Latest andy-tracking , bad crc error

Nelson Castillo nelsoneci at gmail.com
Sat Feb 28 16:39:31 CET 2009


On Sat, Feb 28, 2009 at 9:57 PM, Daniel.Li <lida_mail at 163.com> wrote:
> On Sat, 2009-02-28 at 11:01 +0100, Petr Vanek wrote:
>> DL> > i think you have to either use qi or adjust uboot_env to be able
>> DL> > to load kernel image larger then 2M.
>> DL> >
>> DL> > iirc boot_env could also be flashed if there was an image online -
>> DL> > perhaps somebody uploaded it already somewhere?
>> DL>
>> DL> Thanks pointing me to the right direction.
>> DL> I think I got it.
>>
>> it's not u-boot but it's parameters.
>>
>> here is how to fix it (i did't write this, just copy/paste). i just
>> tested it with success:
>>
>>
>> - bring up the boot menu of NAND u-boot
>> - switch the console to USB
>> - enter the u-boot command line with
>>
>>   neocon /dev/ttyACM{0,1,2}
>>
>>   (or use the serial communication program of your choice)
>>
>> - enter the following two commands:
>>
>>   setenv bootcmd setenv bootargs \${bootargs_base} \${mtdparts}\; nand
>>   read.e 0x32000000 kernel 0x300000\; bootm 0x32000000
>>   saveenv
>>
>> - you can now boot with
>>
>>   boot
>>
>> The change is persistent, so you'll be able to boot large kernels
>> from now on. A small drawback of this change is that loading the
>> kernel now takes a little bit longer.
>>
>> See also
>> http://lists.openmoko.org/pipermail/openmoko-kernel/2008-October/006086.html
>>
>
> Humm,,, Excellent, I have followed above instructions, and it works.
> The only difference is the environment, see below:
>
> Old u-boot environment "bootcmd":
> bootcmd=setenv bootargs ${bootargs_base} ${mtdparts}; nand read.e
> 0x32000000 kernel 0x200000; bootm 0x32000000
>
> New u-boot environment "bootcmd":
> bootcmd=setenv bootargs ${bootargs_base} ${mtdparts}; nand read.e
> 0x32000000 kernel 0x300000; bootm 0x32000000

It would be very nice if somebody contributed this to the wiki FAQ :-)

This has been a FAQ for a while.

A link to an email that explains what needs to be changed would do
(also mentioning that this is not an issue with Qi).



More information about the openmoko-kernel mailing list