SPI设备的驱动
主要包括两个SPI设备步骤:
register_chrdev
spi_register_driver
关键点1:
spi_board_info
可以去已经运行的板子下面找例子:
/sys/bus/spi/drivers
已辰汉电子MX27 MDK 开发板为例:
在/sys/bus/spi/drivers目录:lcd_spi pmic_spi
对应的:mx27mdk27v0.c文件中定义如下:
static struct spi_board_info mxc_spi_board_info[] __initdata = {
{
.modalias = "pmic_spi",
.irq = IOMUX_TO_IRQ(MX27_PIN_TOUT),
.max_speed_hz = 4000000,
.bus_num = 2,
.chip_select = 0,
},
{
.modalias = "lcd_spi",
.max_speed_hz = 1000000,
.bus_num = 1,
.chip_select = 0,
},
};
spi_register_board_info(mxc_spi_board_info,
ARRAY_SIZE(mxc_spi_board_info));
问题:
1. spi_register_board_info函数与 mxc_spi_probe 调用的先后关系?
2 . mxc_spi_board_info 结构体变量 irq 的意义?
3. 在int spi_sync(struct spi_device *spi, struct spi_message *message)
{
DECLARE_COMPLETION_ONSTACK(done);
int status;
message->complete = spi_complete;
message->context = &done;
status = spi_async(spi, message);
if (status == 0)
wait_for_completion(&done);
message->context = NULL;
return status;
}
中有wait_for_completion(&done); 当进入到函数
spi_async --> mxc_spi_transfer 内也有 INIT_COMPLETION(master_drv_data->xfer_done);
wait_for_completion(&master_drv_data->xfer_done);
在mxc_spi_isr函数中 complete(&master_drv_data->xfer_done)
那么在主函数中wait_for_completion(&done); 的done 什么时候被wait到?
答案:还有个spi_bitbang.c bitbang_work函数中处理 m->complete(m->context);
spi驱动中还有一处研究点,其spi发送是通过:
queue_work(bitbang->workqueue, &bitbang->work);
设定的。值得借鉴.
主要包括两个SPI设备,at45db321d和mcp2515,一个是串行的dataflash,一个是can总线设备芯片。前者对于我们来说非常重要,我们可以借助该设备对uboot和kernel以及根文件系统进行更新。
预备知识:设备和驱动是如何匹配的?系统的热插拔是如何实现的?
首先一点,设备和驱动是严格区分的,设备是设备,驱动是驱动,设备通过struct device来定义,当然用户也可以将该结构体封装到自己定义的device结构体中,例如,struct platform_device,这是我们采用platform_bus_type总线的设备定义的结构体形式:
include/linux/platform_device.h文件中:
struct platform_device {
const char * name;
u32 id;
struct device dev;
u32 num_resources;
struct resource * resource;
};
只要是9260的外围模块,就像IIC硬件控制器,SPI硬件控制器,都被完全的定义成这种结构体的格式,这种结构体主要包含了硬件资源和名称,硬件资源分为寄存器和IRQ两种。platform_device通过向内核注册struct device dev这个结构体来告诉内核加载这个设备,
方法就是 device_register(&platform_device->dev)
内核不关心你使用的是platform_device还是spi_device,内核只关心你的struct device结构体,内核通过这个struct device结构体自然能够顺藤摸瓜找到你是platform_device还是spi_device,这就是linux最引以为傲的contian_of()大法。
驱动通过struct driver这个结构体来定义,与struct device一致,你也可以用自己的结构体去封装:例如,struct platform_driver。
include/linux/platform_device.h文件中:
struct platform_driver {
int (*probe)(struct platform_device *);
int (*remove)(struct platform_device *);
void (*shutdown)(struct platform_device *);
int (*suspend)(struct platform_device *, pm_message_t state);
int (*suspend_late)(struct platform_device *, pm_message_t state);
int (*resume_early)(struct platform_device *);
int (*resume)(struct platform_device *);
struct device_driver driver;
};
与device一致,应用程序通过driver_register(&platform_driver->driver)向内核中注册你当前的驱动,而内核不关心你封装成的结构,而内核搜索的方法还是同样的contain_of大法。
系统如何将这两者匹配上的?而不会将iic的设备加载到spi的驱动上面?不会将这个iic设备的驱动加载到那个iic设备上,设备和驱动之间是如何联系的?总线,这就是总线的作用!
include/linux/device.h文件中有总线类型的定义。
struct bus_type {
const char * name;
struct subsystem subsys;
struct kset drivers;
struct kset devices;
struct klist klist_devices;
struct klist klist_drivers;
struct blocking_notifier_head bus_notifier;
struct bus_attribute * bus_attrs;
struct device_attribute * dev_attrs;
struct driver_attribute * drv_attrs;
int (*match)(struct device * dev, struct device_driver * drv);
int (*uevent)(struct device *dev, char **envp,
int num_envp, char *buffer, int buffer_size);
int (*probe)(struct device * dev);
int (*remove)(struct device * dev);
void (*shutdown)(struct device * dev);
int (*suspend)(struct device * dev, pm_message_t state);
int (*suspend_late)(struct device * dev, pm_message_t state);
int (*resume_early)(struct device * dev);
int (*resume)(struct device * dev);
};
这个总线设备中最重要的可能是match成员,由于我们一般很少去建立一个新的总线,所以我们很少涉及总线的编程,我们就只关注我们所关注的。
总线如何将两者关联起来,热插拔大家知道吧,当一个设备被通过device_register注册到内核中时,会导致一个热插拔事件产生,系统会遍历该总线上的所有驱动程序,调用bus的match算法,来寻找与该设备相匹配的驱动程序,当一个驱动注册到内核的时候,处理过程与此相似(这是我理解的阿,大家批评指正),而一般的macth算法都比较简单,例如platform_bus的匹配算法就很简单,就是比较platform_device和platform_driver的name成员,如果匹配成功,就加载相应的设备或者驱动!这就完成了一个连接的过程。。。
那么这两种设备驱动中最重要的类型在linux中如何表现出来,那我们就有必要介绍一下从2.6开始实现的sys文件系统了,
/sys/bus $ cat /etc/fstab
proc /proc proc defaults 0 0
devpts /dev/pts devpts defaults 0 0
tmpfs /dev/shm tmpfs defaults 0 0
sysfs /sys sysfs defaults 0 0
/dev/mtdblock2 /mnt/flash2 yaffs defaults 0 0
加载这个文件系统对于我们分析设备模型是非常有好处的。
sys文件夹下一般有如下的目录:
/sys $ ls -al
drwxr-xr-x 10 root root 0 Jan 1 1970 .
drwxrwxrwx 11 1000 tao 4096 May 22 06:56 ..
drwxr-xr-x 7 root root 0 Oct 27 14:09 block
drwxr-xr-x 8 root root 0 Jan 1 1970 bus
drwxr-xr-x 21 root root 0 Jan 1 1970 class
drwxr-xr-x 4 root root 0 Jan 1 1970 devices
drwxr-xr-x 2 root root 0 Jan 1 1970 firmware
drwxr-xr-x 2 root root 0 Jan 1 1970 fs
drwxr-xr-x 2 root root 0 Jan 1 1970 kernel
drwxr-xr-x 22 root root 0 Oct 27 14:10 module
block是由于历史原因形成的block设备的文件夹。我们关心的是bus文件夹。
我们以spi设备为例:spi部分要包括两种设备,一种是platform_device,一种是spi_device。
在arch/arm/mach-at91/at91sam9260_device.c文件中,定义的SPI硬件控制模块设备,这我们不需要关心。
还有一种是spi_device,定义在arch/arm/mach-at91/board-sam9260ek.c文件中,这就是我们的dataflash和mcp2515设备,
所以如何设备加载成功的话,在bus下面的每个目录里面,都存在devices和drivers两个文件夹,分别对应设备和文件。
/sys/bus/platform/devices $ ls -al
drwxr-xr-x 2 root root 0 Oct 27 16:01 .
drwxr-xr-x 4 root root 0 Jan 1 1970 ..
lrwxrwxrwx 1 root root 0 Oct 27 16:01 at91_i2c -> ../../../devices/platform/at91_i2c
lrwxrwxrwx 1 root root 0 Oct 27 16:01 at91_nand -> ../../../devices/platform/at91_nand
lrwxrwxrwx 1 root root 0 Oct 27 16:01 at91_ohci -> ../../../devices/platform/at91_ohci
lrwxrwxrwx 1 root root 0 Oct 27 16:01 atmel_spi.0 -> ../../../devices/platform/atmel_spi.0
lrwxrwxrwx 1 root root 0 Oct 27 16:01 atmel_spi.1 -> ../../../devices/platform/atmel_spi.1
lrwxrwxrwx 1 root root 0 Oct 27 16:01 atmel_usart.0 -> ../../../devices/platform/atmel_usart.0
lrwxrwxrwx 1 root root 0 Oct 27 16:01 atmel_usart.1 -> ../../../devices/platform/atmel_usart.1
lrwxrwxrwx 1 root root 0 Oct 27 16:01 atmel_usart.2 -> ../../../devices/platform/atmel_usart.2
lrwxrwxrwx 1 root root 0 Oct 27 16:01 macb -> ../../../devices/platform/macb
驱动
/sys/bus/platform/drivers/atmel_spi $ ls -al
drwxr-xr-x 2 root root 0 Jan 1 1970 .
drwxr-xr-x 8 root root 0 Jan 1 1970 ..
lrwxrwxrwx 1 root root 0 Oct 27 16:10 atmel_spi.0 -> ../../../../devices/platform/atmel_spi.0
lrwxrwxrwx 1 root root 0 Oct 27 16:10 atmel_spi.1 -> ../../../../devices/platform/atmel_spi.1
--w------- 1 root root 4096 Oct 27 16:10 bind
--w------- 1 root root 4096 Oct 27 16:10 unbind
如果出现上面的这个情况,说明你的设备(两路spi总线)和驱动都加载成功了,如果你的devices下面没有spi.0设备和spi.1设备的话,说明
board-sam9260ek.c文件中的这个函数出错:
static void __init ek_board_init(void)
{
/* Serial */
at91_add_device_serial();
/* USB Host */
at91_add_device_usbh(&ek_usbh_data);
/* USB Device */
at91_add_device_udc(&ek_udc_data);
/* SPI */
at91_add_device_spi(ek_spi_devices, ARRAY_SIZE(ek_spi_devices));
/* NAND */
at91_add_device_nand(&ek_nand_data);
/* Ethernet */
at91_add_device_eth(&ek_macb_data);
/* MMC */
at91_add_device_mmc(0, &ek_mmc_data);
/* I2C */
at91_add_device_i2c();
}
这里是设备注册的地方,我们还应该在下面这个目录下看到这两个文件。
/sys/bus/spi/devices $ ls -al
drwxr-xr-x 2 root root 0 Oct 27 14:09 .
drwxr-xr-x 4 root root 0 Jan 1 1970 ..
lrwxrwxrwx 1 root root 0 Oct 27 14:09 spi0.1 -> ../../../devices/platform/atmel_spi.0/spi0.1
lrwxrwxrwx 1 root root 0 Oct 27 14:09 spi1.0 -> ../../../devices/platform/atmel_spi.1/spi1.0
这两个链接说明我们的两个spi设备注册都被接受了,剩下来就是驱动的问题。有人看不懂这个sys文件系统的层次关系,其实这里比较好说明,就是spi0.1是atmel_spi.0设备的子设备嘛,很好理解的。
驱动:
platform_driver驱动:
/sys/bus/platform/drivers $ ls -al
drwxr-xr-x 8 root root 0 Jan 1 1970 .
drwxr-xr-x 4 root root 0 Jan 1 1970 ..
drwxr-xr-x 2 root root 0 Jan 1 1970 at91_i2c
drwxr-xr-x 2 root root 0 Jan 1 1970 at91_nand
drwxr-xr-x 2 root root 0 Jan 1 1970 at91_ohci
drwxr-xr-x 2 root root 0 Oct 27 16:10 atmel_spi
drwxr-xr-x 2 root root 0 Jan 1 1970 atmel_usart
drwxr-xr-x 2 root root 0 Jan 1 1970 macb
我们可以看到这个驱动只有一个atmel_spi,这个驱动是在哪加载的?
driver/spi/atmel_spi.c文件加载的。
spi_driver驱动:
/sys/bus/spi/drivers $ ls -al
drwxr-xr-x 4 root root 0 Oct 27 14:10 .
drwxr-xr-x 4 root root 0 Jan 1 1970 ..
drwxr-xr-x 2 root root 0 Oct 27 14:10 mcp2515
drwxr-xr-x 2 root root 0 Oct 27 14:09 mtd_dataflash
这是我们加载的两个驱动,说明驱动也加载正常了。
下面我们来说说我们遇到的问题吧。
在设备和驱动都加载正常之后,出现与dataflash设备通信不上的情况,驱动加载的时候,读取芯片的状态字读出是0xff,说明工作不正常,动用逻辑分析仪监控spi总线的通信,意外的发现,sck信号和cs信号正常,但是mosi无信号输出,开始觉得可能是spi总线适配器有问题,后来仔细观察原理图之后,发现dataflash和mmc/sd是使用同样的io口的,即pa0,pa1,pa2,而我的内核配置中打开了对mmc的支持,所以导致mosi不正常,所以可能9260的mmc与dataflash不能同时使用,但9263的可以。
解决办法:make menuconfig
Device Drivers--->MMC/SD card support,取消其支持,问题解决!
还有一个问题可能大家没有注意到,没有解释清楚,其实是有问题的,我们的at91_add_device_spi函数如下:
static struct spi_board_info ek_spi_devices[] = {
#if !defined(CONFIG_MMC_AT91)
{ /* DataFlash chip */
.modalias = "mtd_dataflash",
.chip_select = 1,
.max_speed_hz = 15 * 1000 * 1000,
.bus_num = 0,
},
#if defined(CONFIG_MTD_AT91_DATAFLASH_CARD)
{ /* DataFlash card */
.modalias = "mtd_dataflash",
.chip_select = 0,
.max_speed_hz = 15 * 1000 * 1000,
.bus_num = 0,
},
#endif
#endif
#if defined(CONFIG_SND_AT73C213) || defined(CONFIG_SND_AT73C213_MODULE)
{ /* AT73C213 DAC */
.modalias = "at73c213",
.chip_select = 0,
.max_speed_hz = 10 * 1000 * 1000,
.bus_num = 1,
},
#endif
/* spi can ,add by mrz */
#if defined(CONFIG_CAN_MCP2515_MODULE) ||defined(CONFIG_CAN_MCP2515)
//defined(CONFIG_CAN_MCP2515)
{
.modalias = "mcp2515",
.chip_select = 0,
// .controller_data = AT91_PIN_PB3,
.irq = AT91_PIN_PC6, //AT91SAM9260_ID_IRQ0,
.platform_data = &mcp251x_data,
.max_speed_hz = 10 * 1000 * 1000,
.bus_num = 1,
.mode = 0,
},
/*
{
.modalias = "mcp2515",
.chip_select = 1,
// .controller_data = AT91_PIN_PC5,
.irq = AT91_PIN_PC7, //AT91SAM9260_ID_IRQ1,
.platform_data = &mcp251x_data,
.max_speed_hz = 10 * 1000 * 1000,
.bus_num = 1,
.mode = 0,
},
*/
#elif defined(CONFIG_CAN_MCP251X)
{
.modalias = "mcp251x",
.chip_select = 0,
// .controller_data = AT91_PIN_PB3,
.irq = AT91_PIN_PC6, //AT91SAM9260_ID_IRQ0,
.platform_data = &mcp251x_data,
.max_speed_hz = 10 * 1000 * 1000,
.bus_num = 1,
.mode = 0,
},
{
.modalias = "mcp251x",
.chip_select = 1,
// .controller_data = AT91_PIN_PC5,
.irq = AT91_PIN_PC7, //AT91SAM9260_ID_IRQ1,
.platform_data = &mcp251x_data,
.max_speed_hz = 10 * 1000 * 1000,
.bus_num = 1,
.mode = 0,
},
#endif
}
void __init at91_add_device_spi(struct spi_board_info *devices, int nr_devices)
{
int i;
unsigned long cs_pin;
short enable_spi0 = 0;
short enable_spi1 = 0;
/* Choose SPI chip-selects */
/*这里加载我们定义的spi_board_info结构体,也就是两个spi设备的信息,注意,他们这里没有使用spi_device结构体来做,而是使用一个板级信息体来完成。*/
for (i = 0; i < nr_devices; i++) {
/*该成员定义的就是cs引脚*/
if (devices[i].controller_data)
cs_pin = (unsigned long) devices[i].controller_data;
else if (devices[i].bus_num == 0)
cs_pin = spi0_standard_cs[devices[i].chip_select];
else
cs_pin = spi1_standard_cs[devices[i].chip_select];
/*根据需要加载的设备,确定需要打开哪几个SPI控制器,我们系统中有两个控制器,所以我们在以模块的方式加载驱动的时候,我们的设备必须在刚开始就被初始化!*/
if (devices[i].bus_num == 0)
enable_spi0 = 1;
else
enable_spi1 = 1;
/* enable chip-select pin */
/*将片选引脚设置为输出*/
at91_set_gpio_output(cs_pin, 1);
/* pass chip-select pin to driver */
devices[i].controller_data = (void *) cs_pin;
}
/*到此,循环执行完毕,向内核注册这些板级信息体*/
spi_register_board_info(devices, nr_devices);
/* Configure SPI bus(es) */
/*如果发现spi0上有设备注册,则打开spi0*/
if (enable_spi0) {
at91_set_A_periph(AT91_PIN_PA0, 0); /* SPI0_MISO */
at91_set_A_periph(AT91_PIN_PA1, 0); /* SPI0_MOSI */
at91_set_A_periph(AT91_PIN_PA2, 0); /* SPI1_SPCK */
at91_clock_associate("spi0_clk", &at91sam9260_spi0_device.dev, "spi_clk");
platform_device_register(&at91sam9260_spi0_device);
}
/*spi0设备也是如此*/
if (enable_spi1) {
at91_set_A_periph(AT91_PIN_PB0, 0); /* SPI1_MISO */
at91_set_A_periph(AT91_PIN_PB1, 0); /* SPI1_MOSI */
at91_set_A_periph(AT91_PIN_PB2, 0); /* SPI1_SPCK */
at91_clock_associate("spi1_clk", &at91sam9260_spi1_device.dev, "spi_clk");
platform_device_register(&at91sam9260_spi1_device);
}
}
从上面这个函数我们可以看出,这个函数就完成了两个功能:
1、向内核完成spi板级信息结构体的注册
2、注册了两个platform_device:spi0与spi1,这两个设备是spi总线控制器!
那么我们客户端spi_device设备的注册是如何完成的?不知道,呵呵
我今天仔细的看代码才发现玄机所在。
内核的注释很清晰的告诉我们,我们的spi设备是不允许热插拔!!这是由于spi设备驱动的框架不允许,我们的spi_device设备注册不是在板级初始化的时候完成的,而是在spi控制器的驱动加载的时候,也就是platform_driver:atmel_spi驱动加载的时候完成。
(atmel9260驱动文件atmel_spi.c)
driver/spi/atmel_spi.c文件中:
static int __init atmel_spi_probe(struct platform_device *pdev)
{
struct resource *regs;
int irq;
struct clk *clk;
int ret;
struct spi_master *master;
struct atmel_spi *as;
regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
if (!regs)
return -ENXIO;
irq = platform_get_irq(pdev, 0);
if (irq < 0)
return irq;
clk = clk_get(&pdev->dev, "spi_clk");
if (IS_ERR(clk))
return PTR_ERR(clk);
/* setup spi core then atmel-specific driver state */
ret = -ENOMEM;
master = spi_alloc_master(&pdev->dev, sizeof *as);
if (!master)
goto out_free;
master->bus_num = pdev->id;
master->num_chipselect = 4;
master->setup = atmel_spi_setup;
master->transfer = atmel_spi_transfer;
master->cleanup = atmel_spi_cleanup;
platform_set_drvdata(pdev, master);
as = spi_master_get_devdata(master);
as->buffer = dma_alloc_coherent(&pdev->dev, BUFFER_SIZE,
&as->buffer_dma, GFP_KERNEL);
if (!as->buffer)
goto out_free;
spin_lock_init(&as->lock);
INIT_LIST_HEAD(&as->queue);
as->pdev = pdev;
as->regs = ioremap(regs->start, (regs->end - regs->start) + 1);
if (!as->regs)
goto out_free_buffer;
as->irq = irq;
as->clk = clk;
#ifdef CONFIG_ARCH_AT91
if (!cpu_is_at91rm9200())
as->new_1 = 1;
#endif
ret = request_irq(irq, atmel_spi_interrupt, 0,
pdev->dev.bus_id, master);
if (ret)
goto out_unmap_regs;
/* Initialize the hardware */
clk_enable(clk);
spi_writel(as, CR, SPI_BIT(SWRST));
spi_writel(as, MR, SPI_BIT(MSTR) | SPI_BIT(MODFDIS));
spi_writel(as, PTCR, SPI_BIT(RXTDIS) | SPI_BIT(TXTDIS));
spi_writel(as, CR, SPI_BIT(SPIEN));
/* go! */
dev_info(&pdev->dev, "Atmel SPI Controller at 0x%08lx (irq %d)\n",
(unsigned long)regs->start, irq);
/*spi注册这个主控制器*/
ret = spi_register_master(master);
if (ret)
goto out_reset_hw;
return 0;
out_reset_hw:
spi_writel(as, CR, SPI_BIT(SWRST));
clk_disable(clk);
free_irq(irq, master);
out_unmap_regs:
iounmap(as->regs);
out_free_buffer:
dma_free_coherent(&pdev->dev, BUFFER_SIZE, as->buffer,
as->buffer_dma);
out_free:
clk_put(clk);
spi_master_put(master);
return ret;
}
而这个spi_register_master位于driver/spi/spi.c文件中,该函数调用了scan_boardinfo(master),扫描该spi master下面设备。该函数就存在于该文件下:该函数调用了spi_new_device(master, chip),这个chip就是一个spi_board_info结构体,这就是at91_add_device_spi第一个作用的用处:向内核的链表注册spi_board_info结构体的用处所在。我们来看函数的调用过程:
atmel_spi_probe----->spi_register_master----->scan_boardinfo
---->spi_new_device
我们来看这个spi_new_device函数:
struct spi_device *spi_new_device(struct spi_master *master,
struct spi_board_info *chip)
{
struct spi_device *proxy;
struct device *dev = master->cdev.dev;
int status;
/* NOTE: caller did any chip->bus_num checks necessary */
if (!spi_master_get(master))
return NULL;
/*靠,终于找到你了,先暴打一顿,舒服了。。这里就分配了我们重要的spi_device结构体*/
proxy = kzalloc(sizeof *proxy, GFP_KERNEL);
if (!proxy) {
dev_err(dev, "can't alloc dev for cs%d\n",
chip->chip_select);
goto fail;
}
/*这就是将我们的信息体中的数据转化为spi_device识别的数据*/
proxy->master = master;
proxy->chip_select = chip->chip_select;
proxy->max_speed_hz = chip->max_speed_hz;
proxy->mode = chip->mode;
proxy->irq = chip->irq;
proxy->modalias = chip->modalias;
snprintf(proxy->dev.bus_id, sizeof proxy->dev.bus_id,
"%s.%u", master->cdev.class_id,
chip->chip_select);
proxy->dev.parent = dev;
proxy->dev.bus = &spi_bus_type;
/*这里很重要,如果你的spi设备是dataflash的话,保存的就是你的分区表!!!所以我们要返回去修改我们的spi_boardinfo结构体*/
proxy->dev.platform_data = (void *) chip->platform_data;
/*片选信号*/
proxy->controller_data = chip->controller_data;
proxy->controller_state = NULL;
proxy->dev.release = spidev_release;
/* drivers may modify this default i/o setup */
status = master->setup(proxy);
if (status < 0) {
dev_dbg(dev, "can't %s %s, status %d\n",
"setup", proxy->dev.bus_id, status);
goto fail;
}
/* driver core catches callers that misbehave by defining
* devices that already exist.
*/
/*看到这句话,大家放心了吧,大家也就知道怎么找到spi_driver驱动的。。。*/
status = device_register(&proxy->dev);
if (status < 0) {
dev_dbg(dev, "can't %s %s, status %d\n",
"add", proxy->dev.bus_id, status);
goto fail;
}
dev_dbg(dev, "registered child %s\n", proxy->dev.bus_id);
return proxy;
fail:
spi_master_put(master);
kfree(proxy);
return NULL;
}
下面我们要解决最后的一个问题,dataflash的分区的问题,看了这么多,大家应该知道怎么解决了吧!
我们看mtd_dataflash.c文件中驱动加载函数调用了下面这个函数来添加flash设备。。
static int __devinit
add_dataflash(struct spi_device *spi, char *name,
int nr_pages, int pagesize, int pageoffset)
{
struct dataflash *priv;
struct mtd_info *device;
/*这里就告诉我们要在spi_boardinfo结构体的platform_data成员指向一个我们需要的flash_platform_data数据!*/
struct flash_platform_data *pdata = spi->dev.platform_data;
priv = kzalloc(sizeof *priv, GFP_KERNEL);
if (!priv)
return -ENOMEM;
init_MUTEX(&priv->lock);
priv->spi = spi;
priv->page_size = pagesize;
priv->page_offset = pageoffset;
/* name must be usable with cmdlinepart */
sprintf(priv->name, "spi%d.%d-%s",
spi->master->bus_num, spi->chip_select,
name);
device = &priv->mtd;
device->name = (pdata && pdata->name) ? pdata->name : priv->name;
device->size = nr_pages * pagesize;
device->erasesize = pagesize;
device->writesize = pagesize;
device->owner = THIS_MODULE;
device->type = MTD_DATAFLASH;
device->flags = MTD_WRITEABLE;
device->erase = dataflash_erase;
device->read = dataflash_read;
device->write = dataflash_write;
device->priv = priv;
dev_info(&spi->dev, "%s (%d KBytes)\n", name, device->size/1024);
dev_set_drvdata(&spi->dev, priv);
if (mtd_has_partitions()) {
struct mtd_partition *parts;
int nr_parts = 0;
/*我们这里没有定义该宏,所以不会在命令行传递分区表*/
#ifdef CONFIG_MTD_CMDLINE_PARTS
static const char *part_probes[] = { "cmdlinepart", NULL, };
nr_parts = parse_mtd_partitions(device, part_probes, &parts, 0);
#endif
if (nr_parts <= 0 && pdata && pdata->parts) {
parts = pdata->parts;
nr_parts = pdata->nr_parts;
}
if (nr_parts > 0) {
priv->partitioned = 1;
return add_mtd_partitions(device, parts, nr_parts);
}
} else if (pdata && pdata->nr_parts)
dev_warn(&spi->dev, "ignoring %d default partitions on %s\n",
pdata->nr_parts, device->name);
return add_mtd_device(device) == 1 ? -ENODEV : 0;
}
所以我们需要修改这个文件:
arch/arm/mach-at91/board-sam9260ek.c文件:
添加如下:
#if !defined(CONFIG_MMC_AT91)
#define SIZE_1PAGE 528
#define SIZE_1M (unsigned long)(1024*1024)
static struct mtd_partition ek_dataflash_partition[] = {
{
.name = "U-boot ENV",
.offset = 0,
.size = 64*SIZE_1PAGE,
},
{
.name = "U-BOOT",
.offset = 64*SIZE_1PAGE,
.size = 400*SIZE_1PAGE,
},
{
.name ="Kernel",
.offset=464*SIZE_1PAGE,
.size = 4000*SIZE_1PAGE,
},
{
.name ="Root fs",
.offset=4464*SIZE_1PAGE,
.size = (8192-4464)*SIZE_1PAGE,
},
};
struct flash_platform_data dataflash_atmel={
.name="AT45DB321",
.parts=ek_dataflash_partition,
.nr_parts=ARRAY_SIZE(ek_dataflash_partition),
};
#endif
修改spi_boardinfo结构体:
static struct spi_board_info ek_spi_devices[] = {
#if !defined(CONFIG_MMC_AT91)
{ /* DataFlash chip */
.modalias = "mtd_dataflash",
.chip_select = 1,
.max_speed_hz = 15 * 1000 * 1000,
.bus_num = 0,
.platform_data=&dataflash_atmel,
},
添加platform_data结构成员。
这里我们建立mtd_partition结构体要注意,由于dataflash是以528字节每页的,其实,at45db321x芯片可以设置为512字节每页,这个操作是不可以逆转的,那个位是一个otp位,用过的人就应该知道,但是出厂的时候默认的528字节每页。
如果我们不是以528个字节为单位的话,内核将出警告,强制将分区加载为readonly格式。
到此,分区加载成功,dmesg输出如下信息:
<6>mtd_dataflash spi0.1: AT45DB321x (4224 KBytes)
<5>Creating 4 MTD partitions on "AT45DB321":
<5>0x00000000-0x00008400 : "U-boot ENV"
<5>0x00008400-0x0003bd00 : "U-BOOT"
<5>0x0003bd00-0x0023f700 : "Kernel"
<5>0x0023f700-0x00420000 : "Root fs"
linux简直太伟大了,使用得越多,就越能体会到其思想的伟大
文章出处:http://www.diybl.com/course/6_system/linux/Linuxjs/200868/123621_3.html