dnotify: reimplement dnotify using fsnotify (3c5119c0) · Commits · e / devices / android_kernel_fairphone_FP4

MAINTAINERS

+3 −3

Original line number	Diff line number	Diff line
		@@ -1802,10 +1802,10 @@ F: drivers/char/epca*
		F: drivers/char/digi*

		DIRECTORY NOTIFICATION (DNOTIFY)
		P: Stephen Rothwell
		M: sfr@canb.auug.org.au
		P: Eric Paris
		M: eparis@parisplace.org
		L: linux-kernel@vger.kernel.org
		S: Supported
		S: Maintained
		F: Documentation/filesystems/dnotify.txt
		F: fs/notify/dnotify/
		F: include/linux/dnotify.h

fs/notify/dnotify/Kconfig

+1 −0

Original line number	Diff line number	Diff line
		config DNOTIFY
		bool "Dnotify support"
		depends on FSNOTIFY
		default y
		help
		Dnotify is a directory-based per-fd file change notification system

fs/notify/dnotify/dnotify.c

+362 −107

Original line number	Diff line number	Diff line
		@@ -3,6 +3,9 @@
		*
		* Copyright (C) 2000,2001,2002 Stephen Rothwell
		*
		* Copyright (C) 2009 Eric Paris <Red Hat Inc>
		* dnotify was largly rewritten to use the new fsnotify infrastructure
		*
		* 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
		@@ -21,24 +24,178 @@
		#include <linux/spinlock.h>
		#include <linux/slab.h>
		#include <linux/fdtable.h>
		#include <linux/fsnotify_backend.h>

		int dir_notify_enable __read_mostly = 1;

		static struct kmem_cache *dn_cache __read_mostly;
		static struct kmem_cache *dnotify_struct_cache __read_mostly;
		static struct kmem_cache *dnotify_mark_entry_cache __read_mostly;
		static struct fsnotify_group *dnotify_group __read_mostly;
		static DEFINE_MUTEX(dnotify_mark_mutex);

		/*
		* dnotify will attach one of these to each inode (i_fsnotify_mark_entries) which
		* is being watched by dnotify. If multiple userspace applications are watching
		* the same directory with dnotify their information is chained in dn
		*/
		struct dnotify_mark_entry {
		struct fsnotify_mark_entry fsn_entry;
		struct dnotify_struct *dn;
		};

		static void redo_inode_mask(struct inode *inode)
		/*
		* When a process starts or stops watching an inode the set of events which
		* dnotify cares about for that inode may change. This function runs the
		* list of everything receiving dnotify events about this directory and calculates
		* the set of all those events. After it updates what dnotify is interested in
		* it calls the fsnotify function so it can update the set of all events relevant
		* to this inode.
		*/
		static void dnotify_recalc_inode_mask(struct fsnotify_mark_entry *entry)
		{
		unsigned long new_mask;
		__u32 new_mask, old_mask;
		struct dnotify_struct *dn;
		struct dnotify_mark_entry *dnentry = container_of(entry,
		struct dnotify_mark_entry,
		fsn_entry);

		assert_spin_locked(&entry->lock);

		old_mask = entry->mask;
		new_mask = 0;
		for (dn = inode->i_dnotify; dn != NULL; dn = dn->dn_next)
		new_mask \|= dn->dn_mask & ~DN_MULTISHOT;
		inode->i_dnotify_mask = new_mask;
		for (dn = dnentry->dn; dn != NULL; dn = dn->dn_next)
		new_mask \|= (dn->dn_mask & ~FS_DN_MULTISHOT);
		entry->mask = new_mask;

		if (old_mask == new_mask)
		return;

		if (entry->inode)
		fsnotify_recalc_inode_mask(entry->inode);
		}

		/*
		* Mains fsnotify call where events are delivered to dnotify.
		* Find the dnotify mark on the relevant inode, run the list of dnotify structs
		* on that mark and determine which of them has expressed interest in receiving
		* events of this type. When found send the correct process and signal and
		* destroy the dnotify struct if it was not registered to receive multiple
		* events.
		*/
		static int dnotify_handle_event(struct fsnotify_group *group,
		struct fsnotify_event *event)
		{
		struct fsnotify_mark_entry *entry = NULL;
		struct dnotify_mark_entry *dnentry;
		struct inode *to_tell;
		struct dnotify_struct *dn;
		struct dnotify_struct **prev;
		struct fown_struct *fown;

		to_tell = event->to_tell;

		spin_lock(&to_tell->i_lock);
		entry = fsnotify_find_mark_entry(group, to_tell);
		spin_unlock(&to_tell->i_lock);

		/* unlikely since we alreay passed dnotify_should_send_event() */
		if (unlikely(!entry))
		return 0;
		dnentry = container_of(entry, struct dnotify_mark_entry, fsn_entry);

		spin_lock(&entry->lock);
		prev = &dnentry->dn;
		while ((dn = *prev) != NULL) {
		if ((dn->dn_mask & event->mask) == 0) {
		prev = &dn->dn_next;
		continue;
		}
		fown = &dn->dn_filp->f_owner;
		send_sigio(fown, dn->dn_fd, POLL_MSG);
		if (dn->dn_mask & FS_DN_MULTISHOT)
		prev = &dn->dn_next;
		else {
		*prev = dn->dn_next;
		kmem_cache_free(dnotify_struct_cache, dn);
		dnotify_recalc_inode_mask(entry);
		}
		}

		spin_unlock(&entry->lock);
		fsnotify_put_mark(entry);

		return 0;
		}

		/*
		* Given an inode and mask determine if dnotify would be interested in sending
		* userspace notification for that pair.
		*/
		static bool dnotify_should_send_event(struct fsnotify_group *group,
		struct inode *inode, __u32 mask)
		{
		struct fsnotify_mark_entry *entry;
		bool send;

		/* !dir_notify_enable should never get here, don't waste time checking
		if (!dir_notify_enable)
		return 0; */

		/* not a dir, dnotify doesn't care */
		if (!S_ISDIR(inode->i_mode))
		return false;

		spin_lock(&inode->i_lock);
		entry = fsnotify_find_mark_entry(group, inode);
		spin_unlock(&inode->i_lock);

		/* no mark means no dnotify watch */
		if (!entry)
		return false;

		spin_lock(&entry->lock);
		send = (mask & entry->mask) ? true : false;
		spin_unlock(&entry->lock);
		fsnotify_put_mark(entry); /* matches fsnotify_find_mark_entry */

		return send;
		}

		static void dnotify_freeing_mark(struct fsnotify_mark_entry *entry,
		struct fsnotify_group *group)
		{
		/* dnotify doesn't care than an inode is on the way out */
		}

		static void dnotify_free_mark(struct fsnotify_mark_entry *entry)
		{
		struct dnotify_mark_entry *dnentry = container_of(entry,
		struct dnotify_mark_entry,
		fsn_entry);

		BUG_ON(dnentry->dn);

		kmem_cache_free(dnotify_mark_entry_cache, dnentry);
		}

		static struct fsnotify_ops dnotify_fsnotify_ops = {
		.handle_event = dnotify_handle_event,
		.should_send_event = dnotify_should_send_event,
		.free_group_priv = NULL,
		.freeing_mark = dnotify_freeing_mark,
		};

		/*
		* Called every time a file is closed. Looks first for a dnotify mark on the
		* inode. If one is found run all of the ->dn entries attached to that
		* mark for one relevant to this process closing the file and remove that
		* dnotify_struct. If that was the last dnotify_struct also remove the
		* fsnotify_mark_entry.
		*/
		void dnotify_flush(struct file *filp, fl_owner_t id)
		{
		struct fsnotify_mark_entry *entry;
		struct dnotify_mark_entry *dnentry;
		struct dnotify_struct *dn;
		struct dnotify_struct **prev;
		struct inode *inode;
		@@ -46,145 +203,243 @@ void dnotify_flush(struct file *filp, fl_owner_t id)
		inode = filp->f_path.dentry->d_inode;
		if (!S_ISDIR(inode->i_mode))
		return;

		spin_lock(&inode->i_lock);
		prev = &inode->i_dnotify;
		entry = fsnotify_find_mark_entry(dnotify_group, inode);
		spin_unlock(&inode->i_lock);
		if (!entry)
		return;
		dnentry = container_of(entry, struct dnotify_mark_entry, fsn_entry);

		mutex_lock(&dnotify_mark_mutex);

		spin_lock(&entry->lock);
		prev = &dnentry->dn;
		while ((dn = *prev) != NULL) {
		if ((dn->dn_owner == id) && (dn->dn_filp == filp)) {
		*prev = dn->dn_next;
		redo_inode_mask(inode);
		kmem_cache_free(dn_cache, dn);
		kmem_cache_free(dnotify_struct_cache, dn);
		dnotify_recalc_inode_mask(entry);
		break;
		}
		prev = &dn->dn_next;
		}
		spin_unlock(&inode->i_lock);

		spin_unlock(&entry->lock);

		/* nothing else could have found us thanks to the dnotify_mark_mutex */
		if (dnentry->dn == NULL)
		fsnotify_destroy_mark_by_entry(entry);

		fsnotify_recalc_group_mask(dnotify_group);

		mutex_unlock(&dnotify_mark_mutex);

		fsnotify_put_mark(entry);
		}

		int fcntl_dirnotify(int fd, struct file *filp, unsigned long arg)
		/* this conversion is done only at watch creation */
		static __u32 convert_arg(unsigned long arg)
		{
		struct dnotify_struct *dn;
		struct dnotify_struct *odn;
		struct dnotify_struct **prev;
		struct inode *inode;
		fl_owner_t id = current->files;
		struct file *f;
		int error = 0;
		__u32 new_mask = FS_EVENT_ON_CHILD;

		if ((arg & ~DN_MULTISHOT) == 0) {
		dnotify_flush(filp, id);
		return 0;
		if (arg & DN_MULTISHOT)
		new_mask \|= FS_DN_MULTISHOT;
		if (arg & DN_DELETE)
		new_mask \|= (FS_DELETE \| FS_MOVED_FROM);
		if (arg & DN_MODIFY)
		new_mask \|= FS_MODIFY;
		if (arg & DN_ACCESS)
		new_mask \|= FS_ACCESS;
		if (arg & DN_ATTRIB)
		new_mask \|= FS_ATTRIB;
		if (arg & DN_RENAME)
		new_mask \|= FS_DN_RENAME;
		if (arg & DN_CREATE)
		new_mask \|= (FS_CREATE \| FS_MOVED_TO);

		return new_mask;
		}
		if (!dir_notify_enable)
		return -EINVAL;
		inode = filp->f_path.dentry->d_inode;
		if (!S_ISDIR(inode->i_mode))
		return -ENOTDIR;
		dn = kmem_cache_alloc(dn_cache, GFP_KERNEL);
		if (dn == NULL)
		return -ENOMEM;
		spin_lock(&inode->i_lock);
		prev = &inode->i_dnotify;
		while ((odn = *prev) != NULL) {

		/*
		* If multiple processes watch the same inode with dnotify there is only one
		* dnotify mark in inode->i_fsnotify_mark_entries but we chain a dnotify_struct
		* onto that mark. This function either attaches the new dnotify_struct onto
		* that list, or it \|= the mask onto an existing dnofiy_struct.
		*/
		static int attach_dn(struct dnotify_struct dn, struct dnotify_mark_entry dnentry,
		fl_owner_t id, int fd, struct file *filp, __u32 mask)
		{
		struct dnotify_struct *odn;

		odn = dnentry->dn;
		while (odn != NULL) {
		/* adding more events to existing dnofiy_struct? */
		if ((odn->dn_owner == id) && (odn->dn_filp == filp)) {
		odn->dn_fd = fd;
		odn->dn_mask \|= arg;
		inode->i_dnotify_mask \|= arg & ~DN_MULTISHOT;
		goto out_free;
		odn->dn_mask \|= mask;
		return -EEXIST;
		}
		prev = &odn->dn_next;
		odn = odn->dn_next;
		}

		rcu_read_lock();
		f = fcheck(fd);
		rcu_read_unlock();
		/* we'd lost the race with close(), sod off silently */
		/* note that inode->i_lock prevents reordering problems
		* between accesses to descriptor table and ->i_dnotify */
		if (f != filp)
		goto out_free;

		error = __f_setown(filp, task_pid(current), PIDTYPE_PID, 0);
		if (error)
		goto out_free;

		dn->dn_mask = arg;
		dn->dn_mask = mask;
		dn->dn_fd = fd;
		dn->dn_filp = filp;
		dn->dn_owner = id;
		inode->i_dnotify_mask \|= arg & ~DN_MULTISHOT;
		dn->dn_next = inode->i_dnotify;
		inode->i_dnotify = dn;
		spin_unlock(&inode->i_lock);
		return 0;
		dn->dn_next = dnentry->dn;
		dnentry->dn = dn;

		out_free:
		spin_unlock(&inode->i_lock);
		kmem_cache_free(dn_cache, dn);
		return error;
		return 0;
		}

		void __inode_dir_notify(struct inode *inode, unsigned long event)
		/*
		* When a process calls fcntl to attach a dnotify watch to a directory it ends
		* up here. Allocate both a mark for fsnotify to add and a dnotify_struct to be
		* attached to the fsnotify_mark.
		*/
		int fcntl_dirnotify(int fd, struct file *filp, unsigned long arg)
		{
		struct dnotify_mark_entry new_dnentry, dnentry;
		struct fsnotify_mark_entry new_entry, entry;
		struct dnotify_struct *dn;
		struct dnotify_struct **prev;
		struct fown_struct * fown;
		int changed = 0;
		struct inode *inode;
		fl_owner_t id = current->files;
		struct file *f;
		int destroy = 0, error = 0;
		__u32 mask;

		spin_lock(&inode->i_lock);
		prev = &inode->i_dnotify;
		while ((dn = *prev) != NULL) {
		if ((dn->dn_mask & event) == 0) {
		prev = &dn->dn_next;
		continue;
		/* we use these to tell if we need to kfree */
		new_entry = NULL;
		dn = NULL;

		if (!dir_notify_enable) {
		error = -EINVAL;
		goto out_err;
		}
		fown = &dn->dn_filp->f_owner;
		send_sigio(fown, dn->dn_fd, POLL_MSG);
		if (dn->dn_mask & DN_MULTISHOT)
		prev = &dn->dn_next;
		else {
		*prev = dn->dn_next;
		changed = 1;
		kmem_cache_free(dn_cache, dn);

		/* a 0 mask means we are explicitly removing the watch */
		if ((arg & ~DN_MULTISHOT) == 0) {
		dnotify_flush(filp, id);
		error = 0;
		goto out_err;
		}

		/* dnotify only works on directories */
		inode = filp->f_path.dentry->d_inode;
		if (!S_ISDIR(inode->i_mode)) {
		error = -ENOTDIR;
		goto out_err;
		}

		/* expect most fcntl to add new rather than augment old */
		dn = kmem_cache_alloc(dnotify_struct_cache, GFP_KERNEL);
		if (!dn) {
		error = -ENOMEM;
		goto out_err;
		}
		if (changed)
		redo_inode_mask(inode);
		spin_unlock(&inode->i_lock);

		/* new fsnotify mark, we expect most fcntl calls to add a new mark */
		new_dnentry = kmem_cache_alloc(dnotify_mark_entry_cache, GFP_KERNEL);
		if (!new_dnentry) {
		error = -ENOMEM;
		goto out_err;
		}

		EXPORT_SYMBOL(__inode_dir_notify);
		/* convert the userspace DN_* "arg" to the internal FS_* defines in fsnotify */
		mask = convert_arg(arg);

		/*
		* This is hopelessly wrong, but unfixable without API changes. At
		* least it doesn't oops the kernel...
		*
		* To safely access ->d_parent we need to keep d_move away from it. Use the
		* dentry's d_lock for this.
		*/
		void dnotify_parent(struct dentry *dentry, unsigned long event)
		{
		struct dentry *parent;
		/* set up the new_entry and new_dnentry */
		new_entry = &new_dnentry->fsn_entry;
		fsnotify_init_mark(new_entry, dnotify_free_mark);
		new_entry->mask = mask;
		new_dnentry->dn = NULL;

		if (!dir_notify_enable)
		return;
		/* this is needed to prevent the fcntl/close race described below */
		mutex_lock(&dnotify_mark_mutex);

		spin_lock(&dentry->d_lock);
		parent = dentry->d_parent;
		if (parent->d_inode->i_dnotify_mask & event) {
		dget(parent);
		spin_unlock(&dentry->d_lock);
		__inode_dir_notify(parent->d_inode, event);
		dput(parent);
		/* add the new_entry or find an old one. */
		spin_lock(&inode->i_lock);
		entry = fsnotify_find_mark_entry(dnotify_group, inode);
		spin_unlock(&inode->i_lock);
		if (entry) {
		dnentry = container_of(entry, struct dnotify_mark_entry, fsn_entry);
		spin_lock(&entry->lock);
		} else {
		spin_unlock(&dentry->d_lock);
		fsnotify_add_mark(new_entry, dnotify_group, inode);
		spin_lock(&new_entry->lock);
		entry = new_entry;
		dnentry = new_dnentry;
		/* we used new_entry, so don't free it */
		new_entry = NULL;
		}

		rcu_read_lock();
		f = fcheck(fd);
		rcu_read_unlock();

		/* if (f != filp) means that we lost a race and another task/thread
		* actually closed the fd we are still playing with before we grabbed
		* the dnotify_mark_mutex and entry->lock. Since closing the fd is the
		* only time we clean up the mark entries we need to get our mark off
		* the list. */
		if (f != filp) {
		/* if we added ourselves, shoot ourselves, it's possible that
		* the flush actually did shoot this entry. That's fine too
		* since multiple calls to destroy_mark is perfectly safe, if
		* we found a dnentry already attached to the inode, just sod
		* off silently as the flush at close time dealt with it.
		*/
		if (dnentry == new_dnentry)
		destroy = 1;
		goto out;
		}

		error = __f_setown(filp, task_pid(current), PIDTYPE_PID, 0);
		if (error) {
		/* if we added, we must shoot */
		if (dnentry == new_dnentry)
		destroy = 1;
		goto out;
		}

		error = attach_dn(dn, dnentry, id, fd, filp, mask);
		/* !error means that we attached the dn to the dnentry, so don't free it */
		if (!error)
		dn = NULL;
		/* -EEXIST means that we didn't add this new dn and used an old one.
		* that isn't an error (and the unused dn should be freed) */
		else if (error == -EEXIST)
		error = 0;

		dnotify_recalc_inode_mask(entry);
		out:
		spin_unlock(&entry->lock);

		if (destroy)
		fsnotify_destroy_mark_by_entry(entry);

		fsnotify_recalc_group_mask(dnotify_group);

		mutex_unlock(&dnotify_mark_mutex);
		fsnotify_put_mark(entry);
		out_err:
		if (new_entry)
		fsnotify_put_mark(new_entry);
		if (dn)
		kmem_cache_free(dnotify_struct_cache, dn);
		return error;
		}
		EXPORT_SYMBOL_GPL(dnotify_parent);

		static int __init dnotify_init(void)
		{
		dn_cache = kmem_cache_create("dnotify_cache",
		sizeof(struct dnotify_struct), 0, SLAB_PANIC, NULL);
		dnotify_struct_cache = KMEM_CACHE(dnotify_struct, SLAB_PANIC);
		dnotify_mark_entry_cache = KMEM_CACHE(dnotify_mark_entry, SLAB_PANIC);

		dnotify_group = fsnotify_obtain_group(DNOTIFY_GROUP_NUM,
		0, &dnotify_fsnotify_ops);
		if (IS_ERR(dnotify_group))
		panic("unable to allocate fsnotify group for dnotify\n");
		return 0;
		}

include/linux/dnotify.h

+8 −21

Original line number	Diff line number	Diff line
		@@ -10,7 +10,7 @@

		struct dnotify_struct {
		struct dnotify_struct * dn_next;
		unsigned long dn_mask;
		__u32 dn_mask;
		int dn_fd;
		struct file * dn_filp;
		fl_owner_t dn_owner;
		@@ -21,23 +21,18 @@ struct dnotify_struct {

		#ifdef CONFIG_DNOTIFY

		extern void __inode_dir_notify(struct inode *, unsigned long);
		#define DNOTIFY_ALL_EVENTS (FS_DELETE \| FS_DELETE_CHILD \|\
		FS_MODIFY \| FS_MODIFY_CHILD \|\
		FS_ACCESS \| FS_ACCESS_CHILD \|\
		FS_ATTRIB \| FS_ATTRIB_CHILD \|\
		FS_CREATE \| FS_DN_RENAME \|\
		FS_MOVED_FROM \| FS_MOVED_TO)

		extern void dnotify_flush(struct file *, fl_owner_t);
		extern int fcntl_dirnotify(int, struct file *, unsigned long);
		extern void dnotify_parent(struct dentry *, unsigned long);

		static inline void inode_dir_notify(struct inode *inode, unsigned long event)
		{
		if (inode->i_dnotify_mask & (event))
		__inode_dir_notify(inode, event);
		}

		#else

		static inline void __inode_dir_notify(struct inode *inode, unsigned long event)
		{
		}

		static inline void dnotify_flush(struct file *filp, fl_owner_t id)
		{
		}
		@@ -47,14 +42,6 @@ static inline int fcntl_dirnotify(int fd, struct file *filp, unsigned long arg)
		return -EINVAL;
		}

		static inline void dnotify_parent(struct dentry *dentry, unsigned long event)
		{
		}

		static inline void inode_dir_notify(struct inode *inode, unsigned long event)
		{
		}

		#endif /* CONFIG_DNOTIFY */

		#endif /* __KERNEL __ */

include/linux/fs.h

+0 −5

Original line number	Diff line number	Diff line
		@@ -760,11 +760,6 @@ struct inode {
		struct hlist_head i_fsnotify_mark_entries; /* fsnotify mark entries */
		#endif

		#ifdef CONFIG_DNOTIFY
		unsigned long i_dnotify_mask; /* Directory notify events */
		struct dnotify_struct i_dnotify; / for directory notifications */
		#endif

		#ifdef CONFIG_INOTIFY
		struct list_head inotify_watches; /* watches on this inode */
		struct mutex inotify_mutex; /* protects the watches list */