[U-Boot] SPL boot with NAND on i.MX6 data abort

[U-Boot] SPL boot with NAND on i.MX6 data abort

[Markus Stelzhammer]

Hi all,


I'm currently working on the port of u-boot for our custom boards.
Previously we were using u-boot based on v2013.04 without SPL for our 
boards (we have boards with SD also) and this worked for all of our devices.

As we now have new boards with different (bigger) RAM, I decided to use 
the newest bootloader (v2015.10) and enable SPL support.
This is already working on a board with SD-Card but I'm not getting it 
to work on the boards with NAND.

When I try to boot, SPL is detecting and loading u-boot but u-boot 
throws a "data abort" during "ubifs mount".
The strange thing for me is that when I disable SPL, the board will boot 
properly.

Could anyone give me a hint how I could trace this error?
I can also provide logs or code snippets if needed!


Best Regards

Markus

[U-Boot] SPL boot with NAND on i.MX6 data abort

[Stefano Babic]

Hi Markus,

On 23/09/2015 10:02, Markus Stelzhammer wrote:
> Hi all,
> 
> 
> I'm currently working on the port of u-boot for our custom boards.
> Previously we were using u-boot based on v2013.04 without SPL for our
> boards (we have boards with SD also) and this worked for all of our
> devices.
> 
> As we now have new boards with different (bigger) RAM, I decided to use
> the newest bootloader (v2015.10) and enable SPL support.
> This is already working on a board with SD-Card but I'm not getting it
> to work on the boards with NAND.
> 
> When I try to boot, SPL is detecting and loading u-boot but u-boot
> throws a "data abort" during "ubifs mount".

u-boot ist also in NAND ? Then SPL has configured correctly the device
else u-boot was not loaded.

> The strange thing for me is that when I disable SPL, the board will boot
> properly.

You can try to follow the data abort and check which addresses are
involved. We cannot underestimate that the bug is still present in
u-boot, but it appears only together with SPL.

> 
> Could anyone give me a hint how I could trace this error?
> I can also provide logs or code snippets if needed!

Some initialization done previously in u-boot, is done now by SPL -
check your code if there is something missing.

As your board is not mainlined, we cannot help a much with code.

Best regards,
Stefano Babic

-- 
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: sbabic at denx.de
=====================================================================

[U-Boot] SPL boot with NAND on i.MX6 data abort

[Markus Stelzhammer]

Hi Stefano,


Today I grabbed a different Hardware with a bigger NAND chip (512 
instead of 256MB).
At this hardware, SPL is not able to jump to U-Boot. So currently I 
assume that I might have some problems when flashing.

Would you be so kind to review my header file if I miss something or if 
there is something completely wrong?! (I attached it)
The commands I use to flash are the following:

 > ./kobs-ng init -v -x SPL
 > nandwrite --start=0x11000 --pad /dev/mtd0 u-boot.uim         (I also 
tried to flash u-boot.bin and u-boot.imx)

The output I get at boot are the following, but then it stucks.
.......
 >>spl:board_init_r()
using memory 0x18300000-0x1b500000 for malloc()
spl_init()
boot device - 4
spl: nand - using hw ecc
NAND : 0x01:0xdc erasesize=1 (>>0)
writesize=-1 (>>0)
oobsize=65535
chipsize=1095216660225
Not support the NAND chips whose oob size is larger then 1024 bytes!
1044479 MiB
nand_spl_load_image offset:0x00011000 len:64 page:69632
mkimage signature not found - ih_magic = 0
nand_spl_load_image offset:0x00011000 len:409600 page:69632
Jumping to U-Boot
loaded - jumping to U-Boot...image entry point: 0x17800000


On 09/24/2015 11:25 AM, Stefano Babic wrote:
> Hi Markus,
>
> On 23/09/2015 10:02, Markus Stelzhammer wrote:
>> Hi all,
>>
>>
>> I'm currently working on the port of u-boot for our custom boards.
>> Previously we were using u-boot based on v2013.04 without SPL for our
>> boards (we have boards with SD also) and this worked for all of our
>> devices.
>>
>> As we now have new boards with different (bigger) RAM, I decided to use
>> the newest bootloader (v2015.10) and enable SPL support.
>> This is already working on a board with SD-Card but I'm not getting it
>> to work on the boards with NAND.
>>
>> When I try to boot, SPL is detecting and loading u-boot but u-boot
>> throws a "data abort" during "ubifs mount".
> u-boot ist also in NAND ? Then SPL has configured correctly the device
> else u-boot was not loaded.
>
>> The strange thing for me is that when I disable SPL, the board will boot
>> properly.
> You can try to follow the data abort and check which addresses are
> involved. We cannot underestimate that the bug is still present in
> u-boot, but it appears only together with SPL.
>
>> Could anyone give me a hint how I could trace this error?
>> I can also provide logs or code snippets if needed!
> Some initialization done previously in u-boot, is done now by SPL -
> check your code if there is something missing.
>
> As your board is not mainlined, we cannot help a much with code.
>
> Best regards,
> Stefano Babic
>

Best regards,
Markus
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sigmatek-mx6.h
Type: text/x-chdr
Size: 7606 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20150924/9597ebe1/attachment.h>

[U-Boot] SPL boot with NAND on i.MX6 data abort

[Heiko Schocher]

Hello Markus,

Am 24.09.2015 um 17:03 schrieb Markus Stelzhammer:
> Hi Stefano,
>
>
> Today I grabbed a different Hardware with a bigger NAND chip (512 instead of 256MB).
> At this hardware, SPL is not able to jump to U-Boot. So currently I assume that I might have some
> problems when flashing.
>
> Would you be so kind to review my header file if I miss something or if there is something
> completely wrong?! (I attached it)
> The commands I use to flash are the following:
>
>  > ./kobs-ng init -v -x SPL
>  > nandwrite --start=0x11000 --pad /dev/mtd0 u-boot.uim         (I also tried to flash u-boot.bin
> and u-boot.imx)
>
> The output I get at boot are the following, but then it stucks.
> .......
>  >>spl:board_init_r()
> using memory 0x18300000-0x1b500000 for malloc()
> spl_init()
> boot device - 4
> spl: nand - using hw ecc
> NAND : 0x01:0xdc erasesize=1 (>>0)

 From here the values seems bogus to me ... erasesize 1 ?

> writesize=-1 (>>0)

negative writesize?

> oobsize=65535
> chipsize=1095216660225

Such a big chip?

> Not support the NAND chips whose oob size is larger then 1024 bytes!
> 1044479 MiB

Wow ...

I think you have problems with detecting the NAND, maybe
wrong bus timings?

bye,
Heiko

> nand_spl_load_image offset:0x00011000 len:64 page:69632
> mkimage signature not found - ih_magic = 0
> nand_spl_load_image offset:0x00011000 len:409600 page:69632
> Jumping to U-Boot
> loaded - jumping to U-Boot...image entry point: 0x17800000
>
>
> On 09/24/2015 11:25 AM, Stefano Babic wrote:
>> Hi Markus,
>>
>> On 23/09/2015 10:02, Markus Stelzhammer wrote:
>>> Hi all,
>>>
>>>
>>> I'm currently working on the port of u-boot for our custom boards.
>>> Previously we were using u-boot based on v2013.04 without SPL for our
>>> boards (we have boards with SD also) and this worked for all of our
>>> devices.
>>>
>>> As we now have new boards with different (bigger) RAM, I decided to use
>>> the newest bootloader (v2015.10) and enable SPL support.
>>> This is already working on a board with SD-Card but I'm not getting it
>>> to work on the boards with NAND.
>>>
>>> When I try to boot, SPL is detecting and loading u-boot but u-boot
>>> throws a "data abort" during "ubifs mount".
>> u-boot ist also in NAND ? Then SPL has configured correctly the device
>> else u-boot was not loaded.
>>
>>> The strange thing for me is that when I disable SPL, the board will boot
>>> properly.
>> You can try to follow the data abort and check which addresses are
>> involved. We cannot underestimate that the bug is still present in
>> u-boot, but it appears only together with SPL.
>>
>>> Could anyone give me a hint how I could trace this error?
>>> I can also provide logs or code snippets if needed!
>> Some initialization done previously in u-boot, is done now by SPL -
>> check your code if there is something missing.
>>
>> As your board is not mainlined, we cannot help a much with code.
>>
>> Best regards,
>> Stefano Babic
>>
>
> Best regards,
> Markus
>
>
> _______________________________________________
> U-Boot mailing list
> U-Boot at lists.denx.de
> http://lists.denx.de/mailman/listinfo/u-boot
>

-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

[U-Boot] SPL boot with NAND on i.MX6 data abort

[Stefano Babic]

Hi Markus,

On 24/09/2015 17:03, Markus Stelzhammer wrote:
> Hi Stefano,
> 
> 
> Today I grabbed a different Hardware with a bigger NAND chip (512
> instead of 256MB).
> At this hardware, SPL is not able to jump to U-Boot. So currently I
> assume that I might have some problems when flashing.
> 

If you are not flashing correctly, SPL does not run. It looks like that
your board boots (SPL runs).

> Would you be so kind to review my header file if I miss something or if
> there is something completely wrong?! (I attached it)
> The commands I use to flash are the following:
> 
>> ./kobs-ng init -v -x SPL
>> nandwrite --start=0x11000 --pad /dev/mtd0 u-boot.uim                
> (I also tried to flash u-boot.bin and u-boot.imx)
> 
> The output I get at boot are the following, but then it stucks.
> .......
>>>spl:board_init_r()
> using memory 0x18300000-0x1b500000 for malloc()
> spl_init()
> boot device - 4
> spl: nand - using hw ecc
> NAND : 0x01:0xdc erasesize=1 (>>0)

erasesize = 1 ??

> writesize=-1 (>>0)
> oobsize=65535
> chipsize=1095216660225
> Not support the NAND chips whose oob size is larger then 1024 bytes!
> 1044479 MiB

This larger NAND cannot be recognized correctly - it is not a problem by
flashing.

You have to investigate why this NAND (even if ONFI) is not recognized,
but it is a different problem as the one you reported previously.

Best regards,
Stefano Babic


-- 
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: sbabic at denx.de
=====================================================================