[linux-2.6-block.git] / kernel / exit.c

/*
 *  linux/kernel/exit.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

#include <linux/config.h>
#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/smp_lock.h>
#include <linux/module.h>
#include <linux/capability.h>
#include <linux/completion.h>
#include <linux/personality.h>
#include <linux/tty.h>
#include <linux/namespace.h>
#include <linux/key.h>
#include <linux/security.h>
#include <linux/cpu.h>
#include <linux/acct.h>
#include <linux/file.h>
#include <linux/binfmts.h>
#include <linux/ptrace.h>
#include <linux/profile.h>
#include <linux/mount.h>
#include <linux/proc_fs.h>
#include <linux/mempolicy.h>
#include <linux/cpuset.h>
#include <linux/syscalls.h>
#include <linux/signal.h>
#include <linux/cn_proc.h>
#include <linux/mutex.h>
#include <linux/futex.h>
#include <linux/compat.h>

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

extern void sem_exit (void);
extern struct task_struct *child_reaper;

int getrusage(struct task_struct *, int, struct rusage __user *);

static void exit_mm(struct task_struct * tsk);

static void __unhash_process(struct task_struct *p)
{
	nr_threads--;
	detach_pid(p, PIDTYPE_PID);
	detach_pid(p, PIDTYPE_TGID);
	if (thread_group_leader(p)) {
		detach_pid(p, PIDTYPE_PGID);
		detach_pid(p, PIDTYPE_SID);

		list_del_init(&p->tasks);
		__get_cpu_var(process_counts)--;
	}

	remove_parent(p);
}

void release_task(struct task_struct * p)
{
	int zap_leader;
	task_t *leader;
	struct dentry *proc_dentry;

repeat:
	atomic_dec(&p->user->processes);
	spin_lock(&p->proc_lock);
	proc_dentry = proc_pid_unhash(p);
	write_lock_irq(&tasklist_lock);
	ptrace_unlink(p);
	BUG_ON(!list_empty(&p->ptrace_list) || !list_empty(&p->ptrace_children));
	__exit_signal(p);
	/*
	 * Note that the fastpath in sys_times depends on __exit_signal having
	 * updated the counters before a task is removed from the tasklist of
	 * the process by __unhash_process.
	 */
	__unhash_process(p);

	/*
	 * If we are the last non-leader member of the thread
	 * group, and the leader is zombie, then notify the
	 * group leader's parent process. (if it wants notification.)
	 */
	zap_leader = 0;
	leader = p->group_leader;
	if (leader != p && thread_group_empty(leader) && leader->exit_state == EXIT_ZOMBIE) {
		BUG_ON(leader->exit_signal == -1);
		do_notify_parent(leader, leader->exit_signal);
		/*
		 * If we were the last child thread and the leader has
		 * exited already, and the leader's parent ignores SIGCHLD,
		 * then we are the one who should release the leader.
		 *
		 * do_notify_parent() will have marked it self-reaping in
		 * that case.
		 */
		zap_leader = (leader->exit_signal == -1);
	}

	sched_exit(p);
	write_unlock_irq(&tasklist_lock);
	spin_unlock(&p->proc_lock);
	proc_pid_flush(proc_dentry);
	release_thread(p);
	put_task_struct(p);

	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...
 */
int session_of_pgrp(int pgrp)
{
	struct task_struct *p;
	int sid = -1;

	read_lock(&tasklist_lock);
	do_each_task_pid(pgrp, PIDTYPE_PGID, p) {
		if (p->signal->session > 0) {
			sid = p->signal->session;
			goto out;
		}
	} while_each_task_pid(pgrp, PIDTYPE_PGID, p);
	p = find_task_by_pid(pgrp);
	if (p)
		sid = p->signal->session;
out:
	read_unlock(&tasklist_lock);
	
	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?"
 */
static int will_become_orphaned_pgrp(int pgrp, task_t *ignored_task)
{
	struct task_struct *p;
	int ret = 1;

	do_each_task_pid(pgrp, PIDTYPE_PGID, p) {
		if (p == ignored_task
				|| p->exit_state
				|| p->real_parent->pid == 1)
			continue;
		if (process_group(p->real_parent) != pgrp
			    && p->real_parent->signal->session == p->signal->session) {
			ret = 0;
			break;
		}
	} while_each_task_pid(pgrp, PIDTYPE_PGID, p);
	return ret;	/* (sighing) "Often!" */
}

int is_orphaned_pgrp(int pgrp)
{
	int retval;

	read_lock(&tasklist_lock);
	retval = will_become_orphaned_pgrp(pgrp, NULL);
	read_unlock(&tasklist_lock);

	return retval;
}

static int has_stopped_jobs(int pgrp)
{
	int retval = 0;
	struct task_struct *p;

	do_each_task_pid(pgrp, PIDTYPE_PGID, p) {
		if (p->state != TASK_STOPPED)
			continue;

		/* If p is stopped by a debugger on a signal that won't
		   stop it, then don't count p as stopped.  This isn't
		   perfect but it's a good approximation.  */
		if (unlikely (p->ptrace)
		    && p->exit_code != SIGSTOP
		    && p->exit_code != SIGTSTP
		    && p->exit_code != SIGTTOU
		    && p->exit_code != SIGTTIN)
			continue;

		retval = 1;
		break;
	} while_each_task_pid(pgrp, PIDTYPE_PGID, p);
	return retval;
}

/**
 * reparent_to_init - Reparent the calling kernel thread to the init task.
 *
 * If a kernel thread is launched as a result of a system call, or if
 * it ever exits, it should generally reparent itself to init so that
 * it is correctly cleaned up on exit.
 *
 * 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.
 *
 * NOTE that reparent_to_init() gives the caller full capabilities.
 */
static void reparent_to_init(void)
{
	write_lock_irq(&tasklist_lock);

	ptrace_unlink(current);
	/* Reparent to init */
	remove_parent(current);
	current->parent = child_reaper;
	current->real_parent = child_reaper;
	add_parent(current);

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

	if ((current->policy == SCHED_NORMAL ||
			current->policy == SCHED_BATCH)
				&& (task_nice(current) < 0))
		set_user_nice(current, 0);
	/* cpus_allowed? */
	/* rt_priority? */
	/* signals? */
	security_task_reparent_to_init(current);
	memcpy(current->signal->rlim, init_task.signal->rlim,
	       sizeof(current->signal->rlim));
	atomic_inc(&(INIT_USER->__count));
	write_unlock_irq(&tasklist_lock);
	switch_uid(INIT_USER);
}

void __set_special_pids(pid_t session, pid_t pgrp)
{
	struct task_struct *curr = current->group_leader;

	if (curr->signal->session != session) {
		detach_pid(curr, PIDTYPE_SID);
		curr->signal->session = session;
		attach_pid(curr, PIDTYPE_SID, session);
	}
	if (process_group(curr) != pgrp) {
		detach_pid(curr, PIDTYPE_PGID);
		curr->signal->pgrp = pgrp;
		attach_pid(curr, PIDTYPE_PGID, pgrp);
	}
}

void set_special_pids(pid_t session, pid_t pgrp)
{
	write_lock_irq(&tasklist_lock);
	__set_special_pids(session, pgrp);
	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)
{
	if (!valid_signal(sig) || sig < 1)
		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)
{
	if (!valid_signal(sig) || sig < 1)
		return -EINVAL;

	spin_lock_irq(&current->sighand->siglock);
	sigaddset(&current->blocked, sig);
	recalc_sigpending();
	spin_unlock_irq(&current->sighand->siglock);
	return 0;
}

EXPORT_SYMBOL(disallow_signal);

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

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

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

	/*
	 * If we were started as result of loading a module, close all of the
	 * user space pages.  We don't need them, and if we didn't close them
	 * they would be locked into memory.
	 */
	exit_mm(current);

	set_special_pids(1, 1);
	mutex_lock(&tty_mutex);
	current->signal->tty = NULL;
	mutex_unlock(&tty_mutex);

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

	/* Become as one with the init task */

	exit_fs(current);	/* current->fs->count--; */
	fs = init_task.fs;
	current->fs = fs;
	atomic_inc(&fs->count);
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/kernel/exit.c
	3	*
	4	* Copyright (C) 1991, 1992 Linus Torvalds
	5	*/
	6
	7	#include <linux/config.h>
	8	#include <linux/mm.h>
	9	#include <linux/slab.h>
	10	#include <linux/interrupt.h>
	11	#include <linux/smp_lock.h>
	12	#include <linux/module.h>
c59ede7b	13	#include <linux/capability.h>
1da177e4 LT	14	#include <linux/completion.h>
	15	#include <linux/personality.h>
	16	#include <linux/tty.h>
	17	#include <linux/namespace.h>
	18	#include <linux/key.h>
	19	#include <linux/security.h>
	20	#include <linux/cpu.h>
	21	#include <linux/acct.h>
	22	#include <linux/file.h>
	23	#include <linux/binfmts.h>
	24	#include <linux/ptrace.h>
	25	#include <linux/profile.h>
	26	#include <linux/mount.h>
	27	#include <linux/proc_fs.h>
	28	#include <linux/mempolicy.h>
	29	#include <linux/cpuset.h>
	30	#include <linux/syscalls.h>
7ed20e1a	31	#include <linux/signal.h>
9f46080c	32	#include <linux/cn_proc.h>
de5097c2	33	#include <linux/mutex.h>
0771dfef	34	#include <linux/futex.h>
34f192c6	35	#include <linux/compat.h>
1da177e4 LT	36
	37	#include <asm/uaccess.h>
	38	#include <asm/unistd.h>
	39	#include <asm/pgtable.h>
	40	#include <asm/mmu_context.h>
	41
	42	extern void sem_exit (void);
	43	extern struct task_struct *child_reaper;
	44
	45	int getrusage(struct task_struct , int, struct rusage __user );
	46
408b664a AB	47	static void exit_mm(struct task_struct * tsk);
408b664a AB	48
1da177e4 LT	49	static void __unhash_process(struct task_struct *p)
	50	{
	51	nr_threads--;
	52	detach_pid(p, PIDTYPE_PID);
	53	detach_pid(p, PIDTYPE_TGID);
	54	if (thread_group_leader(p)) {
	55	detach_pid(p, PIDTYPE_PGID);
	56	detach_pid(p, PIDTYPE_SID);
c97d9893 ON	57
c97d9893 ON	58	list_del_init(&p->tasks);
73b9ebfe	59	__get_cpu_var(process_counts)--;
1da177e4 LT	60	}
1da177e4 LT	61
c97d9893	62	remove_parent(p);
1da177e4 LT	63	}
	64
	65	void release_task(struct task_struct * p)
	66	{
	67	int zap_leader;
	68	task_t *leader;
	69	struct dentry *proc_dentry;
	70
1f09f974	71	repeat:
1da177e4 LT	72	atomic_dec(&p->user->processes);
	73	spin_lock(&p->proc_lock);
	74	proc_dentry = proc_pid_unhash(p);
	75	write_lock_irq(&tasklist_lock);
1f09f974	76	ptrace_unlink(p);
1da177e4 LT	77	BUG_ON(!list_empty(&p->ptrace_list) \|\| !list_empty(&p->ptrace_children));
1da177e4 LT	78	__exit_signal(p);
71a2224d CL	79	/*
	80	* Note that the fastpath in sys_times depends on __exit_signal having
	81	* updated the counters before a task is removed from the tasklist of
	82	* the process by __unhash_process.
	83	*/
1da177e4 LT	84	__unhash_process(p);
	85
	86	/*
	87	* If we are the last non-leader member of the thread
	88	* group, and the leader is zombie, then notify the
	89	* group leader's parent process. (if it wants notification.)
	90	*/
	91	zap_leader = 0;
	92	leader = p->group_leader;
	93	if (leader != p && thread_group_empty(leader) && leader->exit_state == EXIT_ZOMBIE) {
	94	BUG_ON(leader->exit_signal == -1);
	95	do_notify_parent(leader, leader->exit_signal);
	96	/*
	97	* If we were the last child thread and the leader has
	98	* exited already, and the leader's parent ignores SIGCHLD,
	99	* then we are the one who should release the leader.
	100	*
	101	* do_notify_parent() will have marked it self-reaping in
	102	* that case.
	103	*/
	104	zap_leader = (leader->exit_signal == -1);
	105	}
	106
	107	sched_exit(p);
	108	write_unlock_irq(&tasklist_lock);
	109	spin_unlock(&p->proc_lock);
	110	proc_pid_flush(proc_dentry);
	111	release_thread(p);
	112	put_task_struct(p);
	113
	114	p = leader;
	115	if (unlikely(zap_leader))
	116	goto repeat;
	117	}
	118
1da177e4 LT	119	/*
	120	* This checks not only the pgrp, but falls back on the pid if no
	121	* satisfactory pgrp is found. I dunno - gdb doesn't work correctly
	122	* without this...
	123	*/
	124	int session_of_pgrp(int pgrp)
	125	{
	126	struct task_struct *p;
	127	int sid = -1;
	128
	129	read_lock(&tasklist_lock);
	130	do_each_task_pid(pgrp, PIDTYPE_PGID, p) {
	131	if (p->signal->session > 0) {
	132	sid = p->signal->session;
	133	goto out;
	134	}
	135	} while_each_task_pid(pgrp, PIDTYPE_PGID, p);
	136	p = find_task_by_pid(pgrp);
	137	if (p)
	138	sid = p->signal->session;
	139	out:
	140	read_unlock(&tasklist_lock);
	141
	142	return sid;
	143	}
	144
	145	/*
	146	* Determine if a process group is "orphaned", according to the POSIX
	147	* definition in 2.2.2.52. Orphaned process groups are not to be affected
	148	* by terminal-generated stop signals. Newly orphaned process groups are
	149	* to receive a SIGHUP and a SIGCONT.
	150	*
	151	* "I ask you, have you ever known what it is to be an orphan?"
	152	*/
	153	static int will_become_orphaned_pgrp(int pgrp, task_t *ignored_task)
	154	{
	155	struct task_struct *p;
	156	int ret = 1;
	157
	158	do_each_task_pid(pgrp, PIDTYPE_PGID, p) {
	159	if (p == ignored_task
	160	\|\| p->exit_state
	161	\|\| p->real_parent->pid == 1)
	162	continue;
	163	if (process_group(p->real_parent) != pgrp
	164	&& p->real_parent->signal->session == p->signal->session) {
	165	ret = 0;
	166	break;
	167	}
	168	} while_each_task_pid(pgrp, PIDTYPE_PGID, p);
	169	return ret; /* (sighing) "Often!" */
	170	}
	171
	172	int is_orphaned_pgrp(int pgrp)
	173	{
	174	int retval;
	175
	176	read_lock(&tasklist_lock);
	177	retval = will_become_orphaned_pgrp(pgrp, NULL);
	178	read_unlock(&tasklist_lock);
	179
	180	return retval;
	181	}
	182
858119e1	183	static int has_stopped_jobs(int pgrp)
1da177e4 LT	184	{
	185	int retval = 0;
	186	struct task_struct *p;
	187
	188	do_each_task_pid(pgrp, PIDTYPE_PGID, p) {
	189	if (p->state != TASK_STOPPED)
	190	continue;
	191
	192	/* If p is stopped by a debugger on a signal that won't
	193	stop it, then don't count p as stopped. This isn't
	194	perfect but it's a good approximation. */
	195	if (unlikely (p->ptrace)
	196	&& p->exit_code != SIGSTOP
	197	&& p->exit_code != SIGTSTP
	198	&& p->exit_code != SIGTTOU
	199	&& p->exit_code != SIGTTIN)
	200	continue;
	201
	202	retval = 1;
	203	break;
	204	} while_each_task_pid(pgrp, PIDTYPE_PGID, p);
	205	return retval;
	206	}
	207
	208	/**
4dc3b16b	209	* reparent_to_init - Reparent the calling kernel thread to the init task.
1da177e4 LT	210	*
	211	* If a kernel thread is launched as a result of a system call, or if
	212	* it ever exits, it should generally reparent itself to init so that
	213	* it is correctly cleaned up on exit.
	214	*
	215	* The various task state such as scheduling policy and priority may have
	216	* been inherited from a user process, so we reset them to sane values here.
	217	*
	218	* NOTE that reparent_to_init() gives the caller full capabilities.
	219	*/
858119e1	220	static void reparent_to_init(void)
1da177e4 LT	221	{
	222	write_lock_irq(&tasklist_lock);
	223
	224	ptrace_unlink(current);
	225	/* Reparent to init */
9b678ece	226	remove_parent(current);
1da177e4 LT	227	current->parent = child_reaper;
1da177e4 LT	228	current->real_parent = child_reaper;
9b678ece	229	add_parent(current);
1da177e4 LT	230
	231	/* Set the exit signal to SIGCHLD so we signal init on exit */
	232	current->exit_signal = SIGCHLD;
	233
b0a9499c IM	234	if ((current->policy == SCHED_NORMAL \|\|
	235	current->policy == SCHED_BATCH)
	236	&& (task_nice(current) < 0))
1da177e4 LT	237	set_user_nice(current, 0);
	238	/* cpus_allowed? */
	239	/* rt_priority? */
	240	/* signals? */
	241	security_task_reparent_to_init(current);
	242	memcpy(current->signal->rlim, init_task.signal->rlim,
	243	sizeof(current->signal->rlim));
	244	atomic_inc(&(INIT_USER->__count));
	245	write_unlock_irq(&tasklist_lock);
	246	switch_uid(INIT_USER);
	247	}
	248
	249	void __set_special_pids(pid_t session, pid_t pgrp)
	250	{
e19f247a	251	struct task_struct *curr = current->group_leader;
1da177e4 LT	252
	253	if (curr->signal->session != session) {
	254	detach_pid(curr, PIDTYPE_SID);
	255	curr->signal->session = session;
	256	attach_pid(curr, PIDTYPE_SID, session);
	257	}
	258	if (process_group(curr) != pgrp) {
	259	detach_pid(curr, PIDTYPE_PGID);
	260	curr->signal->pgrp = pgrp;
	261	attach_pid(curr, PIDTYPE_PGID, pgrp);
	262	}
	263	}
	264
	265	void set_special_pids(pid_t session, pid_t pgrp)
	266	{
	267	write_lock_irq(&tasklist_lock);
	268	__set_special_pids(session, pgrp);
	269	write_unlock_irq(&tasklist_lock);
	270	}
	271
	272	/*
	273	* Let kernel threads use this to say that they
	274	* allow a certain signal (since daemonize() will
	275	* have disabled all of them by default).
	276	*/
	277	int allow_signal(int sig)
	278	{
7ed20e1a	279	if (!valid_signal(sig) \|\| sig < 1)
1da177e4 LT	280	return -EINVAL;
	281
	282	spin_lock_irq(&current->sighand->siglock);
	283	sigdelset(&current->blocked, sig);
	284	if (!current->mm) {
	285	/* Kernel threads handle their own signals.
	286	Let the signal code know it'll be handled, so
	287	that they don't get converted to SIGKILL or
	288	just silently dropped */
	289	current->sighand->action[(sig)-1].sa.sa_handler = (void __user *)2;
	290	}
	291	recalc_sigpending();
	292	spin_unlock_irq(&current->sighand->siglock);
	293	return 0;
	294	}
	295
	296	EXPORT_SYMBOL(allow_signal);
	297
	298	int disallow_signal(int sig)
	299	{
7ed20e1a	300	if (!valid_signal(sig) \|\| sig < 1)
1da177e4 LT	301	return -EINVAL;
	302
	303	spin_lock_irq(&current->sighand->siglock);
	304	sigaddset(&current->blocked, sig);
	305	recalc_sigpending();
	306	spin_unlock_irq(&current->sighand->siglock);
	307	return 0;
	308	}
	309
	310	EXPORT_SYMBOL(disallow_signal);
	311
	312	/*
	313	* Put all the gunge required to become a kernel thread without
	314	* attached user resources in one place where it belongs.
	315	*/
	316
	317	void daemonize(const char *name, ...)
	318	{
	319	va_list args;
	320	struct fs_struct *fs;
	321	sigset_t blocked;
	322
	323	va_start(args, name);
	324	vsnprintf(current->comm, sizeof(current->comm), name, args);
	325	va_end(args);
	326
	327	/*
	328	* If we were started as result of loading a module, close all of the
	329	* user space pages. We don't need them, and if we didn't close them
	330	* they would be locked into memory.
	331	*/
	332	exit_mm(current);
	333
	334	set_special_pids(1, 1);
70522e12	335	mutex_lock(&tty_mutex);
1da177e4	336	current->signal->tty = NULL;
70522e12	337	mutex_unlock(&tty_mutex);
1da177e4 LT	338
	339	/* Block and flush all signals */
	340	sigfillset(&blocked);
	341	sigprocmask(SIG_BLOCK, &blocked, NULL);
	342	flush_signals(current);
	343
	344	/* Become as one with the init task */
	345
	346	exit_fs(current); /* current->fs->count--; */
	347	fs = init_task.fs;
	348	current->fs = fs;
	349	atomic_inc(&fs->count);