[deliverable/linux.git] / fs / notify / mark.c

/*
 *  Copyright (C) 2008 Red Hat, Inc., Eric Paris <eparis@redhat.com>
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2, or (at your option)
 *  any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; see the file COPYING.  If not, write to
 *  the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
 */

/*
 * fsnotify inode mark locking/lifetime/and refcnting
 *
 * REFCNT:
 * The mark->refcnt tells how many "things" in the kernel currently are
 * referencing this object.  The object typically will live inside the kernel
 * with a refcnt of 2, one for each list it is on (i_list, g_list).  Any task
 * which can find this object holding the appropriete locks, can take a reference
 * and the object itself is guaranteed to survive until the reference is dropped.
 *
 * LOCKING:
 * There are 3 spinlocks involved with fsnotify inode marks and they MUST
 * be taken in order as follows:
 *
 * mark->lock
 * group->mark_lock
 * inode->i_lock
 *
 * mark->lock protects 2 things, mark->group and mark->inode.  You must hold
 * that lock to dereference either of these things (they could be NULL even with
 * the lock)
 *
 * group->mark_lock protects the marks_list anchored inside a given group
 * and each mark is hooked via the g_list.  It also sorta protects the
 * free_g_list, which when used is anchored by a private list on the stack of the
 * task which held the group->mark_lock.
 *
 * inode->i_lock protects the i_fsnotify_marks list anchored inside a
 * given inode and each mark is hooked via the i_list. (and sorta the
 * free_i_list)
 *
 *
 * LIFETIME:
 * Inode marks survive between when they are added to an inode and when their
 * refcnt==0.
 *
 * The inode mark can be cleared for a number of different reasons including:
 * - The inode is unlinked for the last time.  (fsnotify_inode_remove)
 * - The inode is being evicted from cache. (fsnotify_inode_delete)
 * - The fs the inode is on is unmounted.  (fsnotify_inode_delete/fsnotify_unmount_inodes)
 * - Something explicitly requests that it be removed.  (fsnotify_destroy_mark)
 * - The fsnotify_group associated with the mark is going away and all such marks
 *   need to be cleaned up. (fsnotify_clear_marks_by_group)
 *
 * Worst case we are given an inode and need to clean up all the marks on that
 * inode.  We take i_lock and walk the i_fsnotify_marks safely.  For each
 * mark on the list we take a reference (so the mark can't disappear under us).
 * We remove that mark form the inode's list of marks and we add this mark to a
 * private list anchored on the stack using i_free_list;  At this point we no
 * longer fear anything finding the mark using the inode's list of marks.
 *
 * We can safely and locklessly run the private list on the stack of everything
 * we just unattached from the original inode.  For each mark on the private list
 * we grab the mark-> and can thus dereference mark->group and mark->inode.  If
 * we see the group and inode are not NULL we take those locks.  Now holding all
 * 3 locks we can completely remove the mark from other tasks finding it in the
 * future.  Remember, 10 things might already be referencing this mark, but they
 * better be holding a ref.  We drop our reference we took before we unhooked it
 * from the inode.  When the ref hits 0 we can free the mark.
 *
 * Very similarly for freeing by group, except we use free_g_list.
 *
 * This has the very interesting property of being able to run concurrently with
 * any (or all) other directions.
 */

#include <linux/fs.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/kthread.h>
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/srcu.h>

#include <linux/atomic.h>

#include <linux/fsnotify_backend.h>
#include "fsnotify.h"

struct srcu_struct fsnotify_mark_srcu;
static DEFINE_SPINLOCK(destroy_lock);
static LIST_HEAD(destroy_list);
static DECLARE_WAIT_QUEUE_HEAD(destroy_waitq);

void fsnotify_get_mark(struct fsnotify_mark *mark)
{
	atomic_inc(&mark->refcnt);
}

void fsnotify_put_mark(struct fsnotify_mark *mark)
{
	if (atomic_dec_and_test(&mark->refcnt)) {
		if (mark->group)
			fsnotify_put_group(mark->group);
		mark->free_mark(mark);
	}
}

/*
 * Any time a mark is getting freed we end up here.
 * The caller had better be holding a reference to this mark so we don't actually
 * do the final put under the mark->lock
 */
void fsnotify_destroy_mark_locked(struct fsnotify_mark *mark,
				  struct fsnotify_group *group)
{
	struct inode *inode = NULL;

	BUG_ON(!mutex_is_locked(&group->mark_mutex));

	spin_lock(&mark->lock);

	/* something else already called this function on this mark */
	if (!(mark->flags & FSNOTIFY_MARK_FLAG_ALIVE)) {
		spin_unlock(&mark->lock);
		return;
	}

	mark->flags &= ~FSNOTIFY_MARK_FLAG_ALIVE;

	if (mark->flags & FSNOTIFY_MARK_FLAG_INODE) {
		inode = mark->i.inode;
		fsnotify_destroy_inode_mark(mark);
	} else if (mark->flags & FSNOTIFY_MARK_FLAG_VFSMOUNT)
		fsnotify_destroy_vfsmount_mark(mark);
	else
		BUG();

	list_del_init(&mark->g_list);

	spin_unlock(&mark->lock);

	if (inode && (mark->flags & FSNOTIFY_MARK_FLAG_OBJECT_PINNED))
		iput(inode);
	/* release lock temporarily */
	mutex_unlock(&group->mark_mutex);

	spin_lock(&destroy_lock);
	list_add(&mark->destroy_list, &destroy_list);
	spin_unlock(&destroy_lock);
	wake_up(&destroy_waitq);
	/*
	 * We don't necessarily have a ref on mark from caller so the above destroy
	 * may have actually freed it, unless this group provides a 'freeing_mark'
	 * function which must be holding a reference.
	 */

	/*
	 * Some groups like to know that marks are being freed.  This is a
	 * callback to the group function to let it know that this mark
	 * is being freed.
	 */
	if (group->ops->freeing_mark)
		group->ops->freeing_mark(mark, group);

	/*
	 * __fsnotify_update_child_dentry_flags(inode);
	 *
	 * I really want to call that, but we can't, we have no idea if the inode
	 * still exists the second we drop the mark->lock.
	 *
	 * The next time an event arrive to this inode from one of it's children
	 * __fsnotify_parent will see that the inode doesn't care about it's
	 * children and will update all of these flags then.  So really this
	 * is just a lazy update (and could be a perf win...)
	 */

	atomic_dec(&group->num_marks);

	mutex_lock_nested(&group->mark_mutex, SINGLE_DEPTH_NESTING);
}

void fsnotify_destroy_mark(struct fsnotify_mark *mark,
			   struct fsnotify_group *group)
{
	mutex_lock_nested(&group->mark_mutex, SINGLE_DEPTH_NESTING);
	fsnotify_destroy_mark_locked(mark, group);
	mutex_unlock(&group->mark_mutex);
}

void fsnotify_set_mark_mask_locked(struct fsnotify_mark *mark, __u32 mask)
{
	assert_spin_locked(&mark->lock);

	mark->mask = mask;

	if (mark->flags & FSNOTIFY_MARK_FLAG_INODE)
		fsnotify_set_inode_mark_mask_locked(mark, mask);
}

void fsnotify_set_mark_ignored_mask_locked(struct fsnotify_mark *mark, __u32 mask)
{
	assert_spin_locked(&mark->lock);

	mark->ignored_mask = mask;
}

/*
 * Attach an initialized mark to a given group and fs object.
 * These marks may be used for the fsnotify backend to determine which
 * event types should be delivered to which group.
 */
int fsnotify_add_mark_locked(struct fsnotify_mark *mark,
			     struct fsnotify_group *group, struct inode *inode,
			     struct vfsmount *mnt, int allow_dups)
{
	int ret = 0;

	BUG_ON(inode && mnt);
	BUG_ON(!inode && !mnt);
	BUG_ON(!mutex_is_locked(&group->mark_mutex));

	/*
	 * LOCKING ORDER!!!!
	 * group->mark_mutex
	 * mark->lock
	 * inode->i_lock
	 */
	spin_lock(&mark->lock);
	mark->flags |= FSNOTIFY_MARK_FLAG_ALIVE;

	fsnotify_get_group(group);
	mark->group = group;
	list_add(&mark->g_list, &group->marks_list);
	atomic_inc(&group->num_marks);
	fsnotify_get_mark(mark); /* for i_list and g_list */

	if (inode) {
		ret = fsnotify_add_inode_mark(mark, group, inode, allow_dups);
		if (ret)
			goto err;
	} else if (mnt) {
		ret = fsnotify_add_vfsmount_mark(mark, group, mnt, allow_dups);
		if (ret)
			goto err;
	} else {
		BUG();
	}

	/* this will pin the object if appropriate */
	fsnotify_set_mark_mask_locked(mark, mark->mask);
	spin_unlock(&mark->lock);

	if (inode)
		__fsnotify_update_child_dentry_flags(inode);

	return ret;
err:
	mark->flags &= ~FSNOTIFY_MARK_FLAG_ALIVE;
	list_del_init(&mark->g_list);
	fsnotify_put_group(group);
	mark->group = NULL;
	atomic_dec(&group->num_marks);

	spin_unlock(&mark->lock);

	spin_lock(&destroy_lock);
	list_add(&mark->destroy_list, &destroy_list);
	spin_unlock(&destroy_lock);
	wake_up(&destroy_waitq);

	return ret;
}

int fsnotify_add_mark(struct fsnotify_mark *mark, struct fsnotify_group *group,
		      struct inode *inode, struct vfsmount *mnt, int allow_dups)
{
	int ret;
	mutex_lock(&group->mark_mutex);
	ret = fsnotify_add_mark_locked(mark, group, inode, mnt, allow_dups);
	mutex_unlock(&group->mark_mutex);
	return ret;
}

/*
 * clear any marks in a group in which mark->flags & flags is true
 */
void fsnotify_clear_marks_by_group_flags(struct fsnotify_group *group,
					 unsigned int flags)
{
	struct fsnotify_mark *lmark, *mark;

	mutex_lock_nested(&group->mark_mutex, SINGLE_DEPTH_NESTING);
	list_for_each_entry_safe(mark, lmark, &group->marks_list, g_list) {
		if (mark->flags & flags) {
			fsnotify_get_mark(mark);
			fsnotify_destroy_mark_locked(mark, group);
			fsnotify_put_mark(mark);
		}
	}
	mutex_unlock(&group->mark_mutex);
}

/*
 * Given a group, destroy all of the marks associated with that group.
 */
void fsnotify_clear_marks_by_group(struct fsnotify_group *group)
{
	fsnotify_clear_marks_by_group_flags(group, (unsigned int)-1);
}

void fsnotify_duplicate_mark(struct fsnotify_mark *new, struct fsnotify_mark *old)
{
	assert_spin_locked(&old->lock);
	new->i.inode = old->i.inode;
	new->m.mnt = old->m.mnt;
	if (old->group)
		fsnotify_get_group(old->group);
	new->group = old->group;
	new->mask = old->mask;
	new->free_mark = old->free_mark;
}

/*
 * Nothing fancy, just initialize lists and locks and counters.
 */
void fsnotify_init_mark(struct fsnotify_mark *mark,
			void (*free_mark)(struct fsnotify_mark *mark))
{
	memset(mark, 0, sizeof(*mark));
	spin_lock_init(&mark->lock);
	atomic_set(&mark->refcnt, 1);
	mark->free_mark = free_mark;
}

static int fsnotify_mark_destroy(void *ignored)
{
	struct fsnotify_mark *mark, *next;
	LIST_HEAD(private_destroy_list);

	for (;;) {
		spin_lock(&destroy_lock);
		/* exchange the list head */
		list_replace_init(&destroy_list, &private_destroy_list);
		spin_unlock(&destroy_lock);

		synchronize_srcu(&fsnotify_mark_srcu);

		list_for_each_entry_safe(mark, next, &private_destroy_list, destroy_list) {
			list_del_init(&mark->destroy_list);
			fsnotify_put_mark(mark);
		}

		wait_event_interruptible(destroy_waitq, !list_empty(&destroy_list));
	}

	return 0;
}

static int __init fsnotify_mark_init(void)
{
	struct task_struct *thread;

	thread = kthread_run(fsnotify_mark_destroy, NULL,
			     "fsnotify_mark");
	if (IS_ERR(thread))
		panic("unable to start fsnotify mark destruction thread.");

	return 0;
}
device_initcall(fsnotify_mark_init);
Commit	Line	Data
5444e298 EP	1	/*
	2	* Copyright (C) 2008 Red Hat, Inc., Eric Paris <eparis@redhat.com>
	3	*
	4	* This program is free software; you can redistribute it and/or modify
	5	* it under the terms of the GNU General Public License as published by
	6	* the Free Software Foundation; either version 2, or (at your option)
	7	* any later version.
	8	*
	9	* This program is distributed in the hope that it will be useful,
	10	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	11	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	12	* GNU General Public License for more details.
	13	*
	14	* You should have received a copy of the GNU General Public License
	15	* along with this program; see the file COPYING. If not, write to
	16	* the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
	17	*/
	18
	19	/*
	20	* fsnotify inode mark locking/lifetime/and refcnting
	21	*
	22	* REFCNT:
	23	* The mark->refcnt tells how many "things" in the kernel currently are
	24	* referencing this object. The object typically will live inside the kernel
	25	* with a refcnt of 2, one for each list it is on (i_list, g_list). Any task
	26	* which can find this object holding the appropriete locks, can take a reference
25985edc	27	* and the object itself is guaranteed to survive until the reference is dropped.
5444e298 EP	28	*
	29	* LOCKING:
	30	* There are 3 spinlocks involved with fsnotify inode marks and they MUST
	31	* be taken in order as follows:
	32	*
	33	* mark->lock
	34	* group->mark_lock
	35	* inode->i_lock
	36	*
	37	* mark->lock protects 2 things, mark->group and mark->inode. You must hold
	38	* that lock to dereference either of these things (they could be NULL even with
	39	* the lock)
	40	*
	41	* group->mark_lock protects the marks_list anchored inside a given group
	42	* and each mark is hooked via the g_list. It also sorta protects the
	43	* free_g_list, which when used is anchored by a private list on the stack of the
	44	* task which held the group->mark_lock.
	45	*
	46	* inode->i_lock protects the i_fsnotify_marks list anchored inside a
	47	* given inode and each mark is hooked via the i_list. (and sorta the
	48	* free_i_list)
	49	*
	50	*
	51	* LIFETIME:
	52	* Inode marks survive between when they are added to an inode and when their
	53	* refcnt==0.
	54	*
	55	* The inode mark can be cleared for a number of different reasons including:
	56	* - The inode is unlinked for the last time. (fsnotify_inode_remove)
	57	* - The inode is being evicted from cache. (fsnotify_inode_delete)
	58	* - The fs the inode is on is unmounted. (fsnotify_inode_delete/fsnotify_unmount_inodes)
	59	* - Something explicitly requests that it be removed. (fsnotify_destroy_mark)
	60	* - The fsnotify_group associated with the mark is going away and all such marks
	61	* need to be cleaned up. (fsnotify_clear_marks_by_group)
	62	*
	63	* Worst case we are given an inode and need to clean up all the marks on that
	64	* inode. We take i_lock and walk the i_fsnotify_marks safely. For each
	65	* mark on the list we take a reference (so the mark can't disappear under us).
	66	* We remove that mark form the inode's list of marks and we add this mark to a
	67	* private list anchored on the stack using i_free_list; At this point we no
	68	* longer fear anything finding the mark using the inode's list of marks.
	69	*
	70	* We can safely and locklessly run the private list on the stack of everything
	71	* we just unattached from the original inode. For each mark on the private list
	72	* we grab the mark-> and can thus dereference mark->group and mark->inode. If
	73	* we see the group and inode are not NULL we take those locks. Now holding all
	74	* 3 locks we can completely remove the mark from other tasks finding it in the
	75	* future. Remember, 10 things might already be referencing this mark, but they
	76	* better be holding a ref. We drop our reference we took before we unhooked it
	77	* from the inode. When the ref hits 0 we can free the mark.
	78	*
	79	* Very similarly for freeing by group, except we use free_g_list.
	80	*
	81	* This has the very interesting property of being able to run concurrently with
	82	* any (or all) other directions.
	83	*/
	84
	85	#include <linux/fs.h>
	86	#include <linux/init.h>
	87	#include <linux/kernel.h>
75c1be48	88	#include <linux/kthread.h>
5444e298 EP	89	#include <linux/module.h>
	90	#include <linux/mutex.h>
	91	#include <linux/slab.h>
	92	#include <linux/spinlock.h>
75c1be48	93	#include <linux/srcu.h>
5444e298	94
60063497	95	#include <linux/atomic.h>
5444e298 EP	96
	97	#include <linux/fsnotify_backend.h>
	98	#include "fsnotify.h"
	99
75c1be48 EP	100	struct srcu_struct fsnotify_mark_srcu;
	101	static DEFINE_SPINLOCK(destroy_lock);
	102	static LIST_HEAD(destroy_list);
	103	static DECLARE_WAIT_QUEUE_HEAD(destroy_waitq);
	104
5444e298 EP	105	void fsnotify_get_mark(struct fsnotify_mark *mark)
	106	{
	107	atomic_inc(&mark->refcnt);
	108	}
	109
	110	void fsnotify_put_mark(struct fsnotify_mark *mark)
	111	{
23e964c2 LS	112	if (atomic_dec_and_test(&mark->refcnt)) {
	113	if (mark->group)
	114	fsnotify_put_group(mark->group);
5444e298	115	mark->free_mark(mark);
23e964c2	116	}
5444e298 EP	117	}
	118
	119	/*
	120	* Any time a mark is getting freed we end up here.
	121	* The caller had better be holding a reference to this mark so we don't actually
	122	* do the final put under the mark->lock
	123	*/
d5a335b8 LS	124	void fsnotify_destroy_mark_locked(struct fsnotify_mark *mark,
d5a335b8 LS	125	struct fsnotify_group *group)
5444e298	126	{
0d48b7f0	127	struct inode *inode = NULL;
5444e298	128
d5a335b8 LS	129	BUG_ON(!mutex_is_locked(&group->mark_mutex));
d5a335b8 LS	130
104d06f0	131	spin_lock(&mark->lock);
5444e298	132
700307a2 EP	133	/* something else already called this function on this mark */
700307a2 EP	134	if (!(mark->flags & FSNOTIFY_MARK_FLAG_ALIVE)) {
5444e298	135	spin_unlock(&mark->lock);
e2a29943	136	return;
5444e298 EP	137	}
5444e298 EP	138
700307a2 EP	139	mark->flags &= ~FSNOTIFY_MARK_FLAG_ALIVE;
700307a2 EP	140
0d48b7f0	141	if (mark->flags & FSNOTIFY_MARK_FLAG_INODE) {
0d48b7f0	142	inode = mark->i.inode;
b31d397e	143	fsnotify_destroy_inode_mark(mark);
0d48b7f0 EP	144	} else if (mark->flags & FSNOTIFY_MARK_FLAG_VFSMOUNT)
0d48b7f0 EP	145	fsnotify_destroy_vfsmount_mark(mark);
5444e298 EP	146	else
	147	BUG();
	148
	149	list_del_init(&mark->g_list);
5444e298	150
5444e298	151	spin_unlock(&mark->lock);
d5a335b8	152
6960b0d9 LS	153	if (inode && (mark->flags & FSNOTIFY_MARK_FLAG_OBJECT_PINNED))
6960b0d9 LS	154	iput(inode);
d5a335b8	155	/* release lock temporarily */
986ab098	156	mutex_unlock(&group->mark_mutex);
5444e298	157
75c1be48 EP	158	spin_lock(&destroy_lock);
	159	list_add(&mark->destroy_list, &destroy_list);
	160	spin_unlock(&destroy_lock);
	161	wake_up(&destroy_waitq);
6960b0d9 LS	162	/*
	163	* We don't necessarily have a ref on mark from caller so the above destroy
	164	* may have actually freed it, unless this group provides a 'freeing_mark'
	165	* function which must be holding a reference.
	166	*/
75c1be48	167
5444e298 EP	168	/*
	169	* Some groups like to know that marks are being freed. This is a
	170	* callback to the group function to let it know that this mark
	171	* is being freed.
	172	*/
	173	if (group->ops->freeing_mark)
	174	group->ops->freeing_mark(mark, group);
	175
	176	/*
	177	* __fsnotify_update_child_dentry_flags(inode);
	178	*
	179	* I really want to call that, but we can't, we have no idea if the inode
	180	* still exists the second we drop the mark->lock.
	181	*
	182	* The next time an event arrive to this inode from one of it's children
	183	* __fsnotify_parent will see that the inode doesn't care about it's
	184	* children and will update all of these flags then. So really this
	185	* is just a lazy update (and could be a perf win...)
	186	*/
	187
23e964c2	188	atomic_dec(&group->num_marks);
d5a335b8	189
6960b0d9	190	mutex_lock_nested(&group->mark_mutex, SINGLE_DEPTH_NESTING);
d5a335b8 LS	191	}
	192
	193	void fsnotify_destroy_mark(struct fsnotify_mark *mark,
	194	struct fsnotify_group *group)
	195	{
6960b0d9	196	mutex_lock_nested(&group->mark_mutex, SINGLE_DEPTH_NESTING);
d5a335b8 LS	197	fsnotify_destroy_mark_locked(mark, group);
d5a335b8 LS	198	mutex_unlock(&group->mark_mutex);
5444e298 EP	199	}
5444e298 EP	200
90b1e7a5 EP	201	void fsnotify_set_mark_mask_locked(struct fsnotify_mark *mark, __u32 mask)
	202	{
	203	assert_spin_locked(&mark->lock);
	204
	205	mark->mask = mask;
	206
	207	if (mark->flags & FSNOTIFY_MARK_FLAG_INODE)
	208	fsnotify_set_inode_mark_mask_locked(mark, mask);
	209	}
	210
33af5e32 EP	211	void fsnotify_set_mark_ignored_mask_locked(struct fsnotify_mark *mark, __u32 mask)
	212	{
	213	assert_spin_locked(&mark->lock);
	214
	215	mark->ignored_mask = mask;
	216	}
90b1e7a5	217
5444e298 EP	218	/*
	219	* Attach an initialized mark to a given group and fs object.
	220	* These marks may be used for the fsnotify backend to determine which
	221	* event types should be delivered to which group.
	222	*/
d5a335b8 LS	223	int fsnotify_add_mark_locked(struct fsnotify_mark *mark,
	224	struct fsnotify_group group, struct inode inode,
	225	struct vfsmount *mnt, int allow_dups)
5444e298 EP	226	{
	227	int ret = 0;
	228
5444e298 EP	229	BUG_ON(inode && mnt);
5444e298 EP	230	BUG_ON(!inode && !mnt);
d5a335b8	231	BUG_ON(!mutex_is_locked(&group->mark_mutex));
5444e298	232
5444e298 EP	233	/*
5444e298 EP	234	* LOCKING ORDER!!!!
986ab098	235	* group->mark_mutex
104d06f0	236	* mark->lock
5444e298 EP	237	* inode->i_lock
5444e298 EP	238	*/
104d06f0	239	spin_lock(&mark->lock);
700307a2 EP	240	mark->flags \|= FSNOTIFY_MARK_FLAG_ALIVE;
700307a2 EP	241
23e964c2	242	fsnotify_get_group(group);
5444e298 EP	243	mark->group = group;
	244	list_add(&mark->g_list, &group->marks_list);
	245	atomic_inc(&group->num_marks);
	246	fsnotify_get_mark(mark); /* for i_list and g_list */
	247
	248	if (inode) {
	249	ret = fsnotify_add_inode_mark(mark, group, inode, allow_dups);
	250	if (ret)
	251	goto err;
0d48b7f0 EP	252	} else if (mnt) {
	253	ret = fsnotify_add_vfsmount_mark(mark, group, mnt, allow_dups);
	254	if (ret)
	255	goto err;
5444e298 EP	256	} else {
	257	BUG();
	258	}
	259
90b1e7a5 EP	260	/* this will pin the object if appropriate */
90b1e7a5 EP	261	fsnotify_set_mark_mask_locked(mark, mark->mask);
5444e298 EP	262	spin_unlock(&mark->lock);
	263
	264	if (inode)
	265	__fsnotify_update_child_dentry_flags(inode);
	266
	267	return ret;
	268	err:
700307a2	269	mark->flags &= ~FSNOTIFY_MARK_FLAG_ALIVE;
5444e298	270	list_del_init(&mark->g_list);
23e964c2	271	fsnotify_put_group(group);
75c1be48	272	mark->group = NULL;
5444e298	273	atomic_dec(&group->num_marks);
5444e298	274
5444e298 EP	275	spin_unlock(&mark->lock);
5444e298 EP	276
75c1be48 EP	277	spin_lock(&destroy_lock);
	278	list_add(&mark->destroy_list, &destroy_list);
	279	spin_unlock(&destroy_lock);
	280	wake_up(&destroy_waitq);
	281
5444e298 EP	282	return ret;
	283	}
	284
d5a335b8 LS	285	int fsnotify_add_mark(struct fsnotify_mark mark, struct fsnotify_group group,
	286	struct inode inode, struct vfsmount mnt, int allow_dups)
	287	{
	288	int ret;
	289	mutex_lock(&group->mark_mutex);
	290	ret = fsnotify_add_mark_locked(mark, group, inode, mnt, allow_dups);
	291	mutex_unlock(&group->mark_mutex);
	292	return ret;
	293	}
	294
5444e298	295	/*
4d92604c	296	* clear any marks in a group in which mark->flags & flags is true
5444e298	297	*/
4d92604c EP	298	void fsnotify_clear_marks_by_group_flags(struct fsnotify_group *group,
4d92604c EP	299	unsigned int flags)
5444e298 EP	300	{
5444e298 EP	301	struct fsnotify_mark lmark, mark;
5444e298	302
6960b0d9	303	mutex_lock_nested(&group->mark_mutex, SINGLE_DEPTH_NESTING);
5444e298	304	list_for_each_entry_safe(mark, lmark, &group->marks_list, g_list) {
4d92604c	305	if (mark->flags & flags) {
4d92604c	306	fsnotify_get_mark(mark);
64c20d2a LS	307	fsnotify_destroy_mark_locked(mark, group);
64c20d2a LS	308	fsnotify_put_mark(mark);
4d92604c	309	}
5444e298	310	}
986ab098	311	mutex_unlock(&group->mark_mutex);
5444e298 EP	312	}
5444e298 EP	313
4d92604c EP	314	/*
	315	* Given a group, destroy all of the marks associated with that group.
	316	*/
	317	void fsnotify_clear_marks_by_group(struct fsnotify_group *group)
	318	{
	319	fsnotify_clear_marks_by_group_flags(group, (unsigned int)-1);
	320	}
	321
5444e298 EP	322	void fsnotify_duplicate_mark(struct fsnotify_mark new, struct fsnotify_mark old)
	323	{
	324	assert_spin_locked(&old->lock);
	325	new->i.inode = old->i.inode;
	326	new->m.mnt = old->m.mnt;
23e964c2 LS	327	if (old->group)
23e964c2 LS	328	fsnotify_get_group(old->group);
5444e298 EP	329	new->group = old->group;
	330	new->mask = old->mask;
	331	new->free_mark = old->free_mark;
	332	}
	333
	334	/*
	335	* Nothing fancy, just initialize lists and locks and counters.
	336	*/
	337	void fsnotify_init_mark(struct fsnotify_mark *mark,
	338	void (free_mark)(struct fsnotify_mark mark))
	339	{
ba643f04	340	memset(mark, 0, sizeof(*mark));
5444e298 EP	341	spin_lock_init(&mark->lock);
5444e298 EP	342	atomic_set(&mark->refcnt, 1);
5444e298 EP	343	mark->free_mark = free_mark;
5444e298 EP	344	}
75c1be48 EP	345
	346	static int fsnotify_mark_destroy(void *ignored)
	347	{
	348	struct fsnotify_mark mark, next;
	349	LIST_HEAD(private_destroy_list);
	350
	351	for (;;) {
	352	spin_lock(&destroy_lock);
8778abb9 AG	353	/* exchange the list head */
8778abb9 AG	354	list_replace_init(&destroy_list, &private_destroy_list);
75c1be48 EP	355	spin_unlock(&destroy_lock);
	356
	357	synchronize_srcu(&fsnotify_mark_srcu);
	358
	359	list_for_each_entry_safe(mark, next, &private_destroy_list, destroy_list) {
	360	list_del_init(&mark->destroy_list);
	361	fsnotify_put_mark(mark);
	362	}
	363
	364	wait_event_interruptible(destroy_waitq, !list_empty(&destroy_list));
	365	}
	366
	367	return 0;
	368	}
	369
	370	static int __init fsnotify_mark_init(void)
	371	{
	372	struct task_struct *thread;
	373
	374	thread = kthread_run(fsnotify_mark_destroy, NULL,
	375	"fsnotify_mark");
	376	if (IS_ERR(thread))
	377	panic("unable to start fsnotify mark destruction thread.");
	378
	379	return 0;
	380	}
	381	device_initcall(fsnotify_mark_init);