[U-Boot] OneNAND: Continuous locking scheme?

[U-Boot] OneNAND: Continuous locking scheme?

[Stefan Roese]

Hi Kyungmin,

I'm experiencing a problem with one OneNAND device on our board:

KFG1216U2B (512Mbit)

Here the output from the U-Boot bootup:

OneNAND 64MB 2.65/3.3V 16-bit (0x25)
OneNAND version = 0x022e
Lock scheme is Continuous Lock
Scanning device for bad blocks
OneNAND: 64 MB

As you can see the device is reported to have a "Continuous Lock" scheme. But 
this doesn't work. When I try to erase a block I get the following error:

OneNAND erase: offset 0x0, size 0x100000
onenand_wait: controller error = 0x4c00
onenand_wait: it's locked error = 0x4c00
onenand_erase: Failed erase, block 1
erase failed block 1 at 0x20000

When I disable this continuous locking check in onenand_base.c and all blocks 
are unlocked individually then everything is ok. So I suspect that either the 
implementation of the continuous locking scheme is incorrect or the device is 
not really supporting this scheme. I'm still new to OneNAND and I couldn't 
find any reference to this continuous locking scheme in the OneNAND chip used 
here. Do you have any idea why this is not working for us? Do you know for 
sure if the KFG1216U2B (Dev-ID 0x25) really supports this continuous locking 
scheme?

Thanks.

BTW: This is with your latest OneNAND patches applied (Sync with 2.6.27).

Best regards,
Stefan

=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office at denx.de
=====================================================================

[U-Boot] OneNAND: Continuous locking scheme?

[Kyungmin Park]

Hi,

On Tue, Nov 11, 2008 at 5:43 PM, Stefan Roese <sr@denx.de> wrote:
> Hi Kyungmin,
>
> I'm experiencing a problem with one OneNAND device on our board:
>
> KFG1216U2B (512Mbit)
>
> Here the output from the U-Boot bootup:
>
> OneNAND 64MB 2.65/3.3V 16-bit (0x25)
> OneNAND version = 0x022e
> Lock scheme is Continuous Lock
> Scanning device for bad blocks
> OneNAND: 64 MB

It's wrong report, The latest chip use block lock scheme.

>
> As you can see the device is reported to have a "Continuous Lock" scheme. But
> this doesn't work. When I try to erase a block I get the following error:
>
> OneNAND erase: offset 0x0, size 0x100000
> onenand_wait: controller error = 0x4c00
> onenand_wait: it's locked error = 0x4c00
> onenand_erase: Failed erase, block 1
> erase failed block 1 at 0x20000
>
> When I disable this continuous locking check in onenand_base.c and all blocks
> are unlocked individually then everything is ok. So I suspect that either the
> implementation of the continuous locking scheme is incorrect or the device is
> not really supporting this scheme.

Actually device doesn't support continuous lock scheme. The continuous
lock scheme is used early OneNAND chip.

 I'm still new to OneNAND and I couldn't
> find any reference to this continuous locking scheme in the OneNAND chip used
> here. Do you have any idea why this is not working for us? Do you know for
> sure if the KFG1216U2B (Dev-ID 0x25) really supports this continuous locking
> scheme?

In onenand_check_features()

          default:
                /* Some OneNAND has continuous lock scheme */
                if (!process)
                        this->options |= ONENAND_HAS_CONT_LOCK;

It's some strange. As your version id is 0x22e. then process is 0x2.
umm maybe options field is not cleared before. please check this one.

Thank you,
Kyungmin Park

[U-Boot] OneNAND: Continuous locking scheme?

[Stefan Roese]

On Tuesday 11 November 2008, Kyungmin Park wrote:
> > I'm still new to OneNAND and I couldn't
> > find any reference to this continuous locking scheme in the OneNAND chip
> > used here. Do you have any idea why this is not working for us? Do you
> > know for sure if the KFG1216U2B (Dev-ID 0x25) really supports this
> > continuous locking scheme?
>
> In onenand_check_features()
>
>           default:
>                 /* Some OneNAND has continuous lock scheme */
>                 if (!process)
>                         this->options |= ONENAND_HAS_CONT_LOCK;
>
> It's some strange. As your version id is 0x22e. then process is 0x2.
> umm maybe options field is not cleared before. please check this one.

No, that's not the problem. The problem is that version_id was never saved in 
the onenand_chip structure. This is missing:

@@ -1972,6 +1972,7 @@ static int onenand_probe(struct mtd_info *mtd)
        /* Flash device information */
        mtd->name = onenand_print_device_info(dev_id, ver_id);
        this->device_id = dev_id;
+       this->version_id = ver_id;

I'll submit a patch in a short while to fix this.

Thanks.

Best regards,
Stefan

=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office at denx.de
=====================================================================