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);
180

L
Linus Torvalds 已提交
181
	write_lock_irq(&tasklist_lock);
R
Roland McGrath 已提交
182
	tracehook_finish_release_task(p);
L
Linus Torvalds 已提交
183
	__exit_signal(p);
184

L
Linus Torvalds 已提交
185 186 187 188 189 190 191 192
	/*
	 * If we are the last non-leader member of the thread
	 * group, and the leader is zombie, then notify the
	 * group leader's parent process. (if it wants notification.)
	 */
	zap_leader = 0;
	leader = p->group_leader;
	if (leader != p && thread_group_empty(leader) && leader->exit_state == EXIT_ZOMBIE) {
193
		BUG_ON(task_detached(leader));
L
Linus Torvalds 已提交
194 195 196 197 198 199 200 201 202
		do_notify_parent(leader, leader->exit_signal);
		/*
		 * If we were the last child thread and the leader has
		 * exited already, and the leader's parent ignores SIGCHLD,
		 * then we are the one who should release the leader.
		 *
		 * do_notify_parent() will have marked it self-reaping in
		 * that case.
		 */
203
		zap_leader = task_detached(leader);
R
Roland McGrath 已提交
204 205 206 207 208 209 210

		/*
		 * This maintains the invariant that release_task()
		 * only runs on a task in EXIT_DEAD, just for sanity.
		 */
		if (zap_leader)
			leader->exit_state = EXIT_DEAD;
L
Linus Torvalds 已提交
211 212 213 214
	}

	write_unlock_irq(&tasklist_lock);
	release_thread(p);
215
	call_rcu(&p->rcu, delayed_put_task_struct);
L
Linus Torvalds 已提交
216 217 218 219 220 221 222 223 224 225

	p = leader;
	if (unlikely(zap_leader))
		goto repeat;
}

/*
 * This checks not only the pgrp, but falls back on the pid if no
 * satisfactory pgrp is found. I dunno - gdb doesn't work correctly
 * without this...
226 227
 *
 * The caller must hold rcu lock or the tasklist lock.
L
Linus Torvalds 已提交
228
 */
229
struct pid *session_of_pgrp(struct pid *pgrp)
L
Linus Torvalds 已提交
230 231
{
	struct task_struct *p;
232
	struct pid *sid = NULL;
233

234
	p = pid_task(pgrp, PIDTYPE_PGID);
235
	if (p == NULL)
236
		p = pid_task(pgrp, PIDTYPE_PID);
237
	if (p != NULL)
238
		sid = task_session(p);
239

L
Linus Torvalds 已提交
240 241 242 243 244 245 246 247 248 249 250
	return sid;
}

/*
 * Determine if a process group is "orphaned", according to the POSIX
 * definition in 2.2.2.52.  Orphaned process groups are not to be affected
 * by terminal-generated stop signals.  Newly orphaned process groups are
 * to receive a SIGHUP and a SIGCONT.
 *
 * "I ask you, have you ever known what it is to be an orphan?"
 */
251
static int will_become_orphaned_pgrp(struct pid *pgrp, struct task_struct *ignored_task)
L
Linus Torvalds 已提交
252 253 254
{
	struct task_struct *p;

255
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
256 257 258
		if ((p == ignored_task) ||
		    (p->exit_state && thread_group_empty(p)) ||
		    is_global_init(p->real_parent))
L
Linus Torvalds 已提交
259
			continue;
260

261
		if (task_pgrp(p->real_parent) != pgrp &&
262 263
		    task_session(p->real_parent) == task_session(p))
			return 0;
264
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
265 266

	return 1;
L
Linus Torvalds 已提交
267 268
}

269
int is_current_pgrp_orphaned(void)
L
Linus Torvalds 已提交
270 271 272 273
{
	int retval;

	read_lock(&tasklist_lock);
274
	retval = will_become_orphaned_pgrp(task_pgrp(current), NULL);
L
Linus Torvalds 已提交
275 276 277 278 279
	read_unlock(&tasklist_lock);

	return retval;
}

280
static int has_stopped_jobs(struct pid *pgrp)
L
Linus Torvalds 已提交
281 282 283 284
{
	int retval = 0;
	struct task_struct *p;

285
	do_each_pid_task(pgrp, PIDTYPE_PGID, p) {
M
Matthew Wilcox 已提交
286
		if (!task_is_stopped(p))
L
Linus Torvalds 已提交
287 288 289
			continue;
		retval = 1;
		break;
290
	} while_each_pid_task(pgrp, PIDTYPE_PGID, p);
L
Linus Torvalds 已提交
291 292 293
	return retval;
}

294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
/*
 * Check to see if any process groups have become orphaned as
 * a result of our exiting, and if they have any stopped jobs,
 * send them a SIGHUP and then a SIGCONT. (POSIX 3.2.2.2)
 */
static void
kill_orphaned_pgrp(struct task_struct *tsk, struct task_struct *parent)
{
	struct pid *pgrp = task_pgrp(tsk);
	struct task_struct *ignored_task = tsk;

	if (!parent)
		 /* exit: our father is in a different pgrp than
		  * we are and we were the only connection outside.
		  */
		parent = tsk->real_parent;
	else
		/* reparent: our child is in a different pgrp than
		 * we are, and it was the only connection outside.
		 */
		ignored_task = NULL;

	if (task_pgrp(parent) != pgrp &&
	    task_session(parent) == task_session(tsk) &&
	    will_become_orphaned_pgrp(pgrp, ignored_task) &&
	    has_stopped_jobs(pgrp)) {
		__kill_pgrp_info(SIGHUP, SEND_SIG_PRIV, pgrp);
		__kill_pgrp_info(SIGCONT, SEND_SIG_PRIV, pgrp);
	}
}

L
Linus Torvalds 已提交
325
/**
326
 * reparent_to_kthreadd - Reparent the calling kernel thread to kthreadd
L
Linus Torvalds 已提交
327 328
 *
 * If a kernel thread is launched as a result of a system call, or if
329 330
 * it ever exits, it should generally reparent itself to kthreadd so it
 * isn't in the way of other processes and is correctly cleaned up on exit.
L
Linus Torvalds 已提交
331 332 333 334
 *
 * The various task state such as scheduling policy and priority may have
 * been inherited from a user process, so we reset them to sane values here.
 *
335
 * NOTE that reparent_to_kthreadd() gives the caller full capabilities.
L
Linus Torvalds 已提交
336
 */
337
static void reparent_to_kthreadd(void)
L
Linus Torvalds 已提交
338 339 340 341 342
{
	write_lock_irq(&tasklist_lock);

	ptrace_unlink(current);
	/* Reparent to init */
343
	current->real_parent = current->parent = kthreadd_task;
R
Roland McGrath 已提交
344
	list_move_tail(&current->sibling, &current->real_parent->children);
L
Linus Torvalds 已提交
345 346 347 348

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

349
	if (task_nice(current) < 0)
L
Linus Torvalds 已提交
350 351 352 353 354 355
		set_user_nice(current, 0);
	/* cpus_allowed? */
	/* rt_priority? */
	/* signals? */
	memcpy(current->signal->rlim, init_task.signal->rlim,
	       sizeof(current->signal->rlim));
D
David Howells 已提交
356 357 358

	atomic_inc(&init_cred.usage);
	commit_creds(&init_cred);
L
Linus Torvalds 已提交
359 360 361
	write_unlock_irq(&tasklist_lock);
}

362
void __set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
363
{
364
	struct task_struct *curr = current->group_leader;
L
Linus Torvalds 已提交
365

366
	if (task_session(curr) != pid)
367
		change_pid(curr, PIDTYPE_SID, pid);
368 369

	if (task_pgrp(curr) != pid)
370
		change_pid(curr, PIDTYPE_PGID, pid);
L
Linus Torvalds 已提交
371 372
}

373
static void set_special_pids(struct pid *pid)
L
Linus Torvalds 已提交
374 375
{
	write_lock_irq(&tasklist_lock);
376
	__set_special_pids(pid);
L
Linus Torvalds 已提交
377 378 379 380 381 382 383 384 385 386
	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)
{
387
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407
		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)
{
408
	if (!valid_signal(sig) || sig < 1)
L
Linus Torvalds 已提交
409 410 411
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
412
	current->sighand->action[(sig)-1].sa.sa_handler = SIG_IGN;
L
Linus Torvalds 已提交
413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439
	recalc_sigpending();
	spin_unlock_irq(&current->sighand->siglock);
	return 0;
}

EXPORT_SYMBOL(disallow_signal);

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

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

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

	/*
	 * If we were started as result of loading a module, close all of the
	 * user space pages.  We don't need them, and if we didn't close them
	 * they would be locked into memory.
	 */
	exit_mm(current);
440 441 442 443
	/*
	 * We don't want to have TIF_FREEZE set if the system-wide hibernation
	 * or suspend transition begins right now.
	 */
O
Oleg Nesterov 已提交
444
	current->flags |= (PF_NOFREEZE | PF_KTHREAD);
L
Linus Torvalds 已提交
445

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

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

	/* Become as one with the init task */

460
	daemonize_fs_struct();
461
	exit_files(current);
L
Linus Torvalds 已提交
462 463 464
	current->files = init_task.files;
	atomic_inc(&current->files->count);

465
	reparent_to_kthreadd();
L
Linus Torvalds 已提交
466 467 468 469
}

EXPORT_SYMBOL(daemonize);

470
static void close_files(struct files_struct * files)
L
Linus Torvalds 已提交
471 472
{
	int i, j;
473
	struct fdtable *fdt;
L
Linus Torvalds 已提交
474 475

	j = 0;
476 477 478 479 480 481

	/*
	 * It is safe to dereference the fd table without RCU or
	 * ->file_lock because this is the last reference to the
	 * files structure.
	 */
482
	fdt = files_fdtable(files);
L
Linus Torvalds 已提交
483 484 485
	for (;;) {
		unsigned long set;
		i = j * __NFDBITS;
486
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
487
			break;
488
		set = fdt->open_fds->fds_bits[j++];
L
Linus Torvalds 已提交
489 490
		while (set) {
			if (set & 1) {
491
				struct file * file = xchg(&fdt->fd[i], NULL);
492
				if (file) {
L
Linus Torvalds 已提交
493
					filp_close(file, files);
494 495
					cond_resched();
				}
L
Linus Torvalds 已提交
496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515
			}
			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;
}

516
void put_files_struct(struct files_struct *files)
L
Linus Torvalds 已提交
517
{
518 519
	struct fdtable *fdt;

L
Linus Torvalds 已提交
520 521 522 523
	if (atomic_dec_and_test(&files->count)) {
		close_files(files);
		/*
		 * Free the fd and fdset arrays if we expanded them.
524 525 526
		 * 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
527
		 */
528
		fdt = files_fdtable(files);
529
		if (fdt != &files->fdtab)
530
			kmem_cache_free(files_cachep, files);
531
		free_fdtable(fdt);
L
Linus Torvalds 已提交
532 533 534
	}
}

535
void reset_files_struct(struct files_struct *files)
536
{
537
	struct task_struct *tsk = current;
538 539 540 541 542 543 544 545 546
	struct files_struct *old;

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

547
void exit_files(struct task_struct *tsk)
L
Linus Torvalds 已提交
548 549 550 551 552 553 554 555 556 557 558
{
	struct files_struct * files = tsk->files;

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

559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611
#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);
612 613 614
	/*
	 * We found no owner yet mm_users > 1: this implies that we are
	 * most likely racing with swapoff (try_to_unuse()) or /proc or
615
	 * ptrace or page migration (get_task_mm()).  Mark owner as NULL.
616 617
	 */
	mm->owner = NULL;
618 619 620 621 622 623 624 625 626 627
	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);
628 629 630 631 632
	/*
	 * Delay read_unlock() till we have the task_lock()
	 * to ensure that c does not slip away underneath us
	 */
	read_unlock(&tasklist_lock);
633 634 635 636 637 638 639 640 641 642 643
	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
644 645 646 647
/*
 * Turn us into a lazy TLB process if we
 * aren't already..
 */
A
Adrian Bunk 已提交
648
static void exit_mm(struct task_struct * tsk)
L
Linus Torvalds 已提交
649 650
{
	struct mm_struct *mm = tsk->mm;
651
	struct core_state *core_state;
L
Linus Torvalds 已提交
652 653 654 655 656 657

	mm_release(tsk, mm);
	if (!mm)
		return;
	/*
	 * Serialize with any possible pending coredump.
658
	 * We must hold mmap_sem around checking core_state
L
Linus Torvalds 已提交
659
	 * and clearing tsk->mm.  The core-inducing thread
660
	 * will increment ->nr_threads for each thread in the
L
Linus Torvalds 已提交
661 662 663
	 * group with ->mm != NULL.
	 */
	down_read(&mm->mmap_sem);
664 665 666
	core_state = mm->core_state;
	if (core_state) {
		struct core_thread self;
L
Linus Torvalds 已提交
667
		up_read(&mm->mmap_sem);
668

669 670 671 672 673 674 675 676
		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
677

O
Oleg Nesterov 已提交
678 679 680 681 682 683 684
		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
685 686 687
		down_read(&mm->mmap_sem);
	}
	atomic_inc(&mm->mm_count);
E
Eric Sesterhenn 已提交
688
	BUG_ON(mm != tsk->active_mm);
L
Linus Torvalds 已提交
689 690 691 692 693
	/* more a memory barrier than a real lock */
	task_lock(tsk);
	tsk->mm = NULL;
	up_read(&mm->mmap_sem);
	enter_lazy_tlb(mm, current);
694 695
	/* We don't want this task to be frozen prematurely */
	clear_freeze_flag(tsk);
L
Linus Torvalds 已提交
696
	task_unlock(tsk);
697
	mm_update_next_owner(mm);
L
Linus Torvalds 已提交
698 699 700 701 702 703 704
	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
705 706
 * the child reaper process (ie "init") in our pid
 * space.
L
Linus Torvalds 已提交
707
 */
708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739
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;
}

740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775
/*
* 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);
}

776
static void forget_original_parent(struct task_struct *father)
L
Linus Torvalds 已提交
777
{
778
	struct task_struct *p, *n, *reaper;
779
	LIST_HEAD(dead_children);
780

781
	exit_ptrace(father);
782 783

	write_lock_irq(&tasklist_lock);
784
	reaper = find_new_reaper(father);
R
Roland McGrath 已提交
785

786
	list_for_each_entry_safe(p, n, &father->children, sibling) {
787
		p->real_parent = reaper;
R
Roland McGrath 已提交
788 789 790 791
		if (p->parent == father) {
			BUG_ON(p->ptrace);
			p->parent = p->real_parent;
		}
792
		reparent_thread(father, p, &dead_children);
L
Linus Torvalds 已提交
793
	}
794
	write_unlock_irq(&tasklist_lock);
795

796 797
	BUG_ON(!list_empty(&father->children));

798 799
	list_for_each_entry_safe(p, n, &dead_children, sibling) {
		list_del_init(&p->sibling);
800 801
		release_task(p);
	}
L
Linus Torvalds 已提交
802 803 804 805 806 807
}

/*
 * Send signals to all our closest relatives so that they know
 * to properly mourn us..
 */
808
static void exit_notify(struct task_struct *tsk, int group_dead)
L
Linus Torvalds 已提交
809
{
R
Roland McGrath 已提交
810 811
	int signal;
	void *cookie;
L
Linus Torvalds 已提交
812 813 814 815 816 817 818 819 820

	/*
	 * 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)
	 */
821
	forget_original_parent(tsk);
822
	exit_task_namespaces(tsk);
L
Linus Torvalds 已提交
823

824
	write_lock_irq(&tasklist_lock);
825 826
	if (group_dead)
		kill_orphaned_pgrp(tsk->group_leader, NULL);
L
Linus Torvalds 已提交
827

828
	/* Let father know we died
L
Linus Torvalds 已提交
829 830
	 *
	 * Thread signals are configurable, but you aren't going to use
831
	 * that to send signals to arbitary processes.
L
Linus Torvalds 已提交
832 833 834 835 836 837 838 839 840 841
	 * 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.
	 */
842
	if (tsk->exit_signal != SIGCHLD && !task_detached(tsk) &&
843
	    (tsk->parent_exec_id != tsk->real_parent->self_exec_id ||
844
	     tsk->self_exec_id != tsk->parent_exec_id))
L
Linus Torvalds 已提交
845 846
		tsk->exit_signal = SIGCHLD;

R
Roland McGrath 已提交
847
	signal = tracehook_notify_death(tsk, &cookie, group_dead);
848
	if (signal >= 0)
R
Roland McGrath 已提交
849
		signal = do_notify_parent(tsk, signal);
L
Linus Torvalds 已提交
850

851
	tsk->exit_state = signal == DEATH_REAP ? EXIT_DEAD : EXIT_ZOMBIE;
L
Linus Torvalds 已提交
852

853
	/* mt-exec, de_thread() is waiting for us */
854
	if (thread_group_leader(tsk) &&
855 856
	    tsk->signal->group_exit_task &&
	    tsk->signal->notify_count < 0)
857 858
		wake_up_process(tsk->signal->group_exit_task);

L
Linus Torvalds 已提交
859 860
	write_unlock_irq(&tasklist_lock);

R
Roland McGrath 已提交
861 862
	tracehook_report_death(tsk, signal, cookie, group_dead);

L
Linus Torvalds 已提交
863
	/* If the process is dead, release it - nobody will wait for it */
864
	if (signal == DEATH_REAP)
L
Linus Torvalds 已提交
865 866 867
		release_task(tsk);
}

868 869 870 871 872 873 874
#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;

875
	free = stack_not_used(current);
876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892

	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

893
NORET_TYPE void do_exit(long code)
L
Linus Torvalds 已提交
894 895 896 897 898 899
{
	struct task_struct *tsk = current;
	int group_dead;

	profile_task_exit(tsk);

900 901
	WARN_ON(atomic_read(&tsk->fs_excl));

L
Linus Torvalds 已提交
902 903 904 905 906
	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
907
	tracehook_report_exit(&code);
L
Linus Torvalds 已提交
908

909 910 911 912 913 914 915
	/*
	 * 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");
916 917 918 919 920 921 922 923 924 925
		/*
		 * 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;
926 927 928 929
		set_current_state(TASK_UNINTERRUPTIBLE);
		schedule();
	}

930 931
	exit_irq_thread();

O
Oleg Nesterov 已提交
932
	exit_signals(tsk);  /* sets PF_EXITING */
933 934 935 936
	/*
	 * tsk->flags are checked in the futex code to protect against
	 * an exiting task cleaning up the robust pi futexes.
	 */
937 938
	smp_mb();
	spin_unlock_wait(&tsk->pi_lock);
L
Linus Torvalds 已提交
939 940 941

	if (unlikely(in_atomic()))
		printk(KERN_INFO "note: %s[%d] exited with preempt_count %d\n",
942
				current->comm, task_pid_nr(current),
L
Linus Torvalds 已提交
943 944 945
				preempt_count());

	acct_update_integrals(tsk);
946

L
Linus Torvalds 已提交
947
	group_dead = atomic_dec_and_test(&tsk->signal->live);
948
	if (group_dead) {
949
		hrtimer_cancel(&tsk->signal->real_timer);
950
		exit_itimers(tsk->signal);
951
	}
952
	acct_collect(code, group_dead);
M
Miloslav Trmac 已提交
953 954
	if (group_dead)
		tty_audit_exit();
955 956
	if (unlikely(tsk->audit_context))
		audit_free(tsk);
957

958
	tsk->exit_code = code;
959
	taskstats_exit(tsk, group_dead);
960

L
Linus Torvalds 已提交
961 962
	exit_mm(tsk);

963
	if (group_dead)
964
		acct_process();
965 966
	trace_sched_process_exit(tsk);

L
Linus Torvalds 已提交
967
	exit_sem(tsk);
968 969
	exit_files(tsk);
	exit_fs(tsk);
970
	check_stack_usage();
L
Linus Torvalds 已提交
971
	exit_thread();
972
	cgroup_exit(tsk, 1);
L
Linus Torvalds 已提交
973 974 975 976

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

A
Al Viro 已提交
977
	module_put(task_thread_info(tsk)->exec_domain->module);
L
Linus Torvalds 已提交
978 979 980
	if (tsk->binfmt)
		module_put(tsk->binfmt->module);

M
Matt Helsley 已提交
981
	proc_exit_connector(tsk);
982 983 984 985 986 987 988

	/*
	 * Flush inherited counters to the parent - before the parent
	 * gets woken up by child-exit notifications.
	 */
	perf_counter_exit_task(tsk);

989
	exit_notify(tsk, group_dead);
L
Linus Torvalds 已提交
990
#ifdef CONFIG_NUMA
991
	mpol_put(tsk->mempolicy);
L
Linus Torvalds 已提交
992 993
	tsk->mempolicy = NULL;
#endif
994
#ifdef CONFIG_FUTEX
995 996 997 998
	if (unlikely(!list_empty(&tsk->pi_state_list)))
		exit_pi_state_list(tsk);
	if (unlikely(current->pi_state_cache))
		kfree(current->pi_state_cache);
999
#endif
1000
	/*
1001
	 * Make sure we are holding no locks:
1002
	 */
1003
	debug_check_no_locks_held(tsk);
1004 1005 1006 1007 1008 1009
	/*
	 * We can do this unlocked here. The futex code uses this flag
	 * just to verify whether the pi state cleanup has been done
	 * or not. In the worst case it loops once more.
	 */
	tsk->flags |= PF_EXITPIDONE;
L
Linus Torvalds 已提交
1010

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

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

C
Coywolf Qi Hunt 已提交
1017
	preempt_disable();
O
Oleg Nesterov 已提交
1018
	/* causes final put_task_struct in finish_task_switch(). */
1019
	tsk->state = TASK_DEAD;
L
Linus Torvalds 已提交
1020 1021 1022
	schedule();
	BUG();
	/* Avoid "noreturn function does return".  */
A
Alan Cox 已提交
1023 1024
	for (;;)
		cpu_relax();	/* For when BUG is null */
L
Linus Torvalds 已提交
1025 1026
}

1027 1028
EXPORT_SYMBOL_GPL(do_exit);

L
Linus Torvalds 已提交
1029 1030 1031 1032
NORET_TYPE void complete_and_exit(struct completion *comp, long code)
{
	if (comp)
		complete(comp);
O
Oleg Nesterov 已提交
1033

L
Linus Torvalds 已提交
1034 1035 1036 1037 1038
	do_exit(code);
}

EXPORT_SYMBOL(complete_and_exit);

1039
SYSCALL_DEFINE1(exit, int, error_code)
L
Linus Torvalds 已提交
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050
{
	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)
{
1051 1052
	struct signal_struct *sig = current->signal;

L
Linus Torvalds 已提交
1053 1054
	BUG_ON(exit_code & 0x80); /* core dumps don't get here */

1055 1056
	if (signal_group_exit(sig))
		exit_code = sig->group_exit_code;
L
Linus Torvalds 已提交
1057 1058 1059
	else if (!thread_group_empty(current)) {
		struct sighand_struct *const sighand = current->sighand;
		spin_lock_irq(&sighand->siglock);
1060
		if (signal_group_exit(sig))
L
Linus Torvalds 已提交
1061 1062 1063 1064
			/* Another thread got here before we took the lock.  */
			exit_code = sig->group_exit_code;
		else {
			sig->group_exit_code = exit_code;
1065
			sig->flags = SIGNAL_GROUP_EXIT;
L
Linus Torvalds 已提交
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079
			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.
 */
1080
SYSCALL_DEFINE1(exit_group, int, error_code)
L
Linus Torvalds 已提交
1081 1082
{
	do_group_exit((error_code & 0xff) << 8);
1083 1084
	/* NOTREACHED */
	return 0;
L
Linus Torvalds 已提交
1085 1086
}

E
Eric W. Biederman 已提交
1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098
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
1099
{
1100 1101
	int err;

E
Eric W. Biederman 已提交
1102 1103
	if (type < PIDTYPE_MAX) {
		if (task_pid_type(p, type) != pid)
L
Linus Torvalds 已提交
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
			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;

1116
	err = security_task_wait(p);
1117 1118
	if (err)
		return err;
L
Linus Torvalds 已提交
1119

1120
	return 1;
L
Linus Torvalds 已提交
1121 1122
}

1123
static int wait_noreap_copyout(struct task_struct *p, pid_t pid, uid_t uid,
L
Linus Torvalds 已提交
1124 1125 1126 1127 1128
			       int why, int status,
			       struct siginfo __user *infop,
			       struct rusage __user *rusagep)
{
	int retval = rusagep ? getrusage(p, RUSAGE_BOTH, rusagep) : 0;
1129

L
Linus Torvalds 已提交
1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153
	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
1154
static int wait_task_zombie(struct task_struct *p, int options,
L
Linus Torvalds 已提交
1155 1156 1157 1158
			    struct siginfo __user *infop,
			    int __user *stat_addr, struct rusage __user *ru)
{
	unsigned long state;
1159
	int retval, status, traced;
1160
	pid_t pid = task_pid_vnr(p);
1161
	uid_t uid = __task_cred(p)->uid;
L
Linus Torvalds 已提交
1162

R
Roland McGrath 已提交
1163 1164 1165 1166
	if (!likely(options & WEXITED))
		return 0;

	if (unlikely(options & WNOWAIT)) {
L
Linus Torvalds 已提交
1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192
		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;
	}

1193
	traced = ptrace_reparented(p);
1194 1195

	if (likely(!traced)) {
1196 1197
		struct signal_struct *psig;
		struct signal_struct *sig;
1198
		struct task_cputime cputime;
1199

L
Linus Torvalds 已提交
1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213
		/*
		 * 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.
1214 1215 1216 1217
		 *
		 * 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
1218
		 */
1219
		thread_group_cputime(p, &cputime);
L
Linus Torvalds 已提交
1220
		spin_lock_irq(&p->parent->sighand->siglock);
1221 1222 1223 1224
		psig = p->parent->signal;
		sig = p->signal;
		psig->cutime =
			cputime_add(psig->cutime,
1225 1226
			cputime_add(cputime.utime,
				    sig->cutime));
1227 1228
		psig->cstime =
			cputime_add(psig->cstime,
1229 1230
			cputime_add(cputime.stime,
				    sig->cstime));
1231 1232 1233 1234 1235
		psig->cgtime =
			cputime_add(psig->cgtime,
			cputime_add(p->gtime,
			cputime_add(sig->gtime,
				    sig->cgtime)));
1236 1237 1238 1239 1240 1241 1242 1243
		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;
1244 1245 1246 1247 1248 1249
		psig->cinblock +=
			task_io_get_inblock(p) +
			sig->inblock + sig->cinblock;
		psig->coublock +=
			task_io_get_oublock(p) +
			sig->oublock + sig->coublock;
1250 1251
		task_io_accounting_add(&psig->ioac, &p->ioac);
		task_io_accounting_add(&psig->ioac, &sig->ioac);
L
Linus Torvalds 已提交
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284
		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)
1285
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1286
	if (!retval && infop)
1287
		retval = put_user(uid, &infop->si_uid);
1288
	if (!retval)
1289
		retval = pid;
1290 1291

	if (traced) {
L
Linus Torvalds 已提交
1292
		write_lock_irq(&tasklist_lock);
1293 1294 1295 1296 1297 1298 1299
		/* 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.
		 */
1300
		if (!task_detached(p)) {
1301
			do_notify_parent(p, p->exit_signal);
1302
			if (!task_detached(p)) {
1303 1304
				p->exit_state = EXIT_ZOMBIE;
				p = NULL;
L
Linus Torvalds 已提交
1305 1306 1307 1308 1309 1310
			}
		}
		write_unlock_irq(&tasklist_lock);
	}
	if (p != NULL)
		release_task(p);
1311

L
Linus Torvalds 已提交
1312 1313 1314
	return retval;
}

1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
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
1327 1328 1329 1330 1331 1332
/*
 * 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
1333
static int wait_task_stopped(int ptrace, struct task_struct *p,
R
Roland McGrath 已提交
1334
			     int options, struct siginfo __user *infop,
L
Linus Torvalds 已提交
1335 1336
			     int __user *stat_addr, struct rusage __user *ru)
{
1337
	int retval, exit_code, *p_code, why;
1338
	uid_t uid = 0; /* unneeded, required by compiler */
1339
	pid_t pid;
L
Linus Torvalds 已提交
1340

R
Roland McGrath 已提交
1341
	if (!(options & WUNTRACED))
R
Roland McGrath 已提交
1342 1343
		return 0;

1344 1345 1346
	exit_code = 0;
	spin_lock_irq(&p->sighand->siglock);

1347 1348
	p_code = task_stopped_code(p, ptrace);
	if (unlikely(!p_code))
1349 1350
		goto unlock_sig;

1351
	exit_code = *p_code;
1352 1353 1354
	if (!exit_code)
		goto unlock_sig;

R
Roland McGrath 已提交
1355
	if (!unlikely(options & WNOWAIT))
1356
		*p_code = 0;
1357

1358 1359
	/* don't need the RCU readlock here as we're holding a spinlock */
	uid = __task_cred(p)->uid;
1360 1361 1362
unlock_sig:
	spin_unlock_irq(&p->sighand->siglock);
	if (!exit_code)
L
Linus Torvalds 已提交
1363 1364 1365 1366 1367 1368 1369 1370 1371 1372
		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);
1373
	pid = task_pid_vnr(p);
R
Roland McGrath 已提交
1374
	why = ptrace ? CLD_TRAPPED : CLD_STOPPED;
L
Linus Torvalds 已提交
1375 1376
	read_unlock(&tasklist_lock);

R
Roland McGrath 已提交
1377
	if (unlikely(options & WNOWAIT))
L
Linus Torvalds 已提交
1378
		return wait_noreap_copyout(p, pid, uid,
1379
					   why, exit_code,
L
Linus Torvalds 已提交
1380 1381 1382 1383 1384 1385 1386 1387 1388 1389
					   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
1390
		retval = put_user((short)why, &infop->si_code);
L
Linus Torvalds 已提交
1391 1392 1393
	if (!retval && infop)
		retval = put_user(exit_code, &infop->si_status);
	if (!retval && infop)
1394
		retval = put_user(pid, &infop->si_pid);
L
Linus Torvalds 已提交
1395
	if (!retval && infop)
1396
		retval = put_user(uid, &infop->si_uid);
L
Linus Torvalds 已提交
1397
	if (!retval)
1398
		retval = pid;
L
Linus Torvalds 已提交
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410
	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
1411
static int wait_task_continued(struct task_struct *p, int options,
L
Linus Torvalds 已提交
1412 1413 1414 1415 1416 1417 1418
			       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
1419 1420 1421
	if (!unlikely(options & WCONTINUED))
		return 0;

L
Linus Torvalds 已提交
1422 1423 1424 1425 1426 1427 1428 1429 1430
	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
1431
	if (!unlikely(options & WNOWAIT))
L
Linus Torvalds 已提交
1432
		p->signal->flags &= ~SIGNAL_STOP_CONTINUED;
1433
	uid = __task_cred(p)->uid;
L
Linus Torvalds 已提交
1434 1435
	spin_unlock_irq(&p->sighand->siglock);

1436
	pid = task_pid_vnr(p);
L
Linus Torvalds 已提交
1437 1438 1439 1440 1441 1442 1443 1444 1445
	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)
1446
			retval = pid;
L
Linus Torvalds 已提交
1447 1448 1449 1450 1451 1452 1453 1454 1455 1456
	} 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
1457 1458 1459 1460 1461 1462
/*
 * 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;
1463 1464
 * 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
1465
 */
R
Roland McGrath 已提交
1466
static int wait_consider_task(struct task_struct *parent, int ptrace,
R
Roland McGrath 已提交
1467 1468 1469 1470 1471 1472
			      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);
1473
	if (!ret)
R
Roland McGrath 已提交
1474 1475
		return ret;

1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
	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
1488 1489 1490 1491 1492 1493 1494 1495 1496
	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
1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511
	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;

1512
	if (task_stopped_code(p, ptrace))
R
Roland McGrath 已提交
1513 1514
		return wait_task_stopped(ptrace, p, options,
					 infop, stat_addr, ru);
R
Roland McGrath 已提交
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524

	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
1525 1526
 * *@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
1527 1528 1529 1530 1531 1532 1533 1534 1535
 */
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
1536 1537 1538 1539 1540 1541 1542 1543 1544 1545
		/*
		 * 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
1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558
	}

	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
1559
	 * Traditionally we see ptrace'd stopped tasks regardless of options.
R
Roland McGrath 已提交
1560
	 */
R
Roland McGrath 已提交
1561
	options |= WUNTRACED;
R
Roland McGrath 已提交
1562

R
Roland McGrath 已提交
1563 1564 1565 1566 1567
	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
1568 1569 1570 1571 1572 1573
			return ret;
	}

	return 0;
}

E
Eric W. Biederman 已提交
1574 1575 1576
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
1577 1578 1579
{
	DECLARE_WAITQUEUE(wait, current);
	struct task_struct *tsk;
R
Roland McGrath 已提交
1580
	int retval;
L
Linus Torvalds 已提交
1581

1582 1583
	trace_sched_process_wait(pid);

L
Linus Torvalds 已提交
1584 1585
	add_wait_queue(&current->signal->wait_chldexit,&wait);
repeat:
R
Roland McGrath 已提交
1586 1587 1588 1589 1590
	/*
	 * 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
1591 1592 1593 1594
	retval = -ECHILD;
	if ((type < PIDTYPE_MAX) && (!pid || hlist_empty(&pid->tasks[type])))
		goto end;

L
Linus Torvalds 已提交
1595 1596 1597 1598
	current->state = TASK_INTERRUPTIBLE;
	read_lock(&tasklist_lock);
	tsk = current;
	do {
R
Roland McGrath 已提交
1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
		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
1612
		}
R
Roland McGrath 已提交
1613

L
Linus Torvalds 已提交
1614 1615 1616
		if (options & __WNOTHREAD)
			break;
		tsk = next_thread(tsk);
E
Eric Sesterhenn 已提交
1617
		BUG_ON(tsk->signal != current->signal);
L
Linus Torvalds 已提交
1618 1619
	} while (tsk != current);
	read_unlock(&tasklist_lock);
O
Oleg Nesterov 已提交
1620

R
Roland McGrath 已提交
1621
	if (!retval && !(options & WNOHANG)) {
L
Linus Torvalds 已提交
1622
		retval = -ERESTARTSYS;
R
Roland McGrath 已提交
1623 1624 1625 1626
		if (!signal_pending(current)) {
			schedule();
			goto repeat;
		}
L
Linus Torvalds 已提交
1627
	}
R
Roland McGrath 已提交
1628

L
Linus Torvalds 已提交
1629 1630 1631 1632 1633
end:
	current->state = TASK_RUNNING;
	remove_wait_queue(&current->signal->wait_chldexit,&wait);
	if (infop) {
		if (retval > 0)
1634
			retval = 0;
L
Linus Torvalds 已提交
1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657
		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;
}

1658 1659
SYSCALL_DEFINE5(waitid, int, which, pid_t, upid, struct siginfo __user *,
		infop, int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1660
{
E
Eric W. Biederman 已提交
1661 1662
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1663 1664 1665 1666 1667 1668 1669 1670 1671
	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
1672
		type = PIDTYPE_MAX;
L
Linus Torvalds 已提交
1673 1674
		break;
	case P_PID:
E
Eric W. Biederman 已提交
1675 1676
		type = PIDTYPE_PID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1677 1678 1679
			return -EINVAL;
		break;
	case P_PGID:
E
Eric W. Biederman 已提交
1680 1681
		type = PIDTYPE_PGID;
		if (upid <= 0)
L
Linus Torvalds 已提交
1682 1683 1684 1685 1686 1687
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

E
Eric W. Biederman 已提交
1688 1689 1690 1691
	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
1692 1693

	/* avoid REGPARM breakage on x86: */
1694
	asmlinkage_protect(5, ret, which, upid, infop, options, ru);
L
Linus Torvalds 已提交
1695 1696 1697
	return ret;
}

1698 1699
SYSCALL_DEFINE4(wait4, pid_t, upid, int __user *, stat_addr,
		int, options, struct rusage __user *, ru)
L
Linus Torvalds 已提交
1700
{
E
Eric W. Biederman 已提交
1701 1702
	struct pid *pid = NULL;
	enum pid_type type;
L
Linus Torvalds 已提交
1703 1704 1705 1706 1707
	long ret;

	if (options & ~(WNOHANG|WUNTRACED|WCONTINUED|
			__WNOTHREAD|__WCLONE|__WALL))
		return -EINVAL;
E
Eric W. Biederman 已提交
1708 1709 1710 1711 1712 1713 1714 1715

	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;
1716
		pid = get_task_pid(current, PIDTYPE_PGID);
E
Eric W. Biederman 已提交
1717 1718 1719 1720 1721 1722 1723
	} 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
1724 1725

	/* avoid REGPARM breakage on x86: */
1726
	asmlinkage_protect(4, ret, upid, stat_addr, options, ru);
L
Linus Torvalds 已提交
1727 1728 1729 1730 1731 1732 1733 1734 1735
	return ret;
}

#ifdef __ARCH_WANT_SYS_WAITPID

/*
 * sys_waitpid() remains for compatibility. waitpid() should be
 * implemented by calling sys_wait4() from libc.a.
 */
1736
SYSCALL_DEFINE3(waitpid, pid_t, pid, int __user *, stat_addr, int, options)
L
Linus Torvalds 已提交
1737 1738 1739 1740 1741
{
	return sys_wait4(pid, stat_addr, options, NULL);
}

#endif