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

6 files changed, 693 insertions, 171 deletions

diff --git a/fs/fuse/control.c b/fs/fuse/control.c
index 4f3cab321415..99c99dfb0373 100644
--- a/fs/fuse/control.c
+++ b/fs/fuse/control.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -48,11 +48,13 @@ static ssize_t fuse_conn_waiting_read(struct file *file, char __user *buf,
 	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);
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index fba571648a8e..e0c7ada08a1f 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -269,7 +269,7 @@ static void flush_bg_queue(struct fuse_conn *fc)
  * 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;
@@ -293,13 +293,13 @@ static void request_end(struct fuse_conn *fc, struct fuse_req *req)
 	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;
@@ -317,7 +317,8 @@ static void queue_interrupt(struct fuse_conn *fc, struct fuse_req *req)
 }
 
 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 */
@@ -380,7 +381,7 @@ static void request_wait_answer(struct fuse_conn *fc, struct fuse_req *req)
 	}
 }
 
-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);
@@ -399,8 +400,8 @@ void request_send(struct fuse_conn *fc, struct fuse_req *req)
 	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++;
@@ -414,11 +415,11 @@ static void request_send_nowait_locked(struct fuse_conn *fc,
 	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;
@@ -426,16 +427,16 @@ static void request_send_nowait(struct fuse_conn *fc, struct fuse_req *req)
 	}
 }
 
-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);
 }
 
 /*
@@ -443,10 +444,11 @@ void request_send_background(struct fuse_conn *fc, struct fuse_req *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);
 }
 
 /*
@@ -539,8 +541,8 @@ static int fuse_copy_fill(struct fuse_copy_state *cs)
 		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,
@@ -589,9 +591,11 @@ static int fuse_copy_page(struct fuse_copy_state *cs, struct page *page,
 		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;
@@ -631,9 +635,11 @@ static int fuse_copy_pages(struct fuse_copy_state *cs, unsigned nbytes,
 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;
@@ -664,6 +670,8 @@ static int request_pending(struct fuse_conn *fc)
 
 /* 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);
 
@@ -691,7 +699,7 @@ static void request_wait(struct fuse_conn *fc)
  */
 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;
@@ -813,6 +821,34 @@ static ssize_t fuse_dev_read(struct kiocb *iocb, const struct iovec *iov,
 	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)
 {
@@ -876,9 +912,23 @@ static ssize_t fuse_dev_write(struct kiocb *iocb, const struct iovec *iov,
 	err = fuse_copy_one(&cs, &oh, sizeof(oh));
 	if (err)
 		goto err_finish;
+
+	err = -EINVAL;
+	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.unique || oh.error <= -1000 || oh.error > 0 ||
-	    oh.len != nbytes)
+	if (oh.error <= -1000 || oh.error > 0)
 		goto err_finish;
 
 	spin_lock(&fc->lock);
@@ -966,6 +1016,8 @@ static unsigned fuse_dev_poll(struct file *file, poll_table *wait)
  * 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;
@@ -988,7 +1040,8 @@ static void end_requests(struct fuse_conn *fc, struct list_head *head)
  * 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 =
@@ -1002,11 +1055,11 @@ static void end_io_requests(struct fuse_conn *fc)
 		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);
 		}
 	}
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index 95bc22bdd060..fdff346e96fd 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -189,7 +189,7 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
 		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);
@@ -204,7 +204,7 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
 				return 0;
 			}
 			spin_lock(&fc->lock);
-			fi->nlookup ++;
+			fi->nlookup++;
 			spin_unlock(&fc->lock);
 		}
 		fuse_put_request(fc, forget_req);
@@ -283,7 +283,7 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
 	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 */
@@ -369,7 +369,7 @@ static void fuse_sync_release(struct fuse_conn *fc, struct fuse_file *ff,
 {
 	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);
 }
@@ -408,7 +408,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
 		goto out_put_forget_req;
 
 	err = -ENOMEM;
-	ff = fuse_file_alloc();
+	ff = fuse_file_alloc(fc);
 	if (!ff)
 		goto out_put_request;
 
@@ -432,7 +432,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
 	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)
@@ -502,7 +502,7 @@ static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
 	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)
@@ -631,15 +631,17 @@ static int fuse_unlink(struct inode *dir, struct dentry *entry)
 	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);
@@ -662,7 +664,7 @@ static int fuse_rmdir(struct inode *dir, struct dentry *entry)
 	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) {
@@ -695,7 +697,7 @@ static int fuse_rename(struct inode *olddir, struct dentry *oldent,
 	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) {
@@ -811,7 +813,7 @@ static int fuse_do_getattr(struct inode *inode, struct kstat *stat,
 	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) {
@@ -911,7 +913,7 @@ static int fuse_access(struct inode *inode, int mask)
 	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) {
@@ -1033,7 +1035,7 @@ static int fuse_readdir(struct file *file, void *dstbuf, filldir_t filldir)
 	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);
@@ -1067,7 +1069,7 @@ static char *read_link(struct dentry *dentry)
 	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);
@@ -1273,7 +1275,7 @@ static int fuse_do_setattr(struct dentry *entry, struct iattr *attr,
 	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) {
@@ -1367,7 +1369,7 @@ static int fuse_setxattr(struct dentry *entry, const char *name,
 	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) {
@@ -1413,7 +1415,7 @@ static ssize_t fuse_getxattr(struct dentry *entry, const char *name,
 		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;
@@ -1463,7 +1465,7 @@ static ssize_t fuse_listxattr(struct dentry *entry, char *list, size_t 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;
@@ -1496,7 +1498,7 @@ static int fuse_removexattr(struct dentry *entry, const char *name)
 	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) {
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index 4c9ee7011265..e8162646a9b5 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -1,6 +1,6 @@
 /*
   FUSE: Filesystem in Userspace
-  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
+  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
 
   This program can be distributed under the terms of the GNU GPL.
   See the file COPYING.
@@ -39,14 +39,14 @@ static int fuse_send_open(struct inode *inode, struct file *file, int isdir,
 	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);
@@ -58,7 +58,12 @@ struct fuse_file *fuse_file_alloc(void)
 		} 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;
 }
@@ -79,7 +84,6 @@ static void fuse_release_end(struct fuse_conn *fc, struct fuse_req *req)
 {
 	dput(req->misc.release.dentry);
 	mntput(req->misc.release.vfsmount);
-	fuse_put_request(fc, req);
 }
 
 static void fuse_file_put(struct fuse_file *ff)
@@ -89,7 +93,7 @@ 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);
 	}
 }
@@ -109,6 +113,7 @@ void fuse_finish_open(struct inode *inode, struct file *file,
 
 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;
@@ -121,7 +126,7 @@ int fuse_open_common(struct inode *inode, struct file *file, int isdir)
 	if (err)
 		return err;
 
-	ff = fuse_file_alloc();
+	ff = fuse_file_alloc(fc);
 	if (!ff)
 		return -ENOMEM;
 
@@ -167,7 +172,11 @@ int fuse_release_common(struct inode *inode, struct file *file, int isdir)
 
 		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
@@ -280,7 +289,7 @@ static int fuse_flush(struct file *file, fl_owner_t id)
 	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) {
@@ -344,7 +353,7 @@ int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
 	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) {
@@ -396,7 +405,7 @@ static size_t fuse_send_read(struct fuse_req *req, struct file *file,
 		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;
 }
 
@@ -493,7 +502,6 @@ static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
 	}
 	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,
@@ -509,10 +517,11 @@ 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);
 	}
 }
 
@@ -543,7 +552,7 @@ static int fuse_readpages_fill(void *_data, struct page *page)
 		}
 	}
 	req->pages[req->num_pages] = page;
-	req->num_pages ++;
+	req->num_pages++;
 	return 0;
 }
 
@@ -636,7 +645,7 @@ static size_t fuse_send_write(struct fuse_req *req, struct file *file,
 		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;
 }
 
@@ -1042,7 +1051,6 @@ static void fuse_writepage_free(struct fuse_conn *fc, struct fuse_req *req)
 {
 	__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)
@@ -1060,6 +1068,8 @@ 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);
@@ -1079,13 +1089,14 @@ static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
 
 	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);
 }
 
@@ -1096,6 +1107,8 @@ static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
  * 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);
@@ -1325,7 +1338,7 @@ static int fuse_getlk(struct file *file, struct file_lock *fl)
 	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)
@@ -1357,7 +1370,7 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock)
 		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)
@@ -1433,7 +1446,7 @@ static sector_t fuse_bmap(struct address_space *mapping, sector_t block)
 	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)
@@ -1470,6 +1483,406 @@ static loff_t fuse_file_llseek(struct file *file, loff_t offset, int origin)
 	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,