exit.c 46.3 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 67
{
	nr_threads--;
	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);
L
Linus Torvalds 已提交
75
	}
O
Oleg Nesterov 已提交
76
	list_del_rcu(&p->thread_group);
L
Linus Torvalds 已提交
77 78
}

79 80 81 82 83 84
/*
 * This function expects the tasklist_lock write-locked.
 */
static void __exit_signal(struct task_struct *tsk)
{
	struct signal_struct *sig = tsk->signal;
85
	bool group_dead = thread_group_leader(tsk);
86
	struct sighand_struct *sighand;
87
	struct tty_struct *uninitialized_var(tty);
88

89
	sighand = rcu_dereference_check(tsk->sighand,
90
					lockdep_tasklist_lock_is_held());
91 92 93
	spin_lock(&sighand->siglock);

	posix_cpu_timers_exit(tsk);
94
	if (group_dead) {
95
		posix_cpu_timers_exit_group(tsk);
96 97
		tty = sig->tty;
		sig->tty = NULL;
98
	} else {
99 100 101 102 103 104 105 106
		/*
		 * 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);

107 108 109 110
		/*
		 * If there is any task waiting for the group exit
		 * then notify it:
		 */
111
		if (sig->notify_count > 0 && !--sig->notify_count)
112
			wake_up_process(sig->group_exit_task);
113

114 115 116 117 118 119 120 121 122 123 124 125
		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.
		 */
126 127 128
		sig->utime += tsk->utime;
		sig->stime += tsk->stime;
		sig->gtime += tsk->gtime;
129 130 131 132
		sig->min_flt += tsk->min_flt;
		sig->maj_flt += tsk->maj_flt;
		sig->nvcsw += tsk->nvcsw;
		sig->nivcsw += tsk->nivcsw;
133 134
		sig->inblock += task_io_get_inblock(tsk);
		sig->oublock += task_io_get_oublock(tsk);
135
		task_io_accounting_add(&sig->ioac, &tsk->ioac);
136
		sig->sum_sched_runtime += tsk->se.sum_exec_runtime;
137 138
	}

139
	sig->nr_threads--;
140
	__unhash_process(tsk, group_dead);
141

142 143 144 145 146
	/*
	 * Do this under ->siglock, we can race with another thread
	 * doing sigqueue_free() if we have SIGQUEUE_PREALLOC signals.
	 */
	flush_sigqueue(&tsk->pending);
147
	tsk->sighand = NULL;
148 149
	spin_unlock(&sighand->siglock);

150
	__cleanup_sighand(sighand);
151
	clear_tsk_thread_flag(tsk,TIF_SIGPENDING);
152
	if (group_dead) {
153
		flush_sigqueue(&sig->shared_pending);
154
		tty_kref_put(tty);
155 156 157
	}
}

158 159
static void delayed_put_task_struct(struct rcu_head *rhp)
{
160 161
	struct task_struct *tsk = container_of(rhp, struct task_struct, rcu);

162
	perf_event_delayed_put(tsk);
163 164
	trace_sched_process_free(tsk);
	put_task_struct(tsk);
165 166
}

R
Roland McGrath 已提交
167

L
Linus Torvalds 已提交
168 169
void release_task(struct task_struct * p)
{
170
	struct task_struct *leader;
L
Linus Torvalds 已提交
171
	int zap_leader;
172
repeat:
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);
T
Tejun Heo 已提交
182
	ptrace_release_task(p);
L
Linus Torvalds 已提交
183
	__exit_signal(p);
184

L
Linus Torvalds 已提交
185 186 187 188 189 190 191 192 193 194 195 196
	/*
	 * 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
197
		 */
198
		zap_leader = do_notify_parent(leader, leader->exit_signal);
R
Roland McGrath 已提交
199 200
		if (zap_leader)
			leader->exit_state = EXIT_DEAD;
L
Linus Torvalds 已提交
201 202 203 204
	}

	write_unlock_irq(&tasklist_lock);
	release_thread(p);
205
	call_rcu(&p->rcu, delayed_put_task_struct);
L
Linus Torvalds 已提交
206 207 208 209 210 211 212 213 214 215

	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...
216 217
 *
 * The caller must hold rcu lock or the tasklist lock.
L
Linus Torvalds 已提交
218
 */
219
struct pid *session_of_pgrp(struct pid *pgrp)
L
Linus Torvalds 已提交
220 221
{
	struct task_struct *p;
222
	struct pid *sid = NULL;
223

224
	p = pid_task(pgrp, PIDTYPE_PGID);
225
	if (p == NULL)
226
		p = pid_task(pgrp, PIDTYPE_PID);
227
	if (p != NULL)
228
		sid = task_session(p);
229

L
Linus Torvalds 已提交
230 231 232 233 234 235 236 237 238 239 240
	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?"
 */
241
static int will_become_orphaned_pgrp(struct pid *pgrp, struct task_struct *ignored_task)
L
Linus Torvalds 已提交
242 243 244
{
	struct task_struct *p;

245
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
246 247 248
		if ((p == ignored_task) ||
		    (p->exit_state && thread_group_empty(p)) ||
		    is_global_init(p->real_parent))
L
Linus Torvalds 已提交
249
			continue;
250

251
		if (task_pgrp(p->real_parent) != pgrp &&
252 253
		    task_session(p->real_parent) == task_session(p))
			return 0;
254
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
255 256

	return 1;
L
Linus Torvalds 已提交
257 258
}

259
int is_current_pgrp_orphaned(void)
L
Linus Torvalds 已提交
260 261 262 263
{
	int retval;

	read_lock(&tasklist_lock);
264
	retval = will_become_orphaned_pgrp(task_pgrp(current), NULL);
L
Linus Torvalds 已提交
265 266 267 268 269
	read_unlock(&tasklist_lock);

	return retval;
}

270
static bool has_stopped_jobs(struct pid *pgrp)
L
Linus Torvalds 已提交
271 272 273
{
	struct task_struct *p;

274
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
275 276
		if (p->signal->flags & SIGNAL_STOP_STOPPED)
			return true;
277
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
278 279

	return false;
L
Linus Torvalds 已提交
280 281
}

282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
/*
 * 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
313
/**
314
 * reparent_to_kthreadd - Reparent the calling kernel thread to kthreadd
L
Linus Torvalds 已提交
315 316
 *
 * If a kernel thread is launched as a result of a system call, or if
317 318
 * 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
319 320 321 322
 *
 * 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.
 *
323
 * NOTE that reparent_to_kthreadd() gives the caller full capabilities.
L
Linus Torvalds 已提交
324
 */
325
static void reparent_to_kthreadd(void)
L
Linus Torvalds 已提交
326 327 328 329 330
{
	write_lock_irq(&tasklist_lock);

	ptrace_unlink(current);
	/* Reparent to init */
331
	current->real_parent = current->parent = kthreadd_task;
R
Roland McGrath 已提交
332
	list_move_tail(&current->sibling, &current->real_parent->children);
L
Linus Torvalds 已提交
333 334 335 336

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

337
	if (task_nice(current) < 0)
L
Linus Torvalds 已提交
338 339 340 341 342 343
		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
344 345 346

	atomic_inc(&init_cred.usage);
	commit_creds(&init_cred);
L
Linus Torvalds 已提交
347 348 349
	write_unlock_irq(&tasklist_lock);
}

350
void __set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
351
{
352
	struct task_struct *curr = current->group_leader;
L
Linus Torvalds 已提交
353

354
	if (task_session(curr) != pid)
355
		change_pid(curr, PIDTYPE_SID, pid);
356 357

	if (task_pgrp(curr) != pid)
358
		change_pid(curr, PIDTYPE_PGID, pid);
L
Linus Torvalds 已提交
359 360
}

361
static void set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
362 363
{
	write_lock_irq(&tasklist_lock);
364
	__set_special_pids(pid);
L
Linus Torvalds 已提交
365 366 367 368
	write_unlock_irq(&tasklist_lock);
}

/*
369 370
 * 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
371 372 373
 */
int allow_signal(int sig)
{
374
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
375 376 377
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
378
	/* This is only needed for daemonize()'ed kthreads */
L
Linus Torvalds 已提交
379
	sigdelset(&current->blocked, sig);
380 381 382 383 384 385
	/*
	 * 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
386 387 388 389 390 391 392 393 394
	recalc_sigpending();
	spin_unlock_irq(&current->sighand->siglock);
	return 0;
}

EXPORT_SYMBOL(allow_signal);

int disallow_signal(int sig)
{
395
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
396 397 398
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
399
	current->sighand->action[(sig)-1].sa.sa_handler = SIG_IGN;
L
Linus Torvalds 已提交
400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426
	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);
427
	/*
428
	 * We don't want to get frozen, in case system-wide hibernation
429 430
	 * or suspend transition begins right now.
	 */
O
Oleg Nesterov 已提交
431
	current->flags |= (PF_NOFREEZE | PF_KTHREAD);
L
Linus Torvalds 已提交
432

433 434 435 436
	if (current->nsproxy != &init_nsproxy) {
		get_nsproxy(&init_nsproxy);
		switch_task_namespaces(current, &init_nsproxy);
	}
437
	set_special_pids(&init_struct_pid);
438
	proc_clear_tty(current);
L
Linus Torvalds 已提交
439 440 441 442 443 444 445 446

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

	/* Become as one with the init task */

447
	daemonize_fs_struct();
448
	exit_files(current);
L
Linus Torvalds 已提交
449 450 451
	current->files = init_task.files;
	atomic_inc(&current->files->count);

452
	reparent_to_kthreadd();
L
Linus Torvalds 已提交
453 454 455 456
}

EXPORT_SYMBOL(daemonize);

457
static void close_files(struct files_struct * files)
L
Linus Torvalds 已提交
458 459
{
	int i, j;
460
	struct fdtable *fdt;
L
Linus Torvalds 已提交
461 462

	j = 0;
463 464 465 466

	/*
	 * It is safe to dereference the fd table without RCU or
	 * ->file_lock because this is the last reference to the
467
	 * files structure.  But use RCU to shut RCU-lockdep up.
468
	 */
469
	rcu_read_lock();
470
	fdt = files_fdtable(files);
471
	rcu_read_unlock();
L
Linus Torvalds 已提交
472 473 474
	for (;;) {
		unsigned long set;
		i = j * __NFDBITS;
475
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
476
			break;
477
		set = fdt->open_fds->fds_bits[j++];
L
Linus Torvalds 已提交
478 479
		while (set) {
			if (set & 1) {
480
				struct file * file = xchg(&fdt->fd[i], NULL);
481
				if (file) {
L
Linus Torvalds 已提交
482
					filp_close(file, files);
483 484
					cond_resched();
				}
L
Linus Torvalds 已提交
485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504
			}
			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;
}

505
void put_files_struct(struct files_struct *files)
L
Linus Torvalds 已提交
506
{
507 508
	struct fdtable *fdt;

L
Linus Torvalds 已提交
509 510 511 512
	if (atomic_dec_and_test(&files->count)) {
		close_files(files);
		/*
		 * Free the fd and fdset arrays if we expanded them.
513 514 515
		 * 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
516
		 */
517
		rcu_read_lock();
518
		fdt = files_fdtable(files);
519
		if (fdt != &files->fdtab)
520
			kmem_cache_free(files_cachep, files);
521
		free_fdtable(fdt);
522
		rcu_read_unlock();
L
Linus Torvalds 已提交
523 524 525
	}
}

526
void reset_files_struct(struct files_struct *files)
527
{
528
	struct task_struct *tsk = current;
529 530 531 532 533 534 535 536 537
	struct files_struct *old;

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

538
void exit_files(struct task_struct *tsk)
L
Linus Torvalds 已提交
539 540 541 542 543 544 545 546 547 548 549
{
	struct files_struct * files = tsk->files;

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

550 551
#ifdef CONFIG_MM_OWNER
/*
552
 * A task is exiting.   If it owned this mm, find a new owner for the mm.
553 554 555 556 557 558
 */
void mm_update_next_owner(struct mm_struct *mm)
{
	struct task_struct *c, *g, *p = current;

retry:
559 560 561 562 563
	/*
	 * If the exiting or execing task is not the owner, it's
	 * someone else's problem.
	 */
	if (mm->owner != p)
564
		return;
565 566 567 568 569 570 571 572 573
	/*
	 * 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;
	}
574 575 576 577 578 579 580 581 582 583 584 585 586

	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
	 */
587
	list_for_each_entry(c, &p->real_parent->children, sibling) {
588 589 590 591 592 593 594 595 596 597 598 599 600 601
		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);
602 603 604
	/*
	 * We found no owner yet mm_users > 1: this implies that we are
	 * most likely racing with swapoff (try_to_unuse()) or /proc or
605
	 * ptrace or page migration (get_task_mm()).  Mark owner as NULL.
606 607
	 */
	mm->owner = NULL;
608 609 610 611 612 613 614 615 616 617
	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);
618 619 620 621 622
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
623 624 625 626 627 628 629 630 631 632 633
	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
634 635 636 637
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
638
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
639 640
{
	struct mm_struct *mm = tsk->mm;
641
	struct core_state *core_state;
L
Linus Torvalds 已提交
642 643 644 645 646 647

	mm_release(tsk, mm);
	if (!mm)
		return;
	/*
	 * Serialize with any possible pending coredump.
648
	 * We must hold mmap_sem around checking core_state
L
Linus Torvalds 已提交
649
	 * and clearing tsk->mm.  The core-inducing thread
650
	 * will increment ->nr_threads for each thread in the
L
Linus Torvalds 已提交
651 652 653
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
654 655 656
	core_state = mm->core_state;
	if (core_state) {
		struct core_thread self;
L
Linus Torvalds 已提交
657 658
		up_read(&mm->mmap_sem);

659 660 661 662 663 664 665 666
		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
667

O
Oleg Nesterov 已提交
668 669 670 671 672 673 674
		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
675 676 677
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
678
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
679 680 681 682 683 684
	/* 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);
685
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
686 687 688 689 690 691 692
	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
693 694
 * the child reaper process (ie "init") in our pid
 * space.
L
Linus Torvalds 已提交
695
 */
696
static struct task_struct *find_new_reaper(struct task_struct *father)
697 698
	__releases(&tasklist_lock)
	__acquires(&tasklist_lock)
L
Linus Torvalds 已提交
699
{
700 701
	struct pid_namespace *pid_ns = task_active_pid_ns(father);
	struct task_struct *thread;
L
Linus Torvalds 已提交
702

703 704 705 706 707 708 709 710
	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
711

712 713 714 715
	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
716

717 718
		zap_pid_ns_processes(pid_ns);
		write_lock_irq(&tasklist_lock);
L
Linus Torvalds 已提交
719
		/*
720 721 722
		 * 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
723
		 */
724
		pid_ns->child_reaper = init_pid_ns.child_reaper;
L
Linus Torvalds 已提交
725
	}
726

727 728 729
	return pid_ns->child_reaper;
}

730 731 732
/*
* Any that need to be release_task'd are put on the @dead list.
 */
733
static void reparent_leader(struct task_struct *father, struct task_struct *p,
734 735 736 737
				struct list_head *dead)
{
	list_move_tail(&p->sibling, &p->real_parent->children);

738
	if (p->exit_state == EXIT_DEAD)
739 740 741 742 743 744 745 746 747 748 749 750
		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
751
	if (!p->ptrace &&
752
	    p->exit_state == EXIT_ZOMBIE && thread_group_empty(p)) {
753
		if (do_notify_parent(p, p->exit_signal)) {
754 755 756 757 758 759 760 761
			p->exit_state = EXIT_DEAD;
			list_move_tail(&p->sibling, dead);
		}
	}

	kill_orphaned_pgrp(p, father);
}

762
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
763
{
764
	struct task_struct *p, *n, *reaper;
765
	LIST_HEAD(dead_children);
766 767

	write_lock_irq(&tasklist_lock);
768 769 770 771 772
	/*
	 * Note that exit_ptrace() and find_new_reaper() might
	 * drop tasklist_lock and reacquire it.
	 */
	exit_ptrace(father);
773
	reaper = find_new_reaper(father);
R
Roland McGrath 已提交
774

775
	list_for_each_entry_safe(p, n, &father->children, sibling) {
776 777 778 779
		struct task_struct *t = p;
		do {
			t->real_parent = reaper;
			if (t->parent == father) {
T
Tejun Heo 已提交
780
				BUG_ON(t->ptrace);
781 782 783 784 785 786 787
				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
788
	}
789
	write_unlock_irq(&tasklist_lock);
790

791 792
	BUG_ON(!list_empty(&father->children));

793 794
	list_for_each_entry_safe(p, n, &dead_children, sibling) {
		list_del_init(&p->sibling);
795 796
		release_task(p);
	}
L
Linus Torvalds 已提交
797 798 799 800 801 802
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
803
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
804
{
805
	bool autoreap;
L
Linus Torvalds 已提交
806 807 808 809 810 811 812 813 814

	/*
	 * 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)
	 */
815
	forget_original_parent(tsk);
816
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
817

818
	write_lock_irq(&tasklist_lock);
819 820
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
821

O
Oleg Nesterov 已提交
822 823 824 825 826 827 828 829 830 831 832 833
	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
834

835
	tsk->exit_state = autoreap ? EXIT_DEAD : EXIT_ZOMBIE;
L
Linus Torvalds 已提交
836

837 838
	/* mt-exec, de_thread() is waiting for group leader */
	if (unlikely(tsk->signal->notify_count < 0))
839
		wake_up_process(tsk->signal->group_exit_task);
L
Linus Torvalds 已提交
840 841 842
	write_unlock_irq(&tasklist_lock);

	/* If the process is dead, release it - nobody will wait for it */
843
	if (autoreap)
L
Linus Torvalds 已提交
844 845 846
		release_task(tsk);
}

847 848 849 850 851 852 853
#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;

854
	free = stack_not_used(current);
855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871

	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

872
void do_exit(long code)
L
Linus Torvalds 已提交
873 874 875 876 877 878
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

879
	WARN_ON(blk_needs_flush_plug(tsk));
880

L
Linus Torvalds 已提交
881 882 883 884 885
	if (unlikely(in_interrupt()))
		panic("Aiee, killing interrupt handler!");
	if (unlikely(!tsk->pid))
		panic("Attempted to kill the idle task!");

886 887 888 889 890 891 892 893 894
	/*
	 * 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
895
	ptrace_event(PTRACE_EVENT_EXIT, code);
L
Linus Torvalds 已提交
896

897 898
	validate_creds_for_do_exit(tsk);

899 900 901 902 903 904 905
	/*
	 * 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");
906 907 908 909 910 911 912 913 914 915
		/*
		 * 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;
916 917 918 919
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

O
Oleg Nesterov 已提交
920
	exit_signals(tsk);  /* sets PF_EXITING */
921 922 923 924
	/*
	 * tsk->flags are checked in the futex code to protect against
	 * an exiting task cleaning up the robust pi futexes.
	 */
925
	smp_mb();
926
	raw_spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
927

928 929
	exit_irq_thread();

L
Linus Torvalds 已提交
930 931
	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
932
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
933 934 935
				preempt_count());

	acct_update_integrals(tsk);
936
	/* sync mm's RSS info before statistics gathering */
937
	if (tsk->mm)
938
		sync_mm_rss(tsk->mm);
L
Linus Torvalds 已提交
939
	group_dead = atomic_dec_and_test(&tsk->signal->live);
940
	if (group_dead) {
941
		hrtimer_cancel(&tsk->signal->real_timer);
942
		exit_itimers(tsk->signal);
J
Jiri Pirko 已提交
943 944
		if (tsk->mm)
			setmax_mm_hiwater_rss(&tsk->signal->maxrss, tsk->mm);
945
	}
946
	acct_collect(code, group_dead);
M
Miloslav Trmac 已提交
947 948
	if (group_dead)
		tty_audit_exit();
949
	audit_free(tsk);
950

951
	tsk->exit_code = code;
952
	taskstats_exit(tsk, group_dead);
953

L
Linus Torvalds 已提交
954 955
	exit_mm(tsk);

956
	if (group_dead)
957
		acct_process();
958 959
	trace_sched_process_exit(tsk);

L
Linus Torvalds 已提交
960
	exit_sem(tsk);
961
	exit_shm(tsk);
962 963
	exit_files(tsk);
	exit_fs(tsk);
964
	check_stack_usage();
L
Linus Torvalds 已提交
965
	exit_thread();
966 967 968 969 970 971 972 973 974

	/*
	 * 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);

975
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
976

977
	if (group_dead)
L
Linus Torvalds 已提交
978 979
		disassociate_ctty(1);

A
Al Viro 已提交
980
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
981

M
Matt Helsley 已提交
982
	proc_exit_connector(tsk);
983

984 985 986
	/*
	 * FIXME: do that only when needed, using sched_exit tracepoint
	 */
987
	ptrace_put_breakpoints(tsk);
988

989
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
990
#ifdef CONFIG_NUMA
991
	task_lock(tsk);
992
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
993
	tsk->mempolicy = NULL;
994
	task_unlock(tsk);
L
Linus Torvalds 已提交
995
#endif
996
#ifdef CONFIG_FUTEX
997 998
	if (unlikely(current->pi_state_cache))
		kfree(current->pi_state_cache);
999
#endif
1000
	/*
1001
	 * Make sure we are holding no locks:
1002
	 */
1003
	debug_check_no_locks_held(tsk);
1004 1005 1006 1007 1008 1009
	/*
	 * 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
1010

1011
	if (tsk->io_context)
1012
		exit_io_context(tsk);
1013

1014 1015 1016
	if (tsk->splice_pipe)
		__free_pipe_info(tsk->splice_pipe);

1017 1018
	validate_creds_for_do_exit(tsk);

C
Coywolf Qi Hunt 已提交
1019
	preempt_disable();
1020 1021
	if (tsk->nr_dirtied)
		__this_cpu_add(dirty_throttle_leaks, tsk->nr_dirtied);
1022
	exit_rcu();
1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038

	/*
	 * 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
1039
	/* causes final put_task_struct in finish_task_switch(). */
1040
	tsk->state = TASK_DEAD;
1041
	tsk->flags |= PF_NOFREEZE;	/* tell freezer to ignore us */
L
Linus Torvalds 已提交
1042 1043 1044
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
1045 1046
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
1047 1048
}

1049 1050
EXPORT_SYMBOL_GPL(do_exit);

1051
void complete_and_exit(struct completion *comp, long code)
L
Linus Torvalds 已提交
1052 1053 1054
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
1055

L
Linus Torvalds 已提交
1056 1057 1058 1059 1060
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

1061
SYSCALL_DEFINE1(exit, int, error_code)
L
Linus Torvalds 已提交
1062 1063 1064 1065 1066 1067 1068 1069
{
	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).
 */
1070
void
L
Linus Torvalds 已提交
1071 1072
do_group_exit(int exit_code)
{
1073 1074
	struct signal_struct *sig = current->signal;

L
Linus Torvalds 已提交
1075 1076
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

1077 1078
	if (signal_group_exit(sig))
		exit_code = sig->group_exit_code;
L
Linus Torvalds 已提交
1079 1080 1081
	else if (!thread_group_empty(current)) {
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1082
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1083 1084 1085 1086
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1087
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101
			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.
 */
1102
SYSCALL_DEFINE1(exit_group, int, error_code)
L
Linus Torvalds 已提交
1103 1104
{
	do_group_exit((error_code & 0xff) << 8);
1105 1106
	/* NOTREACHED */
	return 0;
L
Linus Torvalds 已提交
1107 1108
}

1109 1110 1111
struct wait_opts {
	enum pid_type		wo_type;
	int			wo_flags;
1112
	struct pid		*wo_pid;
1113 1114 1115 1116 1117

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

1118
	wait_queue_t		child_wait;
1119 1120 1121
	int			notask_error;
};

1122 1123
static inline
struct pid *task_pid_type(struct task_struct *task, enum pid_type type)
E
Eric W. Biederman 已提交
1124
{
1125 1126 1127
	if (type != PIDTYPE_PID)
		task = task->group_leader;
	return task->pids[type].pid;
E
Eric W. Biederman 已提交
1128 1129
}

1130
static int eligible_pid(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1131
{
1132 1133 1134
	return	wo->wo_type == PIDTYPE_MAX ||
		task_pid_type(p, wo->wo_type) == wo->wo_pid;
}
L
Linus Torvalds 已提交
1135

1136 1137 1138 1139
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
1140 1141 1142 1143 1144
	/* 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.) */
1145 1146
	if (((p->exit_signal != SIGCHLD) ^ !!(wo->wo_flags & __WCLONE))
	    && !(wo->wo_flags & __WALL))
L
Linus Torvalds 已提交
1147 1148
		return 0;

1149
	return 1;
L
Linus Torvalds 已提交
1150 1151
}

1152 1153
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
1154
{
1155 1156 1157
	struct siginfo __user *infop;
	int retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
1158

L
Linus Torvalds 已提交
1159
	put_task_struct(p);
1160
	infop = wo->wo_info;
1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174
	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
1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185
	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.
 */
1186
static int wait_task_zombie(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1187 1188
{
	unsigned long state;
1189
	int retval, status, traced;
1190
	pid_t pid = task_pid_vnr(p);
1191
	uid_t uid = __task_cred(p)->uid;
1192
	struct siginfo __user *infop;
L
Linus Torvalds 已提交
1193

1194
	if (!likely(wo->wo_flags & WEXITED))
R
Roland McGrath 已提交
1195 1196
		return 0;

1197
	if (unlikely(wo->wo_flags & WNOWAIT)) {
L
Linus Torvalds 已提交
1198
		int exit_code = p->exit_code;
1199
		int why;
L
Linus Torvalds 已提交
1200 1201 1202 1203 1204 1205 1206 1207 1208 1209

		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;
		}
1210
		return wait_noreap_copyout(wo, p, pid, uid, why, status);
L
Linus Torvalds 已提交
1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222
	}

	/*
	 * 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;
	}

1223
	traced = ptrace_reparented(p);
1224 1225
	/*
	 * It can be ptraced but not reparented, check
O
Oleg Nesterov 已提交
1226
	 * thread_group_leader() to filter out sub-threads.
1227
	 */
O
Oleg Nesterov 已提交
1228
	if (likely(!traced) && thread_group_leader(p)) {
1229 1230
		struct signal_struct *psig;
		struct signal_struct *sig;
J
Jiri Pirko 已提交
1231
		unsigned long maxrss;
1232
		cputime_t tgutime, tgstime;
1233

L
Linus Torvalds 已提交
1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244
		/*
		 * 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
1245
		 * need to protect the access to parent->signal fields,
L
Linus Torvalds 已提交
1246 1247
		 * as other threads in the parent group can be right
		 * here reaping other children at the same time.
1248 1249 1250 1251
		 *
		 * 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
1252
		 */
1253
		thread_group_times(p, &tgutime, &tgstime);
1254 1255
		spin_lock_irq(&p->real_parent->sighand->siglock);
		psig = p->real_parent->signal;
1256
		sig = p->signal;
1257 1258 1259
		psig->cutime += tgutime + sig->cutime;
		psig->cstime += tgstime + sig->cstime;
		psig->cgtime += p->gtime + sig->gtime + sig->cgtime;
1260 1261 1262 1263 1264 1265 1266 1267
		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;
1268 1269 1270 1271 1272 1273
		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
1274 1275 1276
		maxrss = max(sig->maxrss, sig->cmaxrss);
		if (psig->cmaxrss < maxrss)
			psig->cmaxrss = maxrss;
1277 1278
		task_io_accounting_add(&psig->ioac, &p->ioac);
		task_io_accounting_add(&psig->ioac, &sig->ioac);
1279
		spin_unlock_irq(&p->real_parent->sighand->siglock);
L
Linus Torvalds 已提交
1280 1281 1282 1283 1284 1285 1286 1287
	}

	/*
	 * 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);

1288 1289
	retval = wo->wo_rusage
		? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
L
Linus Torvalds 已提交
1290 1291
	status = (p->signal->flags & SIGNAL_GROUP_EXIT)
		? p->signal->group_exit_code : p->exit_code;
1292 1293 1294 1295
	if (!retval && wo->wo_stat)
		retval = put_user(status, wo->wo_stat);

	infop = wo->wo_info;
L
Linus Torvalds 已提交
1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314
	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)
1315
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1316
	if (!retval && infop)
1317
		retval = put_user(uid, &infop->si_uid);
1318
	if (!retval)
1319
		retval = pid;
1320 1321

	if (traced) {
L
Linus Torvalds 已提交
1322
		write_lock_irq(&tasklist_lock);
1323 1324 1325
		/* We dropped tasklist, ptracer could die and untrace */
		ptrace_unlink(p);
		/*
1326 1327
		 * If this is not a sub-thread, notify the parent.
		 * If parent wants a zombie, don't release it now.
1328
		 */
1329 1330 1331 1332
		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
1333 1334 1335 1336 1337
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1338

L
Linus Torvalds 已提交
1339 1340 1341
	return retval;
}

1342 1343 1344
static int *task_stopped_code(struct task_struct *p, bool ptrace)
{
	if (ptrace) {
T
Tejun Heo 已提交
1345 1346
		if (task_is_stopped_or_traced(p) &&
		    !(p->jobctl & JOBCTL_LISTENING))
1347 1348 1349 1350 1351 1352 1353 1354
			return &p->exit_code;
	} else {
		if (p->signal->flags & SIGNAL_STOP_STOPPED)
			return &p->signal->group_exit_code;
	}
	return NULL;
}

1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
/**
 * 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
1372
 */
1373 1374
static int wait_task_stopped(struct wait_opts *wo,
				int ptrace, struct task_struct *p)
L
Linus Torvalds 已提交
1375
{
1376
	struct siginfo __user *infop;
1377
	int retval, exit_code, *p_code, why;
1378
	uid_t uid = 0; /* unneeded, required by compiler */
1379
	pid_t pid;
L
Linus Torvalds 已提交
1380

1381 1382 1383
	/*
	 * Traditionally we see ptrace'd stopped tasks regardless of options.
	 */
1384
	if (!ptrace && !(wo->wo_flags & WUNTRACED))
R
Roland McGrath 已提交
1385 1386
		return 0;

1387 1388 1389
	if (!task_stopped_code(p, ptrace))
		return 0;

1390 1391 1392
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

1393 1394
	p_code = task_stopped_code(p, ptrace);
	if (unlikely(!p_code))
1395 1396
		goto unlock_sig;

1397
	exit_code = *p_code;
1398 1399 1400
	if (!exit_code)
		goto unlock_sig;

1401
	if (!unlikely(wo->wo_flags & WNOWAIT))
1402
		*p_code = 0;
1403

1404
	uid = task_uid(p);
1405 1406 1407
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1408 1409 1410 1411 1412 1413 1414 1415 1416 1417
		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);
1418
	pid = task_pid_vnr(p);
R
Roland McGrath 已提交
1419
	why = ptrace ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1420 1421
	read_unlock(&tasklist_lock);

1422 1423 1424 1425 1426 1427 1428
	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
1429

1430
	infop = wo->wo_info;
L
Linus Torvalds 已提交
1431 1432 1433 1434 1435
	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
1436
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1437 1438 1439
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1440
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1441
	if (!retval && infop)
1442
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1443
	if (!retval)
1444
		retval = pid;
L
Linus Torvalds 已提交
1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
	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.
 */
1457
static int wait_task_continued(struct wait_opts *wo, struct task_struct *p)
L
Linus Torvalds 已提交
1458 1459 1460 1461 1462
{
	int retval;
	pid_t pid;
	uid_t uid;

1463
	if (!unlikely(wo->wo_flags & WCONTINUED))
R
Roland McGrath 已提交
1464 1465
		return 0;

L
Linus Torvalds 已提交
1466 1467 1468 1469 1470 1471 1472 1473 1474
	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;
	}
1475
	if (!unlikely(wo->wo_flags & WNOWAIT))
L
Linus Torvalds 已提交
1476
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
1477
	uid = task_uid(p);
L
Linus Torvalds 已提交
1478 1479
	spin_unlock_irq(&p->sighand->siglock);

1480
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1481 1482 1483
	get_task_struct(p);
	read_unlock(&tasklist_lock);

1484 1485 1486
	if (!wo->wo_info) {
		retval = wo->wo_rusage
			? getrusage(p, RUSAGE_BOTH, wo->wo_rusage) : 0;
L
Linus Torvalds 已提交
1487
		put_task_struct(p);
1488 1489
		if (!retval && wo->wo_stat)
			retval = put_user(0xffff, wo->wo_stat);
L
Linus Torvalds 已提交
1490
		if (!retval)
1491
			retval = pid;
L
Linus Torvalds 已提交
1492
	} else {
1493 1494
		retval = wait_noreap_copyout(wo, p, pid, uid,
					     CLD_CONTINUED, SIGCONT);
L
Linus Torvalds 已提交
1495 1496 1497 1498 1499 1500
		BUG_ON(retval == 0);
	}

	return retval;
}

R
Roland McGrath 已提交
1501 1502 1503
/*
 * Consider @p for a wait by @parent.
 *
1504
 * -ECHILD should be in ->notask_error before the first call.
R
Roland McGrath 已提交
1505 1506
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue;
1507
 * then ->notask_error is 0 if @p is an eligible child,
1508
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1509
 */
1510 1511
static int wait_consider_task(struct wait_opts *wo, int ptrace,
				struct task_struct *p)
R
Roland McGrath 已提交
1512
{
1513
	int ret = eligible_child(wo, p);
1514
	if (!ret)
R
Roland McGrath 已提交
1515 1516
		return ret;

1517
	ret = security_task_wait(p);
1518 1519 1520 1521 1522 1523 1524 1525
	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.
		 */
1526 1527
		if (wo->notask_error)
			wo->notask_error = ret;
1528
		return 0;
1529 1530
	}

1531
	/* dead body doesn't have much to contribute */
1532 1533 1534 1535 1536 1537 1538
	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;
1539
		return 0;
1540
	}
1541

1542 1543
	/* slay zombie? */
	if (p->exit_state == EXIT_ZOMBIE) {
R
Roland McGrath 已提交
1544
		/*
1545 1546 1547
		 * 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
1548
		 */
T
Tejun Heo 已提交
1549
		if (likely(!ptrace) && unlikely(p->ptrace)) {
1550 1551 1552 1553
			/* it will become visible, clear notask_error */
			wo->notask_error = 0;
			return 0;
		}
R
Roland McGrath 已提交
1554

1555 1556 1557
		/* 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
1558

R
Roland McGrath 已提交
1559
		/*
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581
		 * 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 {
1582 1583 1584 1585 1586 1587 1588 1589 1590 1591
		/*
		 * 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.
		 */
1592
		if (likely(!ptrace) && p->ptrace && !ptrace_reparented(p))
1593 1594
			return 0;

1595 1596 1597
		/*
		 * @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
1598
		 */
1599
		wo->notask_error = 0;
R
Roland McGrath 已提交
1600 1601
	}

R
Roland McGrath 已提交
1602
	/*
1603 1604
	 * 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
1605
	 */
1606 1607 1608
	ret = wait_task_stopped(wo, ptrace, p);
	if (ret)
		return ret;
R
Roland McGrath 已提交
1609 1610

	/*
1611 1612 1613
	 * 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
1614
	 */
1615
	return wait_task_continued(wo, p);
R
Roland McGrath 已提交
1616 1617 1618 1619 1620
}

/*
 * Do the work of do_wait() for one thread in the group, @tsk.
 *
1621
 * -ECHILD should be in ->notask_error before the first call.
R
Roland McGrath 已提交
1622 1623
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue; then
1624
 * ->notask_error is 0 if there were any eligible children,
1625
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1626
 */
1627
static int do_wait_thread(struct wait_opts *wo, struct task_struct *tsk)
R
Roland McGrath 已提交
1628 1629 1630 1631
{
	struct task_struct *p;

	list_for_each_entry(p, &tsk->children, sibling) {
1632 1633 1634
		int ret = wait_consider_task(wo, 0, p);
		if (ret)
			return ret;
R
Roland McGrath 已提交
1635 1636 1637 1638 1639
	}

	return 0;
}

1640
static int ptrace_do_wait(struct wait_opts *wo, struct task_struct *tsk)
R
Roland McGrath 已提交
1641 1642 1643
{
	struct task_struct *p;

R
Roland McGrath 已提交
1644
	list_for_each_entry(p, &tsk->ptraced, ptrace_entry) {
1645
		int ret = wait_consider_task(wo, 1, p);
R
Roland McGrath 已提交
1646
		if (ret)
R
Roland McGrath 已提交
1647 1648 1649 1650 1651 1652
			return ret;
	}

	return 0;
}

1653 1654 1655 1656 1657 1658 1659
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;

1660
	if (!eligible_pid(wo, p))
1661 1662
		return 0;

1663 1664 1665
	if ((wo->wo_flags & __WNOTHREAD) && wait->private != p->parent)
		return 0;

1666 1667 1668
	return default_wake_function(wait, mode, sync, key);
}

1669 1670
void __wake_up_parent(struct task_struct *p, struct task_struct *parent)
{
1671 1672
	__wake_up_sync_key(&parent->signal->wait_chldexit,
				TASK_INTERRUPTIBLE, 1, p);
1673 1674
}

1675
static long do_wait(struct wait_opts *wo)
L
Linus Torvalds 已提交
1676 1677
{
	struct task_struct *tsk;
R
Roland McGrath 已提交
1678
	int retval;
L
Linus Torvalds 已提交
1679

1680
	trace_sched_process_wait(wo->wo_pid);
1681

1682 1683 1684
	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
1685
repeat:
R
Roland McGrath 已提交
1686 1687
	/*
	 * If there is nothing that can match our critiera just get out.
1688 1689 1690
	 * 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
1691
	 */
1692
	wo->notask_error = -ECHILD;
1693 1694
	if ((wo->wo_type < PIDTYPE_MAX) &&
	   (!wo->wo_pid || hlist_empty(&wo->wo_pid->tasks[wo->wo_type])))
1695
		goto notask;
E
Eric W. Biederman 已提交
1696

1697
	set_current_state(TASK_INTERRUPTIBLE);
L
Linus Torvalds 已提交
1698 1699 1700
	read_lock(&tasklist_lock);
	tsk = current;
	do {
1701 1702 1703
		retval = do_wait_thread(wo, tsk);
		if (retval)
			goto end;
1704

1705 1706
		retval = ptrace_do_wait(wo, tsk);
		if (retval)
R
Roland McGrath 已提交
1707 1708
			goto end;

1709
		if (wo->wo_flags & __WNOTHREAD)
L
Linus Torvalds 已提交
1710
			break;
1711
	} while_each_thread(current, tsk);
L
Linus Torvalds 已提交
1712
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1713

1714
notask:
1715 1716
	retval = wo->notask_error;
	if (!retval && !(wo->wo_flags & WNOHANG)) {
L
Linus Torvalds 已提交
1717
		retval = -ERESTARTSYS;
R
Roland McGrath 已提交
1718 1719 1720 1721
		if (!signal_pending(current)) {
			schedule();
			goto repeat;
		}
L
Linus Torvalds 已提交
1722 1723
	}
end:
1724
	__set_current_state(TASK_RUNNING);
1725
	remove_wait_queue(&current->signal->wait_chldexit, &wo->child_wait);
L
Linus Torvalds 已提交
1726 1727 1728
	return retval;
}

1729 1730
SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *,
		infop, int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1731
{
1732
	struct wait_opts wo;
E
Eric W. Biederman 已提交
1733 1734
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1735 1736 1737 1738 1739 1740 1741 1742 1743
	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
1744
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1745 1746
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1747 1748
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1749 1750 1751
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1752 1753
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1754 1755 1756 1757 1758 1759
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1760 1761
	if (type < PIDTYPE_MAX)
		pid = find_get_pid(upid);
1762 1763 1764 1765 1766 1767 1768 1769

	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);
1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792

	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
1793
	put_pid(pid);
L
Linus Torvalds 已提交
1794 1795

	/* avoid REGPARM breakage on x86: */
1796
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1797 1798 1799
	return ret;
}

1800 1801
SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr,
		int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1802
{
1803
	struct wait_opts wo;
E
Eric W. Biederman 已提交
1804 1805
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1806 1807 1808 1809 1810
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1811 1812 1813 1814 1815 1816 1817 1818

	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;
1819
		pid = get_task_pid(current, PIDTYPE_PGID);
E
Eric W. Biederman 已提交
1820 1821 1822 1823 1824
	} else /* upid > 0 */ {
		type = PIDTYPE_PID;
		pid = find_get_pid(upid);
	}

1825 1826 1827 1828 1829 1830 1831
	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
1832
	put_pid(pid);
L
Linus Torvalds 已提交
1833 1834

	/* avoid REGPARM breakage on x86: */
1835
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1836 1837 1838 1839 1840 1841 1842 1843 1844
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
1845
SYSCALL_DEFINE3(waitpid, pid_t, pid, int __user *, stat_addr, int, options)
L
Linus Torvalds 已提交
1846 1847 1848 1849 1850
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif