#include <linux/list.h>
#include <linux/writeback.h>
#include <linux/inotify.h>
+#include <linux/syscalls.h>
#include <asm/ioctls.h>
* Lock ordering:
*
* dentry->d_lock (used to keep d_move() away from dentry->d_parent)
- * iprune_sem (synchronize shrink_icache_memory())
+ * iprune_mutex (synchronize shrink_icache_memory())
* inode_lock (protects the super_block->s_inodes list)
- * inode->inotify_sem (protects inode->inotify_watches and watches->i_list)
- * inotify_dev->sem (protects inotify_device and watches->d_list)
+ * inode->inotify_mutex (protects inode->inotify_watches and watches->i_list)
+ * inotify_dev->mutex (protects inotify_device and watches->d_list)
*/
/*
* Lifetimes of the three main data structures--inotify_device, inode, and
* inotify_watch--are managed by reference count.
*
- * inotify_device: Lifetime is from open until release. Additional references
- * can bump the count via get_inotify_dev() and drop the count via
+ * inotify_device: Lifetime is from inotify_init() until release. Additional
+ * references can bump the count via get_inotify_dev() and drop the count via
* put_inotify_dev().
*
* inotify_watch: Lifetime is from create_watch() to destory_watch().
*/
/*
- * struct inotify_device - represents an open instance of an inotify device
+ * struct inotify_device - represents an inotify instance
*
- * This structure is protected by the semaphore 'sem'.
+ * This structure is protected by the mutex 'mutex'.
*/
struct inotify_device {
wait_queue_head_t wq; /* wait queue for i/o */
struct idr idr; /* idr mapping wd -> watch */
- struct semaphore sem; /* protects this bad boy */
+ struct mutex mutex; /* protects this bad boy */
struct list_head events; /* list of queued events */
struct list_head watches; /* list of watches */
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 */
+ u32 last_wd; /* the last wd allocated */
};
/*
* device. In read(), this list is walked and all events that can fit in the
* buffer are returned.
*
- * Protected by dev->sem of the device in which we are queued.
+ * Protected by dev->mutex of the device in which we are queued.
*/
struct inotify_kernel_event {
struct inotify_event event; /* the user-space event */
/*
* struct inotify_watch - represents a watch request on a specific inode
*
- * d_list is protected by dev->sem of the associated watch->dev.
- * i_list and mask are protected by inode->inotify_sem of the associated inode.
+ * d_list is protected by dev->mutex of the associated watch->dev.
+ * i_list and mask are protected by inode->inotify_mutex of the associated inode.
* dev, inode, and wd are never written to once the watch is created.
*/
struct inotify_watch {
if (atomic_dec_and_test(&dev->count)) {
atomic_dec(&dev->user->inotify_devs);
free_uid(dev->user);
+ idr_destroy(&dev->idr);
kfree(dev);
}
}
/*
* inotify_dev_get_event - return the next event in the given dev's queue
*
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
*/
static inline struct inotify_kernel_event *
inotify_dev_get_event(struct inotify_device *dev)
/*
* inotify_dev_queue_event - add a new event to the given device
*
- * Caller must hold dev->sem. Can sleep (calls kernel_event()).
+ * Caller must hold dev->mutex. Can sleep (calls kernel_event()).
*/
static void inotify_dev_queue_event(struct inotify_device *dev,
struct inotify_watch *watch, u32 mask,
/*
* remove_kevent - cleans up and ultimately frees the given kevent
*
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
*/
static void remove_kevent(struct inotify_device *dev,
struct inotify_kernel_event *kevent)
/*
* inotify_dev_event_dequeue - destroy an event on the given device
*
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
*/
static void inotify_dev_event_dequeue(struct inotify_device *dev)
{
/*
* inotify_dev_get_wd - returns the next WD for use by the given dev
*
- * Callers must hold dev->sem. This function can sleep.
+ * Callers must hold dev->mutex. This function can sleep.
*/
static int inotify_dev_get_wd(struct inotify_device *dev,
struct inotify_watch *watch)
do {
if (unlikely(!idr_pre_get(&dev->idr, GFP_KERNEL)))
return -ENOSPC;
- ret = idr_get_new(&dev->idr, watch, &watch->wd);
+ ret = idr_get_new_above(&dev->idr, watch, dev->last_wd+1, &watch->wd);
} while (ret == -EAGAIN);
return ret;
/*
* find_inode - resolve a user-given path to a specific inode and return a nd
*/
-static int find_inode(const char __user *dirname, struct nameidata *nd)
+static int find_inode(const char __user *dirname, struct nameidata *nd,
+ unsigned flags)
{
int error;
- error = __user_walk(dirname, LOOKUP_FOLLOW, nd);
+ error = __user_walk(dirname, flags, nd);
if (error)
return error;
/* you can only watch an inode if you have read permissions on it */
- error = permission(nd->dentry->d_inode, MAY_READ, NULL);
+ error = vfs_permission(nd, MAY_READ);
if (error)
- path_release (nd);
+ path_release(nd);
return error;
}
+/*
+ * inotify_inode_watched - returns nonzero if there are watches on this inode
+ * and zero otherwise. We call this lockless, we do not care if we race.
+ */
+static inline int inotify_inode_watched(struct inode *inode)
+{
+ return !list_empty(&inode->inotify_watches);
+}
+
+/*
+ * Get child dentry flag into synch with parent inode.
+ * Flag should always be clear for negative dentrys.
+ */
+static void set_dentry_child_flags(struct inode *inode, int watched)
+{
+ struct dentry *alias;
+
+ spin_lock(&dcache_lock);
+ list_for_each_entry(alias, &inode->i_dentry, d_alias) {
+ struct dentry *child;
+
+ list_for_each_entry(child, &alias->d_subdirs, d_u.d_child) {
+ if (!child->d_inode) {
+ WARN_ON(child->d_flags & DCACHE_INOTIFY_PARENT_WATCHED);
+ continue;
+ }
+ spin_lock(&child->d_lock);
+ if (watched) {
+ WARN_ON(child->d_flags &
+ DCACHE_INOTIFY_PARENT_WATCHED);
+ child->d_flags |= DCACHE_INOTIFY_PARENT_WATCHED;
+ } else {
+ WARN_ON(!(child->d_flags &
+ DCACHE_INOTIFY_PARENT_WATCHED));
+ child->d_flags&=~DCACHE_INOTIFY_PARENT_WATCHED;
+ }
+ spin_unlock(&child->d_lock);
+ }
+ }
+ spin_unlock(&dcache_lock);
+}
+
/*
* create_watch - creates a watch on the given device.
*
- * Callers must hold dev->sem. Calls inotify_dev_get_wd() so may sleep.
+ * Callers must hold dev->mutex. Calls inotify_dev_get_wd() so may sleep.
* Both 'dev' and 'inode' (by way of nameidata) need to be pinned.
*/
static struct inotify_watch *create_watch(struct inotify_device *dev,
struct inotify_watch *watch;
int ret;
- if (atomic_read(&dev->user->inotify_watches) >= inotify_max_user_watches)
+ if (atomic_read(&dev->user->inotify_watches) >=
+ inotify_max_user_watches)
return ERR_PTR(-ENOSPC);
watch = kmem_cache_alloc(watch_cachep, GFP_KERNEL);
return ERR_PTR(ret);
}
+ dev->last_wd = watch->wd;
watch->mask = mask;
atomic_set(&watch->count, 0);
INIT_LIST_HEAD(&watch->d_list);
/*
* inotify_find_dev - find the watch associated with the given inode and dev
*
- * Callers must hold inode->inotify_sem.
+ * Callers must hold inode->inotify_mutex.
*/
static struct inotify_watch *inode_find_dev(struct inode *inode,
struct inotify_device *dev)
list_del(&watch->i_list);
list_del(&watch->d_list);
+ if (!inotify_inode_watched(watch->inode))
+ set_dentry_child_flags(watch->inode, 0);
+
atomic_dec(&dev->user->inotify_watches);
idr_remove(&dev->idr, watch->wd);
put_inotify_watch(watch);
* the IN_IGNORED event to the given device signifying that the inode is no
* longer watched.
*
- * Callers must hold both inode->inotify_sem and dev->sem. We drop a
+ * Callers must hold both inode->inotify_mutex and dev->mutex. We drop a
* reference to the inode before returning.
*
* The inode is not iput() so as to remain atomic. If the inode needs to be
remove_watch_no_event(watch, dev);
}
+/* Kernel API */
+
/*
- * inotify_inode_watched - returns nonzero if there are watches on this inode
- * and zero otherwise. We call this lockless, we do not care if we race.
+ * inotify_d_instantiate - instantiate dcache entry for inode
*/
-static inline int inotify_inode_watched(struct inode *inode)
+void inotify_d_instantiate(struct dentry *entry, struct inode *inode)
{
- return !list_empty(&inode->inotify_watches);
+ struct dentry *parent;
+
+ if (!inode)
+ return;
+
+ WARN_ON(entry->d_flags & DCACHE_INOTIFY_PARENT_WATCHED);
+ spin_lock(&entry->d_lock);
+ parent = entry->d_parent;
+ if (inotify_inode_watched(parent->d_inode))
+ entry->d_flags |= DCACHE_INOTIFY_PARENT_WATCHED;
+ spin_unlock(&entry->d_lock);
}
-/* Kernel API */
+/*
+ * inotify_d_move - dcache entry has been moved
+ */
+void inotify_d_move(struct dentry *entry)
+{
+ struct dentry *parent;
+
+ parent = entry->d_parent;
+ if (inotify_inode_watched(parent->d_inode))
+ entry->d_flags |= DCACHE_INOTIFY_PARENT_WATCHED;
+ else
+ entry->d_flags &= ~DCACHE_INOTIFY_PARENT_WATCHED;
+}
/**
* inotify_inode_queue_event - queue an event to all watches on this inode
if (!inotify_inode_watched(inode))
return;
- down(&inode->inotify_sem);
+ mutex_lock(&inode->inotify_mutex);
list_for_each_entry_safe(watch, next, &inode->inotify_watches, i_list) {
u32 watch_mask = watch->mask;
if (watch_mask & mask) {
struct inotify_device *dev = watch->dev;
get_inotify_watch(watch);
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
inotify_dev_queue_event(dev, watch, mask, cookie, name);
if (watch_mask & IN_ONESHOT)
remove_watch_no_event(watch, dev);
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
put_inotify_watch(watch);
}
}
- up(&inode->inotify_sem);
+ mutex_unlock(&inode->inotify_mutex);
}
EXPORT_SYMBOL_GPL(inotify_inode_queue_event);
struct dentry *parent;
struct inode *inode;
+ if (!(dentry->d_flags & DCACHE_INOTIFY_PARENT_WATCHED))
+ return;
+
spin_lock(&dentry->d_lock);
parent = dentry->d_parent;
inode = parent->d_inode;
* @list: list of inodes being unmounted (sb->s_inodes)
*
* Called with inode_lock held, protecting the unmounting super block's list
- * of inodes, and with iprune_sem held, keeping shrink_icache_memory() at bay.
+ * of inodes, and with iprune_mutex held, keeping shrink_icache_memory() at bay.
* We temporarily drop inode_lock, however, and CAN block.
*/
void inotify_unmount_inodes(struct list_head *list)
* We can safely drop inode_lock here because we hold
* references on both inode and next_i. Also no new inodes
* will be added since the umount has begun. Finally,
- * iprune_sem keeps shrink_icache_memory() away.
+ * iprune_mutex keeps shrink_icache_memory() away.
*/
spin_unlock(&inode_lock);
iput(need_iput_tmp);
/* for each watch, send IN_UNMOUNT and then remove it */
- down(&inode->inotify_sem);
+ mutex_lock(&inode->inotify_mutex);
watches = &inode->inotify_watches;
list_for_each_entry_safe(watch, next_w, watches, i_list) {
struct inotify_device *dev = watch->dev;
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
inotify_dev_queue_event(dev, watch, IN_UNMOUNT,0,NULL);
remove_watch(watch, dev);
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
}
- up(&inode->inotify_sem);
+ mutex_unlock(&inode->inotify_mutex);
iput(inode);
spin_lock(&inode_lock);
{
struct inotify_watch *watch, *next;
- down(&inode->inotify_sem);
+ mutex_lock(&inode->inotify_mutex);
list_for_each_entry_safe(watch, next, &inode->inotify_watches, i_list) {
struct inotify_device *dev = watch->dev;
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
remove_watch(watch, dev);
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
}
- up(&inode->inotify_sem);
+ mutex_unlock(&inode->inotify_mutex);
}
EXPORT_SYMBOL_GPL(inotify_inode_is_dead);
int ret = 0;
poll_wait(file, &dev->wq, wait);
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
if (!list_empty(&dev->events))
ret = POLLIN | POLLRDNORM;
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
return ret;
}
prepare_to_wait(&dev->wq, &wait, TASK_INTERRUPTIBLE);
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
events = !list_empty(&dev->events);
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
if (events) {
ret = 0;
break;
if (ret)
return ret;
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
while (1) {
struct inotify_kernel_event *kevent;
remove_kevent(dev, kevent);
}
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
return ret;
}
* Destroy all of the watches on this device. Unfortunately, not very
* pretty. We cannot do a simple iteration over the list, because we
* do not know the inode until we iterate to the watch. But we need to
- * hold inode->inotify_sem before dev->sem. The following works.
+ * hold inode->inotify_mutex before dev->mutex. The following works.
*/
while (1) {
struct inotify_watch *watch;
struct list_head *watches;
struct inode *inode;
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
watches = &dev->watches;
if (list_empty(watches)) {
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
break;
}
watch = list_entry(watches->next, struct inotify_watch, d_list);
get_inotify_watch(watch);
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
inode = watch->inode;
- down(&inode->inotify_sem);
- down(&dev->sem);
+ mutex_lock(&inode->inotify_mutex);
+ mutex_lock(&dev->mutex);
remove_watch_no_event(watch, dev);
- up(&dev->sem);
- up(&inode->inotify_sem);
+ mutex_unlock(&dev->mutex);
+ mutex_unlock(&inode->inotify_mutex);
put_inotify_watch(watch);
}
/* destroy all of the events on this device */
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
while (!list_empty(&dev->events))
inotify_dev_event_dequeue(dev);
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
- /* free this device: the put matching the get in inotify_open() */
+ /* free this device: the put matching the get in inotify_init() */
put_inotify_dev(dev);
return 0;
}
/*
- * inotify_ignore - handle the INOTIFY_IGNORE ioctl, asking that a given wd be
- * removed from the device.
+ * inotify_ignore - remove a given wd from this inotify instance.
*
* Can sleep.
*/
struct inotify_watch *watch;
struct inode *inode;
- down(&dev->sem);
+ mutex_lock(&dev->mutex);
watch = idr_find(&dev->idr, wd);
if (unlikely(!watch)) {
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
return -EINVAL;
}
get_inotify_watch(watch);
inode = watch->inode;
- up(&dev->sem);
+ mutex_unlock(&dev->mutex);
- down(&inode->inotify_sem);
- down(&dev->sem);
+ mutex_lock(&inode->inotify_mutex);
+ mutex_lock(&dev->mutex);
/* make sure that we did not race */
watch = idr_find(&dev->idr, wd);
if (likely(watch))
remove_watch(watch, dev);
- up(&dev->sem);
- up(&inode->inotify_sem);
+ mutex_unlock(&dev->mutex);
+ mutex_unlock(&inode->inotify_mutex);
put_inotify_watch(watch);
return 0;
{
struct inotify_device *dev;
struct user_struct *user;
- int ret = -ENOTTY;
- int fd;
- struct file *filp;
+ struct file *filp;
+ int fd, ret;
fd = get_unused_fd();
- if (fd < 0) {
- ret = fd;
- goto out;
- }
+ if (fd < 0)
+ return fd;
filp = get_empty_filp();
if (!filp) {
- put_unused_fd(fd);
ret = -ENFILE;
- goto out;
+ goto out_put_fd;
}
- filp->f_op = &inotify_fops;
- filp->f_vfsmnt = mntget(inotify_mnt);
- filp->f_dentry = dget(inotify_mnt->mnt_root);
- filp->f_mapping = filp->f_dentry->d_inode->i_mapping;
- filp->f_mode = FMODE_READ;
- filp->f_flags = O_RDONLY;
user = get_uid(current->user);
-
- if (unlikely(atomic_read(&user->inotify_devs) >= inotify_max_user_instances)) {
+ if (unlikely(atomic_read(&user->inotify_devs) >=
+ inotify_max_user_instances)) {
ret = -EMFILE;
- goto out_err;
+ goto out_free_uid;
}
dev = kmalloc(sizeof(struct inotify_device), GFP_KERNEL);
if (unlikely(!dev)) {
ret = -ENOMEM;
- goto out_err;
+ goto out_free_uid;
}
+ filp->f_op = &inotify_fops;
+ filp->f_vfsmnt = mntget(inotify_mnt);
+ filp->f_dentry = dget(inotify_mnt->mnt_root);
+ filp->f_mapping = filp->f_dentry->d_inode->i_mapping;
+ filp->f_mode = FMODE_READ;
+ filp->f_flags = O_RDONLY;
+ filp->private_data = dev;
+
idr_init(&dev->idr);
INIT_LIST_HEAD(&dev->events);
INIT_LIST_HEAD(&dev->watches);
init_waitqueue_head(&dev->wq);
- sema_init(&dev->sem, 1);
+ mutex_init(&dev->mutex);
dev->event_count = 0;
dev->queue_size = 0;
dev->max_events = inotify_max_queued_events;
dev->user = user;
+ dev->last_wd = 0;
atomic_set(&dev->count, 0);
get_inotify_dev(dev);
atomic_inc(&user->inotify_devs);
+ fd_install(fd, filp);
- filp->private_data = dev;
- fd_install (fd, filp);
return fd;
-out_err:
- put_unused_fd (fd);
- put_filp (filp);
+out_free_uid:
free_uid(user);
-out:
+ put_filp(filp);
+out_put_fd:
+ put_unused_fd(fd);
return ret;
}
-asmlinkage long sys_inotify_add_watch(int fd, const char *path, u32 mask)
+asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask)
{
struct inotify_watch *watch, *old;
struct inode *inode;
struct inotify_device *dev;
struct nameidata nd;
struct file *filp;
- int ret;
+ int ret, fput_needed;
+ int mask_add = 0;
+ unsigned flags = 0;
- filp = fget(fd);
- if (!filp)
+ filp = fget_light(fd, &fput_needed);
+ if (unlikely(!filp))
return -EBADF;
- dev = filp->private_data;
+ /* verify that this is indeed an inotify instance */
+ if (unlikely(filp->f_op != &inotify_fops)) {
+ ret = -EINVAL;
+ goto fput_and_out;
+ }
- ret = find_inode((const char __user*) path, &nd);
- if (ret)
+ if (!(mask & IN_DONT_FOLLOW))
+ flags |= LOOKUP_FOLLOW;
+ if (mask & IN_ONLYDIR)
+ flags |= LOOKUP_DIRECTORY;
+
+ ret = find_inode(path, &nd, flags);
+ if (unlikely(ret))
goto fput_and_out;
- /* Held in place by reference in nd */
+ /* inode held in place by reference to nd; dev by fget on fd */
inode = nd.dentry->d_inode;
+ dev = filp->private_data;
+
+ mutex_lock(&inode->inotify_mutex);
+ mutex_lock(&dev->mutex);
- down(&inode->inotify_sem);
- down(&dev->sem);
+ if (mask & IN_MASK_ADD)
+ mask_add = 1;
/* don't let user-space set invalid bits: we don't want flags set */
- mask &= IN_ALL_EVENTS;
- if (!mask) {
+ mask &= IN_ALL_EVENTS | IN_ONESHOT;
+ if (unlikely(!mask)) {
ret = -EINVAL;
goto out;
}
*/
old = inode_find_dev(inode, dev);
if (unlikely(old)) {
- old->mask = mask;
+ if (mask_add)
+ old->mask |= mask;
+ else
+ old->mask = mask;
ret = old->wd;
goto out;
}
goto out;
}
+ if (!inotify_inode_watched(inode))
+ set_dentry_child_flags(inode, 1);
+
/* Add the watch to the device's and the inode's list */
list_add(&watch->d_list, &dev->watches);
list_add(&watch->i_list, &inode->inotify_watches);
ret = watch->wd;
out:
- path_release (&nd);
- up(&dev->sem);
- up(&inode->inotify_sem);
+ mutex_unlock(&dev->mutex);
+ mutex_unlock(&inode->inotify_mutex);
+ path_release(&nd);
fput_and_out:
- fput(filp);
+ fput_light(filp, fput_needed);
return ret;
}
{
struct file *filp;
struct inotify_device *dev;
- int ret;
+ int ret, fput_needed;
- filp = fget(fd);
- if (!filp)
+ filp = fget_light(fd, &fput_needed);
+ if (unlikely(!filp))
return -EBADF;
+
+ /* verify that this is indeed an inotify instance */
+ if (unlikely(filp->f_op != &inotify_fops)) {
+ ret = -EINVAL;
+ goto out;
+ }
+
dev = filp->private_data;
ret = inotify_ignore(dev, wd);
- fput(filp);
+out:
+ fput_light(filp, fput_needed);
return ret;
}
};
/*
- * inotify_init - Our initialization function. Note that we cannnot return
+ * inotify_setup - Our initialization function. Note that we cannnot return
* error because we have compiled-in VFS hooks. So an (unlikely) failure here
* must result in panic().
*/
-static int __init inotify_init(void)
+static int __init inotify_setup(void)
{
- register_filesystem(&inotify_fs_type);
+ int ret;
+
+ ret = register_filesystem(&inotify_fs_type);
+ if (unlikely(ret))
+ panic("inotify: register_filesystem returned %d!\n", ret);
+
inotify_mnt = kern_mount(&inotify_fs_type);
+ if (IS_ERR(inotify_mnt))
+ panic("inotify: kern_mount ret %ld!\n", PTR_ERR(inotify_mnt));
- inotify_max_queued_events = 8192;
- inotify_max_user_instances = 8;
+ inotify_max_queued_events = 16384;
+ inotify_max_user_instances = 128;
inotify_max_user_watches = 8192;
atomic_set(&inotify_cookie, 0);
return 0;
}
-module_init(inotify_init);
+module_init(inotify_setup);