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