原文地址:
http://opendevkit.com/?e=37
2. ubifs制作
(1) mkfs.ubifs
mkfs.ubifs -r system -m 2048 -e 126976 -c 1057 -x zlib -o system.ubifs
-r说的目录
-m说的是nand flash的页大小，一般都是2048或者4096，当然也有些是512，硬件特性决定好了，或者说驱动规定好了，有些平台4096页当做2048用也是用的
-e说的是逻辑擦除块大小，大家知道nand flash页读页写块擦，一个设备多个块，一个块多个页，一般也都是一个块是64个页，这样算一下无论擦除块大小就是2048*64=131072，-e的算法是物理擦除块大小-2*页大小，这里就是131072-2*2048=126976
-c说的是最大逻辑块编号，这个很重要，不能大也不能小，太小也要大于image大小，太大mount有问题，计算起点是分区的物理块数量，比如128MiB的mtd分区，物理块数量是128MiB/2048/64 = 1024个，需要减去2个坏块保留块，减去1个wear-leveling块，还要减去1个eba的块，等等，比如最终的值是1022，注意，如果物理上这个分区有坏块的话，kernel会扫描到的，这时候，我们计算的这个值就要减去坏块数了，否则会有逻辑块大于物理块数的内核问题mount失败，确切知道坏块数是比较困难的，一般做法是做一个坏块容忍数，比如20个，这样我们再减去20个坏块，不要担心这个会浪费空间，ubinize的autoresize选项就是解决这个问题的。具体的这个值需要计算。！！！！
-o说的image名字
-x说的是压缩方法，默认是lzo，还支持zlib，zlib压缩率高些，但是lzo压缩解压速度快
(2) ubinize
ubinize -o system.ubi -m 2048 -p 131072 ubinize.cfg
-o说的是输出image
-m还是页大小
-p是物理擦除块大小
ubinize.cfg是volume配置文件，例子如下：
- [ubifs]
- mode=ubi
- image=system.ubifs #说的是mkfs.ubifs的结果
- vol_id=0
- vol_size=100MiB #说的是volume大小，用-e和-c的值做乘法计算，一般不用写，autoresize会自动根据mtd分区大小适应，默认值是image大小，写了这个作用是帮助检查image是否超过了分区限制，制作时候就提示，否则mount会出错。-c的值是经过计算的最大值了，不过autoresize参数会自适应大小，不会浪费空间的。
- vol_type=dynamic
- vol_alignment=1
- vol_name=system #说的是分区名字
- vol_flags=autoresize
(4) uboot支持
-0- 需要打开的配置 - 需要烧写mkfs.ubifs结果时候需要打开，ubinize处理过的不需要
CONFIG_CMD_UBI
CONFIG_CMD_UBIFS
CONFIG_LZO
CONFIG_RBTREE
CONFIG_ZLIB
CONFIG_GZIP
-1- ubifs烧写
mkfs.ubifs工具制作的结果，就是ubifs镜像，不包含volume信息，需要用
nand erase.part system
ubi part system - 激活分区
ubi create system - 创建分区
ubi write 84000000 system xxxxx - xxxxx表示镜像实际大小
-2- ubi volume bin烧写
mkfs.ubifs后，使用ubinize处理了ubifs的镜像后，镜像含有了volume信息，直接
nand write 84000000 system xxxxx - xxxxx表示镜像实际大小
(5) kernel支持
Device Drivers --->
Memory Technology Device(MTD) support --->
UBI - Unsorted block images --->
<*> Enable UBI - Unsorted block images
File systems --->
Miscellaneous filesystems --->
<*> UBIFS file system support
[*] Advanced compression options
[*] LZO compression support
[*] ZLIB compression support
(6) android支持
-1- andriod/system/core/rootdir/init.rc
mount yaffs2 mtd@userdata /data nosuid nodev
改为：
mount ubifs ubi@userdata /data nosuid nodev
-2- 增加对ubifs的mount支持
andriod/system/core/init/builtins.c
在 } else if (!strncmp(source, "loop@", 5)) {
之前加上
+ }else if (!strncmp(source, "ubi@", 4)) {
+ n = ubi_attach_mtd(source + 4);
+ if (n < 0) {
+ return -1;
+ }
+ sprintf(tmp, "/dev/ubi%d_0", n);
+ if (wait)
+ wait_for_file(tmp, COMMAND_RETRY_TIMEOUT);
+ if (mount(tmp, target, system, flags, options) < 0) {
+ ubi_detach_dev(n);
+ return -1;
+ }
+ return 0;
+ }else if (!strncmp(source, "loop@", 5)) {
3. andriod/system/core/init/init.c
在static int property_triggers_enabled = 0;之后加上
+static unsigned ubifs_enabled = 1;
static int set_init_properties_action(int nargs, char **args)
{
property_set("ro.revision", tmp);
+ property_set("ro.ubifs",ubifs_enabled ? "1" : "0");
return 0;
}
int main(int argc, char **argv)
{
action_for_each_trigger("post-fs", action_add_queue_tail);
+ action_for_each_trigger("ubi-fs", action_add_queue_tail);
}
4. andriod/system/core/init/util.h
void get_hardware_name(char *hardware, unsigned int *revision);
+int ubi_attach_mtd(const char *name);
+int ubi_detach_dev(int dev);
5. andriod/system/core/init/util.c
#include <sys/un.h>
+#include <sys/ioctl.h>
#include "util.h"
+#include "ubi-user.h"
+#define UBI_CTRL_DEV "/dev/ubi_ctrl"
+#define UBI_SYS_PATH "/sys/class/ubi"
在最后添加下面三个函数
static int ubi_dev_read_int(int dev, const char *file, int def)
{
int fd, val = def;
char path[128], buf[64];
sprintf(path, UBI_SYS_PATH "/ubi%d/%s", dev, file);
wait_for_file(path, 5);
fd = open(path, O_RDONLY);
if (fd == -1) {
return val;
}
if (read(fd, buf, 64) > 0) {
val = atoi(buf);
}
close(fd);
return val;
}
int ubi_attach_mtd(const char *name)
{
int ret;
int mtd_num, ubi_num;
int ubi_ctrl, ubi_dev;
int vols, avail_lebs, leb_size;
char path[128];
struct ubi_attach_req attach_req;
struct ubi_mkvol_req mkvol_req;
mtd_num = mtd_name_to_number(name);
if (mtd_num == -1) {
return -1;
}
ubi_ctrl = open(UBI_CTRL_DEV, O_RDONLY);
if (ubi_ctrl == -1) {
return -1;
}
memset(&attach_req, 0, sizeof(struct ubi_attach_req));
attach_req.ubi_num = UBI_DEV_NUM_AUTO;
attach_req.mtd_num = mtd_num;
ret = ioctl(ubi_ctrl, UBI_IOCATT, &attach_req);
if (ret == -1) {
close(ubi_ctrl);
return -1;
}
ubi_num = attach_req.ubi_num;
vols = ubi_dev_read_int(ubi_num, "volumes_count", -1);
if (vols == 0) {
sprintf(path, "/dev/ubi%d", ubi_num);
ubi_dev = open(path, O_RDONLY);
if (ubi_dev == -1) {
close(ubi_ctrl);
return ubi_num;
}
avail_lebs = ubi_dev_read_int(ubi_num, "avail_eraseblocks", 0);
leb_size = ubi_dev_read_int(ubi_num, "eraseblock_size", 0);
memset(&mkvol_req, 0, sizeof(struct ubi_mkvol_req));
mkvol_req.vol_id = UBI_VOL_NUM_AUTO;
mkvol_req.alignment = 1;
mkvol_req.bytes = (long long)avail_lebs * leb_size;
mkvol_req.vol_type = UBI_DYNAMIC_VOLUME;
ret = snprintf(mkvol_req.name, UBI_MAX_VOLUME_NAME + 1, "%s", name);
mkvol_req.name_len = ret;
ioctl(ubi_dev, UBI_IOCMKVOL, &mkvol_req);
close(ubi_dev);
}
close(ubi_ctrl);
return ubi_num;
}
int ubi_detach_dev(int dev)
{
int ret, ubi_ctrl;
ubi_ctrl = open(UBI_CTRL_DEV, O_RDONLY);
if (ubi_ctrl == -1) {
return -1;
}
ret = ioctl(ubi_ctrl, UBI_IOCDET, &dev);
close(ubi_ctrl);
return ret;
}
6.增加文件
andriod/system/core/init/ubi-user.h
/*
* Copyright (c) International Business Machines Corp., 2006
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See
* the GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* Author: Artem Bityutskiy (Битюцкий Артём)
*/
#ifndef __UBI_USER_H__
#define __UBI_USER_H__
/*
* UBI device creation (the same as MTD device attachment)
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*
* MTD devices may be attached using %UBI_IOCATT ioctl command of the UBI
* control device. The caller has to properly fill and pass
* &struct ubi_attach_req object - UBI will attach the MTD device specified in
* the request and return the newly created UBI device number as the ioctl
* return value.
*
* UBI device deletion (the same as MTD device detachment)
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*
* An UBI device maybe deleted with %UBI_IOCDET ioctl command of the UBI
* control device.
*
* UBI volume creation
* ~~~~~~~~~~~~~~~~~~~
*
* UBI volumes are created via the %UBI_IOCMKVOL IOCTL command of UBI character
* device. A &struct ubi_mkvol_req object has to be properly filled and a
* pointer to it has to be passed to the IOCTL.
*
* UBI volume deletion
* ~~~~~~~~~~~~~~~~~~~
*
* To delete a volume, the %UBI_IOCRMVOL IOCTL command of the UBI character
* device should be used. A pointer to the 32-bit volume ID hast to be passed
* to the IOCTL.
*
* UBI volume re-size
* ~~~~~~~~~~~~~~~~~~
*
* To re-size a volume, the %UBI_IOCRSVOL IOCTL command of the UBI character
* device should be used. A &struct ubi_rsvol_req object has to be properly
* filled and a pointer to it has to be passed to the IOCTL.
*
* UBI volume update
* ~~~~~~~~~~~~~~~~~
*
* Volume update should be done via the %UBI_IOCVOLUP IOCTL command of the
* corresponding UBI volume character device. A pointer to a 64-bit update
* size should be passed to the IOCTL. After this, UBI expects user to write
* this number of bytes to the volume character device. The update is finished
* when the claimed number of bytes is passed. So, the volume update sequence
* is something like:
*
* fd = open("/dev/my_volume");
* ioctl(fd, UBI_IOCVOLUP, &image_size);
* write(fd, buf, image_size);
* close(fd);
*
* Atomic eraseblock change
* ~~~~~~~~~~~~~~~~~~~~~~~~
*
* Atomic eraseblock change operation is done via the %UBI_IOCEBCH IOCTL
* command of the corresponding UBI volume character device. A pointer to
* &struct ubi_leb_change_req has to be passed to the IOCTL. Then the user is
* expected to write the requested amount of bytes. This is similar to the
* "volume update" IOCTL.
*/
/*
* When a new UBI volume or UBI device is created, users may either specify the
* volume/device number they want to create or to let UBI automatically assign
* the number using these constants.
*/
#define UBI_VOL_NUM_AUTO (-1)
#define UBI_DEV_NUM_AUTO (-1)
/* Maximum volume name length */
#define UBI_MAX_VOLUME_NAME 127
/* IOCTL commands of UBI character devices */
#define UBI_IOC_MAGIC 'o'
/* Create an UBI volume */
#define UBI_IOCMKVOL _IOW(UBI_IOC_MAGIC, 0, struct ubi_mkvol_req)
/* Remove an UBI volume */
#define UBI_IOCRMVOL _IOW(UBI_IOC_MAGIC, 1, int32_t)
/* Re-size an UBI volume */
#define UBI_IOCRSVOL _IOW(UBI_IOC_MAGIC, 2, struct ubi_rsvol_req)
/* IOCTL commands of the UBI control character device */
#define UBI_CTRL_IOC_MAGIC 'o'
/* Attach an MTD device */
#define UBI_IOCATT _IOW(UBI_CTRL_IOC_MAGIC, 64, struct ubi_attach_req)
/* Detach an MTD device */
#define UBI_IOCDET _IOW(UBI_CTRL_IOC_MAGIC, 65, int32_t)
/* IOCTL commands of UBI volume character devices */
#define UBI_VOL_IOC_MAGIC 'O'
/* Start UBI volume update */
#define UBI_IOCVOLUP _IOW(UBI_VOL_IOC_MAGIC, 0, int64_t)
/* An eraseblock erasure command, used for debugging, disabled by default */
#define UBI_IOCEBER _IOW(UBI_VOL_IOC_MAGIC, 1, int32_t)
/* An atomic eraseblock change command */
#define UBI_IOCEBCH _IOW(UBI_VOL_IOC_MAGIC, 2, int32_t)
/* Maximum MTD device name length supported by UBI */
#define MAX_UBI_MTD_NAME_LEN 127
/*
* UBI data type hint constants.
*
* UBI_LONGTERM: long-term data
* UBI_SHORTTERM: short-term data
* UBI_UNKNOWN: data persistence is unknown
*
* These constants are used when data is written to UBI volumes in order to
* help the UBI wear-leveling unit to find more appropriate physical
* eraseblocks.
*/
enum {
UBI_LONGTERM = 1,
UBI_SHORTTERM = 2,
UBI_UNKNOWN = 3,
};
/*
* UBI volume type constants.
*
* @UBI_DYNAMIC_VOLUME: dynamic volume
* @UBI_STATIC_VOLUME: static volume
*/
enum {
UBI_DYNAMIC_VOLUME = 3,
UBI_STATIC_VOLUME = 4,
};
/**
* struct ubi_attach_req - attach MTD device request.
* @ubi_num: UBI device number to create
* @mtd_num: MTD device number to attach
* @vid_hdr_offset: VID header offset (use defaults if %0)
* @padding: reserved for future, not used, has to be zeroed
*
* This data structure is used to specify MTD device UBI has to attach and the
* parameters it has to use. The number which should be assigned to the new UBI
* device is passed in @ubi_num. UBI may automatically assign the number if
* @UBI_DEV_NUM_AUTO is passed. In this case, the device number is returned in
* @ubi_num.
*
* Most applications should pass %0 in @vid_hdr_offset to make UBI use default
* offset of the VID header within physical eraseblocks. The default offset is
* the next min. I/O unit after the EC header. For example, it will be offset
* 512 in case of a 512 bytes page NAND flash with no sub-page support. Or
* it will be 512 in case of a 2KiB page NAND flash with 4 512-byte sub-pages.
*
* But in rare cases, if this optimizes things, the VID header may be placed to
* a different offset. For example, the boot-loader might do things faster if the
* VID header sits at the end of the first 2KiB NAND page with 4 sub-pages. As
* the boot-loader would not normally need to read EC headers (unless it needs
* UBI in RW mode), it might be faster to calculate ECC. This is weird example,
* but it real-life example. So, in this example, @vid_hdr_offer would be
* 2KiB-64 bytes = 1984. Note, that this position is not even 512-bytes
* aligned, which is OK, as UBI is clever enough to realize this is 4th sub-page
* of the first page and add needed padding.
*/
struct ubi_attach_req {
int32_t ubi_num;
int32_t mtd_num;
int32_t vid_hdr_offset;
uint8_t padding[12];
};
/**
* struct ubi_mkvol_req - volume description data structure used in
* volume creation requests.
* @vol_id: volume number
* @alignment: volume alignment
* @bytes: volume size in bytes
* @vol_type: volume type (%UBI_DYNAMIC_VOLUME or %UBI_STATIC_VOLUME)
* @padding1: reserved for future, not used, has to be zeroed
* @name_len: volume name length
* @padding2: reserved for future, not used, has to be zeroed
* @name: volume name
*
* This structure is used by user-space programs when creating new volumes. The
* @used_bytes field is only necessary when creating static volumes.
*
* The @alignment field specifies the required alignment of the volume logical
* eraseblock. This means, that the size of logical eraseblocks will be aligned
* to this number, i.e.,
* (UBI device logical eraseblock size) mod (@alignment) = 0.
*
* To put it differently, the logical eraseblock of this volume may be slightly
* shortened in order to make it properly aligned. The alignment has to be
* multiple of the flash minimal input/output unit, or %1 to utilize the entire
* available space of logical eraseblocks.
*
* The @alignment field may be useful, for example, when one wants to maintain
* a block device on top of an UBI volume. In this case, it is desirable to fit
* an integer number of blocks in logical eraseblocks of this UBI volume. With
* alignment it is possible to update this volume using plane UBI volume image
* BLOBs, without caring about how to properly align them.
*/
struct ubi_mkvol_req {
int32_t vol_id;
int32_t alignment;
int64_t bytes;
int8_t vol_type;
int8_t padding1;
int16_t name_len;
int8_t padding2[4];
char name[UBI_MAX_VOLUME_NAME + 1];
} __attribute__ ((packed));
/**
* struct ubi_rsvol_req - a data structure used in volume re-size requests.
* @vol_id: ID of the volume to re-size
* @bytes: new size of the volume in bytes
*
* Re-sizing is possible for both dynamic and static volumes. But while dynamic
* volumes may be re-sized arbitrarily, static volumes cannot be made to be
* smaller then the number of bytes they bear. To arbitrarily shrink a static
* volume, it must be wiped out first (by means of volume update operation with
* zero number of bytes).
*/
struct ubi_rsvol_req {
int64_t bytes;
int32_t vol_id;
} __attribute__ ((packed));
/**
* struct ubi_leb_change_req - a data structure used in atomic logical
* eraseblock change requests.
* @lnum: logical eraseblock number to change
* @bytes: how many bytes will be written to the logical eraseblock
* @dtype: data type (%UBI_LONGTERM, %UBI_SHORTTERM, %UBI_UNKNOWN)
* @padding: reserved for future, not used, has to be zeroed
*/
struct ubi_leb_change_req {
int32_t lnum;
int32_t bytes;
uint8_t dtype;
uint8_t padding[7];
} __attribute__ ((packed));
#endif /* __UBI_USER_H__ */
3. 常见问题
(1) ubifs_check_node: bad CRC: calculated 0xca82b3d7, read 0x9be0e26
ubifs_check_node: bad node at LEB 51:45312
注意这个crc错误，说的错误并不是LED 0的问题，也就是说第一个逻辑块没问题，CRC是正确的，这个问题的一个解决办法是要精确mkfs.ubifs时候的-c参数的值，不能太大了
(2)
[ 5.433349] UBIFS error (pid 71): ubifs_read_node: bad node type (150 but expected 1)
[ 5.434204] UBIFS error (pid 71): ubifs_read_node: bad node at LEB 524:4072, LEB mapping status 1
[ 5.435241] Not a node, first 24 bytes:
[ 5.435729] 00000000: 34 fb 21 ee 84 18 69 2d 60 b0 33 e6 74 f8 1c 15 da ca a1 c9 96 e3 ac 51 4.!...i-`.3.t..........Q
这个问题是，ubifs给nand驱动的buffer不一定是按照硬件对齐要求的，所以要驱动来判断，dma等地址需要页对齐等特性
(3) crc错误，但是LED 0:0
可能镜像没烧对，或者mtd-utils的版本和内核版本相差较远
(4) 可以擦掉一个分区，并不需要非要烧写image到那个分区，也可以直接mount的，这样可以做实验验证ubi的性能，前提是kernel配置好了，mount正确了
(5) 如果是跟文件系统或者要手动mount，以下步骤供参考:
uboot里, mtd命令后，看到需要用ubifs的mtd分区的编号，比如
device nand0 <rda_nand>, # parts = 10
#: name size offset mask_flags
0: bootloader 0x00200000 0x00000000 0
1: boot 0x00800000 0x00800000 0
2: system 0x08000000 0x02000000 0
system分区mtd编号是2，命令如下:
nand erase.part system
命令行参数增加
ubi.mtd=2，如果有多个，可以增加，如ubi.mtd=2,ubi.mtd=1，这样传递后，内核启动后会做attach的操作，类似android的init中的attach那样，attach之后，/dev/下就建立好设备了，进入系统后用:
内核控制台里，mount -t ubifs /dev/ubi0_0 /mnt 即可mount
