From: Bill Pringlemeir <bpringlemeir@nbsps.com>
To: u-boot@lists.denx.de
Subject: [U-Boot] Unreadable UBIFS partition after power cuts
Date: Fri, 16 Jan 2015 10:42:20 -0500 [thread overview]
Message-ID: <87iog63f6r.fsf@nbsps.com> (raw)
In-Reply-To: 150521CB44B50A4A98049E9C7BAFB0F30D9C15C8@DETENEXMB01.delta.corp
On 16 Jan 2015, Anton.Habegger at delta-es.com wrote:
> What does a "dangling branch" and "dangling match" mean? Are those
> situations handled differently under U-Boot?
... I don't know about this. However, it is easy to think that the
issue is with the UbiFs layer as it doesn't mount. However, it can be
several layer; the MTD driver, MTD, UBI and UbiFS. I guess you have
some raw image captured? If you have a raw NAND image you can try
different analysis.
Here is a nandsim mount to emulate on a PC.
# nandsim with Micron 3.3V 256MiB, 2048 bytes page.
modprobe nandsim first_id_byte=0x2c second_id_byte=0xda third_id_byte=0x90 fourth_id_byte=0x95 parts=2,64,64
flash_erase /dev/mtd3 0 0
ubiformat /dev/mtd3 -f rootfs.ubi # rootfs.ubi is the flash dump
#modprobe ubi mtd=3
#mount -t ubifs -o ro /dev/ubi0_3 /mnt/ubifs
You can put the flash id sequence for your boards flash in the 'id_byte'
and setup different partitions to emulate your actual device (ie, your
parts=... is different).
The last two lines will mount the partition using the Linux PC code.
Grab 'git://git.infradead.org/mtd-utils.git'. There is a file in
ubi-utils called ubinfo.c and it maybe helpful. Also, hujian yang has
been trying to get a 'ubidump' program merged to mtd-utils.
See:
http://lists.infradead.org/pipermail/linux-mtd/2014-December/056828.html
http://lists.infradead.org/pipermail/linux-mtd/2014-December/056829.html
and many others. The program might be helpful. I think there aren't
functional issue with the current patches, but structural issues
(fitting with mtd-utils).
In this message I attached a much simpler parser of just the UBI layer
(this is really dumb, but should be easy to modify).
http://lists.infradead.org/pipermail/linux-mtd/2014-July/054712.html
They might be useful to identify injured blocks/pages. Obviously, the
'recovery needed' shows that fixing some partial write/erase is the
issue. If UBI passes a damaged page/info to UbiFS, then it will act on
bad info.
Fwiw,
Bill Pringlemeir.
next prev parent reply other threads:[~2015-01-16 15:42 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-01-14 12:52 [U-Boot] Unreadable UBIFS partition after power cuts Anton Habegger
2015-01-14 14:57 ` Heiko Schocher
2015-01-14 17:12 ` Anton Habegger
2015-01-15 7:28 ` Heiko Schocher
2015-01-16 15:10 ` Anton Habegger
2015-01-16 15:21 ` Anton Habegger
2015-01-16 15:42 ` Bill Pringlemeir [this message]
2015-01-16 17:47 ` Anton Habegger
2015-01-17 16:06 ` Anton Habegger
2015-01-18 10:07 ` Heiko Schocher
2015-01-19 11:29 ` Anton Habegger
2015-01-20 7:26 ` Heiko Schocher
-- strict thread matches above, loose matches on Subject: below --
2015-01-16 15:38 Anton Habegger
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87iog63f6r.fsf@nbsps.com \
--to=bpringlemeir@nbsps.com \
--cc=u-boot@lists.denx.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox