exit.c 43.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10
/*
 *  linux/kernel/exit.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/module.h>
11
#include <linux/capability.h>
L
Linus Torvalds 已提交
12 13 14
#include <linux/completion.h>
#include <linux/personality.h>
#include <linux/tty.h>
15
#include <linux/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>
49
#include <linux/fs_struct.h>
D
David Howells 已提交
50
#include <linux/init_task.h>
51
#include <trace/sched.h>
L
Linus Torvalds 已提交
52 53 54 55 56

#include <asm/uaccess.h>
#include <asm/unistd.h>
#include <asm/pgtable.h>
#include <asm/mmu_context.h>
D
David Howells 已提交
57
#include "cred-internals.h"
L
Linus Torvalds 已提交
58

59 60 61
DEFINE_TRACE(sched_process_free);
DEFINE_TRACE(sched_process_exit);
DEFINE_TRACE(sched_process_wait);
L
Linus Torvalds 已提交
62

A
Adrian Bunk 已提交
63 64
static void exit_mm(struct task_struct * tsk);

L
Linus Torvalds 已提交
65 66 67 68 69 70 71
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);
72

73
		list_del_rcu(&p->tasks);
74
		__get_cpu_var(process_counts)--;
L
Linus Torvalds 已提交
75
	}
O
Oleg Nesterov 已提交
76
	list_del_rcu(&p->thread_group);
R
Roland McGrath 已提交
77
	list_del_init(&p->sibling);
L
Linus Torvalds 已提交
78 79
}

80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101
/*
 * 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:
		 */
102
		if (sig->group_exit_task && atomic_read(&sig->count) == sig->notify_count)
103
			wake_up_process(sig->group_exit_task);
104

105 106 107 108 109 110 111 112 113 114 115 116
		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.
		 */
117 118
		sig->utime = cputime_add(sig->utime, task_utime(tsk));
		sig->stime = cputime_add(sig->stime, task_stime(tsk));
119
		sig->gtime = cputime_add(sig->gtime, task_gtime(tsk));
120 121 122 123
		sig->min_flt += tsk->min_flt;
		sig->maj_flt += tsk->maj_flt;
		sig->nvcsw += tsk->nvcsw;
		sig->nivcsw += tsk->nivcsw;
124 125
		sig->inblock += task_io_get_inblock(tsk);
		sig->oublock += task_io_get_oublock(tsk);
126
		task_io_accounting_add(&sig->ioac, &tsk->ioac);
127
		sig->sum_sched_runtime += tsk->se.sum_exec_runtime;
128 129 130
		sig = NULL; /* Marker for below. */
	}

131 132
	__unhash_process(tsk);

133 134 135 136 137 138
	/*
	 * Do this under ->siglock, we can race with another thread
	 * doing sigqueue_free() if we have SIGQUEUE_PREALLOC signals.
	 */
	flush_sigqueue(&tsk->pending);

139
	tsk->signal = NULL;
140
	tsk->sighand = NULL;
141 142
	spin_unlock(&sighand->siglock);

143
	__cleanup_sighand(sighand);
144 145 146
	clear_tsk_thread_flag(tsk,TIF_SIGPENDING);
	if (sig) {
		flush_sigqueue(&sig->shared_pending);
147
		taskstats_tgid_free(sig);
148 149 150 151 152
		/*
		 * Make sure ->signal can't go away under rq->lock,
		 * see account_group_exec_runtime().
		 */
		task_rq_unlock_wait(tsk);
153 154 155 156
		__cleanup_signal(sig);
	}
}

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

161 162 163
#ifdef CONFIG_PERF_COUNTERS
	WARN_ON_ONCE(!list_empty(&tsk->perf_counter_ctx.counter_list));
#endif
164 165
	trace_sched_process_free(tsk);
	put_task_struct(tsk);
166 167
}

R
Roland McGrath 已提交
168

L
Linus Torvalds 已提交
169 170
void release_task(struct task_struct * p)
{
171
	struct task_struct *leader;
L
Linus Torvalds 已提交
172
	int zap_leader;
173
repeat:
R
Roland McGrath 已提交
174
	tracehook_prepare_release_task(p);
175 176 177 178
	/* 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);

179
	proc_flush_task(p);
L
Linus Torvalds 已提交
180
	write_lock_irq(&tasklist_lock);
R
Roland McGrath 已提交
181
	tracehook_finish_release_task(p);
L
Linus Torvalds 已提交
182
	__exit_signal(p);
183

L
Linus Torvalds 已提交
184 185 186 187 188 189 190 191
	/*
	 * 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) {
192
		BUG_ON(task_detached(leader));
L
Linus Torvalds 已提交
193 194 195 196 197 198 199 200 201
		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.
		 */
202
		zap_leader = task_detached(leader);
R
Roland McGrath 已提交
203 204 205 206 207 208 209

		/*
		 * 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
210 211 212 213
	}

	write_unlock_irq(&tasklist_lock);
	release_thread(p);
214
	call_rcu(&p->rcu, delayed_put_task_struct);
L
Linus Torvalds 已提交
215 216 217 218 219 220 221 222 223 224

	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...
225 226
 *
 * The caller must hold rcu lock or the tasklist lock.
L
Linus Torvalds 已提交
227
 */
228
struct pid *session_of_pgrp(struct pid *pgrp)
L
Linus Torvalds 已提交
229 230
{
	struct task_struct *p;
231
	struct pid *sid = NULL;
232

233
	p = pid_task(pgrp, PIDTYPE_PGID);
234
	if (p == NULL)
235
		p = pid_task(pgrp, PIDTYPE_PID);
236
	if (p != NULL)
237
		sid = task_session(p);
238

L
Linus Torvalds 已提交
239 240 241 242 243 244 245 246 247 248 249
	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?"
 */
250
static int will_become_orphaned_pgrp(struct pid *pgrp, struct task_struct *ignored_task)
L
Linus Torvalds 已提交
251 252 253
{
	struct task_struct *p;

254
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
255 256 257
		if ((p == ignored_task) ||
		    (p->exit_state && thread_group_empty(p)) ||
		    is_global_init(p->real_parent))
L
Linus Torvalds 已提交
258
			continue;
259

260
		if (task_pgrp(p->real_parent) != pgrp &&
261 262
		    task_session(p->real_parent) == task_session(p))
			return 0;
263
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
264 265

	return 1;
L
Linus Torvalds 已提交
266 267
}

268
int is_current_pgrp_orphaned(void)
L
Linus Torvalds 已提交
269 270 271 272
{
	int retval;

	read_lock(&tasklist_lock);
273
	retval = will_become_orphaned_pgrp(task_pgrp(current), NULL);
L
Linus Torvalds 已提交
274 275 276 277 278
	read_unlock(&tasklist_lock);

	return retval;
}

279
static int has_stopped_jobs(struct pid *pgrp)
L
Linus Torvalds 已提交
280 281 282 283
{
	int retval = 0;
	struct task_struct *p;

284
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
M
Matthew Wilcox 已提交
285
		if (!task_is_stopped(p))
L
Linus Torvalds 已提交
286 287 288
			continue;
		retval = 1;
		break;
289
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
L
Linus Torvalds 已提交
290 291 292
	return retval;
}

293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323
/*
 * 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
324
/**
325
 * reparent_to_kthreadd - Reparent the calling kernel thread to kthreadd
L
Linus Torvalds 已提交
326 327
 *
 * If a kernel thread is launched as a result of a system call, or if
328 329
 * 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
330 331 332 333
 *
 * 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.
 *
334
 * NOTE that reparent_to_kthreadd() gives the caller full capabilities.
L
Linus Torvalds 已提交
335
 */
336
static void reparent_to_kthreadd(void)
L
Linus Torvalds 已提交
337 338 339 340 341
{
	write_lock_irq(&tasklist_lock);

	ptrace_unlink(current);
	/* Reparent to init */
342
	current->real_parent = current->parent = kthreadd_task;
R
Roland McGrath 已提交
343
	list_move_tail(&current->sibling, &current->real_parent->children);
L
Linus Torvalds 已提交
344 345 346 347

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

348
	if (task_nice(current) < 0)
L
Linus Torvalds 已提交
349 350 351 352 353 354
		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
355 356 357

	atomic_inc(&init_cred.usage);
	commit_creds(&init_cred);
L
Linus Torvalds 已提交
358 359 360
	write_unlock_irq(&tasklist_lock);
}

361
void __set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
362
{
363
	struct task_struct *curr = current->group_leader;
L
Linus Torvalds 已提交
364

365
	if (task_session(curr) != pid)
366
		change_pid(curr, PIDTYPE_SID, pid);
367 368

	if (task_pgrp(curr) != pid)
369
		change_pid(curr, PIDTYPE_PGID, pid);
L
Linus Torvalds 已提交
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
	recalc_sigpending();
	spin_unlock_irq(&current->sighand->siglock);
	return 0;
}

EXPORT_SYMBOL(disallow_signal);

/*
 *	Put all the gunge required to become a kernel thread without
 *	attached user resources in one place where it belongs.
 */

void daemonize(const char *name, ...)
{
	va_list args;
	sigset_t blocked;

	va_start(args, name);
	vsnprintf(current->comm, sizeof(current->comm), name, args);
	va_end(args);

	/*
	 * If we were started as result of loading a module, close all of the
	 * user space pages.  We don't need them, and if we didn't close them
	 * they would be locked into memory.
	 */
	exit_mm(current);
439 440 441 442
	/*
	 * We don't want to have TIF_FREEZE set if the system-wide hibernation
	 * or suspend transition begins right now.
	 */
O
Oleg Nesterov 已提交
443
	current->flags |= (PF_NOFREEZE | PF_KTHREAD);
L
Linus Torvalds 已提交
444

445 446 447 448
	if (current->nsproxy != &init_nsproxy) {
		get_nsproxy(&init_nsproxy);
		switch_task_namespaces(current, &init_nsproxy);
	}
449
	set_special_pids(&init_struct_pid);
450
	proc_clear_tty(current);
L
Linus Torvalds 已提交
451 452 453 454 455 456 457 458

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

	/* Become as one with the init task */

459
	daemonize_fs_struct();
460
	exit_files(current);
L
Linus Torvalds 已提交
461 462 463
	current->files = init_task.files;
	atomic_inc(&current->files->count);

464
	reparent_to_kthreadd();
L
Linus Torvalds 已提交
465 466 467 468
}

EXPORT_SYMBOL(daemonize);

469
static void close_files(struct files_struct * files)
L
Linus Torvalds 已提交
470 471
{
	int i, j;
472
	struct fdtable *fdt;
L
Linus Torvalds 已提交
473 474

	j = 0;
475 476 477 478 479 480

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

515
void put_files_struct(struct files_struct *files)
L
Linus Torvalds 已提交
516
{
517 518
	struct fdtable *fdt;

L
Linus Torvalds 已提交
519 520 521 522
	if (atomic_dec_and_test(&files->count)) {
		close_files(files);
		/*
		 * Free the fd and fdset arrays if we expanded them.
523 524 525
		 * 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
526
		 */
527
		fdt = files_fdtable(files);
528
		if (fdt != &files->fdtab)
529
			kmem_cache_free(files_cachep, files);
530
		free_fdtable(fdt);
L
Linus Torvalds 已提交
531 532 533
	}
}

534
void reset_files_struct(struct files_struct *files)
535
{
536
	struct task_struct *tsk = current;
537 538 539 540 541 542 543 544 545
	struct files_struct *old;

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

546
void exit_files(struct task_struct *tsk)
L
Linus Torvalds 已提交
547 548 549 550 551 552 553 554 555 556 557
{
	struct files_struct * files = tsk->files;

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

558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610
#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);
611 612 613
	/*
	 * We found no owner yet mm_users > 1: this implies that we are
	 * most likely racing with swapoff (try_to_unuse()) or /proc or
614
	 * ptrace or page migration (get_task_mm()).  Mark owner as NULL.
615 616
	 */
	mm->owner = NULL;
617 618 619 620 621 622 623 624 625 626
	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);
627 628 629 630 631
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
632 633 634 635 636 637 638 639 640 641 642
	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
643 644 645 646
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
647
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
648 649
{
	struct mm_struct *mm = tsk->mm;
650
	struct core_state *core_state;
L
Linus Torvalds 已提交
651 652 653 654 655 656

	mm_release(tsk, mm);
	if (!mm)
		return;
	/*
	 * Serialize with any possible pending coredump.
657
	 * We must hold mmap_sem around checking core_state
L
Linus Torvalds 已提交
658
	 * and clearing tsk->mm.  The core-inducing thread
659
	 * will increment ->nr_threads for each thread in the
L
Linus Torvalds 已提交
660 661 662
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
663 664 665
	core_state = mm->core_state;
	if (core_state) {
		struct core_thread self;
L
Linus Torvalds 已提交
666
		up_read(&mm->mmap_sem);
667

668 669 670 671 672 673 674 675
		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
676

O
Oleg Nesterov 已提交
677 678 679 680 681 682 683
		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
684 685 686
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
687
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
688 689 690 691 692
	/* more a memory barrier than a real lock */
	task_lock(tsk);
	tsk->mm = NULL;
	up_read(&mm->mmap_sem);
	enter_lazy_tlb(mm, current);
693 694
	/* We don't want this task to be frozen prematurely */
	clear_freeze_flag(tsk);
L
Linus Torvalds 已提交
695
	task_unlock(tsk);
696
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
697 698 699 700 701 702 703
	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
704 705
 * the child reaper process (ie "init") in our pid
 * space.
L
Linus Torvalds 已提交
706
 */
707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738
static struct task_struct *find_new_reaper(struct task_struct *father)
{
	struct pid_namespace *pid_ns = task_active_pid_ns(father);
	struct task_struct *thread;

	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;
	}

	if (unlikely(pid_ns->child_reaper == father)) {
		write_unlock_irq(&tasklist_lock);
		if (unlikely(pid_ns == &init_pid_ns))
			panic("Attempted to kill init!");

		zap_pid_ns_processes(pid_ns);
		write_lock_irq(&tasklist_lock);
		/*
		 * 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.
		 */
		pid_ns->child_reaper = init_pid_ns.child_reaper;
	}

	return pid_ns->child_reaper;
}

739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774
/*
* 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);
}

775
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
776
{
777
	struct task_struct *p, *n, *reaper;
778
	LIST_HEAD(dead_children);
779

780
	exit_ptrace(father);
781 782

	write_lock_irq(&tasklist_lock);
783
	reaper = find_new_reaper(father);
R
Roland McGrath 已提交
784

785
	list_for_each_entry_safe(p, n, &father->children, sibling) {
786
		p->real_parent = reaper;
R
Roland McGrath 已提交
787 788 789 790
		if (p->parent == father) {
			BUG_ON(p->ptrace);
			p->parent = p->real_parent;
		}
791
		reparent_thread(father, p, &dead_children);
L
Linus Torvalds 已提交
792
	}
793
	write_unlock_irq(&tasklist_lock);
794

795 796
	BUG_ON(!list_empty(&father->children));

797 798
	list_for_each_entry_safe(p, n, &dead_children, sibling) {
		list_del_init(&p->sibling);
799 800
		release_task(p);
	}
L
Linus Torvalds 已提交
801 802 803 804 805 806
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
807
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
808
{
R
Roland McGrath 已提交
809 810
	int signal;
	void *cookie;
L
Linus Torvalds 已提交
811 812 813 814 815 816 817 818 819

	/*
	 * 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)
	 */
820
	forget_original_parent(tsk);
821
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
822

823
	write_lock_irq(&tasklist_lock);
824 825
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
826

827
	/* Let father know we died
L
Linus Torvalds 已提交
828 829
	 *
	 * Thread signals are configurable, but you aren't going to use
830
	 * that to send signals to arbitary processes.
L
Linus Torvalds 已提交
831 832 833 834 835 836 837 838 839 840
	 * 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.
	 */
841
	if (tsk->exit_signal != SIGCHLD && !task_detached(tsk) &&
842
	    (tsk->parent_exec_id != tsk->real_parent->self_exec_id ||
843
	     tsk->self_exec_id != tsk->parent_exec_id))
L
Linus Torvalds 已提交
844 845
		tsk->exit_signal = SIGCHLD;

R
Roland McGrath 已提交
846
	signal = tracehook_notify_death(tsk, &cookie, group_dead);
847
	if (signal >= 0)
R
Roland McGrath 已提交
848
		signal = do_notify_parent(tsk, signal);
L
Linus Torvalds 已提交
849

850
	tsk->exit_state = signal == DEATH_REAP ? EXIT_DEAD : EXIT_ZOMBIE;
L
Linus Torvalds 已提交
851

852
	/* mt-exec, de_thread() is waiting for us */
853
	if (thread_group_leader(tsk) &&
854 855
	    tsk->signal->group_exit_task &&
	    tsk->signal->notify_count < 0)
856 857
		wake_up_process(tsk->signal->group_exit_task);

L
Linus Torvalds 已提交
858 859
	write_unlock_irq(&tasklist_lock);

R
Roland McGrath 已提交
860 861
	tracehook_report_death(tsk, signal, cookie, group_dead);

L
Linus Torvalds 已提交
862
	/* If the process is dead, release it - nobody will wait for it */
863
	if (signal == DEATH_REAP)
L
Linus Torvalds 已提交
864 865 866
		release_task(tsk);
}

867 868 869 870 871 872 873
#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;

874
	free = stack_not_used(current);
875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891

	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

892
NORET_TYPE void do_exit(long code)
L
Linus Torvalds 已提交
893 894 895 896 897 898
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

899 900
	WARN_ON(atomic_read(&tsk->fs_excl));

L
Linus Torvalds 已提交
901 902 903 904 905
	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
906
	tracehook_report_exit(&code);
L
Linus Torvalds 已提交
907

908 909 910 911 912 913 914
	/*
	 * 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");
915 916 917 918 919 920 921 922 923 924
		/*
		 * 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;
925 926 927 928
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

929 930
	exit_irq_thread();

O
Oleg Nesterov 已提交
931
	exit_signals(tsk);  /* sets PF_EXITING */
932 933 934 935
	/*
	 * tsk->flags are checked in the futex code to protect against
	 * an exiting task cleaning up the robust pi futexes.
	 */
936 937
	smp_mb();
	spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
938 939 940

	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
941
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
942 943 944
				preempt_count());

	acct_update_integrals(tsk);
945

L
Linus Torvalds 已提交
946
	group_dead = atomic_dec_and_test(&tsk->signal->live);
947
	if (group_dead) {
948
		hrtimer_cancel(&tsk->signal->real_timer);
949
		exit_itimers(tsk->signal);
950
	}
951
	acct_collect(code, group_dead);
M
Miloslav Trmac 已提交
952 953
	if (group_dead)
		tty_audit_exit();
954 955
	if (unlikely(tsk->audit_context))
		audit_free(tsk);
956

957
	tsk->exit_code = code;
958
	taskstats_exit(tsk, group_dead);
959

L
Linus Torvalds 已提交
960 961
	exit_mm(tsk);

962
	if (group_dead)
963
		acct_process();
964 965
	trace_sched_process_exit(tsk);

L
Linus Torvalds 已提交
966
	exit_sem(tsk);
967 968
	exit_files(tsk);
	exit_fs(tsk);
969
	check_stack_usage();
L
Linus Torvalds 已提交
970
	exit_thread();
971
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
972 973 974 975

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

A
Al Viro 已提交
976
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
977 978 979
	if (tsk->binfmt)
		module_put(tsk->binfmt->module);

M
Matt Helsley 已提交
980
	proc_exit_connector(tsk);
981
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
982
#ifdef CONFIG_NUMA
983
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
984 985
	tsk->mempolicy = NULL;
#endif
986
#ifdef CONFIG_FUTEX
987 988 989 990
	if (unlikely(!list_empty(&tsk->pi_state_list)))
		exit_pi_state_list(tsk);
	if (unlikely(current->pi_state_cache))
		kfree(current->pi_state_cache);
991
#endif
992
	/*
993
	 * Make sure we are holding no locks:
994
	 */
995
	debug_check_no_locks_held(tsk);
996 997 998 999 1000 1001
	/*
	 * 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
1002

1003 1004 1005
	if (tsk->io_context)
		exit_io_context();

1006 1007 1008
	if (tsk->splice_pipe)
		__free_pipe_info(tsk->splice_pipe);

C
Coywolf Qi Hunt 已提交
1009
	preempt_disable();
O
Oleg Nesterov 已提交
1010
	/* causes final put_task_struct in finish_task_switch(). */
1011
	tsk->state = TASK_DEAD;
L
Linus Torvalds 已提交
1012 1013 1014
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
1015 1016
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
1017 1018
}

1019 1020
EXPORT_SYMBOL_GPL(do_exit);

L
Linus Torvalds 已提交
1021 1022 1023 1024
NORET_TYPE void complete_and_exit(struct completion *comp, long code)
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
1025

L
Linus Torvalds 已提交
1026 1027 1028 1029 1030
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

1031
SYSCALL_DEFINE1(exit, int, error_code)
L
Linus Torvalds 已提交
1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042
{
	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)
{
1043 1044
	struct signal_struct *sig = current->signal;

L
Linus Torvalds 已提交
1045 1046
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

1047 1048
	if (signal_group_exit(sig))
		exit_code = sig->group_exit_code;
L
Linus Torvalds 已提交
1049 1050 1051
	else if (!thread_group_empty(current)) {
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1052
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1053 1054 1055 1056
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1057
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
			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.
 */
1072
SYSCALL_DEFINE1(exit_group, int, error_code)
L
Linus Torvalds 已提交
1073 1074
{
	do_group_exit((error_code & 0xff) << 8);
1075 1076
	/* NOTREACHED */
	return 0;
L
Linus Torvalds 已提交
1077 1078
}

E
Eric W. Biederman 已提交
1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090
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
1091
{
1092 1093
	int err;

E
Eric W. Biederman 已提交
1094 1095
	if (type < PIDTYPE_MAX) {
		if (task_pid_type(p, type) != pid)
L
Linus Torvalds 已提交
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107
			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;

1108
	err = security_task_wait(p);
1109 1110
	if (err)
		return err;
L
Linus Torvalds 已提交
1111

1112
	return 1;
L
Linus Torvalds 已提交
1113 1114
}

1115
static int wait_noreap_copyout(struct task_struct *p, pid_t pid, uid_t uid,
L
Linus Torvalds 已提交
1116 1117 1118 1119 1120
			       int why, int status,
			       struct siginfo __user *infop,
			       struct rusage __user *rusagep)
{
	int retval = rusagep ? getrusage(p, RUSAGE_BOTH, rusagep) : 0;
1121

L
Linus Torvalds 已提交
1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
	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
1146
static int wait_task_zombie(struct task_struct *p, int options,
L
Linus Torvalds 已提交
1147 1148 1149 1150
			    struct siginfo __user *infop,
			    int __user *stat_addr, struct rusage __user *ru)
{
	unsigned long state;
1151
	int retval, status, traced;
1152
	pid_t pid = task_pid_vnr(p);
1153
	uid_t uid = __task_cred(p)->uid;
L
Linus Torvalds 已提交
1154

R
Roland McGrath 已提交
1155 1156 1157 1158
	if (!likely(options & WEXITED))
		return 0;

	if (unlikely(options & WNOWAIT)) {
L
Linus Torvalds 已提交
1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
		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;
	}

1185
	traced = ptrace_reparented(p);
1186 1187

	if (likely(!traced)) {
1188 1189
		struct signal_struct *psig;
		struct signal_struct *sig;
1190
		struct task_cputime cputime;
1191

L
Linus Torvalds 已提交
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205
		/*
		 * 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.
1206 1207 1208 1209
		 *
		 * 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
1210
		 */
1211
		thread_group_cputime(p, &cputime);
L
Linus Torvalds 已提交
1212
		spin_lock_irq(&p->parent->sighand->siglock);
1213 1214 1215 1216
		psig = p->parent->signal;
		sig = p->signal;
		psig->cutime =
			cputime_add(psig->cutime,
1217 1218
			cputime_add(cputime.utime,
				    sig->cutime));
1219 1220
		psig->cstime =
			cputime_add(psig->cstime,
1221 1222
			cputime_add(cputime.stime,
				    sig->cstime));
1223 1224 1225 1226 1227
		psig->cgtime =
			cputime_add(psig->cgtime,
			cputime_add(p->gtime,
			cputime_add(sig->gtime,
				    sig->cgtime)));
1228 1229 1230 1231 1232 1233 1234 1235
		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;
1236 1237 1238 1239 1240 1241
		psig->cinblock +=
			task_io_get_inblock(p) +
			sig->inblock + sig->cinblock;
		psig->coublock +=
			task_io_get_oublock(p) +
			sig->oublock + sig->coublock;
1242 1243
		task_io_accounting_add(&psig->ioac, &p->ioac);
		task_io_accounting_add(&psig->ioac, &sig->ioac);
L
Linus Torvalds 已提交
1244 1245 1246 1247 1248 1249 1250 1251 1252
		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);

1253 1254 1255 1256 1257 1258
	/*
	 * Flush inherited counters to the parent - before the parent
	 * gets woken up by child-exit notifications.
	 */
	perf_counter_exit_task(p);

L
Linus Torvalds 已提交
1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
	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)
1283
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1284
	if (!retval && infop)
1285
		retval = put_user(uid, &infop->si_uid);
1286
	if (!retval)
1287
		retval = pid;
1288 1289

	if (traced) {
L
Linus Torvalds 已提交
1290
		write_lock_irq(&tasklist_lock);
1291 1292 1293 1294 1295 1296 1297
		/* 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.
		 */
1298
		if (!task_detached(p)) {
1299
			do_notify_parent(p, p->exit_signal);
1300
			if (!task_detached(p)) {
1301 1302
				p->exit_state = EXIT_ZOMBIE;
				p = NULL;
L
Linus Torvalds 已提交
1303 1304 1305 1306 1307 1308
			}
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1309

L
Linus Torvalds 已提交
1310 1311 1312
	return retval;
}

1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324
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
1325 1326 1327 1328 1329 1330
/*
 * 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
1331
static int wait_task_stopped(int ptrace, struct task_struct *p,
R
Roland McGrath 已提交
1332
			     int options, struct siginfo __user *infop,
L
Linus Torvalds 已提交
1333 1334
			     int __user *stat_addr, struct rusage __user *ru)
{
1335
	int retval, exit_code, *p_code, why;
1336
	uid_t uid = 0; /* unneeded, required by compiler */
1337
	pid_t pid;
L
Linus Torvalds 已提交
1338

R
Roland McGrath 已提交
1339
	if (!(options & WUNTRACED))
R
Roland McGrath 已提交
1340 1341
		return 0;

1342 1343 1344
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

1345 1346
	p_code = task_stopped_code(p, ptrace);
	if (unlikely(!p_code))
1347 1348
		goto unlock_sig;

1349
	exit_code = *p_code;
1350 1351 1352
	if (!exit_code)
		goto unlock_sig;

R
Roland McGrath 已提交
1353
	if (!unlikely(options & WNOWAIT))
1354
		*p_code = 0;
1355

1356 1357
	/* don't need the RCU readlock here as we're holding a spinlock */
	uid = __task_cred(p)->uid;
1358 1359 1360
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1361 1362 1363 1364 1365 1366 1367 1368 1369 1370
		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);
1371
	pid = task_pid_vnr(p);
R
Roland McGrath 已提交
1372
	why = ptrace ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1373 1374
	read_unlock(&tasklist_lock);

R
Roland McGrath 已提交
1375
	if (unlikely(options & WNOWAIT))
L
Linus Torvalds 已提交
1376
		return wait_noreap_copyout(p, pid, uid,
1377
					   why, exit_code,
L
Linus Torvalds 已提交
1378 1379 1380 1381 1382 1383 1384 1385 1386 1387
					   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
1388
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1389 1390 1391
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1392
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1393
	if (!retval && infop)
1394
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1395
	if (!retval)
1396
		retval = pid;
L
Linus Torvalds 已提交
1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408
	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
1409
static int wait_task_continued(struct task_struct *p, int options,
L
Linus Torvalds 已提交
1410 1411 1412 1413 1414 1415 1416
			       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
1417 1418 1419
	if (!unlikely(options & WCONTINUED))
		return 0;

L
Linus Torvalds 已提交
1420 1421 1422 1423 1424 1425 1426 1427 1428
	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
1429
	if (!unlikely(options & WNOWAIT))
L
Linus Torvalds 已提交
1430
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
1431
	uid = __task_cred(p)->uid;
L
Linus Torvalds 已提交
1432 1433
	spin_unlock_irq(&p->sighand->siglock);

1434
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1435 1436 1437 1438 1439 1440 1441 1442 1443
	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)
1444
			retval = pid;
L
Linus Torvalds 已提交
1445 1446 1447 1448 1449 1450 1451 1452 1453 1454
	} 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
1455 1456 1457 1458 1459 1460
/*
 * 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;
1461 1462
 * 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
1463
 */
R
Roland McGrath 已提交
1464
static int wait_consider_task(struct task_struct *parent, int ptrace,
R
Roland McGrath 已提交
1465 1466 1467 1468 1469 1470
			      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);
1471
	if (!ret)
R
Roland McGrath 已提交
1472 1473
		return ret;

1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
	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
1486 1487 1488 1489 1490 1491 1492 1493 1494
	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
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509
	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;

1510
	if (task_stopped_code(p, ptrace))
R
Roland McGrath 已提交
1511 1512
		return wait_task_stopped(ptrace, p, options,
					 infop, stat_addr, ru);
R
Roland McGrath 已提交
1513 1514 1515 1516 1517 1518 1519 1520 1521 1522

	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
1523 1524
 * *@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
1525 1526 1527 1528 1529 1530 1531 1532 1533
 */
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
1534 1535 1536 1537 1538 1539 1540 1541 1542 1543
		/*
		 * 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
1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556
	}

	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
1557
	 * Traditionally we see ptrace'd stopped tasks regardless of options.
R
Roland McGrath 已提交
1558
	 */
R
Roland McGrath 已提交
1559
	options |= WUNTRACED;
R
Roland McGrath 已提交
1560

R
Roland McGrath 已提交
1561 1562 1563 1564 1565
	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
1566 1567 1568 1569 1570 1571
			return ret;
	}

	return 0;
}

E
Eric W. Biederman 已提交
1572 1573 1574
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
1575 1576 1577
{
	DECLARE_WAITQUEUE(wait, current);
	struct task_struct *tsk;
R
Roland McGrath 已提交
1578
	int retval;
L
Linus Torvalds 已提交
1579

1580 1581
	trace_sched_process_wait(pid);

L
Linus Torvalds 已提交
1582 1583
	add_wait_queue(&current->signal->wait_chldexit,&wait);
repeat:
R
Roland McGrath 已提交
1584 1585 1586 1587 1588
	/*
	 * 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
1589 1590 1591 1592
	retval = -ECHILD;
	if ((type < PIDTYPE_MAX) && (!pid || hlist_empty(&pid->tasks[type])))
		goto end;

L
Linus Torvalds 已提交
1593 1594 1595 1596
	current->state = TASK_INTERRUPTIBLE;
	read_lock(&tasklist_lock);
	tsk = current;
	do {
R
Roland McGrath 已提交
1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609
		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
1610
		}
R
Roland McGrath 已提交
1611

L
Linus Torvalds 已提交
1612 1613 1614
		if (options & __WNOTHREAD)
			break;
		tsk = next_thread(tsk);
E
Eric Sesterhenn 已提交
1615
		BUG_ON(tsk->signal != current->signal);
L
Linus Torvalds 已提交
1616 1617
	} while (tsk != current);
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1618

R
Roland McGrath 已提交
1619
	if (!retval && !(options & WNOHANG)) {
L
Linus Torvalds 已提交
1620
		retval = -ERESTARTSYS;
R
Roland McGrath 已提交
1621 1622 1623 1624
		if (!signal_pending(current)) {
			schedule();
			goto repeat;
		}
L
Linus Torvalds 已提交
1625
	}
R
Roland McGrath 已提交
1626

L
Linus Torvalds 已提交
1627 1628 1629 1630 1631
end:
	current->state = TASK_RUNNING;
	remove_wait_queue(&current->signal->wait_chldexit,&wait);
	if (infop) {
		if (retval > 0)
1632
			retval = 0;
L
Linus Torvalds 已提交
1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655
		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;
}

1656 1657
SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *,
		infop, int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1658
{
E
Eric W. Biederman 已提交
1659 1660
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1661 1662 1663 1664 1665 1666 1667 1668 1669
	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
1670
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1671 1672
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1673 1674
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1675 1676 1677
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1678 1679
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1680 1681 1682 1683 1684 1685
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1686 1687 1688 1689
	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
1690 1691

	/* avoid REGPARM breakage on x86: */
1692
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1693 1694 1695
	return ret;
}

1696 1697
SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr,
		int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1698
{
E
Eric W. Biederman 已提交
1699 1700
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1701 1702 1703 1704 1705
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1706 1707 1708 1709 1710 1711 1712 1713

	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;
1714
		pid = get_task_pid(current, PIDTYPE_PGID);
E
Eric W. Biederman 已提交
1715 1716 1717 1718 1719 1720 1721
	} 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
1722 1723

	/* avoid REGPARM breakage on x86: */
1724
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1725 1726 1727 1728 1729 1730 1731 1732 1733
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
1734
SYSCALL_DEFINE3(waitpid, pid_t, pid, int __user *, stat_addr, int, options)
L
Linus Torvalds 已提交
1735 1736 1737 1738 1739
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif