Index: uspace/srv/fs/ext4fs/Makefile
===================================================================
--- uspace/srv/fs/ext4fs/Makefile	(revision 4cdac6872dde2afba2d079b508504d95a25f4094)
+++ uspace/srv/fs/ext4fs/Makefile	(revision 4cdac6872dde2afba2d079b508504d95a25f4094)
@@ -0,0 +1,38 @@
+#
+# Copyright (c) 2012 Frantisek Princ
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions
+# are met:
+#
+# - Redistributions of source code must retain the above copyright
+#   notice, this list of conditions and the following disclaimer.
+# - Redistributions in binary form must reproduce the above copyright
+#   notice, this list of conditions and the following disclaimer in the
+#   documentation and/or other materials provided with the distribution.
+# - The name of the author may not be used to endorse or promote products
+#   derived from this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+# OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+# IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+# THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+#
+
+USPACE_PREFIX = ../../..
+LIBS = $(LIBBLOCK_PREFIX)/libblock.a $(LIBFS_PREFIX)/libfs.a $(LIBEXT4_PREFIX)/libext4.a $(LIBPOSIX_PREFIX)/libposix.a 
+EXTRA_CFLAGS += -I$(LIBBLOCK_PREFIX) -I$(LIBFS_PREFIX) -I$(LIBEXT4_PREFIX) -I$(LIBPOSIX_PREFIX)
+BINARY = ext4fs
+
+SOURCES = \
+	ext4fs.c \
+	ext4fs_ops.c
+
+include $(USPACE_PREFIX)/Makefile.common
Index: uspace/srv/fs/ext4fs/ext4fs.c
===================================================================
--- uspace/srv/fs/ext4fs/ext4fs.c	(revision 4cdac6872dde2afba2d079b508504d95a25f4094)
+++ uspace/srv/fs/ext4fs/ext4fs.c	(revision 4cdac6872dde2afba2d079b508504d95a25f4094)
@@ -0,0 +1,101 @@
+/*
+ * Copyright (c) 2012 Frantisek Princ
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * - Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in the
+ *   documentation and/or other materials provided with the distribution.
+ * - The name of the author may not be used to endorse or promote products
+ *   derived from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/** @addtogroup fs
+ * @{
+ */ 
+
+/**
+ * @file	ext4fs.c
+ * @brief	EXT4 file system driver for HelenOS.
+ */
+
+#include <async.h>
+#include <errno.h>
+#include <libfs.h>
+#include <ns.h>
+#include <stdio.h>
+#include <task.h>
+#include <ipc/services.h>
+#include "ext4fs.h"
+#include "../../vfs/vfs.h"
+
+#define NAME	"ext4fs"
+
+vfs_info_t ext4fs_vfs_info = {
+	.name = NAME,
+	.instance = 0,
+};
+
+/**
+ * Entry point of ext4fs server.
+ * Initialize data structures and IPC, then accepts connections in server mode.
+ */
+int main(int argc, char **argv)
+{
+	printf(NAME ": HelenOS EXT4 file system server\n");
+
+	if (argc == 3) {
+		if (!str_cmp(argv[1], "--instance"))
+			ext4fs_vfs_info.instance = strtol(argv[2], NULL, 10);
+		else {
+			printf(NAME " Unrecognized parameters\n");
+			return -1;
+		}
+	}
+
+	async_sess_t *vfs_sess = service_connect_blocking(EXCHANGE_SERIALIZE,
+		SERVICE_VFS, 0, 0);
+	if (!vfs_sess) {
+		printf(NAME ": failed to connect to VFS\n");
+		return -1;
+	}
+
+	int rc = ext4fs_global_init();
+	if (rc != EOK) {
+		printf(NAME ": Failed global initialization\n");
+		return 1;
+	}
+
+	rc = fs_register(vfs_sess, &ext4fs_vfs_info, &ext4fs_ops,
+	    &ext4fs_libfs_ops);
+	if (rc != EOK) {
+		fprintf(stdout, NAME ": Failed to register fs (%d)\n", rc);
+		return 1;
+	}
+
+	printf(NAME ": Accepting connections\n");
+	task_retval(0);
+	async_manager();
+	/* not reached */
+	return 0;
+}
+
+/**
+ * @}
+ */ 
Index: uspace/srv/fs/ext4fs/ext4fs.h
===================================================================
--- uspace/srv/fs/ext4fs/ext4fs.h	(revision 4cdac6872dde2afba2d079b508504d95a25f4094)
+++ uspace/srv/fs/ext4fs/ext4fs.h	(revision 4cdac6872dde2afba2d079b508504d95a25f4094)
@@ -0,0 +1,49 @@
+/*
+ * Copyright (c) 2012 Frantisek Princ
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * - Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in the
+ *   documentation and/or other materials provided with the distribution.
+ * - The name of the author may not be used to endorse or promote products
+ *   derived from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/** @addtogroup fs
+ * @{
+ */ 
+
+#ifndef EXT4FS_EXT4FS_H_
+#define EXT4FS_EXT4FS_H_
+
+#include <libfs.h>
+
+extern vfs_out_ops_t ext4fs_ops;
+extern libfs_ops_t ext4fs_libfs_ops;
+
+extern int ext4fs_global_init(void);
+extern int ext4fs_global_fini(void);
+
+
+#endif
+
+/**
+ * @}
+ */
Index: uspace/srv/fs/ext4fs/ext4fs_ops.c
===================================================================
--- uspace/srv/fs/ext4fs/ext4fs_ops.c	(revision 4cdac6872dde2afba2d079b508504d95a25f4094)
+++ uspace/srv/fs/ext4fs/ext4fs_ops.c	(revision 4cdac6872dde2afba2d079b508504d95a25f4094)
@@ -0,0 +1,1524 @@
+/*
+ * Copyright (c) 2012 Frantisek Princ
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * - Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in the
+ *   documentation and/or other materials provided with the distribution.
+ * - The name of the author may not be used to endorse or promote products
+ *   derived from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/** @addtogroup fs
+ * @{
+ */ 
+
+/**
+ * @file	ext4fs_ops.c
+ * @brief	VFS operations for EXT4 filesystem.
+ */
+
+#include <errno.h>
+#include <fibril_synch.h>
+#include <libext4.h>
+#include <libfs.h>
+#include <macros.h>
+#include <malloc.h>
+#include <string.h>
+#include <adt/hash_table.h>
+#include <ipc/loc.h>
+#include "ext4fs.h"
+#include "../../vfs/vfs.h"
+
+#define EXT4FS_NODE(node)	((node) ? (ext4fs_node_t *) (node)->data : NULL)
+
+#define OPEN_NODES_KEYS 2
+#define OPEN_NODES_DEV_HANDLE_KEY 0
+#define OPEN_NODES_INODE_KEY 1
+#define OPEN_NODES_BUCKETS 256
+
+/**
+ * Type for holding an instance of mounted partition.
+ */
+typedef struct ext4fs_instance {
+	link_t link;
+	service_id_t service_id;
+	ext4_filesystem_t *filesystem;
+	unsigned int open_nodes_count;
+} ext4fs_instance_t;
+
+/**
+ * Type for wrapping common fs_node and add some useful pointers.
+ */
+typedef struct ext4fs_node {
+	ext4fs_instance_t *instance;
+	ext4_inode_ref_t *inode_ref;
+	fs_node_t *fs_node;
+	link_t link;
+	unsigned int references;
+} ext4fs_node_t;
+
+/* Forward declarations of auxiliary functions */
+
+static int ext4fs_read_directory(ipc_callid_t, aoff64_t, size_t,
+    ext4fs_instance_t *, ext4_inode_ref_t *, size_t *);
+static int ext4fs_read_file(ipc_callid_t, aoff64_t, size_t, ext4fs_instance_t *,
+    ext4_inode_ref_t *, size_t *);
+static bool ext4fs_is_dots(const uint8_t *, size_t);
+static int ext4fs_instance_get(service_id_t, ext4fs_instance_t **);
+static int ext4fs_node_get_core(fs_node_t **, ext4fs_instance_t *, fs_index_t);
+static int ext4fs_node_put_core(ext4fs_node_t *);
+
+/* Forward declarations of EXT4 libfs operations. */
+
+static int ext4fs_root_get(fs_node_t **, service_id_t);
+static int ext4fs_match(fs_node_t **, fs_node_t *, const char *);
+static int ext4fs_node_get(fs_node_t **, service_id_t, fs_index_t);
+static int ext4fs_node_open(fs_node_t *);
+static int ext4fs_node_put(fs_node_t *);
+static int ext4fs_create_node(fs_node_t **, service_id_t, int);
+static int ext4fs_destroy_node(fs_node_t *);
+static int ext4fs_link(fs_node_t *, fs_node_t *, const char *);
+static int ext4fs_unlink(fs_node_t *, fs_node_t *, const char *);
+static int ext4fs_has_children(bool *, fs_node_t *);
+static fs_index_t ext4fs_index_get(fs_node_t *);
+static aoff64_t ext4fs_size_get(fs_node_t *);
+static unsigned ext4fs_lnkcnt_get(fs_node_t *);
+static bool ext4fs_is_directory(fs_node_t *);
+static bool ext4fs_is_file(fs_node_t *node);
+static service_id_t ext4fs_service_get(fs_node_t *node);
+
+/* Static variables */
+
+static LIST_INITIALIZE(instance_list);
+static FIBRIL_MUTEX_INITIALIZE(instance_list_mutex);
+static hash_table_t open_nodes;
+static FIBRIL_MUTEX_INITIALIZE(open_nodes_lock);
+
+/* Hash table interface for open nodes hash table */
+static hash_index_t open_nodes_hash(unsigned long key[])
+{
+	/* TODO: This is very simple and probably can be improved */
+	return key[OPEN_NODES_INODE_KEY] % OPEN_NODES_BUCKETS;
+}
+
+/** Compare given item with values in hash table.
+ *
+ * @return	bool result of compare operation
+ */
+static int open_nodes_compare(unsigned long key[], hash_count_t keys,
+    link_t *item)
+{
+	ext4fs_node_t *enode = hash_table_get_instance(item, ext4fs_node_t, link);
+	assert(keys > 0);
+	if (enode->instance->service_id !=
+	    ((service_id_t) key[OPEN_NODES_DEV_HANDLE_KEY])) {
+		return false;
+	}
+	if (keys == 1) {
+		return true;
+	}
+	assert(keys == 2);
+	return (enode->inode_ref->index == key[OPEN_NODES_INODE_KEY]);
+}
+
+/** Empty callback to correct hash table initialization.
+ *
+ */
+static void open_nodes_remove_cb(link_t *link)
+{
+	/* We don't use remove callback for this hash table */
+}
+
+static hash_table_operations_t open_nodes_ops = {
+	.hash = open_nodes_hash,
+	.compare = open_nodes_compare,
+	.remove_callback = open_nodes_remove_cb,
+};
+
+
+/** Basic initialization of the driver.
+ *
+ * There is only needed to create hash table for storing open nodes.
+ *
+ * @return	error code
+ */
+int ext4fs_global_init(void)
+{
+	if (!hash_table_create(&open_nodes, OPEN_NODES_BUCKETS,
+	    OPEN_NODES_KEYS, &open_nodes_ops)) {
+		return ENOMEM;
+	}
+	return EOK;
+}
+
+/* Finalization of the driver.
+ *
+ * There is only needed to destroy hash table.
+ *
+ * @return	error code
+ */
+int ext4fs_global_fini(void)
+{
+	hash_table_destroy(&open_nodes);
+	return EOK;
+}
+
+
+/*
+ * EXT4 libfs operations.
+ */
+
+/** Get instance from internal table by service_id.
+ *
+ * @param service_id	device identifier
+ * @param inst		output instance if successful operation
+ * @return		error code
+ */
+int ext4fs_instance_get(service_id_t service_id, ext4fs_instance_t **inst)
+{
+	fibril_mutex_lock(&instance_list_mutex);
+
+	if (list_empty(&instance_list)) {
+		fibril_mutex_unlock(&instance_list_mutex);
+		return EINVAL;
+	}
+
+	ext4fs_instance_t *tmp;
+	list_foreach(instance_list, link) {
+		tmp = list_get_instance(link, ext4fs_instance_t, link);
+
+		if (tmp->service_id == service_id) {
+			*inst = tmp;
+			fibril_mutex_unlock(&instance_list_mutex);
+			return EOK;
+		}
+	}
+
+	fibril_mutex_unlock(&instance_list_mutex);
+	return EINVAL;
+}
+
+
+/** Get root node of filesystem specified by service_id.
+ * 
+ * @param rfn		output pointer to loaded node
+ * @param service_id	device to load root node from
+ * @return		error code
+ */
+int ext4fs_root_get(fs_node_t **rfn, service_id_t service_id)
+{
+	return ext4fs_node_get(rfn, service_id, EXT4_INODE_ROOT_INDEX);
+}
+
+/** Check if specified name (component) matches with any directory entry.
+ * 
+ * If match is found, load and return matching node.
+ *
+ * @param rfn		output pointer to node if operation successful
+ * @param pfn		parent directory node
+ * @param component	name to check directory for
+ * @return 		error code
+ */
+int ext4fs_match(fs_node_t **rfn, fs_node_t *pfn, const char *component)
+{
+	int rc;
+
+	ext4fs_node_t *eparent = EXT4FS_NODE(pfn);
+	ext4_filesystem_t *fs = eparent->instance->filesystem;
+
+	if (!ext4_inode_is_type(fs->superblock, eparent->inode_ref->inode,
+	    EXT4_INODE_MODE_DIRECTORY)) {
+		return ENOTDIR;
+	}
+
+	/* Try to find entry */
+	ext4_directory_search_result_t result;
+	rc = ext4_directory_find_entry(&result, eparent->inode_ref, component);
+	if (rc != EOK) {
+		if (rc == ENOENT) {
+			*rfn = NULL;
+			return EOK;
+		}
+		return rc;
+	}
+
+	/* Load node from search result */
+	uint32_t inode = ext4_directory_entry_ll_get_inode(result.dentry);
+	rc = ext4fs_node_get_core(rfn, eparent->instance, inode);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	/* Destroy search result structure */
+	rc = ext4_directory_destroy_result(&result);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	return EOK;
+}
+
+/** Get node specified by index
+ *
+ * It's wrapper for node_put_core operation
+ *
+ * @param rfn		output pointer to loaded node if operation successful
+ * @param service_id	device identifier
+ * @param index		node index (here i-node number)
+ * @return		error code
+ */
+int ext4fs_node_get(fs_node_t **rfn, service_id_t service_id, fs_index_t index)
+{
+	int rc;
+
+	ext4fs_instance_t *inst;
+	rc = ext4fs_instance_get(service_id, &inst);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	return ext4fs_node_get_core(rfn, inst, index);
+}
+
+/** Main function for getting node from the filesystem. 
+ *
+ * @param rfn		output point to loaded node if operation successful
+ * @param inst		instance of filesystem
+ * @param index		index of node (i-node number)
+ * @return 		error code
+ */
+int ext4fs_node_get_core(fs_node_t **rfn, ext4fs_instance_t *inst,
+		fs_index_t index)
+{
+	int rc;
+
+	fibril_mutex_lock(&open_nodes_lock);
+
+	/* Check if the node is not already open */
+	unsigned long key[] = {
+		[OPEN_NODES_DEV_HANDLE_KEY] = inst->service_id,
+		[OPEN_NODES_INODE_KEY] = index,
+	};
+
+	link_t *already_open = hash_table_find(&open_nodes, key);
+	ext4fs_node_t *enode = NULL;
+	if (already_open) {
+		enode = hash_table_get_instance(already_open, ext4fs_node_t, link);
+		*rfn = enode->fs_node;
+		enode->references++;
+
+		fibril_mutex_unlock(&open_nodes_lock);
+		return EOK;
+	}
+
+	/* Prepare new enode */
+	enode = malloc(sizeof(ext4fs_node_t));
+	if (enode == NULL) {
+		fibril_mutex_unlock(&open_nodes_lock);
+		return ENOMEM;
+	}
+
+	/* Prepare new fs_node and initialize */
+	fs_node_t *fs_node = malloc(sizeof(fs_node_t));
+	if (fs_node == NULL) {
+		free(enode);
+		fibril_mutex_unlock(&open_nodes_lock);
+		return ENOMEM;
+	}
+	fs_node_initialize(fs_node);
+
+	/* Load i-node from filesystem */
+	ext4_inode_ref_t *inode_ref;
+	rc = ext4_filesystem_get_inode_ref(inst->filesystem, index, &inode_ref);
+	if (rc != EOK) {
+		free(enode);
+		free(fs_node);
+		fibril_mutex_unlock(&open_nodes_lock);
+		return rc;
+	}
+
+	/* Initialize enode */
+	enode->inode_ref = inode_ref;
+	enode->instance = inst;
+	enode->references = 1;
+	enode->fs_node = fs_node;
+	link_initialize(&enode->link);
+
+	fs_node->data = enode;
+	*rfn = fs_node;
+
+	hash_table_insert(&open_nodes, key, &enode->link);
+	inst->open_nodes_count++;
+
+	fibril_mutex_unlock(&open_nodes_lock);
+
+	return EOK;
+}
+
+/** Put previously loaded node.
+ *
+ * @param enode		node to put back
+ * @return		error code
+ */
+int ext4fs_node_put_core(ext4fs_node_t *enode)
+{
+	int rc;
+	unsigned long key[] = {
+		[OPEN_NODES_DEV_HANDLE_KEY] = enode->instance->service_id,
+		[OPEN_NODES_INODE_KEY] = enode->inode_ref->index,
+	};
+
+	hash_table_remove(&open_nodes, key, OPEN_NODES_KEYS);
+	assert(enode->instance->open_nodes_count > 0);
+	enode->instance->open_nodes_count--;
+
+	/* Put inode back in filesystem */
+	rc = ext4_filesystem_put_inode_ref(enode->inode_ref);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	/* Destroy data structure */
+	free(enode->fs_node);
+	free(enode);
+
+	return EOK;
+}
+
+
+/** Open node.
+ *
+ * This operation is stateless in this driver.
+ *
+ * @param fn 	node to open
+ * @return	error code (EOK)
+ */
+int ext4fs_node_open(fs_node_t *fn)
+{
+	/* Stateless operation */
+	return EOK;
+}
+
+
+/** Put previously loaded node.
+ *
+ * It's wrapper for node_put_core operation
+ *
+ * @param fn 	node to put back
+ * @return 	error code
+ */
+int ext4fs_node_put(fs_node_t *fn)
+{
+	int rc;
+
+	fibril_mutex_lock(&open_nodes_lock);
+
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	assert(enode->references > 0);
+	enode->references--;
+	if (enode->references == 0) {
+		rc = ext4fs_node_put_core(enode);
+		if (rc != EOK) {
+			fibril_mutex_unlock(&open_nodes_lock);
+			return rc;
+		}
+	}
+
+	fibril_mutex_unlock(&open_nodes_lock);
+
+	return EOK;
+}
+
+
+/** Create new node in filesystem.
+ *
+ * @param rfn		output pointer to newly created node if successful
+ * @param service_id	device identifier, where the filesystem is
+ * @param flags		flags for specification of new node parameters
+ * @return		error code
+ */
+int ext4fs_create_node(fs_node_t **rfn, service_id_t service_id, int flags)
+{
+	int rc;
+
+	/* Allocate enode */
+	ext4fs_node_t *enode;
+	enode = malloc(sizeof(ext4fs_node_t));
+	if (enode == NULL) {
+		return ENOMEM;
+	}
+
+	/* Allocate fs_node */
+	fs_node_t *fs_node;
+	fs_node = malloc(sizeof(fs_node_t));
+	if (fs_node == NULL) {
+		free(enode);
+		return ENOMEM;
+	}
+
+	/* Load instance */
+	ext4fs_instance_t *inst;
+	rc = ext4fs_instance_get(service_id, &inst);
+	if (rc != EOK) {
+		free(enode);
+		free(fs_node);
+		return rc;
+	}
+
+	/* Allocate new i-node in filesystem */
+	ext4_inode_ref_t *inode_ref;
+	rc = ext4_filesystem_alloc_inode(inst->filesystem, &inode_ref, flags);
+	if (rc != EOK) {
+		free(enode);
+		free(fs_node);
+		return rc;
+	}
+
+	/* Do some interconnections in references */
+	enode->inode_ref = inode_ref;
+	enode->instance = inst;
+	enode->references = 1;
+
+	link_initialize(&enode->link);
+
+	unsigned long key[] = {
+		[OPEN_NODES_DEV_HANDLE_KEY] = inst->service_id,
+		[OPEN_NODES_INODE_KEY] = inode_ref->index,
+	};
+
+	fibril_mutex_lock(&open_nodes_lock);
+	hash_table_insert(&open_nodes, key, &enode->link);
+	fibril_mutex_unlock(&open_nodes_lock);
+	inst->open_nodes_count++;
+
+	enode->inode_ref->dirty = true;
+
+	fs_node_initialize(fs_node);
+	fs_node->data = enode;
+	enode->fs_node = fs_node;
+	*rfn = fs_node;
+
+	return EOK;
+}
+
+
+/** Destroy existing node.
+ *
+ * @param fs	node to destroy
+ * @return	error code
+ */
+int ext4fs_destroy_node(fs_node_t *fn)
+{
+	int rc;
+
+	/* If directory, check for children */
+	bool has_children;
+	rc = ext4fs_has_children(&has_children, fn);
+	if (rc != EOK) {
+		ext4fs_node_put(fn);
+		return rc;
+	}
+
+	if (has_children) {
+		ext4fs_node_put(fn);
+		return EINVAL;
+	}
+
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	ext4_inode_ref_t *inode_ref = enode->inode_ref;
+
+	/* Release data blocks */
+	rc = ext4_filesystem_truncate_inode(inode_ref, 0);
+	if (rc != EOK) {
+		ext4fs_node_put(fn);
+		return rc;
+	}
+
+	/* Handle orphans */
+	// TODO this code should be deleted
+//	ext4_filesystem_t *fs = enode->instance->filesystem;
+//	uint32_t rev_level = ext4_superblock_get_rev_level(fs->superblock);
+//	uint16_t lnk_count = ext4_inode_get_links_count(inode_ref->inode);
+//	if ((rev_level > 0) && (lnk_count == 0)) {
+//		rc = ext4_filesystem_delete_orphan(inode_ref);
+//		if (rc != EOK) {
+//			EXT4FS_DBG("delete orphan error, rc = \%d", rc);
+//		}
+//	}
+
+	// TODO set real deletion time when it will be supported, temporary set fake time
+//	time_t now = time(NULL);
+	ext4_inode_set_deletion_time(inode_ref->inode, 0xdeadbeef);
+	inode_ref->dirty = true;
+
+	/* Free inode */
+	rc = ext4_filesystem_free_inode(inode_ref);
+	if (rc != EOK) {
+		ext4fs_node_put(fn);
+		return rc;
+	}
+
+	return ext4fs_node_put(fn);
+}
+
+
+/** Link the specfied node to directory.
+ *
+ * @param pfn		parent node to link in
+ * @param cfn		node to be linked
+ * @param name		name which will be assigned to directory entry
+ * @return		error code
+ */
+int ext4fs_link(fs_node_t *pfn, fs_node_t *cfn, const char *name)
+{
+	int rc;
+
+	/* Check maximum name length */
+	if (strlen(name) > EXT4_DIRECTORY_FILENAME_LEN) {
+		return ENAMETOOLONG;
+	}
+	ext4fs_node_t *parent = EXT4FS_NODE(pfn);
+	ext4fs_node_t *child = EXT4FS_NODE(cfn);
+	ext4_filesystem_t *fs = parent->instance->filesystem;
+
+	/* Add entry to parent directory */
+	rc = ext4_directory_add_entry(parent->inode_ref, name, child->inode_ref);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	/* Fill new dir -> add '.' and '..' entries */
+	if (ext4_inode_is_type(fs->superblock, child->inode_ref->inode, EXT4_INODE_MODE_DIRECTORY)) {
+
+		rc = ext4_directory_add_entry(child->inode_ref, ".", child->inode_ref);
+		if (rc != EOK) {
+			ext4_directory_remove_entry(parent->inode_ref, name);
+			return rc;
+		}
+
+		rc = ext4_directory_add_entry(child->inode_ref, "..", parent->inode_ref);
+		if (rc != EOK) {
+			ext4_directory_remove_entry(parent->inode_ref, name);
+			ext4_directory_remove_entry(child->inode_ref, ".");
+			return rc;
+		}
+
+		/* Initialize directory index if supported */
+		if (ext4_superblock_has_feature_compatible(
+				fs->superblock, EXT4_FEATURE_COMPAT_DIR_INDEX)) {
+
+			rc = ext4_directory_dx_init(child->inode_ref);
+			if (rc != EOK) {
+				return rc;
+			}
+
+			ext4_inode_set_flag(child->inode_ref->inode, EXT4_INODE_FLAG_INDEX);
+			child->inode_ref->dirty = true;
+		}
+
+		uint16_t parent_links = ext4_inode_get_links_count(parent->inode_ref->inode);
+		parent_links++;
+		ext4_inode_set_links_count(parent->inode_ref->inode, parent_links);
+
+		parent->inode_ref->dirty = true;
+
+	}
+
+	uint16_t child_links = ext4_inode_get_links_count(child->inode_ref->inode);
+	child_links++;
+	ext4_inode_set_links_count(child->inode_ref->inode, child_links);
+
+	child->inode_ref->dirty = true;
+
+	return EOK;
+}
+
+
+/** Unlink node from specified directory.
+ *
+ * @param pfn		parent node to delete node from
+ * @param cfn		child node to be unlinked from directory
+ * @param name		name of entry that will be removed
+ * @return		error code
+ */
+int ext4fs_unlink(fs_node_t *pfn, fs_node_t *cfn, const char *name)
+{
+	int rc;
+
+	bool has_children;
+	rc = ext4fs_has_children(&has_children, cfn);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	/* Cannot unlink non-empty node */
+	if (has_children) {
+		return ENOTEMPTY;
+	}
+
+	/* Remove entry from parent directory */
+	ext4_inode_ref_t *parent = EXT4FS_NODE(pfn)->inode_ref;
+	rc = ext4_directory_remove_entry(parent, name);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	/* Decrement links count */
+	ext4_inode_ref_t * child_inode_ref = EXT4FS_NODE(cfn)->inode_ref;
+
+	uint32_t lnk_count = ext4_inode_get_links_count(child_inode_ref->inode);
+	lnk_count--;
+
+	/* If directory - handle links from parent */
+	if (lnk_count <= 1 && ext4fs_is_directory(cfn)) {
+
+		assert(lnk_count == 1);
+		lnk_count--;
+
+		ext4_inode_ref_t *parent_inode_ref = EXT4FS_NODE(pfn)->inode_ref;
+
+		uint32_t parent_lnk_count = ext4_inode_get_links_count(
+				parent_inode_ref->inode);
+
+		parent_lnk_count--;
+		ext4_inode_set_links_count(parent_inode_ref->inode, parent_lnk_count);
+
+		parent->dirty = true;
+	}
+
+//	ext4_filesystem_t *fs = EXT4FS_NODE(pfn)->instance->filesystem;
+//	uint32_t rev_level = ext4_superblock_get_rev_level(fs->superblock);
+//	if ((rev_level > 0) && (lnk_count == 0)) {
+//		rc = ext4_filesystem_add_orphan(child_inode_ref);
+//		if (rc != EOK) {
+//			EXT4FS_DBG("add orphan error, rc = \%d", rc);
+//		}
+//	}
+
+	// TODO set timestamps for parent (when we have wall-clock time)
+//	time_t now = time(NULL);
+//	ext4_inode_set_change_inode_time(parent->inode, (uint32_t)now);
+//	ext4_inode_set_modification_time(parent->inode, (uint32_t)now);
+//	parent->dirty = true;
+
+	// TODO set timestamp for inode
+//	ext4_inode_set_change_inode_time(child_inode_ref->inode, (uint32_t)now);
+	ext4_inode_set_links_count(child_inode_ref->inode, lnk_count);
+	child_inode_ref->dirty = true;
+
+	return EOK;
+}
+
+
+/** Check if specified node has children.
+ * 
+ * For files is response allways false and check is executed only for directories.
+ *
+ * @param has_children		output value for response
+ * @param fn			node to check
+ * @return 			error code
+ */
+int ext4fs_has_children(bool *has_children, fs_node_t *fn)
+{
+	int rc;
+
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	ext4_filesystem_t *fs = enode->instance->filesystem;
+
+	/* Check if node is directory */
+	if (!ext4_inode_is_type(fs->superblock, enode->inode_ref->inode,
+	    EXT4_INODE_MODE_DIRECTORY)) {
+		*has_children = false;
+		return EOK;
+	}
+
+	ext4_directory_iterator_t it;
+	rc = ext4_directory_iterator_init(&it, enode->inode_ref, 0);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	/* Find a non-empty directory entry */
+	bool found = false;
+	while (it.current != NULL) {
+		if (it.current->inode != 0) {
+			uint16_t name_size = ext4_directory_entry_ll_get_name_length(fs->superblock,
+				it.current);
+			if (!ext4fs_is_dots(it.current->name, name_size)) {
+				found = true;
+				break;
+			}
+		}
+
+		rc = ext4_directory_iterator_next(&it);
+		if (rc != EOK) {
+			ext4_directory_iterator_fini(&it);
+			return rc;
+		}
+	}
+
+	rc = ext4_directory_iterator_fini(&it);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	*has_children = found;
+
+	return EOK;
+}
+
+
+/** Unpack index number from node.
+ *	
+ * @param fn	node to load index from
+ * @return	index number of i-node
+ */
+fs_index_t ext4fs_index_get(fs_node_t *fn)
+{
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	return enode->inode_ref->index;
+}
+
+
+/** Get real size of file / directory.
+ *
+ * @param fn	node to get size of
+ * @return 	real size of node
+ */
+aoff64_t ext4fs_size_get(fs_node_t *fn)
+{
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	ext4_superblock_t *sb = enode->instance->filesystem->superblock;
+	return ext4_inode_get_size(sb, enode->inode_ref->inode);
+}
+
+
+/** Get number of links to specified node.
+ *
+ * @param fn	node to get links to
+ * @return	number of links
+ */
+unsigned ext4fs_lnkcnt_get(fs_node_t *fn)
+{
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	uint32_t lnkcnt = ext4_inode_get_links_count(enode->inode_ref->inode);
+
+	if (ext4fs_is_directory(fn)) {
+		if (lnkcnt > 1) {
+			return 1;
+		} else {
+			return 0;
+		}
+	}
+
+	/* For regular files return real links count */
+	return lnkcnt;
+}
+
+
+/** Check if node is directory.
+ *
+ * @param fn	node to check
+ * @return 	result of check
+ */
+bool ext4fs_is_directory(fs_node_t *fn)
+{
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	ext4_superblock_t *sb = enode->instance->filesystem->superblock;
+	return ext4_inode_is_type(
+			sb, enode->inode_ref->inode, EXT4_INODE_MODE_DIRECTORY);
+}
+
+
+/** Check if node is regular file.
+ *
+ * @param fn	node to check
+ * @return	result of check
+ */
+bool ext4fs_is_file(fs_node_t *fn)
+{
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	ext4_superblock_t *sb = enode->instance->filesystem->superblock;
+	return ext4_inode_is_type(
+			sb, enode->inode_ref->inode, EXT4_INODE_MODE_FILE);
+}
+
+/** Extract device identifier from node.
+ *
+ * @param node	node to extract id from
+ * @return 	id of device, where is the filesystem
+ */
+service_id_t ext4fs_service_get(fs_node_t *fn)
+{
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	return enode->instance->service_id;
+}
+
+/*
+ * libfs operations.
+ */
+libfs_ops_t ext4fs_libfs_ops = {
+	.root_get = ext4fs_root_get,
+	.match = ext4fs_match,
+	.node_get = ext4fs_node_get,
+	.node_open = ext4fs_node_open,
+	.node_put = ext4fs_node_put,
+	.create = ext4fs_create_node,
+	.destroy = ext4fs_destroy_node,
+	.link = ext4fs_link,
+	.unlink = ext4fs_unlink,
+	.has_children = ext4fs_has_children,
+	.index_get = ext4fs_index_get,
+	.size_get = ext4fs_size_get,
+	.lnkcnt_get = ext4fs_lnkcnt_get,
+	.is_directory = ext4fs_is_directory,
+	.is_file = ext4fs_is_file,
+	.service_get = ext4fs_service_get
+};
+
+
+/*
+ * VFS operations.
+ */
+
+/** Mount operation. 
+ *
+ * Try to mount specified filesystem from device.
+ * 
+ * @param service_id	identifier of device
+ * @param opts		mount options
+ * @param index		output value - index of root node
+ * @param size		output value - size of root node
+ * @param lnkcnt	output value - link count of root node
+ * @return		error code
+ */
+static int ext4fs_mounted(service_id_t service_id, const char *opts,
+   fs_index_t *index, aoff64_t *size, unsigned *lnkcnt)
+{
+	int rc;
+
+	/* Allocate libext4 filesystem structure */
+	ext4_filesystem_t *fs;
+	fs = (ext4_filesystem_t *) malloc(sizeof(ext4_filesystem_t));
+	if (fs == NULL) {
+		return ENOMEM;
+	}
+
+	/* Allocate instance structure */
+	ext4fs_instance_t *inst;
+	inst = (ext4fs_instance_t *) malloc(sizeof(ext4fs_instance_t));
+	if (inst == NULL) {
+		free(fs);
+		return ENOMEM;
+	}
+
+	/* Initialize the filesystem */
+	rc = ext4_filesystem_init(fs, service_id);
+	if (rc != EOK) {
+		free(fs);
+		free(inst);
+		return rc;
+	}
+
+	/* Do some sanity checking */
+	rc = ext4_filesystem_check_sanity(fs);
+	if (rc != EOK) {
+		ext4_filesystem_fini(fs);
+		free(fs);
+		free(inst);
+		return rc;
+	}
+
+	/* Check flags */
+	bool read_only;
+	rc = ext4_filesystem_check_features(fs, &read_only);
+	if (rc != EOK) {
+		ext4_filesystem_fini(fs);
+		free(fs);
+		free(inst);
+		return rc;
+	}
+
+	/* Initialize instance */
+	link_initialize(&inst->link);
+	inst->service_id = service_id;
+	inst->filesystem = fs;
+	inst->open_nodes_count = 0;
+
+	/* Read root node */
+	fs_node_t *root_node;
+	rc = ext4fs_node_get_core(&root_node, inst, EXT4_INODE_ROOT_INDEX);
+	if (rc != EOK) {
+		ext4_filesystem_fini(fs);
+		free(fs);
+		free(inst);
+		return rc;
+	}
+
+	/* Add instance to the list */
+	fibril_mutex_lock(&instance_list_mutex);
+	list_append(&inst->link, &instance_list);
+	fibril_mutex_unlock(&instance_list_mutex);
+
+	ext4fs_node_t *enode = EXT4FS_NODE(root_node);
+
+	*index = EXT4_INODE_ROOT_INDEX;
+	*size = ext4_inode_get_size(fs->superblock, enode->inode_ref->inode);
+	*lnkcnt = 1;
+
+	ext4fs_node_put(root_node);
+
+	return EOK;
+}
+
+/** Unmount operation.
+ *
+ * Correctly release the filesystem.
+ *
+ * @param service_id	device to be unmounted
+ * @return 		error code
+ */
+static int ext4fs_unmounted(service_id_t service_id)
+{
+	int rc;
+
+	ext4fs_instance_t *inst;
+	rc = ext4fs_instance_get(service_id, &inst);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	fibril_mutex_lock(&open_nodes_lock);
+
+	if (inst->open_nodes_count != 0) {
+		fibril_mutex_unlock(&open_nodes_lock);
+		return EBUSY;
+	}
+
+	/* Remove the instance from the list */
+	fibril_mutex_lock(&instance_list_mutex);
+	list_remove(&inst->link);
+	fibril_mutex_unlock(&instance_list_mutex);
+
+	fibril_mutex_unlock(&open_nodes_lock);
+
+	return ext4_filesystem_fini(inst->filesystem);
+}
+
+
+/** Read bytes from node.
+ *
+ * @param service_id	device to read data from
+ * @param index		number of node to read from
+ * @param pos		position where the read should be started
+ * @param rbytes	output value, where the real size was returned
+ * @return 		error code
+ */
+static int ext4fs_read(service_id_t service_id, fs_index_t index,
+		aoff64_t pos, size_t *rbytes)
+{
+	int rc;
+
+	/*
+	 * Receive the read request.
+	 */
+	ipc_callid_t callid;
+	size_t size;
+	if (!async_data_read_receive(&callid, &size)) {
+		async_answer_0(callid, EINVAL);
+		return EINVAL;
+	}
+
+	ext4fs_instance_t *inst;
+	rc = ext4fs_instance_get(service_id, &inst);
+	if (rc != EOK) {
+		async_answer_0(callid, rc);
+		return rc;
+	}
+
+	/* Load i-node */
+	ext4_inode_ref_t *inode_ref;
+	rc = ext4_filesystem_get_inode_ref(inst->filesystem, index, &inode_ref);
+	if (rc != EOK) {
+		async_answer_0(callid, rc);
+		return rc;
+	}
+
+	/* Read from i-node by type */
+	if (ext4_inode_is_type(inst->filesystem->superblock, inode_ref->inode,
+			EXT4_INODE_MODE_FILE)) {
+		rc = ext4fs_read_file(callid, pos, size, inst, inode_ref,
+				rbytes);
+	} else if (ext4_inode_is_type(inst->filesystem->superblock,
+			inode_ref->inode, EXT4_INODE_MODE_DIRECTORY)) {
+		rc = ext4fs_read_directory(callid, pos, size, inst, inode_ref,
+				rbytes);
+	} else {
+		/* Other inode types not supported */
+		async_answer_0(callid, ENOTSUP);
+		rc = ENOTSUP;
+	}
+
+	ext4_filesystem_put_inode_ref(inode_ref);
+
+	return rc;
+}
+
+/** Check if filename is dot or dotdot (reserved names).
+ *
+ * @param name		name to check
+ * @param name_size	length of string name
+ * @return 		result of the check
+ */
+bool ext4fs_is_dots(const uint8_t *name, size_t name_size)
+{
+	if (name_size == 1 && name[0] == '.') {
+		return true;
+	}
+
+	if (name_size == 2 && name[0] == '.' && name[1] == '.') {
+		return true;
+	}
+
+	return false;
+}
+
+/** Read data from directory.
+ *
+ * @param callid	IPC id of call (for communication)
+ * @param pos		position to start reading from
+ * @param size		how many bytes to read
+ * @param inst		filesystem instance
+ * @param inode_ref	node to read data from
+ * @param rbytes	output value to return real number of bytes was read
+ * @return 		error code
+ */
+int ext4fs_read_directory(ipc_callid_t callid, aoff64_t pos, size_t size,
+    ext4fs_instance_t *inst, ext4_inode_ref_t *inode_ref, size_t *rbytes)
+{
+	int rc;
+
+	ext4_directory_iterator_t it;
+	rc = ext4_directory_iterator_init(&it, inode_ref, pos);
+	if (rc != EOK) {
+		async_answer_0(callid, rc);
+		return rc;
+	}
+
+	/* Find next interesting directory entry.
+	 * We want to skip . and .. entries
+	 * as these are not used in HelenOS
+	 */
+	bool found = false;
+	while (it.current != NULL) {
+
+		if (it.current->inode == 0) {
+			goto skip;
+		}
+
+		uint16_t name_size = ext4_directory_entry_ll_get_name_length(
+		    inst->filesystem->superblock, it.current);
+
+		/* skip . and .. */
+		if (ext4fs_is_dots(it.current->name, name_size)) {
+			goto skip;
+		}
+
+		/* The on-disk entry does not contain \0 at the end
+		 * end of entry name, so we copy it to new buffer
+		 * and add the \0 at the end
+		 */
+		uint8_t *buf = malloc(name_size+1);
+		if (buf == NULL) {
+			ext4_directory_iterator_fini(&it);
+			async_answer_0(callid, ENOMEM);
+			return ENOMEM;
+		}
+		memcpy(buf, &it.current->name, name_size);
+		*(buf + name_size) = 0;
+		found = true;
+		(void) async_data_read_finalize(callid, buf, name_size + 1);
+		free(buf);
+		break;
+
+skip:
+		rc = ext4_directory_iterator_next(&it);
+		if (rc != EOK) {
+			ext4_directory_iterator_fini(&it);
+			async_answer_0(callid, rc);
+			return rc;
+		}
+	}
+
+	uint64_t next;
+	if (found) {
+		rc = ext4_directory_iterator_next(&it);
+		if (rc != EOK) {
+			return rc;
+		}
+		next = it.current_offset;
+	}
+
+	rc = ext4_directory_iterator_fini(&it);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	/* Prepare return values */
+	if (found) {
+		*rbytes = next - pos;
+		return EOK;
+	} else {
+		async_answer_0(callid, ENOENT);
+		return ENOENT;
+	}
+}
+
+
+/** Read data from file.
+ *
+ * @param callid        IPC id of call (for communication)
+ * @param pos           position to start reading from
+ * @param size          how many bytes to read
+ * @param inst          filesystem instance
+ * @param inode_ref     node to read data from
+ * @param rbytes        output value to return real number of bytes was read
+ * @return              error code
+ */
+int ext4fs_read_file(ipc_callid_t callid, aoff64_t pos, size_t size,
+    ext4fs_instance_t *inst, ext4_inode_ref_t *inode_ref, size_t *rbytes)
+{
+	int rc;
+
+	ext4_superblock_t *sb = inst->filesystem->superblock;
+	uint64_t file_size = ext4_inode_get_size(sb, inode_ref->inode);
+
+	if (pos >= file_size) {
+		/* Read 0 bytes successfully */
+		async_data_read_finalize(callid, NULL, 0);
+		*rbytes = 0;
+		return EOK;
+	}
+
+	/* For now, we only read data from one block at a time */
+	uint32_t block_size = ext4_superblock_get_block_size(sb);
+	aoff64_t file_block = pos / block_size;
+	uint32_t offset_in_block = pos % block_size;
+	uint32_t bytes = min(block_size - offset_in_block, size);
+
+	/* Handle end of file */
+	if (pos + bytes > file_size) {
+		bytes = file_size - pos;
+	}
+
+	/* Get the real block number */
+	uint32_t fs_block;
+	rc = ext4_filesystem_get_inode_data_block_index(inode_ref, file_block, &fs_block);
+	if (rc != EOK) {
+		async_answer_0(callid, rc);
+		return rc;
+	}
+
+	/* Check for sparse file
+	 * If ext4_filesystem_get_inode_data_block_index returned
+	 * fs_block == 0, it means that the given block is not allocated for the
+	 * file and we need to return a buffer of zeros
+	 */
+	uint8_t *buffer;
+	if (fs_block == 0) {
+		buffer = malloc(bytes);
+		if (buffer == NULL) {
+			async_answer_0(callid, ENOMEM);
+			return ENOMEM;
+		}
+
+		memset(buffer, 0, bytes);
+
+		async_data_read_finalize(callid, buffer, bytes);
+		*rbytes = bytes;
+
+		free(buffer);
+
+		return EOK;
+	}
+
+	/* Usual case - we need to read a block from device */
+	block_t *block;
+	rc = block_get(&block, inst->service_id, fs_block, BLOCK_FLAGS_NONE);
+	if (rc != EOK) {
+		async_answer_0(callid, rc);
+		return rc;
+	}
+
+	assert(offset_in_block + bytes <= block_size);
+	async_data_read_finalize(callid, block->data + offset_in_block, bytes);
+
+	rc = block_put(block);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	*rbytes = bytes;
+	return EOK;
+}
+
+
+/** Write bytes to file
+ *
+ * @param service_id	device identifier
+ * @param index		i-node number of file
+ * @param pos		position in file to start reading from
+ * @param wbytes	output value - real number of written bytes
+ * @param nsize		output value - new size of i-node
+ * @return 		error code
+ */
+static int ext4fs_write(service_id_t service_id, fs_index_t index,
+		aoff64_t pos, size_t *wbytes, aoff64_t *nsize)
+{
+	int rc;
+
+	fs_node_t *fn;
+	rc = ext4fs_node_get(&fn, service_id, index);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	ipc_callid_t callid;
+	size_t len;
+	if (!async_data_write_receive(&callid, &len)) {
+		rc = EINVAL;
+		ext4fs_node_put(fn);
+		async_answer_0(callid, rc);
+		return rc;
+	}
+
+
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	ext4_filesystem_t *fs = enode->instance->filesystem;
+
+	uint32_t block_size = ext4_superblock_get_block_size(fs->superblock);
+
+	/* Prevent writing to more than one block */
+	uint32_t bytes = min(len, block_size - (pos % block_size));
+
+	int flags = BLOCK_FLAGS_NONE;
+	if (bytes == block_size) {
+		flags = BLOCK_FLAGS_NOREAD;
+	}
+
+	uint32_t iblock =  pos / block_size;
+	uint32_t fblock;
+
+	/* Load inode */
+	ext4_inode_ref_t *inode_ref = enode->inode_ref;
+	rc = ext4_filesystem_get_inode_data_block_index(inode_ref, iblock, &fblock);
+	if (rc != EOK) {
+		ext4fs_node_put(fn);
+		async_answer_0(callid, rc);
+		return rc;
+	}
+
+	/* Check for sparse file */
+	if (fblock == 0) {
+
+		if (ext4_superblock_has_feature_incompatible(
+				fs->superblock, EXT4_FEATURE_INCOMPAT_EXTENTS) &&
+				(ext4_inode_has_flag(inode_ref->inode, EXT4_INODE_FLAG_EXTENTS))) {
+
+			uint32_t last_iblock = ext4_inode_get_size(fs->superblock, inode_ref->inode) / block_size;
+			while (last_iblock < iblock) {
+				rc = ext4_extent_append_block(inode_ref, &last_iblock, &fblock, true);
+				if (rc != EOK) {
+					ext4fs_node_put(fn);
+					async_answer_0(callid, rc);
+					return rc;
+				}
+			}
+
+			rc = ext4_extent_append_block(inode_ref, &last_iblock, &fblock, false);
+			if (rc != EOK) {
+				ext4fs_node_put(fn);
+				async_answer_0(callid, rc);
+				return rc;
+			}
+
+		} else {
+			rc =  ext4_balloc_alloc_block(inode_ref, &fblock);
+			if (rc != EOK) {
+				ext4fs_node_put(fn);
+				async_answer_0(callid, rc);
+				return rc;
+			}
+
+			rc = ext4_filesystem_set_inode_data_block_index(inode_ref, iblock, fblock);
+			if (rc != EOK) {
+				ext4_balloc_free_block(inode_ref, fblock);
+				ext4fs_node_put(fn);
+				async_answer_0(callid, rc);
+				return rc;
+			}
+		}
+
+		flags = BLOCK_FLAGS_NOREAD;
+		inode_ref->dirty = true;
+	}
+
+	/* Load target block */
+	block_t *write_block;
+	rc = block_get(&write_block, service_id, fblock, flags);
+	if (rc != EOK) {
+		ext4fs_node_put(fn);
+		async_answer_0(callid, rc);
+		return rc;
+	}
+
+	if (flags == BLOCK_FLAGS_NOREAD) {
+		memset(write_block->data, 0, block_size);
+	}
+
+	rc = async_data_write_finalize(callid, write_block->data + (pos % block_size), bytes);
+	if (rc != EOK) {
+		ext4fs_node_put(fn);
+		return rc;
+	}
+
+	write_block->dirty = true;
+
+	rc = block_put(write_block);
+	if (rc != EOK) {
+		ext4fs_node_put(fn);
+		return rc;
+	}
+
+	/* Do some counting */
+	uint32_t old_inode_size = ext4_inode_get_size(fs->superblock, inode_ref->inode);
+	if (pos + bytes > old_inode_size) {
+		ext4_inode_set_size(inode_ref->inode, pos + bytes);
+		inode_ref->dirty = true;
+	}
+
+	*nsize = ext4_inode_get_size(fs->superblock, inode_ref->inode);
+	*wbytes = bytes;
+
+	return ext4fs_node_put(fn);
+}
+
+
+/** Truncate file.
+ *
+ * Only the direction to shorter file is supported.
+ *
+ * @param service_id	device identifier
+ * @param index		index if node to truncated
+ * @param new_size	new size of file
+ * @return		error code
+ */
+static int ext4fs_truncate(service_id_t service_id, fs_index_t index,
+		aoff64_t new_size)
+{
+	int rc;
+
+	fs_node_t *fn;
+	rc = ext4fs_node_get(&fn, service_id, index);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	ext4_inode_ref_t *inode_ref = enode->inode_ref;
+
+	rc = ext4_filesystem_truncate_inode(inode_ref, new_size);
+	ext4fs_node_put(fn);
+
+	return rc;
+}
+
+
+/** Close file.
+ *
+ * @param service_id	device identifier
+ * @param index		i-node number
+ * @return 		error code
+ */
+static int ext4fs_close(service_id_t service_id, fs_index_t index)
+{
+	return EOK;
+}
+
+
+/** Destroy node specified by index.
+ *
+ * @param service_id	device identifier
+ * @param index		number of i-node to destroy
+ * @return		error code
+ */
+static int ext4fs_destroy(service_id_t service_id, fs_index_t index)
+{
+	int rc;
+
+	fs_node_t *fn;
+	rc = ext4fs_node_get(&fn, service_id, index);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	/* Destroy the inode */
+	return ext4fs_destroy_node(fn);
+}
+
+/** Enforce inode synchronization (write) to device. 
+ *
+ * @param service_id 	device identifier
+ * @param index		i-node number.
+ */
+static int ext4fs_sync(service_id_t service_id, fs_index_t index)
+{
+	int rc;
+
+	fs_node_t *fn;
+	rc = ext4fs_node_get(&fn, service_id, index);
+	if (rc != EOK) {
+		return rc;
+	}
+
+	ext4fs_node_t *enode = EXT4FS_NODE(fn);
+	enode->inode_ref->dirty = true;
+
+	return ext4fs_node_put(fn);
+}
+
+/** VFS operations
+ *
+ */
+vfs_out_ops_t ext4fs_ops = {
+	.mounted = ext4fs_mounted,
+	.unmounted = ext4fs_unmounted,
+	.read = ext4fs_read,
+	.write = ext4fs_write,
+	.truncate = ext4fs_truncate,
+	.close = ext4fs_close,
+	.destroy = ext4fs_destroy,
+	.sync = ext4fs_sync,
+};
+
+/**
+ * @}
+ */