[PATCH -v1 11/11] inotify: reimplement inotify using fsnotify

From: Eric Paris
Date: Mon Feb 09 2009 - 16:20:07 EST


Yes, holy shit, I'm trying to reimplement inotify as fsnotify...

Signed-off-by: Eric Paris <eparis@xxxxxxxxxx>
---

MAINTAINERS | 2
fs/notify/inotify/Kconfig | 20 +
fs/notify/inotify/Makefile | 2
fs/notify/inotify/inotify.h | 108 ++++++
fs/notify/inotify/inotify_fsnotify.c | 176 ++++++++++
fs/notify/inotify/inotify_kernel.c | 236 +++++++++++++
fs/notify/inotify/inotify_user.c | 598 +++++++++-------------------------
fs/notify/notification.c | 30 +-
include/linux/fsnotify.h | 39 +-
include/linux/fsnotify_backend.h | 11 +
10 files changed, 744 insertions(+), 478 deletions(-)
create mode 100644 fs/notify/inotify/inotify.h
create mode 100644 fs/notify/inotify/inotify_fsnotify.c
create mode 100644 fs/notify/inotify/inotify_kernel.c

diff --git a/MAINTAINERS b/MAINTAINERS
index daba68f..0bc3f24 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2243,6 +2243,8 @@ P: John McCutchan
M: john@xxxxxxxxxxxxxxxxx
P: Robert Love
M: rlove@xxxxxxxxx
+P: Eric Paris
+M: eparis@xxxxxxxxxxxxxx
L: linux-kernel@xxxxxxxxxxxxxxx
S: Maintained

diff --git a/fs/notify/inotify/Kconfig b/fs/notify/inotify/Kconfig
index 4467928..5356884 100644
--- a/fs/notify/inotify/Kconfig
+++ b/fs/notify/inotify/Kconfig
@@ -1,26 +1,30 @@
config INOTIFY
bool "Inotify file change notification support"
- default y
+ default n
---help---
- Say Y here to enable inotify support. Inotify is a file change
- notification system and a replacement for dnotify. Inotify fixes
- numerous shortcomings in dnotify and introduces several new features
- including multiple file events, one-shot support, and unmount
- notification.
+ Say Y here to enable legacy in kernel inotify support. Inotify is a
+ file change notification system. It is a replacement for dnotify.
+ This option only provides the legacy inotify in kernel API. There
+ are no in tree kernel users of this interface since it is deprecated.
+ You only need this if you are loading an out of tree kernel module
+ that uses inotify.

For more information, see <file:Documentation/filesystems/inotify.txt>

- If unsure, say Y.
+ If unsure, say N.

config INOTIFY_USER
bool "Inotify support for userspace"
- depends on INOTIFY
+ depends on FSNOTIFY
default y
---help---
Say Y here to enable inotify support for userspace, including the
associated system calls. Inotify allows monitoring of both files and
directories via a single open fd. Events are read from the file
descriptor, which is also select()- and poll()-able.
+ Inotify fixes numerous shortcomings in dnotify and introduces several
+ new features including multiple file events, one-shot support, and
+ unmount notification.

For more information, see <file:Documentation/filesystems/inotify.txt>

diff --git a/fs/notify/inotify/Makefile b/fs/notify/inotify/Makefile
index e290f3b..aff7f68 100644
--- a/fs/notify/inotify/Makefile
+++ b/fs/notify/inotify/Makefile
@@ -1,2 +1,2 @@
obj-$(CONFIG_INOTIFY) += inotify.o
-obj-$(CONFIG_INOTIFY_USER) += inotify_user.o
+obj-$(CONFIG_INOTIFY_USER) += inotify_fsnotify.o inotify_kernel.o inotify_user.o
diff --git a/fs/notify/inotify/inotify.h b/fs/notify/inotify/inotify.h
new file mode 100644
index 0000000..e37d24c
--- /dev/null
+++ b/fs/notify/inotify/inotify.h
@@ -0,0 +1,108 @@
+/*
+ * fs/inotify_user.c - inotify support for userspace
+ *
+ * Authors:
+ * John McCutchan <ttb@xxxxxxxxxxxxxxxx>
+ * Robert Love <rml@xxxxxxxxxx>
+ *
+ * Copyright (C) 2005 John McCutchan
+ * Copyright 2006 Hewlett-Packard Development Company, L.P.
+ *
+ * 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.
+ */
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/fs.h>
+#include <linux/file.h>
+#include <linux/fsnotify_backend.h>
+#include <linux/limits.h>
+#include <linux/module.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+#include <linux/poll.h>
+#include <linux/init.h>
+#include <linux/list.h>
+#include <linux/inotify.h>
+#include <linux/syscalls.h>
+#include <linux/string.h>
+#include <linux/magic.h>
+#include <linux/writeback.h>
+#include <linux/fsnotify.h>
+
+#include <asm/ioctls.h>
+
+extern struct kmem_cache *inotify_inode_mark_cachep;
+extern struct kmem_cache *event_priv_cachep;
+extern int inotify_max_user_watches;
+
+struct inotify_event_private_data {
+ struct fsnotify_event_private_data fsnotify_event_priv_data;
+ int wd;
+};
+
+struct inotify_inode_mark_entry {
+ /* fsnotify_mark_entry MUST be the first thing */
+ struct fsnotify_mark_entry fsn_entry;
+ int wd;
+};
+
+static inline __u64 inotify_arg_to_mask(u32 arg)
+{
+ /* everything should accept their own ignored */
+ __u64 mask = FS_IN_IGNORED;
+
+ BUILD_BUG_ON(IN_ACCESS != FS_ACCESS);
+ BUILD_BUG_ON(IN_MODIFY != FS_MODIFY);
+ BUILD_BUG_ON(IN_ATTRIB != FS_ATTRIB);
+ BUILD_BUG_ON(IN_CLOSE_WRITE != FS_CLOSE_WRITE);
+ BUILD_BUG_ON(IN_CLOSE_NOWRITE != FS_CLOSE_NOWRITE);
+ BUILD_BUG_ON(IN_OPEN != FS_OPEN);
+ BUILD_BUG_ON(IN_MOVED_FROM != FS_MOVED_FROM);
+ BUILD_BUG_ON(IN_MOVED_TO != FS_MOVED_TO);
+ BUILD_BUG_ON(IN_CREATE != FS_CREATE);
+ BUILD_BUG_ON(IN_DELETE != FS_DELETE);
+ BUILD_BUG_ON(IN_DELETE_SELF != FS_DELETE_SELF);
+ BUILD_BUG_ON(IN_MOVE_SELF != FS_MOVE_SELF);
+ BUILD_BUG_ON(IN_Q_OVERFLOW != FS_Q_OVERFLOW);
+
+ BUILD_BUG_ON(IN_UNMOUNT != FS_UNMOUNT);
+ BUILD_BUG_ON(IN_ISDIR != FS_IN_ISDIR);
+ BUILD_BUG_ON(IN_IGNORED != FS_IN_IGNORED);
+ BUILD_BUG_ON(IN_ONESHOT != FS_IN_ONESHOT);
+
+ mask |= (arg & (IN_ALL_EVENTS | IN_ONESHOT));
+
+ mask |= ((mask & FS_EVENTS_WITH_CHILD) << 32);
+
+ return mask;
+}
+
+static inline u32 inotify_mask_to_arg(__u64 mask)
+{
+ u32 arg;
+
+ arg = (mask & (IN_ALL_EVENTS | IN_ISDIR | IN_UNMOUNT | IN_IGNORED | IN_Q_OVERFLOW));
+
+ arg |= ((mask >> 32) & FS_EVENTS_WITH_CHILD);
+
+ return arg;
+}
+
+
+int find_inode(const char __user *dirname, struct path *path, unsigned flags);
+void inotify_destroy_mark_entry(struct fsnotify_mark_entry *entry);
+int inotify_update_watch(struct fsnotify_group *group, struct inode *inode, u32 arg);
+struct fsnotify_group *inotify_new_group(struct user_struct *user, unsigned int max_events);
+void __inotify_free_event_priv(struct inotify_event_private_data *event_priv);
+
+extern const struct fsnotify_ops inotify_fsnotify_ops;
diff --git a/fs/notify/inotify/inotify_fsnotify.c b/fs/notify/inotify/inotify_fsnotify.c
new file mode 100644
index 0000000..90dca51
--- /dev/null
+++ b/fs/notify/inotify/inotify_fsnotify.c
@@ -0,0 +1,176 @@
+/*
+ * fs/inotify_user.c - inotify support for userspace
+ *
+ * Authors:
+ * John McCutchan <ttb@xxxxxxxxxxxxxxxx>
+ * Robert Love <rml@xxxxxxxxxx>
+ *
+ * Copyright (C) 2005 John McCutchan
+ * Copyright 2006 Hewlett-Packard Development Company, L.P.
+ *
+ * Copyright (C) 2009 Eric Paris <Red Hat Inc>
+ * inotify was largely rewriten to make use of the 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
+ * 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.
+ */
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/fs.h>
+#include <linux/file.h>
+#include <linux/limits.h>
+#include <linux/module.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+#include <linux/poll.h>
+#include <linux/idr.h>
+#include <linux/init.h>
+#include <linux/inotify.h>
+#include <linux/list.h>
+#include <linux/syscalls.h>
+#include <linux/string.h>
+#include <linux/magic.h>
+#include <linux/writeback.h>
+
+#include "inotify.h"
+
+#include <asm/ioctls.h>
+
+static int inotify_handle_event(struct fsnotify_group *group, struct fsnotify_event *event)
+{
+ struct fsnotify_mark_entry *entry;
+ struct inotify_inode_mark_entry *ientry;
+ struct inode *to_tell;
+ struct inotify_event_private_data *event_priv;
+ int wd;
+
+ 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);
+ /* race with watch removal? */
+ if (!entry)
+ return 0;
+ ientry = (struct inotify_inode_mark_entry *)entry;
+
+ wd = ientry->wd;
+
+ fsnotify_put_mark(entry);
+
+ event_priv = kmem_cache_alloc(event_priv_cachep, GFP_KERNEL);
+ if (unlikely(!event_priv))
+ return -ENOMEM;
+
+ event_priv->fsnotify_event_priv_data.group = group;
+ event_priv->wd = wd;
+
+ return fsnotify_add_notif_event(group, event, (struct fsnotify_event_private_data *)event_priv);
+}
+
+static void inotify_mark_clear_inode(struct fsnotify_mark_entry *entry, struct inode *inode, unsigned int flags)
+{
+ if (unlikely((flags != FSNOTIFY_LAST_DENTRY) && (flags != FSNOTIFY_INODE_DESTROY))) {
+ BUG();
+ return;
+ }
+
+ /*
+ * so no matter what we need to put this entry back on the inode's list.
+ * we need it there so fsnotify can find it to send the ignore message.
+ *
+ * I didn't realize how brilliant this was until I did it. Our caller
+ * blanked the inode->i_fsnotify_mark_entries list so we will be the
+ * only mark on the list when fsnotify runs so only our group will get
+ * this FS_IN_IGNORED.
+ *
+ * Bloody brilliant.
+ */
+ spin_lock(&inode->i_lock);
+ list_add(&entry->i_list, &inode->i_fsnotify_mark_entries);
+ spin_unlock(&inode->i_lock);
+
+ fsnotify(inode, FS_IN_IGNORED, inode, FSNOTIFY_EVENT_INODE, NULL, 0);
+ inotify_destroy_mark_entry(entry);
+}
+
+static int inotify_should_send_event(struct fsnotify_group *group, struct inode *inode, __u64 mask)
+{
+ struct fsnotify_mark_entry *entry;
+ int send;
+
+ spin_lock(&inode->i_lock);
+ entry = fsnotify_find_mark_entry(group, inode);
+ spin_unlock(&inode->i_lock);
+ if (!entry)
+ return 0;
+
+ spin_lock(&entry->lock);
+ send = !!(entry->mask & mask);
+ spin_unlock(&entry->lock);
+
+ /* find took a reference */
+ fsnotify_put_mark(entry);
+
+ return send;
+}
+
+static int idr_callback(int id, void *p, void *data)
+{
+ BUG();
+ return 0;
+}
+
+static void inotify_free_group_priv(struct fsnotify_group *group)
+{
+ /* ideally the idr is empty and we won't hit the BUG in teh callback */
+ idr_for_each(&group->inotify_data.idr, idr_callback, NULL);
+ idr_remove_all(&group->inotify_data.idr);
+ idr_destroy(&group->inotify_data.idr);
+}
+
+void __inotify_free_event_priv(struct inotify_event_private_data *event_priv)
+{
+ list_del_init(&event_priv->fsnotify_event_priv_data.event_list);
+ kmem_cache_free(event_priv_cachep, event_priv);
+}
+
+static void inotify_free_event_priv(struct fsnotify_group *group, struct fsnotify_event *event)
+{
+ struct inotify_event_private_data *event_priv;
+
+ spin_lock(&event->lock);
+
+ event_priv = (struct inotify_event_private_data *)fsnotify_get_priv_from_event(group, event);
+ BUG_ON(!event_priv);
+
+ __inotify_free_event_priv(event_priv);
+
+ spin_unlock(&event->lock);
+}
+
+/* ding dong the mark is dead */
+static void inotify_free_mark(struct fsnotify_mark_entry *entry)
+{
+ struct inotify_inode_mark_entry *ientry = (struct inotify_inode_mark_entry *)entry;
+
+ kmem_cache_free(inotify_inode_mark_cachep, ientry);
+}
+
+const struct fsnotify_ops inotify_fsnotify_ops = {
+ .handle_event = inotify_handle_event,
+ .mark_clear_inode = inotify_mark_clear_inode,
+ .should_send_event = inotify_should_send_event,
+ .free_group_priv = inotify_free_group_priv,
+ .free_event_priv = inotify_free_event_priv,
+ .free_mark = inotify_free_mark,
+};
diff --git a/fs/notify/inotify/inotify_kernel.c b/fs/notify/inotify/inotify_kernel.c
new file mode 100644
index 0000000..1d8673b
--- /dev/null
+++ b/fs/notify/inotify/inotify_kernel.c
@@ -0,0 +1,236 @@
+/*
+ * fs/inotify_user.c - inotify support for userspace
+ *
+ * Authors:
+ * John McCutchan <ttb@xxxxxxxxxxxxxxxx>
+ * Robert Love <rml@xxxxxxxxxx>
+ *
+ * Copyright (C) 2005 John McCutchan
+ * Copyright 2006 Hewlett-Packard Development Company, L.P.
+ *
+ * Copyright (C) 2009 Eric Paris <Red Hat Inc>
+ * inotify was largely rewriten to make use of the 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
+ * 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.
+ */
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/fs.h>
+#include <linux/file.h>
+#include <linux/limits.h>
+#include <linux/module.h>
+#include <linux/mount.h>
+#include <linux/namei.h>
+#include <linux/poll.h>
+#include <linux/idr.h>
+#include <linux/init.h>
+#include <linux/inotify.h>
+#include <linux/list.h>
+#include <linux/syscalls.h>
+#include <linux/string.h>
+#include <linux/magic.h>
+#include <linux/writeback.h>
+
+#include "inotify.h"
+
+#include <asm/ioctls.h>
+
+struct kmem_cache *inotify_inode_mark_cachep __read_mostly;
+struct kmem_cache *event_priv_cachep __read_mostly;
+
+atomic_t inotify_grp_num;
+
+/*
+ * find_inode - resolve a user-given path to a specific inode
+ */
+int find_inode(const char __user *dirname, struct path *path, unsigned flags)
+{
+ int error;
+
+ error = user_path_at(AT_FDCWD, dirname, flags, path);
+ if (error)
+ return error;
+ /* you can only watch an inode if you have read permissions on it */
+ error = inode_permission(path->dentry->d_inode, MAY_READ);
+ if (error)
+ path_put(path);
+ return error;
+}
+
+void inotify_destroy_mark_entry(struct fsnotify_mark_entry *entry)
+{
+ struct inotify_inode_mark_entry *ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry);
+ struct fsnotify_group *group;
+ struct idr *idr;
+
+ spin_lock(&entry->lock);
+
+ group = entry->group;
+ if (!group) {
+ /* racing with group tear down, let it do it */
+ spin_unlock(&entry->lock);
+ return;
+ }
+
+ /* group can't disappear on us, since it has to shoot this entry to disappear
+ * and it's going to block on entry->lock */
+
+ /* remove this entry from the idr */
+ idr = &group->inotify_data.idr;
+ spin_lock(&group->inotify_data.idr_lock);
+ idr_remove(idr, ientry->wd);
+ spin_unlock(&group->inotify_data.idr_lock);
+
+ spin_unlock(&entry->lock);
+
+ /* mark the entry to die */
+ fsnotify_destroy_mark_by_entry(entry);
+
+ /* removed from idr, drop that reference */
+ fsnotify_put_mark(entry);
+}
+
+int inotify_update_watch(struct fsnotify_group *group, struct inode *inode, u32 arg)
+{
+ struct fsnotify_mark_entry *entry = NULL;
+ struct inotify_inode_mark_entry *ientry;
+ int ret = 0;
+ int add = (arg & IN_MASK_ADD);
+ __u64 mask;
+ __u64 old_mask, new_mask;
+
+ /* don't allow invalid bits: we don't want flags set */
+ mask = inotify_arg_to_mask(arg);
+ if (unlikely(!mask))
+ return -EINVAL;
+
+ ientry = kmem_cache_alloc(inotify_inode_mark_cachep, GFP_KERNEL);
+ if (unlikely(!ientry))
+ return -ENOMEM;
+ /* we set the mask at the end after attaching it */
+ fsnotify_init_mark(&ientry->fsn_entry, group, inode, 0);
+ ientry->wd = 0;
+
+find_entry:
+ entry = fsnotify_find_mark_entry(group, inode);
+ if (entry) {
+ kmem_cache_free(inotify_inode_mark_cachep, ientry);
+ ientry = container_of(entry, struct inotify_inode_mark_entry, fsn_entry);
+ } else {
+ if (atomic_read(&group->inotify_data.user->inotify_watches) >= inotify_max_user_watches) {
+ ret = -ENOSPC;
+ goto out_err;
+ }
+
+ ret = fsnotify_add_mark(&ientry->fsn_entry);
+ if (ret == -EEXIST)
+ goto find_entry;
+ else if (ret)
+ goto out_err;
+
+ entry = &ientry->fsn_entry;
+retry:
+ ret = -ENOMEM;
+ if (unlikely(!idr_pre_get(&group->inotify_data.idr, GFP_KERNEL)))
+ goto out_err;
+
+ spin_lock(&group->inotify_data.idr_lock);
+ /* if entry is added to the idr we keep the reference obtained
+ * through fsnotify_mark_add. remember to drop this reference
+ * when entry is removed from idr */
+ ret = idr_get_new_above(&group->inotify_data.idr, entry,
+ ++group->inotify_data.last_wd,
+ &ientry->wd);
+ spin_unlock(&group->inotify_data.idr_lock);
+ if (ret) {
+ if (ret == -EAGAIN)
+ goto retry;
+ goto out_err;
+ }
+ atomic_inc(&group->inotify_data.user->inotify_watches);
+ }
+
+ spin_lock(&entry->lock);
+
+ old_mask = entry->mask;
+ if (add) {
+ entry->mask |= mask;
+ new_mask = entry->mask;
+ } else {
+ entry->mask = mask;
+ new_mask = entry->mask;
+ }
+
+ spin_unlock(&entry->lock);
+
+ if (old_mask != new_mask) {
+ /* more bits in old than in new? */
+ int dropped = (old_mask & ~new_mask);
+ /* more bits in this entry than the inode's mask? */
+ int do_inode = (new_mask & ~inode->i_fsnotify_mask);
+ /* more bits in this entry than the group? */
+ int do_group = (new_mask & ~group->mask);
+
+ /* update the inode with this new entry */
+ if (dropped || do_inode)
+ fsnotify_recalc_inode_mask(inode);
+
+ /* update the group mask with the new mask */
+ if (dropped || do_group)
+ fsnotify_recalc_group_mask(group);
+ }
+
+ return ientry->wd;
+
+out_err:
+ /* see this isn't supposed to happen, just kill the watch */
+ if (entry) {
+ fsnotify_destroy_mark_by_entry(entry);
+ fsnotify_put_mark(entry);
+ }
+ return ret;
+}
+
+struct fsnotify_group *inotify_new_group(struct user_struct *user, unsigned int max_events)
+{
+ struct fsnotify_group *group;
+ unsigned int grp_num;
+
+ /* fsnotify_obtain_group took a reference to group, we put this when we kill the file in the end */
+ grp_num = (INOTIFY_GROUP_NUM - atomic_inc_return(&inotify_grp_num));
+ group = fsnotify_obtain_group(grp_num, grp_num, 0, &inotify_fsnotify_ops);
+ if (IS_ERR(group))
+ return group;
+
+ group->max_events = max_events;
+
+ spin_lock_init(&group->inotify_data.idr_lock);
+ idr_init(&group->inotify_data.idr);
+ group->inotify_data.last_wd = 0;
+ group->inotify_data.user = user;
+ group->inotify_data.fa = NULL;
+
+ return group;
+}
+
+static int __init inotify_kernel_setup(void)
+{
+ inotify_inode_mark_cachep = kmem_cache_create("inotify_mark_entry",
+ sizeof(struct inotify_inode_mark_entry),
+ 0, SLAB_PANIC, NULL);
+ event_priv_cachep = kmem_cache_create("inotify_event_priv_cache",
+ sizeof(struct inotify_event_private_data),
+ 0, SLAB_PANIC, NULL);
+ return 0;
+}
+subsys_initcall(inotify_kernel_setup);
diff --git a/fs/notify/inotify/inotify_user.c b/fs/notify/inotify/inotify_user.c
index bed766e..44b9844 100644
--- a/fs/notify/inotify/inotify_user.c
+++ b/fs/notify/inotify/inotify_user.c
@@ -8,6 +8,9 @@
* Copyright (C) 2005 John McCutchan
* Copyright 2006 Hewlett-Packard Development Company, L.P.
*
+ * Copyright (C) 2009 Eric Paris <Red Hat Inc>
+ * inotify was largely rewriten to make use of the 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
@@ -24,89 +27,32 @@
#include <linux/slab.h>
#include <linux/fs.h>
#include <linux/file.h>
+#include <linux/limits.h>
+#include <linux/module.h>
#include <linux/mount.h>
#include <linux/namei.h>
#include <linux/poll.h>
#include <linux/init.h>
-#include <linux/list.h>
#include <linux/inotify.h>
+#include <linux/list.h>
#include <linux/syscalls.h>
+#include <linux/string.h>
#include <linux/magic.h>
+#include <linux/writeback.h>

-#include <asm/ioctls.h>
+#include "inotify.h"

-static struct kmem_cache *watch_cachep __read_mostly;
-static struct kmem_cache *event_cachep __read_mostly;
+#include <asm/ioctls.h>

static struct vfsmount *inotify_mnt __read_mostly;

+/* this just sits here and wastes global memory. used to just pad userspace messages with zeros */
+static struct inotify_event nul_inotify_event;
+
/* these are configurable via /proc/sys/fs/inotify/ */
static int inotify_max_user_instances __read_mostly;
-static int inotify_max_user_watches __read_mostly;
static int inotify_max_queued_events __read_mostly;
-
-/*
- * Lock ordering:
- *
- * inotify_dev->up_mutex (ensures we don't re-add the same watch)
- * inode->inotify_mutex (protects inode's watch list)
- * inotify_handle->mutex (protects inotify_handle's watch list)
- * inotify_dev->ev_mutex (protects device's event queue)
- */
-
-/*
- * Lifetimes of the main data structures:
- *
- * inotify_device: Lifetime is managed by reference count, from
- * sys_inotify_init() until release. Additional references can bump the count
- * via get_inotify_dev() and drop the count via put_inotify_dev().
- *
- * inotify_user_watch: Lifetime is from create_watch() to the receipt of an
- * IN_IGNORED event from inotify, or when using IN_ONESHOT, to receipt of the
- * first event, or to inotify_destroy().
- */
-
-/*
- * struct inotify_device - represents an inotify instance
- *
- * This structure is protected by the mutex 'mutex'.
- */
-struct inotify_device {
- wait_queue_head_t wq; /* wait queue for i/o */
- struct mutex ev_mutex; /* protects event queue */
- struct mutex up_mutex; /* synchronizes watch updates */
- struct list_head events; /* list of queued events */
- struct user_struct *user; /* user who opened this dev */
- struct inotify_handle *ih; /* inotify handle */
- struct fasync_struct *fa; /* async notification */
- atomic_t count; /* reference count */
- unsigned int queue_size; /* size of the queue (bytes) */
- unsigned int event_count; /* number of pending events */
- unsigned int max_events; /* maximum number of events */
-};
-
-/*
- * struct inotify_kernel_event - An inotify event, originating from a watch and
- * queued for user-space. A list of these is attached to each instance of the
- * device. In read(), this list is walked and all events that can fit in the
- * buffer are returned.
- *
- * Protected by dev->ev_mutex of the device in which we are queued.
- */
-struct inotify_kernel_event {
- struct inotify_event event; /* the user-space event */
- struct list_head list; /* entry in inotify_device's list */
- char *name; /* filename, if any */
-};
-
-/*
- * struct inotify_user_watch - our version of an inotify_watch, we add
- * a reference to the associated inotify_device.
- */
-struct inotify_user_watch {
- struct inotify_device *dev; /* associated device */
- struct inotify_watch wdata; /* inotify watch data */
-};
+int inotify_max_user_watches __read_mostly;

#ifdef CONFIG_SYSCTL

@@ -149,280 +95,17 @@ ctl_table inotify_table[] = {
};
#endif /* CONFIG_SYSCTL */

-static inline void get_inotify_dev(struct inotify_device *dev)
-{
- atomic_inc(&dev->count);
-}
-
-static inline void put_inotify_dev(struct inotify_device *dev)
-{
- if (atomic_dec_and_test(&dev->count)) {
- atomic_dec(&dev->user->inotify_devs);
- free_uid(dev->user);
- kfree(dev);
- }
-}
-
-/*
- * free_inotify_user_watch - cleans up the watch and its references
- */
-static void free_inotify_user_watch(struct inotify_watch *w)
-{
- struct inotify_user_watch *watch;
- struct inotify_device *dev;
-
- watch = container_of(w, struct inotify_user_watch, wdata);
- dev = watch->dev;
-
- atomic_dec(&dev->user->inotify_watches);
- put_inotify_dev(dev);
- kmem_cache_free(watch_cachep, watch);
-}
-
-/*
- * kernel_event - create a new kernel event with the given parameters
- *
- * This function can sleep.
- */
-static struct inotify_kernel_event * kernel_event(s32 wd, u32 mask, u32 cookie,
- const char *name)
-{
- struct inotify_kernel_event *kevent;
-
- kevent = kmem_cache_alloc(event_cachep, GFP_NOFS);
- if (unlikely(!kevent))
- return NULL;
-
- /* we hand this out to user-space, so zero it just in case */
- memset(&kevent->event, 0, sizeof(struct inotify_event));
-
- kevent->event.wd = wd;
- kevent->event.mask = mask;
- kevent->event.cookie = cookie;
-
- INIT_LIST_HEAD(&kevent->list);
-
- if (name) {
- size_t len, rem, event_size = sizeof(struct inotify_event);
-
- /*
- * We need to pad the filename so as to properly align an
- * array of inotify_event structures. Because the structure is
- * small and the common case is a small filename, we just round
- * up to the next multiple of the structure's sizeof. This is
- * simple and safe for all architectures.
- */
- len = strlen(name) + 1;
- rem = event_size - len;
- if (len > event_size) {
- rem = event_size - (len % event_size);
- if (len % event_size == 0)
- rem = 0;
- }
-
- kevent->name = kmalloc(len + rem, GFP_KERNEL);
- if (unlikely(!kevent->name)) {
- kmem_cache_free(event_cachep, kevent);
- return NULL;
- }
- memcpy(kevent->name, name, len);
- if (rem)
- memset(kevent->name + len, 0, rem);
- kevent->event.len = len + rem;
- } else {
- kevent->event.len = 0;
- kevent->name = NULL;
- }
-
- return kevent;
-}
-
-/*
- * inotify_dev_get_event - return the next event in the given dev's queue
- *
- * Caller must hold dev->ev_mutex.
- */
-static inline struct inotify_kernel_event *
-inotify_dev_get_event(struct inotify_device *dev)
-{
- return list_entry(dev->events.next, struct inotify_kernel_event, list);
-}
-
-/*
- * inotify_dev_get_last_event - return the last event in the given dev's queue
- *
- * Caller must hold dev->ev_mutex.
- */
-static inline struct inotify_kernel_event *
-inotify_dev_get_last_event(struct inotify_device *dev)
-{
- if (list_empty(&dev->events))
- return NULL;
- return list_entry(dev->events.prev, struct inotify_kernel_event, list);
-}
-
-/*
- * inotify_dev_queue_event - event handler registered with core inotify, adds
- * a new event to the given device
- *
- * Can sleep (calls kernel_event()).
- */
-static void inotify_dev_queue_event(struct inotify_watch *w, u32 wd, u32 mask,
- u32 cookie, const char *name,
- struct inode *ignored)
-{
- struct inotify_user_watch *watch;
- struct inotify_device *dev;
- struct inotify_kernel_event *kevent, *last;
-
- watch = container_of(w, struct inotify_user_watch, wdata);
- dev = watch->dev;
-
- mutex_lock(&dev->ev_mutex);
-
- /* we can safely put the watch as we don't reference it while
- * generating the event
- */
- if (mask & IN_IGNORED || w->mask & IN_ONESHOT)
- put_inotify_watch(w); /* final put */
-
- /* coalescing: drop this event if it is a dupe of the previous */
- last = inotify_dev_get_last_event(dev);
- if (last && last->event.mask == mask && last->event.wd == wd &&
- last->event.cookie == cookie) {
- const char *lastname = last->name;
-
- if (!name && !lastname)
- goto out;
- if (name && lastname && !strcmp(lastname, name))
- goto out;
- }
-
- /* the queue overflowed and we already sent the Q_OVERFLOW event */
- if (unlikely(dev->event_count > dev->max_events))
- goto out;
-
- /* if the queue overflows, we need to notify user space */
- if (unlikely(dev->event_count == dev->max_events))
- kevent = kernel_event(-1, IN_Q_OVERFLOW, cookie, NULL);
- else
- kevent = kernel_event(wd, mask, cookie, name);
-
- if (unlikely(!kevent))
- goto out;
-
- /* queue the event and wake up anyone waiting */
- dev->event_count++;
- dev->queue_size += sizeof(struct inotify_event) + kevent->event.len;
- list_add_tail(&kevent->list, &dev->events);
- wake_up_interruptible(&dev->wq);
- kill_fasync(&dev->fa, SIGIO, POLL_IN);
-
-out:
- mutex_unlock(&dev->ev_mutex);
-}
-
-/*
- * remove_kevent - cleans up the given kevent
- *
- * Caller must hold dev->ev_mutex.
- */
-static void remove_kevent(struct inotify_device *dev,
- struct inotify_kernel_event *kevent)
-{
- list_del(&kevent->list);
-
- dev->event_count--;
- dev->queue_size -= sizeof(struct inotify_event) + kevent->event.len;
-}
-
-/*
- * free_kevent - frees the given kevent.
- */
-static void free_kevent(struct inotify_kernel_event *kevent)
-{
- kfree(kevent->name);
- kmem_cache_free(event_cachep, kevent);
-}
-
-/*
- * inotify_dev_event_dequeue - destroy an event on the given device
- *
- * Caller must hold dev->ev_mutex.
- */
-static void inotify_dev_event_dequeue(struct inotify_device *dev)
-{
- if (!list_empty(&dev->events)) {
- struct inotify_kernel_event *kevent;
- kevent = inotify_dev_get_event(dev);
- remove_kevent(dev, kevent);
- free_kevent(kevent);
- }
-}
-
-/*
- * find_inode - resolve a user-given path to a specific inode
- */
-static int find_inode(const char __user *dirname, struct path *path,
- unsigned flags)
-{
- int error;
-
- error = user_path_at(AT_FDCWD, dirname, flags, path);
- if (error)
- return error;
- /* you can only watch an inode if you have read permissions on it */
- error = inode_permission(path->dentry->d_inode, MAY_READ);
- if (error)
- path_put(path);
- return error;
-}
-
-/*
- * create_watch - creates a watch on the given device.
- *
- * Callers must hold dev->up_mutex.
- */
-static int create_watch(struct inotify_device *dev, struct inode *inode,
- u32 mask)
-{
- struct inotify_user_watch *watch;
- int ret;
-
- if (atomic_read(&dev->user->inotify_watches) >=
- inotify_max_user_watches)
- return -ENOSPC;
-
- watch = kmem_cache_alloc(watch_cachep, GFP_KERNEL);
- if (unlikely(!watch))
- return -ENOMEM;
-
- /* save a reference to device and bump the count to make it official */
- get_inotify_dev(dev);
- watch->dev = dev;
-
- atomic_inc(&dev->user->inotify_watches);
-
- inotify_init_watch(&watch->wdata);
- ret = inotify_add_watch(dev->ih, &watch->wdata, inode, mask);
- if (ret < 0)
- free_inotify_user_watch(&watch->wdata);
-
- return ret;
-}
-
-/* Device Interface */
-
+/* intofiy userspace file descriptor functions */
static unsigned int inotify_poll(struct file *file, poll_table *wait)
{
- struct inotify_device *dev = file->private_data;
+ struct fsnotify_group *group = file->private_data;
int ret = 0;

- poll_wait(file, &dev->wq, wait);
- mutex_lock(&dev->ev_mutex);
- if (!list_empty(&dev->events))
+ poll_wait(file, &group->notification_waitq, wait);
+ mutex_lock(&group->notification_mutex);
+ if (fsnotify_check_notif_queue(group))
ret = POLLIN | POLLRDNORM;
- mutex_unlock(&dev->ev_mutex);
+ mutex_unlock(&group->notification_mutex);

return ret;
}
@@ -432,26 +115,29 @@ static unsigned int inotify_poll(struct file *file, poll_table *wait)
* enough to fit in "count". Return an error pointer if
* not large enough.
*
- * Called with the device ev_mutex held.
+ * Called with the group->notification_mutex held.
*/
-static struct inotify_kernel_event *get_one_event(struct inotify_device *dev,
- size_t count)
+static struct fsnotify_event *get_one_event(struct fsnotify_group *group,
+ size_t count)
{
size_t event_size = sizeof(struct inotify_event);
- struct inotify_kernel_event *kevent;
+ struct fsnotify_event *event;

- if (list_empty(&dev->events))
+ if (!fsnotify_check_notif_queue(group))
return NULL;

- kevent = inotify_dev_get_event(dev);
- if (kevent->name)
- event_size += kevent->event.len;
+ event = fsnotify_peek_notif_event(group);
+
+ event_size += roundup(event->name_len, event_size);

if (event_size > count)
return ERR_PTR(-EINVAL);

- remove_kevent(dev, kevent);
- return kevent;
+ /* held the notification_mutex the whole time, so this is the
+ * same event we peeked above */
+ fsnotify_remove_notif_event(group);
+
+ return event;
}

/*
@@ -460,51 +146,82 @@ static struct inotify_kernel_event *get_one_event(struct inotify_device *dev,
* We already checked that the event size is smaller than the
* buffer we had in "get_one_event()" above.
*/
-static ssize_t copy_event_to_user(struct inotify_kernel_event *kevent,
+static ssize_t copy_event_to_user(struct fsnotify_group *group,
+ struct fsnotify_event *event,
char __user *buf)
{
+ struct inotify_event inotify_event;
+ struct inotify_event_private_data *priv;
size_t event_size = sizeof(struct inotify_event);
+ size_t name_len;
+
+ /* we get the inotify watch descriptor from the event private data */
+ spin_lock(&event->lock);
+ priv = (struct inotify_event_private_data *)fsnotify_get_priv_from_event(group, event);
+ inotify_event.wd = priv->wd;
+ __inotify_free_event_priv(priv);
+ spin_unlock(&event->lock);
+
+ /* round up event->name_len so it is a multiple of event_size */
+ name_len = roundup(event->name_len, event_size);
+ inotify_event.len = name_len;
+
+ inotify_event.mask = inotify_mask_to_arg(event->mask);
+ inotify_event.cookie = event->sync_cookie;

- if (copy_to_user(buf, &kevent->event, event_size))
+ /* send the main event */
+ if (copy_to_user(buf, &inotify_event, event_size))
return -EFAULT;

- if (kevent->name) {
- buf += event_size;
+ buf += event_size;

- if (copy_to_user(buf, kevent->name, kevent->event.len))
+ /*
+ * fsnotify only stores the pathname, so here we have to send the pathname
+ * and then pad that pathname out to a multiple of sizeof(inotify_event)
+ * with zeros. I get my zeros from the nul_inotify_event.
+ */
+ if (name_len) {
+ unsigned int len_to_zero = name_len - event->name_len;
+ /* copy the path name */
+ if (copy_to_user(buf, event->file_name, event->name_len))
return -EFAULT;
+ buf += event->name_len;

- event_size += kevent->event.len;
+ /* fill userspace with 0's from nul_inotify_event */
+ if (copy_to_user(buf, &nul_inotify_event, len_to_zero))
+ return -EFAULT;
+ buf += len_to_zero;
+ event_size += name_len;
}
+
return event_size;
}

static ssize_t inotify_read(struct file *file, char __user *buf,
size_t count, loff_t *pos)
{
- struct inotify_device *dev;
+ struct fsnotify_group *group;
+ struct fsnotify_event *kevent;
char __user *start;
int ret;
DEFINE_WAIT(wait);

start = buf;
- dev = file->private_data;
+ group = file->private_data;

while (1) {
- struct inotify_kernel_event *kevent;
+ prepare_to_wait(&group->notification_waitq, &wait, TASK_INTERRUPTIBLE);

- prepare_to_wait(&dev->wq, &wait, TASK_INTERRUPTIBLE);
-
- mutex_lock(&dev->ev_mutex);
- kevent = get_one_event(dev, count);
- mutex_unlock(&dev->ev_mutex);
+ mutex_lock(&group->notification_mutex);
+ kevent = get_one_event(group, count);
+ mutex_unlock(&group->notification_mutex);

if (kevent) {
ret = PTR_ERR(kevent);
if (IS_ERR(kevent))
break;
- ret = copy_event_to_user(kevent, buf);
- free_kevent(kevent);
+ ret = copy_event_to_user(group, kevent, buf);
+ fsnotify_put_event(kevent);
if (ret < 0)
break;
buf += ret;
@@ -525,7 +242,7 @@ static ssize_t inotify_read(struct file *file, char __user *buf,
schedule();
}

- finish_wait(&dev->wq, &wait);
+ finish_wait(&group->notification_waitq, &wait);
if (start != buf && ret != -EFAULT)
ret = buf - start;
return ret;
@@ -533,25 +250,35 @@ static ssize_t inotify_read(struct file *file, char __user *buf,

static int inotify_fasync(int fd, struct file *file, int on)
{
- struct inotify_device *dev = file->private_data;
+ struct fsnotify_group *group = file->private_data;

- return fasync_helper(fd, file, on, &dev->fa) >= 0 ? 0 : -EIO;
+ return fasync_helper(fd, file, on, &group->inotify_data.fa) >= 0 ? 0 : -EIO;
}

static int inotify_release(struct inode *ignored, struct file *file)
{
- struct inotify_device *dev = file->private_data;
+ struct fsnotify_group *group = file->private_data;
+ struct fsnotify_mark_entry *entry;

- inotify_destroy(dev->ih);
+ /* run all the entries remove them from the idr and drop that ref */
+ spin_lock(&group->mark_lock);
+ while (!list_empty(&group->mark_entries)) {
+ entry = list_first_entry(&group->mark_entries, struct fsnotify_mark_entry, g_list);

- /* destroy all of the events on this device */
- mutex_lock(&dev->ev_mutex);
- while (!list_empty(&dev->events))
- inotify_dev_event_dequeue(dev);
- mutex_unlock(&dev->ev_mutex);
+ /* make sure entry can't get freed */
+ fsnotify_get_mark(entry);
+ spin_unlock(&group->mark_lock);

- /* free this device: the put matching the get in inotify_init() */
- put_inotify_dev(dev);
+ inotify_destroy_mark_entry(entry);
+
+ /* ok, free it */
+ fsnotify_put_mark(entry);
+ spin_lock(&group->mark_lock);
+ }
+ spin_unlock(&group->mark_lock);
+
+ /* free this group, matching get was inotify_init->fsnotify_obtain_group */
+ fsnotify_put_group(group);

return 0;
}
@@ -559,16 +286,25 @@ static int inotify_release(struct inode *ignored, struct file *file)
static long inotify_ioctl(struct file *file, unsigned int cmd,
unsigned long arg)
{
- struct inotify_device *dev;
+ struct fsnotify_group *group;
+ struct fsnotify_event_holder *holder;
+ struct fsnotify_event *event;
void __user *p;
int ret = -ENOTTY;
+ size_t send_len = 0;

- dev = file->private_data;
+ group = file->private_data;
p = (void __user *) arg;

switch (cmd) {
case FIONREAD:
- ret = put_user(dev->queue_size, (int __user *) p);
+ mutex_lock(&group->notification_mutex);
+ list_for_each_entry(holder, &group->notification_list, event_list) {
+ event = holder->event;
+ send_len += sizeof(struct inotify_event) + event->name_len;
+ }
+ mutex_unlock(&group->notification_mutex);
+ ret = put_user(send_len, (int __user *) p);
break;
}

@@ -576,23 +312,18 @@ static long inotify_ioctl(struct file *file, unsigned int cmd,
}

static const struct file_operations inotify_fops = {
- .poll = inotify_poll,
- .read = inotify_read,
- .fasync = inotify_fasync,
- .release = inotify_release,
- .unlocked_ioctl = inotify_ioctl,
+ .poll = inotify_poll,
+ .read = inotify_read,
+ .fasync = inotify_fasync,
+ .release = inotify_release,
+ .unlocked_ioctl = inotify_ioctl,
.compat_ioctl = inotify_ioctl,
};

-static const struct inotify_operations inotify_user_ops = {
- .handle_event = inotify_dev_queue_event,
- .destroy_watch = free_inotify_user_watch,
-};
-
+/* inotify syscalls */
SYSCALL_DEFINE1(inotify_init1, int, flags)
{
- struct inotify_device *dev;
- struct inotify_handle *ih;
+ struct fsnotify_group *group;
struct user_struct *user;
struct file *filp;
int fd, ret;
@@ -621,45 +352,27 @@ SYSCALL_DEFINE1(inotify_init1, int, flags)
goto out_free_uid;
}

- dev = kmalloc(sizeof(struct inotify_device), GFP_KERNEL);
- if (unlikely(!dev)) {
- ret = -ENOMEM;
+ /* fsnotify_obtain_group took a reference to group, we put this when we kill the file in the end */
+ group = inotify_new_group(user, inotify_max_queued_events);
+ if (IS_ERR(group)) {
+ ret = PTR_ERR(group);
goto out_free_uid;
}

- ih = inotify_init(&inotify_user_ops);
- if (IS_ERR(ih)) {
- ret = PTR_ERR(ih);
- goto out_free_dev;
- }
- dev->ih = ih;
- dev->fa = NULL;
-
filp->f_op = &inotify_fops;
filp->f_path.mnt = mntget(inotify_mnt);
filp->f_path.dentry = dget(inotify_mnt->mnt_root);
filp->f_mapping = filp->f_path.dentry->d_inode->i_mapping;
filp->f_mode = FMODE_READ;
filp->f_flags = O_RDONLY | (flags & O_NONBLOCK);
- filp->private_data = dev;
-
- INIT_LIST_HEAD(&dev->events);
- init_waitqueue_head(&dev->wq);
- mutex_init(&dev->ev_mutex);
- mutex_init(&dev->up_mutex);
- dev->event_count = 0;
- dev->queue_size = 0;
- dev->max_events = inotify_max_queued_events;
- dev->user = user;
- atomic_set(&dev->count, 0);
-
- get_inotify_dev(dev);
+ filp->private_data = group;
+
atomic_inc(&user->inotify_devs);
+
fd_install(fd, filp);

return fd;
-out_free_dev:
- kfree(dev);
+
out_free_uid:
free_uid(user);
put_filp(filp);
@@ -676,8 +389,8 @@ SYSCALL_DEFINE0(inotify_init)
SYSCALL_DEFINE3(inotify_add_watch, int, fd, const char __user *, pathname,
u32, mask)
{
+ struct fsnotify_group *group;
struct inode *inode;
- struct inotify_device *dev;
struct path path;
struct file *filp;
int ret, fput_needed;
@@ -699,19 +412,19 @@ SYSCALL_DEFINE3(inotify_add_watch, int, fd, const char __user *, pathname,
flags |= LOOKUP_DIRECTORY;

ret = find_inode(pathname, &path, flags);
- if (unlikely(ret))
+ if (ret)
goto fput_and_out;

- /* inode held in place by reference to path; dev by fget on fd */
+ /* inode held in place by reference to path; group by fget on fd */
inode = path.dentry->d_inode;
- dev = filp->private_data;
+ group = filp->private_data;

- mutex_lock(&dev->up_mutex);
- ret = inotify_find_update_watch(dev->ih, inode, mask);
- if (ret == -ENOENT)
- ret = create_watch(dev, inode, mask);
- mutex_unlock(&dev->up_mutex);
+ /* create/update an inode mark */
+ ret = inotify_update_watch(group, inode, mask);
+ if (unlikely(ret))
+ goto path_put_and_out;

+path_put_and_out:
path_put(&path);
fput_and_out:
fput_light(filp, fput_needed);
@@ -720,9 +433,10 @@ fput_and_out:

SYSCALL_DEFINE2(inotify_rm_watch, int, fd, __s32, wd)
{
+ struct fsnotify_group *group;
+ struct fsnotify_mark_entry *entry;
struct file *filp;
- struct inotify_device *dev;
- int ret, fput_needed;
+ int ret = 0, fput_needed;

filp = fget_light(fd, &fput_needed);
if (unlikely(!filp))
@@ -734,10 +448,20 @@ SYSCALL_DEFINE2(inotify_rm_watch, int, fd, __s32, wd)
goto out;
}

- dev = filp->private_data;
+ group = filp->private_data;

- /* we free our watch data when we get IN_IGNORED */
- ret = inotify_rm_wd(dev->ih, wd);
+ spin_lock(&group->inotify_data.idr_lock);
+ entry = idr_find(&group->inotify_data.idr, wd);
+ if (unlikely(!entry)) {
+ spin_unlock(&group->inotify_data.idr_lock);
+ ret = -EINVAL;
+ goto out;
+ }
+ fsnotify_get_mark(entry);
+ spin_unlock(&group->inotify_data.idr_lock);
+
+ inotify_destroy_mark_entry(entry);
+ fsnotify_put_mark(entry);

out:
fput_light(filp, fput_needed);
@@ -753,9 +477,9 @@ inotify_get_sb(struct file_system_type *fs_type, int flags,
}

static struct file_system_type inotify_fs_type = {
- .name = "inotifyfs",
- .get_sb = inotify_get_sb,
- .kill_sb = kill_anon_super,
+ .name = "inotifyfs",
+ .get_sb = inotify_get_sb,
+ .kill_sb = kill_anon_super,
};

/*
@@ -779,14 +503,6 @@ static int __init inotify_user_setup(void)
inotify_max_user_instances = 128;
inotify_max_user_watches = 8192;

- watch_cachep = kmem_cache_create("inotify_watch_cache",
- sizeof(struct inotify_user_watch),
- 0, SLAB_PANIC, NULL);
- event_cachep = kmem_cache_create("inotify_event_cache",
- sizeof(struct inotify_kernel_event),
- 0, SLAB_PANIC, NULL);
-
return 0;
}
-
module_init(inotify_user_setup);
diff --git a/fs/notify/notification.c b/fs/notify/notification.c
index a636a1f..14af58d 100644
--- a/fs/notify/notification.c
+++ b/fs/notify/notification.c
@@ -123,7 +123,7 @@ static inline int event_compare(struct fsnotify_event *old, struct fsnotify_even
*/
int fsnotify_add_notif_event(struct fsnotify_group *group, struct fsnotify_event *event, struct fsnotify_event_private_data *priv)
{
- struct fsnotify_event_holder *holder;
+ struct fsnotify_event_holder *holder = NULL;
struct list_head *list = &group->notification_list;
struct fsnotify_event_holder *last_holder;
struct fsnotify_event *last_event;
@@ -141,21 +141,35 @@ int fsnotify_add_notif_event(struct fsnotify_group *group, struct fsnotify_event
* the holder spinlock. If we see it blank we know that once we
* get that lock the in event holder will be ok for us to (re)use.
*/
- if (list_empty(&event->holder.event_list))
- holder = (struct fsnotify_event_holder *)event;
- else
+ if (!list_empty(&event->holder.event_list)) {
holder = alloc_event_holder();
-
- if (!holder)
- return -ENOMEM;
+ if (!holder)
+ return -ENOMEM;
+ }

mutex_lock(&group->notification_mutex);

- if (group->q_len >= group->max_events)
+ if (group->q_len >= group->max_events)
event = &q_overflow_event;

spin_lock(&event->lock);

+ if (list_empty(&event->holder.event_list)) {
+ if (unlikely(holder))
+ fsnotify_destroy_event_holder(holder);
+ holder = &event->holder;
+ } else if (unlikely(!holder)) {
+ /* this only happens if we had room in the original event holder
+ * but we switched to the overflow event and that in event holder
+ * was in use */
+ holder = alloc_event_holder();
+ if (!holder) {
+ spin_unlock(&event->lock);
+ mutex_unlock(&group->notification_mutex);
+ return -ENOMEM;
+ }
+ }
+
if (!list_empty(list)) {
last_holder = list_entry(list->prev, struct fsnotify_event_holder, event_list);
last_event = last_holder->event;
diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h
index bdbc897..e60c943 100644
--- a/include/linux/fsnotify.h
+++ b/include/linux/fsnotify.h
@@ -265,10 +265,10 @@ static inline void fsnotify_mkdir(struct inode *inode, struct dentry *dentry)
static inline void fsnotify_access(struct dentry *dentry)
{
struct inode *inode = dentry->d_inode;
- __u64 mask = IN_ACCESS;
+ __u64 mask = FS_ACCESS;

if (S_ISDIR(inode->i_mode))
- mask |= IN_ISDIR;
+ mask |= FS_IN_ISDIR;

inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name);
inotify_inode_queue_event(inode, mask, 0, NULL, NULL);
@@ -283,10 +283,10 @@ static inline void fsnotify_access(struct dentry *dentry)
static inline void fsnotify_modify(struct dentry *dentry)
{
struct inode *inode = dentry->d_inode;
- __u64 mask = IN_MODIFY;
+ __u64 mask = FS_MODIFY;

if (S_ISDIR(inode->i_mode))
- mask |= IN_ISDIR;
+ mask |= FS_IN_ISDIR;

inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name);
inotify_inode_queue_event(inode, mask, 0, NULL, NULL);
@@ -301,10 +301,10 @@ static inline void fsnotify_modify(struct dentry *dentry)
static inline void fsnotify_open(struct dentry *dentry)
{
struct inode *inode = dentry->d_inode;
- __u64 mask = IN_OPEN;
+ __u64 mask = FS_OPEN;

if (S_ISDIR(inode->i_mode))
- mask |= IN_ISDIR;
+ mask |= FS_IN_ISDIR;

inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name);
inotify_inode_queue_event(inode, mask, 0, NULL, NULL);
@@ -320,14 +320,13 @@ static inline void fsnotify_close(struct file *file)
{
struct dentry *dentry = file->f_path.dentry;
struct inode *inode = dentry->d_inode;
- const char *name = dentry->d_name.name;
fmode_t mode = file->f_mode;
- __u64 mask = (mode & FMODE_WRITE) ? IN_CLOSE_WRITE : IN_CLOSE_NOWRITE;
+ __u64 mask = (mode & FMODE_WRITE) ? FS_CLOSE_WRITE : FS_CLOSE_NOWRITE;

if (S_ISDIR(inode->i_mode))
- mask |= IN_ISDIR;
+ mask |= FS_IN_ISDIR;

- inotify_dentry_parent_queue_event(dentry, mask, 0, name);
+ inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name);
inotify_inode_queue_event(inode, mask, 0, NULL, NULL);

fsnotify_parent(dentry, mask);
@@ -340,10 +339,10 @@ static inline void fsnotify_close(struct file *file)
static inline void fsnotify_xattr(struct dentry *dentry)
{
struct inode *inode = dentry->d_inode;
- __u64 mask = IN_ATTRIB;
+ __u64 mask = FS_ATTRIB;

if (S_ISDIR(inode->i_mode))
- mask |= IN_ISDIR;
+ mask |= FS_IN_ISDIR;

inotify_dentry_parent_queue_event(dentry, mask, 0, dentry->d_name.name);
inotify_inode_queue_event(inode, mask, 0, NULL, NULL);
@@ -362,26 +361,26 @@ static inline void fsnotify_change(struct dentry *dentry, unsigned int ia_valid)
__u64 mask = 0;

if (ia_valid & ATTR_UID)
- mask |= IN_ATTRIB;
+ mask |= FS_ATTRIB;
if (ia_valid & ATTR_GID)
- mask |= IN_ATTRIB;
+ mask |= FS_ATTRIB;
if (ia_valid & ATTR_SIZE)
- mask |= IN_MODIFY;
+ mask |= FS_MODIFY;

/* both times implies a utime(s) call */
if ((ia_valid & (ATTR_ATIME | ATTR_MTIME)) == (ATTR_ATIME | ATTR_MTIME))
- mask |= IN_ATTRIB;
+ mask |= FS_ATTRIB;
else if (ia_valid & ATTR_ATIME)
- mask |= IN_ACCESS;
+ mask |= FS_ACCESS;
else if (ia_valid & ATTR_MTIME)
- mask |= IN_MODIFY;
+ mask |= FS_MODIFY;

if (ia_valid & ATTR_MODE)
- mask |= IN_ATTRIB;
+ mask |= FS_ATTRIB;

if (mask) {
if (S_ISDIR(inode->i_mode))
- mask |= IN_ISDIR;
+ mask |= FS_IN_ISDIR;
inotify_inode_queue_event(inode, mask, 0, NULL, NULL);
inotify_dentry_parent_queue_event(dentry, mask, 0,
dentry->d_name.name);
diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h
index 203842c..0070cc3 100644
--- a/include/linux/fsnotify_backend.h
+++ b/include/linux/fsnotify_backend.h
@@ -9,6 +9,7 @@

#ifdef __KERNEL__

+#include <linux/idr.h> /* inotify uses this */
#include <linux/fs.h> /* struct inode */
#include <linux/list.h>
#include <linux/path.h> /* struct path */
@@ -80,6 +81,7 @@

/* listeners that hard code group numbers near the top */
#define DNOTIFY_GROUP_NUM UINT_MAX
+#define INOTIFY_GROUP_NUM (DNOTIFY_GROUP_NUM-1)

struct fsnotify_group;
struct fsnotify_event;
@@ -118,6 +120,15 @@ struct fsnotify_group {

/* groups can define private fields here */
union {
+#ifdef CONFIG_INOTIFY_USER
+ struct inotify_group_private_data {
+ spinlock_t idr_lock;
+ struct idr idr;
+ u32 last_wd;
+ struct fasync_struct *fa; /* async notification */
+ struct user_struct *user;
+ } inotify_data;
+#endif
};
};


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/