exit.c 43.0 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>
L
Linus Torvalds 已提交
16 17 18 19
#include <linux/key.h>
#include <linux/security.h>
#include <linux/cpu.h>
#include <linux/acct.h>
20
#include <linux/tsacct_kern.h>
L
Linus Torvalds 已提交
21 22
#include <linux/file.h>
#include <linux/binfmts.h>
S
Serge E. Hallyn 已提交
23
#include <linux/nsproxy.h>
24
#include <linux/pid_namespace.h>
L
Linus Torvalds 已提交
25 26 27 28
#include <linux/ptrace.h>
#include <linux/profile.h>
#include <linux/mount.h>
#include <linux/proc_fs.h>
29
#include <linux/kthread.h>
L
Linus Torvalds 已提交
30
#include <linux/mempolicy.h>
31
#include <linux/taskstats_kern.h>
32
#include <linux/delayacct.h>
33
#include <linux/freezer.h>
34
#include <linux/cgroup.h>
L
Linus Torvalds 已提交
35
#include <linux/syscalls.h>
36
#include <linux/signal.h>
37
#include <linux/posix-timers.h>
M
Matt Helsley 已提交
38
#include <linux/cn_proc.h>
39
#include <linux/mutex.h>
40
#include <linux/futex.h>
41
#include <linux/compat.h>
42
#include <linux/pipe_fs_i.h>
43
#include <linux/audit.h> /* for audit_free() */
A
Adrian Bunk 已提交
44
#include <linux/resource.h>
45
#include <linux/blkdev.h>
46
#include <linux/task_io_accounting_ops.h>
L
Linus Torvalds 已提交
47 48 49 50 51 52

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

A
Adrian Bunk 已提交
53 54
static void exit_mm(struct task_struct * tsk);

L
Linus Torvalds 已提交
55 56 57 58 59 60 61
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);
62

63
		list_del_rcu(&p->tasks);
64
		__get_cpu_var(process_counts)--;
L
Linus Torvalds 已提交
65
	}
O
Oleg Nesterov 已提交
66
	list_del_rcu(&p->thread_group);
67
	remove_parent(p);
L
Linus Torvalds 已提交
68 69
}

70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92
/*
 * 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));

	rcu_read_lock();
	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:
		 */
93
		if (sig->group_exit_task && atomic_read(&sig->count) == sig->notify_count)
94
			wake_up_process(sig->group_exit_task);
95

96 97 98 99 100 101 102 103 104 105 106 107 108 109
		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.
		 */
		sig->utime = cputime_add(sig->utime, tsk->utime);
		sig->stime = cputime_add(sig->stime, tsk->stime);
110
		sig->gtime = cputime_add(sig->gtime, tsk->gtime);
111 112 113 114
		sig->min_flt += tsk->min_flt;
		sig->maj_flt += tsk->maj_flt;
		sig->nvcsw += tsk->nvcsw;
		sig->nivcsw += tsk->nivcsw;
115 116
		sig->inblock += task_io_get_inblock(tsk);
		sig->oublock += task_io_get_oublock(tsk);
117
		sig->sum_sched_runtime += tsk->se.sum_exec_runtime;
118 119 120
		sig = NULL; /* Marker for below. */
	}

121 122
	__unhash_process(tsk);

123
	tsk->signal = NULL;
124
	tsk->sighand = NULL;
125 126 127
	spin_unlock(&sighand->siglock);
	rcu_read_unlock();

128
	__cleanup_sighand(sighand);
129 130 131 132
	clear_tsk_thread_flag(tsk,TIF_SIGPENDING);
	flush_sigqueue(&tsk->pending);
	if (sig) {
		flush_sigqueue(&sig->shared_pending);
133
		taskstats_tgid_free(sig);
134 135 136 137
		__cleanup_signal(sig);
	}
}

138 139 140 141 142
static void delayed_put_task_struct(struct rcu_head *rhp)
{
	put_task_struct(container_of(rhp, struct task_struct, rcu));
}

L
Linus Torvalds 已提交
143 144
void release_task(struct task_struct * p)
{
145
	struct task_struct *leader;
L
Linus Torvalds 已提交
146
	int zap_leader;
147
repeat:
L
Linus Torvalds 已提交
148
	atomic_dec(&p->user->processes);
149
	proc_flush_task(p);
L
Linus Torvalds 已提交
150
	write_lock_irq(&tasklist_lock);
151
	ptrace_unlink(p);
L
Linus Torvalds 已提交
152 153
	BUG_ON(!list_empty(&p->ptrace_list) || !list_empty(&p->ptrace_children));
	__exit_signal(p);
154

L
Linus Torvalds 已提交
155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177
	/*
	 * 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) {
		BUG_ON(leader->exit_signal == -1);
		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.
		 */
		zap_leader = (leader->exit_signal == -1);
	}

	write_unlock_irq(&tasklist_lock);
	release_thread(p);
178
	call_rcu(&p->rcu, delayed_put_task_struct);
L
Linus Torvalds 已提交
179 180 181 182 183 184 185 186 187 188

	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...
189 190
 *
 * The caller must hold rcu lock or the tasklist lock.
L
Linus Torvalds 已提交
191
 */
192
struct pid *session_of_pgrp(struct pid *pgrp)
L
Linus Torvalds 已提交
193 194
{
	struct task_struct *p;
195
	struct pid *sid = NULL;
196

197
	p = pid_task(pgrp, PIDTYPE_PGID);
198
	if (p == NULL)
199
		p = pid_task(pgrp, PIDTYPE_PID);
200
	if (p != NULL)
201
		sid = task_session(p);
202

L
Linus Torvalds 已提交
203 204 205 206 207 208 209 210 211 212 213
	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?"
 */
214
static int will_become_orphaned_pgrp(struct pid *pgrp, struct task_struct *ignored_task)
L
Linus Torvalds 已提交
215 216 217
{
	struct task_struct *p;

218
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
219 220 221
		if ((p == ignored_task) ||
		    (p->exit_state && thread_group_empty(p)) ||
		    is_global_init(p->real_parent))
L
Linus Torvalds 已提交
222
			continue;
223

224
		if (task_pgrp(p->real_parent) != pgrp &&
225 226
		    task_session(p->real_parent) == task_session(p))
			return 0;
227
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
228 229

	return 1;
L
Linus Torvalds 已提交
230 231
}

232
int is_current_pgrp_orphaned(void)
L
Linus Torvalds 已提交
233 234 235 236
{
	int retval;

	read_lock(&tasklist_lock);
237
	retval = will_become_orphaned_pgrp(task_pgrp(current), NULL);
L
Linus Torvalds 已提交
238 239 240 241 242
	read_unlock(&tasklist_lock);

	return retval;
}

243
static int has_stopped_jobs(struct pid *pgrp)
L
Linus Torvalds 已提交
244 245 246 247
{
	int retval = 0;
	struct task_struct *p;

248
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
M
Matthew Wilcox 已提交
249
		if (!task_is_stopped(p))
L
Linus Torvalds 已提交
250 251 252
			continue;
		retval = 1;
		break;
253
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
L
Linus Torvalds 已提交
254 255 256
	return retval;
}

257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287
/*
 * 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
288
/**
289
 * reparent_to_kthreadd - Reparent the calling kernel thread to kthreadd
L
Linus Torvalds 已提交
290 291
 *
 * If a kernel thread is launched as a result of a system call, or if
292 293
 * 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
294 295 296 297
 *
 * 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.
 *
298
 * NOTE that reparent_to_kthreadd() gives the caller full capabilities.
L
Linus Torvalds 已提交
299
 */
300
static void reparent_to_kthreadd(void)
L
Linus Torvalds 已提交
301 302 303 304 305
{
	write_lock_irq(&tasklist_lock);

	ptrace_unlink(current);
	/* Reparent to init */
306
	remove_parent(current);
307
	current->real_parent = current->parent = kthreadd_task;
308
	add_parent(current);
L
Linus Torvalds 已提交
309 310 311 312

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

313
	if (task_nice(current) < 0)
L
Linus Torvalds 已提交
314 315 316 317 318 319 320 321 322 323 324 325
		set_user_nice(current, 0);
	/* cpus_allowed? */
	/* rt_priority? */
	/* signals? */
	security_task_reparent_to_init(current);
	memcpy(current->signal->rlim, init_task.signal->rlim,
	       sizeof(current->signal->rlim));
	atomic_inc(&(INIT_USER->__count));
	write_unlock_irq(&tasklist_lock);
	switch_uid(INIT_USER);
}

326
void __set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
327
{
328
	struct task_struct *curr = current->group_leader;
329
	pid_t nr = pid_nr(pid);
L
Linus Torvalds 已提交
330

331
	if (task_session(curr) != pid) {
L
Linus Torvalds 已提交
332
		detach_pid(curr, PIDTYPE_SID);
333 334
		attach_pid(curr, PIDTYPE_SID, pid);
		set_task_session(curr, nr);
L
Linus Torvalds 已提交
335
	}
336
	if (task_pgrp(curr) != pid) {
L
Linus Torvalds 已提交
337
		detach_pid(curr, PIDTYPE_PGID);
338 339
		attach_pid(curr, PIDTYPE_PGID, pid);
		set_task_pgrp(curr, nr);
L
Linus Torvalds 已提交
340 341 342
	}
}

343
static void set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
344 345
{
	write_lock_irq(&tasklist_lock);
346
	__set_special_pids(pid);
L
Linus Torvalds 已提交
347 348 349 350 351 352 353 354 355 356
	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)
{
357
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377
		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)
{
378
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
379 380 381
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
382
	current->sighand->action[(sig)-1].sa.sa_handler = SIG_IGN;
L
Linus Torvalds 已提交
383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410
	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);
411 412 413 414 415
	/*
	 * We don't want to have TIF_FREEZE set if the system-wide hibernation
	 * or suspend transition begins right now.
	 */
	current->flags |= PF_NOFREEZE;
L
Linus Torvalds 已提交
416

417 418 419 420
	if (current->nsproxy != &init_nsproxy) {
		get_nsproxy(&init_nsproxy);
		switch_task_namespaces(current, &init_nsproxy);
	}
421
	set_special_pids(&init_struct_pid);
422
	proc_clear_tty(current);
L
Linus Torvalds 已提交
423 424 425 426 427 428 429 430 431 432 433 434

	/* 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
435

436
	exit_files(current);
L
Linus Torvalds 已提交
437 438 439
	current->files = init_task.files;
	atomic_inc(&current->files->count);

440
	reparent_to_kthreadd();
L
Linus Torvalds 已提交
441 442 443 444
}

EXPORT_SYMBOL(daemonize);

445
static void close_files(struct files_struct * files)
L
Linus Torvalds 已提交
446 447
{
	int i, j;
448
	struct fdtable *fdt;
L
Linus Torvalds 已提交
449 450

	j = 0;
451 452 453 454 455 456

	/*
	 * It is safe to dereference the fd table without RCU or
	 * ->file_lock because this is the last reference to the
	 * files structure.
	 */
457
	fdt = files_fdtable(files);
L
Linus Torvalds 已提交
458 459 460
	for (;;) {
		unsigned long set;
		i = j * __NFDBITS;
461
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
462
			break;
463
		set = fdt->open_fds->fds_bits[j++];
L
Linus Torvalds 已提交
464 465
		while (set) {
			if (set & 1) {
466
				struct file * file = xchg(&fdt->fd[i], NULL);
467
				if (file) {
L
Linus Torvalds 已提交
468
					filp_close(file, files);
469 470
					cond_resched();
				}
L
Linus Torvalds 已提交
471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490
			}
			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;
}

491
void put_files_struct(struct files_struct *files)
L
Linus Torvalds 已提交
492
{
493 494
	struct fdtable *fdt;

L
Linus Torvalds 已提交
495 496 497 498
	if (atomic_dec_and_test(&files->count)) {
		close_files(files);
		/*
		 * Free the fd and fdset arrays if we expanded them.
499 500 501
		 * 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
502
		 */
503
		fdt = files_fdtable(files);
504
		if (fdt != &files->fdtab)
505
			kmem_cache_free(files_cachep, files);
506
		free_fdtable(fdt);
L
Linus Torvalds 已提交
507 508 509
	}
}

510
void reset_files_struct(struct files_struct *files)
511
{
512
	struct task_struct *tsk = current;
513 514 515 516 517 518 519 520 521
	struct files_struct *old;

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

522
void exit_files(struct task_struct *tsk)
L
Linus Torvalds 已提交
523 524 525 526 527 528 529 530 531 532 533
{
	struct files_struct * files = tsk->files;

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

534
void put_fs_struct(struct fs_struct *fs)
L
Linus Torvalds 已提交
535 536 537
{
	/* No need to hold fs->lock if we are killing it */
	if (atomic_dec_and_test(&fs->count)) {
J
Jan Blunck 已提交
538 539 540 541
		path_put(&fs->root);
		path_put(&fs->pwd);
		if (fs->altroot.dentry)
			path_put(&fs->altroot);
L
Linus Torvalds 已提交
542 543 544 545
		kmem_cache_free(fs_cachep, fs);
	}
}

546
void exit_fs(struct task_struct *tsk)
L
Linus Torvalds 已提交
547 548 549 550 551 552 553
{
	struct fs_struct * fs = tsk->fs;

	if (fs) {
		task_lock(tsk);
		tsk->fs = NULL;
		task_unlock(tsk);
554
		put_fs_struct(fs);
L
Linus Torvalds 已提交
555 556 557 558 559
	}
}

EXPORT_SYMBOL_GPL(exit_fs);

560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641
#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 (!mm)
		return 0;
	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);
	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);
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
	if (c->mm != mm) {
		task_unlock(c);
		put_task_struct(c);
		goto retry;
	}
	cgroup_mm_owner_callbacks(mm->owner, c);
	mm->owner = c;
	task_unlock(c);
	put_task_struct(c);
}
#endif /* CONFIG_MM_OWNER */

L
Linus Torvalds 已提交
642 643 644 645
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
646
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671
{
	struct mm_struct *mm = tsk->mm;

	mm_release(tsk, mm);
	if (!mm)
		return;
	/*
	 * Serialize with any possible pending coredump.
	 * We must hold mmap_sem around checking core_waiters
	 * and clearing tsk->mm.  The core-inducing thread
	 * will increment core_waiters for each thread in the
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
	if (mm->core_waiters) {
		up_read(&mm->mmap_sem);
		down_write(&mm->mmap_sem);
		if (!--mm->core_waiters)
			complete(mm->core_startup_done);
		up_write(&mm->mmap_sem);

		wait_for_completion(&mm->core_done);
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
672
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
673 674 675 676 677
	/* more a memory barrier than a real lock */
	task_lock(tsk);
	tsk->mm = NULL;
	up_read(&mm->mmap_sem);
	enter_lazy_tlb(mm, current);
678 679
	/* We don't want this task to be frozen prematurely */
	clear_freeze_flag(tsk);
L
Linus Torvalds 已提交
680
	task_unlock(tsk);
681
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
682 683 684
	mmput(mm);
}

685 686
static void
reparent_thread(struct task_struct *p, struct task_struct *father, int traced)
L
Linus Torvalds 已提交
687
{
688 689 690 691
	if (p->pdeath_signal)
		/* We already hold the tasklist_lock here.  */
		group_send_sig_info(p->pdeath_signal, SEND_SIG_NOINFO, p);

L
Linus Torvalds 已提交
692 693 694 695 696 697 698 699 700 701 702
	/* Move the child from its dying parent to the new one.  */
	if (unlikely(traced)) {
		/* Preserve ptrace links if someone else is tracing this child.  */
		list_del_init(&p->ptrace_list);
		if (p->parent != p->real_parent)
			list_add(&p->ptrace_list, &p->real_parent->ptrace_children);
	} else {
		/* If this child is being traced, then we're the one tracing it
		 * anyway, so let go of it.
		 */
		p->ptrace = 0;
703
		remove_parent(p);
L
Linus Torvalds 已提交
704
		p->parent = p->real_parent;
705
		add_parent(p);
L
Linus Torvalds 已提交
706

M
Matthew Wilcox 已提交
707
		if (task_is_traced(p)) {
L
Linus Torvalds 已提交
708 709 710 711 712 713 714 715 716
			/*
			 * If it was at a trace stop, turn it into
			 * a normal stop since it's no longer being
			 * traced.
			 */
			ptrace_untrace(p);
		}
	}

717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
	/* If this is a threaded reparent there is no need to
	 * notify anyone anything has happened.
	 */
	if (p->real_parent->group_leader == father->group_leader)
		return;

	/* We don't want people slaying init.  */
	if (p->exit_signal != -1)
		p->exit_signal = SIGCHLD;

	/* If we'd notified the old parent about this child's death,
	 * also notify the new parent.
	 */
	if (!traced && p->exit_state == EXIT_ZOMBIE &&
	    p->exit_signal != -1 && thread_group_empty(p))
		do_notify_parent(p, p->exit_signal);

734
	kill_orphaned_pgrp(p, father);
L
Linus Torvalds 已提交
735 736 737 738 739 740
}

/*
 * 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
741 742
 * the child reaper process (ie "init") in our pid
 * space.
L
Linus Torvalds 已提交
743
 */
744
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
745
{
746
	struct task_struct *p, *n, *reaper = father;
747 748 749 750 751
	struct list_head ptrace_dead;

	INIT_LIST_HEAD(&ptrace_dead);

	write_lock_irq(&tasklist_lock);
L
Linus Torvalds 已提交
752 753 754 755

	do {
		reaper = next_thread(reaper);
		if (reaper == father) {
756
			reaper = task_child_reaper(father);
L
Linus Torvalds 已提交
757 758
			break;
		}
759
	} while (reaper->flags & PF_EXITING);
L
Linus Torvalds 已提交
760 761 762 763 764 765 766 767 768

	/*
	 * There are only two places where our children can be:
	 *
	 * - in our child list
	 * - in our ptraced child list
	 *
	 * Search them and reparent children.
	 */
769
	list_for_each_entry_safe(p, n, &father->children, sibling) {
L
Linus Torvalds 已提交
770 771 772 773 774 775 776 777 778
		int ptrace;

		ptrace = p->ptrace;

		/* if father isn't the real parent, then ptrace must be enabled */
		BUG_ON(father != p->real_parent && !ptrace);

		if (father == p->real_parent) {
			/* reparent with a reaper, real father it's us */
779
			p->real_parent = reaper;
L
Linus Torvalds 已提交
780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
			reparent_thread(p, father, 0);
		} else {
			/* reparent ptraced task to its real parent */
			__ptrace_unlink (p);
			if (p->exit_state == EXIT_ZOMBIE && p->exit_signal != -1 &&
			    thread_group_empty(p))
				do_notify_parent(p, p->exit_signal);
		}

		/*
		 * if the ptraced child is a zombie with exit_signal == -1
		 * we must collect it before we exit, or it will remain
		 * zombie forever since we prevented it from self-reap itself
		 * while it was being traced by us, to be able to see it in wait4.
		 */
		if (unlikely(ptrace && p->exit_state == EXIT_ZOMBIE && p->exit_signal == -1))
796
			list_add(&p->ptrace_list, &ptrace_dead);
L
Linus Torvalds 已提交
797
	}
798

799
	list_for_each_entry_safe(p, n, &father->ptrace_children, ptrace_list) {
800
		p->real_parent = reaper;
L
Linus Torvalds 已提交
801 802
		reparent_thread(p, father, 1);
	}
803 804 805 806 807 808 809 810 811 812

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

	list_for_each_entry_safe(p, n, &ptrace_dead, ptrace_list) {
		list_del_init(&p->ptrace_list);
		release_task(p);
	}

L
Linus Torvalds 已提交
813 814 815 816 817 818
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
819
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
820 821 822 823 824 825 826 827 828 829 830
{
	int state;

	/*
	 * 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)
	 */
831
	forget_original_parent(tsk);
832
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
833

834
	write_lock_irq(&tasklist_lock);
835 836
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
837

838
	/* Let father know we died
L
Linus Torvalds 已提交
839 840
	 *
	 * Thread signals are configurable, but you aren't going to use
841
	 * that to send signals to arbitary processes.
L
Linus Torvalds 已提交
842 843 844 845 846 847 848 849 850 851 852
	 * 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.
	 */
	if (tsk->exit_signal != SIGCHLD && tsk->exit_signal != -1 &&
853 854
	    (tsk->parent_exec_id != tsk->real_parent->self_exec_id ||
	     tsk->self_exec_id != tsk->parent_exec_id)
L
Linus Torvalds 已提交
855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
	    && !capable(CAP_KILL))
		tsk->exit_signal = SIGCHLD;


	/* If something other than our normal parent is ptracing us, then
	 * send it a SIGCHLD instead of honoring exit_signal.  exit_signal
	 * only has special meaning to our real parent.
	 */
	if (tsk->exit_signal != -1 && thread_group_empty(tsk)) {
		int signal = tsk->parent == tsk->real_parent ? tsk->exit_signal : SIGCHLD;
		do_notify_parent(tsk, signal);
	} else if (tsk->ptrace) {
		do_notify_parent(tsk, SIGCHLD);
	}

	state = EXIT_ZOMBIE;
871
	if (tsk->exit_signal == -1 && likely(!tsk->ptrace))
L
Linus Torvalds 已提交
872 873 874
		state = EXIT_DEAD;
	tsk->exit_state = state;

875 876 877 878 879
	if (thread_group_leader(tsk) &&
	    tsk->signal->notify_count < 0 &&
	    tsk->signal->group_exit_task)
		wake_up_process(tsk->signal->group_exit_task);

L
Linus Torvalds 已提交
880 881 882 883 884 885 886
	write_unlock_irq(&tasklist_lock);

	/* If the process is dead, release it - nobody will wait for it */
	if (state == EXIT_DEAD)
		release_task(tsk);
}

887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914
#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 *n = end_of_stack(current);
	unsigned long free;

	while (*n == 0)
		n++;
	free = (unsigned long)n - (unsigned long)end_of_stack(current);

	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

915 916
static inline void exit_child_reaper(struct task_struct *tsk)
{
917
	if (likely(tsk->group_leader != task_child_reaper(tsk)))
918 919
		return;

920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945
	if (tsk->nsproxy->pid_ns == &init_pid_ns)
		panic("Attempted to kill init!");

	/*
	 * @tsk is the last thread in the 'cgroup-init' and is exiting.
	 * Terminate all remaining processes in the namespace and reap them
	 * before exiting @tsk.
	 *
	 * Note that @tsk (last thread of cgroup-init) may not necessarily
	 * be the child-reaper (i.e main thread of cgroup-init) of the
	 * namespace i.e the child_reaper may have already exited.
	 *
	 * Even after a child_reaper exits, we let it inherit orphaned children,
	 * because, pid_ns->child_reaper remains valid as long as there is
	 * at least one living sub-thread in the cgroup init.

	 * This living sub-thread of the cgroup-init will be notified when
	 * a child inherited by the 'child-reaper' exits (do_notify_parent()
	 * uses __group_send_sig_info()). Further, when reaping child processes,
	 * do_wait() iterates over children of all living sub threads.

	 * i.e even though 'child_reaper' thread is listed as the parent of the
	 * orphaned children, any living sub-thread in the cgroup-init can
	 * perform the role of the child_reaper.
	 */
	zap_pid_ns_processes(tsk->nsproxy->pid_ns);
946 947
}

948
NORET_TYPE void do_exit(long code)
L
Linus Torvalds 已提交
949 950 951 952 953 954
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

955 956
	WARN_ON(atomic_read(&tsk->fs_excl));

L
Linus Torvalds 已提交
957 958 959 960 961 962 963 964 965 966
	if (unlikely(in_interrupt()))
		panic("Aiee, killing interrupt handler!");
	if (unlikely(!tsk->pid))
		panic("Attempted to kill the idle task!");

	if (unlikely(current->ptrace & PT_TRACE_EXIT)) {
		current->ptrace_message = code;
		ptrace_notify((PTRACE_EVENT_EXIT << 8) | SIGTRAP);
	}

967 968 969 970 971 972 973
	/*
	 * 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");
974 975 976 977 978 979 980 981 982 983
		/*
		 * 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;
984 985
		if (tsk->io_context)
			exit_io_context();
986 987 988 989
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

O
Oleg Nesterov 已提交
990
	exit_signals(tsk);  /* sets PF_EXITING */
991 992 993 994
	/*
	 * tsk->flags are checked in the futex code to protect against
	 * an exiting task cleaning up the robust pi futexes.
	 */
995 996
	smp_mb();
	spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
997 998 999

	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
1000
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
1001 1002 1003
				preempt_count());

	acct_update_integrals(tsk);
1004 1005 1006 1007
	if (tsk->mm) {
		update_hiwater_rss(tsk->mm);
		update_hiwater_vm(tsk->mm);
	}
L
Linus Torvalds 已提交
1008
	group_dead = atomic_dec_and_test(&tsk->signal->live);
1009
	if (group_dead) {
1010
		exit_child_reaper(tsk);
1011
		hrtimer_cancel(&tsk->signal->real_timer);
1012
		exit_itimers(tsk->signal);
1013
	}
1014
	acct_collect(code, group_dead);
1015
#ifdef CONFIG_FUTEX
1016 1017
	if (unlikely(tsk->robust_list))
		exit_robust_list(tsk);
1018
#ifdef CONFIG_COMPAT
1019 1020
	if (unlikely(tsk->compat_robust_list))
		compat_exit_robust_list(tsk);
1021
#endif
1022
#endif
M
Miloslav Trmac 已提交
1023 1024
	if (group_dead)
		tty_audit_exit();
1025 1026
	if (unlikely(tsk->audit_context))
		audit_free(tsk);
1027

1028
	tsk->exit_code = code;
1029
	taskstats_exit(tsk, group_dead);
1030

L
Linus Torvalds 已提交
1031 1032
	exit_mm(tsk);

1033
	if (group_dead)
1034
		acct_process();
L
Linus Torvalds 已提交
1035
	exit_sem(tsk);
1036 1037
	exit_files(tsk);
	exit_fs(tsk);
1038
	check_stack_usage();
L
Linus Torvalds 已提交
1039
	exit_thread();
1040
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
1041 1042 1043 1044 1045
	exit_keys(tsk);

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

A
Al Viro 已提交
1046
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
1047 1048 1049
	if (tsk->binfmt)
		module_put(tsk->binfmt->module);

M
Matt Helsley 已提交
1050
	proc_exit_connector(tsk);
1051
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
1052
#ifdef CONFIG_NUMA
1053
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
1054 1055
	tsk->mempolicy = NULL;
#endif
1056
#ifdef CONFIG_FUTEX
1057 1058 1059 1060 1061 1062 1063 1064
	/*
	 * 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);
1065
#endif
1066
	/*
1067
	 * Make sure we are holding no locks:
1068
	 */
1069
	debug_check_no_locks_held(tsk);
1070 1071 1072 1073 1074 1075
	/*
	 * 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
1076

1077 1078 1079
	if (tsk->io_context)
		exit_io_context();

1080 1081 1082
	if (tsk->splice_pipe)
		__free_pipe_info(tsk->splice_pipe);

C
Coywolf Qi Hunt 已提交
1083
	preempt_disable();
O
Oleg Nesterov 已提交
1084
	/* causes final put_task_struct in finish_task_switch(). */
1085
	tsk->state = TASK_DEAD;
C
Coywolf Qi Hunt 已提交
1086

L
Linus Torvalds 已提交
1087 1088 1089
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
1090 1091
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
1092 1093
}

1094 1095
EXPORT_SYMBOL_GPL(do_exit);

L
Linus Torvalds 已提交
1096 1097 1098 1099
NORET_TYPE void complete_and_exit(struct completion *comp, long code)
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
1100

L
Linus Torvalds 已提交
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

asmlinkage long sys_exit(int error_code)
{
	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)
{
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

	if (current->signal->flags & SIGNAL_GROUP_EXIT)
		exit_code = current->signal->group_exit_code;
	else if (!thread_group_empty(current)) {
		struct signal_struct *const sig = current->signal;
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1126
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1127 1128 1129 1130
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1131
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150
			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.
 */
asmlinkage void sys_exit_group(int error_code)
{
	do_group_exit((error_code & 0xff) << 8);
}

E
Eric W. Biederman 已提交
1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
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
1163
{
1164 1165
	int err;

E
Eric W. Biederman 已提交
1166 1167
	if (type < PIDTYPE_MAX) {
		if (task_pid_type(p, type) != pid)
L
Linus Torvalds 已提交
1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186
			return 0;
	}

	/*
	 * Do not consider detached threads that are
	 * not ptraced:
	 */
	if (p->exit_signal == -1 && !p->ptrace)
		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;

1187
	err = security_task_wait(p);
O
Oleg Nesterov 已提交
1188 1189
	if (likely(!err))
		return 1;
L
Linus Torvalds 已提交
1190

E
Eric W. Biederman 已提交
1191
	if (type != PIDTYPE_PID)
O
Oleg Nesterov 已提交
1192 1193 1194 1195
		return 0;
	/* This child was explicitly requested, abort */
	read_unlock(&tasklist_lock);
	return err;
L
Linus Torvalds 已提交
1196 1197
}

1198
static int wait_noreap_copyout(struct task_struct *p, pid_t pid, uid_t uid,
L
Linus Torvalds 已提交
1199 1200 1201 1202 1203
			       int why, int status,
			       struct siginfo __user *infop,
			       struct rusage __user *rusagep)
{
	int retval = rusagep ? getrusage(p, RUSAGE_BOTH, rusagep) : 0;
1204

L
Linus Torvalds 已提交
1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228
	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.
 */
1229
static int wait_task_zombie(struct task_struct *p, int noreap,
L
Linus Torvalds 已提交
1230 1231 1232 1233
			    struct siginfo __user *infop,
			    int __user *stat_addr, struct rusage __user *ru)
{
	unsigned long state;
1234
	int retval, status, traced;
1235
	pid_t pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264

	if (unlikely(noreap)) {
		uid_t uid = p->uid;
		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;
	}

1265 1266 1267 1268
	/* traced means p->ptrace, but not vice versa */
	traced = (p->real_parent != p->parent);

	if (likely(!traced)) {
1269 1270 1271
		struct signal_struct *psig;
		struct signal_struct *sig;

L
Linus Torvalds 已提交
1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
		/*
		 * 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.
		 */
		spin_lock_irq(&p->parent->sighand->siglock);
1288 1289 1290 1291
		psig = p->parent->signal;
		sig = p->signal;
		psig->cutime =
			cputime_add(psig->cutime,
L
Linus Torvalds 已提交
1292
			cputime_add(p->utime,
1293 1294 1295 1296
			cputime_add(sig->utime,
				    sig->cutime)));
		psig->cstime =
			cputime_add(psig->cstime,
L
Linus Torvalds 已提交
1297
			cputime_add(p->stime,
1298 1299
			cputime_add(sig->stime,
				    sig->cstime)));
1300 1301 1302 1303 1304
		psig->cgtime =
			cputime_add(psig->cgtime,
			cputime_add(p->gtime,
			cputime_add(sig->gtime,
				    sig->cgtime)));
1305 1306 1307 1308 1309 1310 1311 1312
		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;
1313 1314 1315 1316 1317 1318
		psig->cinblock +=
			task_io_get_inblock(p) +
			sig->inblock + sig->cinblock;
		psig->coublock +=
			task_io_get_oublock(p) +
			sig->oublock + sig->coublock;
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351
		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)
1352
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1353 1354
	if (!retval && infop)
		retval = put_user(p->uid, &infop->si_uid);
1355
	if (!retval)
1356
		retval = pid;
1357 1358

	if (traced) {
L
Linus Torvalds 已提交
1359
		write_lock_irq(&tasklist_lock);
1360 1361 1362 1363 1364 1365 1366 1367 1368
		/* 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.
		 */
		if (p->exit_signal != -1) {
			do_notify_parent(p, p->exit_signal);
L
Linus Torvalds 已提交
1369
			if (p->exit_signal != -1) {
1370 1371
				p->exit_state = EXIT_ZOMBIE;
				p = NULL;
L
Linus Torvalds 已提交
1372 1373 1374 1375 1376 1377
			}
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1378

L
Linus Torvalds 已提交
1379 1380 1381 1382 1383 1384 1385 1386 1387
	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.
 */
1388
static int wait_task_stopped(struct task_struct *p,
1389
			     int noreap, struct siginfo __user *infop,
L
Linus Torvalds 已提交
1390 1391
			     int __user *stat_addr, struct rusage __user *ru)
{
1392 1393
	int retval, exit_code, why;
	uid_t uid = 0; /* unneeded, required by compiler */
1394
	pid_t pid;
L
Linus Torvalds 已提交
1395

1396 1397 1398 1399 1400 1401
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

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

1402
	if (!(p->ptrace & PT_PTRACED) && p->signal->group_stop_count > 0)
L
Linus Torvalds 已提交
1403 1404 1405 1406
		/*
		 * A group stop is in progress and this is the group leader.
		 * We won't report until all threads have stopped.
		 */
1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419
		goto unlock_sig;

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

	if (!noreap)
		p->exit_code = 0;

	uid = p->uid;
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1420 1421 1422 1423 1424 1425 1426 1427 1428 1429
		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);
1430
	pid = task_pid_vnr(p);
1431
	why = (p->ptrace & PT_PTRACED) ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1432 1433
	read_unlock(&tasklist_lock);

1434
	if (unlikely(noreap))
L
Linus Torvalds 已提交
1435
		return wait_noreap_copyout(p, pid, uid,
1436
					   why, exit_code,
L
Linus Torvalds 已提交
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446
					   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
1447
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1448 1449 1450
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1451
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1452
	if (!retval && infop)
1453
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1454
	if (!retval)
1455
		retval = pid;
L
Linus Torvalds 已提交
1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
	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.
 */
1468
static int wait_task_continued(struct task_struct *p, int noreap,
L
Linus Torvalds 已提交
1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488
			       struct siginfo __user *infop,
			       int __user *stat_addr, struct rusage __user *ru)
{
	int retval;
	pid_t pid;
	uid_t uid;

	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;
	}
	if (!noreap)
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
	spin_unlock_irq(&p->sighand->siglock);

1489
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1490 1491 1492 1493 1494 1495 1496 1497 1498 1499
	uid = p->uid;
	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)
1500
			retval = pid;
L
Linus Torvalds 已提交
1501 1502 1503 1504 1505 1506 1507 1508 1509 1510
	} else {
		retval = wait_noreap_copyout(p, pid, uid,
					     CLD_CONTINUED, SIGCONT,
					     infop, ru);
		BUG_ON(retval == 0);
	}

	return retval;
}

E
Eric W. Biederman 已提交
1511 1512 1513
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
1514 1515 1516 1517 1518 1519 1520
{
	DECLARE_WAITQUEUE(wait, current);
	struct task_struct *tsk;
	int flag, retval;

	add_wait_queue(&current->signal->wait_chldexit,&wait);
repeat:
E
Eric W. Biederman 已提交
1521 1522 1523 1524 1525
	/* If there is nothing that can match our critier just get out */
	retval = -ECHILD;
	if ((type < PIDTYPE_MAX) && (!pid || hlist_empty(&pid->tasks[type])))
		goto end;

L
Linus Torvalds 已提交
1526 1527 1528 1529
	/*
	 * We will set this flag if we see any child that might later
	 * match our criteria, even if we are not able to reap it yet.
	 */
O
Oleg Nesterov 已提交
1530
	flag = retval = 0;
L
Linus Torvalds 已提交
1531 1532 1533 1534 1535 1536
	current->state = TASK_INTERRUPTIBLE;
	read_lock(&tasklist_lock);
	tsk = current;
	do {
		struct task_struct *p;

1537
		list_for_each_entry(p, &tsk->children, sibling) {
E
Eric W. Biederman 已提交
1538
			int ret = eligible_child(type, pid, options, p);
L
Linus Torvalds 已提交
1539 1540 1541
			if (!ret)
				continue;

1542
			if (unlikely(ret < 0)) {
O
Oleg Nesterov 已提交
1543 1544
				retval = ret;
			} else if (task_is_stopped_or_traced(p)) {
L
Linus Torvalds 已提交
1545 1546 1547 1548 1549
				/*
				 * It's stopped now, so it might later
				 * continue, exit, or stop again.
				 */
				flag = 1;
O
Oleg Nesterov 已提交
1550 1551 1552
				if (!(p->ptrace & PT_PTRACED) &&
				    !(options & WUNTRACED))
					continue;
M
Matthew Wilcox 已提交
1553

1554
				retval = wait_task_stopped(p,
M
Matthew Wilcox 已提交
1555 1556
						(options & WNOWAIT), infop,
						stat_addr, ru);
1557 1558
			} else if (p->exit_state == EXIT_ZOMBIE &&
					!delay_group_leader(p)) {
M
Matthew Wilcox 已提交
1559
				/*
1560
				 * We don't reap group leaders with subthreads.
M
Matthew Wilcox 已提交
1561 1562
				 */
				if (!likely(options & WEXITED))
L
Linus Torvalds 已提交
1563
					continue;
M
Matthew Wilcox 已提交
1564 1565 1566
				retval = wait_task_zombie(p,
						(options & WNOWAIT), infop,
						stat_addr, ru);
1567
			} else if (p->exit_state != EXIT_DEAD) {
L
Linus Torvalds 已提交
1568 1569 1570 1571 1572 1573 1574
				/*
				 * It's running now, so it might later
				 * exit, stop, or stop and then continue.
				 */
				flag = 1;
				if (!unlikely(options & WCONTINUED))
					continue;
M
Matthew Wilcox 已提交
1575 1576 1577
				retval = wait_task_continued(p,
						(options & WNOWAIT), infop,
						stat_addr, ru);
L
Linus Torvalds 已提交
1578
			}
1579 1580
			if (retval != 0) /* tasklist_lock released */
				goto end;
L
Linus Torvalds 已提交
1581 1582
		}
		if (!flag) {
1583
			list_for_each_entry(p, &tsk->ptrace_children,
O
Oleg Nesterov 已提交
1584
								ptrace_list) {
E
Eric W. Biederman 已提交
1585
				flag = eligible_child(type, pid, options, p);
O
Oleg Nesterov 已提交
1586
				if (!flag)
L
Linus Torvalds 已提交
1587
					continue;
O
Oleg Nesterov 已提交
1588 1589 1590 1591
				if (likely(flag > 0))
					break;
				retval = flag;
				goto end;
L
Linus Torvalds 已提交
1592 1593 1594 1595 1596
			}
		}
		if (options & __WNOTHREAD)
			break;
		tsk = next_thread(tsk);
E
Eric Sesterhenn 已提交
1597
		BUG_ON(tsk->signal != current->signal);
L
Linus Torvalds 已提交
1598 1599
	} while (tsk != current);
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1600

L
Linus Torvalds 已提交
1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615
	if (flag) {
		if (options & WNOHANG)
			goto end;
		retval = -ERESTARTSYS;
		if (signal_pending(current))
			goto end;
		schedule();
		goto repeat;
	}
	retval = -ECHILD;
end:
	current->state = TASK_RUNNING;
	remove_wait_queue(&current->signal->wait_chldexit,&wait);
	if (infop) {
		if (retval > 0)
1616
			retval = 0;
L
Linus Torvalds 已提交
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639
		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;
}

E
Eric W. Biederman 已提交
1640
asmlinkage long sys_waitid(int which, pid_t upid,
L
Linus Torvalds 已提交
1641 1642 1643
			   struct siginfo __user *infop, int options,
			   struct rusage __user *ru)
{
E
Eric W. Biederman 已提交
1644 1645
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1646 1647 1648 1649 1650 1651 1652 1653 1654
	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
1655
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1656 1657
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1658 1659
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1660 1661 1662
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1663 1664
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1665 1666 1667 1668 1669 1670
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1671 1672 1673 1674
	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
1675 1676

	/* avoid REGPARM breakage on x86: */
1677
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1678 1679 1680
	return ret;
}

E
Eric W. Biederman 已提交
1681
asmlinkage long sys_wait4(pid_t upid, int __user *stat_addr,
L
Linus Torvalds 已提交
1682 1683
			  int options, struct rusage __user *ru)
{
E
Eric W. Biederman 已提交
1684 1685
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1686 1687 1688 1689 1690
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706

	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
1707 1708

	/* avoid REGPARM breakage on x86: */
1709
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
asmlinkage long sys_waitpid(pid_t pid, int __user *stat_addr, int options)
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif