Ubifs power cuts errors (oe-v2.6) on Colibri T20

I’ve got installation with about 50 devices working on Colibri T20 modules (working in city buses). In every day each device is powered on in the morning and powered off in the evening (power cut). Every week I’ve got few modules which can’t mount ubifs and have to reflash Colibri module. I’ve tried mounting ubifs in sync mode (in /etc/fstab file) but still have corrupted modules. Is there any solution to fix this?

What you are seeing is probably related to the following issue with a current ETA of Q1 2017.

Please note that beta versions of BSPs have not passed full validation & verification and are therefore not recommended for production use.

I see in the release details of 2.8.b3, that problem with UBIFS error after power cuts on T20 persist. Because of this we use V2.3, but need the features of the latest image. Is there any chance that it will be fixed?

After cca 300 power cuts we got this message during boot (2.8b2):

[    6.124909] CPU rate: 1000 MHz
[    6.129959] DVFS: Got RTC device name:rtc0
[    6.139384] Enabling Tegra protected aperture at 0x16700000
[    6.147915] registered taskstats version 1
[    6.154963] regulator_init_complete: REG-LDO_9: incomplete constraints, leaving on
[    6.169568] UBI: default fastmap pool size: 200
[    6.176238] UBI: default fastmap WL pool size: 25
[    6.183019] UBI: attaching mtd3 to ubi0
[    6.526441] UBI: attached by fastmap
[    6.532478] UBI: fastmap pool size: 200
[    6.539879] UBI: fastmap WL pool size: 25
[    6.568013] UBI: attached mtd3 (name "ubi", size 1020 MiB) to ubi0
[    6.577801] UBI: PEB size: 262144 bytes (256 KiB), LEB size: 253952 bytes
[    6.588269] UBI: min./max. I/O unit sizes: 4096/4096, sub-page size 4096
[    6.598832] UBI: VID header offset: 4096 (aligned 4096), data offset: 8192
[    6.609772] UBI: good PEBs: 4080, bad PEBs: 0, corrupted PEBs: 0
[    6.620011] UBI: user volume: 3, internal volumes: 1, max. volumes count: 128
[    6.631626] UBI: max/mean erase counter: 2/1, WL threshold: 4096, image sequence number: 0
[    6.649128] UBI: available PEBs: 0, total reserved PEBs: 4080, PEBs reserved for bad PEB handling: 80
[    6.668653] UBI: background thread "ubi_bgt0d" started, PID 77
[    6.669988] input: gpio-keys as /devices/platform/gpio-keys.0/input/input1
[    6.693917] rtc-ds1307 0-0068: setting system clock to 2018-04-30 08:34:34 UTC (1525077274)
[    6.716661] UBIFS: background thread "ubifs_bgt0_2" started, PID 78
[    6.795528] UBIFS: recovery needed
[    7.050470] UBIFS error (pid 1): ubifs_recover_leb: corruption -3
[    7.062766] UBIFS error (pid 1): ubifs_scanned_corruption: corruption at LEB 1019:57248
[    7.082650] UBIFS error (pid 1): ubifs_scanned_corruption: first 8192 bytes from LEB 1019:57248
[    7.127729] UBIFS error (pid 1): ubifs_recover_leb: LEB 1019 scanning failed
[    7.144232] UBIFS: background thread "ubifs_bgt0_2" stops
[    7.209437] VFS: Cannot open root device "ubi0:rootfs" or unknown-block(0,0)
[    7.225426] Please append a correct "root=" boot option; here are the available partitions:
[    7.251387] 1f00            2048 mtdblock0  (driver?)
[    7.265308] 1f01            1024 mtdblock1  (driver?)
[    7.279002] 1f02            1024 mtdblock2  (driver?)

I see in the release details of 2.8.b3, that problem with UBIFS error after power cuts on T20 persist.


Because of this we use V2.3,

Yes, using stable BSP 2.3 is still recommended.

but need the features of the latest image.

What exact features are you talking about?

Is there any chance that it will be fixed?

Yes, there is always a chance.

I need a support for LM816 (RTL8188EU) which was added in 2.7b4

As e.g. explained here you may actually use the latest downstream Linux kernel sources even with BSP 2.3 and with it whatever matching backports as well.