1. 前言

前面分析i2c-tool测试工具就是基于drivers/i2c/i2c-dev.c驱动来实现的。i2c-dev驱动在加载时会遍历所有的I2C总线(i2c_bus_type)上所有注册的adapter，并且在linux系统创建对应的字符设备，如：/dev/i2c-0、/dev/i2c-1、/dev/i2c-2等。应用程序通过open打开对应的i2c字符设备，通过ioctl来收发数据。具体的架构如下图：

2. i2c-dev驱动的注册过程

在i2c-dev.c的驱动入口i2c_dev_init函数具体操作如下：

1. register_chrdev_region：注册i2c字符设备
2. class_create：创建i2c-dev的class，为在linux文件系统中创建字符设备做准备。
3. i2cdev_attach_adapter：通过函数i2c_for_each_dev遍历已经绑定的adapter，有多少个adapter就调用i2cdev_attach_adapter函数几次。
   
4. to_i2c_adapter：通过dev获取对应的i2c adapter。
5. cdev_init：它会初始化一个重要的结构体，file_operations。
6. dev_set_name：设置device name为i2c-x，也就是我们在字符设备创建成功后看到的/dev/i2c-x设备。
7. cdev_device_add：添加设备到系统，并且创建对应的字符设备到用户空间。
   

备注：
cdev_device_add这里其实调用了cdev_add和device_add。然而，device_create()是device_register()的封装，而device_register()则是device_add()的封装。

3. open_i2c_dev函数分析

open_i2c_dev是i2c-tool工具open i2c-dev驱动的函数，根据传递的参数最终重要函数应该如下：
file = open("/dev/i2c-0", O_RDWR);
该函数的主要代码如下：

/* File Path = i2c-tools-4.3/tools/i2cbusses.c */
int open_i2c_dev(int i2cbus, char *filename, size_t size, int quiet)
{
	int file, len;

	len = snprintf(filename, size, "/dev/i2c/%d", i2cbus);
	if (len >= (int)size) {
		fprintf(stderr, "%s: path truncated\n", filename);
		return -EOVERFLOW;
	}
	file = open(filename, O_RDWR);

	if (file < 0 && (errno == ENOENT || errno == ENOTDIR)) {
		len = snprintf(filename, size, "/dev/i2c-%d", i2cbus);
		if (len >= (int)size) {
			fprintf(stderr, "%s: path truncated\n", filename);
			return -EOVERFLOW;
		}
		file = open(filename, O_RDWR);
	}
	...
	return file;
}

应用层调用open后，会对应调用i2c-dev通用驱动的open函数。主要是如下几个步骤：

1. to_i2c_adapter：通过minor次设备号，其实这里等同于i2c总线编号。通过它来获取对应总线的adapter。
2. kzalloc：申请一个i2c client表示I2C设备，并且初始化该client的name和保存adapter与其建立联系。但是，整个open函数这里没有对I2C地址进行初始化。
3. file->private_data：通过private_data保存申请的client地址，为了后面read/write/ioctl可以通过file->private_data很方便的拿到当前dev的client。

/* File Path = kernel/drivers/i2c/i2c-dev.c */
static int i2cdev_open(struct inode *inode, struct file *file)
{
	unsigned int minor = iminor(inode);
	struct i2c_client *client;
	struct i2c_adapter *adap;

	adap = i2c_get_adapter(minor);
	if (!adap)
		return -ENODEV;

	/* This creates an anonymous i2c_client, which may later be
	 * pointed to some address using I2C_SLAVE or I2C_SLAVE_FORCE.
	 *
	 * This client is ** NEVER REGISTERED ** with the driver model
	 * or I2C core code!!  It just holds private copies of addressing
	 * information and maybe a PEC flag.
	 */
	client = kzalloc(sizeof(*client), GFP_KERNEL);
	if (!client) {
		i2c_put_adapter(adap);
		return -ENOMEM;
	}
	snprintf(client->name, I2C_NAME_SIZE, "i2c-dev %d", adap->nr);

	client->adapter = adap;
	file->private_data = client;

	return 0;
}

4. set_slave_addr函数分析

open_i2c_dev是i2c-tool工具来设置I2C的设备地址。具体代码如下：

/* File Path = i2c-tools-4.3/tools/i2cbusses.c */
int set_slave_addr(int file, int address, int force)
{
	/* With force, let the user read from/write to the registers
	   even when a driver is also running */
	if (ioctl(file, force ? I2C_SLAVE_FORCE : I2C_SLAVE, address) < 0) {
		fprintf(stderr,
			"Error: Could not set address to 0x%02x: %s\n",
			address, strerror(errno));
		return -errno;
	}

	return 0;
}

应用层调用ioctl I2C_SLAVE_FORCE后，会对应调用i2c-dev通用驱动对应的ioctl I2C_SLAVE_FORCE。主要是如下2个步骤：

1. client = file->private_data：从private_data中获取当前I2C设备。
2. client->addr = arg：设置当前I2C设备的地址信息，方便后面的read/write操作I2C设备。

/* File Path = kernel/drivers/i2c/i2c-dev.c */
static long i2cdev_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
	struct i2c_client *client = file->private_data;
	unsigned long funcs;

	switch (cmd) {
	case I2C_SLAVE:
	case I2C_SLAVE_FORCE:
		if ((arg > 0x3ff) ||
		    (((client->flags & I2C_M_TEN) == 0) && arg > 0x7f))
			return -EINVAL;
		if (cmd == I2C_SLAVE && i2cdev_check_addr(client->adapter, arg))
			return -EBUSY;
		/* REVISIT: address could become busy later */
		client->addr = arg;
		return 0;

	...
}

5. i2c_read_bytes函数分析

i2c_read_bytes这个函数是我自己写的demo code，这里只列了read部分来分析，完整的demo code可以查看上一篇我的博客。主要是如下2个步骤：

1. struct i2c_msg messages[2]：定义2个i2c_msg的消息体，并且初始化它。它包含了：I2C的从机设备地址、read/write的flag、数据的长度、存放数据的地址。msg[0]主要是为了发送需要读取从机的哪个寄存器，msg[1]主要的配置读取的数据长度的和存放数据的地址。
2. struct i2c_rdwr_ioctl_data packets：它主要是存放i2c_msg消息地址和i2c_msg消息个数。
3. ioctl：通过ioctl发送到内核。

static int i2c_read_bytes(int fd, uint8_t slave_addr, uint8_t reg_addr, uint8_t *values, uint8_t len)
{
    uint8_t outbuf[1];
    struct i2c_rdwr_ioctl_data packets;
    struct i2c_msg messages[2];

    outbuf[0] = reg_addr;
    messages[0].addr = slave_addr;
    messages[0].flags = 0;
    messages[0].len = sizeof(outbuf);
    messages[0].buf = outbuf;
    
    /* The data will get returned in this structure */
    messages[1].addr = slave_addr;
    messages[1].flags = I2C_M_RD/* | I2C_M_NOSTART*/;
    messages[1].len = len;
    messages[1].buf = values;
    
    /* Send the request to the kernel and get the result back */
    packets.msgs = messages;
    packets.nmsgs = 2;
    if(ioctl(fd, I2C_RDWR, &packets) < 0)
    {
        printf("Error: Unable to send data");
        return -1;
    }
    
    return 0;
}

应用层调用ioctl I2C_RDWR后，会对应调用i2c-dev通用驱动对应的ioctl I2C_RDWR。它将应用层的数据到拷贝的内核，具体如下：

/* File Path = kernel/drivers/i2c/i2c-dev.c */
case I2C_RDWR: {
	struct i2c_rdwr_ioctl_data rdwr_arg;
	struct i2c_msg *rdwr_pa;
	
	//从用户空间拷贝数据到内核空间
	copy_from_user(&rdwr_arg,(struct i2c_rdwr_ioctl_data __user *)arg,sizeof(rdwr_arg));
	//分配一块内存空间，将用户空间的数据拷贝进去
	rdwr_pa = memdup_user(rdwr_arg.msgs, rdwr_arg.nmsgs * sizeof(struct i2c_msg));

	return i2cdev_ioctl_rdwr(client, rdwr_arg.nmsgs, rdwr_pa);
}

static int i2cdev_ioctl_rdwr(struct i2c_client *client, unsigned nmsgs, struct i2c_msg *msgs)
{
	***
	res = i2c_transfer(client->adapter, msgs, nmsgs);
	***
}

总结：其实应用层初始化的i2c_msg就是直接给内核i2c_transfer将I2C消息发送出去的。