對linux的devfs類型的驅動程序的編寫可以從以下幾大內容理解和入手:
通過分析驅動程序源代碼可以發(fā)現(xiàn)驅動程序一般可分三部分:
核心數(shù)據(jù)結構;核心數(shù)據(jù)和資源的初始化,注冊以及注消,釋放;底層設備操作函數(shù);

A.核心數(shù)據(jù)結構
struct file_operations fops 設備驅動程序接口
struct file_operations {
struct module *owner;
loff_t (*llseek) (struct file *, loff_t, int);
ssize_t (*read) (struct file *, char *, size_t, loff_t *);
ssize_t (*write) (struct file *, const char *, size_t, loff_t *);
int (*readdir) (struct file *, void *, filldir_t);
unsigned int (*poll) (struct file *, struct poll_table_struct *);
int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
int (*mmap) (struct file *, struct vm_area_struct *);
int (*open) (struct inode *, struct file *);
int (*flush) (struct file *);
int (*release) (struct inode *, struct file *);
int (*fsync) (struct file *, struct dentry *, int datasync);
int (*fasync) (int, struct file *, int);
int (*lock) (struct file *, int, struct file_lock *);
ssize_t (*readv) (struct file *, const struct iovec *, unsigned long, loff_t *);
ssize_t (*writev) (struct file *, const struct iovec *, unsigned long, loff_t *);
ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
};


block_device_operations 塊設備驅動程序接口
{ int (*open) (struct inode *, struct file *);
int (*release) (struct inode *, struct file *);
int (*ioctl) (struct inode *, struct file *, unsigned, unsigned long);
int (*check_media_change) (kdev_t);
int (*revalidate) (kdev_t);
struct module *owner;
};塊設備的READ().WRITE()不在這里注冊,而是在設備的讀寫請求隊列里注冊,內核在這里將調用通用的blk_read(),blk_write().向讀寫隊列
發(fā)出讀寫請求.

Linux 利用這些數(shù)據(jù)結構向內核注冊open(),release(),ioctl(),check_media_change(),rvalidate()等函數(shù)的入口句柄.
我們將要編寫的open(),release(),ioctl(),check_media_change(),revalidate()等函數(shù),將在驅動初始化的時候,
通過一個此結構類型的變量向內核提供函數(shù)的 入口.

struct request_queue_t 設備請求隊列的數(shù)據(jù)結構
struct request_list {
unsigned int count;
unsigned int pending[2];
struct list_head free;
};
struct request {
struct list_head queue;
int elevator_sequence;
kdev_t rq_dev;
int cmd; /* READ or WRITE */
int errors;
unsigned long start_time;
unsigned long sector;
unsigned long nr_sectors;
unsigned long hard_sector, hard_nr_sectors;
unsigned int nr_segments;
unsigned int nr_hw_segments;
unsigned long current_nr_sectors, hard_cur_sectors;
void * special;
char * buffer;
struct completion * waiting;
struct buffer_head * bh;
struct buffer_head * bhtail;
request_queue_t *q;
};

struct request_queue
{
/*
* the queue request freelist, one for reads and one for writes
*/
struct request_list rq;

/*
* The total number of requests on each queue
*/
int nr_requests;

/*
* Batching threshold for sleep/wakeup decisions
*/
int batch_requests;

/*
* The total number of 512byte blocks on each queue
*/
atomic_t nr_sectors;

/*
* Batching threshold for sleep/wakeup decisions
*/
int batch_sectors;

/*
* The max number of 512byte blocks on each queue
*/
int max_queue_sectors;

/*
* Together with queue_head for cacheline sharing
*/
struct list_head queue_head;
elevator_t elevator;

request_fn_proc * request_fn;
merge_request_fn * back_merge_fn;
merge_request_fn * front_merge_fn;
merge_requests_fn * merge_requests_fn;
make_request_fn * make_request_fn;
plug_device_fn * plug_device_fn;
/*
* The queue owner gets to use this for whatever they like.
* ll_rw_blk doesn't touch it.
*/
void * queuedata;

/*
* This is used to remove the plug when tq_disk runs.
*/
struct tq_struct plug_tq;

/*
* Boolean that indicates whether this queue is plugged or not.
*/
int plugged:1;

/*
* Boolean that indicates whether current_request is active or
* not.
*/
int head_active:1;

/*
* Boolean that indicates you will use blk_started_sectors
* and blk_finished_sectors in addition to blk_started_io
* and blk_finished_io. It enables the throttling code to
* help keep the sectors in flight to a reasonable value
*/
int can_throttle:1;

unsigned long bounce_pfn;

/*
* Is meant to protect the queue in the future instead of
* io_request_lock
*/
spinlock_t queue_lock;

/*
* Tasks wait here for free read and write requests
*/
wait_queue_head_t wait_for_requests;

struct request *last_request;
};

緩沖區(qū)和對緩沖區(qū)相應的I/O操作在此任務隊列中相關聯(lián),等待內核的調度.如果是字符設備就不需要此數(shù)據(jù)結構.而
塊設備的read(),write()函數(shù)則在buffer_queue的initize和設備請求隊列進行處理請求時候傳遞給request_fn().
struct request_queue_t{}設備請求隊列的變量類型,驅動程序在初始化的時候需要填寫request_fn().

其他的數(shù)據(jù)結構還有 I/O port,Irq,DMA 資源分配,符合POSIX標準的ioctl的cmd的構造和定義,以及描述設備自身的
相關數(shù)據(jù)結構定義-如設備的控制寄存器的相關數(shù)據(jù)結構定義,BIOS里的參數(shù)定義,設備類型定義等.

B.初始化和注冊和注消,模塊方式驅動程序的加載和卸載.
設備驅動程序在定義了數(shù)據(jù)結構后 ,首先開始初始化:
如I/O 端口的檢查和登記,內核對 I/O PORT的檢查和登記提供了兩個 函數(shù)check_region(int io_port, int off_set)
和request_region(int io_port, int off_set,char *devname).I/O Port登記后,就可以用inb()和outb()進行操作了 .

還有DMA和Irq的初始化檢查和 登記,
int request_irq(unsigned int irq ,void(*handle)(int,void *,struct pt_regs *),unsigned int long flags,

const char *device);

irq: 是要申請的中斷。

handle：中斷處理函數(shù)指針。

flags：SA_INTERRUPT 請求一個快速中斷，0 正常中斷。

device：設備名。

如果登記成功，返回0，這時在/proc/interrupts文件中可以看你請求的中斷。

DMA主要是在內存中分配交換內存空間.還有緩沖區(qū),設備請求隊列的初始化.

還有設備控制寄存器的檢查和初始化,還有對設備自身相關的數(shù)據(jù)結構的初始化,填寫一些設備特定的數(shù)據(jù)等.

然后,開始注冊
devfs_register()向VFS注冊統(tǒng)一的設備操作函數(shù).
static struct file_operations XXX_fops = {
owner: THIS_MODULE, XXX_fops所屬的設備模塊
read: XXX_read, 讀設備操作
write: XXX_write, 寫設備操作
ioctl: XXX_ioctl, 控制設備操作
mmap: XXX_mmap, 內存重映射操作
open: XXX_open, 打開設備操作
release: XXX_release 釋放設備操作
/* ... */
};


blk_init_queue()隊列初始化函數(shù).
request_irq()中斷注冊函數(shù)
相應的注消函數(shù):
devfs_unregister (devfs_handle_t de){};
free_irq()釋放中斷,I/O資源,釋放緩沖區(qū),釋放設備,請求隊列,VFS節(jié)點等.

模塊方式驅動程序的加載和卸載.

static int __init _init_module (void)
{
/* ... */
}

static void __exit _cleanup_module (void)
{

}

/* 加載驅動程序模塊入口 */
module_init(_init_module);

/* 卸載驅動程序模塊入口 */
module_exit(_cleanup_module);


_intrrupt()
設備發(fā)生中斷時的處理程序.
{
1.對共享中斷的處理;
2.對spinlock以及其他的事務的處理;
}

C. 底層設備操作函數(shù)的編寫
read().write(),open(),release(),check_media_change(),revalidate()等.


open()和release()

打開設備是通過調用file_operations結構中的函數(shù)open( )來完成的，它是驅動程序用來為今后的操作完成初始化準備工作的。在大部分驅動程序中，open( )通常需要完成下列工作：

1. 檢查設備相關錯誤，如設備尚未準備好等。
2. 如果是次打開，則初始化硬件設備。
3. 識別次設備號，如果有必要則更新讀寫操作的當前位置指針f_ops。
4. 分配和填寫要放在file->private_data里的數(shù)據(jù)結構。
5. 使用計數(shù)增1。

釋放設備是通過調用file_operations結構中的函數(shù)release( )來完成的，這個設備方法有時也被稱為close( )，它的作用正好與open( )相反，通常要完成下列工作：

1. 使用計數(shù)減1。
2. 釋放在file->private_data中分配的內存。
3. 如果使用計算為0，則關閉設備。

read()和 write()

字符設備的讀寫操作相對比較簡單，直接使用函數(shù)read( )和write( )就可以了。但如果是塊設備的話，則需要調用函數(shù)block_read( )和block_write( )來進行數(shù)據(jù)讀寫，這兩個函數(shù)將向設備請求表中增加讀寫請求，以便Linux內核可以對請求順序進行優(yōu)化。由于是對內存緩沖區(qū)而不是直接對設備進行操作的，因此能很大程度上加快讀寫速度。如果內存緩沖區(qū)中沒有所要讀入的數(shù)據(jù)，或者需要執(zhí)行寫操作將數(shù)據(jù)寫入設備，那么就要執(zhí)行真正的數(shù)據(jù)傳輸，這是通過調用數(shù)據(jù)結構blk_dev_struct中的函數(shù)request_fn( )來完成的。

ioctl()--將cmd進行解釋,并送到設備的控制寄存器.事實上,read()和write()也要通過ioctl()來完成操作的 .
ioctl(){
CASE CMD{
SWITCH CASE1:{...};
SWITCH CASE2:{...};
SWITCH CASE N:{...};
.
.
DEFAULT : {...};
}
END CASE

總結:
我們可以看出一個linux的驅動程序通常包含如下:
初始化設備模塊、
{I/O port ,DMA.Irq,內存 buffer,初始化并且填寫具體設備數(shù)據(jù)結構,注冊 fops的具體函數(shù)等等 }
中斷處理模塊、設備釋放模塊、設備卸載模塊
設備打開模塊、數(shù)據(jù)讀寫和控制模塊、
驅動裝載模塊、驅動釋放模塊.