exit.c 43.7 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);

L
Linus Torvalds 已提交
64 65 66 67 68 69 70
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);
71

72
		list_del_rcu(&p->tasks);
73
		__get_cpu_var(process_counts)--;
L
Linus Torvalds 已提交
74
	}
O
Oleg Nesterov 已提交
75
	list_del_rcu(&p->thread_group);
R
Roland McGrath 已提交
76
	list_del_init(&p->sibling);
L
Linus Torvalds 已提交
77 78
}

79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
/*
 * 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:
		 */
101
		if (sig->group_exit_task && atomic_read(&sig->count) == sig->notify_count)
102
			wake_up_process(sig->group_exit_task);
103

104 105 106 107 108 109 110 111 112 113 114 115
		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.
		 */
116 117
		sig->utime = cputime_add(sig->utime, task_utime(tsk));
		sig->stime = cputime_add(sig->stime, task_stime(tsk));
118
		sig->gtime = cputime_add(sig->gtime, task_gtime(tsk));
119 120 121 122
		sig->min_flt += tsk->min_flt;
		sig->maj_flt += tsk->maj_flt;
		sig->nvcsw += tsk->nvcsw;
		sig->nivcsw += tsk->nivcsw;
123 124
		sig->inblock += task_io_get_inblock(tsk);
		sig->oublock += task_io_get_oublock(tsk);
125
		task_io_accounting_add(&sig->ioac, &tsk->ioac);
126
		sig->sum_sched_runtime += tsk->se.sum_exec_runtime;
127 128 129
		sig = NULL; /* Marker for below. */
	}

130 131
	__unhash_process(tsk);

132 133 134 135 136 137
	/*
	 * Do this under ->siglock, we can race with another thread
	 * doing sigqueue_free() if we have SIGQUEUE_PREALLOC signals.
	 */
	flush_sigqueue(&tsk->pending);

138
	tsk->signal = NULL;
139
	tsk->sighand = NULL;
140 141
	spin_unlock(&sighand->siglock);

142
	__cleanup_sighand(sighand);
143 144 145
	clear_tsk_thread_flag(tsk,TIF_SIGPENDING);
	if (sig) {
		flush_sigqueue(&sig->shared_pending);
146
		taskstats_tgid_free(sig);
147 148 149 150 151
		/*
		 * Make sure ->signal can't go away under rq->lock,
		 * see account_group_exec_runtime().
		 */
		task_rq_unlock_wait(tsk);
152 153 154 155
		__cleanup_signal(sig);
	}
}

156 157
static void delayed_put_task_struct(struct rcu_head *rhp)
{
158 159 160 161
	struct task_struct *tsk = container_of(rhp, struct task_struct, rcu);

	trace_sched_process_free(tsk);
	put_task_struct(tsk);
162 163
}

R
Roland McGrath 已提交
164

L
Linus Torvalds 已提交
165 166
void release_task(struct task_struct * p)
{
167
	struct task_struct *leader;
L
Linus Torvalds 已提交
168
	int zap_leader;
169
repeat:
R
Roland McGrath 已提交
170
	tracehook_prepare_release_task(p);
171 172 173 174
	/* 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);

175
	proc_flush_task(p);
L
Linus Torvalds 已提交
176
	write_lock_irq(&tasklist_lock);
R
Roland McGrath 已提交
177
	tracehook_finish_release_task(p);
L
Linus Torvalds 已提交
178
	__exit_signal(p);
179

L
Linus Torvalds 已提交
180 181 182 183 184 185 186 187
	/*
	 * 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) {
188
		BUG_ON(task_detached(leader));
L
Linus Torvalds 已提交
189 190 191 192 193 194 195 196 197
		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.
		 */
198
		zap_leader = task_detached(leader);
R
Roland McGrath 已提交
199 200 201 202 203 204 205

		/*
		 * 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
206 207 208 209
	}

	write_unlock_irq(&tasklist_lock);
	release_thread(p);
210
	call_rcu(&p->rcu, delayed_put_task_struct);
L
Linus Torvalds 已提交
211 212 213 214 215 216 217 218 219 220

	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...
221 222
 *
 * The caller must hold rcu lock or the tasklist lock.
L
Linus Torvalds 已提交
223
 */
224
struct pid *session_of_pgrp(struct pid *pgrp)
L
Linus Torvalds 已提交
225 226
{
	struct task_struct *p;
227
	struct pid *sid = NULL;
228

229
	p = pid_task(pgrp, PIDTYPE_PGID);
230
	if (p == NULL)
231
		p = pid_task(pgrp, PIDTYPE_PID);
232
	if (p != NULL)
233
		sid = task_session(p);
234

L
Linus Torvalds 已提交
235 236 237 238 239 240 241 242 243 244 245
	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?"
 */
246
static int will_become_orphaned_pgrp(struct pid *pgrp, struct task_struct *ignored_task)
L
Linus Torvalds 已提交
247 248 249
{
	struct task_struct *p;

250
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
251 252 253
		if ((p == ignored_task) ||
		    (p->exit_state && thread_group_empty(p)) ||
		    is_global_init(p->real_parent))
L
Linus Torvalds 已提交
254
			continue;
255

256
		if (task_pgrp(p->real_parent) != pgrp &&
257 258
		    task_session(p->real_parent) == task_session(p))
			return 0;
259
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
260 261

	return 1;
L
Linus Torvalds 已提交
262 263
}

264
int is_current_pgrp_orphaned(void)
L
Linus Torvalds 已提交
265 266 267 268
{
	int retval;

	read_lock(&tasklist_lock);
269
	retval = will_become_orphaned_pgrp(task_pgrp(current), NULL);
L
Linus Torvalds 已提交
270 271 272 273 274
	read_unlock(&tasklist_lock);

	return retval;
}

275
static int has_stopped_jobs(struct pid *pgrp)
L
Linus Torvalds 已提交
276 277 278 279
{
	int retval = 0;
	struct task_struct *p;

280
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
M
Matthew Wilcox 已提交
281
		if (!task_is_stopped(p))
L
Linus Torvalds 已提交
282 283 284
			continue;
		retval = 1;
		break;
285
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
L
Linus Torvalds 已提交
286 287 288
	return retval;
}

289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319
/*
 * 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
320
/**
321
 * reparent_to_kthreadd - Reparent the calling kernel thread to kthreadd
L
Linus Torvalds 已提交
322 323
 *
 * If a kernel thread is launched as a result of a system call, or if
324 325
 * 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
326 327 328 329
 *
 * 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.
 *
330
 * NOTE that reparent_to_kthreadd() gives the caller full capabilities.
L
Linus Torvalds 已提交
331
 */
332
static void reparent_to_kthreadd(void)
L
Linus Torvalds 已提交
333 334 335 336 337
{
	write_lock_irq(&tasklist_lock);

	ptrace_unlink(current);
	/* Reparent to init */
338
	current->real_parent = current->parent = kthreadd_task;
R
Roland McGrath 已提交
339
	list_move_tail(&current->sibling, &current->real_parent->children);
L
Linus Torvalds 已提交
340 341 342 343

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

344
	if (task_nice(current) < 0)
L
Linus Torvalds 已提交
345 346 347 348 349 350
		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
351 352 353

	atomic_inc(&init_cred.usage);
	commit_creds(&init_cred);
L
Linus Torvalds 已提交
354 355 356
	write_unlock_irq(&tasklist_lock);
}

357
void __set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
358
{
359
	struct task_struct *curr = current->group_leader;
360
	pid_t nr = pid_nr(pid);
L
Linus Torvalds 已提交
361

362
	if (task_session(curr) != pid) {
363
		change_pid(curr, PIDTYPE_SID, pid);
364
		set_task_session(curr, nr);
L
Linus Torvalds 已提交
365
	}
366
	if (task_pgrp(curr) != pid) {
367
		change_pid(curr, PIDTYPE_PGID, pid);
368
		set_task_pgrp(curr, nr);
L
Linus Torvalds 已提交
369 370 371
	}
}

372
static void set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
373 374
{
	write_lock_irq(&tasklist_lock);
375
	__set_special_pids(pid);
L
Linus Torvalds 已提交
376 377 378 379 380 381 382 383 384 385
	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)
{
386
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406
		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)
{
407
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
408 409 410
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
411
	current->sighand->action[(sig)-1].sa.sa_handler = SIG_IGN;
L
Linus Torvalds 已提交
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439
	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);
440 441 442 443
	/*
	 * 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
444
	current->flags |= (PF_NOFREEZE | PF_KTHREAD);
L
Linus Torvalds 已提交
445

446 447 448 449
	if (current->nsproxy != &init_nsproxy) {
		get_nsproxy(&init_nsproxy);
		switch_task_namespaces(current, &init_nsproxy);
	}
450
	set_special_pids(&init_struct_pid);
451
	proc_clear_tty(current);
L
Linus Torvalds 已提交
452 453 454 455 456 457 458 459 460 461 462 463

	/* 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
464

465
	exit_files(current);
L
Linus Torvalds 已提交
466 467 468
	current->files = init_task.files;
	atomic_inc(&current->files->count);

469
	reparent_to_kthreadd();
L
Linus Torvalds 已提交
470 471 472 473
}

EXPORT_SYMBOL(daemonize);

474
static void close_files(struct files_struct * files)
L
Linus Torvalds 已提交
475 476
{
	int i, j;
477
	struct fdtable *fdt;
L
Linus Torvalds 已提交
478 479

	j = 0;
480 481 482 483 484 485

	/*
	 * It is safe to dereference the fd table without RCU or
	 * ->file_lock because this is the last reference to the
	 * files structure.
	 */
486
	fdt = files_fdtable(files);
L
Linus Torvalds 已提交
487 488 489
	for (;;) {
		unsigned long set;
		i = j * __NFDBITS;
490
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
491
			break;
492
		set = fdt->open_fds->fds_bits[j++];
L
Linus Torvalds 已提交
493 494
		while (set) {
			if (set & 1) {
495
				struct file * file = xchg(&fdt->fd[i], NULL);
496
				if (file) {
L
Linus Torvalds 已提交
497
					filp_close(file, files);
498 499
					cond_resched();
				}
L
Linus Torvalds 已提交
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519
			}
			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;
}

520
void put_files_struct(struct files_struct *files)
L
Linus Torvalds 已提交
521
{
522 523
	struct fdtable *fdt;

L
Linus Torvalds 已提交
524 525 526 527
	if (atomic_dec_and_test(&files->count)) {
		close_files(files);
		/*
		 * Free the fd and fdset arrays if we expanded them.
528 529 530
		 * 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
531
		 */
532
		fdt = files_fdtable(files);
533
		if (fdt != &files->fdtab)
534
			kmem_cache_free(files_cachep, files);
535
		free_fdtable(fdt);
L
Linus Torvalds 已提交
536 537 538
	}
}

539
void reset_files_struct(struct files_struct *files)
540
{
541
	struct task_struct *tsk = current;
542 543 544 545 546 547 548 549 550
	struct files_struct *old;

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

551
void exit_files(struct task_struct *tsk)
L
Linus Torvalds 已提交
552 553 554 555 556 557 558 559 560 561 562
{
	struct files_struct * files = tsk->files;

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

563
void put_fs_struct(struct fs_struct *fs)
L
Linus Torvalds 已提交
564 565 566
{
	/* No need to hold fs->lock if we are killing it */
	if (atomic_dec_and_test(&fs->count)) {
J
Jan Blunck 已提交
567 568
		path_put(&fs->root);
		path_put(&fs->pwd);
L
Linus Torvalds 已提交
569 570 571 572
		kmem_cache_free(fs_cachep, fs);
	}
}

573
void exit_fs(struct task_struct *tsk)
L
Linus Torvalds 已提交
574 575 576 577 578 579 580
{
	struct fs_struct * fs = tsk->fs;

	if (fs) {
		task_lock(tsk);
		tsk->fs = NULL;
		task_unlock(tsk);
581
		put_fs_struct(fs);
L
Linus Torvalds 已提交
582 583 584 585 586
	}
}

EXPORT_SYMBOL_GPL(exit_fs);

587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
#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);
640 641 642
	/*
	 * We found no owner yet mm_users > 1: this implies that we are
	 * most likely racing with swapoff (try_to_unuse()) or /proc or
643
	 * ptrace or page migration (get_task_mm()).  Mark owner as NULL.
644 645
	 */
	mm->owner = NULL;
646 647 648 649 650 651 652 653 654 655
	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);
656 657 658 659 660
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
661 662 663 664 665 666 667 668 669 670 671
	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
672 673 674 675
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
676
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
677 678
{
	struct mm_struct *mm = tsk->mm;
679
	struct core_state *core_state;
L
Linus Torvalds 已提交
680 681 682 683 684 685

	mm_release(tsk, mm);
	if (!mm)
		return;
	/*
	 * Serialize with any possible pending coredump.
686
	 * We must hold mmap_sem around checking core_state
L
Linus Torvalds 已提交
687
	 * and clearing tsk->mm.  The core-inducing thread
688
	 * will increment ->nr_threads for each thread in the
L
Linus Torvalds 已提交
689 690 691
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
692 693 694
	core_state = mm->core_state;
	if (core_state) {
		struct core_thread self;
L
Linus Torvalds 已提交
695 696
		up_read(&mm->mmap_sem);

697 698 699 700 701 702 703 704
		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
705

O
Oleg Nesterov 已提交
706 707 708 709 710 711 712
		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
713 714 715
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
716
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
717 718 719 720 721
	/* more a memory barrier than a real lock */
	task_lock(tsk);
	tsk->mm = NULL;
	up_read(&mm->mmap_sem);
	enter_lazy_tlb(mm, current);
722 723
	/* We don't want this task to be frozen prematurely */
	clear_freeze_flag(tsk);
L
Linus Torvalds 已提交
724
	task_unlock(tsk);
725
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
726 727 728 729 730 731 732
	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
733 734
 * the child reaper process (ie "init") in our pid
 * space.
L
Linus Torvalds 已提交
735
 */
736
static struct task_struct *find_new_reaper(struct task_struct *father)
L
Linus Torvalds 已提交
737
{
738 739
	struct pid_namespace *pid_ns = task_active_pid_ns(father);
	struct task_struct *thread;
L
Linus Torvalds 已提交
740

741 742 743 744 745 746 747 748
	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
749

750 751 752 753
	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
754

755 756
		zap_pid_ns_processes(pid_ns);
		write_lock_irq(&tasklist_lock);
L
Linus Torvalds 已提交
757
		/*
758 759 760
		 * 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
761
		 */
762
		pid_ns->child_reaper = init_pid_ns.child_reaper;
L
Linus Torvalds 已提交
763
	}
764

765 766 767
	return pid_ns->child_reaper;
}

768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803
/*
* Any that need to be release_task'd are put on the @dead list.
 */
static void reparent_thread(struct task_struct *father, struct task_struct *p,
				struct list_head *dead)
{
	if (p->pdeath_signal)
		group_send_sig_info(p->pdeath_signal, SEND_SIG_NOINFO, p);

	list_move_tail(&p->sibling, &p->real_parent->children);

	if (task_detached(p))
		return;
	/*
	 * If this is a threaded reparent there is no need to
	 * notify anyone anything has happened.
	 */
	if (same_thread_group(p->real_parent, father))
		return;

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

	/* If it has exited notify the new parent about this child's death. */
	if (!p->ptrace &&
	    p->exit_state == EXIT_ZOMBIE && thread_group_empty(p)) {
		do_notify_parent(p, p->exit_signal);
		if (task_detached(p)) {
			p->exit_state = EXIT_DEAD;
			list_move_tail(&p->sibling, dead);
		}
	}

	kill_orphaned_pgrp(p, father);
}

804
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
805
{
806
	struct task_struct *p, *n, *reaper;
807
	LIST_HEAD(dead_children);
808

809 810
	exit_ptrace(father);

811
	write_lock_irq(&tasklist_lock);
812
	reaper = find_new_reaper(father);
R
Roland McGrath 已提交
813

814
	list_for_each_entry_safe(p, n, &father->children, sibling) {
815
		p->real_parent = reaper;
R
Roland McGrath 已提交
816 817 818 819
		if (p->parent == father) {
			BUG_ON(p->ptrace);
			p->parent = p->real_parent;
		}
820
		reparent_thread(father, p, &dead_children);
L
Linus Torvalds 已提交
821
	}
822
	write_unlock_irq(&tasklist_lock);
823

824 825
	BUG_ON(!list_empty(&father->children));

826 827
	list_for_each_entry_safe(p, n, &dead_children, sibling) {
		list_del_init(&p->sibling);
828 829
		release_task(p);
	}
L
Linus Torvalds 已提交
830 831 832 833 834 835
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
836
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
837
{
R
Roland McGrath 已提交
838 839
	int signal;
	void *cookie;
L
Linus Torvalds 已提交
840 841 842 843 844 845 846 847 848

	/*
	 * 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)
	 */
849
	forget_original_parent(tsk);
850
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
851

852
	write_lock_irq(&tasklist_lock);
853 854
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
855

856
	/* Let father know we died
L
Linus Torvalds 已提交
857 858
	 *
	 * Thread signals are configurable, but you aren't going to use
859
	 * that to send signals to arbitary processes.
L
Linus Torvalds 已提交
860 861 862 863 864 865 866 867 868 869
	 * 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.
	 */
870
	if (tsk->exit_signal != SIGCHLD && !task_detached(tsk) &&
871
	    (tsk->parent_exec_id != tsk->real_parent->self_exec_id ||
872 873
	     tsk->self_exec_id != tsk->parent_exec_id) &&
	    !capable(CAP_KILL))
L
Linus Torvalds 已提交
874 875
		tsk->exit_signal = SIGCHLD;

R
Roland McGrath 已提交
876
	signal = tracehook_notify_death(tsk, &cookie, group_dead);
877
	if (signal >= 0)
R
Roland McGrath 已提交
878
		signal = do_notify_parent(tsk, signal);
L
Linus Torvalds 已提交
879

880
	tsk->exit_state = signal == DEATH_REAP ? EXIT_DEAD : EXIT_ZOMBIE;
L
Linus Torvalds 已提交
881

882
	/* mt-exec, de_thread() is waiting for us */
883
	if (thread_group_leader(tsk) &&
884 885
	    tsk->signal->group_exit_task &&
	    tsk->signal->notify_count < 0)
886 887
		wake_up_process(tsk->signal->group_exit_task);

L
Linus Torvalds 已提交
888 889
	write_unlock_irq(&tasklist_lock);

R
Roland McGrath 已提交
890 891
	tracehook_report_death(tsk, signal, cookie, group_dead);

L
Linus Torvalds 已提交
892
	/* If the process is dead, release it - nobody will wait for it */
893
	if (signal == DEATH_REAP)
L
Linus Torvalds 已提交
894 895 896
		release_task(tsk);
}

897 898 899 900 901 902 903
#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;

904
	free = stack_not_used(current);
905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921

	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

922
NORET_TYPE void do_exit(long code)
L
Linus Torvalds 已提交
923 924 925 926 927 928
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

929 930
	WARN_ON(atomic_read(&tsk->fs_excl));

L
Linus Torvalds 已提交
931 932 933 934 935
	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
936
	tracehook_report_exit(&code);
L
Linus Torvalds 已提交
937

938 939 940 941 942 943 944
	/*
	 * 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");
945 946 947 948 949 950 951 952 953 954
		/*
		 * 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;
955 956 957 958
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

O
Oleg Nesterov 已提交
959
	exit_signals(tsk);  /* sets PF_EXITING */
960 961 962 963
	/*
	 * tsk->flags are checked in the futex code to protect against
	 * an exiting task cleaning up the robust pi futexes.
	 */
964 965
	smp_mb();
	spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
966 967 968

	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
969
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
970 971 972
				preempt_count());

	acct_update_integrals(tsk);
973

L
Linus Torvalds 已提交
974
	group_dead = atomic_dec_and_test(&tsk->signal->live);
975
	if (group_dead) {
976
		hrtimer_cancel(&tsk->signal->real_timer);
977
		exit_itimers(tsk->signal);
978
	}
979
	acct_collect(code, group_dead);
M
Miloslav Trmac 已提交
980 981
	if (group_dead)
		tty_audit_exit();
982 983
	if (unlikely(tsk->audit_context))
		audit_free(tsk);
984

985
	tsk->exit_code = code;
986
	taskstats_exit(tsk, group_dead);
987

L
Linus Torvalds 已提交
988 989
	exit_mm(tsk);

990
	if (group_dead)
991
		acct_process();
992 993
	trace_sched_process_exit(tsk);

L
Linus Torvalds 已提交
994
	exit_sem(tsk);
995 996
	exit_files(tsk);
	exit_fs(tsk);
997
	check_stack_usage();
L
Linus Torvalds 已提交
998
	exit_thread();
999
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
1000 1001 1002 1003

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

A
Al Viro 已提交
1004
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
1005 1006 1007
	if (tsk->binfmt)
		module_put(tsk->binfmt->module);

M
Matt Helsley 已提交
1008
	proc_exit_connector(tsk);
1009
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
1010
#ifdef CONFIG_NUMA
1011
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
1012 1013
	tsk->mempolicy = NULL;
#endif
1014
#ifdef CONFIG_FUTEX
1015 1016 1017 1018 1019 1020 1021 1022
	/*
	 * 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);
1023
#endif
1024
	/*
1025
	 * Make sure we are holding no locks:
1026
	 */
1027
	debug_check_no_locks_held(tsk);
1028 1029 1030 1031 1032 1033
	/*
	 * 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
1034

1035 1036 1037
	if (tsk->io_context)
		exit_io_context();

1038 1039 1040
	if (tsk->splice_pipe)
		__free_pipe_info(tsk->splice_pipe);

C
Coywolf Qi Hunt 已提交
1041
	preempt_disable();
O
Oleg Nesterov 已提交
1042
	/* causes final put_task_struct in finish_task_switch(). */
1043
	tsk->state = TASK_DEAD;
L
Linus Torvalds 已提交
1044 1045 1046
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
1047 1048
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
1049 1050
}

1051 1052
EXPORT_SYMBOL_GPL(do_exit);

L
Linus Torvalds 已提交
1053 1054 1055 1056
NORET_TYPE void complete_and_exit(struct completion *comp, long code)
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
1057

L
Linus Torvalds 已提交
1058 1059 1060 1061 1062
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

1063
SYSCALL_DEFINE1(exit, int, error_code)
L
Linus Torvalds 已提交
1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074
{
	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)
{
1075 1076
	struct signal_struct *sig = current->signal;

L
Linus Torvalds 已提交
1077 1078
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

1079 1080
	if (signal_group_exit(sig))
		exit_code = sig->group_exit_code;
L
Linus Torvalds 已提交
1081 1082 1083
	else if (!thread_group_empty(current)) {
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1084
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1085 1086 1087 1088
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1089
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103
			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.
 */
1104
SYSCALL_DEFINE1(exit_group, int, error_code)
L
Linus Torvalds 已提交
1105 1106
{
	do_group_exit((error_code & 0xff) << 8);
1107 1108
	/* NOTREACHED */
	return 0;
L
Linus Torvalds 已提交
1109 1110
}

E
Eric W. Biederman 已提交
1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122
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
1123
{
1124 1125
	int err;

E
Eric W. Biederman 已提交
1126 1127
	if (type < PIDTYPE_MAX) {
		if (task_pid_type(p, type) != pid)
L
Linus Torvalds 已提交
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
			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;

1140
	err = security_task_wait(p);
1141 1142
	if (err)
		return err;
L
Linus Torvalds 已提交
1143

1144
	return 1;
L
Linus Torvalds 已提交
1145 1146
}

1147
static int wait_noreap_copyout(struct task_struct *p, pid_t pid, uid_t uid,
L
Linus Torvalds 已提交
1148 1149 1150 1151 1152
			       int why, int status,
			       struct siginfo __user *infop,
			       struct rusage __user *rusagep)
{
	int retval = rusagep ? getrusage(p, RUSAGE_BOTH, rusagep) : 0;
1153

L
Linus Torvalds 已提交
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177
	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
1178
static int wait_task_zombie(struct task_struct *p, int options,
L
Linus Torvalds 已提交
1179 1180 1181 1182
			    struct siginfo __user *infop,
			    int __user *stat_addr, struct rusage __user *ru)
{
	unsigned long state;
1183
	int retval, status, traced;
1184
	pid_t pid = task_pid_vnr(p);
1185
	uid_t uid = __task_cred(p)->uid;
L
Linus Torvalds 已提交
1186

R
Roland McGrath 已提交
1187 1188 1189 1190
	if (!likely(options & WEXITED))
		return 0;

	if (unlikely(options & WNOWAIT)) {
L
Linus Torvalds 已提交
1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216
		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;
	}

1217
	traced = ptrace_reparented(p);
1218 1219

	if (likely(!traced)) {
1220 1221
		struct signal_struct *psig;
		struct signal_struct *sig;
1222
		struct task_cputime cputime;
1223

L
Linus Torvalds 已提交
1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237
		/*
		 * 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.
1238 1239 1240 1241
		 *
		 * 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
1242
		 */
1243
		thread_group_cputime(p, &cputime);
L
Linus Torvalds 已提交
1244
		spin_lock_irq(&p->parent->sighand->siglock);
1245 1246 1247 1248
		psig = p->parent->signal;
		sig = p->signal;
		psig->cutime =
			cputime_add(psig->cutime,
1249 1250
			cputime_add(cputime.utime,
				    sig->cutime));
1251 1252
		psig->cstime =
			cputime_add(psig->cstime,
1253 1254
			cputime_add(cputime.stime,
				    sig->cstime));
1255 1256 1257 1258 1259
		psig->cgtime =
			cputime_add(psig->cgtime,
			cputime_add(p->gtime,
			cputime_add(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;
1274 1275
		task_io_accounting_add(&psig->ioac, &p->ioac);
		task_io_accounting_add(&psig->ioac, &sig->ioac);
L
Linus Torvalds 已提交
1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308
		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)
1309
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1310
	if (!retval && infop)
1311
		retval = put_user(uid, &infop->si_uid);
1312
	if (!retval)
1313
		retval = pid;
1314 1315

	if (traced) {
L
Linus Torvalds 已提交
1316
		write_lock_irq(&tasklist_lock);
1317 1318 1319 1320 1321 1322 1323
		/* 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.
		 */
1324
		if (!task_detached(p)) {
1325
			do_notify_parent(p, p->exit_signal);
1326
			if (!task_detached(p)) {
1327 1328
				p->exit_state = EXIT_ZOMBIE;
				p = NULL;
L
Linus Torvalds 已提交
1329 1330 1331 1332 1333 1334
			}
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1335

L
Linus Torvalds 已提交
1336 1337 1338
	return retval;
}

1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
static int *task_stopped_code(struct task_struct *p, bool ptrace)
{
	if (ptrace) {
		if (task_is_stopped_or_traced(p))
			return &p->exit_code;
	} else {
		if (p->signal->flags & SIGNAL_STOP_STOPPED)
			return &p->signal->group_exit_code;
	}
	return NULL;
}

L
Linus Torvalds 已提交
1351 1352 1353 1354 1355 1356
/*
 * 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
1357
static int wait_task_stopped(int ptrace, struct task_struct *p,
R
Roland McGrath 已提交
1358
			     int options, struct siginfo __user *infop,
L
Linus Torvalds 已提交
1359 1360
			     int __user *stat_addr, struct rusage __user *ru)
{
1361
	int retval, exit_code, *p_code, why;
1362
	uid_t uid = 0; /* unneeded, required by compiler */
1363
	pid_t pid;
L
Linus Torvalds 已提交
1364

R
Roland McGrath 已提交
1365
	if (!(options & WUNTRACED))
R
Roland McGrath 已提交
1366 1367
		return 0;

1368 1369 1370
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

1371 1372
	p_code = task_stopped_code(p, ptrace);
	if (unlikely(!p_code))
1373 1374
		goto unlock_sig;

1375
	exit_code = *p_code;
1376 1377 1378
	if (!exit_code)
		goto unlock_sig;

R
Roland McGrath 已提交
1379
	if (!unlikely(options & WNOWAIT))
1380
		*p_code = 0;
1381

1382 1383
	/* don't need the RCU readlock here as we're holding a spinlock */
	uid = __task_cred(p)->uid;
1384 1385 1386
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
		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);
1397
	pid = task_pid_vnr(p);
R
Roland McGrath 已提交
1398
	why = ptrace ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1399 1400
	read_unlock(&tasklist_lock);

R
Roland McGrath 已提交
1401
	if (unlikely(options & WNOWAIT))
L
Linus Torvalds 已提交
1402
		return wait_noreap_copyout(p, pid, uid,
1403
					   why, exit_code,
L
Linus Torvalds 已提交
1404 1405 1406 1407 1408 1409 1410 1411 1412 1413
					   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
1414
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1415 1416 1417
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1418
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1419
	if (!retval && infop)
1420
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1421
	if (!retval)
1422
		retval = pid;
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434
	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
1435
static int wait_task_continued(struct task_struct *p, int options,
L
Linus Torvalds 已提交
1436 1437 1438 1439 1440 1441 1442
			       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
1443 1444 1445
	if (!unlikely(options & WCONTINUED))
		return 0;

L
Linus Torvalds 已提交
1446 1447 1448 1449 1450 1451 1452 1453 1454
	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
1455
	if (!unlikely(options & WNOWAIT))
L
Linus Torvalds 已提交
1456
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
1457
	uid = __task_cred(p)->uid;
L
Linus Torvalds 已提交
1458 1459
	spin_unlock_irq(&p->sighand->siglock);

1460
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1461 1462 1463 1464 1465 1466 1467 1468 1469
	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)
1470
			retval = pid;
L
Linus Torvalds 已提交
1471 1472 1473 1474 1475 1476 1477 1478 1479 1480
	} 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
1481 1482 1483 1484 1485 1486
/*
 * 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;
1487 1488
 * 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
1489
 */
R
Roland McGrath 已提交
1490
static int wait_consider_task(struct task_struct *parent, int ptrace,
R
Roland McGrath 已提交
1491 1492 1493 1494 1495 1496
			      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);
1497
	if (!ret)
R
Roland McGrath 已提交
1498 1499
		return ret;

1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511
	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
1512 1513 1514 1515 1516 1517 1518 1519 1520
	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
1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
	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;

1536
	if (task_stopped_code(p, ptrace))
R
Roland McGrath 已提交
1537 1538
		return wait_task_stopped(ptrace, p, options,
					 infop, stat_addr, ru);
R
Roland McGrath 已提交
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548

	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
1549 1550
 * *@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
1551 1552 1553 1554 1555 1556 1557 1558 1559
 */
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
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569
		/*
		 * 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
1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582
	}

	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
1583
	 * Traditionally we see ptrace'd stopped tasks regardless of options.
R
Roland McGrath 已提交
1584
	 */
R
Roland McGrath 已提交
1585
	options |= WUNTRACED;
R
Roland McGrath 已提交
1586

R
Roland McGrath 已提交
1587 1588 1589 1590 1591
	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
1592 1593 1594 1595 1596 1597
			return ret;
	}

	return 0;
}

E
Eric W. Biederman 已提交
1598 1599 1600
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
1601 1602 1603
{
	DECLARE_WAITQUEUE(wait, current);
	struct task_struct *tsk;
R
Roland McGrath 已提交
1604
	int retval;
L
Linus Torvalds 已提交
1605

1606 1607
	trace_sched_process_wait(pid);

L
Linus Torvalds 已提交
1608 1609
	add_wait_queue(&current->signal->wait_chldexit,&wait);
repeat:
R
Roland McGrath 已提交
1610 1611 1612 1613 1614
	/*
	 * 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
1615 1616 1617 1618
	retval = -ECHILD;
	if ((type < PIDTYPE_MAX) && (!pid || hlist_empty(&pid->tasks[type])))
		goto end;

L
Linus Torvalds 已提交
1619 1620 1621 1622
	current->state = TASK_INTERRUPTIBLE;
	read_lock(&tasklist_lock);
	tsk = current;
	do {
R
Roland McGrath 已提交
1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635
		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
1636
		}
R
Roland McGrath 已提交
1637

L
Linus Torvalds 已提交
1638 1639 1640
		if (options & __WNOTHREAD)
			break;
		tsk = next_thread(tsk);
E
Eric Sesterhenn 已提交
1641
		BUG_ON(tsk->signal != current->signal);
L
Linus Torvalds 已提交
1642 1643
	} while (tsk != current);
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1644

R
Roland McGrath 已提交
1645
	if (!retval && !(options & WNOHANG)) {
L
Linus Torvalds 已提交
1646
		retval = -ERESTARTSYS;
R
Roland McGrath 已提交
1647 1648 1649 1650
		if (!signal_pending(current)) {
			schedule();
			goto repeat;
		}
L
Linus Torvalds 已提交
1651
	}
R
Roland McGrath 已提交
1652

L
Linus Torvalds 已提交
1653 1654 1655 1656 1657
end:
	current->state = TASK_RUNNING;
	remove_wait_queue(&current->signal->wait_chldexit,&wait);
	if (infop) {
		if (retval > 0)
1658
			retval = 0;
L
Linus Torvalds 已提交
1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681
		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;
}

1682 1683
SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *,
		infop, int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1684
{
E
Eric W. Biederman 已提交
1685 1686
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1687 1688 1689 1690 1691 1692 1693 1694 1695
	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
1696
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1697 1698
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1699 1700
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1701 1702 1703
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1704 1705
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1706 1707 1708 1709 1710 1711
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1712 1713 1714 1715
	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
1716 1717

	/* avoid REGPARM breakage on x86: */
1718
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1719 1720 1721
	return ret;
}

1722 1723
SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr,
		int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1724
{
E
Eric W. Biederman 已提交
1725 1726
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1727 1728 1729 1730 1731
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1732 1733 1734 1735 1736 1737 1738 1739

	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;
1740
		pid = get_task_pid(current, PIDTYPE_PGID);
E
Eric W. Biederman 已提交
1741 1742 1743 1744 1745 1746 1747
	} 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
1748 1749

	/* avoid REGPARM breakage on x86: */
1750
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1751 1752 1753 1754 1755 1756 1757 1758 1759
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
1760
SYSCALL_DEFINE3(waitpid, pid_t, pid, int __user *, stat_addr, int, options)
L
Linus Torvalds 已提交
1761 1762 1763 1764 1765
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif