drivers/esoc/esoc_dev.c - kernel/msm-4.9 - Gitiles

 /* Copyright (c) 2013-2014, 2017, The Linux Foundation. All rights reserved.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 and
  * only version 2 as published by the Free Software Foundation.
  *
  * This program is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  * GNU General Public License for more details.
  */
 #include <linux/kfifo.h>
 #include <linux/list.h>
 #include <linux/sched.h>
 #include <linux/uaccess.h>
 #include <linux/wait.h>
 #include "esoc.h"

 /**
  * struct esoc_udev: Userspace char interface
  * @dev: interface device.
  * @req_fifio: fifo for clink requests.
  * @req_wait: signal availability of request from clink
  * @req_fifo_lock: serialize access to req fifo
  * @evt_fito: fifo for clink events
  * @evt_wait: signal availability of clink event
  * @evt_fifo_lock: serialize access to event fifo
  * @list: entry in esoc dev list.
  * @clink: reference to contorl link
  */
 struct esoc_udev {
 	struct device *dev;
 	struct kfifo req_fifo;
 	wait_queue_head_t req_wait;
 	spinlock_t req_fifo_lock;
 	struct kfifo evt_fifo;
 	wait_queue_head_t evt_wait;
 	spinlock_t evt_fifo_lock;
 	struct list_head list;
 	struct esoc_clink *clink;
 };

 /**
  * struct esoc_uhandle: Userspace handle of esoc
  * @esoc_clink: esoc control link.
  * @eng: esoc engine for commands/ requests.
  * @esoc_udev: user interface device.
  * @req_eng_reg: indicates if engine is registered as request eng
  * @cmd_eng_reg: indicates if engine is registered as cmd eng
  */
 struct esoc_uhandle {
 	struct esoc_clink *esoc_clink;
 	struct esoc_eng eng;
 	struct esoc_udev *esoc_udev;
 	bool req_eng_reg;
 	bool cmd_eng_reg;
 };

 #define ESOC_MAX_MINOR	256
 #define ESOC_MAX_REQ	8
 #define ESOC_MAX_EVT	4

 static LIST_HEAD(esoc_udev_list);
 static DEFINE_SPINLOCK(esoc_udev_list_lock);
 struct class *esoc_class;
 static int esoc_major;

 static struct esoc_udev *get_free_esoc_udev(struct esoc_clink *esoc_clink)
 {
 	struct esoc_udev *esoc_udev;
 	int err;

 	if (esoc_clink->id > ESOC_MAX_MINOR) {
 		pr_err("too many esoc devices\n");
 		return ERR_PTR(-ENODEV);
 	}
 	esoc_udev = kzalloc(sizeof(*esoc_udev), GFP_KERNEL);
 	if (!esoc_udev)
 		return ERR_PTR(-ENOMEM);
 	err = kfifo_alloc(&esoc_udev->req_fifo, (sizeof(u32)) * ESOC_MAX_REQ,
 								GFP_KERNEL);
 	if (err) {
 		pr_err("unable to allocate request fifo for %s\n",
 							esoc_clink->name);
 		goto req_fifo_fail;
 	}
 	err = kfifo_alloc(&esoc_udev->evt_fifo, (sizeof(u32)) * ESOC_MAX_EVT,
 								GFP_KERNEL);
 	if (err) {
 		pr_err("unable to allocate evt fifo for %s\n",
 							esoc_clink->name);
 		goto evt_fifo_fail;
 	}
 	init_waitqueue_head(&esoc_udev->req_wait);
 	init_waitqueue_head(&esoc_udev->evt_wait);
 	spin_lock_init(&esoc_udev->req_fifo_lock);
 	spin_lock_init(&esoc_udev->evt_fifo_lock);
 	esoc_udev->clink = esoc_clink;
 	spin_lock(&esoc_udev_list_lock);
 	list_add_tail(&esoc_udev->list, &esoc_udev_list);
 	spin_unlock(&esoc_udev_list_lock);
 	return esoc_udev;
 evt_fifo_fail:
 	kfifo_free(&esoc_udev->req_fifo);
 req_fifo_fail:
 	kfree(esoc_udev);
 	return ERR_PTR(-ENODEV);
 }

 static void return_esoc_udev(struct esoc_udev *esoc_udev)
 {
 	spin_lock(&esoc_udev_list_lock);
 	list_del(&esoc_udev->list);
 	spin_unlock(&esoc_udev_list_lock);
 	kfifo_free(&esoc_udev->req_fifo);
 	kfifo_free(&esoc_udev->evt_fifo);
 	kfree(esoc_udev);
 }

 static struct esoc_udev *esoc_udev_get_by_minor(unsigned int index)
 {
 	struct esoc_udev *esoc_udev;

 	spin_lock(&esoc_udev_list_lock);
 	list_for_each_entry(esoc_udev, &esoc_udev_list, list) {
 		if (esoc_udev->clink->id == index)
 			goto found;
 	}
 	esoc_udev = NULL;
 found:
 	spin_unlock(&esoc_udev_list_lock);
 	return esoc_udev;
 }

 void esoc_udev_handle_clink_req(enum esoc_req req, struct esoc_eng *eng)
 {
 	int err;
 	u32 clink_req;
 	struct esoc_clink *esoc_clink = eng->esoc_clink;
 	struct esoc_udev *esoc_udev = esoc_udev_get_by_minor(esoc_clink->id);

 	if (!esoc_udev)
 		return;
 	clink_req = (u32)req;
 	err = kfifo_in_spinlocked(&esoc_udev->req_fifo, &clink_req,
 						sizeof(clink_req),
 						&esoc_udev->req_fifo_lock);
 	if (err != sizeof(clink_req)) {
 		pr_err("unable to queue request for %s\n", esoc_clink->name);
 		return;
 	}
 	wake_up_interruptible(&esoc_udev->req_wait);
 }

 void esoc_udev_handle_clink_evt(enum esoc_evt evt, struct esoc_eng *eng)
 {
 	int err;
 	u32 clink_evt;
 	struct esoc_clink *esoc_clink = eng->esoc_clink;
 	struct esoc_udev *esoc_udev = esoc_udev_get_by_minor(esoc_clink->id);

 	if (!esoc_udev)
 		return;
 	clink_evt = (u32)evt;
 	err = kfifo_in_spinlocked(&esoc_udev->evt_fifo, &clink_evt,
 						sizeof(clink_evt),
 						&esoc_udev->evt_fifo_lock);
 	if (err != sizeof(clink_evt)) {
 		pr_err("unable to queue event for %s\n", esoc_clink->name);
 		return;
 	}
 	wake_up_interruptible(&esoc_udev->evt_wait);
 }

 static long esoc_dev_ioctl(struct file *file, unsigned int cmd,
 						unsigned long arg)
 {
 	int err;
 	u32 esoc_cmd, status, req, evt;
 	struct esoc_uhandle *uhandle = file->private_data;
 	struct esoc_udev *esoc_udev = uhandle->esoc_udev;
 	struct esoc_clink *esoc_clink = uhandle->esoc_clink;
 	const struct esoc_clink_ops * const clink_ops = esoc_clink->clink_ops;
 	void __user *uarg = (void __user *)arg;

 	switch (cmd) {
 	case ESOC_REG_REQ_ENG:
 		err = esoc_clink_register_req_eng(esoc_clink, &uhandle->eng);
 		if (err)
 			return err;
 		uhandle->req_eng_reg = true;
 		break;
 	case ESOC_REG_CMD_ENG:
 		err = esoc_clink_register_cmd_eng(esoc_clink, &uhandle->eng);
 		if (err)
 			return err;
 		uhandle->cmd_eng_reg = true;
 		break;
 	case ESOC_CMD_EXE:
 		if (esoc_clink->cmd_eng != &uhandle->eng)
 			return -EACCES;
 		get_user(esoc_cmd, (u32 __user *)arg);
 		return clink_ops->cmd_exe(esoc_cmd, esoc_clink);
 	case ESOC_WAIT_FOR_REQ:
 		if (esoc_clink->req_eng != &uhandle->eng)
 			return -EACCES;
 		err = wait_event_interruptible(esoc_udev->req_wait,
 					!kfifo_is_empty(&esoc_udev->req_fifo));
 		if (!err) {
 			err = kfifo_out_spinlocked(&esoc_udev->req_fifo, &req,
 								sizeof(req),
 						&esoc_udev->req_fifo_lock);
 			if (err != sizeof(req)) {
 				pr_err("read from clink %s req q failed\n",
 							esoc_clink->name);
 				return -EIO;
 			}
 			put_user(req, (unsigned int __user *)uarg);

 		}
 		return err;
 	case ESOC_NOTIFY:
 		get_user(esoc_cmd, (u32 __user *)arg);
 		clink_ops->notify(esoc_cmd, esoc_clink);
 		break;
 	case ESOC_GET_STATUS:
 		err = clink_ops->get_status(&status, esoc_clink);
 		if (err)
 			return err;
 		put_user(status, (unsigned int __user *)uarg);
 		break;
 	case ESOC_WAIT_FOR_CRASH:
 		err = wait_event_interruptible(esoc_udev->evt_wait,
 					!kfifo_is_empty(&esoc_udev->evt_fifo));
 		if (!err) {
 			err = kfifo_out_spinlocked(&esoc_udev->evt_fifo, &evt,
 								sizeof(evt),
 						&esoc_udev->evt_fifo_lock);
 			if (err != sizeof(evt)) {
 				pr_err("read from clink %s evt q failed\n",
 							esoc_clink->name);
 				return -EIO;
 			}
 			put_user(evt, (unsigned int __user *)uarg);
 		}
 		return err;
 	default:
 		return -EINVAL;
 	};
 	return 0;
 }

 static int esoc_dev_open(struct inode *inode, struct file *file)
 {
 	struct esoc_uhandle *uhandle;
 	struct esoc_udev *esoc_udev;
 	struct esoc_clink *esoc_clink;
 	struct esoc_eng *eng;
 	unsigned int minor = iminor(inode);

 	esoc_udev = esoc_udev_get_by_minor(minor);
 	if (!esoc_udev) {
 		pr_err("failed to get udev\n");
 		return -ENOMEM;
 	}

 	esoc_clink = get_esoc_clink(esoc_udev->clink->id);
 	if (!esoc_clink) {
 		pr_err("failed to get clink\n");
 		return -ENOMEM;
 	}

 	uhandle = kzalloc(sizeof(*uhandle), GFP_KERNEL);
 	if (!uhandle) {
 		put_esoc_clink(esoc_clink);
 		return -ENOMEM;
 	}
 	uhandle->esoc_udev = esoc_udev;
 	uhandle->esoc_clink = esoc_clink;
 	eng = &uhandle->eng;
 	eng->handle_clink_req = esoc_udev_handle_clink_req;
 	eng->handle_clink_evt = esoc_udev_handle_clink_evt;
 	file->private_data = uhandle;
 	return 0;
 }

 static int esoc_dev_release(struct inode *inode, struct file *file)
 {
 	struct esoc_clink *esoc_clink;
 	struct esoc_uhandle *uhandle = file->private_data;

 	esoc_clink = uhandle->esoc_clink;
 	if (uhandle->req_eng_reg)
 		esoc_clink_unregister_req_eng(esoc_clink, &uhandle->eng);
 	if (uhandle->cmd_eng_reg)
 		esoc_clink_unregister_cmd_eng(esoc_clink, &uhandle->eng);
 	uhandle->req_eng_reg = false;
 	uhandle->cmd_eng_reg = false;
 	put_esoc_clink(esoc_clink);
 	kfree(uhandle);
 	return 0;
 }
 static const struct file_operations esoc_dev_fops = {
 	.owner		= THIS_MODULE,
 	.open		= esoc_dev_open,
 	.unlocked_ioctl = esoc_dev_ioctl,
 	.release	= esoc_dev_release,
 };

 int esoc_clink_add_device(struct device *dev, void *dummy)
 {
 	struct esoc_udev *esoc_udev;
 	struct esoc_clink *esoc_clink = to_esoc_clink(dev);

 	esoc_udev = get_free_esoc_udev(esoc_clink);
 	if (IS_ERR_OR_NULL(esoc_udev))
 		return PTR_ERR(esoc_udev);
 	esoc_udev->dev = device_create(esoc_class, &esoc_clink->dev,
 					MKDEV(esoc_major, esoc_clink->id),
 					esoc_clink, "esoc-%d", esoc_clink->id);
 	if (IS_ERR_OR_NULL(esoc_udev->dev)) {
 		pr_err("failed to create user device\n");
 		goto dev_err;
 	}
 	return 0;
 dev_err:
 	return_esoc_udev(esoc_udev);
 	return -ENODEV;
 }

 int esoc_clink_del_device(struct device *dev, void *dummy)
 {
 	struct esoc_udev *esoc_udev;
 	struct esoc_clink *esoc_clink = to_esoc_clink(dev);

 	esoc_udev = esoc_udev_get_by_minor(esoc_clink->id);
 	if (!esoc_udev)
 		return 0;
 	return_esoc_udev(esoc_udev);
 	device_destroy(esoc_class, MKDEV(esoc_major, esoc_clink->id));
 	return_esoc_udev(esoc_udev);
 	return 0;
 }

 static int esoc_dev_notifier_call(struct notifier_block *nb,
 					unsigned long action,
 					void *data)
 {
 	struct device *dev = data;

 	switch (action) {
 	case BUS_NOTIFY_ADD_DEVICE:
 		return esoc_clink_add_device(dev, NULL);
 	case BUS_NOTIFY_DEL_DEVICE:
 		return esoc_clink_del_device(dev, NULL);
 	};
 	return 0;
 }

 static struct notifier_block esoc_dev_notifier = {
 	.notifier_call = esoc_dev_notifier_call,
 };

 int __init esoc_dev_init(void)
 {
 	int ret = 0;

 	esoc_class = class_create(THIS_MODULE, "esoc-dev");
 	if (IS_ERR_OR_NULL(esoc_class)) {
 		pr_err("coudn't create class");
 		return PTR_ERR(esoc_class);
 	}
 	esoc_major = register_chrdev(0, "esoc", &esoc_dev_fops);
 	if (esoc_major < 0) {
 		pr_err("failed to allocate char dev\n");
 		ret = esoc_major;
 		goto class_unreg;
 	}
 	ret = bus_register_notifier(&esoc_bus_type, &esoc_dev_notifier);
 	if (ret)
 		goto chrdev_unreg;
 	esoc_for_each_dev(NULL, esoc_clink_add_device);
 	return ret;
 chrdev_unreg:
 	unregister_chrdev(esoc_major, "esoc");
 class_unreg:
 	class_destroy(esoc_class);
 	return 0;
 }

 void __exit esoc_dev_exit(void)
 {
 	bus_unregister_notifier(&esoc_bus_type, &esoc_dev_notifier);
 	class_destroy(esoc_class);
 	unregister_chrdev(esoc_major, "esoc-dev");
 }

 MODULE_LICENSE("GPL v2");
 module_init(esoc_dev_init);
 module_exit(esoc_dev_exit);
	/* Copyright (c) 2013-2014, 2017, The Linux Foundation. All rights reserved.
	*
	* This program is free software; you can redistribute it and/or modify
	* it under the terms of the GNU General Public License version 2 and
	* only version 2 as published by the Free Software Foundation.
	*
	* This program is distributed in the hope that it will be useful,
	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	* GNU General Public License for more details.
	*/
	#include <linux/kfifo.h>
	#include <linux/list.h>
	#include <linux/sched.h>
	#include <linux/uaccess.h>
	#include <linux/wait.h>
	#include "esoc.h"

	/**
	* struct esoc_udev: Userspace char interface
	* @dev: interface device.
	* @req_fifio: fifo for clink requests.
	* @req_wait: signal availability of request from clink
	* @req_fifo_lock: serialize access to req fifo
	* @evt_fito: fifo for clink events
	* @evt_wait: signal availability of clink event
	* @evt_fifo_lock: serialize access to event fifo
	* @list: entry in esoc dev list.
	* @clink: reference to contorl link
	*/
	struct esoc_udev {
	struct device *dev;
	struct kfifo req_fifo;
	wait_queue_head_t req_wait;
	spinlock_t req_fifo_lock;
	struct kfifo evt_fifo;
	wait_queue_head_t evt_wait;
	spinlock_t evt_fifo_lock;
	struct list_head list;
	struct esoc_clink *clink;
	};

	/**
	* struct esoc_uhandle: Userspace handle of esoc
	* @esoc_clink: esoc control link.
	* @eng: esoc engine for commands/ requests.
	* @esoc_udev: user interface device.
	* @req_eng_reg: indicates if engine is registered as request eng
	* @cmd_eng_reg: indicates if engine is registered as cmd eng
	*/
	struct esoc_uhandle {
	struct esoc_clink *esoc_clink;
	struct esoc_eng eng;
	struct esoc_udev *esoc_udev;
	bool req_eng_reg;
	bool cmd_eng_reg;
	};

	#define ESOC_MAX_MINOR 256
	#define ESOC_MAX_REQ 8
	#define ESOC_MAX_EVT 4

	static LIST_HEAD(esoc_udev_list);
	static DEFINE_SPINLOCK(esoc_udev_list_lock);
	struct class *esoc_class;
	static int esoc_major;

	static struct esoc_udev get_free_esoc_udev(struct esoc_clink esoc_clink)
	{
	struct esoc_udev *esoc_udev;
	int err;

	if (esoc_clink->id > ESOC_MAX_MINOR) {
	pr_err("too many esoc devices\n");
	return ERR_PTR(-ENODEV);
	}
	esoc_udev = kzalloc(sizeof(*esoc_udev), GFP_KERNEL);
	if (!esoc_udev)
	return ERR_PTR(-ENOMEM);
	err = kfifo_alloc(&esoc_udev->req_fifo, (sizeof(u32)) * ESOC_MAX_REQ,
	GFP_KERNEL);
	if (err) {
	pr_err("unable to allocate request fifo for %s\n",
	esoc_clink->name);
	goto req_fifo_fail;
	}
	err = kfifo_alloc(&esoc_udev->evt_fifo, (sizeof(u32)) * ESOC_MAX_EVT,
	GFP_KERNEL);
	if (err) {
	pr_err("unable to allocate evt fifo for %s\n",
	esoc_clink->name);
	goto evt_fifo_fail;
	}
	init_waitqueue_head(&esoc_udev->req_wait);
	init_waitqueue_head(&esoc_udev->evt_wait);
	spin_lock_init(&esoc_udev->req_fifo_lock);
	spin_lock_init(&esoc_udev->evt_fifo_lock);
	esoc_udev->clink = esoc_clink;
	spin_lock(&esoc_udev_list_lock);
	list_add_tail(&esoc_udev->list, &esoc_udev_list);
	spin_unlock(&esoc_udev_list_lock);
	return esoc_udev;
	evt_fifo_fail:
	kfifo_free(&esoc_udev->req_fifo);
	req_fifo_fail:
	kfree(esoc_udev);
	return ERR_PTR(-ENODEV);
	}

	static void return_esoc_udev(struct esoc_udev *esoc_udev)
	{
	spin_lock(&esoc_udev_list_lock);
	list_del(&esoc_udev->list);
	spin_unlock(&esoc_udev_list_lock);
	kfifo_free(&esoc_udev->req_fifo);
	kfifo_free(&esoc_udev->evt_fifo);
	kfree(esoc_udev);
	}

	static struct esoc_udev *esoc_udev_get_by_minor(unsigned int index)
	{
	struct esoc_udev *esoc_udev;

	spin_lock(&esoc_udev_list_lock);
	list_for_each_entry(esoc_udev, &esoc_udev_list, list) {
	if (esoc_udev->clink->id == index)
	goto found;
	}
	esoc_udev = NULL;
	found:
	spin_unlock(&esoc_udev_list_lock);
	return esoc_udev;
	}

	void esoc_udev_handle_clink_req(enum esoc_req req, struct esoc_eng *eng)
	{
	int err;
	u32 clink_req;
	struct esoc_clink *esoc_clink = eng->esoc_clink;
	struct esoc_udev *esoc_udev = esoc_udev_get_by_minor(esoc_clink->id);

	if (!esoc_udev)
	return;
	clink_req = (u32)req;
	err = kfifo_in_spinlocked(&esoc_udev->req_fifo, &clink_req,
	sizeof(clink_req),
	&esoc_udev->req_fifo_lock);
	if (err != sizeof(clink_req)) {
	pr_err("unable to queue request for %s\n", esoc_clink->name);
	return;
	}
	wake_up_interruptible(&esoc_udev->req_wait);
	}

	void esoc_udev_handle_clink_evt(enum esoc_evt evt, struct esoc_eng *eng)
	{
	int err;
	u32 clink_evt;
	struct esoc_clink *esoc_clink = eng->esoc_clink;
	struct esoc_udev *esoc_udev = esoc_udev_get_by_minor(esoc_clink->id);

	if (!esoc_udev)
	return;
	clink_evt = (u32)evt;
	err = kfifo_in_spinlocked(&esoc_udev->evt_fifo, &clink_evt,
	sizeof(clink_evt),
	&esoc_udev->evt_fifo_lock);
	if (err != sizeof(clink_evt)) {
	pr_err("unable to queue event for %s\n", esoc_clink->name);
	return;
	}
	wake_up_interruptible(&esoc_udev->evt_wait);
	}

	static long esoc_dev_ioctl(struct file *file, unsigned int cmd,
	unsigned long arg)
	{
	int err;
	u32 esoc_cmd, status, req, evt;
	struct esoc_uhandle *uhandle = file->private_data;
	struct esoc_udev *esoc_udev = uhandle->esoc_udev;
	struct esoc_clink *esoc_clink = uhandle->esoc_clink;
	const struct esoc_clink_ops * const clink_ops = esoc_clink->clink_ops;
	void __user uarg = (void __user )arg;

	switch (cmd) {
	case ESOC_REG_REQ_ENG:
	err = esoc_clink_register_req_eng(esoc_clink, &uhandle->eng);
	if (err)
	return err;
	uhandle->req_eng_reg = true;
	break;
	case ESOC_REG_CMD_ENG:
	err = esoc_clink_register_cmd_eng(esoc_clink, &uhandle->eng);
	if (err)
	return err;
	uhandle->cmd_eng_reg = true;
	break;
	case ESOC_CMD_EXE:
	if (esoc_clink->cmd_eng != &uhandle->eng)
	return -EACCES;
	get_user(esoc_cmd, (u32 __user *)arg);
	return clink_ops->cmd_exe(esoc_cmd, esoc_clink);
	case ESOC_WAIT_FOR_REQ:
	if (esoc_clink->req_eng != &uhandle->eng)
	return -EACCES;
	err = wait_event_interruptible(esoc_udev->req_wait,
	!kfifo_is_empty(&esoc_udev->req_fifo));
	if (!err) {
	err = kfifo_out_spinlocked(&esoc_udev->req_fifo, &req,
	sizeof(req),
	&esoc_udev->req_fifo_lock);
	if (err != sizeof(req)) {
	pr_err("read from clink %s req q failed\n",
	esoc_clink->name);
	return -EIO;
	}
	put_user(req, (unsigned int __user *)uarg);

	}
	return err;
	case ESOC_NOTIFY:
	get_user(esoc_cmd, (u32 __user *)arg);
	clink_ops->notify(esoc_cmd, esoc_clink);
	break;
	case ESOC_GET_STATUS:
	err = clink_ops->get_status(&status, esoc_clink);
	if (err)
	return err;
	put_user(status, (unsigned int __user *)uarg);
	break;
	case ESOC_WAIT_FOR_CRASH:
	err = wait_event_interruptible(esoc_udev->evt_wait,
	!kfifo_is_empty(&esoc_udev->evt_fifo));
	if (!err) {
	err = kfifo_out_spinlocked(&esoc_udev->evt_fifo, &evt,
	sizeof(evt),
	&esoc_udev->evt_fifo_lock);
	if (err != sizeof(evt)) {
	pr_err("read from clink %s evt q failed\n",
	esoc_clink->name);
	return -EIO;
	}
	put_user(evt, (unsigned int __user *)uarg);
	}
	return err;
	default:
	return -EINVAL;
	};
	return 0;
	}

	static int esoc_dev_open(struct inode inode, struct file file)
	{
	struct esoc_uhandle *uhandle;
	struct esoc_udev *esoc_udev;
	struct esoc_clink *esoc_clink;
	struct esoc_eng *eng;
	unsigned int minor = iminor(inode);

	esoc_udev = esoc_udev_get_by_minor(minor);
	if (!esoc_udev) {
	pr_err("failed to get udev\n");
	return -ENOMEM;
	}

	esoc_clink = get_esoc_clink(esoc_udev->clink->id);
	if (!esoc_clink) {
	pr_err("failed to get clink\n");
	return -ENOMEM;
	}

	uhandle = kzalloc(sizeof(*uhandle), GFP_KERNEL);
	if (!uhandle) {
	put_esoc_clink(esoc_clink);
	return -ENOMEM;
	}
	uhandle->esoc_udev = esoc_udev;
	uhandle->esoc_clink = esoc_clink;
	eng = &uhandle->eng;
	eng->handle_clink_req = esoc_udev_handle_clink_req;
	eng->handle_clink_evt = esoc_udev_handle_clink_evt;
	file->private_data = uhandle;
	return 0;
	}

	static int esoc_dev_release(struct inode inode, struct file file)
	{
	struct esoc_clink *esoc_clink;
	struct esoc_uhandle *uhandle = file->private_data;

	esoc_clink = uhandle->esoc_clink;
	if (uhandle->req_eng_reg)
	esoc_clink_unregister_req_eng(esoc_clink, &uhandle->eng);
	if (uhandle->cmd_eng_reg)
	esoc_clink_unregister_cmd_eng(esoc_clink, &uhandle->eng);
	uhandle->req_eng_reg = false;
	uhandle->cmd_eng_reg = false;
	put_esoc_clink(esoc_clink);
	kfree(uhandle);
	return 0;
	}
	static const struct file_operations esoc_dev_fops = {
	.owner = THIS_MODULE,
	.open = esoc_dev_open,
	.unlocked_ioctl = esoc_dev_ioctl,
	.release = esoc_dev_release,
	};

	int esoc_clink_add_device(struct device dev, void dummy)
	{
	struct esoc_udev *esoc_udev;
	struct esoc_clink *esoc_clink = to_esoc_clink(dev);

	esoc_udev = get_free_esoc_udev(esoc_clink);
	if (IS_ERR_OR_NULL(esoc_udev))
	return PTR_ERR(esoc_udev);
	esoc_udev->dev = device_create(esoc_class, &esoc_clink->dev,
	MKDEV(esoc_major, esoc_clink->id),
	esoc_clink, "esoc-%d", esoc_clink->id);
	if (IS_ERR_OR_NULL(esoc_udev->dev)) {
	pr_err("failed to create user device\n");
	goto dev_err;
	}
	return 0;
	dev_err:
	return_esoc_udev(esoc_udev);
	return -ENODEV;
	}

	int esoc_clink_del_device(struct device dev, void dummy)
	{
	struct esoc_udev *esoc_udev;
	struct esoc_clink *esoc_clink = to_esoc_clink(dev);

	esoc_udev = esoc_udev_get_by_minor(esoc_clink->id);
	if (!esoc_udev)
	return 0;
	return_esoc_udev(esoc_udev);
	device_destroy(esoc_class, MKDEV(esoc_major, esoc_clink->id));
	return_esoc_udev(esoc_udev);
	return 0;
	}

	static int esoc_dev_notifier_call(struct notifier_block *nb,
	unsigned long action,
	void *data)
	{
	struct device *dev = data;

	switch (action) {
	case BUS_NOTIFY_ADD_DEVICE:
	return esoc_clink_add_device(dev, NULL);
	case BUS_NOTIFY_DEL_DEVICE:
	return esoc_clink_del_device(dev, NULL);
	};
	return 0;
	}

	static struct notifier_block esoc_dev_notifier = {
	.notifier_call = esoc_dev_notifier_call,
	};

	int __init esoc_dev_init(void)
	{
	int ret = 0;

	esoc_class = class_create(THIS_MODULE, "esoc-dev");
	if (IS_ERR_OR_NULL(esoc_class)) {
	pr_err("coudn't create class");
	return PTR_ERR(esoc_class);
	}
	esoc_major = register_chrdev(0, "esoc", &esoc_dev_fops);
	if (esoc_major < 0) {
	pr_err("failed to allocate char dev\n");
	ret = esoc_major;
	goto class_unreg;
	}
	ret = bus_register_notifier(&esoc_bus_type, &esoc_dev_notifier);
	if (ret)
	goto chrdev_unreg;
	esoc_for_each_dev(NULL, esoc_clink_add_device);
	return ret;
	chrdev_unreg:
	unregister_chrdev(esoc_major, "esoc");
	class_unreg:
	class_destroy(esoc_class);
	return 0;
	}

	void __exit esoc_dev_exit(void)
	{
	bus_unregister_notifier(&esoc_bus_type, &esoc_dev_notifier);
	class_destroy(esoc_class);
	unregister_chrdev(esoc_major, "esoc-dev");
	}

	MODULE_LICENSE("GPL v2");
	module_init(esoc_dev_init);
	module_exit(esoc_dev_exit);