file.c

/*
  FUSE: Filesystem in Userspace
  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.
*/

#include "fuse_i.h"

#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/module.h>
#include <linux/compat.h>
#include <linux/swap.h>

static const struct file_operations fuse_direct_io_file_operations;

static int fuse_send_open(struct fuse_conn *fc, u64 nodeid, struct file *file,
			  int opcode, struct fuse_open_out *outargp)
{
	struct fuse_open_in inarg;
	struct fuse_req *req;
	int err;

	req = fuse_get_req_nopages(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);

	memset(&inarg, 0, sizeof(inarg));
	inarg.flags = file->f_flags & ~(O_CREAT | O_EXCL | O_NOCTTY);
	if (!fc->atomic_o_trunc)
		inarg.flags &= ~O_TRUNC;
	req->in.h.opcode = opcode;
	req->in.h.nodeid = nodeid;
	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(*outargp);
	req->out.args[0].value = outargp;
	fuse_request_send(fc, req);
	err = req->out.h.error;
	fuse_put_request(fc, req);

	return err;
}

struct fuse_file *fuse_file_alloc(struct fuse_conn *fc)
{
	struct fuse_file *ff;

	ff = kmalloc(sizeof(struct fuse_file), GFP_KERNEL);
	if (unlikely(!ff))
		return NULL;

	ff->fc = fc;
	ff->reserved_req = fuse_request_alloc(0);
	if (unlikely(!ff->reserved_req)) {
		kfree(ff);
		return NULL;
	}

	INIT_LIST_HEAD(&ff->write_entry);
	atomic_set(&ff->count, 0);
	RB_CLEAR_NODE(&ff->polled_node);
	init_waitqueue_head(&ff->poll_wait);

	spin_lock(&fc->lock);
	ff->kh = ++fc->khctr;
	spin_unlock(&fc->lock);

	return ff;
}

void fuse_file_free(struct fuse_file *ff)
{
	fuse_request_free(ff->reserved_req);
	kfree(ff);
}

struct fuse_file *fuse_file_get(struct fuse_file *ff)
{
	atomic_inc(&ff->count);
	return ff;
}

static void fuse_release_async(struct work_struct *work)
{
	struct fuse_req *req;
	struct fuse_conn *fc;
	struct path path;

	req = container_of(work, struct fuse_req, misc.release.work);
	path = req->misc.release.path;
	fc = get_fuse_conn(path.dentry->d_inode);

	fuse_put_request(fc, req);
	path_put(&path);
}

static void fuse_release_end(struct fuse_conn *fc, struct fuse_req *req)
{
	if (fc->destroy_req) {
		/*
		 * If this is a fuseblk mount, then it's possible that
		 * releasing the path will result in releasing the
		 * super block and sending the DESTROY request.  If
		 * the server is single threaded, this would hang.
		 * For this reason do the path_put() in a separate
		 * thread.
		 */
		atomic_inc(&req->count);
		INIT_WORK(&req->misc.release.work, fuse_release_async);
		schedule_work(&req->misc.release.work);
	} else {
		path_put(&req->misc.release.path);
	}
}

static void fuse_file_put(struct fuse_file *ff, bool sync)
{
	if (atomic_dec_and_test(&ff->count)) {
		struct fuse_req *req = ff->reserved_req;

		if (sync) {
			fuse_request_send(ff->fc, req);
			path_put(&req->misc.release.path);
			fuse_put_request(ff->fc, req);
		} else {
			req->end = fuse_release_end;
			fuse_request_send_background(ff->fc, req);
		}
		kfree(ff);
	}
}

int fuse_do_open(struct fuse_conn *fc, u64 nodeid, struct file *file,
		 bool isdir)
{
	struct fuse_open_out outarg;
	struct fuse_file *ff;
	int err;
	int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN;

	ff = fuse_file_alloc(fc);
	if (!ff)
		return -ENOMEM;

	err = fuse_send_open(fc, nodeid, file, opcode, &outarg);
	if (err) {
		fuse_file_free(ff);
		return err;
	}

	if (isdir)
		outarg.open_flags &= ~FOPEN_DIRECT_IO;

	ff->fh = outarg.fh;
	ff->nodeid = nodeid;
	ff->open_flags = outarg.open_flags;
	file->private_data = fuse_file_get(ff);

	return 0;
}
EXPORT_SYMBOL_GPL(fuse_do_open);

void fuse_finish_open(struct inode *inode, struct file *file)
{
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = get_fuse_conn(inode);

	if (ff->open_flags & FOPEN_DIRECT_IO)
		file->f_op = &fuse_direct_io_file_operations;
	if (!(ff->open_flags & FOPEN_KEEP_CACHE))
		invalidate_inode_pages2(inode->i_mapping);
	if (ff->open_flags & FOPEN_NONSEEKABLE)
		nonseekable_open(inode, file);
	if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) {
		struct fuse_inode *fi = get_fuse_inode(inode);

		spin_lock(&fc->lock);
		fi->attr_version = ++fc->attr_version;
		i_size_write(inode, 0);
		spin_unlock(&fc->lock);
		fuse_invalidate_attr(inode);
	}
}

int fuse_open_common(struct inode *inode, struct file *file, bool isdir)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	int err;

	err = generic_file_open(inode, file);
	if (err)
		return err;

	err = fuse_do_open(fc, get_node_id(inode), file, isdir);
	if (err)
		return err;

	fuse_finish_open(inode, file);

	return 0;
}

static void fuse_prepare_release(struct fuse_file *ff, int flags, int opcode)
{
	struct fuse_conn *fc = ff->fc;
	struct fuse_req *req = ff->reserved_req;
	struct fuse_release_in *inarg = &req->misc.release.in;

	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_all(&ff->poll_wait);

	inarg->fh = ff->fh;
	inarg->flags = flags;
	req->in.h.opcode = opcode;
	req->in.h.nodeid = ff->nodeid;
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(struct fuse_release_in);
	req->in.args[0].value = inarg;
}

void fuse_release_common(struct file *file, int opcode)
{
	struct fuse_file *ff;
	struct fuse_req *req;

	ff = file->private_data;
	if (unlikely(!ff))
		return;

	req = ff->reserved_req;
	fuse_prepare_release(ff, file->f_flags, opcode);

	if (ff->flock) {
		struct fuse_release_in *inarg = &req->misc.release.in;
		inarg->release_flags |= FUSE_RELEASE_FLOCK_UNLOCK;
		inarg->lock_owner = fuse_lock_owner_id(ff->fc,
						       (fl_owner_t) file);
	}
	/* Hold vfsmount and dentry until release is finished */
	path_get(&file->f_path);
	req->misc.release.path = file->f_path;

	/*
	 * Normally this will send the RELEASE request, however if
	 * some asynchronous READ or WRITE requests are outstanding,
	 * the sending will be delayed.
	 *
	 * Make the release synchronous if this is a fuseblk mount,
	 * synchronous RELEASE is allowed (and desirable) in this case
	 * because the server can be trusted not to screw up.
	 */
	fuse_file_put(ff, ff->fc->destroy_req != NULL);
}

static int fuse_open(struct inode *inode, struct file *file)
{
	return fuse_open_common(inode, file, false);
}

static int fuse_release(struct inode *inode, struct file *file)
{
	fuse_release_common(file, FUSE_RELEASE);

	/* return value is ignored by VFS */
	return 0;
}

void fuse_sync_release(struct fuse_file *ff, int flags)
{
	WARN_ON(atomic_read(&ff->count) > 1);
	fuse_prepare_release(ff, flags, FUSE_RELEASE);
	ff->reserved_req->force = 1;
	fuse_request_send(ff->fc, ff->reserved_req);
	fuse_put_request(ff->fc, ff->reserved_req);
	kfree(ff);
}
EXPORT_SYMBOL_GPL(fuse_sync_release);

/*
 * Scramble the ID space with XTEA, so that the value of the files_struct
 * pointer is not exposed to userspace.
 */
u64 fuse_lock_owner_id(struct fuse_conn *fc, fl_owner_t id)
{
	u32 *k = fc->scramble_key;
	u64 v = (unsigned long) id;
	u32 v0 = v;
	u32 v1 = v >> 32;
	u32 sum = 0;
	int i;

	for (i = 0; i < 32; i++) {
		v0 += ((v1 << 4 ^ v1 >> 5) + v1) ^ (sum + k[sum & 3]);
		sum += 0x9E3779B9;
		v1 += ((v0 << 4 ^ v0 >> 5) + v0) ^ (sum + k[sum>>11 & 3]);
	}

	return (u64) v0 + ((u64) v1 << 32);
}

/*
 * Check if page is under writeback
 *
 * This is currently done by walking the list of writepage requests
 * for the inode, which can be pretty inefficient.
 */
static bool fuse_page_is_writeback(struct inode *inode, pgoff_t index)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);
	struct fuse_req *req;
	bool found = false;

	spin_lock(&fc->lock);
	list_for_each_entry(req, &fi->writepages, writepages_entry) {
		pgoff_t curr_index;

		BUG_ON(req->inode != inode);
		curr_index = req->misc.write.in.offset >> PAGE_CACHE_SHIFT;
		if (curr_index == index) {
			found = true;
			break;
		}
	}
	spin_unlock(&fc->lock);

	return found;
}

/*
 * Wait for page writeback to be completed.
 *
 * Since fuse doesn't rely on the VM writeback tracking, this has to
 * use some other means.
 */
static int fuse_wait_on_page_writeback(struct inode *inode, pgoff_t index)
{
	struct fuse_inode *fi = get_fuse_inode(inode);

	wait_event(fi->page_waitq, !fuse_page_is_writeback(inode, index));
	return 0;
}

static int fuse_flush(struct file *file, fl_owner_t id)
{
	struct inode *inode = file->f_path.dentry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_file *ff = file->private_data;
	struct fuse_req *req;
	struct fuse_flush_in inarg;
	int err;

	if (is_bad_inode(inode))
		return -EIO;

	if (fc->no_flush)
		return 0;

	req = fuse_get_req_nofail_nopages(fc, file);
	memset(&inarg, 0, sizeof(inarg));
	inarg.fh = ff->fh;
	inarg.lock_owner = fuse_lock_owner_id(fc, id);
	req->in.h.opcode = FUSE_FLUSH;
	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->force = 1;
	fuse_request_send(fc, req);
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
		fc->no_flush = 1;
		err = 0;
	}
	return err;
}

/*
 * Wait for all pending writepages on the inode to finish.
 *
 * This is currently done by blocking further writes with FUSE_NOWRITE
 * and waiting for all sent writes to complete.
 *
 * This must be called under i_mutex, otherwise the FUSE_NOWRITE usage
 * could conflict with truncation.
 */
static void fuse_sync_writes(struct inode *inode)
{
	fuse_set_nowrite(inode);
	fuse_release_nowrite(inode);
}

int fuse_fsync_common(struct file *file, loff_t start, loff_t end,
		      int datasync, int isdir)
{
	struct inode *inode = file->f_mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_file *ff = file->private_data;
	struct fuse_req *req;
	struct fuse_fsync_in inarg;
	int err;

	if (is_bad_inode(inode))
		return -EIO;

	err = filemap_write_and_wait_range(inode->i_mapping, start, end);
	if (err)
		return err;

	if ((!isdir && fc->no_fsync) || (isdir && fc->no_fsyncdir))
		return 0;

	mutex_lock(&inode->i_mutex);

	/*
	 * Start writeback against all dirty pages of the inode, then
	 * wait for all outstanding writes, before sending the FSYNC
	 * request.
	 */
	err = write_inode_now(inode, 0);
	if (err)
		goto out;

	fuse_sync_writes(inode);

	req = fuse_get_req_nopages(fc);
	if (IS_ERR(req)) {
		err = PTR_ERR(req);
		goto out;
	}

	memset(&inarg, 0, sizeof(inarg));
	inarg.fh = ff->fh;
	inarg.fsync_flags = datasync ? 1 : 0;
	req->in.h.opcode = isdir ? FUSE_FSYNCDIR : FUSE_FSYNC;
	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;
	fuse_request_send(fc, req);
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
		if (isdir)
			fc->no_fsyncdir = 1;
		else
			fc->no_fsync = 1;
		err = 0;
	}
out:
	mutex_unlock(&inode->i_mutex);
	return err;
}

static int fuse_fsync(struct file *file, loff_t start, loff_t end,
		      int datasync)
{
	return fuse_fsync_common(file, start, end, datasync, 0);
}

void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos,
		    size_t count, int opcode)
{
	struct fuse_read_in *inarg = &req->misc.read.in;
	struct fuse_file *ff = file->private_data;

	inarg->fh = ff->fh;
	inarg->offset = pos;
	inarg->size = count;
	inarg->flags = file->f_flags;
	req->in.h.opcode = opcode;
	req->in.h.nodeid = ff->nodeid;
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(struct fuse_read_in);
	req->in.args[0].value = inarg;
	req->out.argvar = 1;
	req->out.numargs = 1;
	req->out.args[0].size = count;
}

static size_t fuse_send_read(struct fuse_req *req, struct file *file,
			     loff_t pos, size_t count, fl_owner_t owner)
{
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = ff->fc;

	fuse_read_fill(req, file, pos, count, FUSE_READ);
	if (owner != NULL) {
		struct fuse_read_in *inarg = &req->misc.read.in;

		inarg->read_flags |= FUSE_READ_LOCKOWNER;
		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
	}
	fuse_request_send(fc, req);
	return req->out.args[0].size;
}

static void fuse_read_update_size(struct inode *inode, loff_t size,
				  u64 attr_ver)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);

	spin_lock(&fc->lock);
	if (attr_ver == fi->attr_version && size < inode->i_size) {
		fi->attr_version = ++fc->attr_version;
		i_size_write(inode, size);
	}
	spin_unlock(&fc->lock);
}

static int fuse_readpage(struct file *file, struct page *page)
{
	struct inode *inode = page->mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	size_t num_read;
	loff_t pos = page_offset(page);
	size_t count = PAGE_CACHE_SIZE;
	u64 attr_ver;
	int err;

	err = -EIO;
	if (is_bad_inode(inode))
		goto out;

	/*
	 * Page writeback can extend beyond the lifetime of the
	 * page-cache page, so make sure we read a properly synced
	 * page.
	 */
	fuse_wait_on_page_writeback(inode, page->index);

	req = fuse_get_req(fc, 1);
	err = PTR_ERR(req);
	if (IS_ERR(req))
		goto out;

	attr_ver = fuse_get_attr_version(fc);

	req->out.page_zeroing = 1;
	req->out.argpages = 1;
	req->num_pages = 1;
	req->pages[0] = page;
	req->page_descs[0].length = count;
	num_read = fuse_send_read(req, file, pos, count, NULL);
	err = req->out.h.error;
	fuse_put_request(fc, req);

	if (!err) {
		/*
		 * Short read means EOF.  If file size is larger, truncate it
		 */
		if (num_read < count)
			fuse_read_update_size(inode, pos + num_read, attr_ver);

		SetPageUptodate(page);
	}

	fuse_invalidate_attr(inode); /* atime changed */
 out:
	unlock_page(page);
	return err;
}

static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
{
	int i;
	size_t count = req->misc.read.in.size;
	size_t num_read = req->out.args[0].size;
	struct address_space *mapping = NULL;

	for (i = 0; mapping == NULL && i < req->num_pages; i++)
		mapping = req->pages[i]->mapping;

	if (mapping) {
		struct inode *inode = mapping->host;

		/*
		 * Short read means EOF. If file size is larger, truncate it
		 */
		if (!req->out.h.error && num_read < count) {
			loff_t pos;

			pos = page_offset(req->pages[0]) + num_read;
			fuse_read_update_size(inode, pos,
					      req->misc.read.attr_ver);
		}
		fuse_invalidate_attr(inode); /* atime changed */
	}

	for (i = 0; i < req->num_pages; i++) {
		struct page *page = req->pages[i];
		if (!req->out.h.error)
			SetPageUptodate(page);
		else
			SetPageError(page);
		unlock_page(page);
		page_cache_release(page);
	}
	if (req->ff)
		fuse_file_put(req->ff, false);
}

static void fuse_send_readpages(struct fuse_req *req, struct file *file)
{
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = ff->fc;
	loff_t pos = page_offset(req->pages[0]);
	size_t count = req->num_pages << PAGE_CACHE_SHIFT;

	req->out.argpages = 1;
	req->out.page_zeroing = 1;
	req->out.page_replace = 1;
	fuse_read_fill(req, file, pos, count, FUSE_READ);
	req->misc.read.attr_ver = fuse_get_attr_version(fc);
	if (fc->async_read) {
		req->ff = fuse_file_get(ff);
		req->end = fuse_readpages_end;
		fuse_request_send_background(fc, req);
	} else {
		fuse_request_send(fc, req);
		fuse_readpages_end(fc, req);
		fuse_put_request(fc, req);
	}
}

struct fuse_fill_data {
	struct fuse_req *req;
	struct file *file;
	struct inode *inode;
	unsigned nr_pages;
};

static int fuse_readpages_fill(void *_data, struct page *page)
{
	struct fuse_fill_data *data = _data;
	struct fuse_req *req = data->req;
	struct inode *inode = data->inode;
	struct fuse_conn *fc = get_fuse_conn(inode);

	fuse_wait_on_page_writeback(inode, page->index);

	if (req->num_pages &&
	    (req->num_pages == FUSE_MAX_PAGES_PER_REQ ||
	     (req->num_pages + 1) * PAGE_CACHE_SIZE > fc->max_read ||
	     req->pages[req->num_pages - 1]->index + 1 != page->index)) {
		int nr_alloc = min_t(unsigned, data->nr_pages,
				     FUSE_MAX_PAGES_PER_REQ);
		fuse_send_readpages(req, data->file);
		data->req = req = fuse_get_req(fc, nr_alloc);
		if (IS_ERR(req)) {
			unlock_page(page);
			return PTR_ERR(req);
		}
	}

	if (WARN_ON(req->num_pages >= req->max_pages)) {
		fuse_put_request(fc, req);
		return -EIO;
	}

	page_cache_get(page);
	req->pages[req->num_pages] = page;
	req->page_descs[req->num_pages].length = PAGE_SIZE;
	req->num_pages++;
	data->nr_pages--;
	return 0;
}

static int fuse_readpages(struct file *file, struct address_space *mapping,
			  struct list_head *pages, unsigned nr_pages)
{
	struct inode *inode = mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_fill_data data;
	int err;
	int nr_alloc = min_t(unsigned, nr_pages, FUSE_MAX_PAGES_PER_REQ);

	err = -EIO;
	if (is_bad_inode(inode))
		goto out;

	data.file = file;
	data.inode = inode;
	data.req = fuse_get_req(fc, nr_alloc);
	data.nr_pages = nr_pages;
	err = PTR_ERR(data.req);
	if (IS_ERR(data.req))
		goto out;

	err = read_cache_pages(mapping, pages, fuse_readpages_fill, &data);
	if (!err) {
		if (data.req->num_pages)
			fuse_send_readpages(data.req, file);
		else
			fuse_put_request(fc, data.req);
	}
out:
	return err;
}

static ssize_t fuse_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
				  unsigned long nr_segs, loff_t pos)
{
	struct inode *inode = iocb->ki_filp->f_mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);

	/*
	 * In auto invalidate mode, always update attributes on read.
	 * Otherwise, only update if we attempt to read past EOF (to ensure
	 * i_size is up to date).
	 */
	if (fc->auto_inval_data ||
	    (pos + iov_length(iov, nr_segs) > i_size_read(inode))) {
		int err;
		err = fuse_update_attributes(inode, NULL, iocb->ki_filp, NULL);
		if (err)
			return err;
	}

	return generic_file_aio_read(iocb, iov, nr_segs, pos);
}

static void fuse_write_fill(struct fuse_req *req, struct fuse_file *ff,
			    loff_t pos, size_t count)
{
	struct fuse_write_in *inarg = &req->misc.write.in;
	struct fuse_write_out *outarg = &req->misc.write.out;

	inarg->fh = ff->fh;
	inarg->offset = pos;
	inarg->size = count;
	req->in.h.opcode = FUSE_WRITE;
	req->in.h.nodeid = ff->nodeid;
	req->in.numargs = 2;
	if (ff->fc->minor < 9)
		req->in.args[0].size = FUSE_COMPAT_WRITE_IN_SIZE;
	else
		req->in.args[0].size = sizeof(struct fuse_write_in);
	req->in.args[0].value = inarg;
	req->in.args[1].size = count;
	req->out.numargs = 1;
	req->out.args[0].size = sizeof(struct fuse_write_out);
	req->out.args[0].value = outarg;
}

static size_t fuse_send_write(struct fuse_req *req, struct file *file,
			      loff_t pos, size_t count, fl_owner_t owner)
{
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = ff->fc;
	struct fuse_write_in *inarg = &req->misc.write.in;

	fuse_write_fill(req, ff, pos, count);
	inarg->flags = file->f_flags;
	if (owner != NULL) {
		inarg->write_flags |= FUSE_WRITE_LOCKOWNER;
		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
	}
	fuse_request_send(fc, req);
	return req->misc.write.out.size;
}

void fuse_write_update_size(struct inode *inode, loff_t pos)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);

	spin_lock(&fc->lock);
	fi->attr_version = ++fc->attr_version;
	if (pos > inode->i_size)
		i_size_write(inode, pos);
	spin_unlock(&fc->lock);
}

static size_t fuse_send_write_pages(struct fuse_req *req, struct file *file,
				    struct inode *inode, loff_t pos,
				    size_t count)
{
	size_t res;
	unsigned offset;
	unsigned i;

	for (i = 0; i < req->num_pages; i++)
		fuse_wait_on_page_writeback(inode, req->pages[i]->index);

	res = fuse_send_write(req, file, pos, count, NULL);

	offset = req->page_descs[0].offset;
	count = res;
	for (i = 0; i < req->num_pages; i++) {
		struct page *page = req->pages[i];

		if (!req->out.h.error && !offset && count >= PAGE_CACHE_SIZE)
			SetPageUptodate(page);

		if (count > PAGE_CACHE_SIZE - offset)
			count -= PAGE_CACHE_SIZE - offset;
		else
			count = 0;
		offset = 0;

		unlock_page(page);
		page_cache_release(page);
	}

	return res;
}

static ssize_t fuse_fill_write_pages(struct fuse_req *req,
			       struct address_space *mapping,
			       struct iov_iter *ii, loff_t pos)
{
	struct fuse_conn *fc = get_fuse_conn(mapping->host);
	unsigned offset = pos & (PAGE_CACHE_SIZE - 1);
	size_t count = 0;
	int err;

	req->in.argpages = 1;
	req->page_descs[0].offset = offset;

	do {
		size_t tmp;
		struct page *page;
		pgoff_t index = pos >> PAGE_CACHE_SHIFT;
		size_t bytes = min_t(size_t, PAGE_CACHE_SIZE - offset,
				     iov_iter_count(ii));

		bytes = min_t(size_t, bytes, fc->max_write - count);

 again:
		err = -EFAULT;
		if (iov_iter_fault_in_readable(ii, bytes))
			break;

		err = -ENOMEM;
		page = grab_cache_page_write_begin(mapping, index, 0);
		if (!page)
			break;

		if (mapping_writably_mapped(mapping))
			flush_dcache_page(page);

		pagefault_disable();
		tmp = iov_iter_copy_from_user_atomic(page, ii, offset, bytes);
		pagefault_enable();
		flush_dcache_page(page);

		mark_page_accessed(page);

		if (!tmp) {
			unlock_page(page);
			page_cache_release(page);
			bytes = min(bytes, iov_iter_single_seg_count(ii));
			goto again;
		}

		err = 0;
		req->pages[req->num_pages] = page;
		req->page_descs[req->num_pages].length = tmp;
		req->num_pages++;

		iov_iter_advance(ii, tmp);
		count += tmp;
		pos += tmp;
		offset += tmp;
		if (offset == PAGE_CACHE_SIZE)
			offset = 0;

		if (!fc->big_writes)
			break;
	} while (iov_iter_count(ii) && count < fc->max_write &&
		 req->num_pages < req->max_pages && offset == 0);

	return count > 0 ? count : err;
}

static inline unsigned fuse_wr_pages(loff_t pos, size_t len)
{
	return min_t(unsigned,
		     ((pos + len - 1) >> PAGE_CACHE_SHIFT) -
		     (pos >> PAGE_CACHE_SHIFT) + 1,
		     FUSE_MAX_PAGES_PER_REQ);
}

static ssize_t fuse_perform_write(struct file *file,
				  struct address_space *mapping,
				  struct iov_iter *ii, loff_t pos)
{
	struct inode *inode = mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
	int err = 0;
	ssize_t res = 0;

	if (is_bad_inode(inode))
		return -EIO;

	do {
		struct fuse_req *req;
		ssize_t count;
		unsigned nr_pages = fuse_wr_pages(pos, iov_iter_count(ii));

		req = fuse_get_req(fc, nr_pages);
		if (IS_ERR(req)) {
			err = PTR_ERR(req);
			break;
		}

		count = fuse_fill_write_pages(req, mapping, ii, pos);
		if (count <= 0) {
			err = count;
		} else {
			size_t num_written;

			num_written = fuse_send_write_pages(req, file, inode,
							    pos, count);
			err = req->out.h.error;
			if (!err) {
				res += num_written;
				pos += num_written;

				/* break out of the loop on short write */
				if (num_written != count)
					err = -EIO;
			}
		}
		fuse_put_request(fc, req);
	} while (!err && iov_iter_count(ii));

	if (res > 0)
		fuse_write_update_size(inode, pos);

	fuse_invalidate_attr(inode);

	return res > 0 ? res : err;
}

static ssize_t fuse_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
				   unsigned long nr_segs, loff_t pos)
{
	struct file *file = iocb->ki_filp;
	struct address_space *mapping = file->f_mapping;
	size_t count = 0;
	size_t ocount = 0;
	ssize_t written = 0;
	ssize_t written_buffered = 0;
	struct inode *inode = mapping->host;
	ssize_t err;
	struct iov_iter i;
	loff_t endbyte = 0;

	WARN_ON(iocb->ki_pos != pos);

	ocount = 0;
	err = generic_segment_checks(iov, &nr_segs, &ocount, VERIFY_READ);
	if (err)
		return err;

	count = ocount;
	sb_start_write(inode->i_sb);
	mutex_lock(&inode->i_mutex);

	/* We can write back this queue in page reclaim */
	current->backing_dev_info = mapping->backing_dev_info;

	err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
	if (err)
		goto out;

	if (count == 0)
		goto out;

	err = file_remove_suid(file);
	if (err)
		goto out;

	err = file_update_time(file);
	if (err)
		goto out;

	if (file->f_flags & O_DIRECT) {
		written = generic_file_direct_write(iocb, iov, &nr_segs,
						    pos, &iocb->ki_pos,
						    count, ocount);
		if (written < 0 || written == count)
			goto out;

		pos += written;
		count -= written;

		iov_iter_init(&i, iov, nr_segs, count, written);
		written_buffered = fuse_perform_write(file, mapping, &i, pos);
		if (written_buffered < 0) {
			err = written_buffered;
			goto out;
		}
		endbyte = pos + written_buffered - 1;

		err = filemap_write_and_wait_range(file->f_mapping, pos,
						   endbyte);
		if (err)
			goto out;

		invalidate_mapping_pages(file->f_mapping,
					 pos >> PAGE_CACHE_SHIFT,
					 endbyte >> PAGE_CACHE_SHIFT);

		written += written_buffered;
		iocb->ki_pos = pos + written_buffered;
	} else {
		iov_iter_init(&i, iov, nr_segs, count, 0);
		written = fuse_perform_write(file, mapping, &i, pos);
		if (written >= 0)
			iocb->ki_pos = pos + written;
	}
out:
	current->backing_dev_info = NULL;
	mutex_unlock(&inode->i_mutex);
	sb_end_write(inode->i_sb);

	return written ? written : err;
}

static void fuse_release_user_pages(struct fuse_req *req, int write)
{
	unsigned i;

	for (i = 0; i < req->num_pages; i++) {
		struct page *page = req->pages[i];
		if (write)
			set_page_dirty_lock(page);
		put_page(page);
	}
}

static inline void fuse_page_descs_length_init(struct fuse_req *req)
{
	int i;

	for (i = 0; i < req->num_pages; i++)
		req->page_descs[i].length = PAGE_SIZE -
			req->page_descs[i].offset;
}

static int fuse_get_user_pages(struct fuse_req *req, struct iov_iter *ii,
			       size_t *nbytesp, int write)
{
	size_t nbytes = *nbytesp;
	size_t frag_size = min(iov_iter_single_seg_count(ii), nbytes);
	unsigned long user_addr;
	unsigned offset;
	int npages;

	user_addr = (unsigned long)ii->iov->iov_base + ii->iov_offset;
	offset = user_addr & ~PAGE_MASK;

	/* Special case for kernel I/O: can copy directly into the buffer */
	if (segment_eq(get_fs(), KERNEL_DS)) {
		if (write)
			req->in.args[1].value = (void *) user_addr;
		else
			req->out.args[0].value = (void *) user_addr;

		iov_iter_advance(ii, frag_size);
		*nbytesp = frag_size;
		return 0;
	}

	nbytes = min_t(size_t, frag_size, FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT);
	npages = (nbytes + offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
	npages = clamp(npages, 1, FUSE_MAX_PAGES_PER_REQ);
	npages = get_user_pages_fast(user_addr, npages, !write, req->pages);
	if (npages < 0)
		return npages;

	req->num_pages = npages;
	req->page_descs[0].offset = offset;
	fuse_page_descs_length_init(req);

	if (write)
		req->in.argpages = 1;
	else
		req->out.argpages = 1;

	nbytes = (req->num_pages << PAGE_SHIFT) - req->page_descs[0].offset;

	if (frag_size < nbytes)
		req->page_descs[req->num_pages - 1].length -=
			nbytes - frag_size;

	*nbytesp = min(frag_size, nbytes);
	iov_iter_advance(ii, *nbytesp);

	return 0;
}

static ssize_t __fuse_direct_io(struct file *file, const struct iovec *iov,
				unsigned long nr_segs, size_t count,
				loff_t *ppos, int write)
{
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = ff->fc;
	size_t nmax = write ? fc->max_write : fc->max_read;
	loff_t pos = *ppos;
	ssize_t res = 0;
	struct fuse_req *req;
	struct iov_iter ii;

	iov_iter_init(&ii, iov, nr_segs, count, 0);

	req = fuse_get_req(fc, FUSE_MAX_PAGES_PER_REQ);
	if (IS_ERR(req))
		return PTR_ERR(req);

	while (count) {
		size_t nres;
		fl_owner_t owner = current->files;
		size_t nbytes = min(count, nmax);
		int err = fuse_get_user_pages(req, &ii, &nbytes, write);
		if (err) {
			res = err;
			break;
		}

		if (write)
			nres = fuse_send_write(req, file, pos, nbytes, owner);
		else
			nres = fuse_send_read(req, file, pos, nbytes, owner);