Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
fuse: clean up annotations of fc->lock
fuse: fix sparse warning in ioctl
fuse: update interface version
fuse: add fuse_conn->release()
fuse: separate out fuse_conn_init() from new_conn()
fuse: add fuse_ prefix to several functions
fuse: implement poll support
fuse: implement unsolicited notification
fuse: add file kernel handle
fuse: implement ioctl support
fuse: don't let fuse_req->end() put the base reference
fuse: move FUSE_MINOR to miscdevice.h
fuse: style fixes

fs/fuse/control.c

···
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <[e-mail]>
+  Copyright (C) 2001-2008  Miklos Szeredi <[e-mail]>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
···
 	size_t size;
 
 	if (!*ppos) {
+		long value;
 		struct fuse_conn *fc = fuse_ctl_file_conn_get(file);
 		if (!fc)
 			return 0;
 
-		file->private_data=(void *)(long)atomic_read(&fc->num_waiting);
+		value = atomic_read(&fc->num_waiting);
+		file->private_data = (void *)value;
 		fuse_conn_put(fc);
 	}
 	size = sprintf(tmp, "%ld\n", (long)file->private_data);

fs/fuse/dev.c

···
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <[e-mail]>
+  Copyright (C) 2001-2008  Miklos Szeredi <[e-mail]>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
···
  * Called with fc->lock, unlocks it
  */
 static void request_end(struct fuse_conn *fc, struct fuse_req *req)
-	__releases(fc->lock)
+__releases(&fc->lock)
 {
 	void (*end) (struct fuse_conn *, struct fuse_req *) = req->end;
 	req->end = NULL;
···
 	wake_up(&req->waitq);
 	if (end)
 		end(fc, req);
-	else
-		fuse_put_request(fc, req);
+	fuse_put_request(fc, req);
 }
 
 static void wait_answer_interruptible(struct fuse_conn *fc,
 				      struct fuse_req *req)
-	__releases(fc->lock) __acquires(fc->lock)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	if (signal_pending(current))
 		return;
···
 }
 
 static void request_wait_answer(struct fuse_conn *fc, struct fuse_req *req)
-	__releases(fc->lock) __acquires(fc->lock)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	if (!fc->no_interrupt) {
 		/* Any signal may interrupt this */
···
 	}
 }
 
-void request_send(struct fuse_conn *fc, struct fuse_req *req)
+void fuse_request_send(struct fuse_conn *fc, struct fuse_req *req)
 {
 	req->isreply = 1;
 	spin_lock(&fc->lock);
···
 	spin_unlock(&fc->lock);
 }
 
-static void request_send_nowait_locked(struct fuse_conn *fc,
-				       struct fuse_req *req)
+static void fuse_request_send_nowait_locked(struct fuse_conn *fc,
+					    struct fuse_req *req)
 {
 	req->background = 1;
 	fc->num_background++;
···
 	flush_bg_queue(fc);
 }
 
-static void request_send_nowait(struct fuse_conn *fc, struct fuse_req *req)
+static void fuse_request_send_nowait(struct fuse_conn *fc, struct fuse_req *req)
 {
 	spin_lock(&fc->lock);
 	if (fc->connected) {
-		request_send_nowait_locked(fc, req);
+		fuse_request_send_nowait_locked(fc, req);
 		spin_unlock(&fc->lock);
 	} else {
 		req->out.h.error = -ENOTCONN;
···
 	}
 }
 
-void request_send_noreply(struct fuse_conn *fc, struct fuse_req *req)
+void fuse_request_send_noreply(struct fuse_conn *fc, struct fuse_req *req)
 {
 	req->isreply = 0;
-	request_send_nowait(fc, req);
+	fuse_request_send_nowait(fc, req);
 }
 
-void request_send_background(struct fuse_conn *fc, struct fuse_req *req)
+void fuse_request_send_background(struct fuse_conn *fc, struct fuse_req *req)
 {
 	req->isreply = 1;
-	request_send_nowait(fc, req);
+	fuse_request_send_nowait(fc, req);
 }
 
 /*
···
  *
  * fc->connected must have been checked previously
  */
-void request_send_background_locked(struct fuse_conn *fc, struct fuse_req *req)
+void fuse_request_send_background_locked(struct fuse_conn *fc,
+					 struct fuse_req *req)
 {
 	req->isreply = 1;
-	request_send_nowait_locked(fc, req);
+	fuse_request_send_nowait_locked(fc, req);
 }
 
 /*
···
 		BUG_ON(!cs->nr_segs);
 		cs->seglen = cs->iov[0].iov_len;
 		cs->addr = (unsigned long) cs->iov[0].iov_base;
-		cs->iov ++;
-		cs->nr_segs --;
+		cs->iov++;
+		cs->nr_segs--;
 	}
 	down_read(&current->mm->mmap_sem);
 	err = get_user_pages(current, current->mm, cs->addr, 1, cs->write, 0,
···
 		kunmap_atomic(mapaddr, KM_USER1);
 	}
 	while (count) {
-		int err;
-		if (!cs->len && (err = fuse_copy_fill(cs)))
-			return err;
+		if (!cs->len) {
+			int err = fuse_copy_fill(cs);
+			if (err)
+				return err;
+		}
 		if (page) {
 			void *mapaddr = kmap_atomic(page, KM_USER1);
 			void *buf = mapaddr + offset;
···
 static int fuse_copy_one(struct fuse_copy_state *cs, void *val, unsigned size)
 {
 	while (size) {
-		int err;
-		if (!cs->len && (err = fuse_copy_fill(cs)))
-			return err;
+		if (!cs->len) {
+			int err = fuse_copy_fill(cs);
+			if (err)
+				return err;
+		}
 		fuse_copy_do(cs, &val, &size);
 	}
 	return 0;
···
 
 /* Wait until a request is available on the pending list */
 static void request_wait(struct fuse_conn *fc)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	DECLARE_WAITQUEUE(wait, current);
 
···
  */
 static int fuse_read_interrupt(struct fuse_conn *fc, struct fuse_req *req,
 			       const struct iovec *iov, unsigned long nr_segs)
-	__releases(fc->lock)
+__releases(&fc->lock)
 {
 	struct fuse_copy_state cs;
 	struct fuse_in_header ih;
···
 	return err;
 }
 
+static int fuse_notify_poll(struct fuse_conn *fc, unsigned int size,
+			    struct fuse_copy_state *cs)
+{
+	struct fuse_notify_poll_wakeup_out outarg;
+	int err;
+
+	if (size != sizeof(outarg))
+		return -EINVAL;
+
+	err = fuse_copy_one(cs, &outarg, sizeof(outarg));
+	if (err)
+		return err;
+
+	return fuse_notify_poll_wakeup(fc, &outarg);
+}
+
+static int fuse_notify(struct fuse_conn *fc, enum fuse_notify_code code,
+		       unsigned int size, struct fuse_copy_state *cs)
+{
+	switch (code) {
+	case FUSE_NOTIFY_POLL:
+		return fuse_notify_poll(fc, size, cs);
+
+	default:
+		return -EINVAL;
+	}
+}
+
 /* Look up request on processing list by unique ID */
 static struct fuse_req *request_find(struct fuse_conn *fc, u64 unique)
 {
···
 	err = fuse_copy_one(&cs, &oh, sizeof(oh));
 	if (err)
 		goto err_finish;
+
 	err = -EINVAL;
-	if (!oh.unique || oh.error <= -1000 || oh.error > 0 ||
-	    oh.len != nbytes)
+	if (oh.len != nbytes)
+		goto err_finish;
+
+	/*
+	 * Zero oh.unique indicates unsolicited notification message
+	 * and error contains notification code.
+	 */
+	if (!oh.unique) {
+		err = fuse_notify(fc, oh.error, nbytes - sizeof(oh), &cs);
+		fuse_copy_finish(&cs);
+		return err ? err : nbytes;
+	}
+
+	err = -EINVAL;
+	if (oh.error <= -1000 || oh.error > 0)
 		goto err_finish;
 
 	spin_lock(&fc->lock);
···
  * This function releases and reacquires fc->lock
  */
 static void end_requests(struct fuse_conn *fc, struct list_head *head)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	while (!list_empty(head)) {
 		struct fuse_req *req;
···
  * locked).
  */
 static void end_io_requests(struct fuse_conn *fc)
-	__releases(fc->lock) __acquires(fc->lock)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	while (!list_empty(&fc->io)) {
 		struct fuse_req *req =
···
 		wake_up(&req->waitq);
 		if (end) {
 			req->end = NULL;
-			/* The end function will consume this reference */
 			__fuse_get_request(req);
 			spin_unlock(&fc->lock);
 			wait_event(req->waitq, !req->locked);
 			end(fc, req);
+			fuse_put_request(fc, req);
 			spin_lock(&fc->lock);
 		}
 	}

fs/fuse/dir.c

···
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <[e-mail]>
+  Copyright (C) 2001-2008  Miklos Szeredi <[e-mail]>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
···
 		parent = dget_parent(entry);
 		fuse_lookup_init(fc, req, get_node_id(parent->d_inode),
 				 &entry->d_name, &outarg);
-		request_send(fc, req);
+		fuse_request_send(fc, req);
 		dput(parent);
 		err = req->out.h.error;
 		fuse_put_request(fc, req);
···
 				return 0;
 			}
 			spin_lock(&fc->lock);
-			fi->nlookup ++;
+			fi->nlookup++;
 			spin_unlock(&fc->lock);
 		}
 		fuse_put_request(fc, forget_req);
···
 	attr_version = fuse_get_attr_version(fc);
 
 	fuse_lookup_init(fc, req, nodeid, name, outarg);
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	/* Zero nodeid is same as -ENOENT, but with valid timeout */
···
 {
 	fuse_release_fill(ff, nodeid, flags, FUSE_RELEASE);
 	ff->reserved_req->force = 1;
-	request_send(fc, ff->reserved_req);
+	fuse_request_send(fc, ff->reserved_req);
 	fuse_put_request(fc, ff->reserved_req);
 	kfree(ff);
 }
···
 		goto out_put_forget_req;
 
 	err = -ENOMEM;
-	ff = fuse_file_alloc();
+	ff = fuse_file_alloc(fc);
 	if (!ff)
 		goto out_put_request;
 
···
 	req->out.args[0].value = &outentry;
 	req->out.args[1].size = sizeof(outopen);
 	req->out.args[1].value = &outopen;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	if (err) {
 		if (err == -ENOSYS)
···
 	else
 		req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err)
···
 	req->in.numargs = 1;
 	req->in.args[0].size = entry->d_name.len + 1;
 	req->in.args[0].value = entry->d_name.name;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err) {
 		struct inode *inode = entry->d_inode;
 
-		/* Set nlink to zero so the inode can be cleared, if
-                   the inode does have more links this will be
-                   discovered at the next lookup/getattr */
+		/*
+		 * Set nlink to zero so the inode can be cleared, if the inode
+		 * does have more links this will be discovered at the next
+		 * lookup/getattr.
+		 */
 		clear_nlink(inode);
 		fuse_invalidate_attr(inode);
 		fuse_invalidate_attr(dir);
···
 	req->in.numargs = 1;
 	req->in.args[0].size = entry->d_name.len + 1;
 	req->in.args[0].value = entry->d_name.name;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err) {
···
 	req->in.args[1].value = oldent->d_name.name;
 	req->in.args[2].size = newent->d_name.len + 1;
 	req->in.args[2].value = newent->d_name.name;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err) {
···
 	else
 		req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err) {
···
 	req->in.numargs = 1;
 	req->in.args[0].size = sizeof(inarg);
 	req->in.args[0].value = &inarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
···
 	req->num_pages = 1;
 	req->pages[0] = page;
 	fuse_read_fill(req, file, inode, file->f_pos, PAGE_SIZE, FUSE_READDIR);
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	nbytes = req->out.args[0].size;
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
···
 	req->out.numargs = 1;
 	req->out.args[0].size = PAGE_SIZE - 1;
 	req->out.args[0].value = link;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	if (req->out.h.error) {
 		free_page((unsigned long) link);
 		link = ERR_PTR(req->out.h.error);
···
 	else
 		req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err) {
···
 	req->in.args[1].value = name;
 	req->in.args[2].size = size;
 	req->in.args[2].value = value;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
···
 		req->out.args[0].size = sizeof(outarg);
 		req->out.args[0].value = &outarg;
 	}
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	ret = req->out.h.error;
 	if (!ret)
 		ret = size ? req->out.args[0].size : outarg.size;
···
 		req->out.args[0].size = sizeof(outarg);
 		req->out.args[0].value = &outarg;
 	}
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	ret = req->out.h.error;
 	if (!ret)
 		ret = size ? req->out.args[0].size : outarg.size;
···
 	req->in.numargs = 1;
 	req->in.args[0].size = strlen(name) + 1;
 	req->in.args[0].value = name;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {

fs/fuse/file.c

···
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <[e-mail]>
+  Copyright (C) 2001-2008  Miklos Szeredi <[e-mail]>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
···
 	req->out.numargs = 1;
 	req->out.args[0].size = sizeof(*outargp);
 	req->out.args[0].value = outargp;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 
 	return err;
 }
 
-struct fuse_file *fuse_file_alloc(void)
+struct fuse_file *fuse_file_alloc(struct fuse_conn *fc)
 {
 	struct fuse_file *ff;
 	ff = kmalloc(sizeof(struct fuse_file), GFP_KERNEL);
···
 		} else {
 			INIT_LIST_HEAD(&ff->write_entry);
 			atomic_set(&ff->count, 0);
+			spin_lock(&fc->lock);
+			ff->kh = ++fc->khctr;
+			spin_unlock(&fc->lock);
 		}
+		RB_CLEAR_NODE(&ff->polled_node);
+		init_waitqueue_head(&ff->poll_wait);
 	}
 	return ff;
 }
···
 {
 	dput(req->misc.release.dentry);
 	mntput(req->misc.release.vfsmount);
-	fuse_put_request(fc, req);
 }
 
 static void fuse_file_put(struct fuse_file *ff)
···
 		struct inode *inode = req->misc.release.dentry->d_inode;
 		struct fuse_conn *fc = get_fuse_conn(inode);
 		req->end = fuse_release_end;
-		request_send_background(fc, req);
+		fuse_request_send_background(fc, req);
 		kfree(ff);
 	}
 }
···
 
 int fuse_open_common(struct inode *inode, struct file *file, int isdir)
 {
+	struct fuse_conn *fc = get_fuse_conn(inode);
 	struct fuse_open_out outarg;
 	struct fuse_file *ff;
 	int err;
···
 	if (err)
 		return err;
 
-	ff = fuse_file_alloc();
+	ff = fuse_file_alloc(fc);
 	if (!ff)
 		return -ENOMEM;
 
···
 
 		spin_lock(&fc->lock);
 		list_del(&ff->write_entry);
+		if (!RB_EMPTY_NODE(&ff->polled_node))
+			rb_erase(&ff->polled_node, &fc->polled_files);
 		spin_unlock(&fc->lock);
+
+		wake_up_interruptible_sync(&ff->poll_wait);
 		/*
 		 * Normally this will send the RELEASE request,
 		 * however if some asynchronous READ or WRITE requests
···
 	req->in.args[0].size = sizeof(inarg);
 	req->in.args[0].value = &inarg;
 	req->force = 1;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
···
 	req->in.numargs = 1;
 	req->in.args[0].size = sizeof(inarg);
 	req->in.args[0].value = &inarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS) {
···
 		inarg->read_flags |= FUSE_READ_LOCKOWNER;
 		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
 	}
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	return req->out.args[0].size;
 }
 
···
 	}
 	if (req->ff)
 		fuse_file_put(req->ff);
-	fuse_put_request(fc, req);
 }
 
 static void fuse_send_readpages(struct fuse_req *req, struct file *file,
···
 		struct fuse_file *ff = file->private_data;
 		req->ff = fuse_file_get(ff);
 		req->end = fuse_readpages_end;
-		request_send_background(fc, req);
+		fuse_request_send_background(fc, req);
 	} else {
-		request_send(fc, req);
+		fuse_request_send(fc, req);
 		fuse_readpages_end(fc, req);
+		fuse_put_request(fc, req);
 	}
 }
 
···
 		}
 	}
 	req->pages[req->num_pages] = page;
-	req->num_pages ++;
+	req->num_pages++;
 	return 0;
 }
 
···
 		inarg->write_flags |= FUSE_WRITE_LOCKOWNER;
 		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
 	}
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	return req->misc.write.out.size;
 }
 
···
 {
 	__free_page(req->pages[0]);
 	fuse_file_put(req->ff);
-	fuse_put_request(fc, req);
 }
 
 static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req)
···
 
 /* Called under fc->lock, may release and reacquire it */
 static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	struct fuse_inode *fi = get_fuse_inode(req->inode);
 	loff_t size = i_size_read(req->inode);
···
 
 	req->in.args[1].size = inarg->size;
 	fi->writectr++;
-	request_send_background_locked(fc, req);
+	fuse_request_send_background_locked(fc, req);
 	return;
 
  out_free:
 	fuse_writepage_finish(fc, req);
 	spin_unlock(&fc->lock);
 	fuse_writepage_free(fc, req);
+	fuse_put_request(fc, req);
 	spin_lock(&fc->lock);
 }
 
···
  * Called with fc->lock
  */
 void fuse_flush_writepages(struct inode *inode)
+__releases(&fc->lock)
+__acquires(&fc->lock)
 {
 	struct fuse_conn *fc = get_fuse_conn(inode);
 	struct fuse_inode *fi = get_fuse_inode(inode);
···
 	req->out.numargs = 1;
 	req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (!err)
···
 		return PTR_ERR(req);
 
 	fuse_lk_fill(req, file, fl, opcode, pid, flock);
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	/* locking is restartable */
 	if (err == -EINTR)
···
 	req->out.numargs = 1;
 	req->out.args[0].size = sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	fuse_put_request(fc, req);
 	if (err == -ENOSYS)
···
 	return retval;
 }
 
+static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov,
+			unsigned int nr_segs, size_t bytes, bool to_user)
+{
+	struct iov_iter ii;
+	int page_idx = 0;
+
+	if (!bytes)
+		return 0;
+
+	iov_iter_init(&ii, iov, nr_segs, bytes, 0);
+
+	while (iov_iter_count(&ii)) {
+		struct page *page = pages[page_idx++];
+		size_t todo = min_t(size_t, PAGE_SIZE, iov_iter_count(&ii));
+		void *kaddr, *map;
+
+		kaddr = map = kmap(page);
+
+		while (todo) {
+			char __user *uaddr = ii.iov->iov_base + ii.iov_offset;
+			size_t iov_len = ii.iov->iov_len - ii.iov_offset;
+			size_t copy = min(todo, iov_len);
+			size_t left;
+
+			if (!to_user)
+				left = copy_from_user(kaddr, uaddr, copy);
+			else
+				left = copy_to_user(uaddr, kaddr, copy);
+
+			if (unlikely(left))
+				return -EFAULT;
+
+			iov_iter_advance(&ii, copy);
+			todo -= copy;
+			kaddr += copy;
+		}
+
+		kunmap(map);
+	}
+
+	return 0;
+}
+
+/*
+ * For ioctls, there is no generic way to determine how much memory
+ * needs to be read and/or written.  Furthermore, ioctls are allowed
+ * to dereference the passed pointer, so the parameter requires deep
+ * copying but FUSE has no idea whatsoever about what to copy in or
+ * out.
+ *
+ * This is solved by allowing FUSE server to retry ioctl with
+ * necessary in/out iovecs.  Let's assume the ioctl implementation
+ * needs to read in the following structure.
+ *
+ * struct a {
+ *	char	*buf;
+ *	size_t	buflen;
+ * }
+ *
+ * On the first callout to FUSE server, inarg->in_size and
+ * inarg->out_size will be NULL; then, the server completes the ioctl
+ * with FUSE_IOCTL_RETRY set in out->flags, out->in_iovs set to 1 and
+ * the actual iov array to
+ *
+ * { { .iov_base = inarg.arg,	.iov_len = sizeof(struct a) } }
+ *
+ * which tells FUSE to copy in the requested area and retry the ioctl.
+ * On the second round, the server has access to the structure and
+ * from that it can tell what to look for next, so on the invocation,
+ * it sets FUSE_IOCTL_RETRY, out->in_iovs to 2 and iov array to
+ *
+ * { { .iov_base = inarg.arg,	.iov_len = sizeof(struct a)	},
+ *   { .iov_base = a.buf,	.iov_len = a.buflen		} }
+ *
+ * FUSE will copy both struct a and the pointed buffer from the
+ * process doing the ioctl and retry ioctl with both struct a and the
+ * buffer.
+ *
+ * This time, FUSE server has everything it needs and completes ioctl
+ * without FUSE_IOCTL_RETRY which finishes the ioctl call.
+ *
+ * Copying data out works the same way.
+ *
+ * Note that if FUSE_IOCTL_UNRESTRICTED is clear, the kernel
+ * automatically initializes in and out iovs by decoding @cmd with
+ * _IOC_* macros and the server is not allowed to request RETRY.  This
+ * limits ioctl data transfers to well-formed ioctls and is the forced
+ * behavior for all FUSE servers.
+ */
+static long fuse_file_do_ioctl(struct file *file, unsigned int cmd,
+			       unsigned long arg, unsigned int flags)
+{
+	struct inode *inode = file->f_dentry->d_inode;
+	struct fuse_file *ff = file->private_data;
+	struct fuse_conn *fc = get_fuse_conn(inode);
+	struct fuse_ioctl_in inarg = {
+		.fh = ff->fh,
+		.cmd = cmd,
+		.arg = arg,
+		.flags = flags
+	};
+	struct fuse_ioctl_out outarg;
+	struct fuse_req *req = NULL;
+	struct page **pages = NULL;
+	struct page *iov_page = NULL;
+	struct iovec *in_iov = NULL, *out_iov = NULL;
+	unsigned int in_iovs = 0, out_iovs = 0, num_pages = 0, max_pages;
+	size_t in_size, out_size, transferred;
+	int err;
+
+	/* assume all the iovs returned by client always fits in a page */
+	BUILD_BUG_ON(sizeof(struct iovec) * FUSE_IOCTL_MAX_IOV > PAGE_SIZE);
+
+	if (!fuse_allow_task(fc, current))
+		return -EACCES;
+
+	err = -EIO;
+	if (is_bad_inode(inode))
+		goto out;
+
+	err = -ENOMEM;
+	pages = kzalloc(sizeof(pages[0]) * FUSE_MAX_PAGES_PER_REQ, GFP_KERNEL);
+	iov_page = alloc_page(GFP_KERNEL);
+	if (!pages || !iov_page)
+		goto out;
+
+	/*
+	 * If restricted, initialize IO parameters as encoded in @cmd.
+	 * RETRY from server is not allowed.
+	 */
+	if (!(flags & FUSE_IOCTL_UNRESTRICTED)) {
+		struct iovec *iov = page_address(iov_page);
+
+		iov->iov_base = (void __user *)arg;
+		iov->iov_len = _IOC_SIZE(cmd);
+
+		if (_IOC_DIR(cmd) & _IOC_WRITE) {
+			in_iov = iov;
+			in_iovs = 1;
+		}
+
+		if (_IOC_DIR(cmd) & _IOC_READ) {
+			out_iov = iov;
+			out_iovs = 1;
+		}
+	}
+
+ retry:
+	inarg.in_size = in_size = iov_length(in_iov, in_iovs);
+	inarg.out_size = out_size = iov_length(out_iov, out_iovs);
+
+	/*
+	 * Out data can be used either for actual out data or iovs,
+	 * make sure there always is at least one page.
+	 */
+	out_size = max_t(size_t, out_size, PAGE_SIZE);
+	max_pages = DIV_ROUND_UP(max(in_size, out_size), PAGE_SIZE);
+
+	/* make sure there are enough buffer pages and init request with them */
+	err = -ENOMEM;
+	if (max_pages > FUSE_MAX_PAGES_PER_REQ)
+		goto out;
+	while (num_pages < max_pages) {
+		pages[num_pages] = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
+		if (!pages[num_pages])
+			goto out;
+		num_pages++;
+	}
+
+	req = fuse_get_req(fc);
+	if (IS_ERR(req)) {
+		err = PTR_ERR(req);
+		req = NULL;
+		goto out;
+	}
+	memcpy(req->pages, pages, sizeof(req->pages[0]) * num_pages);
+	req->num_pages = num_pages;
+
+	/* okay, let's send it to the client */
+	req->in.h.opcode = FUSE_IOCTL;
+	req->in.h.nodeid = get_node_id(inode);
+	req->in.numargs = 1;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	if (in_size) {
+		req->in.numargs++;
+		req->in.args[1].size = in_size;
+		req->in.argpages = 1;
+
+		err = fuse_ioctl_copy_user(pages, in_iov, in_iovs, in_size,
+					   false);
+		if (err)
+			goto out;
+	}
+
+	req->out.numargs = 2;
+	req->out.args[0].size = sizeof(outarg);
+	req->out.args[0].value = &outarg;
+	req->out.args[1].size = out_size;
+	req->out.argpages = 1;
+	req->out.argvar = 1;
+
+	fuse_request_send(fc, req);
+	err = req->out.h.error;
+	transferred = req->out.args[1].size;
+	fuse_put_request(fc, req);
+	req = NULL;
+	if (err)
+		goto out;
+
+	/* did it ask for retry? */
+	if (outarg.flags & FUSE_IOCTL_RETRY) {
+		char *vaddr;
+
+		/* no retry if in restricted mode */
+		err = -EIO;
+		if (!(flags & FUSE_IOCTL_UNRESTRICTED))
+			goto out;
+
+		in_iovs = outarg.in_iovs;
+		out_iovs = outarg.out_iovs;
+
+		/*
+		 * Make sure things are in boundary, separate checks
+		 * are to protect against overflow.
+		 */
+		err = -ENOMEM;
+		if (in_iovs > FUSE_IOCTL_MAX_IOV ||
+		    out_iovs > FUSE_IOCTL_MAX_IOV ||
+		    in_iovs + out_iovs > FUSE_IOCTL_MAX_IOV)
+			goto out;
+
+		err = -EIO;
+		if ((in_iovs + out_iovs) * sizeof(struct iovec) != transferred)
+			goto out;
+
+		/* okay, copy in iovs and retry */
+		vaddr = kmap_atomic(pages[0], KM_USER0);
+		memcpy(page_address(iov_page), vaddr, transferred);
+		kunmap_atomic(vaddr, KM_USER0);
+
+		in_iov = page_address(iov_page);
+		out_iov = in_iov + in_iovs;
+
+		goto retry;
+	}
+
+	err = -EIO;
+	if (transferred > inarg.out_size)
+		goto out;
+
+	err = fuse_ioctl_copy_user(pages, out_iov, out_iovs, transferred, true);
+ out:
+	if (req)
+		fuse_put_request(fc, req);
+	if (iov_page)
+		__free_page(iov_page);
+	while (num_pages)
+		__free_page(pages[--num_pages]);
+	kfree(pages);
+
+	return err ? err : outarg.result;
+}
+
+static long fuse_file_ioctl(struct file *file, unsigned int cmd,
+			    unsigned long arg)
+{
+	return fuse_file_do_ioctl(file, cmd, arg, 0);
+}
+
+static long fuse_file_compat_ioctl(struct file *file, unsigned int cmd,
+				   unsigned long arg)
+{
+	return fuse_file_do_ioctl(file, cmd, arg, FUSE_IOCTL_COMPAT);
+}
+
+/*
+ * All files which have been polled are linked to RB tree
+ * fuse_conn->polled_files which is indexed by kh.  Walk the tree and
+ * find the matching one.
+ */
+static struct rb_node **fuse_find_polled_node(struct fuse_conn *fc, u64 kh,
+					      struct rb_node **parent_out)
+{
+	struct rb_node **link = &fc->polled_files.rb_node;
+	struct rb_node *last = NULL;
+
+	while (*link) {
+		struct fuse_file *ff;
+
+		last = *link;
+		ff = rb_entry(last, struct fuse_file, polled_node);
+
+		if (kh < ff->kh)
+			link = &last->rb_left;
+		else if (kh > ff->kh)
+			link = &last->rb_right;
+		else
+			return link;
+	}
+
+	if (parent_out)
+		*parent_out = last;
+	return link;
+}
+
+/*
+ * The file is about to be polled.  Make sure it's on the polled_files
+ * RB tree.  Note that files once added to the polled_files tree are
+ * not removed before the file is released.  This is because a file
+ * polled once is likely to be polled again.
+ */
+static void fuse_register_polled_file(struct fuse_conn *fc,
+				      struct fuse_file *ff)
+{
+	spin_lock(&fc->lock);
+	if (RB_EMPTY_NODE(&ff->polled_node)) {
+		struct rb_node **link, *parent;
+
+		link = fuse_find_polled_node(fc, ff->kh, &parent);
+		BUG_ON(*link);
+		rb_link_node(&ff->polled_node, parent, link);
+		rb_insert_color(&ff->polled_node, &fc->polled_files);
+	}
+	spin_unlock(&fc->lock);
+}
+
+static unsigned fuse_file_poll(struct file *file, poll_table *wait)
+{
+	struct inode *inode = file->f_dentry->d_inode;
+	struct fuse_file *ff = file->private_data;
+	struct fuse_conn *fc = get_fuse_conn(inode);
+	struct fuse_poll_in inarg = { .fh = ff->fh, .kh = ff->kh };
+	struct fuse_poll_out outarg;
+	struct fuse_req *req;
+	int err;
+
+	if (fc->no_poll)
+		return DEFAULT_POLLMASK;
+
+	poll_wait(file, &ff->poll_wait, wait);
+
+	/*
+	 * Ask for notification iff there's someone waiting for it.
+	 * The client may ignore the flag and always notify.
+	 */
+	if (waitqueue_active(&ff->poll_wait)) {
+		inarg.flags |= FUSE_POLL_SCHEDULE_NOTIFY;
+		fuse_register_polled_file(fc, ff);
+	}
+
+	req = fuse_get_req(fc);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	req->in.h.opcode = FUSE_POLL;
+	req->in.h.nodeid = get_node_id(inode);
+	req->in.numargs = 1;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	req->out.numargs = 1;
+	req->out.args[0].size = sizeof(outarg);
+	req->out.args[0].value = &outarg;
+	fuse_request_send(fc, req);
+	err = req->out.h.error;
+	fuse_put_request(fc, req);
+
+	if (!err)
+		return outarg.revents;
+	if (err == -ENOSYS) {
+		fc->no_poll = 1;
+		return DEFAULT_POLLMASK;
+	}
+	return POLLERR;
+}
+
+/*
+ * This is called from fuse_handle_notify() on FUSE_NOTIFY_POLL and
+ * wakes up the poll waiters.
+ */
+int fuse_notify_poll_wakeup(struct fuse_conn *fc,
+			    struct fuse_notify_poll_wakeup_out *outarg)
+{
+	u64 kh = outarg->kh;
+	struct rb_node **link;
+
+	spin_lock(&fc->lock);
+
+	link = fuse_find_polled_node(fc, kh, NULL);
+	if (*link) {
+		struct fuse_file *ff;
+
+		ff = rb_entry(*link, struct fuse_file, polled_node);
+		wake_up_interruptible_sync(&ff->poll_wait);
+	}
+
+	spin_unlock(&fc->lock);
+	return 0;
+}
+
 static const struct file_operations fuse_file_operations = {
 	.llseek		= fuse_file_llseek,
 	.read		= do_sync_read,
···
 	.lock		= fuse_file_lock,
 	.flock		= fuse_file_flock,
 	.splice_read	= generic_file_splice_read,
+	.unlocked_ioctl	= fuse_file_ioctl,
+	.compat_ioctl	= fuse_file_compat_ioctl,
+	.poll		= fuse_file_poll,
 };
 
 static const struct file_operations fuse_direct_io_file_operations = {
···
 	.fsync		= fuse_fsync,
 	.lock		= fuse_file_lock,
 	.flock		= fuse_file_flock,
+	.unlocked_ioctl	= fuse_file_ioctl,
+	.compat_ioctl	= fuse_file_compat_ioctl,
+	.poll		= fuse_file_poll,
 	/* no mmap and splice_read */
 };
 

fs/fuse/fuse_i.h

···
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <[e-mail]>
+  Copyright (C) 2001-2008  Miklos Szeredi <[e-mail]>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
···
 #include <linux/backing-dev.h>
 #include <linux/mutex.h>
 #include <linux/rwsem.h>
+#include <linux/rbtree.h>
+#include <linux/poll.h>
 
 /** Max number of pages that can be used in a single read request */
 #define FUSE_MAX_PAGES_PER_REQ 32
···
 	/** Request reserved for flush and release */
 	struct fuse_req *reserved_req;
 
+	/** Kernel file handle guaranteed to be unique */
+	u64 kh;
+
 	/** File handle used by userspace */
 	u64 fh;
 
···
 
 	/** Entry on inode's write_files list */
 	struct list_head write_entry;
+
+	/** RB node to be linked on fuse_conn->polled_files */
+	struct rb_node polled_node;
+
+	/** Wait queue head for poll */
+	wait_queue_head_t poll_wait;
 };
 
 /** One input argument of a request */
···
 	/** The list of requests under I/O */
 	struct list_head io;
 
+	/** The next unique kernel file handle */
+	u64 khctr;
+
+	/** rbtree of fuse_files waiting for poll events indexed by ph */
+	struct rb_root polled_files;
+
 	/** Number of requests currently in the background */
 	unsigned num_background;
 
···
 	/** Connection failed (version mismatch).  Cannot race with
 	    setting other bitfields since it is only set once in INIT
 	    reply, before any other request, and never cleared */
-	unsigned conn_error : 1;
+	unsigned conn_error:1;
 
 	/** Connection successful.  Only set in INIT */
-	unsigned conn_init : 1;
+	unsigned conn_init:1;
 
 	/** Do readpages asynchronously?  Only set in INIT */
-	unsigned async_read : 1;
+	unsigned async_read:1;
 
 	/** Do not send separate SETATTR request before open(O_TRUNC)  */
-	unsigned atomic_o_trunc : 1;
+	unsigned atomic_o_trunc:1;
 
 	/** Filesystem supports NFS exporting.  Only set in INIT */
-	unsigned export_support : 1;
+	unsigned export_support:1;
 
 	/*
 	 * The following bitfields are only for optimization purposes
···
 	 */
 
 	/** Is fsync not implemented by fs? */
-	unsigned no_fsync : 1;
+	unsigned no_fsync:1;
 
 	/** Is fsyncdir not implemented by fs? */
-	unsigned no_fsyncdir : 1;
+	unsigned no_fsyncdir:1;
 
 	/** Is flush not implemented by fs? */
-	unsigned no_flush : 1;
+	unsigned no_flush:1;
 
 	/** Is setxattr not implemented by fs? */
-	unsigned no_setxattr : 1;
+	unsigned no_setxattr:1;
 
 	/** Is getxattr not implemented by fs? */
-	unsigned no_getxattr : 1;
+	unsigned no_getxattr:1;
 
 	/** Is listxattr not implemented by fs? */
-	unsigned no_listxattr : 1;
+	unsigned no_listxattr:1;
 
 	/** Is removexattr not implemented by fs? */
-	unsigned no_removexattr : 1;
+	unsigned no_removexattr:1;
 
 	/** Are file locking primitives not implemented by fs? */
-	unsigned no_lock : 1;
+	unsigned no_lock:1;
 
 	/** Is access not implemented by fs? */
-	unsigned no_access : 1;
+	unsigned no_access:1;
 
 	/** Is create not implemented by fs? */
-	unsigned no_create : 1;
+	unsigned no_create:1;
 
 	/** Is interrupt not implemented by fs? */
-	unsigned no_interrupt : 1;
+	unsigned no_interrupt:1;
 
 	/** Is bmap not implemented by fs? */
-	unsigned no_bmap : 1;
+	unsigned no_bmap:1;
+
+	/** Is poll not implemented by fs? */
+	unsigned no_poll:1;
 
 	/** Do multi-page cached writes */
-	unsigned big_writes : 1;
+	unsigned big_writes:1;
 
 	/** The number of requests waiting for completion */
 	atomic_t num_waiting;
···
 
 	/** Version counter for attribute changes */
 	u64 attr_version;
+
+	/** Called on final put */
+	void (*release)(struct fuse_conn *);
 };
 
 static inline struct fuse_conn *get_fuse_conn_super(struct super_block *sb)
···
  */
 int fuse_open_common(struct inode *inode, struct file *file, int isdir);
 
-struct fuse_file *fuse_file_alloc(void);
+struct fuse_file *fuse_file_alloc(struct fuse_conn *fc);
 void fuse_file_free(struct fuse_file *ff);
 void fuse_finish_open(struct inode *inode, struct file *file,
 		      struct fuse_file *ff, struct fuse_open_out *outarg);
···
  */
 int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
 		      int isdir);
+
+/**
+ * Notify poll wakeup
+ */
+int fuse_notify_poll_wakeup(struct fuse_conn *fc,
+			    struct fuse_notify_poll_wakeup_out *outarg);
 
 /**
  * Initialize file operations on a regular file
···
 /**
  * Send a request (synchronous)
  */
-void request_send(struct fuse_conn *fc, struct fuse_req *req);
+void fuse_request_send(struct fuse_conn *fc, struct fuse_req *req);
 
 /**
  * Send a request with no reply
  */
-void request_send_noreply(struct fuse_conn *fc, struct fuse_req *req);
+void fuse_request_send_noreply(struct fuse_conn *fc, struct fuse_req *req);
 
 /**
  * Send a request in the background
  */
-void request_send_background(struct fuse_conn *fc, struct fuse_req *req);
+void fuse_request_send_background(struct fuse_conn *fc, struct fuse_req *req);
 
-void request_send_background_locked(struct fuse_conn *fc, struct fuse_req *req);
+void fuse_request_send_background_locked(struct fuse_conn *fc,
+					 struct fuse_req *req);
 
 /* Abort all requests */
 void fuse_abort_conn(struct fuse_conn *fc);
···
  * Acquire reference to fuse_conn
  */
 struct fuse_conn *fuse_conn_get(struct fuse_conn *fc);
+
+/**
+ * Initialize fuse_conn
+ */
+int fuse_conn_init(struct fuse_conn *fc, struct super_block *sb);
 
 /**
  * Release reference to fuse_conn

fs/fuse/inode.c

···
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <[e-mail]>
+  Copyright (C) 2001-2008  Miklos Szeredi <[e-mail]>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
···
 	unsigned rootmode;
 	unsigned user_id;
 	unsigned group_id;
-	unsigned fd_present : 1;
-	unsigned rootmode_present : 1;
-	unsigned user_id_present : 1;
-	unsigned group_id_present : 1;
+	unsigned fd_present:1;
+	unsigned rootmode_present:1;
+	unsigned user_id_present:1;
+	unsigned group_id_present:1;
 	unsigned flags;
 	unsigned max_read;
 	unsigned blksize;
···
 	req->in.numargs = 1;
 	req->in.args[0].size = sizeof(struct fuse_forget_in);
 	req->in.args[0].value = inarg;
-	request_send_noreply(fc, req);
+	fuse_request_send_noreply(fc, req);
 }
 
 static void fuse_clear_inode(struct inode *inode)
···
 
 	fi = get_fuse_inode(inode);
 	spin_lock(&fc->lock);
-	fi->nlookup ++;
+	fi->nlookup++;
 	spin_unlock(&fc->lock);
 	fuse_change_attributes(inode, attr, attr_valid, attr_version);
 
···
 		fc->destroy_req = NULL;
 		req->in.h.opcode = FUSE_DESTROY;
 		req->force = 1;
-		request_send(fc, req);
+		fuse_request_send(fc, req);
 		fuse_put_request(fc, req);
 	}
 }
···
 	req->out.args[0].size =
 		fc->minor < 4 ? FUSE_COMPAT_STATFS_SIZE : sizeof(outarg);
 	req->out.args[0].value = &outarg;
-	request_send(fc, req);
+	fuse_request_send(fc, req);
 	err = req->out.h.error;
 	if (!err)
 		convert_fuse_statfs(buf, &outarg.st);
···
 	return 0;
 }
 
-static struct fuse_conn *new_conn(struct super_block *sb)
+int fuse_conn_init(struct fuse_conn *fc, struct super_block *sb)
 {
-	struct fuse_conn *fc;
 	int err;
 
-	fc = kzalloc(sizeof(*fc), GFP_KERNEL);
-	if (fc) {
-		spin_lock_init(&fc->lock);
-		mutex_init(&fc->inst_mutex);
-		atomic_set(&fc->count, 1);
-		init_waitqueue_head(&fc->waitq);
-		init_waitqueue_head(&fc->blocked_waitq);
-		init_waitqueue_head(&fc->reserved_req_waitq);
-		INIT_LIST_HEAD(&fc->pending);
-		INIT_LIST_HEAD(&fc->processing);
-		INIT_LIST_HEAD(&fc->io);
-		INIT_LIST_HEAD(&fc->interrupts);
-		INIT_LIST_HEAD(&fc->bg_queue);
-		atomic_set(&fc->num_waiting, 0);
-		fc->bdi.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
-		fc->bdi.unplug_io_fn = default_unplug_io_fn;
-		/* fuse does it's own writeback accounting */
-		fc->bdi.capabilities = BDI_CAP_NO_ACCT_WB;
-		fc->dev = sb->s_dev;
-		err = bdi_init(&fc->bdi);
-		if (err)
-			goto error_kfree;
-		if (sb->s_bdev) {
-			err = bdi_register(&fc->bdi, NULL, "%u:%u-fuseblk",
-					   MAJOR(fc->dev), MINOR(fc->dev));
-		} else {
-			err = bdi_register_dev(&fc->bdi, fc->dev);
-		}
-		if (err)
-			goto error_bdi_destroy;
-		/*
-		 * For a single fuse filesystem use max 1% of dirty +
-		 * writeback threshold.
-		 *
-		 * This gives about 1M of write buffer for memory maps on a
-		 * machine with 1G and 10% dirty_ratio, which should be more
-		 * than enough.
-		 *
-		 * Privileged users can raise it by writing to
-		 *
-		 *    /sys/class/bdi/<bdi>/max_ratio
-		 */
-		bdi_set_max_ratio(&fc->bdi, 1);
-		fc->reqctr = 0;
-		fc->blocked = 1;
-		fc->attr_version = 1;
-		get_random_bytes(&fc->scramble_key, sizeof(fc->scramble_key));
+	memset(fc, 0, sizeof(*fc));
+	spin_lock_init(&fc->lock);
+	mutex_init(&fc->inst_mutex);
+	atomic_set(&fc->count, 1);
+	init_waitqueue_head(&fc->waitq);
+	init_waitqueue_head(&fc->blocked_waitq);
+	init_waitqueue_head(&fc->reserved_req_waitq);
+	INIT_LIST_HEAD(&fc->pending);
+	INIT_LIST_HEAD(&fc->processing);
+	INIT_LIST_HEAD(&fc->io);
+	INIT_LIST_HEAD(&fc->interrupts);
+	INIT_LIST_HEAD(&fc->bg_queue);
+	INIT_LIST_HEAD(&fc->entry);
+	atomic_set(&fc->num_waiting, 0);
+	fc->bdi.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
+	fc->bdi.unplug_io_fn = default_unplug_io_fn;
+	/* fuse does it's own writeback accounting */
+	fc->bdi.capabilities = BDI_CAP_NO_ACCT_WB;
+	fc->khctr = 0;
+	fc->polled_files = RB_ROOT;
+	fc->dev = sb->s_dev;
+	err = bdi_init(&fc->bdi);
+	if (err)
+		goto error_mutex_destroy;
+	if (sb->s_bdev) {
+		err = bdi_register(&fc->bdi, NULL, "%u:%u-fuseblk",
+				   MAJOR(fc->dev), MINOR(fc->dev));
+	} else {
+		err = bdi_register_dev(&fc->bdi, fc->dev);
 	}
-	return fc;
+	if (err)
+		goto error_bdi_destroy;
+	/*
+	 * For a single fuse filesystem use max 1% of dirty +
+	 * writeback threshold.
+	 *
+	 * This gives about 1M of write buffer for memory maps on a
+	 * machine with 1G and 10% dirty_ratio, which should be more
+	 * than enough.
+	 *
+	 * Privileged users can raise it by writing to
+	 *
+	 *    /sys/class/bdi/<bdi>/max_ratio
+	 */
+	bdi_set_max_ratio(&fc->bdi, 1);
+	fc->reqctr = 0;
+	fc->blocked = 1;
+	fc->attr_version = 1;
+	get_random_bytes(&fc->scramble_key, sizeof(fc->scramble_key));
 
-error_bdi_destroy:
+	return 0;
+
+ error_bdi_destroy:
 	bdi_destroy(&fc->bdi);
-error_kfree:
+ error_mutex_destroy:
 	mutex_destroy(&fc->inst_mutex);
-	kfree(fc);
-	return NULL;
+	return err;
 }
+EXPORT_SYMBOL_GPL(fuse_conn_init);
 
 void fuse_conn_put(struct fuse_conn *fc)
 {
···
 			fuse_request_free(fc->destroy_req);
 		mutex_destroy(&fc->inst_mutex);
 		bdi_destroy(&fc->bdi);
-		kfree(fc);
+		fc->release(fc);
 	}
 }
 
···
 	return fc;
 }
 
-static struct inode *get_root_inode(struct super_block *sb, unsigned mode)
+static struct inode *fuse_get_root_inode(struct super_block *sb, unsigned mode)
 {
 	struct fuse_attr attr;
 	memset(&attr, 0, sizeof(attr));
···
 	return fuse_iget(sb, 1, 0, &attr, 0, 0);
 }
 
-struct fuse_inode_handle
-{
+struct fuse_inode_handle {
 	u64 nodeid;
 	u32 generation;
 };
···
 		fc->max_write = max_t(unsigned, 4096, fc->max_write);
 		fc->conn_init = 1;
 	}
-	fuse_put_request(fc, req);
 	fc->blocked = 0;
 	wake_up_all(&fc->blocked_waitq);
 }
···
 	req->out.args[0].size = sizeof(struct fuse_init_out);
 	req->out.args[0].value = &req->misc.init_out;
 	req->end = process_init_reply;
-	request_send_background(fc, req);
+	fuse_request_send_background(fc, req);
+}
+
+static void fuse_free_conn(struct fuse_conn *fc)
+{
+	kfree(fc);
 }
 
 static int fuse_fill_super(struct super_block *sb, void *data, int silent)
···
 	if (file->f_op != &fuse_dev_operations)
 		return -EINVAL;
 
-	fc = new_conn(sb);
+	fc = kmalloc(sizeof(*fc), GFP_KERNEL);
 	if (!fc)
 		return -ENOMEM;
 
+	err = fuse_conn_init(fc, sb);
+	if (err) {
+		kfree(fc);
+		return err;
+	}
+
+	fc->release = fuse_free_conn;
 	fc->flags = d.flags;
 	fc->user_id = d.user_id;
 	fc->group_id = d.group_id;
···
 	sb->s_fs_info = fc;
 
 	err = -ENOMEM;
-	root = get_root_inode(sb, d.rootmode);
+	root = fuse_get_root_inode(sb, d.rootmode);
 	if (!root)
 		goto err;
 
···
 
 static void fuse_inode_init_once(void *foo)
 {
-	struct inode * inode = foo;
+	struct inode *inode = foo;
 
 	inode_init_once(inode);
 }
···
 {
 	int res;
 
-	printk("fuse init (API version %i.%i)\n",
+	printk(KERN_INFO "fuse init (API version %i.%i)\n",
 	       FUSE_KERNEL_VERSION, FUSE_KERNEL_MINOR_VERSION);
 
 	INIT_LIST_HEAD(&fuse_conn_list);

include/linux/fuse.h

···
 /*
     FUSE: Filesystem in Userspace
-    Copyright (C) 2001-2006  Miklos Szeredi <[e-mail]>
+    Copyright (C) 2001-2008  Miklos Szeredi <[e-mail]>
 
     This program can be distributed under the terms of the GNU GPL.
     See the file COPYING.
···
  *
  * 7.10
  *  - add nonseekable open flag
+ *
+ * 7.11
+ *  - add IOCTL message
+ *  - add unsolicited notification support
+ *  - add POLL message and NOTIFY_POLL notification
  */
 
 #ifndef _LINUX_FUSE_H
 #define _LINUX_FUSE_H
 
-#include <asm/types.h>
-#include <linux/major.h>
+#include <linux/types.h>
 
 /** Version number of this interface */
 #define FUSE_KERNEL_VERSION 7
 
 /** Minor version number of this interface */
-#define FUSE_KERNEL_MINOR_VERSION 10
+#define FUSE_KERNEL_MINOR_VERSION 11
 
 /** The node ID of the root inode */
 #define FUSE_ROOT_ID 1
-
-/** The major number of the fuse character device */
-#define FUSE_MAJOR MISC_MAJOR
-
-/** The minor number of the fuse character device */
-#define FUSE_MINOR 229
 
 /* Make sure all structures are padded to 64bit boundary, so 32bit
    userspace works under 64bit kernels */
···
  */
 #define FUSE_READ_LOCKOWNER	(1 << 1)
 
+/**
+ * Ioctl flags
+ *
+ * FUSE_IOCTL_COMPAT: 32bit compat ioctl on 64bit machine
+ * FUSE_IOCTL_UNRESTRICTED: not restricted to well-formed ioctls, retry allowed
+ * FUSE_IOCTL_RETRY: retry with new iovecs
+ *
+ * FUSE_IOCTL_MAX_IOV: maximum of in_iovecs + out_iovecs
+ */
+#define FUSE_IOCTL_COMPAT	(1 << 0)
+#define FUSE_IOCTL_UNRESTRICTED	(1 << 1)
+#define FUSE_IOCTL_RETRY	(1 << 2)
+
+#define FUSE_IOCTL_MAX_IOV	256
+
+/**
+ * Poll flags
+ *
+ * FUSE_POLL_SCHEDULE_NOTIFY: request poll notify
+ */
+#define FUSE_POLL_SCHEDULE_NOTIFY (1 << 0)
+
 enum fuse_opcode {
 	FUSE_LOOKUP	   = 1,
 	FUSE_FORGET	   = 2,  /* no reply */
···
 	FUSE_INTERRUPT     = 36,
 	FUSE_BMAP          = 37,
 	FUSE_DESTROY       = 38,
+	FUSE_IOCTL         = 39,
+	FUSE_POLL          = 40,
+};
+
+enum fuse_notify_code {
+	FUSE_NOTIFY_POLL   = 1,
+	FUSE_NOTIFY_CODE_MAX,
 };
 
 /* The read buffer is required to be at least 8k, but may be much larger */
···
 
 struct fuse_bmap_out {
 	__u64	block;
+};
+
+struct fuse_ioctl_in {
+	__u64	fh;
+	__u32	flags;
+	__u32	cmd;
+	__u64	arg;
+	__u32	in_size;
+	__u32	out_size;
+};
+
+struct fuse_ioctl_out {
+	__s32	result;
+	__u32	flags;
+	__u32	in_iovs;
+	__u32	out_iovs;
+};
+
+struct fuse_poll_in {
+	__u64	fh;
+	__u64	kh;
+	__u32	flags;
+	__u32   padding;
+};
+
+struct fuse_poll_out {
+	__u32	revents;
+	__u32	padding;
+};
+
+struct fuse_notify_poll_wakeup_out {
+	__u64	kh;
 };
 
 struct fuse_in_header {

include/linux/miscdevice.h

···
 #include <linux/module.h>
 #include <linux/major.h>
 
-#define PSMOUSE_MINOR  1
-#define MS_BUSMOUSE_MINOR 2
-#define ATIXL_BUSMOUSE_MINOR 3
-/*#define AMIGAMOUSE_MINOR 4	FIXME OBSOLETE */
-#define ATARIMOUSE_MINOR 5
-#define SUN_MOUSE_MINOR 6
-#define APOLLO_MOUSE_MINOR 7
-#define PC110PAD_MINOR 9
-/*#define ADB_MOUSE_MINOR 10	FIXME OBSOLETE */
+#define PSMOUSE_MINOR		1
+#define MS_BUSMOUSE_MINOR	2
+#define ATIXL_BUSMOUSE_MINOR	3
+/*#define AMIGAMOUSE_MINOR	4	FIXME OBSOLETE */
+#define ATARIMOUSE_MINOR	5
+#define SUN_MOUSE_MINOR		6
+#define APOLLO_MOUSE_MINOR	7
+#define PC110PAD_MINOR		9
+/*#define ADB_MOUSE_MINOR	10	FIXME OBSOLETE */
 #define WATCHDOG_MINOR		130	/* Watchdog timer     */
 #define TEMP_MINOR		131	/* Temperature Sensor */
-#define RTC_MINOR 135
+#define RTC_MINOR		135
 #define EFI_RTC_MINOR		136	/* EFI Time services */
-#define SUN_OPENPROM_MINOR 139
+#define SUN_OPENPROM_MINOR	139
 #define DMAPI_MINOR		140	/* DMAPI */
-#define NVRAM_MINOR 144
-#define SGI_MMTIMER        153
+#define NVRAM_MINOR		144
+#define SGI_MMTIMER		153
 #define STORE_QUEUE_MINOR	155
-#define I2O_MINOR 166
+#define I2O_MINOR		166
 #define MICROCODE_MINOR		184
-#define MWAVE_MINOR	219		/* ACP/Mwave Modem */
-#define MPT_MINOR	220
-#define MISC_DYNAMIC_MINOR 255
-
-#define TUN_MINOR	     200
-#define	HPET_MINOR	     228
-#define KVM_MINOR            232
+#define TUN_MINOR		200
+#define MWAVE_MINOR		219	/* ACP/Mwave Modem */
+#define MPT_MINOR		220
+#define HPET_MINOR		228
+#define FUSE_MINOR		229
+#define KVM_MINOR		232
+#define MISC_DYNAMIC_MINOR	255
 
 struct device;