631 1632
	 * 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
1633
	 */
1634
	return wait_task_continued(wo, p);
R
Roland McGrath 已提交
1635 1636 1637 1638 1639
}

/*
 * Do the work of do_wait() for one thread in the group, @tsk.
 *
1640
 * -ECHILD should be in ->notask_error before the first call.
R
Roland McGrath 已提交
1641 1642
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue; then
1643
 * ->notask_error is 0 if there were any eligible children,
1644
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1645
 */
1646
static int do_wait_thread(struct wait_opts *wo, struct task_struct *tsk)
R
Roland McGrath 已提交
1647 1648 1649 1650
{
	struct task_struct *p;

	list_for_each_entry(p, &tsk->children, sibling) {
1651 1652 1653
		int ret = wait_consider_task(wo, 0, p);
		if (ret)
			return ret;
R
Roland McGrath 已提交
1654 1655 1656 1657 1658
	}

	return 0;
}

1659
static int ptrace_do_wait(struct wait_opts *wo, struct task_struct *tsk)
R
Roland McGrath 已提交
1660 1661 1662
{
	struct task_struct *p;

R
Roland McGrath 已提交
1663
	list_for_each_entry(p, &tsk->ptraced, ptrace_entry) {
1664
		int ret = wait_consider_task(wo, 1, p);
R
Roland McGrath 已提交
1665
		if (ret)
R
Roland McGrath 已提交
1666 1667 1668 1669 1670 1671
			return ret;
	}

	return 0;
}

1672 1673 1674 1675 1676 1677 1678
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;

1679
	if (!eligible_pid(wo, p))
1680 1681
		return 0;

1682 1683 1684
	if ((wo->wo_flags & __WNOTHREAD) && wait->private != p->parent)
		return 0;

1685 1686 1687
	return default_wake_function(wait, mode, sync, key);
}

1688 1689
void __wake_up_parent(struct task_struct *p, struct task_struct *parent)
{
1690 1691
	__wake_up_sync_key(&parent->signal->wait_chldexit,
				TASK_INTERRUPTIBLE, 1, p);
1692 1693
}

1694
static long do_wait(struct wait_opts *wo)
L
Linus Torvalds 已提交
1695 1696
{
	struct task_struct *tsk;
R
Roland McGrath 已提交
1697
	int retval;
L
Linus Torvalds 已提交
1698

1699
	trace_sched_process_wait(wo->wo_pid);
1700

1701 1702 1703
	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
1704
repeat:
R
Roland McGrath 已提交
1705 1706
	/*
	 * If there is nothing that can match our critiera just get out.
1707 1708 1709
	 * 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
1710
	 */
1711
	wo->notask_error = -ECHILD;
1712 1713
	if ((wo->wo_type < PIDTYPE_MAX) &&
	   (!wo->wo_pid || hlist_empty(&wo->wo_pid->tasks[wo->wo_type])))
1714
		goto notask;
E
Eric W. Biederman 已提交
1715

1716
	set_current_state(TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
1717 1718 1719
	read_lock(&tasklist_lock);
	tsk = current;
	do {
1720 1721 1722
		retval = do_wait_thread(wo, tsk);
		if (retval)
			goto end;
1723

1724 1725
		retval = ptrace_do_wait(wo, tsk);
		if (retval)
R
Roland McGrath 已提交
1726 1727
			goto end;

1728
		if (wo->wo_flags & __WNOTHREAD)
L
Linus Torvalds 已提交
1729
			break;
1730
	} while_each_thread(current, tsk);
L
Linus Torvalds 已提交
1731
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1732

1733
notask:
1734 1735
	retval = wo->notask_error;
	if (!retval && !(wo->wo_flags & WNOHANG)) {
L
Linus Torvalds 已提交
1736
		retval = -ERESTARTSYS;
R
Roland McGrath 已提交
1737 1738 1739 1740
		if (!signal_pending(current)) {
			schedule();
			goto repeat;
		}
L
Linus Torvalds 已提交
1741 1742
	}
end:
1743
	__set_current_state(TASK_RUNNING);
1744
	remove_wait_queue(&current->signal->wait_chldexit, &wo->child_wait);
L
Linus Torvalds 已提交
1745 1746 1747
	return retval;
}

1748 1749
SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *,
		infop, int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1750
{
1751
	struct wait_opts wo;
E
Eric W. Biederman 已提交
1752 1753
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1754 1755 1756 1757 1758 1759 1760 1761 1762
	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
1763
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1764 1765
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1766 1767
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1768 1769 1770
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1771 1772
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1773 1774 1775 1776 1777 1778
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1779 1780
	if (type < PIDTYPE_MAX)
		pid = find_get_pid(upid);
1781 1782 1783 1784 1785 1786 1787 1788

	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);
1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811

	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
1812
	put_pid(pid);
L
Linus Torvalds 已提交
1813 1814

	/* avoid REGPARM breakage on x86: */
1815
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1816 1817 1818
	return ret;
}

1819 1820
SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr,
		int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1821
{
1822
	struct wait_opts wo;
E
Eric W. Biederman 已提交
1823 1824
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1825 1826 1827 1828 1829
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1830 1831 1832 1833 1834 1835 1836 1837

	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;
1838
		pid = get_task_pid(current, PIDTYPE_PGID);
E
Eric W. Biederman 已提交
1839 1840 1841 1842 1843
	} else /* upid > 0 */ {
		type = PIDTYPE_PID;
		pid = find_get_pid(upid);
	}

1844 1845 1846 1847 1848 1849 1850
	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
1851
	put_pid(pid);
L
Linus Torvalds 已提交
1852 1853

	/* avoid REGPARM breakage on x86: */
1854
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1855 1856 1857 1858 1859 1860 1861 1862 1863
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
1864
SYSCALL_DEFINE3(waitpid, pid_t, pid, int __user *, stat_addr, int, options)
L
Linus Torvalds 已提交
1865 1866 1867 1868 1869
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif