[U-Boot] Problems upgrading NAND

[U-Boot] Problems upgrading NAND

[Alemao]

Hi all,

I was using a NAND flash from ST Micro, 512 Mbit, with no problems in
U-boot-1.2.0.

So I upgraded this nand to a 1 Gbit, also from ST Micro. But now U-boot
read only zeros. In linux 2.6.26 both NANDs works.

NAND is connect in Freescale Local Bus UPM.

U-boot finds the NAND normally:

NAND:  128 MiB
In:    serial
Out:   serial
Err:   serial

And the Chip ID is correct in nand_ids.c (dont know the other options):

{"NAND 128MiB 3,3V 8-bit",    0xF1, 0, 128, 0,
NAND_SAMSUNG_LP_OPTIONS | NAND_NO_AUTOINCR}

Before I start patching, any hints if could be in nand core or fsl upm driver?
Or if NAND1G have some peculiarity? Cause I didnt see in datasheet
anything that called my attention.

NAND part numbers:

NAND512W3A
NAND01GW3B


Cheers,

--
Alemao

[U-Boot] Problems upgrading NAND

[Scott Wood]

On Thu, Jan 15, 2009 at 02:46:28PM -0200, Alemao wrote:
> I was using a NAND flash from ST Micro, 512 Mbit, with no problems in
> U-boot-1.2.0.

I'm guessing that 1.2.0 was a BSP u-boot, that was not upstream.  I don't
think upstream 1.2.0 had UPM NAND support.

Could you try the latest upstream U-boot?

What board are you using?

-Scott

[U-Boot] Problems upgrading NAND

[Alemao]

You are right, there were no UPM support on U-boot-1.2.0.
I used patches sended by Anton from here:

http://www.nabble.com/-PATCH-v2-0-5--mpc83xx%3A-MPC8360E-RDK-related-patches-td14717360.html#a14717487

If I use the latest upstream I will have to do a lot of validations.
If a patch nand core, also I will have to validate, but just nand
related things.

So if it just a detail in fsl upm driver, will save a lot of time.

Im using a MPC8343 board.

On Thu, Jan 15, 2009 at 4:23 PM, Scott Wood <scottwood@freescale.com> wrote:
> On Thu, Jan 15, 2009 at 02:46:28PM -0200, Alemao wrote:
>> I was using a NAND flash from ST Micro, 512 Mbit, with no problems in
>> U-boot-1.2.0.
>
> I'm guessing that 1.2.0 was a BSP u-boot, that was not upstream.  I don't
> think upstream 1.2.0 had UPM NAND support.
>
> Could you try the latest upstream U-boot?
>
> What board are you using?
>
> -Scott
>

[U-Boot] Problems upgrading NAND

[Alemao]

> I'm guessing that 1.2.0 was a BSP u-boot, that was not upstream.  I don't
> think upstream 1.2.0 had UPM NAND support.

You are right, there were no UPM support on U-boot-1.2.0.
I used patches sended by Anton from here:

http://www.nabble.com/-PATCH-v2-0-5--mpc83xx%3A-MPC8360E-RDK-related-patches-td14717360.html#a14717487

> Could you try the latest upstream U-boot?

If I use the latest upstream I will have to do a lot of validations.
If a patch nand core, also I will have to validate, but just nand
related things.

So if it just a detail in fsl upm driver, will save a lot of time.

> What board are you using?

Im using a MPC8343 board

[U-Boot] Problems upgrading NAND

[Anton Vorontsov]

On Fri, Jan 16, 2009 at 02:02:26PM -0200, Alemao wrote:
> > I'm guessing that 1.2.0 was a BSP u-boot, that was not upstream.  I don't
> > think upstream 1.2.0 had UPM NAND support.
> 
> You are right, there were no UPM support on U-boot-1.2.0.
> I used patches sended by Anton from here:
> 
> http://www.nabble.com/-PATCH-v2-0-5--mpc83xx%3A-MPC8360E-RDK-related-patches-td14717360.html#a14717487

There were known issues with this version (it only worked for the
small page NANDs, IIRC).

This is already fixed in the upstream U-Boot, via this patch set:

http://www.nabble.com/-PATCH-v3-0-2--NAND-FSL-UPM-driver-update-td17667335.html

Thanks,

-- 
Anton Vorontsov
email: cbouatmailru at gmail.com
irc://irc.freenode.net/bd2

[U-Boot] Problems upgrading NAND

[Alemao]

On Fri, Jan 16, 2009 at 8:46 PM, Anton Vorontsov
<avorontsov@ru.mvista.com> wrote:
> On Fri, Jan 16, 2009 at 02:02:26PM -0200, Alemao wrote:
>> > I'm guessing that 1.2.0 was a BSP u-boot, that was not upstream.  I don't
>> > think upstream 1.2.0 had UPM NAND support.
>>
>> You are right, there were no UPM support on U-boot-1.2.0.
>> I used patches sended by Anton from here:
>>
>> http://www.nabble.com/-PATCH-v2-0-5--mpc83xx%3A-MPC8360E-RDK-related-patches-td14717360.html#a14717487
>
> There were known issues with this version (it only worked for the
> small page NANDs, IIRC).
>
> This is already fixed in the upstream U-Boot, via this patch set:
>
> http://www.nabble.com/-PATCH-v3-0-2--NAND-FSL-UPM-driver-update-td17667335.html
>

It worked. Now here u-boot-1.2.0 supports NAND 512 and 1Gbit.

The only problem is that kernel is complaining about empty spaces on
nand flash. It seems that u-boot when writing to nand is not using all
page/block space:

[  131.220948] Empty flash at 0x00033ff8 ends at 0x00034000
[  131.309014] Empty flash at 0x00063f68 ends at 0x00064000
[  131.354234] Empty flash at 0x00093f68 ends at 0x00094000
...
[  139.266081] Empty flash at 0x0076bffc ends at 0x0076c000

With NAND 512 Gbits I usually got only one "Empty flash at..."

Is this a known issue? Or should I patch nand core? I didnt see
relevant changes in drivers/mtd/nand/* that could cause this behavior
(based on U-Boot-1.3.4, U-Boot from head differs a lot)

Thanks in advance,

--
Alemao

[U-Boot] Problems upgrading NAND

[Alemao]

On Tue, Jan 20, 2009 at 8:56 AM, Alemao <xcarandiru@gmail.com> wrote:
> On Fri, Jan 16, 2009 at 8:46 PM, Anton Vorontsov
> <avorontsov@ru.mvista.com> wrote:
>> On Fri, Jan 16, 2009 at 02:02:26PM -0200, Alemao wrote:
>>> > I'm guessing that 1.2.0 was a BSP u-boot, that was not upstream.  I don't
>>> > think upstream 1.2.0 had UPM NAND support.
>>>
>>> You are right, there were no UPM support on U-boot-1.2.0.
>>> I used patches sended by Anton from here:
>>>
>>> http://www.nabble.com/-PATCH-v2-0-5--mpc83xx%3A-MPC8360E-RDK-related-patches-td14717360.html#a14717487
>>
>> There were known issues with this version (it only worked for the
>> small page NANDs, IIRC).
>>
>> This is already fixed in the upstream U-Boot, via this patch set:
>>
>> http://www.nabble.com/-PATCH-v3-0-2--NAND-FSL-UPM-driver-update-td17667335.html
>>
>
> It worked. Now here u-boot-1.2.0 supports NAND 512 and 1Gbit.
>
> The only problem is that kernel is complaining about empty spaces on
> nand flash. It seems that u-boot when writing to nand is not using all
> page/block space:
>
> [  131.220948] Empty flash at 0x00033ff8 ends at 0x00034000
> [  131.309014] Empty flash at 0x00063f68 ends at 0x00064000
> [  131.354234] Empty flash at 0x00093f68 ends at 0x00094000
> ...
> [  139.266081] Empty flash at 0x0076bffc ends at 0x0076c000
>
> With NAND 512 Gbits I usually got only one "Empty flash at..."
>
> Is this a known issue? Or should I patch nand core? I didnt see
> relevant changes in drivers/mtd/nand/* that could cause this behavior
> (based on U-Boot-1.3.4, U-Boot from head differs a lot)
>
> Thanks in advance,
>
> --
> Alemao
>

Sorry, I didnt change the eraseblock size in mkfs.jffs2 command.

My apologies,

--
Alemao