Ubi fastmap是ubi的一个可配置属性, 这个属性是通过内核的CONFIG_MTD_UBI_FASTMAP这个配置来使能的,

使能ubi fastmap之后, 就会在ubi内部新建一个分卷来存储fastmap的数据.

使能ubi fastmap后, 当执行一次ubiattach之后, 会把ubi各个分卷的扫描信息存储在ubi fastmap的分卷上,

下一次再执行ubiattach后, 就可以使用fastmap分卷上的信息, 不用再扫描整个ubi分区,

所以在使用fastmap可以减少ubi flash的扫描时间,

在我们系统上使用ubifastmap功能后会比不使用ubifastmap启动快3~4s.

更多ubi fastmap信息参考官方文档介绍:

http://www.linux-mtd.infradead.org/doc/ubi.html#L_fastmap


BAREBOX中使能fastmap功能

barebox中ubi fastmap特性介绍:

https://www.barebox.org/doc/latest/user/ubi.html

在kernel中是配置了CONFIG_MTD_UBI_FASTMAP的配置后, 启动kernel后就会生成一个ubi fastmap的内部分卷,

在新建ubi fastmap内部分卷之后, 需要在barebox中使用ubiformat格式化整个nand0.root分区, 然后在attach整个ubi分区ubi fastmap才能起作用.

MX6核心板版本出厂测试时BSP版本CONFIG_MTD_UBI_FASTMAP
老SOMPD16n
新SOMPD18y

在MX6的BSP中, PD16的版本是不使能CONFIG_MTD_UBI_FASTMAP配置的, 所以老版本的SOM出厂后如果要使能ubi fastmap功能的话,

需要先更新一遍PD18的镜像, 然后启动进入系统后, 此时ubi内部会新建一个ubi fastmap的分卷,

然后再重启进入barebox, 使用ubiformat格式化nand0.root分区, 然后在重新烧写ubi分区的镜像后就可以使能fastmap的功能.

而新SOM在出厂时是使用PD18的镜像测试的, 此时ubi内部已经有ubi fastmap的分卷, 出厂后直接烧写PD18版本的镜像就可以使能ubi fastmap的功能.


MX6 PD18.1.2 Ubi Fastmap测试

使能ubi fastmap的情况

barebox ubiattach的log:

booting 'nand'
ubi0: attached by fastmap
ubi0: fastmap pool size: 256
ubi0: fastmap WL pool size: 128
ubi0: registering /dev/nand0.root.ubi
ubi0: registering kernel as /dev/nand0.root.ubi.kernel
ubi0: registering oftree as /dev/nand0.root.ubi.oftree
ubi0: registering root as /dev/nand0.root.ubi.root
ubi0: attached mtd0 (name "nand0.root", size 1007 MiB) to ubi0
ubi0: PEB size: 131072 bytes (128 KiB), LEB size: 126976 bytes
ubi0: min./max. I/O unit sizes: 2048/2048, sub-page size 2048
ubi0: VID header offset: 2048 (aligned 2048), data offset: 4096
ubi0: good PEBs: 8052, bad PEBs: 4, corrupted PEBs: 0
ubi0: user volume: 3, internal volumes: 1, max. volumes count: 128
ubi0: max/mean erase counter: 180/106, WL threshold: 65536, image sequence number: 21814
ubi0: available PEBs: 0, total reserved PEBs: 8052, PEBs reserved for bad PEB handling: 156

kernel中ubi attach的log:

[    4.250283] ubi0: default fastmap pool size: 256
[    4.254988] ubi0: default fastmap WL pool size: 128
[    4.259924] ubi0: attaching mtd2
[    4.493051] ubi0: attached by fastmap
[    4.496727] ubi0: fastmap pool size: 256
[    4.500654] ubi0: fastmap WL pool size: 128
[    4.519748] ubi0: attached mtd2 (name "root", size 1007 MiB)
[    4.525497] ubi0: PEB size: 131072 bytes (128 KiB), LEB size: 126976 bytes
[    4.532395] ubi0: min./max. I/O unit sizes: 2048/2048, sub-page size 2048
[    4.539186] ubi0: VID header offset: 2048 (aligned 2048), data offset: 4096
[    4.546179] ubi0: good PEBs: 8052, bad PEBs: 4, corrupted PEBs: 0
[    4.552290] ubi0: user volume: 3, internal volumes: 1, max. volumes count: 128
[    4.559530] ubi0: max/mean erase counter: 180/106, WL threshold: 4096, image sequence number: 21814
[    4.568592] ubi0: available PEBs: 0, total reserved PEBs: 8052, PEBs reserved for bad PEB handling: 156
[    4.578032] ubi0: background thread "ubi_bgt0d" started, PID 135

从kernel log中看出ubiattach时大概花了0.2s.

系统启动时间:

root@phyboard-mira-imx6-3:~# systemd-analyze
Startup finished in 4.920s (kernel) + 3.869s (userspace) = 8.790s

上电到phytec-qtdemo启动时间约为18.45s

不使能ubi fastmap的情况

barebox ubiattach的log:

booting 'nand'
ubi0: scanning is finished
ubi0 warning: print_rsvd_warning: cannot reserve enough PEBs for bad PEB handling, reserved 60, need 156
ubi0: registering /dev/nand0.root.ubi
ubi0: registering kernel as /dev/nand0.root.ubi.kernel
ubi0: registering oftree as /dev/nand0.root.ubi.oftree
ubi0: registering root as /dev/nand0.root.ubi.root
ubi0: attached mtd0 (name "nand0.root", size 1007 MiB) to ubi0
ubi0: PEB size: 131072 bytes (128 KiB), LEB size: 126976 bytes
ubi0: min./max. I/O unit sizes: 2048/2048, sub-page size 2048
ubi0: VID header offset: 2048 (aligned 2048), data offset: 4096
ubi0: good PEBs: 8052, bad PEBs: 4, corrupted PEBs: 0
ubi0: user volume: 3, internal volumes: 1, max. volumes count: 128
ubi0: max/mean erase counter: 178/105, WL threshold: 65536, image sequence number: 5862
ubi0: available PEBs: 0, total reserved PEBs: 8052, PEBs reserved for bad PEB handling: 60

kernel中ubi attach的log:

[    4.252653] ubi0: default fastmap pool size: 256
[    4.257276] ubi0: default fastmap WL pool size: 128
[    4.262200] ubi0: attaching mtd2
[    4.750678] random: crng init done
[    7.099309] ubi0: scanning is finished
[    7.114345] ubi0 warning: ubi_eba_init: cannot reserve enough PEBs for bad PEB handling, reserved 60, need 156
[    7.128042] ubi0: attached mtd2 (name "root", size 1007 MiB)
[    7.133753] ubi0: PEB size: 131072 bytes (128 KiB), LEB size: 126976 bytes
[    7.140635] ubi0: min./max. I/O unit sizes: 2048/2048, sub-page size 2048
[    7.147473] ubi0: VID header offset: 2048 (aligned 2048), data offset: 4096
[    7.154460] ubi0: good PEBs: 8052, bad PEBs: 4, corrupted PEBs: 0
[    7.160575] ubi0: user volume: 3, internal volumes: 1, max. volumes count: 128
[    7.167822] ubi0: max/mean erase counter: 178/105, WL threshold: 4096, image sequence number: 5862
[    7.176817] ubi0: available PEBs: 0, total reserved PEBs: 8052, PEBs reserved for bad PEB handling: 60
[    7.186169] ubi0: background thread "ubi_bgt0d" started, PID 135

从kernel log中看出ubiattach时大概花了2.8s.

系统启动时间:

root@phyboard-mira-imx6-3:~# systemd-analyze 
Startup finished in 7.504s (kernel) + 3.785s (userspace) = 11.290s

上电到phytec-qtdemo启动时间约为22.22s.




  • No labels

support@phytec.cn
热线:0755-61802110-803