UBI: infinite error loop

UBI: infinite error loop

[Patrick]

Dear mailing-list,

I am still trying to use UBI and UBIFS on my board running a 3.0.21 kernel.

When I am doing an the ubiattach -p /dev/mtd4 command I have the following infinite error loop:


UBI error: ubi_io_read: error -74 (ECC error) while reading 512 bytes from PEB 4059:512, read 512 bytes
UBI warning: ubi_io_read_vid_hdr: bad CRC at PEB 4059, calculated 0x9f1c9c0d, read 0x9f1c1c0d
UBI warning: ubi_eba_copy_leb: error 4 while reading VID header back from PEB 4059
UBI: run torture test for PEB 4059
UBI: PEB 4059 passed torture test, do not mark it as bad
UBI error: ubi_io_read: error -74 (ECC error) while reading 512 bytes from PEB 4059:512, read 512 bytes
UBI: run torture test for PEB 4059
UBI: PEB 4059 passed torture test, do not mark it as bad
UBI warning: ubi_io_read_vid_hdr: bad CRC at PEB 4059, calculated 0x42cdb8ed, read 0xff402946
UBI warning: ubi_eba_copy_leb: error 4 while reading VID header back from PEB 4059
UBI: run torture test for PEB 4059
UBI: PEB 4059 passed torture test, do not mark it as bad
UBI error: ubi_io_read: error -74 (ECC error) while reading 512 bytes from PEB 4059:512, read 512 bytes
UBI: run torture test for PEB 4059
....


I don't think it's normal that this error loop infinitely again and again...

Any help is appreciated !

Thanks in advance

Patrick

Re: UBI: infinite error loop

[Matthieu CASTET]

Patrick a écrit :
> Dear mailing-list,
> 
> I am still trying to use UBI and UBIFS on my board running a 3.0.21 kernel.
> 
> When I am doing an the ubiattach -p /dev/mtd4 command I have the following infinite error loop:
> 
> 
> UBI error: ubi_io_read: error -74 (ECC error) while reading 512 bytes from PEB 4059:512, read 512 bytes
> UBI warning: ubi_io_read_vid_hdr: bad CRC at PEB 4059, calculated 0x9f1c9c0d, read 0x9f1c1c0d
> UBI warning: ubi_eba_copy_leb: error 4 while reading VID header back from PEB 4059
> UBI: run torture test for PEB 4059
> UBI: PEB 4059 passed torture test, do not mark it as bad
> UBI error: ubi_io_read: error -74 (ECC error) while reading 512 bytes from PEB 4059:512, read 512 bytes
> UBI: run torture test for PEB 4059
> UBI: PEB 4059 passed torture test, do not mark it as bad
> UBI warning: ubi_io_read_vid_hdr: bad CRC at PEB 4059, calculated 0x42cdb8ed, read 0xff402946
> UBI warning: ubi_eba_copy_leb: error 4 while reading VID header back from PEB 4059
> UBI: run torture test for PEB 4059
> UBI: PEB 4059 passed torture test, do not mark it as bad
> UBI error: ubi_io_read: error -74 (ECC error) while reading 512 bytes from PEB 4059:512, read 512 bytes
> UBI: run torture test for PEB 4059
> .....
> 
> 
> I don't think it's normal that this error loop infinitely again and again...
> 
> Any help is appreciated !
> 
> Thanks in advance
> 
What's your nand driver ?
What's your nand device ?

Look like it doesn't handle subpage reading : you have ecc error when reading
512 bytes at offset 512.

Maybe you should pass -O 2048 to ubiattach

Re: Re : UBI: infinite error loop

[Matthieu CASTET]

Please keep mtd ML in CC

Patrick a écrit :
> 
>>> Dear mailing-list,
>>>
>>> I am still trying to use UBI and UBIFS on my board running a 3.0.21 kernel.
>>>
>>> When I am doing an the ubiattach -p /dev/mtd4 command I have the following infinite error loop:
>>>
>>>
>>> .....
>>>
>>>
>>> I don't think it's normal that this error loop infinitely again and again...
>>>
>>> Any help is appreciated !
>>>
>>> Thanks in advance
>>>
>> What's your nand driver ?
>> What's your nand device ?
>>
>> Look like it doesn't handle subpage reading : you have ecc error when reading
>> 512 bytes at offset 512.
>>
>> Maybe you should pass -O 2048 to ubiattach
>>
> 
> Hello Matthieu,
> 
> Thanks for your answer.
> 
> I have a Micron MT29F4G08ABBDAH4 and I use the MTD_NAND driver.
But what driver is using your nand controller ?


> I have made a try with the -O 2048 flag but it's worst:
You need to use before ubiformat with the same flag.

Matthieu

Re: Re : Re : UBI: infinite error loop

[Matthieu CASTET]

Patrick a écrit :
> 
>>>>> Dear mailing-list,
>>>>>
>>>>> I am still trying to use UBI and UBIFS on my board running a 3.0.21 kernel.
>>>>>
>>>>> When I am doing an the ubiattach -p /dev/mtd4 command I have the following infinite error loop:
>>>>>
>>>>>
>>>>> .....
>>>>>
>>>>>
>>>>> I don't think it's normal that this error loop infinitely again and again...
>>>>>
>>>>> Any help is appreciated !
>>>>>
>>>>> Thanks in advance
>>>>>
>>>> What's your nand driver ?
>>>> What's your nand device ?
>>>>
>>>> Look like it doesn't handle subpage reading : you have ecc error when reading
>>>> 512 bytes at offset 512.
>>>>
>>>> Maybe you should pass -O 2048 to ubiattach
>>>>
>>> Hello Matthieu,
>>>
>>> Thanks for your answer.
>>>
>>> I have a Micron MT29F4G08ABBDAH4 and I use the MTD_NAND driver.
>> But what driver is using your nand controller ?
> 
> MTD_NAND_OMAP2
> 
>>> I have made a try with the -O 2048 flag but it's worst:
>> You need to use before ubiformat with the same flag.
> 
> It's seems to help. I could attach without error, but now I couldn't mount the volume:
> "LEB size mismatch: 129024 in superblock, 126976 real"
> I will try to rebuild my fs image with this size of LEB
> 
Yes now your leb size is 62*2K instead of 63*2K.

Matthieu

Re: UBI: infinite error loop

[Artem Bityutskiy]

On Fri, 2012-02-24 at 13:35 +0000, Patrick wrote:
> I don't think it's normal that this error loop infinitely again and again...
> 
> Any help is appreciated !

You've got help AFAICS. I agree that an infinite loop is not normal. But
the right thing for UBI to do in this situation would be to mark all
your eraseblocs as bad, which is not what you want I guess, so it is
probably better to have this infinite loop?

-- 
Best Regards,
Artem Bityutskiy

Re: UBI: infinite error loop

[Brian Foster]

matthieu.castet@parrot.com wrote:
> Patrick a écrit :
> > When I am doing an the ubiattach -p /dev/mtd4 command
> > I have the following infinite error loop:
>[ ... ]
> Look like it doesn't handle subpage reading :
> you have ecc error when reading 512 bytes at offset 512.

 Also, keep in mind read-after-write checks don't work with subpages,
 http://www.linux-mtd.infradead.org/faq/ubi.html#L_subpage_verify_fail
───────────────────────────────────────────────────────────
I get "ubi_io_write: error -5 while writing 512 bytes to PEB 5:512"

If you have a 2048 bytes per NAND page device, and have
CONFIG_MTD_NAND_VERIFY_WRITE enabled in your kernel,
you will need to turn it off.  The code does not currently
(as of 2.6.26) perform verification of sub-page writes
correctly.  As UBI is one of the few users of sub-page
writes, not much else seems to be affected by this bug.
───────────────────────────────────────────────────────────

 YES, I realize that is not your error, but this was
 something that badly tripped us up until an intense
 search-engine session ....  ;-\  (I have no idea if
 this has since been fixed or not?)

cheers!
	-blf-

-- 
Brian Foster
Principal MTS, Software        |  La Ciotat, France
Maxim Integrated Products      |  Web:  http://www.maxim-ic.com/

Re: UBI: infinite error loop

[Artem Bityutskiy]

On Fri, 2012-02-24 at 17:16 +0100, Brian Foster wrote:
>  Also, keep in mind read-after-write checks don't work with subpages,
>  http://www.linux-mtd.infradead.org/faq/ubi.html#L_subpage_verify_fail
> ───────────────────────────────────────────────────────────
> I get "ubi_io_write: error -5 while writing 512 bytes to PEB 5:512"
> 
> If you have a 2048 bytes per NAND page device, and have
> CONFIG_MTD_NAND_VERIFY_WRITE enabled in your kernel,
> you will need to turn it off.  The code does not currently
> (as of 2.6.26) perform verification of sub-page writes
> correctly.  As UBI is one of the few users of sub-page
> writes, not much else seems to be affected by this bug.
> ───────────────────────────────────────────────────────────
> 
>  YES, I realize that is not your error, but this was
>  something that badly tripped us up until an intense
>  search-engine session ....  ;-\  (I have no idea if
>  this has since been fixed or not?)

Well, no one cares enough to investigate this and fix the issue. I think
this article should be removed - may be this will inspire someone to
send a fix...

-- 
Best Regards,
Artem Bityutskiy