Patchwork [U-Boot] Tricorder: Fix boot after making software ecc configurable

login
register
mail settings
Submitter Thomas Weber
Date Feb. 13, 2012, 8:56 a.m.
Message ID <1329123412-16339-1-git-send-email-weber@corscience.de>
Download mbox | patch
Permalink /patch/140872/
State Deferred
Delegated to: Scott Wood
Headers show

Comments

Thomas Weber - Feb. 13, 2012, 8:56 a.m.
The tricorder stops booting with the following message:

U-Boot 2011.12-00325-ga661206 (Feb 13 2012 - 09:29:50)

OMAP3503-GP ES3.1, CPU-OPP2, L3-165MHz, Max CPU Clock 600 mHz
OMAP3 Tricorder + LPDDR/NAND
I2C:   ready
DRAM:  128 MiB
NAND:  CONFIG_MTD_ECC_SOFT not enabled
0 MiB
MMC:   OMAP SD/MMC: 0

Adding CONFIG_MTD_ECC_SOFT to config file fixes this problem.

Signed-off-by: Thomas Weber <weber@corscience.de>
---
 include/configs/tricorder.h |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
Scott Wood - Feb. 13, 2012, 9:44 p.m.
On 02/13/2012 02:56 AM, Thomas Weber wrote:
> The tricorder stops booting with the following message:
> 
> U-Boot 2011.12-00325-ga661206 (Feb 13 2012 - 09:29:50)
> 
> OMAP3503-GP ES3.1, CPU-OPP2, L3-165MHz, Max CPU Clock 600 mHz
> OMAP3 Tricorder + LPDDR/NAND
> I2C:   ready
> DRAM:  128 MiB
> NAND:  CONFIG_MTD_ECC_SOFT not enabled
> 0 MiB
> MMC:   OMAP SD/MMC: 0
> 
> Adding CONFIG_MTD_ECC_SOFT to config file fixes this problem.
> 
> Signed-off-by: Thomas Weber <weber@corscience.de>
> ---
>  include/configs/tricorder.h |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)

Sorry about that -- I meant to exclude "nand: make 1-bit software ECC
configurable" from the patchset I applied, due to the potential for this
sort of breakage.

We should revert that patch until the issues raised here are resolved:
http://patchwork.ozlabs.org/patch/119149/

-Scott
Tom Rini - Feb. 13, 2012, 10:14 p.m.
On Mon, Feb 13, 2012 at 2:44 PM, Scott Wood <scottwood@freescale.com> wrote:
> On 02/13/2012 02:56 AM, Thomas Weber wrote:
>> The tricorder stops booting with the following message:
>>
>> U-Boot 2011.12-00325-ga661206 (Feb 13 2012 - 09:29:50)
>>
>> OMAP3503-GP ES3.1, CPU-OPP2, L3-165MHz, Max CPU Clock 600 mHz
>> OMAP3 Tricorder + LPDDR/NAND
>> I2C:   ready
>> DRAM:  128 MiB
>> NAND:  CONFIG_MTD_ECC_SOFT not enabled
>> 0 MiB
>> MMC:   OMAP SD/MMC: 0
>>
>> Adding CONFIG_MTD_ECC_SOFT to config file fixes this problem.
>>
>> Signed-off-by: Thomas Weber <weber@corscience.de>
>> ---
>>  include/configs/tricorder.h |    1 +
>>  1 files changed, 1 insertions(+), 0 deletions(-)
>
> Sorry about that -- I meant to exclude "nand: make 1-bit software ECC
> configurable" from the patchset I applied, due to the potential for this
> sort of breakage.
>
> We should revert that patch until the issues raised here are resolved:
> http://patchwork.ozlabs.org/patch/119149/

I had / have plan to take the tricorder patch tomorrow, unless you
really want to push a revert from your tree instead.  Either way is
fine with me.
Scott Wood - Feb. 13, 2012, 11:05 p.m.
On 02/13/2012 04:14 PM, Tom Rini wrote:
> On Mon, Feb 13, 2012 at 2:44 PM, Scott Wood <scottwood@freescale.com> wrote:
>> On 02/13/2012 02:56 AM, Thomas Weber wrote:
>>> The tricorder stops booting with the following message:
>>>
>>> U-Boot 2011.12-00325-ga661206 (Feb 13 2012 - 09:29:50)
>>>
>>> OMAP3503-GP ES3.1, CPU-OPP2, L3-165MHz, Max CPU Clock 600 mHz
>>> OMAP3 Tricorder + LPDDR/NAND
>>> I2C:   ready
>>> DRAM:  128 MiB
>>> NAND:  CONFIG_MTD_ECC_SOFT not enabled
>>> 0 MiB
>>> MMC:   OMAP SD/MMC: 0
>>>
>>> Adding CONFIG_MTD_ECC_SOFT to config file fixes this problem.
>>>
>>> Signed-off-by: Thomas Weber <weber@corscience.de>
>>> ---
>>>  include/configs/tricorder.h |    1 +
>>>  1 files changed, 1 insertions(+), 0 deletions(-)
>>
>> Sorry about that -- I meant to exclude "nand: make 1-bit software ECC
>> configurable" from the patchset I applied, due to the potential for this
>> sort of breakage.
>>
>> We should revert that patch until the issues raised here are resolved:
>> http://patchwork.ozlabs.org/patch/119149/
> 
> I had / have plan to take the tricorder patch tomorrow, unless you
> really want to push a revert from your tree instead.  Either way is
> fine with me.

This affects more than just tricorder -- any board that depends on soft
ECC will be broken, since none of them define this symbol.  I thought I
had left this patch out, but apparently didn't.

-Scott

Patch

diff --git a/include/configs/tricorder.h b/include/configs/tricorder.h
index f87696b..0e8003f 100644
--- a/include/configs/tricorder.h
+++ b/include/configs/tricorder.h
@@ -110,6 +110,7 @@ 
 /* Board NAND Info */
 #define CONFIG_SYS_NO_FLASH		/* no NOR flash */
 #define CONFIG_MTD_DEVICE		/* needed for mtdparts commands */
+#define CONFIG_MTD_ECC_SOFT
 #define MTDIDS_DEFAULT			"nand0=nand"
 #define MTDPARTS_DEFAULT		"mtdparts=nand:" \
 						"512k(u-boot-spl)," \