[okl4-developer] okl4 on arm1136jfs goes haywire [fixed]

[Ashish Bijlani]

Hi Geoffrey,

I was just about to draft a mail. I've written decompression routines
to decompress the okl4 kernel on the fly. The decompression routines
are loaded at addr 0x8000 0000, that's why I was loading the kernel at
0x8000 8000. I've now moved the decompression to routines to upper
half of the ram so that I can load the kernel at 0x8000 0000 (the
start of the ram, which is 1MB aligned). This works fine for me.
Thanks much!!

-Ashish

On Tue, Jul 22, 2008 at 12:28 AM, Geoffrey Lee <glee at ok-labs.com> wrote:
> On Mon, Jul 21, 2008 at 11:54:31PM -0500, Ashish Bijlani wrote:
>> Hi Geoffrey,
>
> Hi Ashish
>
>>
>> I tried loading the kernel to 0x8010 0000 (this is the start of the
>> ram in machines.py file under the platform) which now i believe maps
>> to f000 0000, still I observe the same behavior? Am I missing
>> something? What should be the load addr in RAM if it starts at 0x8000
>> 0000?
>
> Just as a safety measure, can you print the output of
> arm-linux-objdump -x images/image.boot to ensure it is indeed loading
> at the correct address of 0x80100000 (Look at the LOAD sections
> and the entry point)?
>
> Can you also let us know whether you have made any modifications to
> OKL4 apart from adding in a machine description file for your
> platform?
>
>
>
>>
>> Thanks,
>> Ashish
>
>        -gl
>