diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index 254df56b847b..f3f783dc4f75 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -180,6 +180,8 @@ u64 fuse_get_attr_version(struct fuse_conn *fc)
 static int fuse_dentry_revalidate(struct dentry *entry, unsigned int flags)
 {
 	struct inode *inode;
+	struct dentry *parent;
+	struct fuse_conn *fc;
 
 	inode = ACCESS_ONCE(entry->d_inode);
 	if (inode && is_bad_inode(inode))
@@ -187,10 +189,8 @@ static int fuse_dentry_revalidate(struct dentry *entry, unsigned int flags)
 	else if (fuse_dentry_time(entry) < get_jiffies_64()) {
 		int err;
 		struct fuse_entry_out outarg;
-		struct fuse_conn *fc;
 		struct fuse_req *req;
 		struct fuse_forget_link *forget;
-		struct dentry *parent;
 		u64 attr_version;
 
 		/* For negative dentries, always do a fresh lookup */
@@ -241,8 +241,14 @@ static int fuse_dentry_revalidate(struct dentry *entry, unsigned int flags)
 				       entry_attr_timeout(&outarg),
 				       attr_version);
 		fuse_change_entry_timeout(entry, &outarg);
+	} else if (inode) {
+		fc = get_fuse_conn(inode);
+		if (fc->readdirplus_auto) {
+			parent = dget_parent(entry);
+			fuse_advise_use_readdirplus(parent->d_inode);
+			dput(parent);
+		}
 	}
-	fuse_advise_use_readdirplus(inode);
 	return 1;
 }
 
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index d1c9b85b3f58..e570081f9f76 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -16,6 +16,7 @@
 #include <linux/compat.h>
 #include <linux/swap.h>
 #include <linux/aio.h>
+#include <linux/falloc.h>
 
 static const struct file_operations fuse_direct_io_file_operations;
 
@@ -1278,7 +1279,10 @@ ssize_t fuse_direct_io(struct fuse_io_priv *io, const struct iovec *iov,
 
 	iov_iter_init(&ii, iov, nr_segs, count, 0);
 
-	req = fuse_get_req(fc, fuse_iter_npages(&ii));
+	if (io->async)
+		req = fuse_get_req_for_background(fc, fuse_iter_npages(&ii));
+	else
+		req = fuse_get_req(fc, fuse_iter_npages(&ii));
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
@@ -1314,7 +1318,11 @@ ssize_t fuse_direct_io(struct fuse_io_priv *io, const struct iovec *iov,
 			break;
 		if (count) {
 			fuse_put_request(fc, req);
-			req = fuse_get_req(fc, fuse_iter_npages(&ii));
+			if (io->async)
+				req = fuse_get_req_for_background(fc,
+					fuse_iter_npages(&ii));
+			else
+				req = fuse_get_req(fc, fuse_iter_npages(&ii));
 			if (IS_ERR(req))
 				break;
 		}
@@ -2365,6 +2373,11 @@ static void fuse_do_truncate(struct file *file)
 	fuse_do_setattr(inode, &attr, file);
 }
 
+static inline loff_t fuse_round_up(loff_t off)
+{
+	return round_up(off, FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT);
+}
+
 static ssize_t
 fuse_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 			loff_t offset, unsigned long nr_segs)
@@ -2372,6 +2385,7 @@ fuse_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	ssize_t ret = 0;
 	struct file *file = iocb->ki_filp;
 	struct fuse_file *ff = file->private_data;
+	bool async_dio = ff->fc->async_dio;
 	loff_t pos = 0;
 	struct inode *inode;
 	loff_t i_size;
@@ -2383,10 +2397,10 @@ fuse_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	i_size = i_size_read(inode);
 
 	/* optimization for short read */
-	if (rw != WRITE && offset + count > i_size) {
+	if (async_dio && rw != WRITE && offset + count > i_size) {
 		if (offset >= i_size)
 			return 0;
-		count = i_size - offset;
+		count = min_t(loff_t, count, fuse_round_up(i_size - offset));
 	}
 
 	io = kmalloc(sizeof(struct fuse_io_priv), GFP_KERNEL);
@@ -2404,7 +2418,7 @@ fuse_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	 * By default, we want to optimize all I/Os with async request
 	 * submission to the client filesystem if supported.
 	 */
-	io->async = ff->fc->async_dio;
+	io->async = async_dio;
 	io->iocb = iocb;
 
 	/*
@@ -2412,7 +2426,7 @@ fuse_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	 * to wait on real async I/O requests, so we must submit this request
 	 * synchronously.
 	 */
-	if (!is_sync_kiocb(iocb) && (offset + count > i_size))
+	if (!is_sync_kiocb(iocb) && (offset + count > i_size) && rw == WRITE)
 		io->async = false;
 
 	if (rw == WRITE)
@@ -2424,7 +2438,7 @@ fuse_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 		fuse_aio_complete(io, ret < 0 ? ret : 0, -1);
 
 		/* we have a non-extending, async request, so return */
-		if (ret > 0 && !is_sync_kiocb(iocb))
+		if (!is_sync_kiocb(iocb))
 			return -EIOCBQUEUED;
 
 		ret = wait_on_sync_kiocb(iocb);
@@ -2446,6 +2460,7 @@ static long fuse_file_fallocate(struct file *file, int mode, loff_t offset,
 				loff_t length)
 {
 	struct fuse_file *ff = file->private_data;
+	struct inode *inode = file->f_inode;
 	struct fuse_conn *fc = ff->fc;
 	struct fuse_req *req;
 	struct fuse_fallocate_in inarg = {
@@ -2459,9 +2474,16 @@ static long fuse_file_fallocate(struct file *file, int mode, loff_t offset,
 	if (fc->no_fallocate)
 		return -EOPNOTSUPP;
 
+	if (mode & FALLOC_FL_PUNCH_HOLE) {
+		mutex_lock(&inode->i_mutex);
+		fuse_set_nowrite(inode);
+	}
+
 	req = fuse_get_req_nopages(fc);
-	if (IS_ERR(req))
-		return PTR_ERR(req);
+	if (IS_ERR(req)) {
+		err = PTR_ERR(req);
+		goto out;
+	}
 
 	req->in.h.opcode = FUSE_FALLOCATE;
 	req->in.h.nodeid = ff->nodeid;
@@ -2476,6 +2498,24 @@ static long fuse_file_fallocate(struct file *file, int mode, loff_t offset,
 	}
 	fuse_put_request(fc, req);
 
+	if (err)
+		goto out;
+
+	/* we could have extended the file */
+	if (!(mode & FALLOC_FL_KEEP_SIZE))
+		fuse_write_update_size(inode, offset + length);
+
+	if (mode & FALLOC_FL_PUNCH_HOLE)
+		truncate_pagecache_range(inode, offset, offset + length - 1);
+
+	fuse_invalidate_attr(inode);
+
+out:
+	if (mode & FALLOC_FL_PUNCH_HOLE) {
+		fuse_release_nowrite(inode);
+		mutex_unlock(&inode->i_mutex);
+	}
+
 	return err;
 }
 
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index 6201f81e4d3a..9a0cdde14a08 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -867,10 +867,11 @@ static void process_init_reply(struct fuse_conn *fc, struct fuse_req *req)
 				fc->dont_mask = 1;
 			if (arg->flags & FUSE_AUTO_INVAL_DATA)
 				fc->auto_inval_data = 1;
-			if (arg->flags & FUSE_DO_READDIRPLUS)
+			if (arg->flags & FUSE_DO_READDIRPLUS) {
 				fc->do_readdirplus = 1;
-			if (arg->flags & FUSE_READDIRPLUS_AUTO)
-				fc->readdirplus_auto = 1;
+				if (arg->flags & FUSE_READDIRPLUS_AUTO)
+					fc->readdirplus_auto = 1;
+			}
 			if (arg->flags & FUSE_ASYNC_DIO)
 				fc->async_dio = 1;
 		} else {