SPI protocol驱动编写实例

内核版本:3.9.5

Linux中SPI驱动有俩个部分组成:controller驱动,直接和底层硬件打交道,protocol驱动,针对特定的设备,也是我们要做的.

这里只考虑SPI protocol驱动

在主线内核中有一个通用的字符型驱动范例spidev.本文不讨论spidev,而是探讨如何编写一个自定义SPI设备驱动.

为何要编写一个自定义驱动:

• 可以掌控性能;
• 无需给内核打补丁.SPI框架完整支持在可加载内核模块中进行驱动设置.这也就允许我们在内核源码树外进行编写代码;

Kernel Configuration

关掉SPIDEV选项以支持自定义的驱动

Device Drivers-->

SPI support-->

*** SPI Protocol Masters ***

<> User mode SPIdevice driver support

为使linuxSPI系统识别你的驱动,需要做两步准备.

1. 在特定的SPI总线(一个master就是一条总线)上注册SPI slave设备.可以是内核初始化的时候亦或是在你的驱动代码中动态的进行.设备名称在这步中必须指定;

2. 注册SPI protocol驱动.使用的名字必须和步骤1中的一致,这样kernel就可以将它们链接在一起了;

两个步骤可以以任意次序进行,但是得知道当他们完成之后,SPI框架就可以响应你驱动中的probe()调用了.你将得到一个spi_device用来和SPI系统进行交互了.一旦你的probe函数调用成功,你就可以开始使用SPI总线了.

静态注册SPI设备涵盖在spidev中了.

动态主册SPI设备的步骤如下:

1. 得到管理总线的spi_master控制器指针(句柄);

2. 为总线分配spi_device结构;

3. 验证没有其他的设备已经在这条总线bus.cs上注册过了;

4. 使用设备特定的值(speed,datasize,etc)来填充spi_device;

5. 将新的spi_device添加到总线;

有个范例代码:

 static int __init add_spike_device_to_bus(void)
 {
     struct spi_master *spi_master;
     struct spi_device *spi_device;
     struct device *pdev;
     char buff[];
     int status = ;
 
     spi_master = spi_busnum_to_master(SPI_BUS);
     if (!spi_master) {
         printk(KERN_ALERT "spi_busnum_to_master(%d) returned NULL\n",
             SPI_BUS);
         printk(KERN_ALERT "Missing modprobe omap2_mcspi?\n");
         return -;
     }
 
     spi_device = spi_alloc_device(spi_master);
     if (!spi_device) {
         put_device(&spi_master->dev);
         printk(KERN_ALERT "spi_alloc_device() failed\n");
         return -;
     }
 
     /* specify a chip select line */
     spi_device->chip_select = SPI_BUS_CS1;
 
     /* Check whether this SPI bus.cs is already claimed */
     snprintf(buff, sizeof(buff), "%s.%u",
             dev_name(&spi_device->master->dev),
             spi_device->chip_select);
 
     pdev = bus_find_device_by_name(spi_device->dev.bus, NULL, buff);
      if (pdev) {
         /* We are not going to use this spi_device, so free it */
         spi_dev_put(spi_device);
 
         /*
          * There is already a device configured for this bus.cs combination.
          * It's okay if it's us. This happens if we previously loaded then
                  * unloaded our driver.
                  * If it is not us, we complain and fail.
          */
         if (pdev->driver && pdev->driver->name &&
                 strcmp(this_driver_name, pdev->driver->name)) {
             printk(KERN_ALERT
                 "Driver [%s] already registered for %s\n",
                 pdev->driver->name, buff);
             status = -;
         }
     } else {
         spi_device->max_speed_hz = SPI_BUS_SPEED;
         spi_device->mode = SPI_MODE_0;
         spi_device->bits_per_word = ;
         spi_device->irq = -;
         spi_device->controller_state = NULL;
         spi_device->controller_data = NULL;
         strlcpy(spi_device->modalias, this_driver_name, SPI_NAME_SIZE);
         status = spi_add_device(spi_device);
 
         if (status < ) {
             spi_dev_put(spi_device);
             printk(KERN_ALERT "spi_add_device() failed: %d\n",
                 status);
         }
     }
 
     put_device(&spi_master->dev);
 
     return status;
 }

稍微有点复杂,这是个完整的样板.你完全可以只更改spi_device中指定域来服务你自己的驱动.

注册spi_driver,首先要用驱动名字还有些回调函数来初始化他.至少你得提供一个probe()函数,让他能够找到他的另一半spi_device.

范例:

 static struct spi_driver spike_driver = {
     .driver = {
         .name =this_driver_name,
         .owner =THIS_MODULE,
     },
     .probe =spike_probe,
     .remove =spike_remove,
 };

具体的也可参考linux/spi/spi.h文件

接着就可一在某个地方调用spi_register_driver()进行注册了.

 spi_register_driver(&spike_driver); 

本文转自:http://blog.csdn.net/wuhzossibility/article/details/7831652