Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

mtd errors on boot OpenRD "ultimate"

5 views
Skip to first unread message

Rick Thomas

unread,
Aug 6, 2011, 1:50:01 PM8/6/11
to

For a few seconds during the boot process on my OpenRD "ultimate",
running Debian Squeeze, I get hundreds of messages in syslog like this:

Buffer I/O error on device mtdblock0, logical block 0
Buffer I/O error on device mtdblock0, logical block 8
end_request: I/O error, dev mtdblock0, sector 0
end_request: I/O error, dev mtdblock0, sector 128
end_request: I/O error, dev mtdblock0, sector 16
end_request: I/O error, dev mtdblock0, sector 64
end_request: I/O error, dev mtdblock0, sector 8
loop: module loaded
uncorrectable error :


Does anybody know what causes them? Is there anything I can (or
should) do about them?


Thanks!

Rick


--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Archive: http://lists.debian.org/425E049C-4C6F-4B7B...@pobox.com

Rick Thomas

unread,
Aug 6, 2011, 2:10:02 PM8/6/11
to
Interestingly, and confusingly...

I get no such messages on an OpenRD "base" sitting right beside the
"ultimate" running an essentially identical software load.

Any thoughts?

Rick

Archive: http://lists.debian.org/ADE1F4B1-18FC-456D...@pobox.com

Wouter Verhelst

unread,
Aug 6, 2011, 5:50:01 PM8/6/11
to
On Sat, Aug 06, 2011 at 01:51:11PM -0400, Rick Thomas wrote:
> Interestingly, and confusingly...
>
> I get no such messages on an OpenRD "base" sitting right beside the
> "ultimate" running an essentially identical software load.
>
> Any thoughts?

I guess you've got some broken bits in your flash. There usually isn't a
wear-leveling controller in whatever sits behind the mtd driver, so
these errors are exposed rather than hidden (as would be the case on a
removable flash medium such as a USB stick or an sd card or some such).

The fact that it happens with one installation but not with the other
suggests that it's not your entire system that's fried (which is a good
thing, I guess).

--
The volume of a pizza of thickness a and radius z can be described by
the following formula:

pi zz a


--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org

Archive: http://lists.debian.org/20110806214...@grep.be

Philippe Clérié

unread,
Aug 6, 2011, 7:10:01 PM8/6/11
to
When I first encountered those messages I was told that they were completely
harmless and have always been there. You'd best ignore them.

--


Philippe

------
The trouble with common sense is that it is so uncommon.
<Anonymous>

Archive: http://lists.debian.org/201108061802....@gcal.net

Rick Thomas

unread,
Aug 7, 2011, 12:10:01 AM8/7/11
to
Hi again,

Some googling on the subject offers a suggestion that is attractive, even if not easily verified:

The suggestion is that Uboot and Linux do different and incompatible checksums on the write-blocks in the NAND flash. So you either get messages in your syslog from Linux saying it doesn't like the Uboot-style checksums, or you get messages from Uboot saying it doesn't like the Linux-style checksums. The Uboot messages won't show up in syslog, but they will pollute your serial console during boot-up. Pay your money and take your pick. Both are unpleasant.

Does anybody out there have true and certain knowledge if this suggestion is even close?

If so, would the new Debian Uboot package have Linux compatible checksums?

Rick


On Aug 6, 2011, at 6:02 PM, Philippe Clérié wrote:

> When I first encountered those messages I was told that they were completely
> harmless and have always been there. You'd best ignore them.
>

> On Saturday 06 August 2011 13:31:24 Rick Thomas wrote:
>> For a few seconds during the boot process on my OpenRD "ultimate",
>> running Debian Squeeze, I get hundreds of messages in syslog like this:
>>
>> Buffer I/O error on device mtdblock0, logical block 0
>> Buffer I/O error on device mtdblock0, logical block 8
>> end_request: I/O error, dev mtdblock0, sector 0
>> end_request: I/O error, dev mtdblock0, sector 128
>> end_request: I/O error, dev mtdblock0, sector 16
>> end_request: I/O error, dev mtdblock0, sector 64
>> end_request: I/O error, dev mtdblock0, sector 8
>> loop: module loaded
>> uncorrectable error :
>>
>>
>> Does anybody know what causes them? Is there anything I can (or
>> should) do about them?
>>
>>
>> Thanks!
>>
>> Rick


--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org

Archive: http://lists.debian.org/28F1BB66-7309-4BA3...@pobox.com

Alexander Clouter

unread,
Aug 7, 2011, 7:10:01 AM8/7/11
to
Rick Thomas <rbth...@pobox.com> wrote:
>
> For a few seconds during the boot process on my OpenRD "ultimate",
> running Debian Squeeze, I get hundreds of messages in syslog like this:
>
> Buffer I/O error on device mtdblock0, logical block 0
> Buffer I/O error on device mtdblock0, logical block 8
> end_request: I/O error, dev mtdblock0, sector 0
> end_request: I/O error, dev mtdblock0, sector 128
> end_request: I/O error, dev mtdblock0, sector 16
> end_request: I/O error, dev mtdblock0, sector 64
> end_request: I/O error, dev mtdblock0, sector 8
> loop: module loaded
> uncorrectable error :
>
> Does anybody know what causes them? Is there anything I can (or
> should) do about them?
>
If I am right it's because u-boot does not support (or uses a different
ECC method) for the NAND to Linux. When Linux reads blocks from the
NAND the ECC check fails and you get an IO error.

The messages are harmless though for mtdblock0 and mtdblock1, it's only
if they appear on mtdblock2 should you get worried :)

Cheers

--
Alexander Clouter
.sigmonster says: Anything is possible on paper.
-- Ron McAfee


--
To UNSUBSCRIBE, email to debian-ar...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org

Archive: http://lists.debian.org/ugg0h8-...@chipmunk.wormnet.eu

Rick Thomas

unread,
Aug 7, 2011, 4:40:02 PM8/7/11
to

Thanks for the confirmation, Alexander!

Would there be any point in trying to generate a Linux kernel patch to
accept both kinds of ECC?

Rick

PS: Can you explain (for us less knowledgeable folks) why they are
harmless for mtdblock[01] but not for mtdblock[2]?


Alexander Coulter says:

> If I am right it's because u-boot does not support (or uses a
> different
> ECC method) for the NAND to Linux. When Linux reads blocks from the
> NAND the ECC check fails and you get an IO error.
>
> The messages are harmless though for mtdblock0 and mtdblock1, it's
> only
> if they appear on mtdblock2 should you get worried :)

Archive: http://lists.debian.org/39EB0A21-DE12-4096...@pobox.com

0 new messages