硬件环境： 飞凌OK6410，256MB DDR，2GB NAND Flash、 NAND Flash 型号：K9G8G08U9A 、 分析源码：Linux 2.6.36.2 内核源码。
一、 MTD 设备驱动。
1、先来简单介绍一下MTD
在Linux 系统中， 提供了MTD(Memory Technology Device , 内存技术设备)系统来建立 Flash 针对 Linux 的系统、抽象的接口， MTD 将文件系统 与 底层的Flash
存储器进行了隔离， 使 Flash 驱动工程师 无需关心Flash 作为字符设备和 块 设备与 LInux内核的接口。
2、在引入MTD 后Linux 系统中的Flash 设备及接口可分为4层, 从上到下依次是：设备节点、MTD 设备层、MTD原始设备层 和 硬件驱动层。 这 4 层的作用定义如下：
1-> 硬件驱动层： Flash 硬件驱动层负责 Flash 硬件设备的读、写、擦除， LInux MTD 设备的 NOR Flash 芯片驱动位于 drivers/mtd/chips 子目录下， NAND Flash
的驱动程序则 位于 drivers/mtd/nand 子目录下。
2->MTD 原始设备层： MTD原始设备层由两部分组成， 一部分是MTD 原始设备的通用代码， 另一部分是各个特定 Flash 的数据，例如分区。
3->MTD设备层： 基于MTD 原始设备，Linux 系统可以定义出 MTD 的块设备的结构(主设备号 31) 和 字符设备 (设备号 90) ，构成MTD 设备层， MTD 字符设备定义
在mtdchar.c 中实现，MTD 块设备则是定义在一个描述MTD 块设备的结构 mtdblk_dev ，并声明了一个名为 mtdblks 的指针数组，这个数组 中的每个mtdblk_dev
和 mtd_table 中的每一个mtd_info 一一对应。
4->设备节点： 通过mknod 在/dev 子目录下建立MTD字符设备节点 和 块设备节点，用户通过访问此此设备节点即可访问 MTD 字符设备和块设备。
3、分析Linux MTD 系统接口 mtd_info 结构体代码分析 此结构体定义在 ./include/linux/mtd/mtd.h 中
关键词词解析：
XIP ：XIP eXecute In Place，即芯片内执行，指应用程序可以直接在flash闪存内运行，不必再把代码读到系统RAM中。flash内执行
是指nor flash 不需要初始化，可以直接在flash内执行代码。但往往只执行部分代码，比如初始化RAM.
OOB ：Out Of Brower 传输层协议使用带外数据(out-of-band,OOB)来发送一些重要的数据,如果通信一方有重要的数据需要通知对方时,协议能够将这些数据
快速地发送到对方.为了发送这些数据
iovec-base : iovec 结构体基础。struct iovec定义了一个向量元素。通常，这个结构用作一个多元素的数组。对于每一个传输的元素，指针成员iov_base指向
一个缓冲区，这个缓冲区是存放的是readv所接收的数据或是writev将要发送的数据。成员iov_len在各种情况下分别确定了接收的最大长度以及实际写入的长度。
Sync : 函数， 函数说明：此函数负责将系统缓冲区的内容写回磁盘，以确保数据同步。
structmtd_info {
u_char type;// 内存技术的类型
uint32_t flags;// 标志位
uint64_t size;// Total size of the MTD 、mtd 设备的大小
/* "Major" erase size for the device. Na茂ve users may take this
* to be the only erase size available, or may use the more detailed
* information below if they desire
*/
uint32_t erasesize;// 主要的擦除块大小 erase size of main block
/* Minimal writable flash unit size. In case of NOR flash it is 1 (even
* though individual bits can be cleared), in case of NAND flash it is
* one NAND page (or half, or one-fourths of it), in case of ECC-ed NOR
* it is of ECC block size, etc. It is illegal to have writesize = 0.
* Any driver registering a struct mtd_info must ensure a writesize of
* 1 or larger.
*/
uint32_t writesize;// 最小的可写单元的字节数
uint32_t oobsize;// Amount of OOB data per block (e.g. 16) OOB 字节数
uint32_t oobavail;// Available OOB bytes per block 可用OBB 字节数
/*
* If erasesize is a power of 2 then the shift is stored in
* erasesize_shift otherwise erasesize_shift is zero. Ditto writesize.
*/
unsignedinterasesize_shift;
unsignedintwritesize_shift;
/* Masks based on erasesize_shift and writesize_shift */
unsignedinterasesize_mask;
unsignedintwritesize_mask;
// Kernel-only stuff starts here.
constchar*name;
intindex;
/* ecc layout structure pointer - read only ! */
structnand_ecclayout *ecclayout;// ECC 布局结构体指针
/* Data for variable erase regions. If numeraseregions is zero,
* it means that the whole device has erasesize as given above.
*/
intnumeraseregions;// 不同的erasesize 的区域 数目通常是1
structmtd_erase_region_info *eraseregions;
/*
* Erase is an asynchronous operation. Device drivers are supposed
* to call instr->callback() whenever the operation completes, even
* if it completes with a failure.
* Callers are supposed to pass a callback function and wait for it
* to be called before writing to the block.
*/
int(*erase) (structmtd_info *mtd,structerase_info *instr);
/* This stuff for eXecute-In-Place */
/* phys is optional and may be set to NULL */
int(*point) (structmtd_info *mtd, loff_t from,size_tlen,// 针对 eXecute-In- Place
size_t*retlen,void**virt, resource_size_t *phys);
/* We probably shouldn't allow XIP if the unpoint isn't a NULL */
void(*unpoint) (structmtd_info *mtd, loff_t from,size_tlen);// 如果unpoint 为空，不允许 XIP
/* Allow NOMMU mmap() to directly map the device (if not NULL)
* - return the address to which the offset maps
* - return -ENOSYS to indicate refusal to do the mapping
*/
unsignedlong(*get_unmapped_area) (structmtd_info *mtd,
unsignedlonglen,
unsignedlongoffset,
unsignedlongflags);
/* Backing device capabilities for this device
* - provides mmap capabilities
*/
structbacking_dev_info *backing_dev_info;
int(*read) (structmtd_info *mtd, loff_t from,size_tlen,size_t*retlen, u_char *buf);// 读 flash
int(*write) (structmtd_info *mtd, loff_t to,size_tlen,size_t*retlen,constu_char *buf);// 写 flash
/* In blackbox flight recorder like scenarios we want to make successful
writes in interrupt context. panic_write() is only intended to be
called when its known the kernel is about to panic and we need the
write to succeed. Since the kernel is not going to be running for much
longer, this function can break locks and delay to ensure the write
succeeds (but not sleep). */
int(*panic_write) (structmtd_info *mtd, loff_t to,size_tlen,size_t*retlen,constu_char *buf);// Kernel panic 时序读写
int(*read_oob) (structmtd_info *mtd, loff_t from,// 读 out-of-band
structmtd_oob_ops *ops);
int(*write_oob) (structmtd_info *mtd, loff_t to,// 写 out-of-band
structmtd_oob_ops *ops);
/*
* Methods to access the protection register area, present in some
* flash devices. The user data is one time programmable but the
* factory data is read only.
*/
int(*get_fact_prot_info) (structmtd_info *mtd,structotp_info *buf,size_tlen);
int(*read_fact_prot_reg) (structmtd_info *mtd, loff_t from,size_tlen,size_t*retlen, u_char *buf);
int(*get_user_prot_info) (structmtd_info *mtd,structotp_info *buf,size_tlen);
int(*read_user_prot_reg) (structmtd_info *mtd, loff_t from,size_tlen,size_t*retlen, u_char *buf);
int(*write_user_prot_reg) (structmtd_info *mtd, loff_t from,size_tlen,size_t*retlen, u_char *buf);
int(*lock_user_prot_reg) (structmtd_info *mtd, loff_t from,size_tlen);
/* kvec-based read/write methods.
NB: The 'count' parameter is the number of _vectors_, each of
which contains an (ofs, len) tuple.
*/
int(*writev) (structmtd_info *mtd,conststructkvec *vecs, unsignedlongcount, loff_t to,size_t*retlen);// iovec-based 读写函数
/* Sync */
void(*sync) (structmtd_info *mtd);// Sync
/* Chip-supported device locking */
int(*lock) (structmtd_info *mtd, loff_t ofs, uint64_t len);// 设备锁
int(*unlock) (structmtd_info *mtd, loff_t ofs, uint64_t len);
int(*is_locked) (structmtd_info *mtd, loff_t ofs, uint64_t len);
/* Power Management functions */
int(*suspend) (structmtd_info *mtd);// 电源管理函数
void(*resume) (structmtd_info *mtd);
/* Bad block management functions */
int(*block_isbad) (structmtd_info *mtd, loff_t ofs);// 坏块管理函数
int(*block_markbad) (structmtd_info *mtd, loff_t ofs);
structnotifier_block reboot_notifier;/* default mode before reboot */
/* ECC status information */
structmtd_ecc_stats ecc_stats;
/* Subpage shift (NAND) */
intsubpage_sft;
void*priv;// 私有函数
structmodule *owner;
structdevice dev;
intusecount;
/* If the driver is something smart, like UBI, it may need to maintain
* its own reference counting. The below functions are only for driver.
* The driver may register its callbacks. These callbacks are not
* supposed to be called by MTD users */
int(*get_device) (structmtd_info *mtd);
void(*put_device) (structmtd_info *mtd);
};
mtd_info 中的 read(). write(). read_oob(). write_oob(). erase() 是 MTD 设备驱动主要实现的函数。在在后面我将要介绍的nand flahs 驱动中几乎看不到mtd_info
的成员函数(也就是说这些函数对于Flash 芯片来说是透明的)，这是因为在Linux MTD 下层实现了针对 NOR、NAND Flsh 的同游mtd_info 成员函数。
Flash 驱动中使用如下两个函数来注册 和注销MTD 设备：
int add_mtd_device(struct mtd_info *mtd);
int del_mtd_device (struct mtd_info *mtd)