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

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

#include <asm/uaccess.h>
#include <asm/unistd.h>
#include <asm/pgtable.h>
#include <asm/mmu_context.h>
D
David Howells 已提交
56
#include "cred-internals.h"
L
Linus Torvalds 已提交
57

58 59 60
DEFINE_TRACE(sched_process_free);
DEFINE_TRACE(sched_process_exit);
DEFINE_TRACE(sched_process_wait);
L
Linus Torvalds 已提交
61

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

64 65 66 67 68
static inline int task_detached(struct task_struct *p)
{
	return p->exit_signal == -1;
}

L
Linus Torvalds 已提交
69 70 71 72 73 74 75
static void __unhash_process(struct task_struct *p)
{
	nr_threads--;
	detach_pid(p, PIDTYPE_PID);
	if (thread_group_leader(p)) {
		detach_pid(p, PIDTYPE_PGID);
		detach_pid(p, PIDTYPE_SID);
76

77
		list_del_rcu(&p->tasks);
78
		__get_cpu_var(process_counts)--;
L
Linus Torvalds 已提交
79
	}
O
Oleg Nesterov 已提交
80
	list_del_rcu(&p->thread_group);
R
Roland McGrath 已提交
81
	list_del_init(&p->sibling);
L
Linus Torvalds 已提交
82 83
}

84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
/*
 * This function expects the tasklist_lock write-locked.
 */
static void __exit_signal(struct task_struct *tsk)
{
	struct signal_struct *sig = tsk->signal;
	struct sighand_struct *sighand;

	BUG_ON(!sig);
	BUG_ON(!atomic_read(&sig->count));

	sighand = rcu_dereference(tsk->sighand);
	spin_lock(&sighand->siglock);

	posix_cpu_timers_exit(tsk);
	if (atomic_dec_and_test(&sig->count))
		posix_cpu_timers_exit_group(tsk);
	else {
		/*
		 * If there is any task waiting for the group exit
		 * then notify it:
		 */
106
		if (sig->group_exit_task && atomic_read(&sig->count) == sig->notify_count)
107
			wake_up_process(sig->group_exit_task);
108

109 110 111 112 113 114 115 116 117 118 119 120
		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.
		 */
121 122
		sig->utime = cputime_add(sig->utime, task_utime(tsk));
		sig->stime = cputime_add(sig->stime, task_stime(tsk));
123
		sig->gtime = cputime_add(sig->gtime, task_gtime(tsk));
124 125 126 127
		sig->min_flt += tsk->min_flt;
		sig->maj_flt += tsk->maj_flt;
		sig->nvcsw += tsk->nvcsw;
		sig->nivcsw += tsk->nivcsw;
128 129
		sig->inblock += task_io_get_inblock(tsk);
		sig->oublock += task_io_get_oublock(tsk);
130
		task_io_accounting_add(&sig->ioac, &tsk->ioac);
131
		sig->sum_sched_runtime += tsk->se.sum_exec_runtime;
132 133 134
		sig = NULL; /* Marker for below. */
	}

135 136
	__unhash_process(tsk);

137 138 139 140 141 142
	/*
	 * Do this under ->siglock, we can race with another thread
	 * doing sigqueue_free() if we have SIGQUEUE_PREALLOC signals.
	 */
	flush_sigqueue(&tsk->pending);

143
	tsk->signal = NULL;
144
	tsk->sighand = NULL;
145 146
	spin_unlock(&sighand->siglock);

147
	__cleanup_sighand(sighand);
148 149 150
	clear_tsk_thread_flag(tsk,TIF_SIGPENDING);
	if (sig) {
		flush_sigqueue(&sig->shared_pending);
151
		taskstats_tgid_free(sig);
152 153 154 155 156
		/*
		 * Make sure ->signal can't go away under rq->lock,
		 * see account_group_exec_runtime().
		 */
		task_rq_unlock_wait(tsk);
157 158 159 160
		__cleanup_signal(sig);
	}
}

161 162
static void delayed_put_task_struct(struct rcu_head *rhp)
{
163 164 165 166
	struct task_struct *tsk = container_of(rhp, struct task_struct, rcu);

	trace_sched_process_free(tsk);
	put_task_struct(tsk);
167 168
}

R
Roland McGrath 已提交
169

L
Linus Torvalds 已提交
170 171
void release_task(struct task_struct * p)
{
172
	struct task_struct *leader;
L
Linus Torvalds 已提交
173
	int zap_leader;
174
repeat:
R
Roland McGrath 已提交
175
	tracehook_prepare_release_task(p);
176 177 178 179
	/* don't need to get the RCU readlock here - the process is dead and
	 * can't be modifying its own credentials */
	atomic_dec(&__task_cred(p)->user->processes);

180
	proc_flush_task(p);
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;
365
	pid_t nr = pid_nr(pid);
L
Linus Torvalds 已提交
366

367
	if (task_session(curr) != pid) {
368
		change_pid(curr, PIDTYPE_SID, pid);
369
		set_task_session(curr, nr);
L
Linus Torvalds 已提交
370
	}
371
	if (task_pgrp(curr) != pid) {
372
		change_pid(curr, PIDTYPE_PGID, pid);
373
		set_task_pgrp(curr, nr);
L
Linus Torvalds 已提交
374 375 376
	}
}

377
static void set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
378 379
{
	write_lock_irq(&tasklist_lock);
380
	__set_special_pids(pid);
L
Linus Torvalds 已提交
381 382 383 384 385 386 387 388 389 390
	write_unlock_irq(&tasklist_lock);
}

/*
 * Let kernel threads use this to say that they
 * allow a certain signal (since daemonize() will
 * have disabled all of them by default).
 */
int allow_signal(int sig)
{
391
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
	sigdelset(&current->blocked, sig);
	if (!current->mm) {
		/* 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;
	}
	recalc_sigpending();
	spin_unlock_irq(&current->sighand->siglock);
	return 0;
}

EXPORT_SYMBOL(allow_signal);

int disallow_signal(int sig)
{
412
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
413 414 415
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
416
	current->sighand->action[(sig)-1].sa.sa_handler = SIG_IGN;
L
Linus Torvalds 已提交
417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444
	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;
	struct fs_struct *fs;
	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);
445 446 447 448
	/*
	 * 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
449
	current->flags |= (PF_NOFREEZE | PF_KTHREAD);
L
Linus Torvalds 已提交
450

451 452 453 454
	if (current->nsproxy != &init_nsproxy) {
		get_nsproxy(&init_nsproxy);
		switch_task_namespaces(current, &init_nsproxy);
	}
455
	set_special_pids(&init_struct_pid);
456
	proc_clear_tty(current);
L
Linus Torvalds 已提交
457 458 459 460 461 462 463 464 465 466 467 468

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

	/* Become as one with the init task */

	exit_fs(current);	/* current->fs->count--; */
	fs = init_task.fs;
	current->fs = fs;
	atomic_inc(&fs->count);
S
Serge E. Hallyn 已提交
469

470
	exit_files(current);
L
Linus Torvalds 已提交
471 472 473
	current->files = init_task.files;
	atomic_inc(&current->files->count);

474
	reparent_to_kthreadd();
L
Linus Torvalds 已提交
475 476 477 478
}

EXPORT_SYMBOL(daemonize);

479
static void close_files(struct files_struct * files)
L
Linus Torvalds 已提交
480 481
{
	int i, j;
482
	struct fdtable *fdt;
L
Linus Torvalds 已提交
483 484

	j = 0;
485 486 487 488 489 490

	/*
	 * It is safe to dereference the fd table without RCU or
	 * ->file_lock because this is the last reference to the
	 * files structure.
	 */
491
	fdt = files_fdtable(files);
L
Linus Torvalds 已提交
492 493 494
	for (;;) {
		unsigned long set;
		i = j * __NFDBITS;
495
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
496
			break;
497
		set = fdt->open_fds->fds_bits[j++];
L
Linus Torvalds 已提交
498 499
		while (set) {
			if (set & 1) {
500
				struct file * file = xchg(&fdt->fd[i], NULL);
501
				if (file) {
L
Linus Torvalds 已提交
502
					filp_close(file, files);
503 504
					cond_resched();
				}
L
Linus Torvalds 已提交
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524
			}
			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;
}

525
void put_files_struct(struct files_struct *files)
L
Linus Torvalds 已提交
526
{
527 528
	struct fdtable *fdt;

L
Linus Torvalds 已提交
529 530 531 532
	if (atomic_dec_and_test(&files->count)) {
		close_files(files);
		/*
		 * Free the fd and fdset arrays if we expanded them.
533 534 535
		 * 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
536
		 */
537
		fdt = files_fdtable(files);
538
		if (fdt != &files->fdtab)
539
			kmem_cache_free(files_cachep, files);
540
		free_fdtable(fdt);
L
Linus Torvalds 已提交
541 542 543
	}
}

544
void reset_files_struct(struct files_struct *files)
545
{
546
	struct task_struct *tsk = current;
547 548 549 550 551 552 553 554 555
	struct files_struct *old;

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

556
void exit_files(struct task_struct *tsk)
L
Linus Torvalds 已提交
557 558 559 560 561 562 563 564 565 566 567
{
	struct files_struct * files = tsk->files;

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

568
void put_fs_struct(struct fs_struct *fs)
L
Linus Torvalds 已提交
569 570 571
{
	/* No need to hold fs->lock if we are killing it */
	if (atomic_dec_and_test(&fs->count)) {
J
Jan Blunck 已提交
572 573
		path_put(&fs->root);
		path_put(&fs->pwd);
L
Linus Torvalds 已提交
574 575 576 577
		kmem_cache_free(fs_cachep, fs);
	}
}

578
void exit_fs(struct task_struct *tsk)
L
Linus Torvalds 已提交
579 580 581 582 583 584 585
{
	struct fs_struct * fs = tsk->fs;

	if (fs) {
		task_lock(tsk);
		tsk->fs = NULL;
		task_unlock(tsk);
586
		put_fs_struct(fs);
L
Linus Torvalds 已提交
587 588 589 590 591
	}
}

EXPORT_SYMBOL_GPL(exit_fs);

592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644
#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
	 */
	list_for_each_entry(c, &p->parent->children, sibling) {
		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);
645 646 647
	/*
	 * We found no owner yet mm_users > 1: this implies that we are
	 * most likely racing with swapoff (try_to_unuse()) or /proc or
648
	 * ptrace or page migration (get_task_mm()).  Mark owner as NULL.
649 650
	 */
	mm->owner = NULL;
651 652 653 654 655 656 657 658 659 660
	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);
661 662 663 664 665
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
666 667 668 669 670 671 672 673 674 675 676
	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
677 678 679 680
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
681
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
682 683
{
	struct mm_struct *mm = tsk->mm;
684
	struct core_state *core_state;
L
Linus Torvalds 已提交
685 686 687 688 689 690

	mm_release(tsk, mm);
	if (!mm)
		return;
	/*
	 * Serialize with any possible pending coredump.
691
	 * We must hold mmap_sem around checking core_state
L
Linus Torvalds 已提交
692
	 * and clearing tsk->mm.  The core-inducing thread
693
	 * will increment ->nr_threads for each thread in the
L
Linus Torvalds 已提交
694 695 696
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
697 698 699
	core_state = mm->core_state;
	if (core_state) {
		struct core_thread self;
L
Linus Torvalds 已提交
700 701
		up_read(&mm->mmap_sem);

702 703 704 705 706 707 708 709
		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
710

O
Oleg Nesterov 已提交
711 712 713 714 715 716 717
		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
718 719 720
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
721
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
722 723 724 725 726
	/* more a memory barrier than a real lock */
	task_lock(tsk);
	tsk->mm = NULL;
	up_read(&mm->mmap_sem);
	enter_lazy_tlb(mm, current);
727 728
	/* We don't want this task to be frozen prematurely */
	clear_freeze_flag(tsk);
L
Linus Torvalds 已提交
729
	task_unlock(tsk);
730
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
731 732 733
	mmput(mm);
}

734 735 736 737 738 739
/*
 * Return nonzero if @parent's children should reap themselves.
 *
 * Called with write_lock_irq(&tasklist_lock) held.
 */
static int ignoring_children(struct task_struct *parent)
L
Linus Torvalds 已提交
740
{
741 742 743 744 745 746 747 748 749
	int ret;
	struct sighand_struct *psig = parent->sighand;
	unsigned long flags;
	spin_lock_irqsave(&psig->siglock, flags);
	ret = (psig->action[SIGCHLD-1].sa.sa_handler == SIG_IGN ||
	       (psig->action[SIGCHLD-1].sa.sa_flags & SA_NOCLDWAIT));
	spin_unlock_irqrestore(&psig->siglock, flags);
	return ret;
}
750

R
Roland McGrath 已提交
751 752 753 754 755 756 757
/*
 * Detach all tasks we were using ptrace on.
 * Any that need to be release_task'd are put on the @dead list.
 *
 * Called with write_lock(&tasklist_lock) held.
 */
static void ptrace_exit(struct task_struct *parent, struct list_head *dead)
L
Linus Torvalds 已提交
758
{
R
Roland McGrath 已提交
759
	struct task_struct *p, *n;
760
	int ign = -1;
761

R
Roland McGrath 已提交
762 763 764 765 766 767 768 769 770 771 772 773 774 775
	list_for_each_entry_safe(p, n, &parent->ptraced, ptrace_entry) {
		__ptrace_unlink(p);

		if (p->exit_state != EXIT_ZOMBIE)
			continue;

		/*
		 * If it's a zombie, our attachedness prevented normal
		 * parent notification or self-reaping.  Do notification
		 * now if it would have happened earlier.  If it should
		 * reap itself, add it to the @dead list.  We can't call
		 * release_task() here because we already hold tasklist_lock.
		 *
		 * If it's our own child, there is no notification to do.
776 777
		 * But if our normal children self-reap, then this child
		 * was prevented by ptrace and we must reap it now.
L
Linus Torvalds 已提交
778
		 */
R
Roland McGrath 已提交
779 780 781
		if (!task_detached(p) && thread_group_empty(p)) {
			if (!same_thread_group(p->real_parent, parent))
				do_notify_parent(p, p->exit_signal);
782 783 784 785 786 787
			else {
				if (ign < 0)
					ign = ignoring_children(parent);
				if (ign)
					p->exit_signal = -1;
			}
R
Roland McGrath 已提交
788
		}
L
Linus Torvalds 已提交
789

R
Roland McGrath 已提交
790
		if (task_detached(p)) {
L
Linus Torvalds 已提交
791
			/*
R
Roland McGrath 已提交
792
			 * Mark it as in the process of being reaped.
L
Linus Torvalds 已提交
793
			 */
R
Roland McGrath 已提交
794 795
			p->exit_state = EXIT_DEAD;
			list_add(&p->ptrace_entry, dead);
L
Linus Torvalds 已提交
796 797
		}
	}
R
Roland McGrath 已提交
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
}

/*
 * Finish up exit-time ptrace cleanup.
 *
 * Called without locks.
 */
static void ptrace_exit_finish(struct task_struct *parent,
			       struct list_head *dead)
{
	struct task_struct *p, *n;

	BUG_ON(!list_empty(&parent->ptraced));

	list_for_each_entry_safe(p, n, dead, ptrace_entry) {
		list_del_init(&p->ptrace_entry);
		release_task(p);
	}
}

static void reparent_thread(struct task_struct *p, struct task_struct *father)
{
	if (p->pdeath_signal)
		/* We already hold the tasklist_lock here.  */
		group_send_sig_info(p->pdeath_signal, SEND_SIG_NOINFO, p);

	list_move_tail(&p->sibling, &p->real_parent->children);
L
Linus Torvalds 已提交
825

826 827 828
	/* If this is a threaded reparent there is no need to
	 * notify anyone anything has happened.
	 */
829
	if (same_thread_group(p->real_parent, father))
830 831 832
		return;

	/* We don't want people slaying init.  */
833
	if (!task_detached(p))
834 835 836 837 838
		p->exit_signal = SIGCHLD;

	/* If we'd notified the old parent about this child's death,
	 * also notify the new parent.
	 */
R
Roland McGrath 已提交
839 840
	if (!ptrace_reparented(p) &&
	    p->exit_state == EXIT_ZOMBIE &&
841
	    !task_detached(p) && thread_group_empty(p))
842 843
		do_notify_parent(p, p->exit_signal);

844
	kill_orphaned_pgrp(p, father);
L
Linus Torvalds 已提交
845 846 847 848 849 850
}

/*
 * 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
851 852
 * the child reaper process (ie "init") in our pid
 * space.
L
Linus Torvalds 已提交
853
 */
854
static struct task_struct *find_new_reaper(struct task_struct *father)
L
Linus Torvalds 已提交
855
{
856 857
	struct pid_namespace *pid_ns = task_active_pid_ns(father);
	struct task_struct *thread;
L
Linus Torvalds 已提交
858

859 860 861 862 863 864 865 866
	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
867

868 869 870 871
	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
872

873 874
		zap_pid_ns_processes(pid_ns);
		write_lock_irq(&tasklist_lock);
L
Linus Torvalds 已提交
875
		/*
876 877 878
		 * 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
879
		 */
880
		pid_ns->child_reaper = init_pid_ns.child_reaper;
L
Linus Torvalds 已提交
881
	}
882

883 884 885
	return pid_ns->child_reaper;
}

886
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
887
{
888
	struct task_struct *p, *n, *reaper;
R
Roland McGrath 已提交
889
	LIST_HEAD(ptrace_dead);
890 891

	write_lock_irq(&tasklist_lock);
892
	reaper = find_new_reaper(father);
R
Roland McGrath 已提交
893 894 895 896 897
	/*
	 * First clean up ptrace if we were using it.
	 */
	ptrace_exit(father, &ptrace_dead);

898
	list_for_each_entry_safe(p, n, &father->children, sibling) {
899
		p->real_parent = reaper;
R
Roland McGrath 已提交
900 901 902 903 904
		if (p->parent == father) {
			BUG_ON(p->ptrace);
			p->parent = p->real_parent;
		}
		reparent_thread(p, father);
L
Linus Torvalds 已提交
905
	}
906 907 908 909

	write_unlock_irq(&tasklist_lock);
	BUG_ON(!list_empty(&father->children));

R
Roland McGrath 已提交
910
	ptrace_exit_finish(father, &ptrace_dead);
L
Linus Torvalds 已提交
911 912 913 914 915 916
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
917
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
918
{
R
Roland McGrath 已提交
919 920
	int signal;
	void *cookie;
L
Linus Torvalds 已提交
921 922 923 924 925 926 927 928 929

	/*
	 * 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)
	 */
930
	forget_original_parent(tsk);
931
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
932

933
	write_lock_irq(&tasklist_lock);
934 935
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
936

937
	/* Let father know we died
L
Linus Torvalds 已提交
938 939
	 *
	 * Thread signals are configurable, but you aren't going to use
940
	 * that to send signals to arbitary processes.
L
Linus Torvalds 已提交
941 942 943 944 945 946 947 948 949 950
	 * 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.
	 */
951
	if (tsk->exit_signal != SIGCHLD && !task_detached(tsk) &&
952
	    (tsk->parent_exec_id != tsk->real_parent->self_exec_id ||
953 954
	     tsk->self_exec_id != tsk->parent_exec_id) &&
	    !capable(CAP_KILL))
L
Linus Torvalds 已提交
955 956
		tsk->exit_signal = SIGCHLD;

R
Roland McGrath 已提交
957
	signal = tracehook_notify_death(tsk, &cookie, group_dead);
958
	if (signal >= 0)
R
Roland McGrath 已提交
959
		signal = do_notify_parent(tsk, signal);
L
Linus Torvalds 已提交
960

961
	tsk->exit_state = signal == DEATH_REAP ? EXIT_DEAD : EXIT_ZOMBIE;
L
Linus Torvalds 已提交
962

963
	/* mt-exec, de_thread() is waiting for us */
964
	if (thread_group_leader(tsk) &&
965 966
	    tsk->signal->group_exit_task &&
	    tsk->signal->notify_count < 0)
967 968
		wake_up_process(tsk->signal->group_exit_task);

L
Linus Torvalds 已提交
969 970
	write_unlock_irq(&tasklist_lock);

R
Roland McGrath 已提交
971 972
	tracehook_report_death(tsk, signal, cookie, group_dead);

L
Linus Torvalds 已提交
973
	/* If the process is dead, release it - nobody will wait for it */
974
	if (signal == DEATH_REAP)
L
Linus Torvalds 已提交
975 976 977
		release_task(tsk);
}

978 979 980 981 982 983 984
#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;

985
	free = stack_not_used(current);
986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002

	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

1003
NORET_TYPE void do_exit(long code)
L
Linus Torvalds 已提交
1004 1005 1006 1007 1008 1009
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

1010 1011
	WARN_ON(atomic_read(&tsk->fs_excl));

L
Linus Torvalds 已提交
1012 1013 1014 1015 1016
	if (unlikely(in_interrupt()))
		panic("Aiee, killing interrupt handler!");
	if (unlikely(!tsk->pid))
		panic("Attempted to kill the idle task!");

R
Roland McGrath 已提交
1017
	tracehook_report_exit(&code);
L
Linus Torvalds 已提交
1018

1019 1020 1021 1022 1023 1024 1025
	/*
	 * 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");
1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
		/*
		 * 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;
1036 1037 1038 1039
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

O
Oleg Nesterov 已提交
1040
	exit_signals(tsk);  /* sets PF_EXITING */
1041 1042 1043 1044
	/*
	 * tsk->flags are checked in the futex code to protect against
	 * an exiting task cleaning up the robust pi futexes.
	 */
1045 1046
	smp_mb();
	spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
1047 1048 1049

	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
1050
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
1051 1052 1053
				preempt_count());

	acct_update_integrals(tsk);
1054

L
Linus Torvalds 已提交
1055
	group_dead = atomic_dec_and_test(&tsk->signal->live);
1056
	if (group_dead) {
1057
		hrtimer_cancel(&tsk->signal->real_timer);
1058
		exit_itimers(tsk->signal);
1059
	}
1060
	acct_collect(code, group_dead);
M
Miloslav Trmac 已提交
1061 1062
	if (group_dead)
		tty_audit_exit();
1063 1064
	if (unlikely(tsk->audit_context))
		audit_free(tsk);
1065

1066
	tsk->exit_code = code;
1067
	taskstats_exit(tsk, group_dead);
1068

L
Linus Torvalds 已提交
1069 1070
	exit_mm(tsk);

1071
	if (group_dead)
1072
		acct_process();
1073 1074
	trace_sched_process_exit(tsk);

L
Linus Torvalds 已提交
1075
	exit_sem(tsk);
1076 1077
	exit_files(tsk);
	exit_fs(tsk);
1078
	check_stack_usage();
L
Linus Torvalds 已提交
1079
	exit_thread();
1080
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
1081 1082 1083 1084

	if (group_dead && tsk->signal->leader)
		disassociate_ctty(1);

A
Al Viro 已提交
1085
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
1086 1087 1088
	if (tsk->binfmt)
		module_put(tsk->binfmt->module);

M
Matt Helsley 已提交
1089
	proc_exit_connector(tsk);
1090
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
1091
#ifdef CONFIG_NUMA
1092
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
1093 1094
	tsk->mempolicy = NULL;
#endif
1095
#ifdef CONFIG_FUTEX
1096 1097 1098 1099 1100 1101 1102 1103
	/*
	 * This must happen late, after the PID is not
	 * hashed anymore:
	 */
	if (unlikely(!list_empty(&tsk->pi_state_list)))
		exit_pi_state_list(tsk);
	if (unlikely(current->pi_state_cache))
		kfree(current->pi_state_cache);
1104
#endif
1105
	/*
1106
	 * Make sure we are holding no locks:
1107
	 */
1108
	debug_check_no_locks_held(tsk);
1109 1110 1111 1112 1113 1114
	/*
	 * 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
1115

1116 1117 1118
	if (tsk->io_context)
		exit_io_context();

1119 1120 1121
	if (tsk->splice_pipe)
		__free_pipe_info(tsk->splice_pipe);

C
Coywolf Qi Hunt 已提交
1122
	preempt_disable();
O
Oleg Nesterov 已提交
1123
	/* causes final put_task_struct in finish_task_switch(). */
1124
	tsk->state = TASK_DEAD;
L
Linus Torvalds 已提交
1125 1126 1127
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
1128 1129
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
1130 1131
}

1132 1133
EXPORT_SYMBOL_GPL(do_exit);

L
Linus Torvalds 已提交
1134 1135 1136 1137
NORET_TYPE void complete_and_exit(struct completion *comp, long code)
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
1138

L
Linus Torvalds 已提交
1139 1140 1141 1142 1143
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

1144
SYSCALL_DEFINE1(exit, int, error_code)
L
Linus Torvalds 已提交
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
{
	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)
{
1156 1157
	struct signal_struct *sig = current->signal;

L
Linus Torvalds 已提交
1158 1159
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

1160 1161
	if (signal_group_exit(sig))
		exit_code = sig->group_exit_code;
L
Linus Torvalds 已提交
1162 1163 1164
	else if (!thread_group_empty(current)) {
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1165
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1166 1167 1168 1169
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1170
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
			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.
 */
1185
SYSCALL_DEFINE1(exit_group, int, error_code)
L
Linus Torvalds 已提交
1186 1187
{
	do_group_exit((error_code & 0xff) << 8);
1188 1189
	/* NOTREACHED */
	return 0;
L
Linus Torvalds 已提交
1190 1191
}

E
Eric W. Biederman 已提交
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203
static struct pid *task_pid_type(struct task_struct *task, enum pid_type type)
{
	struct pid *pid = NULL;
	if (type == PIDTYPE_PID)
		pid = task->pids[type].pid;
	else if (type < PIDTYPE_MAX)
		pid = task->group_leader->pids[type].pid;
	return pid;
}

static int eligible_child(enum pid_type type, struct pid *pid, int options,
			  struct task_struct *p)
L
Linus Torvalds 已提交
1204
{
1205 1206
	int err;

E
Eric W. Biederman 已提交
1207 1208
	if (type < PIDTYPE_MAX) {
		if (task_pid_type(p, type) != pid)
L
Linus Torvalds 已提交
1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220
			return 0;
	}

	/* 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.) */
	if (((p->exit_signal != SIGCHLD) ^ ((options & __WCLONE) != 0))
	    && !(options & __WALL))
		return 0;

1221
	err = security_task_wait(p);
1222 1223
	if (err)
		return err;
L
Linus Torvalds 已提交
1224

1225
	return 1;
L
Linus Torvalds 已提交
1226 1227
}

1228
static int wait_noreap_copyout(struct task_struct *p, pid_t pid, uid_t uid,
L
Linus Torvalds 已提交
1229 1230 1231 1232 1233
			       int why, int status,
			       struct siginfo __user *infop,
			       struct rusage __user *rusagep)
{
	int retval = rusagep ? getrusage(p, RUSAGE_BOTH, rusagep) : 0;
1234

L
Linus Torvalds 已提交
1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258
	put_task_struct(p);
	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);
	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.
 */
R
Roland McGrath 已提交
1259
static int wait_task_zombie(struct task_struct *p, int options,
L
Linus Torvalds 已提交
1260 1261 1262 1263
			    struct siginfo __user *infop,
			    int __user *stat_addr, struct rusage __user *ru)
{
	unsigned long state;
1264
	int retval, status, traced;
1265
	pid_t pid = task_pid_vnr(p);
1266
	uid_t uid = __task_cred(p)->uid;
L
Linus Torvalds 已提交
1267

R
Roland McGrath 已提交
1268 1269 1270 1271
	if (!likely(options & WEXITED))
		return 0;

	if (unlikely(options & WNOWAIT)) {
L
Linus Torvalds 已提交
1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297
		int exit_code = p->exit_code;
		int why, status;

		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;
		}
		return wait_noreap_copyout(p, pid, uid, why,
					   status, infop, ru);
	}

	/*
	 * 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;
	}

1298
	traced = ptrace_reparented(p);
1299 1300

	if (likely(!traced)) {
1301 1302
		struct signal_struct *psig;
		struct signal_struct *sig;
1303
		struct task_cputime cputime;
1304

L
Linus Torvalds 已提交
1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318
		/*
		 * 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
		 * need to protect the access to p->parent->signal fields,
		 * as other threads in the parent group can be right
		 * here reaping other children at the same time.
1319 1320 1321 1322
		 *
		 * We use thread_group_cputime() 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
1323
		 */
1324
		thread_group_cputime(p, &cputime);
L
Linus Torvalds 已提交
1325
		spin_lock_irq(&p->parent->sighand->siglock);
1326 1327 1328 1329
		psig = p->parent->signal;
		sig = p->signal;
		psig->cutime =
			cputime_add(psig->cutime,
1330 1331
			cputime_add(cputime.utime,
				    sig->cutime));
1332 1333
		psig->cstime =
			cputime_add(psig->cstime,
1334 1335
			cputime_add(cputime.stime,
				    sig->cstime));
1336 1337 1338 1339 1340
		psig->cgtime =
			cputime_add(psig->cgtime,
			cputime_add(p->gtime,
			cputime_add(sig->gtime,
				    sig->cgtime)));
1341 1342 1343 1344 1345 1346 1347 1348
		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;
1349 1350 1351 1352 1353 1354
		psig->cinblock +=
			task_io_get_inblock(p) +
			sig->inblock + sig->cinblock;
		psig->coublock +=
			task_io_get_oublock(p) +
			sig->oublock + sig->coublock;
1355 1356
		task_io_accounting_add(&psig->ioac, &p->ioac);
		task_io_accounting_add(&psig->ioac, &sig->ioac);
L
Linus Torvalds 已提交
1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389
		spin_unlock_irq(&p->parent->sighand->siglock);
	}

	/*
	 * 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);

	retval = ru ? getrusage(p, RUSAGE_BOTH, ru) : 0;
	status = (p->signal->flags & SIGNAL_GROUP_EXIT)
		? p->signal->group_exit_code : p->exit_code;
	if (!retval && stat_addr)
		retval = put_user(status, stat_addr);
	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)
1390
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1391
	if (!retval && infop)
1392
		retval = put_user(uid, &infop->si_uid);
1393
	if (!retval)
1394
		retval = pid;
1395 1396

	if (traced) {
L
Linus Torvalds 已提交
1397
		write_lock_irq(&tasklist_lock);
1398 1399 1400 1401 1402 1403 1404
		/* 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.
		 */
1405
		if (!task_detached(p)) {
1406
			do_notify_parent(p, p->exit_signal);
1407
			if (!task_detached(p)) {
1408 1409
				p->exit_state = EXIT_ZOMBIE;
				p = NULL;
L
Linus Torvalds 已提交
1410 1411 1412 1413 1414 1415
			}
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1416

L
Linus Torvalds 已提交
1417 1418 1419 1420 1421 1422 1423 1424 1425
	return retval;
}

/*
 * Handle sys_wait4 work for one task in state TASK_STOPPED.  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.
 */
R
Roland McGrath 已提交
1426
static int wait_task_stopped(int ptrace, struct task_struct *p,
R
Roland McGrath 已提交
1427
			     int options, struct siginfo __user *infop,
L
Linus Torvalds 已提交
1428 1429
			     int __user *stat_addr, struct rusage __user *ru)
{
1430 1431
	int retval, exit_code, why;
	uid_t uid = 0; /* unneeded, required by compiler */
1432
	pid_t pid;
L
Linus Torvalds 已提交
1433

R
Roland McGrath 已提交
1434
	if (!(options & WUNTRACED))
R
Roland McGrath 已提交
1435 1436
		return 0;

1437 1438 1439 1440 1441 1442
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

	if (unlikely(!task_is_stopped_or_traced(p)))
		goto unlock_sig;

R
Roland McGrath 已提交
1443
	if (!ptrace && p->signal->group_stop_count > 0)
L
Linus Torvalds 已提交
1444 1445 1446 1447
		/*
		 * A group stop is in progress and this is the group leader.
		 * We won't report until all threads have stopped.
		 */
1448 1449 1450 1451 1452 1453
		goto unlock_sig;

	exit_code = p->exit_code;
	if (!exit_code)
		goto unlock_sig;

R
Roland McGrath 已提交
1454
	if (!unlikely(options & WNOWAIT))
1455 1456
		p->exit_code = 0;

1457 1458
	/* don't need the RCU readlock here as we're holding a spinlock */
	uid = __task_cred(p)->uid;
1459 1460 1461
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1462 1463 1464 1465 1466 1467 1468 1469 1470 1471
		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);
1472
	pid = task_pid_vnr(p);
R
Roland McGrath 已提交
1473
	why = ptrace ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1474 1475
	read_unlock(&tasklist_lock);

R
Roland McGrath 已提交
1476
	if (unlikely(options & WNOWAIT))
L
Linus Torvalds 已提交
1477
		return wait_noreap_copyout(p, pid, uid,
1478
					   why, exit_code,
L
Linus Torvalds 已提交
1479 1480 1481 1482 1483 1484 1485 1486 1487 1488
					   infop, ru);

	retval = ru ? getrusage(p, RUSAGE_BOTH, ru) : 0;
	if (!retval && stat_addr)
		retval = put_user((exit_code << 8) | 0x7f, stat_addr);
	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
1489
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1490 1491 1492
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1493
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1494
	if (!retval && infop)
1495
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1496
	if (!retval)
1497
		retval = pid;
L
Linus Torvalds 已提交
1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509
	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.
 */
R
Roland McGrath 已提交
1510
static int wait_task_continued(struct task_struct *p, int options,
L
Linus Torvalds 已提交
1511 1512 1513 1514 1515 1516 1517
			       struct siginfo __user *infop,
			       int __user *stat_addr, struct rusage __user *ru)
{
	int retval;
	pid_t pid;
	uid_t uid;

R
Roland McGrath 已提交
1518 1519 1520
	if (!unlikely(options & WCONTINUED))
		return 0;

L
Linus Torvalds 已提交
1521 1522 1523 1524 1525 1526 1527 1528 1529
	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;
	}
R
Roland McGrath 已提交
1530
	if (!unlikely(options & WNOWAIT))
L
Linus Torvalds 已提交
1531
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
1532
	uid = __task_cred(p)->uid;
L
Linus Torvalds 已提交
1533 1534
	spin_unlock_irq(&p->sighand->siglock);

1535
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1536 1537 1538 1539 1540 1541 1542 1543 1544
	get_task_struct(p);
	read_unlock(&tasklist_lock);

	if (!infop) {
		retval = ru ? getrusage(p, RUSAGE_BOTH, ru) : 0;
		put_task_struct(p);
		if (!retval && stat_addr)
			retval = put_user(0xffff, stat_addr);
		if (!retval)
1545
			retval = pid;
L
Linus Torvalds 已提交
1546 1547 1548 1549 1550 1551 1552 1553 1554 1555
	} else {
		retval = wait_noreap_copyout(p, pid, uid,
					     CLD_CONTINUED, SIGCONT,
					     infop, ru);
		BUG_ON(retval == 0);
	}

	return retval;
}

R
Roland McGrath 已提交
1556 1557 1558 1559 1560 1561
/*
 * Consider @p for a wait by @parent.
 *
 * -ECHILD should be in *@notask_error before the first call.
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue;
1562 1563
 * then *@notask_error is 0 if @p is an eligible child,
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1564
 */
R
Roland McGrath 已提交
1565
static int wait_consider_task(struct task_struct *parent, int ptrace,
R
Roland McGrath 已提交
1566 1567 1568 1569 1570 1571
			      struct task_struct *p, int *notask_error,
			      enum pid_type type, struct pid *pid, int options,
			      struct siginfo __user *infop,
			      int __user *stat_addr, struct rusage __user *ru)
{
	int ret = eligible_child(type, pid, options, p);
1572
	if (!ret)
R
Roland McGrath 已提交
1573 1574
		return ret;

1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
	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.
		 */
		if (*notask_error)
			*notask_error = ret;
	}

R
Roland McGrath 已提交
1587 1588 1589 1590 1591 1592 1593 1594 1595
	if (likely(!ptrace) && unlikely(p->ptrace)) {
		/*
		 * This child is hidden by ptrace.
		 * We aren't allowed to see it now, but eventually we will.
		 */
		*notask_error = 0;
		return 0;
	}

R
Roland McGrath 已提交
1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
	if (p->exit_state == EXIT_DEAD)
		return 0;

	/*
	 * We don't reap group leaders with subthreads.
	 */
	if (p->exit_state == EXIT_ZOMBIE && !delay_group_leader(p))
		return wait_task_zombie(p, options, infop, stat_addr, ru);

	/*
	 * It's stopped or running now, so it might
	 * later continue, exit, or stop again.
	 */
	*notask_error = 0;

	if (task_is_stopped_or_traced(p))
R
Roland McGrath 已提交
1612 1613
		return wait_task_stopped(ptrace, p, options,
					 infop, stat_addr, ru);
R
Roland McGrath 已提交
1614 1615 1616 1617 1618 1619 1620 1621 1622 1623

	return wait_task_continued(p, options, infop, stat_addr, ru);
}

/*
 * Do the work of do_wait() for one thread in the group, @tsk.
 *
 * -ECHILD should be in *@notask_error before the first call.
 * Returns nonzero for a final return, when we have unlocked tasklist_lock.
 * Returns zero if the search for a child should continue; then
1624 1625
 * *@notask_error is 0 if there were any eligible children,
 * or another error from security_task_wait(), or still -ECHILD.
R
Roland McGrath 已提交
1626 1627 1628 1629 1630 1631 1632 1633 1634
 */
static int do_wait_thread(struct task_struct *tsk, int *notask_error,
			  enum pid_type type, struct pid *pid, int options,
			  struct siginfo __user *infop, int __user *stat_addr,
			  struct rusage __user *ru)
{
	struct task_struct *p;

	list_for_each_entry(p, &tsk->children, sibling) {
R
Roland McGrath 已提交
1635 1636 1637 1638 1639 1640 1641 1642 1643 1644
		/*
		 * Do not consider detached threads.
		 */
		if (!task_detached(p)) {
			int ret = wait_consider_task(tsk, 0, p, notask_error,
						     type, pid, options,
						     infop, stat_addr, ru);
			if (ret)
				return ret;
		}
R
Roland McGrath 已提交
1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657
	}

	return 0;
}

static int ptrace_do_wait(struct task_struct *tsk, int *notask_error,
			  enum pid_type type, struct pid *pid, int options,
			  struct siginfo __user *infop, int __user *stat_addr,
			  struct rusage __user *ru)
{
	struct task_struct *p;

	/*
R
Roland McGrath 已提交
1658
	 * Traditionally we see ptrace'd stopped tasks regardless of options.
R
Roland McGrath 已提交
1659
	 */
R
Roland McGrath 已提交
1660
	options |= WUNTRACED;
R
Roland McGrath 已提交
1661

R
Roland McGrath 已提交
1662 1663 1664 1665 1666
	list_for_each_entry(p, &tsk->ptraced, ptrace_entry) {
		int ret = wait_consider_task(tsk, 1, p, notask_error,
					     type, pid, options,
					     infop, stat_addr, ru);
		if (ret)
R
Roland McGrath 已提交
1667 1668 1669 1670 1671 1672
			return ret;
	}

	return 0;
}

E
Eric W. Biederman 已提交
1673 1674 1675
static long do_wait(enum pid_type type, struct pid *pid, int options,
		    struct siginfo __user *infop, int __user *stat_addr,
		    struct rusage __user *ru)
L
Linus Torvalds 已提交
1676 1677 1678
{
	DECLARE_WAITQUEUE(wait, current);
	struct task_struct *tsk;
R
Roland McGrath 已提交
1679
	int retval;
L
Linus Torvalds 已提交
1680

1681 1682
	trace_sched_process_wait(pid);

L
Linus Torvalds 已提交
1683 1684
	add_wait_queue(&current->signal->wait_chldexit,&wait);
repeat:
R
Roland McGrath 已提交
1685 1686 1687 1688 1689
	/*
	 * If there is nothing that can match our critiera just get out.
	 * We will clear @retval to zero if we see any child that might later
	 * match our criteria, even if we are not able to reap it yet.
	 */
E
Eric W. Biederman 已提交
1690 1691 1692 1693
	retval = -ECHILD;
	if ((type < PIDTYPE_MAX) && (!pid || hlist_empty(&pid->tasks[type])))
		goto end;

L
Linus Torvalds 已提交
1694 1695 1696 1697
	current->state = TASK_INTERRUPTIBLE;
	read_lock(&tasklist_lock);
	tsk = current;
	do {
R
Roland McGrath 已提交
1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710
		int tsk_result = do_wait_thread(tsk, &retval,
						type, pid, options,
						infop, stat_addr, ru);
		if (!tsk_result)
			tsk_result = ptrace_do_wait(tsk, &retval,
						    type, pid, options,
						    infop, stat_addr, ru);
		if (tsk_result) {
			/*
			 * tasklist_lock is unlocked and we have a final result.
			 */
			retval = tsk_result;
			goto end;
L
Linus Torvalds 已提交
1711
		}
R
Roland McGrath 已提交
1712

L
Linus Torvalds 已提交
1713 1714 1715
		if (options & __WNOTHREAD)
			break;
		tsk = next_thread(tsk);
E
Eric Sesterhenn 已提交
1716
		BUG_ON(tsk->signal != current->signal);
L
Linus Torvalds 已提交
1717 1718
	} while (tsk != current);
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1719

R
Roland McGrath 已提交
1720
	if (!retval && !(options & WNOHANG)) {
L
Linus Torvalds 已提交
1721
		retval = -ERESTARTSYS;
R
Roland McGrath 已提交
1722 1723 1724 1725
		if (!signal_pending(current)) {
			schedule();
			goto repeat;
		}
L
Linus Torvalds 已提交
1726
	}
R
Roland McGrath 已提交
1727

L
Linus Torvalds 已提交
1728 1729 1730 1731 1732
end:
	current->state = TASK_RUNNING;
	remove_wait_queue(&current->signal->wait_chldexit,&wait);
	if (infop) {
		if (retval > 0)
1733
			retval = 0;
L
Linus Torvalds 已提交
1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756
		else {
			/*
			 * For a WNOHANG return, clear out all the fields
			 * we would set so the user can easily tell the
			 * difference.
			 */
			if (!retval)
				retval = put_user(0, &infop->si_signo);
			if (!retval)
				retval = put_user(0, &infop->si_errno);
			if (!retval)
				retval = put_user(0, &infop->si_code);
			if (!retval)
				retval = put_user(0, &infop->si_pid);
			if (!retval)
				retval = put_user(0, &infop->si_uid);
			if (!retval)
				retval = put_user(0, &infop->si_status);
		}
	}
	return retval;
}

1757 1758
SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *,
		infop, int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1759
{
E
Eric W. Biederman 已提交
1760 1761
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1762 1763 1764 1765 1766 1767 1768 1769 1770
	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
1771
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1772 1773
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1774 1775
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1776 1777 1778
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1779 1780
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1781 1782 1783 1784 1785 1786
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1787 1788 1789 1790
	if (type < PIDTYPE_MAX)
		pid = find_get_pid(upid);
	ret = do_wait(type, pid, options, infop, NULL, ru);
	put_pid(pid);
L
Linus Torvalds 已提交
1791 1792

	/* avoid REGPARM breakage on x86: */
1793
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1794 1795 1796
	return ret;
}

1797 1798
SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr,
		int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1799
{
E
Eric W. Biederman 已提交
1800 1801
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1802 1803 1804 1805 1806
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822

	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;
		pid = get_pid(task_pgrp(current));
	} else /* upid > 0 */ {
		type = PIDTYPE_PID;
		pid = find_get_pid(upid);
	}

	ret = do_wait(type, pid, options | WEXITED, NULL, stat_addr, ru);
	put_pid(pid);
L
Linus Torvalds 已提交
1823 1824

	/* avoid REGPARM breakage on x86: */
1825
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1826 1827 1828 1829 1830 1831 1832 1833 1834
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
1835
SYSCALL_DEFINE3(waitpid, pid_t, pid, int __user *, stat_addr, int, options)
L
Linus Torvalds 已提交
1836 1837 1838 1839 1840
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif